5 Changes between 0.9.8zb and 0.9.8zc [xx XXX xxxx]
7 *) Add support for TLS_FALLBACK_SCSV.
8 Client applications doing fallback retries should call
9 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
11 [Adam Langley, Bodo Moeller]
13 *) Add additional DigestInfo checks.
15 Reencode DigestInto in DER and check against the original when
16 verifying RSA signature: this will reject any improperly encoded
17 DigestInfo structures.
19 Note: this is a precautionary measure and no attacks are currently known.
23 Changes between 0.9.8za and 0.9.8zb [6 Aug 2014]
25 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
26 to a denial of service attack. A malicious server can crash the client
27 with a null pointer dereference (read) by specifying an anonymous (EC)DH
28 ciphersuite and sending carefully crafted handshake messages.
30 Thanks to Felix Gröbert (Google) for discovering and researching this
35 *) By sending carefully crafted DTLS packets an attacker could cause openssl
36 to leak memory. This can be exploited through a Denial of Service attack.
37 Thanks to Adam Langley for discovering and researching this issue.
41 *) An attacker can force openssl to consume large amounts of memory whilst
42 processing DTLS handshake messages. This can be exploited through a
43 Denial of Service attack.
44 Thanks to Adam Langley for discovering and researching this issue.
48 *) An attacker can force an error condition which causes openssl to crash
49 whilst processing DTLS packets due to memory being freed twice. This
50 can be exploited through a Denial of Service attack.
51 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
56 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
57 X509_name_oneline, X509_name_print_ex et al. to leak some information
58 from the stack. Applications may be affected if they echo pretty printing
59 output to the attacker.
61 Thanks to Ivan Fratric (Google) for discovering this issue.
63 [Emilia Käsper, and Steve Henson]
65 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
66 for corner cases. (Certain input points at infinity could lead to
67 bogus results, with non-infinity inputs mapped to infinity too.)
70 Changes between 0.9.8y and 0.9.8za [5 Jun 2014]
72 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
73 handshake can force the use of weak keying material in OpenSSL
74 SSL/TLS clients and servers.
76 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
77 researching this issue. (CVE-2014-0224)
78 [KIKUCHI Masashi, Steve Henson]
80 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
81 OpenSSL DTLS client the code can be made to recurse eventually crashing
84 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
86 [Imre Rad, Steve Henson]
88 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
89 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
90 client or server. This is potentially exploitable to run arbitrary
91 code on a vulnerable client or server.
93 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
94 [Jüri Aedla, Steve Henson]
96 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
97 are subject to a denial of service attack.
99 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
100 this issue. (CVE-2014-3470)
101 [Felix Gröbert, Ivan Fratric, Steve Henson]
103 *) Fix for the attack described in the paper "Recovering OpenSSL
104 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
105 by Yuval Yarom and Naomi Benger. Details can be obtained from:
106 http://eprint.iacr.org/2014/140
108 Thanks to Yuval Yarom and Naomi Benger for discovering this
109 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
110 [Yuval Yarom and Naomi Benger]
112 Thanks to mancha for backporting the fix to the 0.9.8 branch.
114 *) Fix handling of warning-level alerts in SSL23 client mode so they
115 don't cause client-side termination (eg. on SNI unrecognized_name
116 warnings). Add client and server support for six additional alerts
117 per RFC 6066 and RFC 4279.
120 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
121 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
122 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
123 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
124 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
125 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
126 [Rob Stradling, Adam Langley]
128 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
130 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
132 This addresses the flaw in CBC record processing discovered by
133 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
134 at: http://www.isg.rhul.ac.uk/tls/
136 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
137 Security Group at Royal Holloway, University of London
138 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
139 Emilia Käsper for the initial patch.
141 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
143 *) Return an error when checking OCSP signatures when key is NULL.
144 This fixes a DoS attack. (CVE-2013-0166)
147 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
148 the right response is stapled. Also change SSL_get_certificate()
149 so it returns the certificate actually sent.
150 See http://rt.openssl.org/Ticket/Display.html?id=2836.
152 [Rob Stradling <rob.stradling@comodo.com>]
154 *) Fix possible deadlock when decoding public keys.
157 Changes between 0.9.8w and 0.9.8x [10 May 2012]
159 *) Sanity check record length before skipping explicit IV in DTLS
162 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
163 fuzzing as a service testing platform.
167 *) Initialise tkeylen properly when encrypting CMS messages.
168 Thanks to Solar Designer of Openwall for reporting this issue.
171 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
173 *) The fix for CVE-2012-2110 did not take into account that the
174 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
175 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
176 rejecting negative len parameter. (CVE-2012-2131)
177 [Tomas Hoger <thoger@redhat.com>]
179 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
181 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
182 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
183 in CRYPTO_realloc_clean.
185 Thanks to Tavis Ormandy, Google Security Team, for discovering this
186 issue and to Adam Langley <agl@chromium.org> for fixing it.
188 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
190 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
192 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
193 in CMS and PKCS7 code. When RSA decryption fails use a random key for
194 content decryption and always return the same error. Note: this attack
195 needs on average 2^20 messages so it only affects automated senders. The
196 old behaviour can be reenabled in the CMS code by setting the
197 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
198 an MMA defence is not necessary.
199 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
200 this issue. (CVE-2012-0884)
203 *) Fix CVE-2011-4619: make sure we really are receiving a
204 client hello before rejecting multiple SGC restarts. Thanks to
205 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
208 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
210 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
211 Thanks to Antonio Martin, Enterprise Secure Access Research and
212 Development, Cisco Systems, Inc. for discovering this bug and
213 preparing a fix. (CVE-2012-0050)
216 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
218 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
219 of the Vaudenay padding oracle attack on CBC mode encryption
220 which enables an efficient plaintext recovery attack against
221 the OpenSSL implementation of DTLS. Their attack exploits timing
222 differences arising during decryption processing. A research
223 paper describing this attack can be found at:
224 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
225 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
226 Security Group at Royal Holloway, University of London
227 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
228 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
229 for preparing the fix. (CVE-2011-4108)
230 [Robin Seggelmann, Michael Tuexen]
232 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
233 [Ben Laurie, Kasper <ekasper@google.com>]
235 *) Clear bytes used for block padding of SSL 3.0 records.
237 [Adam Langley (Google)]
239 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
240 Kadianakis <desnacked@gmail.com> for discovering this issue and
241 Adam Langley for preparing the fix. (CVE-2011-4619)
242 [Adam Langley (Google)]
244 *) Prevent malformed RFC3779 data triggering an assertion failure.
245 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
246 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
247 [Rob Austein <sra@hactrn.net>]
249 *) Fix ssl_ciph.c set-up race.
250 [Adam Langley (Google)]
252 *) Fix spurious failures in ecdsatest.c.
253 [Emilia Käsper (Google)]
255 *) Fix the BIO_f_buffer() implementation (which was mixing different
256 interpretations of the '..._len' fields).
257 [Adam Langley (Google)]
259 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
260 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
261 threads won't reuse the same blinding coefficients.
263 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
264 lock to call BN_BLINDING_invert_ex, and avoids one use of
265 BN_BLINDING_update for each BN_BLINDING structure (previously,
266 the last update always remained unused).
267 [Emilia Käsper (Google)]
269 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
270 for multi-threaded use of ECDH.
271 [Adam Langley (Google)]
273 *) Fix x509_name_ex_d2i memory leak on bad inputs.
276 *) Add protection against ECDSA timing attacks as mentioned in the paper
277 by Billy Bob Brumley and Nicola Tuveri, see:
279 http://eprint.iacr.org/2011/232.pdf
281 [Billy Bob Brumley and Nicola Tuveri]
283 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
285 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
286 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
288 *) Fix bug in string printing code: if *any* escaping is enabled we must
289 escape the escape character (backslash) or the resulting string is
293 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
295 *) Disable code workaround for ancient and obsolete Netscape browsers
296 and servers: an attacker can use it in a ciphersuite downgrade attack.
297 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
300 *) Fixed J-PAKE implementation error, originally discovered by
301 Sebastien Martini, further info and confirmation from Stefan
302 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
305 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
307 *) Fix extension code to avoid race conditions which can result in a buffer
308 overrun vulnerability: resumed sessions must not be modified as they can
309 be shared by multiple threads. CVE-2010-3864
312 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
315 *) Don't reencode certificate when calculating signature: cache and use
316 the original encoding instead. This makes signature verification of
317 some broken encodings work correctly.
320 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
321 is also one of the inputs.
322 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
324 *) Don't repeatedly append PBE algorithms to table if they already exist.
325 Sort table on each new add. This effectively makes the table read only
326 after all algorithms are added and subsequent calls to PKCS12_pbe_add
330 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
332 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
335 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
336 access or freeing data twice (CVE-2010-0742)
337 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
339 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
340 common in certificates and some applications which only call
341 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
345 Reduce copying into .apps and .test in makevms.com
346 Don't try to use blank CA certificate in CA.com
347 Allow use of C files from original directories in maketests.com
348 [Steven M. Schweda" <sms@antinode.info>]
350 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
352 *) When rejecting SSL/TLS records due to an incorrect version number, never
353 update s->server with a new major version number. As of
354 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
355 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
356 the previous behavior could result in a read attempt at NULL when
357 receiving specific incorrect SSL/TLS records once record payload
358 protection is active. (CVE-2010-0740)
359 [Bodo Moeller, Adam Langley <agl@chromium.org>]
361 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
362 could be crashed if the relevant tables were not present (e.g. chrooted).
363 [Tomas Hoger <thoger@redhat.com>]
365 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
367 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
368 [Martin Olsson, Neel Mehta]
370 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
371 accommodate for stack sorting, always a write lock!).
374 *) On some versions of WIN32 Heap32Next is very slow. This can cause
375 excessive delays in the RAND_poll(): over a minute. As a workaround
376 include a time check in the inner Heap32Next loop too.
379 *) The code that handled flushing of data in SSL/TLS originally used the
380 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
381 the problem outlined in PR#1949. The fix suggested there however can
382 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
383 of Apache). So instead simplify the code to flush unconditionally.
384 This should be fine since flushing with no data to flush is a no op.
387 *) Handle TLS versions 2.0 and later properly and correctly use the
388 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
389 off ancient servers have a habit of sticking around for a while...
392 *) Modify compression code so it frees up structures without using the
393 ex_data callbacks. This works around a problem where some applications
394 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
395 restarting) then use compression (e.g. SSL with compression) later.
396 This results in significant per-connection memory leaks and
397 has caused some security issues including CVE-2008-1678 and
401 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
402 change when encrypting or decrypting.
405 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
406 connect and renegotiate with servers which do not support RI.
407 Until RI is more widely deployed this option is enabled by default.
410 *) Add "missing" ssl ctrls to clear options and mode.
413 *) If client attempts to renegotiate and doesn't support RI respond with
414 a no_renegotiation alert as required by RFC5746. Some renegotiating
415 TLS clients will continue a connection gracefully when they receive
416 the alert. Unfortunately OpenSSL mishandled this alert and would hang
417 waiting for a server hello which it will never receive. Now we treat a
418 received no_renegotiation alert as a fatal error. This is because
419 applications requesting a renegotiation might well expect it to succeed
420 and would have no code in place to handle the server denying it so the
421 only safe thing to do is to terminate the connection.
424 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
425 peer supports secure renegotiation and 0 otherwise. Print out peer
426 renegotiation support in s_client/s_server.
429 *) Replace the highly broken and deprecated SPKAC certification method with
430 the updated NID creation version. This should correctly handle UTF8.
433 *) Implement RFC5746. Re-enable renegotiation but require the extension
434 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
435 turns out to be a bad idea. It has been replaced by
436 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
437 SSL_CTX_set_options(). This is really not recommended unless you
438 know what you are doing.
439 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
441 *) Fixes to stateless session resumption handling. Use initial_ctx when
442 issuing and attempting to decrypt tickets in case it has changed during
443 servername handling. Use a non-zero length session ID when attempting
444 stateless session resumption: this makes it possible to determine if
445 a resumption has occurred immediately after receiving server hello
446 (several places in OpenSSL subtly assume this) instead of later in
450 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
451 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
452 fixes for a few places where the return code is not checked
454 [Julia Lawall <julia@diku.dk>]
456 *) Add --strict-warnings option to Configure script to include devteam
457 warnings in other configurations.
460 *) Add support for --libdir option and LIBDIR variable in makefiles. This
461 makes it possible to install openssl libraries in locations which
462 have names other than "lib", for example "/usr/lib64" which some
464 [Steve Henson, based on patch from Jeremy Utley]
466 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
467 X690 8.9.12 and can produce some misleading textual output of OIDs.
468 [Steve Henson, reported by Dan Kaminsky]
470 *) Delete MD2 from algorithm tables. This follows the recommendation in
471 several standards that it is not used in new applications due to
472 several cryptographic weaknesses. For binary compatibility reasons
473 the MD2 API is still compiled in by default.
476 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
480 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
481 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
483 [Guenter <lists@gknw.net>]
485 *) Fix the server certificate chain building code to use X509_verify_cert(),
486 it used to have an ad-hoc builder which was unable to cope with anything
487 other than a simple chain.
488 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
490 *) Don't check self signed certificate signatures in X509_verify_cert()
491 by default (a flag can override this): it just wastes time without
492 adding any security. As a useful side effect self signed root CAs
493 with non-FIPS digests are now usable in FIPS mode.
496 *) In dtls1_process_out_of_seq_message() the check if the current message
497 is already buffered was missing. For every new message was memory
498 allocated, allowing an attacker to perform an denial of service attack
499 with sending out of seq handshake messages until there is no memory
500 left. Additionally every future messege was buffered, even if the
501 sequence number made no sense and would be part of another handshake.
502 So only messages with sequence numbers less than 10 in advance will be
503 buffered. (CVE-2009-1378)
504 [Robin Seggelmann, discovered by Daniel Mentz]
506 *) Records are buffered if they arrive with a future epoch to be
507 processed after finishing the corresponding handshake. There is
508 currently no limitation to this buffer allowing an attacker to perform
509 a DOS attack with sending records with future epochs until there is no
510 memory left. This patch adds the pqueue_size() function to detemine
511 the size of a buffer and limits the record buffer to 100 entries.
513 [Robin Seggelmann, discovered by Daniel Mentz]
515 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
516 parent structure is freed. (CVE-2009-1379)
519 *) Handle non-blocking I/O properly in SSL_shutdown() call.
520 [Darryl Miles <darryl-mailinglists@netbauds.net>]
523 [Ilya O. <vrghost@gmail.com>]
525 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
527 *) Disable renegotiation completely - this fixes a severe security
528 problem (CVE-2009-3555) at the cost of breaking all
529 renegotiation. Renegotiation can be re-enabled by setting
530 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
531 run-time. This is really not recommended unless you know what
535 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
537 *) Don't set val to NULL when freeing up structures, it is freed up by
538 underlying code. If sizeof(void *) > sizeof(long) this can result in
539 zeroing past the valid field. (CVE-2009-0789)
540 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
542 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
543 checked correctly. This would allow some invalid signed attributes to
544 appear to verify correctly. (CVE-2009-0591)
545 [Ivan Nestlerode <inestlerode@us.ibm.com>]
547 *) Reject UniversalString and BMPString types with invalid lengths. This
548 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
549 a legal length. (CVE-2009-0590)
552 *) Set S/MIME signing as the default purpose rather than setting it
553 unconditionally. This allows applications to override it at the store
557 *) Permit restricted recursion of ASN1 strings. This is needed in practice
558 to handle some structures.
561 *) Improve efficiency of mem_gets: don't search whole buffer each time
563 [Jeremy Shapiro <jnshapir@us.ibm.com>]
565 *) New -hex option for openssl rand.
568 *) Print out UTF8String and NumericString when parsing ASN1.
571 *) Support NumericString type for name components.
574 *) Allow CC in the environment to override the automatically chosen
575 compiler. Note that nothing is done to ensure flags work with the
579 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
581 *) Properly check EVP_VerifyFinal() and similar return values
583 [Ben Laurie, Bodo Moeller, Google Security Team]
585 *) Enable TLS extensions by default.
588 *) Allow the CHIL engine to be loaded, whether the application is
589 multithreaded or not. (This does not release the developer from the
590 obligation to set up the dynamic locking callbacks.)
591 [Sander Temme <sander@temme.net>]
593 *) Use correct exit code if there is an error in dgst command.
594 [Steve Henson; problem pointed out by Roland Dirlewanger]
596 *) Tweak Configure so that you need to say "experimental-jpake" to enable
597 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
600 *) Add experimental JPAKE support, including demo authentication in
601 s_client and s_server.
604 *) Set the comparison function in v3_addr_canonize().
605 [Rob Austein <sra@hactrn.net>]
607 *) Add support for XMPP STARTTLS in s_client.
608 [Philip Paeps <philip@freebsd.org>]
610 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
611 to ensure that even with this option, only ciphersuites in the
612 server's preference list will be accepted. (Note that the option
613 applies only when resuming a session, so the earlier behavior was
614 just about the algorithm choice for symmetric cryptography.)
617 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
619 *) Fix NULL pointer dereference if a DTLS server received
620 ChangeCipherSpec as first record (CVE-2009-1386).
623 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
624 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
627 *) The fix in 0.9.8c that supposedly got rid of unsafe
628 double-checked locking was incomplete for RSA blinding,
629 addressing just one layer of what turns out to have been
630 doubly unsafe triple-checked locking.
632 So now fix this for real by retiring the MONT_HELPER macro
633 in crypto/rsa/rsa_eay.c.
635 [Bodo Moeller; problem pointed out by Marius Schilder]
637 *) Various precautionary measures:
639 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
641 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
642 (NB: This would require knowledge of the secret session ticket key
643 to exploit, in which case you'd be SOL either way.)
645 - Change bn_nist.c so that it will properly handle input BIGNUMs
646 outside the expected range.
648 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
651 [Neel Mehta, Bodo Moeller]
653 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
654 the load fails. Useful for distros.
655 [Ben Laurie and the FreeBSD team]
657 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
660 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
663 *) Expand ENGINE to support engine supplied SSL client certificate functions.
665 This work was sponsored by Logica.
668 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
669 keystores. Support for SSL/TLS client authentication too.
670 Not compiled unless enable-capieng specified to Configure.
672 This work was sponsored by Logica.
675 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
676 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
677 attribute creation routines such as certifcate requests and PKCS#12
681 Changes between 0.9.8g and 0.9.8h [28 May 2008]
683 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
684 handshake which could lead to a cilent crash as found using the
685 Codenomicon TLS test suite (CVE-2008-1672)
686 [Steve Henson, Mark Cox]
688 *) Fix double free in TLS server name extensions which could lead to
689 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
692 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
694 Clear the error queue to ensure that error entries left from
695 older function calls do not interfere with the correct operation.
696 [Lutz Jaenicke, Erik de Castro Lopo]
698 *) Remove root CA certificates of commercial CAs:
700 The OpenSSL project does not recommend any specific CA and does not
701 have any policy with respect to including or excluding any CA.
702 Therefore it does not make any sense to ship an arbitrary selection
703 of root CA certificates with the OpenSSL software.
706 *) RSA OAEP patches to fix two separate invalid memory reads.
707 The first one involves inputs when 'lzero' is greater than
708 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
709 before the beginning of from). The second one involves inputs where
710 the 'db' section contains nothing but zeroes (there is a one-byte
711 invalid read after the end of 'db').
712 [Ivan Nestlerode <inestlerode@us.ibm.com>]
714 *) Partial backport from 0.9.9-dev:
716 Introduce bn_mul_mont (dedicated Montgomery multiplication
717 procedure) as a candidate for BIGNUM assembler implementation.
718 While 0.9.9-dev uses assembler for various architectures, only
719 x86_64 is available by default here in the 0.9.8 branch, and
720 32-bit x86 is available through a compile-time setting.
722 To try the 32-bit x86 assembler implementation, use Configure
723 option "enable-montasm" (which exists only for this backport).
725 As "enable-montasm" for 32-bit x86 disclaims code stability
726 anyway, in this constellation we activate additional code
727 backported from 0.9.9-dev for further performance improvements,
728 namely BN_from_montgomery_word. (To enable this otherwise,
729 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
731 [Andy Polyakov (backport partially by Bodo Moeller)]
733 *) Add TLS session ticket callback. This allows an application to set
734 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
735 values. This is useful for key rollover for example where several key
736 sets may exist with different names.
739 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
740 This was broken until now in 0.9.8 releases, such that the only way
741 a registered ENGINE could be used (assuming it initialises
742 successfully on the host) was to explicitly set it as the default
743 for the relevant algorithms. This is in contradiction with 0.9.7
744 behaviour and the documentation. With this fix, when an ENGINE is
745 registered into a given algorithm's table of implementations, the
746 'uptodate' flag is reset so that auto-discovery will be used next
747 time a new context for that algorithm attempts to select an
749 [Ian Lister (tweaked by Geoff Thorpe)]
751 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
752 implemention in the following ways:
754 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
757 Lack of BER streaming support means one pass streaming processing is
758 only supported if data is detached: setting the streaming flag is
759 ignored for embedded content.
761 CMS support is disabled by default and must be explicitly enabled
762 with the enable-cms configuration option.
765 *) Update the GMP engine glue to do direct copies between BIGNUM and
766 mpz_t when openssl and GMP use the same limb size. Otherwise the
767 existing "conversion via a text string export" trick is still used.
768 [Paul Sheer <paulsheer@gmail.com>]
770 *) Zlib compression BIO. This is a filter BIO which compressed and
771 uncompresses any data passed through it.
774 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
775 RFC3394 compatible AES key wrapping.
778 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
779 sets string data without copying. X509_ALGOR_set0() and
780 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
781 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
782 from an X509_ATTRIBUTE structure optionally checking it occurs only
783 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
787 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
788 to get the expected BN_FLG_CONSTTIME behavior.
789 [Bodo Moeller (Google)]
793 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
794 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
795 - added some more tests to do_tests.pl
796 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
797 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
798 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
799 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
800 - various changes to netware.pl to enable gcc-cross builds on Win32
802 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
803 - various changes to fix missing prototype warnings
804 - fixed x86nasm.pl to create correct asm files for NASM COFF output
805 - added AES, WHIRLPOOL and CPUID assembler code to build files
806 - added missing AES assembler make rules to mk1mf.pl
807 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
808 [Guenter Knauf <eflash@gmx.net>]
810 *) Implement certificate status request TLS extension defined in RFC3546.
811 A client can set the appropriate parameters and receive the encoded
812 OCSP response via a callback. A server can query the supplied parameters
813 and set the encoded OCSP response in the callback. Add simplified examples
814 to s_client and s_server.
817 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
820 + Binary incompatibility of ssl_ctx_st structure
821 + DTLS interoperation with non-compliant servers
822 + Don't call get_session_cb() without proposed session
823 + Fix ia64 assembler code
824 [Andy Polyakov, Steve Henson]
826 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
828 *) DTLS Handshake overhaul. There were longstanding issues with
829 OpenSSL DTLS implementation, which were making it impossible for
830 RFC 4347 compliant client to communicate with OpenSSL server.
831 Unfortunately just fixing these incompatibilities would "cut off"
832 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
833 server keeps tolerating non RFC compliant syntax. The opposite is
834 not true, 0.9.8f client can not communicate with earlier server.
835 This update even addresses CVE-2007-4995.
838 *) Changes to avoid need for function casts in OpenSSL: some compilers
839 (gcc 4.2 and later) reject their use.
840 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
843 *) Add RFC4507 support to OpenSSL. This includes the corrections in
844 RFC4507bis. The encrypted ticket format is an encrypted encoded
845 SSL_SESSION structure, that way new session features are automatically
848 If a client application caches session in an SSL_SESSION structure
849 support is transparent because tickets are now stored in the encoded
852 The SSL_CTX structure automatically generates keys for ticket
853 protection in servers so again support should be possible
854 with no application modification.
856 If a client or server wishes to disable RFC4507 support then the option
857 SSL_OP_NO_TICKET can be set.
859 Add a TLS extension debugging callback to allow the contents of any client
860 or server extensions to be examined.
862 This work was sponsored by Google.
865 *) Add initial support for TLS extensions, specifically for the server_name
866 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
867 have new members for a host name. The SSL data structure has an
868 additional member SSL_CTX *initial_ctx so that new sessions can be
869 stored in that context to allow for session resumption, even after the
870 SSL has been switched to a new SSL_CTX in reaction to a client's
871 server_name extension.
873 New functions (subject to change):
876 SSL_get_servername_type()
879 New CTRL codes and macros (subject to change):
881 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
882 - SSL_CTX_set_tlsext_servername_callback()
883 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
884 - SSL_CTX_set_tlsext_servername_arg()
885 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
887 openssl s_client has a new '-servername ...' option.
889 openssl s_server has new options '-servername_host ...', '-cert2 ...',
890 '-key2 ...', '-servername_fatal' (subject to change). This allows
891 testing the HostName extension for a specific single host name ('-cert'
892 and '-key' remain fallbacks for handshakes without HostName
893 negotiation). If the unrecogninzed_name alert has to be sent, this by
894 default is a warning; it becomes fatal with the '-servername_fatal'
897 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
899 *) Add AES and SSE2 assembly language support to VC++ build.
902 *) Mitigate attack on final subtraction in Montgomery reduction.
905 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
906 (which previously caused an internal error).
909 *) Squeeze another 10% out of IGE mode when in != out.
912 *) AES IGE mode speedup.
913 [Dean Gaudet (Google)]
915 *) Add the Korean symmetric 128-bit cipher SEED (see
916 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
917 add SEED ciphersuites from RFC 4162:
919 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
920 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
921 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
922 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
924 To minimize changes between patchlevels in the OpenSSL 0.9.8
925 series, SEED remains excluded from compilation unless OpenSSL
926 is configured with 'enable-seed'.
929 *) Mitigate branch prediction attacks, which can be practical if a
930 single processor is shared, allowing a spy process to extract
931 information. For detailed background information, see
932 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
933 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
934 and Necessary Software Countermeasures"). The core of the change
935 are new versions BN_div_no_branch() and
936 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
937 respectively, which are slower, but avoid the security-relevant
938 conditional branches. These are automatically called by BN_div()
939 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
940 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
941 remove a conditional branch.
943 BN_FLG_CONSTTIME is the new name for the previous
944 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
945 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
946 in the exponent causes BN_mod_exp_mont() to use the alternative
947 implementation in BN_mod_exp_mont_consttime().) The old name
948 remains as a deprecated alias.
950 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
951 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
952 constant-time implementations for more than just exponentiation.
953 Here too the old name is kept as a deprecated alias.
955 BN_BLINDING_new() will now use BN_dup() for the modulus so that
956 the BN_BLINDING structure gets an independent copy of the
957 modulus. This means that the previous "BIGNUM *m" argument to
958 BN_BLINDING_new() and to BN_BLINDING_create_param() now
959 essentially becomes "const BIGNUM *m", although we can't actually
960 change this in the header file before 0.9.9. It allows
961 RSA_setup_blinding() to use BN_with_flags() on the modulus to
962 enable BN_FLG_CONSTTIME.
964 [Matthew D Wood (Intel Corp)]
966 *) In the SSL/TLS server implementation, be strict about session ID
967 context matching (which matters if an application uses a single
968 external cache for different purposes). Previously,
969 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
970 set. This did ensure strict client verification, but meant that,
971 with applications using a single external cache for quite
972 different requirements, clients could circumvent ciphersuite
973 restrictions for a given session ID context by starting a session
974 in a different context.
977 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
978 a ciphersuite string such as "DEFAULT:RSA" cannot enable
979 authentication-only ciphersuites.
982 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
983 not complete and could lead to a possible single byte overflow
984 (CVE-2007-5135) [Ben Laurie]
986 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
988 *) Since AES128 and AES256 (and similarly Camellia128 and
989 Camellia256) share a single mask bit in the logic of
990 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
991 kludge to work properly if AES128 is available and AES256 isn't
992 (or if Camellia128 is available and Camellia256 isn't).
995 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
996 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
997 When a point or a seed is encoded in a BIT STRING, we need to
998 prevent the removal of trailing zero bits to get the proper DER
999 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1000 of a NamedBitList, for which trailing 0 bits need to be removed.)
1003 *) Have SSL/TLS server implementation tolerate "mismatched" record
1004 protocol version while receiving ClientHello even if the
1005 ClientHello is fragmented. (The server can't insist on the
1006 particular protocol version it has chosen before the ServerHello
1007 message has informed the client about his choice.)
1010 *) Add RFC 3779 support.
1011 [Rob Austein for ARIN, Ben Laurie]
1013 *) Load error codes if they are not already present instead of using a
1014 static variable. This allows them to be cleanly unloaded and reloaded.
1015 Improve header file function name parsing.
1018 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1019 or CAPABILITY handshake as required by RFCs.
1022 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1024 *) Introduce limits to prevent malicious keys being able to
1025 cause a denial of service. (CVE-2006-2940)
1026 [Steve Henson, Bodo Moeller]
1028 *) Fix ASN.1 parsing of certain invalid structures that can result
1029 in a denial of service. (CVE-2006-2937) [Steve Henson]
1031 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1032 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1034 *) Fix SSL client code which could crash if connecting to a
1035 malicious SSLv2 server. (CVE-2006-4343)
1036 [Tavis Ormandy and Will Drewry, Google Security Team]
1038 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1039 match only those. Before that, "AES256-SHA" would be interpreted
1040 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1041 the same strength classification in 0.9.7h) as we currently only
1042 have a single AES bit in the ciphersuite description bitmap.
1043 That change, however, also applied to ciphersuite strings such as
1044 "RC4-MD5" that intentionally matched multiple ciphersuites --
1045 namely, SSL 2.0 ciphersuites in addition to the more common ones
1046 from SSL 3.0/TLS 1.0.
1048 So we change the selection algorithm again: Naming an explicit
1049 ciphersuite selects this one ciphersuite, and any other similar
1050 ciphersuite (same bitmap) from *other* protocol versions.
1051 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1052 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1054 Since SSL 2.0 does not have any ciphersuites for which the
1055 128/256 bit distinction would be relevant, this works for now.
1056 The proper fix will be to use different bits for AES128 and
1057 AES256, which would have avoided the problems from the beginning;
1058 however, bits are scarce, so we can only do this in a new release
1059 (not just a patchlevel) when we can change the SSL_CIPHER
1060 definition to split the single 'unsigned long mask' bitmap into
1061 multiple values to extend the available space.
1065 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1067 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1068 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1070 *) Add AES IGE and biIGE modes.
1073 *) Change the Unix randomness entropy gathering to use poll() when
1074 possible instead of select(), since the latter has some
1075 undesirable limitations.
1076 [Darryl Miles via Richard Levitte and Bodo Moeller]
1078 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1079 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1080 cannot be implicitly activated as part of, e.g., the "AES" alias.
1081 However, please upgrade to OpenSSL 0.9.9[-dev] for
1082 non-experimental use of the ECC ciphersuites to get TLS extension
1083 support, which is required for curve and point format negotiation
1084 to avoid potential handshake problems.
1087 *) Disable rogue ciphersuites:
1089 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1090 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1091 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1093 The latter two were purportedly from
1094 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1097 Also deactivate the remaining ciphersuites from
1098 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1099 unofficial, and the ID has long expired.
1102 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1103 dual-core machines) and other potential thread-safety issues.
1106 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1107 versions), which is now available for royalty-free use
1108 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1109 Also, add Camellia TLS ciphersuites from RFC 4132.
1111 To minimize changes between patchlevels in the OpenSSL 0.9.8
1112 series, Camellia remains excluded from compilation unless OpenSSL
1113 is configured with 'enable-camellia'.
1116 *) Disable the padding bug check when compression is in use. The padding
1117 bug check assumes the first packet is of even length, this is not
1118 necessarily true if compresssion is enabled and can result in false
1119 positives causing handshake failure. The actual bug test is ancient
1120 code so it is hoped that implementations will either have fixed it by
1121 now or any which still have the bug do not support compression.
1124 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1126 *) When applying a cipher rule check to see if string match is an explicit
1127 cipher suite and only match that one cipher suite if it is.
1130 *) Link in manifests for VC++ if needed.
1131 [Austin Ziegler <halostatue@gmail.com>]
1133 *) Update support for ECC-based TLS ciphersuites according to
1134 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1135 TLS extensions, which are supported starting with the 0.9.9
1136 branch, not in the OpenSSL 0.9.8 branch).
1139 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1140 opaque EVP_CIPHER_CTX handling.
1143 *) Fixes and enhancements to zlib compression code. We now only use
1144 "zlib1.dll" and use the default __cdecl calling convention on Win32
1145 to conform with the standards mentioned here:
1146 http://www.zlib.net/DLL_FAQ.txt
1147 Static zlib linking now works on Windows and the new --with-zlib-include
1148 --with-zlib-lib options to Configure can be used to supply the location
1149 of the headers and library. Gracefully handle case where zlib library
1153 *) Several fixes and enhancements to the OID generation code. The old code
1154 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1155 handle numbers larger than ULONG_MAX, truncated printing and had a
1156 non standard OBJ_obj2txt() behaviour.
1159 *) Add support for building of engines under engine/ as shared libraries
1160 under VC++ build system.
1163 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1164 Hopefully, we will not see any false combination of paths any more.
1167 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1169 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1170 (part of SSL_OP_ALL). This option used to disable the
1171 countermeasure against man-in-the-middle protocol-version
1172 rollback in the SSL 2.0 server implementation, which is a bad
1173 idea. (CVE-2005-2969)
1175 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1176 for Information Security, National Institute of Advanced Industrial
1177 Science and Technology [AIST], Japan)]
1179 *) Add two function to clear and return the verify parameter flags.
1182 *) Keep cipherlists sorted in the source instead of sorting them at
1183 runtime, thus removing the need for a lock.
1186 *) Avoid some small subgroup attacks in Diffie-Hellman.
1187 [Nick Mathewson and Ben Laurie]
1189 *) Add functions for well-known primes.
1192 *) Extended Windows CE support.
1193 [Satoshi Nakamura and Andy Polyakov]
1195 *) Initialize SSL_METHOD structures at compile time instead of during
1196 runtime, thus removing the need for a lock.
1199 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1200 attempting to decrypt each encrypted key in turn. Add support to
1204 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1206 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1209 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1212 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1213 key into the same file any more.
1216 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1219 *) Add -utf8 command line and config file option to 'ca'.
1220 [Stefan <stf@udoma.org]
1222 *) Removed the macro des_crypt(), as it seems to conflict with some
1223 libraries. Use DES_crypt().
1226 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1227 involves renaming the source and generated shared-libs for
1228 both. The engines will accept the corrected or legacy ids
1229 ('ncipher' and '4758_cca' respectively) when binding. NB,
1230 this only applies when building 'shared'.
1231 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1233 *) Add attribute functions to EVP_PKEY structure. Modify
1234 PKCS12_create() to recognize a CSP name attribute and
1235 use it. Make -CSP option work again in pkcs12 utility.
1238 *) Add new functionality to the bn blinding code:
1239 - automatic re-creation of the BN_BLINDING parameters after
1240 a fixed number of uses (currently 32)
1241 - add new function for parameter creation
1242 - introduce flags to control the update behaviour of the
1243 BN_BLINDING parameters
1244 - hide BN_BLINDING structure
1245 Add a second BN_BLINDING slot to the RSA structure to improve
1246 performance when a single RSA object is shared among several
1250 *) Add support for DTLS.
1251 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1253 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1254 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1257 *) Remove buggy and incompletet DH cert support from
1258 ssl/ssl_rsa.c and ssl/s3_both.c
1261 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1262 the apps/openssl applications.
1265 *) Compile clean with "-Wall -Wmissing-prototypes
1266 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1267 DEBUG_SAFESTACK must also be set.
1270 *) Change ./Configure so that certain algorithms can be disabled by default.
1271 The new counterpiece to "no-xxx" is "enable-xxx".
1273 The patented RC5 and MDC2 algorithms will now be disabled unless
1274 "enable-rc5" and "enable-mdc2", respectively, are specified.
1276 (IDEA remains enabled despite being patented. This is because IDEA
1277 is frequently required for interoperability, and there is no license
1278 fee for non-commercial use. As before, "no-idea" can be used to
1279 avoid this algorithm.)
1283 *) Add processing of proxy certificates (see RFC 3820). This work was
1284 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1285 EGEE (Enabling Grids for E-science in Europe).
1288 *) RC4 performance overhaul on modern architectures/implementations, such
1289 as Intel P4, IA-64 and AMD64.
1292 *) New utility extract-section.pl. This can be used specify an alternative
1293 section number in a pod file instead of having to treat each file as
1294 a separate case in Makefile. This can be done by adding two lines to the
1297 =for comment openssl_section:XXX
1299 The blank line is mandatory.
1303 *) New arguments -certform, -keyform and -pass for s_client and s_server
1304 to allow alternative format key and certificate files and passphrase
1308 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1309 update associated structures and add various utility functions.
1311 Add new policy related verify parameters, include policy checking in
1312 standard verify code. Enhance 'smime' application with extra parameters
1313 to support policy checking and print out.
1316 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1317 Nehemiah processors. These extensions support AES encryption in hardware
1318 as well as RNG (though RNG support is currently disabled).
1319 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1321 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1324 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1325 [Andy Polyakov and a number of other people]
1327 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1328 implementation contributed by IBM.
1329 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1331 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1332 exponent rather than 'unsigned long'. There is a corresponding change to
1333 the new 'rsa_keygen' element of the RSA_METHOD structure.
1334 [Jelte Jansen, Geoff Thorpe]
1336 *) Functionality for creating the initial serial number file is now
1337 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1339 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1340 number file to 1, which is bound to cause problems. To avoid
1341 the problems while respecting compatibility between different 0.9.7
1342 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1343 CA.pl for serial number initialization. With the new release 0.9.8,
1344 we can fix the problem directly in the 'ca' utility.)
1347 *) Reduced header interdepencies by declaring more opaque objects in
1348 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1349 give fewer recursive includes, which could break lazy source code - so
1350 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1351 developers should define this symbol when building and using openssl to
1352 ensure they track the recommended behaviour, interfaces, [etc], but
1353 backwards-compatible behaviour prevails when this isn't defined.
1356 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1359 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1360 This will generate a random key of the appropriate length based on the
1361 cipher context. The EVP_CIPHER can provide its own random key generation
1362 routine to support keys of a specific form. This is used in the des and
1363 3des routines to generate a key of the correct parity. Update S/MIME
1364 code to use new functions and hence generate correct parity DES keys.
1365 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1366 valid (weak or incorrect parity).
1369 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1370 as looking them up. This is useful when the verified structure may contain
1371 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1372 present unless the new PKCS7_NO_CRL flag is asserted.
1375 *) Extend ASN1 oid configuration module. It now additionally accepts the
1378 shortName = some long name, 1.2.3.4
1381 *) Reimplemented the BN_CTX implementation. There is now no more static
1382 limitation on the number of variables it can handle nor the depth of the
1383 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1384 information can now expand as required, and rather than having a single
1385 static array of bignums, BN_CTX now uses a linked-list of such arrays
1386 allowing it to expand on demand whilst maintaining the usefulness of
1387 BN_CTX's "bundling".
1390 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1391 to allow all RSA operations to function using a single BN_CTX.
1394 *) Preliminary support for certificate policy evaluation and checking. This
1395 is initially intended to pass the tests outlined in "Conformance Testing
1396 of Relying Party Client Certificate Path Processing Logic" v1.07.
1399 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1400 remained unused and not that useful. A variety of other little bignum
1401 tweaks and fixes have also been made continuing on from the audit (see
1405 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1406 associated ASN1, EVP and SSL functions and old ASN1 macros.
1409 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1410 and this should never fail. So the return value from the use of
1411 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1412 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1415 *) BN_CTX_get() should return zero-valued bignums, providing the same
1416 initialised value as BN_new().
1417 [Geoff Thorpe, suggested by Ulf Möller]
1419 *) Support for inhibitAnyPolicy certificate extension.
1422 *) An audit of the BIGNUM code is underway, for which debugging code is
1423 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1424 is considered valid when processing BIGNUMs, and causes execution to
1425 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1426 further steps are taken to deliberately pollute unused data in BIGNUM
1427 structures to try and expose faulty code further on. For now, openssl will
1428 (in its default mode of operation) continue to tolerate the inconsistent
1429 forms that it has tolerated in the past, but authors and packagers should
1430 consider trying openssl and their own applications when compiled with
1431 these debugging symbols defined. It will help highlight potential bugs in
1432 their own code, and will improve the test coverage for OpenSSL itself. At
1433 some point, these tighter rules will become openssl's default to improve
1434 maintainability, though the assert()s and other overheads will remain only
1435 in debugging configurations. See bn.h for more details.
1436 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1438 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1439 that can only be obtained through BN_CTX_new() (which implicitly
1440 initialises it). The presence of this function only made it possible
1441 to overwrite an existing structure (and cause memory leaks).
1444 *) Because of the callback-based approach for implementing LHASH as a
1445 template type, lh_insert() adds opaque objects to hash-tables and
1446 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1447 to clean up those corresponding objects before destroying the hash table
1448 (and losing the object pointers). So some over-zealous constifications in
1449 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1450 objects as "const" and the lh_doall[_arg] callback wrappers are not
1451 prototyped to have "const" restrictions on the object pointers they are
1452 given (and so aren't required to cast them away any more).
1455 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1456 (speed) prefers to use its own implementation. The two implementations
1457 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1458 its object type properly exposed (MS_TM) instead of casting to/from "char
1459 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1460 aren't necessarily the greatest nomenclatures - but this is what was used
1461 internally to the implementation so I've used that for now.
1464 *) Ensure that deprecated functions do not get compiled when
1465 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1466 the self-tests were still using deprecated key-generation functions so
1467 these have been updated also.
1470 *) Reorganise PKCS#7 code to separate the digest location functionality
1471 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1472 New function PKCS7_set_digest() to set the digest type for PKCS#7
1473 digestedData type. Add additional code to correctly generate the
1474 digestedData type and add support for this type in PKCS7 initialization
1478 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1479 structure of type "other".
1482 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1483 sure the loop does correctly stop and breaking ("division by zero")
1484 modulus operations are not performed. The (pre-generated) prime
1485 table crypto/bn/bn_prime.h was already correct, but it could not be
1486 re-generated on some platforms because of the "division by zero"
1487 situation in the script.
1488 [Ralf S. Engelschall]
1490 *) Update support for ECC-based TLS ciphersuites according to
1491 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1492 SHA-1 now is only used for "small" curves (where the
1493 representation of a field element takes up to 24 bytes); for
1494 larger curves, the field element resulting from ECDH is directly
1495 used as premaster secret.
1496 [Douglas Stebila (Sun Microsystems Laboratories)]
1498 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1499 curve secp160r1 to the tests.
1500 [Douglas Stebila (Sun Microsystems Laboratories)]
1502 *) Add the possibility to load symbols globally with DSO.
1503 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1505 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1506 control of the error stack.
1509 *) Add support for STORE in ENGINE.
1512 *) Add the STORE type. The intention is to provide a common interface
1513 to certificate and key stores, be they simple file-based stores, or
1514 HSM-type store, or LDAP stores, or...
1515 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1518 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1519 pass a list of arguments to any function as well as provide a way
1520 for a function to pass data back to the caller.
1523 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1524 works like BUF_strdup() but can be used to duplicate a portion of
1525 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1529 *) Add the function sk_find_ex() which works like sk_find(), but will
1530 return an index to an element even if an exact match couldn't be
1531 found. The index is guaranteed to point at the element where the
1532 searched-for key would be inserted to preserve sorting order.
1535 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1536 takes an extra flags argument for optional functionality. Currently,
1537 the following flags are defined:
1539 OBJ_BSEARCH_VALUE_ON_NOMATCH
1540 This one gets OBJ_bsearch_ex() to return a pointer to the first
1541 element where the comparing function returns a negative or zero
1544 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1545 This one gets OBJ_bsearch_ex() to return a pointer to the first
1546 element where the comparing function returns zero. This is useful
1547 if there are more than one element where the comparing function
1551 *) Make it possible to create self-signed certificates with 'openssl ca'
1552 in such a way that the self-signed certificate becomes part of the
1553 CA database and uses the same mechanisms for serial number generation
1554 as all other certificate signing. The new flag '-selfsign' enables
1555 this functionality. Adapt CA.sh and CA.pl.in.
1558 *) Add functionality to check the public key of a certificate request
1559 against a given private. This is useful to check that a certificate
1560 request can be signed by that key (self-signing).
1563 *) Make it possible to have multiple active certificates with the same
1564 subject in the CA index file. This is done only if the keyword
1565 'unique_subject' is set to 'no' in the main CA section (default
1566 if 'CA_default') of the configuration file. The value is saved
1567 with the database itself in a separate index attribute file,
1568 named like the index file with '.attr' appended to the name.
1571 *) Generate muti valued AVAs using '+' notation in config files for
1575 *) Support for nameConstraints certificate extension.
1578 *) Support for policyConstraints certificate extension.
1581 *) Support for policyMappings certificate extension.
1584 *) Make sure the default DSA_METHOD implementation only uses its
1585 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1586 and change its own handlers to be NULL so as to remove unnecessary
1587 indirection. This lets alternative implementations fallback to the
1588 default implementation more easily.
1591 *) Support for directoryName in GeneralName related extensions
1595 *) Make it possible to link applications using Makefile.shared.
1596 Make that possible even when linking against static libraries!
1599 *) Support for single pass processing for S/MIME signing. This now
1600 means that S/MIME signing can be done from a pipe, in addition
1601 cleartext signing (multipart/signed type) is effectively streaming
1602 and the signed data does not need to be all held in memory.
1604 This is done with a new flag PKCS7_STREAM. When this flag is set
1605 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1606 is done after the data is output (and digests calculated) in
1607 SMIME_write_PKCS7().
1610 *) Add full support for -rpath/-R, both in shared libraries and
1611 applications, at least on the platforms where it's known how
1615 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1616 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1617 will now compute a table of multiples of the generator that
1618 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1619 faster (notably in the case of a single point multiplication,
1620 scalar * generator).
1621 [Nils Larsch, Bodo Moeller]
1623 *) IPv6 support for certificate extensions. The various extensions
1624 which use the IP:a.b.c.d can now take IPv6 addresses using the
1625 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1629 *) Added an ENGINE that implements RSA by performing private key
1630 exponentiations with the GMP library. The conversions to and from
1631 GMP's mpz_t format aren't optimised nor are any montgomery forms
1632 cached, and on x86 it appears OpenSSL's own performance has caught up.
1633 However there are likely to be other architectures where GMP could
1634 provide a boost. This ENGINE is not built in by default, but it can be
1635 specified at Configure time and should be accompanied by the necessary
1636 linker additions, eg;
1637 ./config -DOPENSSL_USE_GMP -lgmp
1640 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1641 testing availability of engines with "-t" - the old behaviour is
1642 produced by increasing the feature's verbosity with "-tt".
1645 *) ECDSA routines: under certain error conditions uninitialized BN objects
1646 could be freed. Solution: make sure initialization is performed early
1647 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1651 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1652 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1653 software implementations. For DSA and DH, parameter generation can
1654 also be overriden by providing the appropriate method callbacks.
1657 *) Change the "progress" mechanism used in key-generation and
1658 primality testing to functions that take a new BN_GENCB pointer in
1659 place of callback/argument pairs. The new API functions have "_ex"
1660 postfixes and the older functions are reimplemented as wrappers for
1661 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1662 declarations of the old functions to help (graceful) attempts to
1663 migrate to the new functions. Also, the new key-generation API
1664 functions operate on a caller-supplied key-structure and return
1665 success/failure rather than returning a key or NULL - this is to
1666 help make "keygen" another member function of RSA_METHOD etc.
1668 Example for using the new callback interface:
1670 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1674 BN_GENCB_set(&my_cb, my_callback, my_arg);
1676 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1677 /* For the meaning of a, b in calls to my_callback(), see the
1678 * documentation of the function that calls the callback.
1679 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1680 * my_callback should return 1 if it wants BN_is_prime_ex()
1681 * to continue, or 0 to stop.
1686 *) Change the ZLIB compression method to be stateful, and make it
1687 available to TLS with the number defined in
1688 draft-ietf-tls-compression-04.txt.
1691 *) Add the ASN.1 structures and functions for CertificatePair, which
1692 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1694 CertificatePair ::= SEQUENCE {
1695 forward [0] Certificate OPTIONAL,
1696 reverse [1] Certificate OPTIONAL,
1697 -- at least one of the pair shall be present -- }
1699 Also implement the PEM functions to read and write certificate
1700 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1702 This needed to be defined, mostly for the sake of the LDAP
1703 attribute crossCertificatePair, but may prove useful elsewhere as
1707 *) Make it possible to inhibit symlinking of shared libraries in
1708 Makefile.shared, for Cygwin's sake.
1711 *) Extend the BIGNUM API by creating a function
1712 void BN_set_negative(BIGNUM *a, int neg);
1713 and a macro that behave like
1714 int BN_is_negative(const BIGNUM *a);
1716 to avoid the need to access 'a->neg' directly in applications.
1719 *) Implement fast modular reduction for pseudo-Mersenne primes
1720 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1721 EC_GROUP_new_curve_GFp() will now automatically use this
1723 [Nils Larsch <nla@trustcenter.de>]
1725 *) Add new lock type (CRYPTO_LOCK_BN).
1728 *) Change the ENGINE framework to automatically load engines
1729 dynamically from specific directories unless they could be
1730 found to already be built in or loaded. Move all the
1731 current engines except for the cryptodev one to a new
1733 The engines in engines/ are built as shared libraries if
1734 the "shared" options was given to ./Configure or ./config.
1735 Otherwise, they are inserted in libcrypto.a.
1736 /usr/local/ssl/engines is the default directory for dynamic
1737 engines, but that can be overriden at configure time through
1738 the usual use of --prefix and/or --openssldir, and at run
1739 time with the environment variable OPENSSL_ENGINES.
1740 [Geoff Thorpe and Richard Levitte]
1742 *) Add Makefile.shared, a helper makefile to build shared
1743 libraries. Addapt Makefile.org.
1746 *) Add version info to Win32 DLLs.
1747 [Peter 'Luna' Runestig" <peter@runestig.com>]
1749 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1750 can be added using this API to created arbitrary PKCS#12
1751 files while avoiding the low level API.
1753 New options to PKCS12_create(), key or cert can be NULL and
1754 will then be omitted from the output file. The encryption
1755 algorithm NIDs can be set to -1 for no encryption, the mac
1756 iteration count can be set to 0 to omit the mac.
1758 Enhance pkcs12 utility by making the -nokeys and -nocerts
1759 options work when creating a PKCS#12 file. New option -nomac
1760 to omit the mac, NONE can be set for an encryption algorithm.
1761 New code is modified to use the enhanced PKCS12_create()
1762 instead of the low level API.
1765 *) Extend ASN1 encoder to support indefinite length constructed
1766 encoding. This can output sequences tags and octet strings in
1767 this form. Modify pk7_asn1.c to support indefinite length
1768 encoding. This is experimental and needs additional code to
1769 be useful, such as an ASN1 bio and some enhanced streaming
1772 Extend template encode functionality so that tagging is passed
1773 down to the template encoder.
1776 *) Let 'openssl req' fail if an argument to '-newkey' is not
1777 recognized instead of using RSA as a default.
1780 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1781 As these are not official, they are not included in "ALL";
1782 the "ECCdraft" ciphersuite group alias can be used to select them.
1783 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1785 *) Add ECDH engine support.
1786 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1788 *) Add ECDH in new directory crypto/ecdh/.
1789 [Douglas Stebila (Sun Microsystems Laboratories)]
1791 *) Let BN_rand_range() abort with an error after 100 iterations
1792 without success (which indicates a broken PRNG).
1795 *) Change BN_mod_sqrt() so that it verifies that the input value
1796 is really the square of the return value. (Previously,
1797 BN_mod_sqrt would show GIGO behaviour.)
1800 *) Add named elliptic curves over binary fields from X9.62, SECG,
1801 and WAP/WTLS; add OIDs that were still missing.
1803 [Sheueling Chang Shantz and Douglas Stebila
1804 (Sun Microsystems Laboratories)]
1806 *) Extend the EC library for elliptic curves over binary fields
1807 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1810 EC_GF2m_simple_method
1814 EC_GROUP_new_curve_GF2m
1815 EC_GROUP_set_curve_GF2m
1816 EC_GROUP_get_curve_GF2m
1817 EC_POINT_set_affine_coordinates_GF2m
1818 EC_POINT_get_affine_coordinates_GF2m
1819 EC_POINT_set_compressed_coordinates_GF2m
1821 Point compression for binary fields is disabled by default for
1822 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1825 As binary polynomials are represented as BIGNUMs, various members
1826 of the EC_GROUP and EC_POINT data structures can be shared
1827 between the implementations for prime fields and binary fields;
1828 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1829 are essentially identical to their ..._GFp counterparts.
1830 (For simplicity, the '..._GFp' prefix has been dropped from
1831 various internal method names.)
1833 An internal 'field_div' method (similar to 'field_mul' and
1834 'field_sqr') has been added; this is used only for binary fields.
1836 [Sheueling Chang Shantz and Douglas Stebila
1837 (Sun Microsystems Laboratories)]
1839 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1840 through methods ('mul', 'precompute_mult').
1842 The generic implementations (now internally called 'ec_wNAF_mul'
1843 and 'ec_wNAF_precomputed_mult') remain the default if these
1844 methods are undefined.
1846 [Sheueling Chang Shantz and Douglas Stebila
1847 (Sun Microsystems Laboratories)]
1849 *) New function EC_GROUP_get_degree, which is defined through
1850 EC_METHOD. For curves over prime fields, this returns the bit
1851 length of the modulus.
1853 [Sheueling Chang Shantz and Douglas Stebila
1854 (Sun Microsystems Laboratories)]
1856 *) New functions EC_GROUP_dup, EC_POINT_dup.
1857 (These simply call ..._new and ..._copy).
1859 [Sheueling Chang Shantz and Douglas Stebila
1860 (Sun Microsystems Laboratories)]
1862 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1863 Polynomials are represented as BIGNUMs (where the sign bit is not
1864 used) in the following functions [macros]:
1867 BN_GF2m_sub [= BN_GF2m_add]
1868 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1869 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1870 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1872 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1873 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1874 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1875 BN_GF2m_cmp [= BN_ucmp]
1877 (Note that only the 'mod' functions are actually for fields GF(2^m).
1878 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1880 For some functions, an the irreducible polynomial defining a
1881 field can be given as an 'unsigned int[]' with strictly
1882 decreasing elements giving the indices of those bits that are set;
1883 i.e., p[] represents the polynomial
1884 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1886 p[0] > p[1] > ... > p[k] = 0.
1887 This applies to the following functions:
1892 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1893 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1895 BN_GF2m_mod_sqrt_arr
1896 BN_GF2m_mod_solve_quad_arr
1900 Conversion can be performed by the following functions:
1905 bntest.c has additional tests for binary polynomial arithmetic.
1907 Two implementations for BN_GF2m_mod_div() are available.
1908 The default algorithm simply uses BN_GF2m_mod_inv() and
1909 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1910 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1911 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1913 [Sheueling Chang Shantz and Douglas Stebila
1914 (Sun Microsystems Laboratories)]
1916 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1917 functionality is disabled at compile-time.
1918 [Douglas Stebila <douglas.stebila@sun.com>]
1920 *) Change default behaviour of 'openssl asn1parse' so that more
1921 information is visible when viewing, e.g., a certificate:
1923 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1924 mode the content of non-printable OCTET STRINGs is output in a
1925 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1926 avoid the appearance of a printable string.
1927 [Nils Larsch <nla@trustcenter.de>]
1929 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1931 EC_GROUP_set_asn1_flag()
1932 EC_GROUP_get_asn1_flag()
1933 EC_GROUP_set_point_conversion_form()
1934 EC_GROUP_get_point_conversion_form()
1935 These control ASN1 encoding details:
1936 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1937 has been set to OPENSSL_EC_NAMED_CURVE.
1938 - Points are encoded in uncompressed form by default; options for
1939 asn1_for are as for point2oct, namely
1940 POINT_CONVERSION_COMPRESSED
1941 POINT_CONVERSION_UNCOMPRESSED
1942 POINT_CONVERSION_HYBRID
1944 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1947 EC_GROUP_get0_seed()
1948 EC_GROUP_get_seed_len()
1949 This is used only for ASN1 purposes (so far).
1950 [Nils Larsch <nla@trustcenter.de>]
1952 *) Add 'field_type' member to EC_METHOD, which holds the NID
1953 of the appropriate field type OID. The new function
1954 EC_METHOD_get_field_type() returns this value.
1955 [Nils Larsch <nla@trustcenter.de>]
1960 EC_POINT_point2hex()
1961 EC_POINT_hex2point()
1962 providing useful interfaces to EC_POINT_point2oct() and
1963 EC_POINT_oct2point().
1964 [Nils Larsch <nla@trustcenter.de>]
1966 *) Change internals of the EC library so that the functions
1967 EC_GROUP_set_generator()
1968 EC_GROUP_get_generator()
1969 EC_GROUP_get_order()
1970 EC_GROUP_get_cofactor()
1971 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1972 to methods, which would lead to unnecessary code duplication when
1973 adding different types of curves.
1974 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1976 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1977 arithmetic, and such that modified wNAFs are generated
1978 (which avoid length expansion in many cases).
1981 *) Add a function EC_GROUP_check_discriminant() (defined via
1982 EC_METHOD) that verifies that the curve discriminant is non-zero.
1984 Add a function EC_GROUP_check() that makes some sanity tests
1985 on a EC_GROUP, its generator and order. This includes
1986 EC_GROUP_check_discriminant().
1987 [Nils Larsch <nla@trustcenter.de>]
1989 *) Add ECDSA in new directory crypto/ecdsa/.
1991 Add applications 'openssl ecparam' and 'openssl ecdsa'
1992 (these are based on 'openssl dsaparam' and 'openssl dsa').
1994 ECDSA support is also included in various other files across the
1995 library. Most notably,
1996 - 'openssl req' now has a '-newkey ecdsa:file' option;
1997 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1998 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1999 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2000 them suitable for ECDSA where domain parameters must be
2001 extracted before the specific public key;
2002 - ECDSA engine support has been added.
2003 [Nils Larsch <nla@trustcenter.de>]
2005 *) Include some named elliptic curves, and add OIDs from X9.62,
2006 SECG, and WAP/WTLS. Each curve can be obtained from the new
2008 EC_GROUP_new_by_curve_name(),
2009 and the list of available named curves can be obtained with
2010 EC_get_builtin_curves().
2011 Also add a 'curve_name' member to EC_GROUP objects, which can be
2013 EC_GROUP_set_curve_name()
2014 EC_GROUP_get_curve_name()
2015 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2017 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2018 was actually never needed) and in BN_mul(). The removal in BN_mul()
2019 required a small change in bn_mul_part_recursive() and the addition
2020 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2021 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2022 bn_sub_words() and bn_add_words() except they take arrays with
2026 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2028 *) Cleanse PEM buffers before freeing them since they may contain
2030 [Benjamin Bennett <ben@psc.edu>]
2032 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2033 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2034 authentication-only ciphersuites.
2037 *) Since AES128 and AES256 share a single mask bit in the logic of
2038 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2039 kludge to work properly if AES128 is available and AES256 isn't.
2042 *) Expand security boundary to match 1.1.1 module.
2045 *) Remove redundant features: hash file source, editing of test vectors
2046 modify fipsld to use external fips_premain.c signature.
2049 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2050 run algorithm test programs.
2053 *) Make algorithm test programs more tolerant of whitespace.
2056 *) Have SSL/TLS server implementation tolerate "mismatched" record
2057 protocol version while receiving ClientHello even if the
2058 ClientHello is fragmented. (The server can't insist on the
2059 particular protocol version it has chosen before the ServerHello
2060 message has informed the client about his choice.)
2063 *) Load error codes if they are not already present instead of using a
2064 static variable. This allows them to be cleanly unloaded and reloaded.
2067 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2069 *) Introduce limits to prevent malicious keys being able to
2070 cause a denial of service. (CVE-2006-2940)
2071 [Steve Henson, Bodo Moeller]
2073 *) Fix ASN.1 parsing of certain invalid structures that can result
2074 in a denial of service. (CVE-2006-2937) [Steve Henson]
2076 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2077 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2079 *) Fix SSL client code which could crash if connecting to a
2080 malicious SSLv2 server. (CVE-2006-4343)
2081 [Tavis Ormandy and Will Drewry, Google Security Team]
2083 *) Change ciphersuite string processing so that an explicit
2084 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2085 will no longer include "AES128-SHA"), and any other similar
2086 ciphersuite (same bitmap) from *other* protocol versions (so that
2087 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2088 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2089 changes from 0.9.8b and 0.9.8d.
2092 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2094 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2095 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2097 *) Change the Unix randomness entropy gathering to use poll() when
2098 possible instead of select(), since the latter has some
2099 undesirable limitations.
2100 [Darryl Miles via Richard Levitte and Bodo Moeller]
2102 *) Disable rogue ciphersuites:
2104 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2105 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2106 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2108 The latter two were purportedly from
2109 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2112 Also deactive the remaining ciphersuites from
2113 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2114 unofficial, and the ID has long expired.
2117 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2118 dual-core machines) and other potential thread-safety issues.
2121 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2123 *) Adapt fipsld and the build system to link against the validated FIPS
2124 module in FIPS mode.
2127 *) Fixes for VC++ 2005 build under Windows.
2130 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2131 from a Windows bash shell such as MSYS. It is autodetected from the
2132 "config" script when run from a VC++ environment. Modify standard VC++
2133 build to use fipscanister.o from the GNU make build.
2136 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2138 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2139 The value now differs depending on if you build for FIPS or not.
2140 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2141 safely run with a non-FIPSed libcrypto, as it may crash because of
2142 the difference induced by this change.
2145 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2147 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2148 (part of SSL_OP_ALL). This option used to disable the
2149 countermeasure against man-in-the-middle protocol-version
2150 rollback in the SSL 2.0 server implementation, which is a bad
2151 idea. (CVE-2005-2969)
2153 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2154 for Information Security, National Institute of Advanced Industrial
2155 Science and Technology [AIST], Japan)]
2157 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2158 mainly for FIPS compliance and not fully integrated at this stage.
2161 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2162 the exponentiation using a fixed-length exponent. (Otherwise,
2163 the information leaked through timing could expose the secret key
2164 after many signatures; cf. Bleichenbacher's attack on DSA with
2168 *) Make a new fixed-window mod_exp implementation the default for
2169 RSA, DSA, and DH private-key operations so that the sequence of
2170 squares and multiplies and the memory access pattern are
2171 independent of the particular secret key. This will mitigate
2172 cache-timing and potential related attacks.
2174 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2175 and this is automatically used by BN_mod_exp_mont() if the new flag
2176 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2177 will use this BN flag for private exponents unless the flag
2178 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2179 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2181 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2183 *) Change the client implementation for SSLv23_method() and
2184 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2185 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2186 (Previously, the SSL 2.0 backwards compatible Client Hello
2187 message format would be used even with SSL_OP_NO_SSLv2.)
2190 *) Add support for smime-type MIME parameter in S/MIME messages which some
2194 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2195 a threadsafe manner. Modify rsa code to use new function and add calls
2196 to dsa and dh code (which had race conditions before).
2199 *) Include the fixed error library code in the C error file definitions
2200 instead of fixing them up at runtime. This keeps the error code
2201 structures constant.
2204 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2206 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2209 *) Fixes for newer kerberos headers. NB: the casts are needed because
2210 the 'length' field is signed on one version and unsigned on another
2211 with no (?) obvious way to tell the difference, without these VC++
2212 complains. Also the "definition" of FAR (blank) is no longer included
2213 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2214 some needed definitions.
2217 *) Undo Cygwin change.
2220 *) Added support for proxy certificates according to RFC 3820.
2221 Because they may be a security thread to unaware applications,
2222 they must be explicitely allowed in run-time. See
2223 docs/HOWTO/proxy_certificates.txt for further information.
2226 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2228 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2229 server and client random values. Previously
2230 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2231 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2233 This change has negligible security impact because:
2235 1. Server and client random values still have 24 bytes of pseudo random
2238 2. Server and client random values are sent in the clear in the initial
2241 3. The master secret is derived using the premaster secret (48 bytes in
2242 size for static RSA ciphersuites) as well as client server and random
2245 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2248 [Stephen Henson, reported by UK NISCC]
2250 *) Use Windows randomness collection on Cygwin.
2253 *) Fix hang in EGD/PRNGD query when communication socket is closed
2254 prematurely by EGD/PRNGD.
2255 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2257 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2260 *) Back-port of selected performance improvements from development
2261 branch, as well as improved support for PowerPC platforms.
2264 *) Add lots of checks for memory allocation failure, error codes to indicate
2265 failure and freeing up memory if a failure occurs.
2266 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2268 *) Add new -passin argument to dgst.
2271 *) Perform some character comparisons of different types in X509_NAME_cmp:
2272 this is needed for some certificates that reencode DNs into UTF8Strings
2273 (in violation of RFC3280) and can't or wont issue name rollover
2277 *) Make an explicit check during certificate validation to see that
2278 the CA setting in each certificate on the chain is correct. As a
2279 side effect always do the following basic checks on extensions,
2280 not just when there's an associated purpose to the check:
2282 - if there is an unhandled critical extension (unless the user
2283 has chosen to ignore this fault)
2284 - if the path length has been exceeded (if one is set at all)
2285 - that certain extensions fit the associated purpose (if one has
2289 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2291 *) Avoid a race condition when CRLs are checked in a multi threaded
2292 environment. This would happen due to the reordering of the revoked
2293 entries during signature checking and serial number lookup. Now the
2294 encoding is cached and the serial number sort performed under a lock.
2295 Add new STACK function sk_is_sorted().
2298 *) Add Delta CRL to the extension code.
2301 *) Various fixes to s3_pkt.c so alerts are sent properly.
2302 [David Holmes <d.holmes@f5.com>]
2304 *) Reduce the chances of duplicate issuer name and serial numbers (in
2305 violation of RFC3280) using the OpenSSL certificate creation utilities.
2306 This is done by creating a random 64 bit value for the initial serial
2307 number when a serial number file is created or when a self signed
2308 certificate is created using 'openssl req -x509'. The initial serial
2309 number file is created using 'openssl x509 -next_serial' in CA.pl
2310 rather than being initialized to 1.
2313 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2315 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2316 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2317 [Joe Orton, Steve Henson]
2319 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2321 [Joe Orton, Steve Henson]
2323 *) Make it possible to have multiple active certificates with the same
2324 subject in the CA index file. This is done only if the keyword
2325 'unique_subject' is set to 'no' in the main CA section (default
2326 if 'CA_default') of the configuration file. The value is saved
2327 with the database itself in a separate index attribute file,
2328 named like the index file with '.attr' appended to the name.
2331 *) X509 verify fixes. Disable broken certificate workarounds when
2332 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2333 keyUsage extension present. Don't accept CRLs with unhandled critical
2334 extensions: since verify currently doesn't process CRL extensions this
2335 rejects a CRL with *any* critical extensions. Add new verify error codes
2339 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2340 A clarification of RFC2560 will require the use of OCTET STRINGs and
2341 some implementations cannot handle the current raw format. Since OpenSSL
2342 copies and compares OCSP nonces as opaque blobs without any attempt at
2343 parsing them this should not create any compatibility issues.
2346 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2347 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2348 this HMAC (and other) operations are several times slower than OpenSSL
2352 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2353 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2355 *) Use the correct content when signing type "other".
2358 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2360 *) Fix various bugs revealed by running the NISCC test suite:
2362 Stop out of bounds reads in the ASN1 code when presented with
2363 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2365 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2367 If verify callback ignores invalid public key errors don't try to check
2368 certificate signature with the NULL public key.
2372 *) New -ignore_err option in ocsp application to stop the server
2373 exiting on the first error in a request.
2376 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2377 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2381 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2382 extra data after the compression methods not only for TLS 1.0
2383 but also for SSL 3.0 (as required by the specification).
2384 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2386 *) Change X509_certificate_type() to mark the key as exported/exportable
2387 when it's 512 *bits* long, not 512 bytes.
2390 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2391 blocks during encryption.
2394 *) Various fixes to base64 BIO and non blocking I/O. On write
2395 flushes were not handled properly if the BIO retried. On read
2396 data was not being buffered properly and had various logic bugs.
2397 This also affects blocking I/O when the data being decoded is a
2401 *) Various S/MIME bugfixes and compatibility changes:
2402 output correct application/pkcs7 MIME type if
2403 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2404 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2405 of files as .eml work). Correctly handle very long lines in MIME
2409 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2411 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2412 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2413 a protocol version number mismatch like a decryption error
2414 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2417 *) Turn on RSA blinding by default in the default implementation
2418 to avoid a timing attack. Applications that don't want it can call
2419 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2420 They would be ill-advised to do so in most cases.
2421 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2423 *) Change RSA blinding code so that it works when the PRNG is not
2424 seeded (in this case, the secret RSA exponent is abused as
2425 an unpredictable seed -- if it is not unpredictable, there
2426 is no point in blinding anyway). Make RSA blinding thread-safe
2427 by remembering the creator's thread ID in rsa->blinding and
2428 having all other threads use local one-time blinding factors
2429 (this requires more computation than sharing rsa->blinding, but
2430 avoids excessive locking; and if an RSA object is not shared
2431 between threads, blinding will still be very fast).
2434 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2435 ENGINE as defaults for all supported algorithms irrespective of
2436 the 'flags' parameter. 'flags' is now honoured, so applications
2437 should make sure they are passing it correctly.
2440 *) Target "mingw" now allows native Windows code to be generated in
2441 the Cygwin environment as well as with the MinGW compiler.
2444 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2446 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2447 via timing by performing a MAC computation even if incorrrect
2448 block cipher padding has been found. This is a countermeasure
2449 against active attacks where the attacker has to distinguish
2450 between bad padding and a MAC verification error. (CVE-2003-0078)
2452 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2453 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2454 Martin Vuagnoux (EPFL, Ilion)]
2456 *) Make the no-err option work as intended. The intention with no-err
2457 is not to have the whole error stack handling routines removed from
2458 libcrypto, it's only intended to remove all the function name and
2459 reason texts, thereby removing some of the footprint that may not
2460 be interesting if those errors aren't displayed anyway.
2462 NOTE: it's still possible for any application or module to have it's
2463 own set of error texts inserted. The routines are there, just not
2464 used by default when no-err is given.
2467 *) Add support for FreeBSD on IA64.
2468 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2470 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2471 Kerberos function mit_des_cbc_cksum(). Before this change,
2472 the value returned by DES_cbc_cksum() was like the one from
2473 mit_des_cbc_cksum(), except the bytes were swapped.
2474 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2476 *) Allow an application to disable the automatic SSL chain building.
2477 Before this a rather primitive chain build was always performed in
2478 ssl3_output_cert_chain(): an application had no way to send the
2479 correct chain if the automatic operation produced an incorrect result.
2481 Now the chain builder is disabled if either:
2483 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2485 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2487 The reasoning behind this is that an application would not want the
2488 auto chain building to take place if extra chain certificates are
2489 present and it might also want a means of sending no additional
2490 certificates (for example the chain has two certificates and the
2494 *) Add the possibility to build without the ENGINE framework.
2495 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2497 *) Under Win32 gmtime() can return NULL: check return value in
2498 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2501 *) DSA routines: under certain error conditions uninitialized BN objects
2502 could be freed. Solution: make sure initialization is performed early
2503 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2504 Nils Larsch <nla@trustcenter.de> via PR#459)
2507 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2508 checked on reconnect on the client side, therefore session resumption
2509 could still fail with a "ssl session id is different" error. This
2510 behaviour is masked when SSL_OP_ALL is used due to
2511 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2512 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2513 followup to PR #377.
2516 *) IA-32 assembler support enhancements: unified ELF targets, support
2517 for SCO/Caldera platforms, fix for Cygwin shared build.
2520 *) Add support for FreeBSD on sparc64. As a consequence, support for
2521 FreeBSD on non-x86 processors is separate from x86 processors on
2522 the config script, much like the NetBSD support.
2523 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2525 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2527 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2530 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2531 code (06) was taken as the first octet of the session ID and the last
2532 octet was ignored consequently. As a result SSLv2 client side session
2533 caching could not have worked due to the session ID mismatch between
2535 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2539 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2540 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2544 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2545 seems that in spite of existing for more than a year, many application
2546 author have done nothing to provide the necessary callbacks, which
2547 means that this particular engine will not work properly anywhere.
2548 This is a very unfortunate situation which forces us, in the name
2549 of usability, to give the hw_ncipher.c a static lock, which is part
2551 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2552 appear in 0.9.8 or later. We EXPECT application authors to have
2553 dealt properly with this when 0.9.8 is released (unless we actually
2554 make such changes in the libcrypto locking code that changes will
2555 have to be made anyway).
2558 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2559 octets have been read, EOF or an error occurs. Without this change
2560 some truncated ASN1 structures will not produce an error.
2563 *) Disable Heimdal support, since it hasn't been fully implemented.
2564 Still give the possibility to force the use of Heimdal, but with
2565 warnings and a request that patches get sent to openssl-dev.
2568 *) Add the VC-CE target, introduce the WINCE sysname, and add
2569 INSTALL.WCE and appropriate conditionals to make it build.
2570 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2572 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2573 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2574 edit numbers of the version.
2575 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2577 *) Introduce safe string copy and catenation functions
2578 (BUF_strlcpy() and BUF_strlcat()).
2579 [Ben Laurie (CHATS) and Richard Levitte]
2581 *) Avoid using fixed-size buffers for one-line DNs.
2582 [Ben Laurie (CHATS)]
2584 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2585 resizing buffers containing secrets, and use where appropriate.
2586 [Ben Laurie (CHATS)]
2588 *) Avoid using fixed size buffers for configuration file location.
2589 [Ben Laurie (CHATS)]
2591 *) Avoid filename truncation for various CA files.
2592 [Ben Laurie (CHATS)]
2594 *) Use sizeof in preference to magic numbers.
2595 [Ben Laurie (CHATS)]
2597 *) Avoid filename truncation in cert requests.
2598 [Ben Laurie (CHATS)]
2600 *) Add assertions to check for (supposedly impossible) buffer
2602 [Ben Laurie (CHATS)]
2604 *) Don't cache truncated DNS entries in the local cache (this could
2605 potentially lead to a spoofing attack).
2606 [Ben Laurie (CHATS)]
2608 *) Fix various buffers to be large enough for hex/decimal
2609 representations in a platform independent manner.
2610 [Ben Laurie (CHATS)]
2612 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2613 resizing buffers containing secrets, and use where appropriate.
2614 [Ben Laurie (CHATS)]
2616 *) Add BIO_indent() to avoid much slightly worrying code to do
2618 [Ben Laurie (CHATS)]
2620 *) Convert sprintf()/BIO_puts() to BIO_printf().
2621 [Ben Laurie (CHATS)]
2623 *) buffer_gets() could terminate with the buffer only half
2625 [Ben Laurie (CHATS)]
2627 *) Add assertions to prevent user-supplied crypto functions from
2628 overflowing internal buffers by having large block sizes, etc.
2629 [Ben Laurie (CHATS)]
2631 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2633 [Ben Laurie (CHATS)]
2635 *) Eliminate unused copy of key in RC4.
2636 [Ben Laurie (CHATS)]
2638 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2639 [Ben Laurie (CHATS)]
2641 *) Fix off-by-one error in EGD path.
2642 [Ben Laurie (CHATS)]
2644 *) If RANDFILE path is too long, ignore instead of truncating.
2645 [Ben Laurie (CHATS)]
2647 *) Eliminate unused and incorrectly sized X.509 structure
2649 [Ben Laurie (CHATS)]
2651 *) Eliminate unused and dangerous function knumber().
2652 [Ben Laurie (CHATS)]
2654 *) Eliminate unused and dangerous structure, KSSL_ERR.
2655 [Ben Laurie (CHATS)]
2657 *) Protect against overlong session ID context length in an encoded
2658 session object. Since these are local, this does not appear to be
2660 [Ben Laurie (CHATS)]
2662 *) Change from security patch (see 0.9.6e below) that did not affect
2663 the 0.9.6 release series:
2665 Remote buffer overflow in SSL3 protocol - an attacker could
2666 supply an oversized master key in Kerberos-enabled versions.
2668 [Ben Laurie (CHATS)]
2670 *) Change the SSL kerb5 codes to match RFC 2712.
2673 *) Make -nameopt work fully for req and add -reqopt switch.
2674 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2676 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2677 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2679 *) Make sure tests can be performed even if the corresponding algorithms
2680 have been removed entirely. This was also the last step to make
2681 OpenSSL compilable with DJGPP under all reasonable conditions.
2682 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2684 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2685 to allow version independent disabling of normally unselected ciphers,
2686 which may be activated as a side-effect of selecting a single cipher.
2688 (E.g., cipher list string "RSA" enables ciphersuites that are left
2689 out of "ALL" because they do not provide symmetric encryption.
2690 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2691 [Lutz Jaenicke, Bodo Moeller]
2693 *) Add appropriate support for separate platform-dependent build
2694 directories. The recommended way to make a platform-dependent
2695 build directory is the following (tested on Linux), maybe with
2698 # Place yourself outside of the OpenSSL source tree. In
2699 # this example, the environment variable OPENSSL_SOURCE
2700 # is assumed to contain the absolute OpenSSL source directory.
2701 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2702 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2703 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2704 mkdir -p `dirname $F`
2705 ln -s $OPENSSL_SOURCE/$F $F
2708 To be absolutely sure not to disturb the source tree, a "make clean"
2709 is a good thing. If it isn't successfull, don't worry about it,
2710 it probably means the source directory is very clean.
2713 *) Make sure any ENGINE control commands make local copies of string
2714 pointers passed to them whenever necessary. Otherwise it is possible
2715 the caller may have overwritten (or deallocated) the original string
2716 data when a later ENGINE operation tries to use the stored values.
2717 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2719 *) Improve diagnostics in file reading and command-line digests.
2720 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2722 *) Add AES modes CFB and OFB to the object database. Correct an
2723 error in AES-CFB decryption.
2726 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2727 allows existing EVP_CIPHER_CTX structures to be reused after
2728 calling EVP_*Final(). This behaviour is used by encryption
2729 BIOs and some applications. This has the side effect that
2730 applications must explicitly clean up cipher contexts with
2731 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2734 *) Check the values of dna and dnb in bn_mul_recursive before calling
2735 bn_mul_comba (a non zero value means the a or b arrays do not contain
2736 n2 elements) and fallback to bn_mul_normal if either is not zero.
2739 *) Fix escaping of non-ASCII characters when using the -subj option
2740 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2743 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2744 form for "surname", serialNumber has no short form.
2745 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2746 therefore remove "mail" short name for "internet 7".
2747 The OID for unique identifiers in X509 certificates is
2748 x500UniqueIdentifier, not uniqueIdentifier.
2749 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2752 *) Add an "init" command to the ENGINE config module and auto initialize
2753 ENGINEs. Without any "init" command the ENGINE will be initialized
2754 after all ctrl commands have been executed on it. If init=1 the
2755 ENGINE is initailized at that point (ctrls before that point are run
2756 on the uninitialized ENGINE and after on the initialized one). If
2757 init=0 then the ENGINE will not be iniatialized at all.
2760 *) Fix the 'app_verify_callback' interface so that the user-defined
2761 argument is actually passed to the callback: In the
2762 SSL_CTX_set_cert_verify_callback() prototype, the callback
2763 declaration has been changed from
2766 int (*cb)(X509_STORE_CTX *,void *);
2767 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2768 i=s->ctx->app_verify_callback(&ctx)
2769 has been changed into
2770 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2772 To update applications using SSL_CTX_set_cert_verify_callback(),
2773 a dummy argument can be added to their callback functions.
2774 [D. K. Smetters <smetters@parc.xerox.com>]
2776 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2777 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2779 *) Add and OPENSSL_LOAD_CONF define which will cause
2780 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2781 This allows older applications to transparently support certain
2782 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2783 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2784 load the config file and OPENSSL_add_all_algorithms_conf() which will
2785 always load it have also been added.
2788 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2789 Adjust NIDs and EVP layer.
2790 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2792 *) Config modules support in openssl utility.
2794 Most commands now load modules from the config file,
2795 though in a few (such as version) this isn't done
2796 because it couldn't be used for anything.
2798 In the case of ca and req the config file used is
2799 the same as the utility itself: that is the -config
2800 command line option can be used to specify an
2804 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2805 use "openssl_conf" if filename is NULL use default openssl config file.
2808 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2809 config section name. Add a new flag to tolerate a missing config file
2810 and move code to CONF_modules_load_file().
2813 *) Support for crypto accelerator cards from Accelerated Encryption
2814 Processing, www.aep.ie. (Use engine 'aep')
2815 The support was copied from 0.9.6c [engine] and adapted/corrected
2816 to work with the new engine framework.
2817 [AEP Inc. and Richard Levitte]
2819 *) Support for SureWare crypto accelerator cards from Baltimore
2820 Technologies. (Use engine 'sureware')
2821 The support was copied from 0.9.6c [engine] and adapted
2822 to work with the new engine framework.
2825 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2826 make the newer ENGINE framework commands for the CHIL engine work.
2827 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2829 *) Make it possible to produce shared libraries on ReliantUNIX.
2830 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2832 *) Add the configuration target debug-linux-ppro.
2833 Make 'openssl rsa' use the general key loading routines
2834 implemented in apps.c, and make those routines able to
2835 handle the key format FORMAT_NETSCAPE and the variant
2837 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2839 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2840 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2842 *) Add -keyform to rsautl, and document -engine.
2843 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2845 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2846 BIO_R_NO_SUCH_FILE error code rather than the generic
2847 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2850 *) Add new functions
2852 ERR_peek_last_error_line
2853 ERR_peek_last_error_line_data.
2854 These are similar to
2857 ERR_peek_error_line_data,
2858 but report on the latest error recorded rather than the first one
2859 still in the error queue.
2860 [Ben Laurie, Bodo Moeller]
2862 *) default_algorithms option in ENGINE config module. This allows things
2864 default_algorithms = ALL
2865 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2868 *) Prelminary ENGINE config module.
2871 *) New experimental application configuration code.
2874 *) Change the AES code to follow the same name structure as all other
2875 symmetric ciphers, and behave the same way. Move everything to
2876 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2877 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2879 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2880 [Ben Laurie and Theo de Raadt]
2882 *) Add option to output public keys in req command.
2883 [Massimiliano Pala madwolf@openca.org]
2885 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2886 (up to about 10% better than before for P-192 and P-224).
2889 *) New functions/macros
2891 SSL_CTX_set_msg_callback(ctx, cb)
2892 SSL_CTX_set_msg_callback_arg(ctx, arg)
2893 SSL_set_msg_callback(ssl, cb)
2894 SSL_set_msg_callback_arg(ssl, arg)
2896 to request calling a callback function
2898 void cb(int write_p, int version, int content_type,
2899 const void *buf, size_t len, SSL *ssl, void *arg)
2901 whenever a protocol message has been completely received
2902 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2903 protocol version according to which the SSL library interprets
2904 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2905 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2906 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2907 specification (change_cipher_spec(20), alert(21), handshake(22)).
2908 'buf' and 'len' point to the actual message, 'ssl' to the
2909 SSL object, and 'arg' is the application-defined value set by
2910 SSL[_CTX]_set_msg_callback_arg().
2912 'openssl s_client' and 'openssl s_server' have new '-msg' options
2913 to enable a callback that displays all protocol messages.
2916 *) Change the shared library support so shared libraries are built as
2917 soon as the corresponding static library is finished, and thereby get
2918 openssl and the test programs linked against the shared library.
2919 This still only happens when the keyword "shard" has been given to
2920 the configuration scripts.
2922 NOTE: shared library support is still an experimental thing, and
2923 backward binary compatibility is still not guaranteed.
2924 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2926 *) Add support for Subject Information Access extension.
2927 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2929 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2930 additional bytes when new memory had to be allocated, not just
2931 when reusing an existing buffer.
2934 *) New command line and configuration option 'utf8' for the req command.
2935 This allows field values to be specified as UTF8 strings.
2938 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2939 runs for the former and machine-readable output for the latter.
2942 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2943 of the e-mail address in the DN (i.e., it will go into a certificate
2944 extension only). The new configuration file option 'email_in_dn = no'
2945 has the same effect.
2946 [Massimiliano Pala madwolf@openca.org]
2948 *) Change all functions with names starting with des_ to be starting
2949 with DES_ instead. Add wrappers that are compatible with libdes,
2950 but are named _ossl_old_des_*. Finally, add macros that map the
2951 des_* symbols to the corresponding _ossl_old_des_* if libdes
2952 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2953 desired, the des_* symbols will be mapped to DES_*, with one
2956 Since we provide two compatibility mappings, the user needs to
2957 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2958 compatibility is desired. The default (i.e., when that macro
2959 isn't defined) is OpenSSL 0.9.6c compatibility.
2961 There are also macros that enable and disable the support of old
2962 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2963 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2964 are defined, the default will apply: to support the old des routines.
2966 In either case, one must include openssl/des.h to get the correct
2967 definitions. Do not try to just include openssl/des_old.h, that
2970 NOTE: This is a major break of an old API into a new one. Software
2971 authors are encouraged to switch to the DES_ style functions. Some
2972 time in the future, des_old.h and the libdes compatibility functions
2973 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2974 default), and then completely removed.
2977 *) Test for certificates which contain unsupported critical extensions.
2978 If such a certificate is found during a verify operation it is
2979 rejected by default: this behaviour can be overridden by either
2980 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2981 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2982 X509_supported_extension() has also been added which returns 1 if a
2983 particular extension is supported.
2986 *) Modify the behaviour of EVP cipher functions in similar way to digests
2987 to retain compatibility with existing code.
2990 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2991 compatibility with existing code. In particular the 'ctx' parameter does
2992 not have to be to be initialized before the call to EVP_DigestInit() and
2993 it is tidied up after a call to EVP_DigestFinal(). New function
2994 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2995 EVP_MD_CTX_copy() changed to not require the destination to be
2996 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2997 requires the destination to be valid.
2999 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3000 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3003 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3004 so that complete 'Handshake' protocol structures are kept in memory
3005 instead of overwriting 'msg_type' and 'length' with 'body' data.
3008 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3009 [Massimo Santin via Richard Levitte]
3011 *) Major restructuring to the underlying ENGINE code. This includes
3012 reduction of linker bloat, separation of pure "ENGINE" manipulation
3013 (initialisation, etc) from functionality dealing with implementations
3014 of specific crypto iterfaces. This change also introduces integrated
3015 support for symmetric ciphers and digest implementations - so ENGINEs
3016 can now accelerate these by providing EVP_CIPHER and EVP_MD
3017 implementations of their own. This is detailed in crypto/engine/README
3018 as it couldn't be adequately described here. However, there are a few
3019 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3020 were changed in the original introduction of ENGINE code have now
3021 reverted back - the hooking from this code to ENGINE is now a good
3022 deal more passive and at run-time, operations deal directly with
3023 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3024 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3025 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3026 they were not being used by the framework as there is no concept of a
3027 BIGNUM_METHOD and they could not be generalised to the new
3028 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3029 ENGINE_cpy() has been removed as it cannot be consistently defined in
3033 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3036 *) Change mkdef.pl to sort symbols that get the same entry number,
3037 and make sure the automatically generated functions ERR_load_*
3038 become part of libeay.num as well.
3041 *) New function SSL_renegotiate_pending(). This returns true once
3042 renegotiation has been requested (either SSL_renegotiate() call
3043 or HelloRequest/ClientHello receveived from the peer) and becomes
3044 false once a handshake has been completed.
3045 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3046 sends a HelloRequest, but does not ensure that a handshake takes
3047 place. SSL_renegotiate_pending() is useful for checking if the
3048 client has followed the request.)
3051 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3052 By default, clients may request session resumption even during
3053 renegotiation (if session ID contexts permit); with this option,
3054 session resumption is possible only in the first handshake.
3056 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3057 more bits available for options that should not be part of
3058 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3061 *) Add some demos for certificate and certificate request creation.
3064 *) Make maximum certificate chain size accepted from the peer application
3065 settable (SSL*_get/set_max_cert_list()), as proposed by
3066 "Douglas E. Engert" <deengert@anl.gov>.
3069 *) Add support for shared libraries for Unixware-7
3070 (Boyd Lynn Gerber <gerberb@zenez.com>).
3073 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3074 be done prior to destruction. Use this to unload error strings from
3075 ENGINEs that load their own error strings. NB: This adds two new API
3076 functions to "get" and "set" this destroy handler in an ENGINE.
3079 *) Alter all existing ENGINE implementations (except "openssl" and
3080 "openbsd") to dynamically instantiate their own error strings. This
3081 makes them more flexible to be built both as statically-linked ENGINEs
3082 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3083 Also, add stub code to each that makes building them as self-contained
3084 shared-libraries easier (see README.ENGINE).
3087 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3088 implementations into applications that are completely implemented in
3089 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3090 commands that can be used to configure what shared-library to load and
3091 to control aspects of the way it is handled. Also, made an update to
3092 the README.ENGINE file that brings its information up-to-date and
3093 provides some information and instructions on the "dynamic" ENGINE
3094 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3097 *) Make it possible to unload ranges of ERR strings with a new
3098 "ERR_unload_strings" function.
3101 *) Add a copy() function to EVP_MD.
3104 *) Make EVP_MD routines take a context pointer instead of just the
3105 md_data void pointer.
3108 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3109 that the digest can only process a single chunk of data
3110 (typically because it is provided by a piece of
3111 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3112 is only going to provide a single chunk of data, and hence the
3113 framework needn't accumulate the data for oneshot drivers.
3116 *) As with "ERR", make it possible to replace the underlying "ex_data"
3117 functions. This change also alters the storage and management of global
3118 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3119 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3120 index counters. The API functions that use this state have been changed
3121 to take a "class_index" rather than pointers to the class's local STACK
3122 and counter, and there is now an API function to dynamically create new
3123 classes. This centralisation allows us to (a) plug a lot of the
3124 thread-safety problems that existed, and (b) makes it possible to clean
3125 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3126 such data would previously have always leaked in application code and
3127 workarounds were in place to make the memory debugging turn a blind eye
3128 to it. Application code that doesn't use this new function will still
3129 leak as before, but their memory debugging output will announce it now
3130 rather than letting it slide.
3132 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3133 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3134 has a return value to indicate success or failure.
3137 *) Make it possible to replace the underlying "ERR" functions such that the
3138 global state (2 LHASH tables and 2 locks) is only used by the "default"
3139 implementation. This change also adds two functions to "get" and "set"
3140 the implementation prior to it being automatically set the first time
3141 any other ERR function takes place. Ie. an application can call "get",
3142 pass the return value to a module it has just loaded, and that module
3143 can call its own "set" function using that value. This means the
3144 module's "ERR" operations will use (and modify) the error state in the
3145 application and not in its own statically linked copy of OpenSSL code.
3148 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3149 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3150 the operation, and provides a more encapsulated way for external code
3151 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3152 to use these functions rather than manually incrementing the counts.
3154 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3157 *) Add EVP test program.
3160 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3163 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3164 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3165 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3166 These allow a CRL to be built without having to access X509_CRL fields
3167 directly. Modify 'ca' application to use new functions.
3170 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3171 bug workarounds. Rollback attack detection is a security feature.
3172 The problem will only arise on OpenSSL servers when TLSv1 is not
3173 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3174 Software authors not wanting to support TLSv1 will have special reasons
3175 for their choice and can explicitly enable this option.
3176 [Bodo Moeller, Lutz Jaenicke]
3178 *) Rationalise EVP so it can be extended: don't include a union of
3179 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3180 (similar to those existing for EVP_CIPHER_CTX).
3185 EVP_MD_CTX_init(&md); /* new function call */
3186 EVP_DigestInit(&md, EVP_sha1());
3187 EVP_DigestUpdate(&md, in, len);
3188 EVP_DigestFinal(&md, out, NULL);
3189 EVP_MD_CTX_cleanup(&md); /* new function call */
3193 *) Make DES key schedule conform to the usual scheme, as well as
3194 correcting its structure. This means that calls to DES functions
3195 now have to pass a pointer to a des_key_schedule instead of a
3196 plain des_key_schedule (which was actually always a pointer
3199 des_key_schedule ks;
3201 des_set_key_checked(..., &ks);
3202 des_ncbc_encrypt(..., &ks, ...);
3204 (Note that a later change renames 'des_...' into 'DES_...'.)
3207 *) Initial reduction of linker bloat: the use of some functions, such as
3208 PEM causes large amounts of unused functions to be linked in due to
3209 poor organisation. For example pem_all.c contains every PEM function
3210 which has a knock on effect of linking in large amounts of (unused)
3211 ASN1 code. Grouping together similar functions and splitting unrelated
3212 functions prevents this.
3215 *) Cleanup of EVP macros.
3218 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3219 correct _ecb suffix.
3222 *) Add initial OCSP responder support to ocsp application. The
3223 revocation information is handled using the text based index
3224 use by the ca application. The responder can either handle
3225 requests generated internally, supplied in files (for example
3226 via a CGI script) or using an internal minimal server.
3229 *) Add configuration choices to get zlib compression for TLS.
3232 *) Changes to Kerberos SSL for RFC 2712 compliance:
3233 1. Implemented real KerberosWrapper, instead of just using
3234 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3235 2. Implemented optional authenticator field of KerberosWrapper.
3237 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3238 and authenticator structs; see crypto/krb5/.
3240 Generalized Kerberos calls to support multiple Kerberos libraries.
3241 [Vern Staats <staatsvr@asc.hpc.mil>,
3242 Jeffrey Altman <jaltman@columbia.edu>
3243 via Richard Levitte]
3245 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3246 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3247 values for each of the key sizes rather than having just
3248 parameters (and 'speed' generating keys each time).
3251 *) Speed up EVP routines.
3254 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3255 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3256 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3257 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3259 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3260 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3261 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3264 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3266 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3269 *) Added the OS2-EMX target.
3270 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3272 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3273 to support NCONF routines in extension code. New function CONF_set_nconf()
3274 to allow functions which take an NCONF to also handle the old LHASH
3275 structure: this means that the old CONF compatible routines can be
3276 retained (in particular wrt extensions) without having to duplicate the
3277 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3280 *) Enhance the general user interface with mechanisms for inner control
3281 and with possibilities to have yes/no kind of prompts.
3284 *) Change all calls to low level digest routines in the library and
3285 applications to use EVP. Add missing calls to HMAC_cleanup() and
3286 don't assume HMAC_CTX can be copied using memcpy().
3287 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3289 *) Add the possibility to control engines through control names but with
3290 arbitrary arguments instead of just a string.
3291 Change the key loaders to take a UI_METHOD instead of a callback
3292 function pointer. NOTE: this breaks binary compatibility with earlier
3293 versions of OpenSSL [engine].
3294 Adapt the nCipher code for these new conditions and add a card insertion
3298 *) Enhance the general user interface with mechanisms to better support
3299 dialog box interfaces, application-defined prompts, the possibility
3300 to use defaults (for example default passwords from somewhere else)
3301 and interrupts/cancellations.
3304 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3305 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3308 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3309 tidy up some unnecessarily weird code in 'sk_new()').
3310 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3312 *) Change the key loading routines for ENGINEs to use the same kind
3313 callback (pem_password_cb) as all other routines that need this
3317 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3318 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3319 than this minimum value is recommended.
3322 *) New random seeder for OpenVMS, using the system process statistics
3323 that are easily reachable.
3326 *) Windows apparently can't transparently handle global
3327 variables defined in DLLs. Initialisations such as:
3329 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3331 wont compile. This is used by the any applications that need to
3332 declare their own ASN1 modules. This was fixed by adding the option
3333 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3334 needed for static libraries under Win32.
3337 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3338 setting of purpose and trust fields. New X509_STORE trust and
3339 purpose functions and tidy up setting in other SSL functions.
3342 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3343 structure. These are inherited by X509_STORE_CTX when it is
3344 initialised. This allows various defaults to be set in the
3345 X509_STORE structure (such as flags for CRL checking and custom
3346 purpose or trust settings) for functions which only use X509_STORE_CTX
3347 internally such as S/MIME.
3349 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3350 trust settings if they are not set in X509_STORE. This allows X509_STORE
3351 purposes and trust (in S/MIME for example) to override any set by default.
3353 Add command line options for CRL checking to smime, s_client and s_server
3357 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3358 are set then the CRL is looked up in the X509_STORE structure and
3359 its validity and signature checked, then if the certificate is found
3360 in the CRL the verify fails with a revoked error.
3362 Various new CRL related callbacks added to X509_STORE_CTX structure.
3364 Command line options added to 'verify' application to support this.
3366 This needs some additional work, such as being able to handle multiple
3367 CRLs with different times, extension based lookup (rather than just
3368 by subject name) and ultimately more complete V2 CRL extension
3372 *) Add a general user interface API (crypto/ui/). This is designed
3373 to replace things like des_read_password and friends (backward
3374 compatibility functions using this new API are provided).
3375 The purpose is to remove prompting functions from the DES code
3376 section as well as provide for prompting through dialog boxes in
3377 a window system and the like.
3380 *) Add "ex_data" support to ENGINE so implementations can add state at a
3381 per-structure level rather than having to store it globally.
3384 *) Make it possible for ENGINE structures to be copied when retrieved by
3385 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3386 This causes the "original" ENGINE structure to act like a template,
3387 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3388 operational state can be localised to each ENGINE structure, despite the
3389 fact they all share the same "methods". New ENGINE structures returned in
3390 this case have no functional references and the return value is the single
3391 structural reference. This matches the single structural reference returned
3392 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3396 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3397 needs to match any other type at all we need to manually clear the
3401 *) Changes to the "openssl engine" utility to include;
3402 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3403 about an ENGINE's available control commands.
3404 - executing control commands from command line arguments using the
3405 '-pre' and '-post' switches. '-post' is only used if '-t' is
3406 specified and the ENGINE is successfully initialised. The syntax for
3407 the individual commands are colon-separated, for example;
3408 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3411 *) New dynamic control command support for ENGINEs. ENGINEs can now
3412 declare their own commands (numbers), names (strings), descriptions,
3413 and input types for run-time discovery by calling applications. A
3414 subset of these commands are implicitly classed as "executable"
3415 depending on their input type, and only these can be invoked through
3416 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3417 can be based on user input, config files, etc). The distinction is
3418 that "executable" commands cannot return anything other than a boolean
3419 result and can only support numeric or string input, whereas some
3420 discoverable commands may only be for direct use through
3421 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3422 pointers, or other custom uses. The "executable" commands are to
3423 support parameterisations of ENGINE behaviour that can be
3424 unambiguously defined by ENGINEs and used consistently across any
3425 OpenSSL-based application. Commands have been added to all the
3426 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3427 control over shared-library paths without source code alterations.
3430 *) Changed all ENGINE implementations to dynamically allocate their
3431 ENGINEs rather than declaring them statically. Apart from this being
3432 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3433 this also allows the implementations to compile without using the
3434 internal engine_int.h header.
3437 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3438 'const' value. Any code that should be able to modify a RAND_METHOD
3439 should already have non-const pointers to it (ie. they should only
3440 modify their own ones).
3443 *) Made a variety of little tweaks to the ENGINE code.
3444 - "atalla" and "ubsec" string definitions were moved from header files
3445 to C code. "nuron" string definitions were placed in variables
3446 rather than hard-coded - allowing parameterisation of these values
3447 later on via ctrl() commands.
3448 - Removed unused "#if 0"'d code.
3449 - Fixed engine list iteration code so it uses ENGINE_free() to release
3450 structural references.
3451 - Constified the RAND_METHOD element of ENGINE structures.
3452 - Constified various get/set functions as appropriate and added
3453 missing functions (including a catch-all ENGINE_cpy that duplicates
3454 all ENGINE values onto a new ENGINE except reference counts/state).
3455 - Removed NULL parameter checks in get/set functions. Setting a method
3456 or function to NULL is a way of cancelling out a previously set
3457 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3458 and doesn't justify the extra error symbols and code.
3459 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3460 flags from engine_int.h to engine.h.
3461 - Changed prototypes for ENGINE handler functions (init(), finish(),
3462 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3465 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3466 to the algorithm using long division. The binary algorithm can be
3467 used only if the modulus is odd. On 32-bit systems, it is faster
3468 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3469 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3470 up to 450 bits. In 64-bit environments, the binary algorithm
3471 appears to be advantageous for much longer moduli; here we use it
3472 for moduli up to 2048 bits.
3475 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3476 could not support the combine flag in choice fields.
3479 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3480 extensions from a certificate request to the certificate.
3483 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3484 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3485 file: this allows the display of the certificate about to be
3486 signed to be customised, to allow certain fields to be included
3487 or excluded and extension details. The old system didn't display
3488 multicharacter strings properly, omitted fields not in the policy
3489 and couldn't display additional details such as extensions.
3492 *) Function EC_POINTs_mul for multiple scalar multiplication
3493 of an arbitrary number of elliptic curve points
3494 \sum scalars[i]*points[i],
3495 optionally including the generator defined for the EC_GROUP:
3496 scalar*generator + \sum scalars[i]*points[i].
3498 EC_POINT_mul is a simple wrapper function for the typical case
3499 that the point list has just one item (besides the optional
3503 *) First EC_METHODs for curves over GF(p):
3505 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3506 operations and provides various method functions that can also
3507 operate with faster implementations of modular arithmetic.
3509 EC_GFp_mont_method() reuses most functions that are part of
3510 EC_GFp_simple_method, but uses Montgomery arithmetic.
3512 [Bodo Moeller; point addition and point doubling
3513 implementation directly derived from source code provided by
3514 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3516 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3517 crypto/ec/ec_lib.c):
3519 Curves are EC_GROUP objects (with an optional group generator)
3520 based on EC_METHODs that are built into the library.
3522 Points are EC_POINT objects based on EC_GROUP objects.
3524 Most of the framework would be able to handle curves over arbitrary
3525 finite fields, but as there are no obvious types for fields other
3526 than GF(p), some functions are limited to that for now.
3529 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3530 that the file contains a complete HTTP response.
3533 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3534 change the def and num file printf format specifier from "%-40sXXX"
3535 to "%-39s XXX". The latter will always guarantee a space after the
3536 field while the former will cause them to run together if the field
3537 is 40 of more characters long.
3540 *) Constify the cipher and digest 'method' functions and structures
3541 and modify related functions to take constant EVP_MD and EVP_CIPHER
3545 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3546 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3549 *) Modify EVP_Digest*() routines so they now return values. Although the
3550 internal software routines can never fail additional hardware versions
3554 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3556 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3557 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3560 ERR_R_NESTED_ASN1_ERROR
3562 ERR_R_MISSING_ASN1_EOS
3563 were 4 .. 9, conflicting with
3564 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3566 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3567 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3569 Add new error code 'ERR_R_INTERNAL_ERROR'.
3572 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3576 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3577 sets the subject name for a new request or supersedes the
3578 subject name in a given request. Formats that can be parsed are
3579 'CN=Some Name, OU=myOU, C=IT'
3581 'CN=Some Name/OU=myOU/C=IT'.
3583 Add options '-batch' and '-verbose' to 'openssl req'.
3584 [Massimiliano Pala <madwolf@hackmasters.net>]
3586 *) Introduce the possibility to access global variables through
3587 functions on platform were that's the best way to handle exporting
3588 global variables in shared libraries. To enable this functionality,
3589 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3590 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3591 is normally done by Configure or something similar).
3593 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3594 in the source file (foo.c) like this:
3596 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3597 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3599 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3600 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3602 OPENSSL_DECLARE_GLOBAL(int,foo);
3603 #define foo OPENSSL_GLOBAL_REF(foo)
3604 OPENSSL_DECLARE_GLOBAL(double,bar);
3605 #define bar OPENSSL_GLOBAL_REF(bar)
3607 The #defines are very important, and therefore so is including the
3608 header file everywhere where the defined globals are used.
3610 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3611 of ASN.1 items, but that structure is a bit different.
3613 The largest change is in util/mkdef.pl which has been enhanced with
3614 better and easier to understand logic to choose which symbols should
3615 go into the Windows .def files as well as a number of fixes and code
3616 cleanup (among others, algorithm keywords are now sorted
3617 lexicographically to avoid constant rewrites).
3620 *) In BN_div() keep a copy of the sign of 'num' before writing the
3621 result to 'rm' because if rm==num the value will be overwritten
3622 and produce the wrong result if 'num' is negative: this caused
3623 problems with BN_mod() and BN_nnmod().
3626 *) Function OCSP_request_verify(). This checks the signature on an
3627 OCSP request and verifies the signer certificate. The signer
3628 certificate is just checked for a generic purpose and OCSP request
3632 *) Add OCSP_check_validity() function to check the validity of OCSP
3633 responses. OCSP responses are prepared in real time and may only
3634 be a few seconds old. Simply checking that the current time lies
3635 between thisUpdate and nextUpdate max reject otherwise valid responses
3636 caused by either OCSP responder or client clock inaccuracy. Instead
3637 we allow thisUpdate and nextUpdate to fall within a certain period of
3638 the current time. The age of the response can also optionally be
3639 checked. Two new options -validity_period and -status_age added to
3643 *) If signature or public key algorithm is unrecognized print out its
3644 OID rather that just UNKNOWN.
3647 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3648 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3649 ID to be generated from the issuer certificate alone which can then be
3650 passed to OCSP_id_issuer_cmp().
3653 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3654 ASN1 modules to export functions returning ASN1_ITEM pointers
3655 instead of the ASN1_ITEM structures themselves. This adds several
3656 new macros which allow the underlying ASN1 function/structure to
3657 be accessed transparently. As a result code should not use ASN1_ITEM
3658 references directly (such as &X509_it) but instead use the relevant
3659 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3660 use of the new ASN1 code on platforms where exporting structures
3661 is problematical (for example in shared libraries) but exporting
3662 functions returning pointers to structures is not.
3665 *) Add support for overriding the generation of SSL/TLS session IDs.
3666 These callbacks can be registered either in an SSL_CTX or per SSL.
3667 The purpose of this is to allow applications to control, if they wish,
3668 the arbitrary values chosen for use as session IDs, particularly as it
3669 can be useful for session caching in multiple-server environments. A
3670 command-line switch for testing this (and any client code that wishes
3671 to use such a feature) has been added to "s_server".
3672 [Geoff Thorpe, Lutz Jaenicke]
3674 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3675 of the form '#if defined(...) || defined(...) || ...' and
3676 '#if !defined(...) && !defined(...) && ...'. This also avoids
3677 the growing number of special cases it was previously handling.
3680 *) Make all configuration macros available for application by making
3681 sure they are available in opensslconf.h, by giving them names starting
3682 with "OPENSSL_" to avoid conflicts with other packages and by making
3683 sure e_os2.h will cover all platform-specific cases together with
3685 Additionally, it is now possible to define configuration/platform-
3686 specific names (called "system identities"). In the C code, these
3687 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3688 macro with the name beginning with "OPENSSL_SYS_", which is determined
3689 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3693 *) New option -set_serial to 'req' and 'x509' this allows the serial
3694 number to use to be specified on the command line. Previously self
3695 signed certificates were hard coded with serial number 0 and the
3696 CA options of 'x509' had to use a serial number in a file which was
3700 *) New options to 'ca' utility to support V2 CRL entry extensions.
3701 Currently CRL reason, invalidity date and hold instruction are
3702 supported. Add new CRL extensions to V3 code and some new objects.
3705 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3706 disable standard block padding (aka PKCS#5 padding) in the EVP
3707 API, which was previously mandatory. This means that the data is
3708 not padded in any way and so the total length much be a multiple
3709 of the block size, otherwise an error occurs.
3712 *) Initial (incomplete) OCSP SSL support.
3715 *) New function OCSP_parse_url(). This splits up a URL into its host,
3716 port and path components: primarily to parse OCSP URLs. New -url
3717 option to ocsp utility.
3720 *) New nonce behavior. The return value of OCSP_check_nonce() now
3721 reflects the various checks performed. Applications can decide
3722 whether to tolerate certain situations such as an absent nonce
3723 in a response when one was present in a request: the ocsp application
3724 just prints out a warning. New function OCSP_add1_basic_nonce()
3725 this is to allow responders to include a nonce in a response even if
3726 the request is nonce-less.
3729 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3730 skipped when using openssl x509 multiple times on a single input file,
3731 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3734 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3735 set string type: to handle setting ASN1_TIME structures. Fix ca
3736 utility to correctly initialize revocation date of CRLs.
3739 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3740 the clients preferred ciphersuites and rather use its own preferences.
3741 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3742 Internet Explorer by ensuring unchanged hash method during stepup.
3743 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3746 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3747 to aes and add a new 'exist' option to print out symbols that don't
3751 *) Additional options to ocsp utility to allow flags to be set and
3752 additional certificates supplied.
3755 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3756 OCSP client a number of certificate to only verify the response
3760 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3761 handle the new API. Currently only ECB, CBC modes supported. Add new
3764 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3765 Encryption Standard (AES) Ciphersuites for Transport Layer
3766 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3767 not enabled by default and were not part of the "ALL" ciphersuite
3768 alias because they were not yet official; they could be
3769 explicitly requested by specifying the "AESdraft" ciphersuite
3770 group alias. In the final release of OpenSSL 0.9.7, the group
3771 alias is called "AES" and is part of "ALL".)
3772 [Ben Laurie, Steve Henson, Bodo Moeller]
3774 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3775 request to response.
3778 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3779 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3780 extract information from a certificate request. OCSP_response_create()
3781 creates a response and optionally adds a basic response structure.
3782 OCSP_basic_add1_status() adds a complete single response to a basic
3783 response and returns the OCSP_SINGLERESP structure just added (to allow
3784 extensions to be included for example). OCSP_basic_add1_cert() adds a
3785 certificate to a basic response and OCSP_basic_sign() signs a basic
3786 response with various flags. New helper functions ASN1_TIME_check()
3787 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3788 (converts ASN1_TIME to GeneralizedTime).
3791 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3792 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3793 structure from a certificate. X509_pubkey_digest() digests the public_key
3794 contents: this is used in various key identifiers.
3797 *) Make sk_sort() tolerate a NULL argument.
3798 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3800 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3801 passed by the function are trusted implicitly. If any of them signed the
3802 response then it is assumed to be valid and is not verified.
3805 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3806 to data. This was previously part of the PKCS7 ASN1 code. This
3807 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3808 [Steve Henson, reported by Kenneth R. Robinette
3809 <support@securenetterm.com>]
3811 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3812 routines: without these tracing memory leaks is very painful.
3813 Fix leaks in PKCS12 and PKCS7 routines.
3816 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3817 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3818 effectively meant GeneralizedTime would never be used. Now it
3819 is initialised to -1 but X509_time_adj() now has to check the value
3820 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3821 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3822 [Steve Henson, reported by Kenneth R. Robinette
3823 <support@securenetterm.com>]
3825 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3826 result in a zero length in the ASN1_INTEGER structure which was
3827 not consistent with the structure when d2i_ASN1_INTEGER() was used
3828 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3829 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3830 where it did not print out a minus for negative ASN1_INTEGER.
3833 *) Add summary printout to ocsp utility. The various functions which
3834 convert status values to strings have been renamed to:
3835 OCSP_response_status_str(), OCSP_cert_status_str() and
3836 OCSP_crl_reason_str() and are no longer static. New options
3837 to verify nonce values and to disable verification. OCSP response
3838 printout format cleaned up.
3841 *) Add additional OCSP certificate checks. These are those specified
3842 in RFC2560. This consists of two separate checks: the CA of the
3843 certificate being checked must either be the OCSP signer certificate
3844 or the issuer of the OCSP signer certificate. In the latter case the
3845 OCSP signer certificate must contain the OCSP signing extended key
3846 usage. This check is performed by attempting to match the OCSP
3847 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3848 in the OCSP_CERTID structures of the response.
3851 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3852 and related routines. This uses the standard OpenSSL certificate
3853 verify routines to perform initial checks (just CA validity) and
3854 to obtain the certificate chain. Then additional checks will be
3855 performed on the chain. Currently the root CA is checked to see
3856 if it is explicitly trusted for OCSP signing. This is used to set
3857 a root CA as a global signing root: that is any certificate that
3858 chains to that CA is an acceptable OCSP signing certificate.
3861 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3862 extensions from a separate configuration file.
3863 As when reading extensions from the main configuration file,
3864 the '-extensions ...' option may be used for specifying the
3866 [Massimiliano Pala <madwolf@comune.modena.it>]
3868 *) New OCSP utility. Allows OCSP requests to be generated or
3869 read. The request can be sent to a responder and the output
3870 parsed, outputed or printed in text form. Not complete yet:
3871 still needs to check the OCSP response validity.
3874 *) New subcommands for 'openssl ca':
3875 'openssl ca -status <serial>' prints the status of the cert with
3876 the given serial number (according to the index file).
3877 'openssl ca -updatedb' updates the expiry status of certificates
3879 [Massimiliano Pala <madwolf@comune.modena.it>]
3881 *) New '-newreq-nodes' command option to CA.pl. This is like
3882 '-newreq', but calls 'openssl req' with the '-nodes' option
3883 so that the resulting key is not encrypted.
3884 [Damien Miller <djm@mindrot.org>]
3886 *) New configuration for the GNU Hurd.
3887 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3889 *) Initial code to implement OCSP basic response verify. This
3890 is currently incomplete. Currently just finds the signer's
3891 certificate and verifies the signature on the response.
3894 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3895 value of OPENSSLDIR. This is available via the new '-d' option
3896 to 'openssl version', and is also included in 'openssl version -a'.
3899 *) Allowing defining memory allocation callbacks that will be given
3900 file name and line number information in additional arguments
3901 (a const char* and an int). The basic functionality remains, as
3902 well as the original possibility to just replace malloc(),
3903 realloc() and free() by functions that do not know about these
3904 additional arguments. To register and find out the current
3905 settings for extended allocation functions, the following
3906 functions are provided:
3908 CRYPTO_set_mem_ex_functions
3909 CRYPTO_set_locked_mem_ex_functions
3910 CRYPTO_get_mem_ex_functions
3911 CRYPTO_get_locked_mem_ex_functions
3913 These work the same way as CRYPTO_set_mem_functions and friends.
3914 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3915 extended allocation function is enabled.
3916 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3917 a conventional allocation function is enabled.
3918 [Richard Levitte, Bodo Moeller]
3920 *) Finish off removing the remaining LHASH function pointer casts.
3921 There should no longer be any prototype-casting required when using
3922 the LHASH abstraction, and any casts that remain are "bugs". See
3923 the callback types and macros at the head of lhash.h for details
3924 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3927 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3928 If /dev/[u]random devices are not available or do not return enough
3929 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3931 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3932 /etc/entropy will be queried once each in this sequence, quering stops
3933 when enough entropy was collected without querying more sockets.
3936 *) Change the Unix RAND_poll() variant to be able to poll several
3937 random devices, as specified by DEVRANDOM, until a sufficient amount
3938 of data has been collected. We spend at most 10 ms on each file
3939 (select timeout) and read in non-blocking mode. DEVRANDOM now
3940 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3941 (previously it was just the string "/dev/urandom"), so on typical
3942 platforms the 10 ms delay will never occur.
3943 Also separate out the Unix variant to its own file, rand_unix.c.
3944 For VMS, there's a currently-empty rand_vms.c.
3947 *) Move OCSP client related routines to ocsp_cl.c. These
3948 provide utility functions which an application needing
3949 to issue a request to an OCSP responder and analyse the
3950 response will typically need: as opposed to those which an
3951 OCSP responder itself would need which will be added later.
3953 OCSP_request_sign() signs an OCSP request with an API similar
3954 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3955 response. OCSP_response_get1_basic() extracts basic response
3956 from response. OCSP_resp_find_status(): finds and extracts status
3957 information from an OCSP_CERTID structure (which will be created
3958 when the request structure is built). These are built from lower
3959 level functions which work on OCSP_SINGLERESP structures but
3960 wont normally be used unless the application wishes to examine
3961 extensions in the OCSP response for example.
3963 Replace nonce routines with a pair of functions.
3964 OCSP_request_add1_nonce() adds a nonce value and optionally
3965 generates a random value. OCSP_check_nonce() checks the
3966 validity of the nonce in an OCSP response.
3969 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3970 This doesn't copy the supplied OCSP_CERTID and avoids the
3971 need to free up the newly created id. Change return type
3972 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3973 This can then be used to add extensions to the request.
3974 Deleted OCSP_request_new(), since most of its functionality
3975 is now in OCSP_REQUEST_new() (and the case insensitive name
3976 clash) apart from the ability to set the request name which
3977 will be added elsewhere.
3980 *) Update OCSP API. Remove obsolete extensions argument from
3981 various functions. Extensions are now handled using the new
3982 OCSP extension code. New simple OCSP HTTP function which
3983 can be used to send requests and parse the response.
3986 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3987 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3988 uses the special reorder version of SET OF to sort the attributes
3989 and reorder them to match the encoded order. This resolves a long
3990 standing problem: a verify on a PKCS7 structure just after signing
3991 it used to fail because the attribute order did not match the
3992 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3993 it uses the received order. This is necessary to tolerate some broken
3994 software that does not order SET OF. This is handled by encoding
3995 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3996 to produce the required SET OF.
3999 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4000 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4001 files to get correct declarations of the ASN.1 item variables.
4004 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4005 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4006 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4007 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4008 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4009 ASN1_ITEM and no wrapper functions.
4012 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4013 replace the old function pointer based I/O routines. Change most of
4014 the *_d2i_bio() and *_d2i_fp() functions to use these.
4017 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4018 lines, recognice more "algorithms" that can be deselected, and make
4019 it complain about algorithm deselection that isn't recognised.
4022 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4023 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4024 to use new functions. Add NO_ASN1_OLD which can be set to remove
4025 some old style ASN1 functions: this can be used to determine if old
4026 code will still work when these eventually go away.
4029 *) New extension functions for OCSP structures, these follow the
4030 same conventions as certificates and CRLs.
4033 *) New function X509V3_add1_i2d(). This automatically encodes and
4034 adds an extension. Its behaviour can be customised with various
4035 flags to append, replace or delete. Various wrappers added for
4036 certifcates and CRLs.
4039 *) Fix to avoid calling the underlying ASN1 print routine when
4040 an extension cannot be parsed. Correct a typo in the
4041 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4044 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4045 entries for variables.
4048 *) Add functionality to apps/openssl.c for detecting locking
4049 problems: As the program is single-threaded, all we have
4050 to do is register a locking callback using an array for
4051 storing which locks are currently held by the program.
4054 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4055 SSL_get_ex_data_X509_STORE_idx(), which is used in
4056 ssl_verify_cert_chain() and thus can be called at any time
4057 during TLS/SSL handshakes so that thread-safety is essential.
4058 Unfortunately, the ex_data design is not at all suited
4059 for multi-threaded use, so it probably should be abolished.
4062 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4063 [Broadcom, tweaked and integrated by Geoff Thorpe]
4065 *) Move common extension printing code to new function
4066 X509V3_print_extensions(). Reorganise OCSP print routines and
4067 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4070 *) New function X509_signature_print() to remove duplication in some
4074 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4075 set (this was treated exactly the same as SET OF previously). This
4076 is used to reorder the STACK representing the structure to match the
4077 encoding. This will be used to get round a problem where a PKCS7
4078 structure which was signed could not be verified because the STACK
4079 order did not reflect the encoded order.
4082 *) Reimplement the OCSP ASN1 module using the new code.
4085 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4086 for its ASN1 operations. The old style function pointers still exist
4087 for now but they will eventually go away.
4090 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4091 completely replaces the old ASN1 functionality with a table driven
4092 encoder and decoder which interprets an ASN1_ITEM structure describing
4093 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4094 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4095 has also been converted to the new form.
4098 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4099 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4100 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4101 for negative moduli.
4104 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4105 of not touching the result's sign bit.
4108 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4112 *) Changed the LHASH code to use prototypes for callbacks, and created
4113 macros to declare and implement thin (optionally static) functions
4114 that provide type-safety and avoid function pointer casting for the
4115 type-specific callbacks.
4118 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4120 [Veers Staats <staatsvr@asc.hpc.mil>,
4121 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4123 *) Reformat the FAQ so the different questions and answers can be divided
4124 in sections depending on the subject.
4127 *) Have the zlib compression code load ZLIB.DLL dynamically under
4131 *) New function BN_mod_sqrt for computing square roots modulo a prime
4132 (using the probabilistic Tonelli-Shanks algorithm unless
4133 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4134 be handled deterministically).
4135 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4137 *) Make BN_mod_inverse faster by explicitly handling small quotients
4138 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4139 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4142 *) New function BN_kronecker.
4145 *) Fix BN_gcd so that it works on negative inputs; the result is
4146 positive unless both parameters are zero.
4147 Previously something reasonably close to an infinite loop was
4148 possible because numbers could be growing instead of shrinking
4149 in the implementation of Euclid's algorithm.
4152 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4153 sign of the number in question.
4155 Fix BN_is_word(a,w) to work correctly for w == 0.
4157 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4158 because its test if the absolute value of 'a' equals 'w'.
4159 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4160 it exists mostly for use in the implementations of BN_is_zero(),
4161 BN_is_one(), and BN_is_word().
4164 *) New function BN_swap.
4167 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4168 the exponentiation functions are more likely to produce reasonable
4169 results on negative inputs.
4172 *) Change BN_mod_mul so that the result is always non-negative.
4173 Previously, it could be negative if one of the factors was negative;
4174 I don't think anyone really wanted that behaviour.
4177 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4178 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4179 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4180 and add new functions:
4189 BN_mod_lshift1_quick
4193 These functions always generate non-negative results.
4195 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4196 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4198 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4199 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4200 be reduced modulo m.
4201 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4204 The following entry accidentily appeared in the CHANGES file
4205 distributed with OpenSSL 0.9.7. The modifications described in
4206 it do *not* apply to OpenSSL 0.9.7.
4208 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4209 was actually never needed) and in BN_mul(). The removal in BN_mul()
4210 required a small change in bn_mul_part_recursive() and the addition
4211 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4212 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4213 bn_sub_words() and bn_add_words() except they take arrays with
4218 *) In 'openssl passwd', verify passwords read from the terminal
4219 unless the '-salt' option is used (which usually means that
4220 verification would just waste user's time since the resulting
4221 hash is going to be compared with some given password hash)
4222 or the new '-noverify' option is used.
4224 This is an incompatible change, but it does not affect
4225 non-interactive use of 'openssl passwd' (passwords on the command
4226 line, '-stdin' option, '-in ...' option) and thus should not
4230 *) Remove all references to RSAref, since there's no more need for it.
4233 *) Make DSO load along a path given through an environment variable
4234 (SHLIB_PATH) with shl_load().
4237 *) Constify the ENGINE code as a result of BIGNUM constification.
4238 Also constify the RSA code and most things related to it. In a
4239 few places, most notable in the depth of the ASN.1 code, ugly
4240 casts back to non-const were required (to be solved at a later
4244 *) Make it so the openssl application has all engines loaded by default.
4247 *) Constify the BIGNUM routines a little more.
4250 *) Add the following functions:
4252 ENGINE_load_cswift()
4254 ENGINE_load_atalla()
4256 ENGINE_load_builtin_engines()
4258 That way, an application can itself choose if external engines that
4259 are built-in in OpenSSL shall ever be used or not. The benefit is
4260 that applications won't have to be linked with libdl or other dso
4261 libraries unless it's really needed.
4263 Changed 'openssl engine' to load all engines on demand.
4264 Changed the engine header files to avoid the duplication of some
4265 declarations (they differed!).
4268 *) 'openssl engine' can now list capabilities.
4271 *) Better error reporting in 'openssl engine'.
4274 *) Never call load_dh_param(NULL) in s_server.
4277 *) Add engine application. It can currently list engines by name and
4278 identity, and test if they are actually available.
4281 *) Improve RPM specification file by forcing symbolic linking and making
4282 sure the installed documentation is also owned by root.root.
4283 [Damien Miller <djm@mindrot.org>]
4285 *) Give the OpenSSL applications more possibilities to make use of
4286 keys (public as well as private) handled by engines.
4289 *) Add OCSP code that comes from CertCo.
4292 *) Add VMS support for the Rijndael code.
4295 *) Added untested support for Nuron crypto accelerator.
4298 *) Add support for external cryptographic devices. This code was
4299 previously distributed separately as the "engine" branch.
4300 [Geoff Thorpe, Richard Levitte]
4302 *) Rework the filename-translation in the DSO code. It is now possible to
4303 have far greater control over how a "name" is turned into a filename
4304 depending on the operating environment and any oddities about the
4305 different shared library filenames on each system.
4308 *) Support threads on FreeBSD-elf in Configure.
4311 *) Fix for SHA1 assembly problem with MASM: it produces
4312 warnings about corrupt line number information when assembling
4313 with debugging information. This is caused by the overlapping
4315 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4318 NCONF_get_number() has no error checking at all. As a replacement,
4319 NCONF_get_number_e() is defined (_e for "error checking") and is
4320 promoted strongly. The old NCONF_get_number is kept around for
4321 binary backward compatibility.
4322 Make it possible for methods to load from something other than a BIO,
4323 by providing a function pointer that is given a name instead of a BIO.
4324 For example, this could be used to load configuration data from an
4328 *) Fix for non blocking accept BIOs. Added new I/O special reason
4329 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4330 with non blocking I/O was not possible because no retry code was
4331 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4335 *) Added the beginnings of Rijndael support.
4338 *) Fix for bug in DirectoryString mask setting. Add support for
4339 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4340 to allow certificate printing to more controllable, additional
4341 'certopt' option to 'x509' to allow new printing options to be
4345 *) Clean old EAY MD5 hack from e_os.h.
4348 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4350 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4351 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4352 [Joe Orton, Steve Henson]
4354 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4356 *) Fix additional bug revealed by the NISCC test suite:
4358 Stop bug triggering large recursion when presented with
4359 certain ASN.1 tags (CVE-2003-0851)
4362 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4364 *) Fix various bugs revealed by running the NISCC test suite:
4366 Stop out of bounds reads in the ASN1 code when presented with
4367 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4369 If verify callback ignores invalid public key errors don't try to check
4370 certificate signature with the NULL public key.
4374 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4375 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4379 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4380 extra data after the compression methods not only for TLS 1.0
4381 but also for SSL 3.0 (as required by the specification).
4382 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4384 *) Change X509_certificate_type() to mark the key as exported/exportable
4385 when it's 512 *bits* long, not 512 bytes.
4388 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4390 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4391 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4392 a protocol version number mismatch like a decryption error
4393 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4396 *) Turn on RSA blinding by default in the default implementation
4397 to avoid a timing attack. Applications that don't want it can call
4398 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4399 They would be ill-advised to do so in most cases.
4400 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4402 *) Change RSA blinding code so that it works when the PRNG is not
4403 seeded (in this case, the secret RSA exponent is abused as
4404 an unpredictable seed -- if it is not unpredictable, there
4405 is no point in blinding anyway). Make RSA blinding thread-safe
4406 by remembering the creator's thread ID in rsa->blinding and
4407 having all other threads use local one-time blinding factors
4408 (this requires more computation than sharing rsa->blinding, but
4409 avoids excessive locking; and if an RSA object is not shared
4410 between threads, blinding will still be very fast).
4413 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4415 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4416 via timing by performing a MAC computation even if incorrrect
4417 block cipher padding has been found. This is a countermeasure
4418 against active attacks where the attacker has to distinguish
4419 between bad padding and a MAC verification error. (CVE-2003-0078)
4421 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4422 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4423 Martin Vuagnoux (EPFL, Ilion)]
4425 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4427 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4428 memory from it's contents. This is done with a counter that will
4429 place alternating values in each byte. This can be used to solve
4430 two issues: 1) the removal of calls to memset() by highly optimizing
4431 compilers, and 2) cleansing with other values than 0, since those can
4432 be read through on certain media, for example a swap space on disk.
4435 *) Bugfix: client side session caching did not work with external caching,
4436 because the session->cipher setting was not restored when reloading
4437 from the external cache. This problem was masked, when
4438 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4439 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4442 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4443 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4444 [Zeev Lieber <zeev-l@yahoo.com>]
4446 *) Undo an undocumented change introduced in 0.9.6e which caused
4447 repeated calls to OpenSSL_add_all_ciphers() and
4448 OpenSSL_add_all_digests() to be ignored, even after calling
4452 *) Change the default configuration reader to deal with last line not
4453 being properly terminated.
4456 *) Change X509_NAME_cmp() so it applies the special rules on handling
4457 DN values that are of type PrintableString, as well as RDNs of type
4458 emailAddress where the value has the type ia5String.
4459 [stefank@valicert.com via Richard Levitte]
4461 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4462 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4463 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4464 the bitwise-OR of the two for use by the majority of applications
4465 wanting this behaviour, and update the docs. The documented
4466 behaviour and actual behaviour were inconsistent and had been
4467 changing anyway, so this is more a bug-fix than a behavioural
4469 [Geoff Thorpe, diagnosed by Nadav Har'El]
4471 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4472 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4475 *) Fix initialization code race conditions in
4476 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4477 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4478 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4479 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4480 ssl2_get_cipher_by_char(),
4481 ssl3_get_cipher_by_char().
4482 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4484 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4485 the cached sessions are flushed, as the remove_cb() might use ex_data
4486 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4487 (see [openssl.org #212]).
4488 [Geoff Thorpe, Lutz Jaenicke]
4490 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4491 length, instead of the encoding length to d2i_ASN1_OBJECT.
4494 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4496 *) [In 0.9.6g-engine release:]
4497 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4498 [Lynn Gazis <lgazis@rainbow.com>]
4500 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4502 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4503 and get fix the header length calculation.
4504 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4505 Alon Kantor <alonk@checkpoint.com> (and others),
4508 *) Use proper error handling instead of 'assertions' in buffer
4509 overflow checks added in 0.9.6e. This prevents DoS (the
4510 assertions could call abort()).
4511 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4513 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4515 *) Add various sanity checks to asn1_get_length() to reject
4516 the ASN1 length bytes if they exceed sizeof(long), will appear
4517 negative or the content length exceeds the length of the
4519 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4521 *) Fix cipher selection routines: ciphers without encryption had no flags
4522 for the cipher strength set and where therefore not handled correctly
4523 by the selection routines (PR #130).
4526 *) Fix EVP_dsa_sha macro.
4530 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4531 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4532 that was added in OpenSSL 0.9.6d.
4534 As the countermeasure turned out to be incompatible with some
4535 broken SSL implementations, the new option is part of SSL_OP_ALL.
4536 SSL_OP_ALL is usually employed when compatibility with weird SSL
4537 implementations is desired (e.g. '-bugs' option to 's_client' and
4538 's_server'), so the new option is automatically set in many
4542 *) Changes in security patch:
4544 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4545 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4546 Air Force Materiel Command, USAF, under agreement number
4549 *) Add various sanity checks to asn1_get_length() to reject
4550 the ASN1 length bytes if they exceed sizeof(long), will appear
4551 negative or the content length exceeds the length of the
4552 supplied buffer. (CVE-2002-0659)
4553 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4555 *) Assertions for various potential buffer overflows, not known to
4557 [Ben Laurie (CHATS)]
4559 *) Various temporary buffers to hold ASCII versions of integers were
4560 too small for 64 bit platforms. (CVE-2002-0655)
4561 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4563 *) Remote buffer overflow in SSL3 protocol - an attacker could
4564 supply an oversized session ID to a client. (CVE-2002-0656)
4565 [Ben Laurie (CHATS)]
4567 *) Remote buffer overflow in SSL2 protocol - an attacker could
4568 supply an oversized client master key. (CVE-2002-0656)
4569 [Ben Laurie (CHATS)]
4571 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4573 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4574 encoded as NULL) with id-dsa-with-sha1.
4575 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4577 *) Check various X509_...() return values in apps/req.c.
4578 [Nils Larsch <nla@trustcenter.de>]
4580 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4581 an end-of-file condition would erronously be flagged, when the CRLF
4582 was just at the end of a processed block. The bug was discovered when
4583 processing data through a buffering memory BIO handing the data to a
4584 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4585 <ptsekov@syntrex.com> and Nedelcho Stanev.
4588 *) Implement a countermeasure against a vulnerability recently found
4589 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4590 before application data chunks to avoid the use of known IVs
4591 with data potentially chosen by the attacker.
4594 *) Fix length checks in ssl3_get_client_hello().
4597 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4598 to prevent ssl3_read_internal() from incorrectly assuming that
4599 ssl3_read_bytes() found application data while handshake
4600 processing was enabled when in fact s->s3->in_read_app_data was
4601 merely automatically cleared during the initial handshake.
4602 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4604 *) Fix object definitions for Private and Enterprise: they were not
4605 recognized in their shortname (=lowercase) representation. Extend
4606 obj_dat.pl to issue an error when using undefined keywords instead
4607 of silently ignoring the problem (Svenning Sorensen
4608 <sss@sss.dnsalias.net>).
4611 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4612 generators, i.e. generators other than 2 and 5. (Previously, the
4613 code did not properly initialise the 'add' and 'rem' values to
4614 BN_generate_prime().)
4616 In the new general case, we do not insist that 'generator' is
4617 actually a primitive root: This requirement is rather pointless;
4618 a generator of the order-q subgroup is just as good, if not
4622 *) Map new X509 verification errors to alerts. Discovered and submitted by
4623 Tom Wu <tom@arcot.com>.
4626 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4627 returning non-zero before the data has been completely received
4628 when using non-blocking I/O.
4629 [Bodo Moeller; problem pointed out by John Hughes]
4631 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4632 [Ben Laurie, Lutz Jaenicke]
4634 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4635 Yoram Zahavi <YoramZ@gilian.com>).
4638 *) Add information about CygWin 1.3 and on, and preserve proper
4639 configuration for the versions before that.
4640 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4642 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4643 check whether we deal with a copy of a session and do not delete from
4644 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4645 <izhar@checkpoint.com>.
4648 *) Do not store session data into the internal session cache, if it
4649 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4650 flag is set). Proposed by Aslam <aslam@funk.com>.
4653 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4657 *) [In 0.9.6d-engine release:]
4658 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4659 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4661 *) Add the configuration target linux-s390x.
4662 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4664 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4665 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4666 variable as an indication that a ClientHello message has been
4667 received. As the flag value will be lost between multiple
4668 invocations of ssl3_accept when using non-blocking I/O, the
4669 function may not be aware that a handshake has actually taken
4670 place, thus preventing a new session from being added to the
4673 To avoid this problem, we now set s->new_session to 2 instead of
4674 using a local variable.
4675 [Lutz Jaenicke, Bodo Moeller]
4677 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4678 if the SSL_R_LENGTH_MISMATCH error is detected.
4679 [Geoff Thorpe, Bodo Moeller]
4681 *) New 'shared_ldflag' column in Configure platform table.
4684 *) Fix EVP_CIPHER_mode macro.
4685 ["Dan S. Camper" <dan@bti.net>]
4687 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4688 type, we must throw them away by setting rr->length to 0.
4689 [D P Chang <dpc@qualys.com>]
4691 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4693 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4694 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4695 worked incorrectly for those cases where range = 10..._2 and
4696 3*range is two bits longer than range.)
4699 *) Only add signing time to PKCS7 structures if it is not already
4703 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4704 OBJ_ld_ce should be OBJ_id_ce.
4705 Also some ip-pda OIDs in crypto/objects/objects.txt were
4706 incorrect (cf. RFC 3039).
4707 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4709 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4710 returns early because it has nothing to do.
4711 [Andy Schneider <andy.schneider@bjss.co.uk>]
4713 *) [In 0.9.6c-engine release:]
4714 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4715 [Andy Schneider <andy.schneider@bjss.co.uk>]
4717 *) [In 0.9.6c-engine release:]
4718 Add support for Cryptographic Appliance's keyserver technology.
4719 (Use engine 'keyclient')
4720 [Cryptographic Appliances and Geoff Thorpe]
4722 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4723 is called via tools/c89.sh because arguments have to be
4724 rearranged (all '-L' options must appear before the first object
4726 [Richard Shapiro <rshapiro@abinitio.com>]
4728 *) [In 0.9.6c-engine release:]
4729 Add support for Broadcom crypto accelerator cards, backported
4731 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4733 *) [In 0.9.6c-engine release:]
4734 Add support for SureWare crypto accelerator cards from
4735 Baltimore Technologies. (Use engine 'sureware')
4736 [Baltimore Technologies and Mark Cox]
4738 *) [In 0.9.6c-engine release:]
4739 Add support for crypto accelerator cards from Accelerated
4740 Encryption Processing, www.aep.ie. (Use engine 'aep')
4741 [AEP Inc. and Mark Cox]
4743 *) Add a configuration entry for gcc on UnixWare.
4744 [Gary Benson <gbenson@redhat.com>]
4746 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4747 messages are stored in a single piece (fixed-length part and
4748 variable-length part combined) and fix various bugs found on the way.
4751 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4752 instead. BIO_gethostbyname() does not know what timeouts are
4753 appropriate, so entries would stay in cache even when they have
4755 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4757 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4758 faced with a pathologically small ClientHello fragment that does
4759 not contain client_version: Instead of aborting with an error,
4760 simply choose the highest available protocol version (i.e.,
4761 TLS 1.0 unless it is disabled). In practice, ClientHello
4762 messages are never sent like this, but this change gives us
4763 strictly correct behaviour at least for TLS.
4766 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4767 never resets s->method to s->ctx->method when called from within
4768 one of the SSL handshake functions.
4769 [Bodo Moeller; problem pointed out by Niko Baric]
4771 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4772 (sent using the client's version number) if client_version is
4773 smaller than the protocol version in use. Also change
4774 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4775 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4776 the client will at least see that alert.
4779 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4783 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4784 client receives HelloRequest while in a handshake.
4785 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4787 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4788 should end in 'break', not 'goto end' which circuments various
4789 cleanups done in state SSL_ST_OK. But session related stuff
4790 must be disabled for SSL_ST_OK in the case that we just sent a
4793 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4794 before just sending a HelloRequest.
4795 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4797 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4798 reveal whether illegal block cipher padding was found or a MAC
4799 verification error occured. (Neither SSLerr() codes nor alerts
4800 are directly visible to potential attackers, but the information
4801 may leak via logfiles.)
4803 Similar changes are not required for the SSL 2.0 implementation
4804 because the number of padding bytes is sent in clear for SSL 2.0,
4805 and the extra bytes are just ignored. However ssl/s2_pkt.c
4806 failed to verify that the purported number of padding bytes is in
4810 *) Add OpenUNIX-8 support including shared libraries
4811 (Boyd Lynn Gerber <gerberb@zenez.com>).
4814 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4815 'wristwatch attack' using huge encoding parameters (cf.
4816 James H. Manger's CRYPTO 2001 paper). Note that the
4817 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4818 encoding parameters and hence was not vulnerable.
4821 *) BN_sqr() bug fix.
4822 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4824 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4825 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4826 followed by modular reduction.
4827 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4829 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4830 equivalent based on BN_pseudo_rand() instead of BN_rand().
4833 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4834 This function was broken, as the check for a new client hello message
4835 to handle SGC did not allow these large messages.
4836 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4839 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4842 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4843 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4846 *) Rework the configuration and shared library support for Tru64 Unix.
4847 The configuration part makes use of modern compiler features and
4848 still retains old compiler behavior for those that run older versions
4849 of the OS. The shared library support part includes a variant that
4850 uses the RPATH feature, and is available through the special
4851 configuration target "alpha-cc-rpath", which will never be selected
4853 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4855 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4856 with the same message size as in ssl3_get_certificate_request().
4857 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4858 messages might inadvertently be reject as too long.
4859 [Petr Lampa <lampa@fee.vutbr.cz>]
4861 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4864 *) Modified SSL library such that the verify_callback that has been set
4865 specificly for an SSL object with SSL_set_verify() is actually being
4866 used. Before the change, a verify_callback set with this function was
4867 ignored and the verify_callback() set in the SSL_CTX at the time of
4868 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4869 to allow the necessary settings.
4872 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4873 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4874 done automatically (in contradiction to the requirements of the C
4875 standard). This made problems when used from OpenSSH.
4878 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4879 dh->length and always used
4881 BN_rand_range(priv_key, dh->p).
4883 BN_rand_range() is not necessary for Diffie-Hellman, and this
4884 specific range makes Diffie-Hellman unnecessarily inefficient if
4885 dh->length (recommended exponent length) is much smaller than the
4886 length of dh->p. We could use BN_rand_range() if the order of
4887 the subgroup was stored in the DH structure, but we only have
4892 BN_rand(priv_key, l, ...)
4894 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4900 RSA_eay_public_encrypt
4901 RSA_eay_private_decrypt
4902 RSA_eay_private_encrypt (signing)
4903 RSA_eay_public_decrypt (signature verification)
4905 (default implementations for RSA_public_encrypt,
4906 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4907 always reject numbers >= n.
4910 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4911 to synchronize access to 'locking_thread'. This is necessary on
4912 systems where access to 'locking_thread' (an 'unsigned long'
4913 variable) is not atomic.
4916 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4917 *before* setting the 'crypto_lock_rand' flag. The previous code had
4918 a race condition if 0 is a valid thread ID.
4919 [Travis Vitek <vitek@roguewave.com>]
4921 *) Add support for shared libraries under Irix.
4922 [Albert Chin-A-Young <china@thewrittenword.com>]
4924 *) Add configuration option to build on Linux on both big-endian and
4926 [Ralf Baechle <ralf@uni-koblenz.de>]
4928 *) Add the possibility to create shared libraries on HP-UX.
4931 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4933 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4934 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4935 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4936 PRNG state recovery was possible based on the output of
4937 one PRNG request appropriately sized to gain knowledge on
4938 'md' followed by enough consecutive 1-byte PRNG requests
4939 to traverse all of 'state'.
4941 1. When updating 'md_local' (the current thread's copy of 'md')
4942 during PRNG output generation, hash all of the previous
4943 'md_local' value, not just the half used for PRNG output.
4945 2. Make the number of bytes from 'state' included into the hash
4946 independent from the number of PRNG bytes requested.
4948 The first measure alone would be sufficient to avoid
4949 Markku-Juhani's attack. (Actually it had never occurred
4950 to me that the half of 'md_local' used for chaining was the
4951 half from which PRNG output bytes were taken -- I had always
4952 assumed that the secret half would be used.) The second
4953 measure makes sure that additional data from 'state' is never
4954 mixed into 'md_local' in small portions; this heuristically
4955 further strengthens the PRNG.
4958 *) Fix crypto/bn/asm/mips3.s.
4961 *) When only the key is given to "enc", the IV is undefined. Print out
4962 an error message in this case.
4965 *) Handle special case when X509_NAME is empty in X509 printing routines.
4968 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4969 positive and less than q.
4972 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4973 used: it isn't thread safe and the add_lock_callback should handle
4975 [Paul Rose <Paul.Rose@bridge.com>]
4977 *) Verify that incoming data obeys the block size in
4978 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4982 [Ulf Möller, Bodo Möller]
4984 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4985 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4986 when fixing the server behaviour for backwards-compatible 'client
4987 hello' messages. (Note that the attack is impractical against
4988 SSL 3.0 and TLS 1.0 anyway because length and version checking
4989 means that the probability of guessing a valid ciphertext is
4990 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4993 Before 0.9.5, the countermeasure (hide the error by generating a
4994 random 'decryption result') did not work properly because
4995 ERR_clear_error() was missing, meaning that SSL_get_error() would
4996 detect the supposedly ignored error.
4998 Both problems are now fixed.
5001 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5002 (previously it was 1024).
5005 *) Fix for compatibility mode trust settings: ignore trust settings
5006 unless some valid trust or reject settings are present.
5009 *) Fix for blowfish EVP: its a variable length cipher.
5012 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5013 parameters in DSA public key structures and return an error in the
5014 DSA routines if parameters are absent.
5017 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5018 in the current directory if neither $RANDFILE nor $HOME was set.
5019 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5020 caused some confusion to Windows users who haven't defined $HOME.
5021 Thus RAND_file_name() is changed again: e_os.h can define a
5022 DEFAULT_HOME, which will be used if $HOME is not set.
5023 For Windows, we use "C:"; on other platforms, we still require
5024 environment variables.
5026 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5027 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5028 having multiple threads call RAND_poll() concurrently.
5031 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5032 combination of a flag and a thread ID variable.
5033 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5034 flag), *other* threads can enter ssleay_add_bytes without obeying
5035 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5036 that they do not hold after the first thread unsets add_do_not_lock).
5039 *) Change bctest again: '-x' expressions are not available in all
5043 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5045 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5046 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5048 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5049 the default extension for executables, if any. Also, make the perl
5050 scripts that use symlink() to test if it really exists and use "cp"
5051 if it doesn't. All this made OpenSSL compilable and installable in
5055 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5056 If SEQUENCE is length is indefinite just set c->slen to the total
5057 amount of data available.
5058 [Steve Henson, reported by shige@FreeBSD.org]
5059 [This change does not apply to 0.9.7.]
5061 *) Change bctest to avoid here-documents inside command substitution
5062 (workaround for FreeBSD /bin/sh bug).
5063 For compatibility with Ultrix, avoid shell functions (introduced
5064 in the bctest version that searches along $PATH).
5067 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5068 with des_encrypt() defined on some operating systems, like Solaris
5072 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5073 On the Importance of Eliminating Errors in Cryptographic
5074 Computations, J. Cryptology 14 (2001) 2, 101-119,
5075 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5078 *) MIPS assembler BIGNUM division bug fix.
5081 *) Disabled incorrect Alpha assembler code.
5084 *) Fix PKCS#7 decode routines so they correctly update the length
5085 after reading an EOC for the EXPLICIT tag.
5087 [This change does not apply to 0.9.7.]
5089 *) Fix bug in PKCS#12 key generation routines. This was triggered
5090 if a 3DES key was generated with a 0 initial byte. Include
5091 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5092 (but broken) behaviour.
5095 *) Enhance bctest to search for a working bc along $PATH and print
5097 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5099 *) Fix memory leaks in err.c: free err_data string if necessary;
5100 don't write to the wrong index in ERR_set_error_data.
5103 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5107 *) Replace rdtsc with _emit statements for VC++ version 5.
5108 [Jeremy Cooper <jeremy@baymoo.org>]
5110 *) Make it possible to reuse SSLv2 sessions.
5113 *) In copy_email() check for >= 0 as a return value for
5114 X509_NAME_get_index_by_NID() since 0 is a valid index.
5115 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5117 *) Avoid coredump with unsupported or invalid public keys by checking if
5118 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5119 PKCS7_verify() fails with non detached data.
5122 *) Don't use getenv in library functions when run as setuid/setgid.
5123 New function OPENSSL_issetugid().
5126 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5127 due to incorrect handling of multi-threading:
5129 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5131 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5133 3. Count how many times MemCheck_off() has been called so that
5134 nested use can be treated correctly. This also avoids
5135 inband-signalling in the previous code (which relied on the
5136 assumption that thread ID 0 is impossible).
5139 *) Add "-rand" option also to s_client and s_server.
5142 *) Fix CPU detection on Irix 6.x.
5143 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5144 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5146 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5149 [This change does not apply to 0.9.7.]
5151 *) Use the cached encoding of an X509_NAME structure rather than
5152 copying it. This is apparently the reason for the libsafe "errors"
5153 but the code is actually correct.
5156 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5157 Bleichenbacher's DSA attack.
5158 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5159 to be set and top=0 forces the highest bit to be set; top=-1 is new
5160 and leaves the highest bit random.
5161 [Ulf Moeller, Bodo Moeller]
5163 *) In the NCONF_...-based implementations for CONF_... queries
5164 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5165 a temporary CONF structure with the data component set to NULL
5166 (which gives segmentation faults in lh_retrieve).
5167 Instead, use NULL for the CONF pointer in CONF_get_string and
5168 CONF_get_number (which may use environment variables) and directly
5169 return NULL from CONF_get_section.
5172 *) Fix potential buffer overrun for EBCDIC.
5175 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5176 keyUsage if basicConstraints absent for a CA.
5179 *) Make SMIME_write_PKCS7() write mail header values with a format that
5180 is more generally accepted (no spaces before the semicolon), since
5181 some programs can't parse those values properly otherwise. Also make
5182 sure BIO's that break lines after each write do not create invalid
5186 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5187 macros previously used would not encode an empty SEQUENCE OF
5188 and break the signature.
5190 [This change does not apply to 0.9.7.]
5192 *) Zero the premaster secret after deriving the master secret in
5196 *) Add some EVP_add_digest_alias registrations (as found in
5197 OpenSSL_add_all_digests()) to SSL_library_init()
5198 aka OpenSSL_add_ssl_algorithms(). This provides improved
5199 compatibility with peers using X.509 certificates
5200 with unconventional AlgorithmIdentifier OIDs.
5203 *) Fix for Irix with NO_ASM.
5204 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5206 *) ./config script fixes.
5207 [Ulf Moeller, Richard Levitte]
5209 *) Fix 'openssl passwd -1'.
5212 *) Change PKCS12_key_gen_asc() so it can cope with non null
5213 terminated strings whose length is passed in the passlen
5214 parameter, for example from PEM callbacks. This was done
5215 by adding an extra length parameter to asc2uni().
5216 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5218 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5219 call failed, free the DSA structure.
5222 *) Fix to uni2asc() to cope with zero length Unicode strings.
5223 These are present in some PKCS#12 files.
5226 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5227 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5228 when writing a 32767 byte record.
5229 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5231 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5232 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5234 (RSA objects have a reference count access to which is protected
5235 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5236 so they are meant to be shared between threads.)
5237 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5238 "Reddie, Steven" <Steven.Reddie@ca.com>]
5240 *) Fix a deadlock in CRYPTO_mem_leaks().
5243 *) Use better test patterns in bntest.
5246 *) rand_win.c fix for Borland C.
5249 *) BN_rshift bugfix for n == 0.
5252 *) Add a 'bctest' script that checks for some known 'bc' bugs
5253 so that 'make test' does not abort just because 'bc' is broken.
5256 *) Store verify_result within SSL_SESSION also for client side to
5257 avoid potential security hole. (Re-used sessions on the client side
5258 always resulted in verify_result==X509_V_OK, not using the original
5259 result of the server certificate verification.)
5262 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5263 SSL3_RT_APPLICATION_DATA, return 0.
5264 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5268 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5269 releases, have been re-implemented by renaming the previous
5270 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5271 and ssl3_read_internal, respectively, and adding 'peek' parameters
5272 to them. The new ssl[23]_{read,peek} functions are calls to
5273 ssl[23]_read_internal with the 'peek' flag set appropriately.
5274 A 'peek' parameter has also been added to ssl3_read_bytes, which
5275 does the actual work for ssl3_read_internal.
5278 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5279 the method-specific "init()" handler. Also clean up ex_data after
5280 calling the method-specific "finish()" handler. Previously, this was
5281 happening the other way round.
5284 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5285 The previous value, 12, was not always sufficient for BN_mod_exp().
5288 *) Make sure that shared libraries get the internal name engine with
5289 the full version number and not just 0. This should mark the
5290 shared libraries as not backward compatible. Of course, this should
5291 be changed again when we can guarantee backward binary compatibility.
5294 *) Fix typo in get_cert_by_subject() in by_dir.c
5295 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5297 *) Rework the system to generate shared libraries:
5299 - Make note of the expected extension for the shared libraries and
5300 if there is a need for symbolic links from for example libcrypto.so.0
5301 to libcrypto.so.0.9.7. There is extended info in Configure for
5304 - Make as few rebuilds of the shared libraries as possible.
5306 - Still avoid linking the OpenSSL programs with the shared libraries.
5308 - When installing, install the shared libraries separately from the
5312 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5314 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5315 and not in SSL_clear because the latter is also used by the
5316 accept/connect functions; previously, the settings made by
5317 SSL_set_read_ahead would be lost during the handshake.
5318 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5320 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5321 Previously, it would create entries for disableed algorithms no
5325 *) Added several new manual pages for SSL_* function.
5328 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5330 *) In ssl23_get_client_hello, generate an error message when faced
5331 with an initial SSL 3.0/TLS record that is too small to contain the
5332 first two bytes of the ClientHello message, i.e. client_version.
5333 (Note that this is a pathologic case that probably has never happened
5334 in real life.) The previous approach was to use the version number
5335 from the record header as a substitute; but our protocol choice
5336 should not depend on that one because it is not authenticated
5337 by the Finished messages.
5340 *) More robust randomness gathering functions for Windows.
5341 [Jeffrey Altman <jaltman@columbia.edu>]
5343 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5344 not set then we don't setup the error code for issuer check errors
5345 to avoid possibly overwriting other errors which the callback does
5346 handle. If an application does set the flag then we assume it knows
5347 what it is doing and can handle the new informational codes
5351 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5352 a general "ANY" type, as such it should be able to decode anything
5353 including tagged types. However it didn't check the class so it would
5354 wrongly interpret tagged types in the same way as their universal
5355 counterpart and unknown types were just rejected. Changed so that the
5356 tagged and unknown types are handled in the same way as a SEQUENCE:
5357 that is the encoding is stored intact. There is also a new type
5358 "V_ASN1_OTHER" which is used when the class is not universal, in this
5359 case we have no idea what the actual type is so we just lump them all
5363 *) On VMS, stdout may very well lead to a file that is written to
5364 in a record-oriented fashion. That means that every write() will
5365 write a separate record, which will be read separately by the
5366 programs trying to read from it. This can be very confusing.
5368 The solution is to put a BIO filter in the way that will buffer
5369 text until a linefeed is reached, and then write everything a
5370 line at a time, so every record written will be an actual line,
5371 not chunks of lines and not (usually doesn't happen, but I've
5372 seen it once) several lines in one record. BIO_f_linebuffer() is
5375 Currently, it's a VMS-only method, because that's where it has
5376 been tested well enough.
5379 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5380 it can return incorrect results.
5381 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5382 but it was in 0.9.6-beta[12].)
5385 *) Disable the check for content being present when verifying detached
5386 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5387 include zero length content when signing messages.
5390 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5391 BIO_ctrl (for BIO pairs).
5394 *) Add DSO method for VMS.
5397 *) Bug fix: Montgomery multiplication could produce results with the
5401 *) Add RPM specification openssl.spec and modify it to build three
5402 packages. The default package contains applications, application
5403 documentation and run-time libraries. The devel package contains
5404 include files, static libraries and function documentation. The
5405 doc package contains the contents of the doc directory. The original
5406 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5409 *) Add a large number of documentation files for many SSL routines.
5410 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5412 *) Add a configuration entry for Sony News 4.
5413 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5415 *) Don't set the two most significant bits to one when generating a
5416 random number < q in the DSA library.
5419 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5420 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5421 the underlying transport is blocking) if a handshake took place.
5422 (The default behaviour is needed by applications such as s_client
5423 and s_server that use select() to determine when to use SSL_read;
5424 but for applications that know in advance when to expect data, it
5425 just makes things more complicated.)
5428 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5432 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5433 work better on such systems.
5434 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5436 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5437 Update PKCS12_parse() so it copies the friendlyName and the
5438 keyid to the certificates aux info.
5441 *) Fix bug in PKCS7_verify() which caused an infinite loop
5442 if there was more than one signature.
5443 [Sven Uszpelkat <su@celocom.de>]
5445 *) Major change in util/mkdef.pl to include extra information
5446 about each symbol, as well as presentig variables as well
5447 as functions. This change means that there's n more need
5448 to rebuild the .num files when some algorithms are excluded.
5451 *) Allow the verify time to be set by an application,
5452 rather than always using the current time.
5455 *) Phase 2 verify code reorganisation. The certificate
5456 verify code now looks up an issuer certificate by a
5457 number of criteria: subject name, authority key id
5458 and key usage. It also verifies self signed certificates
5459 by the same criteria. The main comparison function is
5460 X509_check_issued() which performs these checks.
5462 Lot of changes were necessary in order to support this
5463 without completely rewriting the lookup code.
5465 Authority and subject key identifier are now cached.
5467 The LHASH 'certs' is X509_STORE has now been replaced
5468 by a STACK_OF(X509_OBJECT). This is mainly because an
5469 LHASH can't store or retrieve multiple objects with
5470 the same hash value.
5472 As a result various functions (which were all internal
5473 use only) have changed to handle the new X509_STORE
5474 structure. This will break anything that messed round
5475 with X509_STORE internally.
5477 The functions X509_STORE_add_cert() now checks for an
5478 exact match, rather than just subject name.
5480 The X509_STORE API doesn't directly support the retrieval
5481 of multiple certificates matching a given criteria, however
5482 this can be worked round by performing a lookup first
5483 (which will fill the cache with candidate certificates)
5484 and then examining the cache for matches. This is probably
5485 the best we can do without throwing out X509_LOOKUP
5486 entirely (maybe later...).
5488 The X509_VERIFY_CTX structure has been enhanced considerably.
5490 All certificate lookup operations now go via a get_issuer()
5491 callback. Although this currently uses an X509_STORE it
5492 can be replaced by custom lookups. This is a simple way
5493 to bypass the X509_STORE hackery necessary to make this
5494 work and makes it possible to use more efficient techniques
5495 in future. A very simple version which uses a simple
5496 STACK for its trusted certificate store is also provided
5497 using X509_STORE_CTX_trusted_stack().
5499 The verify_cb() and verify() callbacks now have equivalents
5500 in the X509_STORE_CTX structure.
5502 X509_STORE_CTX also has a 'flags' field which can be used
5503 to customise the verify behaviour.
5506 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5507 excludes S/MIME capabilities.
5510 *) When a certificate request is read in keep a copy of the
5511 original encoding of the signed data and use it when outputing
5512 again. Signatures then use the original encoding rather than
5513 a decoded, encoded version which may cause problems if the
5514 request is improperly encoded.
5517 *) For consistency with other BIO_puts implementations, call
5518 buffer_write(b, ...) directly in buffer_puts instead of calling
5521 In BIO_puts, increment b->num_write as in BIO_write.
5522 [Peter.Sylvester@EdelWeb.fr]
5524 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5525 BN_zero, we may not return a BIGNUM with an array consisting of
5529 *) Avoid calling abort() from within the library when problems are
5530 detected, except if preprocessor symbols have been defined
5531 (such as REF_CHECK, BN_DEBUG etc.).
5534 *) New openssl application 'rsautl'. This utility can be
5535 used for low level RSA operations. DER public key
5536 BIO/fp routines also added.
5539 *) New Configure entry and patches for compiling on QNX 4.
5540 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5542 *) A demo state-machine implementation was sponsored by
5543 Nuron (http://www.nuron.com/) and is now available in
5544 demos/state_machine.
5547 *) New options added to the 'dgst' utility for signature
5548 generation and verification.
5551 *) Unrecognized PKCS#7 content types are now handled via a
5552 catch all ASN1_TYPE structure. This allows unsupported
5553 types to be stored as a "blob" and an application can
5554 encode and decode it manually.
5557 *) Fix various signed/unsigned issues to make a_strex.c
5559 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5561 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5562 length if passed a buffer. ASN1_INTEGER_to_BN failed
5563 if passed a NULL BN and its argument was negative.
5564 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5566 *) Modification to PKCS#7 encoding routines to output definite
5567 length encoding. Since currently the whole structures are in
5568 memory there's not real point in using indefinite length
5569 constructed encoding. However if OpenSSL is compiled with
5570 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5573 *) Added BIO_vprintf() and BIO_vsnprintf().
5576 *) Added more prefixes to parse for in the the strings written
5577 through a logging bio, to cover all the levels that are available
5578 through syslog. The prefixes are now:
5580 PANIC, EMERG, EMR => LOG_EMERG
5581 ALERT, ALR => LOG_ALERT
5582 CRIT, CRI => LOG_CRIT
5583 ERROR, ERR => LOG_ERR
5584 WARNING, WARN, WAR => LOG_WARNING
5585 NOTICE, NOTE, NOT => LOG_NOTICE
5586 INFO, INF => LOG_INFO
5587 DEBUG, DBG => LOG_DEBUG
5589 and as before, if none of those prefixes are present at the
5590 beginning of the string, LOG_ERR is chosen.
5592 On Win32, the LOG_* levels are mapped according to this:
5594 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5595 LOG_WARNING => EVENTLOG_WARNING_TYPE
5596 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5600 *) Made it possible to reconfigure with just the configuration
5601 argument "reconf" or "reconfigure". The command line arguments
5602 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5603 and are retrieved from there when reconfiguring.
5607 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5609 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5612 *) The obj_dat.pl script was messing up the sorting of object
5613 names. The reason was that it compared the quoted version
5614 of strings as a result "OCSP" > "OCSP Signing" because
5615 " > SPACE. Changed script to store unquoted versions of
5616 names and add quotes on output. It was also omitting some
5617 names from the lookup table if they were given a default
5618 value (that is if SN is missing it is given the same
5619 value as LN and vice versa), these are now added on the
5620 grounds that if an object has a name we should be able to
5621 look it up. Finally added warning output when duplicate
5622 short or long names are found.
5625 *) Changes needed for Tandem NSK.
5626 [Scott Uroff <scott@xypro.com>]
5628 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5629 RSA_padding_check_SSLv23(), special padding was never detected
5630 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5631 version rollback attacks was not effective.
5633 In s23_clnt.c, don't use special rollback-attack detection padding
5634 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5635 client; similarly, in s23_srvr.c, don't do the rollback check if
5636 SSL 2.0 is the only protocol enabled in the server.
5639 *) Make it possible to get hexdumps of unprintable data with 'openssl
5640 asn1parse'. By implication, the functions ASN1_parse_dump() and
5641 BIO_dump_indent() are added.
5644 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5645 these print out strings and name structures based on various
5646 flags including RFC2253 support and proper handling of
5647 multibyte characters. Added options to the 'x509' utility
5648 to allow the various flags to be set.
5651 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5652 Also change the functions X509_cmp_current_time() and
5653 X509_gmtime_adj() work with an ASN1_TIME structure,
5654 this will enable certificates using GeneralizedTime in validity
5655 dates to be checked.
5658 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5659 negative public key encodings) on by default,
5660 NO_NEG_PUBKEY_BUG can be set to disable it.
5663 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5664 content octets. An i2c_ASN1_OBJECT is unnecessary because
5665 the encoding can be trivially obtained from the structure.
5668 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5669 not read locks (CRYPTO_r_[un]lock).
5672 *) A first attempt at creating official support for shared
5673 libraries through configuration. I've kept it so the
5674 default is static libraries only, and the OpenSSL programs
5675 are always statically linked for now, but there are
5676 preparations for dynamic linking in place.
5677 This has been tested on Linux and Tru64.
5680 *) Randomness polling function for Win9x, as described in:
5681 Peter Gutmann, Software Generation of Practically Strong
5685 *) Fix so PRNG is seeded in req if using an already existing
5689 *) New options to smime application. -inform and -outform
5690 allow alternative formats for the S/MIME message including
5691 PEM and DER. The -content option allows the content to be
5692 specified separately. This should allow things like Netscape
5693 form signing output easier to verify.
5696 *) Fix the ASN1 encoding of tags using the 'long form'.
5699 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5700 STRING types. These convert content octets to and from the
5701 underlying type. The actual tag and length octets are
5702 already assumed to have been read in and checked. These
5703 are needed because all other string types have virtually
5704 identical handling apart from the tag. By having versions
5705 of the ASN1 functions that just operate on content octets
5706 IMPLICIT tagging can be handled properly. It also allows
5707 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5708 and ASN1_INTEGER are identical apart from the tag.
5711 *) Change the handling of OID objects as follows:
5713 - New object identifiers are inserted in objects.txt, following
5714 the syntax given in objects.README.
5715 - objects.pl is used to process obj_mac.num and create a new
5717 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5720 This is currently kind of a hack, and the perl code in objects.pl
5721 isn't very elegant, but it works as I intended. The simplest way
5722 to check that it worked correctly is to look in obj_dat.h and
5723 check the array nid_objs and make sure the objects haven't moved
5724 around (this is important!). Additions are OK, as well as
5725 consistent name changes.
5728 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5731 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5732 The given file adds to whatever has already been seeded into the
5733 random pool through the RANDFILE configuration file option or
5734 environment variable, or the default random state file.
5737 *) mkstack.pl now sorts each macro group into lexical order.
5738 Previously the output order depended on the order the files
5739 appeared in the directory, resulting in needless rewriting
5743 *) Patches to make OpenSSL compile under Win32 again. Mostly
5744 work arounds for the VC++ problem that it treats func() as
5745 func(void). Also stripped out the parts of mkdef.pl that
5746 added extra typesafe functions: these no longer exist.
5749 *) Reorganisation of the stack code. The macros are now all
5750 collected in safestack.h . Each macro is defined in terms of
5751 a "stack macro" of the form SKM_<name>(type, a, b). The
5752 DEBUG_SAFESTACK is now handled in terms of function casts,
5753 this has the advantage of retaining type safety without the
5754 use of additional functions. If DEBUG_SAFESTACK is not defined
5755 then the non typesafe macros are used instead. Also modified the
5756 mkstack.pl script to handle the new form. Needs testing to see
5757 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5758 the default if no major problems. Similar behaviour for ASN1_SET_OF
5759 and PKCS12_STACK_OF.
5762 *) When some versions of IIS use the 'NET' form of private key the
5763 key derivation algorithm is different. Normally MD5(password) is
5764 used as a 128 bit RC4 key. In the modified case
5765 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5766 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5767 as the old Netscape_RSA functions except they have an additional
5768 'sgckey' parameter which uses the modified algorithm. Also added
5769 an -sgckey command line option to the rsa utility. Thanks to
5770 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5771 algorithm to openssl-dev.
5774 *) The evp_local.h macros were using 'c.##kname' which resulted in
5775 invalid expansion on some systems (SCO 5.0.5 for example).
5776 Corrected to 'c.kname'.
5777 [Phillip Porch <root@theporch.com>]
5779 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5780 a STACK of email addresses from a certificate or request, these look
5781 in the subject name and the subject alternative name extensions and
5782 omit any duplicate addresses.
5785 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5786 This makes DSA verification about 2 % faster.
5789 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5790 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5791 plus overhead for 1024 bit moduli).
5792 This makes exponentiations about 0.5 % faster for 1024 bit
5793 exponents (as measured by "openssl speed rsa2048").
5796 *) Rename memory handling macros to avoid conflicts with other
5798 Malloc => OPENSSL_malloc
5799 Malloc_locked => OPENSSL_malloc_locked
5800 Realloc => OPENSSL_realloc
5801 Free => OPENSSL_free
5804 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5805 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5808 *) CygWin32 support.
5809 [John Jarvie <jjarvie@newsguy.com>]
5811 *) The type-safe stack code has been rejigged. It is now only compiled
5812 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5813 by default all type-specific stack functions are "#define"d back to
5814 standard stack functions. This results in more streamlined output
5815 but retains the type-safety checking possibilities of the original
5819 *) The STACK code has been cleaned up, and certain type declarations
5820 that didn't make a lot of sense have been brought in line. This has
5821 also involved a cleanup of sorts in safestack.h to more correctly
5822 map type-safe stack functions onto their plain stack counterparts.
5823 This work has also resulted in a variety of "const"ifications of
5824 lots of the code, especially "_cmp" operations which should normally
5825 be prototyped with "const" parameters anyway.
5828 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5829 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5830 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5831 where all of 'md' is used each time the PRNG is used, but 'state'
5832 is used only indexed by a cyclic counter. As entropy may not be
5833 well distributed from the beginning, 'md' is important as a
5834 chaining variable. However, the output function chains only half
5835 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5836 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5837 in all of 'state' being rewritten, with the new values depending
5838 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5841 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5842 the handshake is continued after ssl_verify_cert_chain();
5843 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5844 can lead to 'unexplainable' connection aborts later.
5845 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5847 *) Major EVP API cipher revision.
5848 Add hooks for extra EVP features. This allows various cipher
5849 parameters to be set in the EVP interface. Support added for variable
5850 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5851 setting of RC2 and RC5 parameters.
5853 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5856 Remove lots of duplicated code from the EVP library. For example *every*
5857 cipher init() function handles the 'iv' in the same way according to the
5858 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5859 for CFB and OFB modes they zero ctx->num.
5861 New functionality allows removal of S/MIME code RC2 hack.
5863 Most of the routines have the same form and so can be declared in terms
5866 By shifting this to the top level EVP_CipherInit() it can be removed from
5867 all individual ciphers. If the cipher wants to handle IVs or keys
5868 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5871 Change lots of functions like EVP_EncryptUpdate() to now return a
5872 value: although software versions of the algorithms cannot fail
5873 any installed hardware versions can.
5876 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5877 this option is set, tolerate broken clients that send the negotiated
5878 protocol version number instead of the requested protocol version
5882 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5883 i.e. non-zero for export ciphersuites, zero otherwise.
5884 Previous versions had this flag inverted, inconsistent with
5885 rsa_tmp_cb (..._TMP_RSA_CB).
5886 [Bodo Moeller; problem reported by Amit Chopra]
5888 *) Add missing DSA library text string. Work around for some IIS
5889 key files with invalid SEQUENCE encoding.
5892 *) Add a document (doc/standards.txt) that list all kinds of standards
5893 and so on that are implemented in OpenSSL.
5896 *) Enhance c_rehash script. Old version would mishandle certificates
5897 with the same subject name hash and wouldn't handle CRLs at all.
5898 Added -fingerprint option to crl utility, to support new c_rehash
5902 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5905 *) Fix for SSL server purpose checking. Server checking was
5906 rejecting certificates which had extended key usage present
5907 but no ssl client purpose.
5908 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5910 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5911 is a little unclear about how a blank password is handled.
5912 Since the password in encoded as a BMPString with terminating
5913 double NULL a zero length password would end up as just the
5914 double NULL. However no password at all is different and is
5915 handled differently in the PKCS#12 key generation code. NS
5916 treats a blank password as zero length. MSIE treats it as no
5917 password on export: but it will try both on import. We now do
5918 the same: PKCS12_parse() tries zero length and no password if
5919 the password is set to "" or NULL (NULL is now a valid password:
5920 it wasn't before) as does the pkcs12 application.
5923 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5924 perror when PEM_read_bio_X509_REQ fails, the error message must
5925 be obtained from the error queue.
5928 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5929 it in ERR_remove_state if appropriate, and change ERR_get_state
5930 accordingly to avoid race conditions (this is necessary because
5931 thread_hash is no longer constant once set).
5934 *) Bugfix for linux-elf makefile.one.
5937 *) RSA_get_default_method() will now cause a default
5938 RSA_METHOD to be chosen if one doesn't exist already.
5939 Previously this was only set during a call to RSA_new()
5940 or RSA_new_method(NULL) meaning it was possible for
5941 RSA_get_default_method() to return NULL.
5944 *) Added native name translation to the existing DSO code
5945 that will convert (if the flag to do so is set) filenames
5946 that are sufficiently small and have no path information
5947 into a canonical native form. Eg. "blah" converted to
5948 "libblah.so" or "blah.dll" etc.
5951 *) New function ERR_error_string_n(e, buf, len) which is like
5952 ERR_error_string(e, buf), but writes at most 'len' bytes
5953 including the 0 terminator. For ERR_error_string_n, 'buf'
5955 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5957 *) CONF library reworked to become more general. A new CONF
5958 configuration file reader "class" is implemented as well as a
5959 new functions (NCONF_*, for "New CONF") to handle it. The now
5960 old CONF_* functions are still there, but are reimplemented to
5961 work in terms of the new functions. Also, a set of functions
5962 to handle the internal storage of the configuration data is
5963 provided to make it easier to write new configuration file
5964 reader "classes" (I can definitely see something reading a
5965 configuration file in XML format, for example), called _CONF_*,
5966 or "the configuration storage API"...
5968 The new configuration file reading functions are:
5970 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5971 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5973 NCONF_default, NCONF_WIN32
5975 NCONF_dump_fp, NCONF_dump_bio
5977 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5978 NCONF_new creates a new CONF object. This works in the same way
5979 as other interfaces in OpenSSL, like the BIO interface.
5980 NCONF_dump_* dump the internal storage of the configuration file,
5981 which is useful for debugging. All other functions take the same
5982 arguments as the old CONF_* functions wth the exception of the
5983 first that must be a `CONF *' instead of a `LHASH *'.
5985 To make it easer to use the new classes with the old CONF_* functions,
5986 the function CONF_set_default_method is provided.
5989 *) Add '-tls1' option to 'openssl ciphers', which was already
5990 mentioned in the documentation but had not been implemented.
5991 (This option is not yet really useful because even the additional
5992 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5995 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5996 OpenSSL-based applications) load shared libraries and bind to
5997 them in a portable way.
5998 [Geoff Thorpe, with contributions from Richard Levitte]
6000 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6002 *) Make sure _lrotl and _lrotr are only used with MSVC.
6004 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6005 (the default implementation of RAND_status).
6007 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6008 to '-clrext' (= clear extensions), as intended and documented.
6009 [Bodo Moeller; inconsistency pointed out by Michael Attili
6012 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6013 was larger than the MD block size.
6014 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6016 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6017 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6018 using the passed key: if the passed key was a private key the result
6019 of X509_print(), for example, would be to print out all the private key
6023 *) des_quad_cksum() byte order bug fix.
6024 [Ulf Möller, using the problem description in krb4-0.9.7, where
6025 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6027 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6029 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6031 *) For easily testing in shell scripts whether some command
6032 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6033 returns with exit code 0 iff no command of the given name is available.
6034 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6035 the output goes to stdout and nothing is printed to stderr.
6036 Additional arguments are always ignored.
6038 Since for each cipher there is a command of the same name,
6039 the 'no-cipher' compilation switches can be tested this way.
6041 ('openssl no-XXX' is not able to detect pseudo-commands such
6042 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6045 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6048 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6049 is set; it will be thrown away anyway because each handshake creates
6051 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6052 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6053 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6054 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6057 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6058 'Q' and 'R' lose their special meanings (quit/renegotiate).
6059 This is part of what -quiet does; unlike -quiet, -ign_eof
6060 does not suppress any output.
6063 *) Add compatibility options to the purpose and trust code. The
6064 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6065 accepts a certificate or CA, this was the previous behaviour,
6066 with all the associated security issues.
6068 X509_TRUST_COMPAT is the old trust behaviour: only and
6069 automatically trust self signed roots in certificate store. A
6070 new trust setting X509_TRUST_DEFAULT is used to specify that
6071 a purpose has no associated trust setting and it should instead
6072 use the value in the default purpose.
6075 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6076 and fix a memory leak.
6079 *) In util/mkerr.pl (which implements 'make errors'), preserve
6080 reason strings from the previous version of the .c file, as
6081 the default to have only downcase letters (and digits) in
6082 automatically generated reasons codes is not always appropriate.
6085 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6086 using strerror. Previously, ERR_reason_error_string() returned
6087 library names as reason strings for SYSerr; but SYSerr is a special
6088 case where small numbers are errno values, not library numbers.
6091 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6092 converts DSA parameters into DH parameters. (When creating parameters,
6093 DSA_generate_parameters is used.)
6096 *) Include 'length' (recommended exponent length) in C code generated
6097 by 'openssl dhparam -C'.
6100 *) The second argument to set_label in perlasm was already being used
6101 so couldn't be used as a "file scope" flag. Moved to third argument
6105 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6106 instead of RAND_bytes for encryption IVs and salts.
6109 *) Include RAND_status() into RAND_METHOD instead of implementing
6110 it only for md_rand.c Otherwise replacing the PRNG by calling
6111 RAND_set_rand_method would be impossible.
6114 *) Don't let DSA_generate_key() enter an infinite loop if the random
6115 number generation fails.
6118 *) New 'rand' application for creating pseudo-random output.
6121 *) Added configuration support for Linux/IA64
6122 [Rolf Haberrecker <rolf@suse.de>]
6124 *) Assembler module support for Mingw32.
6127 *) Shared library support for HPUX (in shlib/).
6128 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6130 *) Shared library support for Solaris gcc.
6131 [Lutz Behnke <behnke@trustcenter.de>]
6133 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6135 *) PKCS7_encrypt() was adding text MIME headers twice because they
6136 were added manually and by SMIME_crlf_copy().
6139 *) In bntest.c don't call BN_rand with zero bits argument.
6140 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6142 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6143 case was implemented. This caused BN_div_recp() to fail occasionally.
6146 *) Add an optional second argument to the set_label() in the perl
6147 assembly language builder. If this argument exists and is set
6148 to 1 it signals that the assembler should use a symbol whose
6149 scope is the entire file, not just the current function. This
6150 is needed with MASM which uses the format label:: for this scope.
6151 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6153 *) Change the ASN1 types so they are typedefs by default. Before
6154 almost all types were #define'd to ASN1_STRING which was causing
6155 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6159 *) Change names of new functions to the new get1/get0 naming
6160 convention: After 'get1', the caller owns a reference count
6161 and has to call ..._free; 'get0' returns a pointer to some
6162 data structure without incrementing reference counters.
6163 (Some of the existing 'get' functions increment a reference
6164 counter, some don't.)
6165 Similarly, 'set1' and 'add1' functions increase reference
6166 counters or duplicate objects.
6169 *) Allow for the possibility of temp RSA key generation failure:
6170 the code used to assume it always worked and crashed on failure.
6173 *) Fix potential buffer overrun problem in BIO_printf().
6174 [Ulf Möller, using public domain code by Patrick Powell; problem
6175 pointed out by David Sacerdote <das33@cornell.edu>]
6177 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6178 RAND_egd() and RAND_status(). In the command line application,
6179 the EGD socket can be specified like a seed file using RANDFILE
6183 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6184 Some CAs (e.g. Verisign) distribute certificates in this form.
6187 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6188 list to exclude them. This means that no special compilation option
6189 is needed to use anonymous DH: it just needs to be included in the
6193 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6194 EVP_MD_type. The old functionality is available in a new macro called
6195 EVP_MD_md(). Change code that uses it and update docs.
6198 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6199 where the 'void *' argument is replaced by a function pointer argument.
6200 Previously 'void *' was abused to point to functions, which works on
6201 many platforms, but is not correct. As these functions are usually
6202 called by macros defined in OpenSSL header files, most source code
6203 should work without changes.
6206 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6207 sections with information on -D... compiler switches used for
6208 compiling the library so that applications can see them. To enable
6209 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6210 must be defined. E.g.,
6211 #define OPENSSL_ALGORITHM_DEFINES
6212 #include <openssl/opensslconf.h>
6213 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6214 [Richard Levitte, Ulf and Bodo Möller]
6216 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6220 *) Change the 'other' type in certificate aux info to a STACK_OF
6221 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6222 the required ASN1 format: arbitrary types determined by an OID.
6225 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6226 argument to 'req'. This is not because the function is newer or
6227 better than others it just uses the work 'NEW' in the certificate
6228 request header lines. Some software needs this.
6231 *) Reorganise password command line arguments: now passwords can be
6232 obtained from various sources. Delete the PEM_cb function and make
6233 it the default behaviour: i.e. if the callback is NULL and the
6234 usrdata argument is not NULL interpret it as a null terminated pass
6235 phrase. If usrdata and the callback are NULL then the pass phrase
6236 is prompted for as usual.
6239 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6240 the support is automatically enabled. The resulting binaries will
6241 autodetect the card and use it if present.
6242 [Ben Laurie and Compaq Inc.]
6244 *) Work around for Netscape hang bug. This sends certificate request
6245 and server done in one record. Since this is perfectly legal in the
6246 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6247 the bugs/SSLv3 entry for more info.
6250 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6253 *) Add -rand argument to smime and pkcs12 applications and read/write
6257 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6260 *) Add command line password options to the remaining applications.
6263 *) Bug fix for BN_div_recp() for numerators with an even number of
6267 *) More tests in bntest.c, and changed test_bn output.
6270 *) ./config recognizes MacOS X now.
6273 *) Bug fix for BN_div() when the first words of num and divsor are
6274 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6277 *) Add support for various broken PKCS#8 formats, and command line
6278 options to produce them.
6281 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6282 get temporary BIGNUMs from a BN_CTX.
6285 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6289 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6290 include a #define from the old name to the new. The original intent
6291 was that statically linked binaries could for example just call
6292 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6293 link with digests. This never worked becayse SSLeay_add_all_digests()
6294 and SSLeay_add_all_ciphers() were in the same source file so calling
6295 one would link with the other. They are now in separate source files.
6298 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6301 *) Use a less unusual form of the Miller-Rabin primality test (it used
6302 a binary algorithm for exponentiation integrated into the Miller-Rabin
6303 loop, our standard modexp algorithms are faster).
6306 *) Support for the EBCDIC character set completed.
6307 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6309 *) Source code cleanups: use const where appropriate, eliminate casts,
6310 use void * instead of char * in lhash.
6313 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6314 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6315 this the server could overwrite ephemeral keys that the client
6319 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6320 using 50 iterations of the Rabin-Miller test.
6322 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6323 iterations of the Rabin-Miller test as required by the appendix
6324 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6325 As BN_is_prime_fasttest includes trial division, DSA parameter
6326 generation becomes much faster.
6328 This implies a change for the callback functions in DSA_is_prime
6329 and DSA_generate_parameters: The callback function is called once
6330 for each positive witness in the Rabin-Miller test, not just
6331 occasionally in the inner loop; and the parameters to the
6332 callback function now provide an iteration count for the outer
6333 loop rather than for the current invocation of the inner loop.
6334 DSA_generate_parameters additionally can call the callback
6335 function with an 'iteration count' of -1, meaning that a
6336 candidate has passed the trial division test (when q is generated
6337 from an application-provided seed, trial division is skipped).
6340 *) New function BN_is_prime_fasttest that optionally does trial
6341 division before starting the Rabin-Miller test and has
6342 an additional BN_CTX * argument (whereas BN_is_prime always
6343 has to allocate at least one BN_CTX).
6344 'callback(1, -1, cb_arg)' is called when a number has passed the
6345 trial division stage.
6348 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6352 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6355 *) New function BN_pseudo_rand().
6358 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6359 bignum version of BN_from_montgomery() with the working code from
6360 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6364 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6365 made it impossible to use the same SSL_SESSION data structure in
6366 SSL2 clients in multiple threads.
6369 *) The return value of RAND_load_file() no longer counts bytes obtained
6370 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6371 to seed the PRNG (previously an explicit byte count was required).
6372 [Ulf Möller, Bodo Möller]
6374 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6375 used (char *) instead of (void *) and had casts all over the place.
6378 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6381 *) Retain source code compatibility for BN_prime_checks macro:
6382 BN_is_prime(..., BN_prime_checks, ...) now uses
6383 BN_prime_checks_for_size to determine the appropriate number of
6384 Rabin-Miller iterations.
6387 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6388 DH_CHECK_P_NOT_SAFE_PRIME.
6389 (Check if this is true? OpenPGP calls them "strong".)
6392 *) Merge the functionality of "dh" and "gendh" programs into a new program
6393 "dhparam". The old programs are retained for now but will handle DH keys
6394 (instead of parameters) in future.
6397 *) Make the ciphers, s_server and s_client programs check the return values
6398 when a new cipher list is set.
6401 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6402 ciphers. Before when the 56bit ciphers were enabled the sorting was
6405 The syntax for the cipher sorting has been extended to support sorting by
6406 cipher-strength (using the strength_bits hard coded in the tables).
6407 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6409 Fix a bug in the cipher-command parser: when supplying a cipher command
6410 string with an "undefined" symbol (neither command nor alphanumeric
6411 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6412 an error is flagged.
6414 Due to the strength-sorting extension, the code of the
6415 ssl_create_cipher_list() function was completely rearranged. I hope that
6416 the readability was also increased :-)
6417 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6419 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6420 for the first serial number and places 2 in the serial number file. This
6421 avoids problems when the root CA is created with serial number zero and
6422 the first user certificate has the same issuer name and serial number
6426 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6427 the new code. Add documentation for this stuff.
6430 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6431 X509_*() to X509at_*() on the grounds that they don't handle X509
6432 structures and behave in an analagous way to the X509v3 functions:
6433 they shouldn't be called directly but wrapper functions should be used
6436 So we also now have some wrapper functions that call the X509at functions
6437 when passed certificate requests. (TO DO: similar things can be done with
6438 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6439 things. Some of these need some d2i or i2d and print functionality
6440 because they handle more complex structures.)
6443 *) Add missing #ifndefs that caused missing symbols when building libssl
6444 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6445 NO_RSA in ssl/s2*.c.
6446 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6448 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6449 has a return value which indicates the quality of the random data
6450 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6451 error queue. New function RAND_pseudo_bytes() generates output that is
6452 guaranteed to be unique but not unpredictable. RAND_add is like
6453 RAND_seed, but takes an extra argument for an entropy estimate
6454 (RAND_seed always assumes full entropy).
6457 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6458 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6459 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6460 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6461 false-positive rate of at most 2^-80 for random input.
6464 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6467 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6468 in the 0.9.5 release), this returns the chain
6469 from an X509_CTX structure with a dup of the stack and all
6470 the X509 reference counts upped: so the stack will exist
6471 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6474 Also make SSL_SESSION_print() print out the verify return
6478 *) Add manpage for the pkcs12 command. Also change the default
6479 behaviour so MAC iteration counts are used unless the new
6480 -nomaciter option is used. This improves file security and
6481 only older versions of MSIE (4.0 for example) need it.
6484 *) Honor the no-xxx Configure options when creating .DEF files.
6487 *) Add PKCS#10 attributes to field table: challengePassword,
6488 unstructuredName and unstructuredAddress. These are taken from
6489 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6490 international characters are used.
6492 More changes to X509_ATTRIBUTE code: allow the setting of types
6493 based on strings. Remove the 'loc' parameter when adding
6494 attributes because these will be a SET OF encoding which is sorted
6498 *) Initial changes to the 'req' utility to allow request generation
6499 automation. This will allow an application to just generate a template
6500 file containing all the field values and have req construct the
6503 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6504 used all over the place including certificate requests and PKCS#7
6505 structures. They are currently handled manually where necessary with
6506 some primitive wrappers for PKCS#7. The new functions behave in a
6507 manner analogous to the X509 extension functions: they allow
6508 attributes to be looked up by NID and added.
6510 Later something similar to the X509V3 code would be desirable to
6511 automatically handle the encoding, decoding and printing of the
6512 more complex types. The string types like challengePassword can
6513 be handled by the string table functions.
6515 Also modified the multi byte string table handling. Now there is
6516 a 'global mask' which masks out certain types. The table itself
6517 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6518 is useful when for example there is only one permissible type
6519 (as in countryName) and using the mask might result in no valid
6523 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6524 SSL_get_peer_finished to allow applications to obtain the latest
6525 Finished messages sent to the peer or expected from the peer,
6526 respectively. (SSL_get_peer_finished is usually the Finished message
6527 actually received from the peer, otherwise the protocol will be aborted.)
6529 As the Finished message are message digests of the complete handshake
6530 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6531 be used for external authentication procedures when the authentication
6532 provided by SSL/TLS is not desired or is not enough.
6535 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6536 the host supports BWX extension and if Compaq C is present on the
6537 $PATH. Just exploiting of the BWX extension results in 20-30%
6538 performance kick for some algorithms, e.g. DES and RC4 to mention
6539 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6543 *) Add support for MS "fast SGC". This is arguably a violation of the
6544 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6545 weak crypto and after checking the certificate is SGC a second one
6546 with strong crypto. MS SGC stops the first handshake after receiving
6547 the server certificate message and sends a second client hello. Since
6548 a server will typically do all the time consuming operations before
6549 expecting any further messages from the client (server key exchange
6550 is the most expensive) there is little difference between the two.
6552 To get OpenSSL to support MS SGC we have to permit a second client
6553 hello message after we have sent server done. In addition we have to
6554 reset the MAC if we do get this second client hello.
6557 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6558 if a DER encoded private key is RSA or DSA traditional format. Changed
6559 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6560 format DER encoded private key. Newer code should use PKCS#8 format which
6561 has the key type encoded in the ASN1 structure. Added DER private key
6562 support to pkcs8 application.
6565 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6566 ciphersuites has been selected (as required by the SSL 3/TLS 1
6567 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6568 is set, we interpret this as a request to violate the specification
6569 (the worst that can happen is a handshake failure, and 'correct'
6570 behaviour would result in a handshake failure anyway).
6573 *) In SSL_CTX_add_session, take into account that there might be multiple
6574 SSL_SESSION structures with the same session ID (e.g. when two threads
6575 concurrently obtain them from an external cache).
6576 The internal cache can handle only one SSL_SESSION with a given ID,
6577 so if there's a conflict, we now throw out the old one to achieve
6581 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6582 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6583 some routines that use cipher OIDs: some ciphers do not have OIDs
6584 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6588 *) Simplify the trust setting structure and code. Now we just have
6589 two sequences of OIDs for trusted and rejected settings. These will
6590 typically have values the same as the extended key usage extension
6591 and any application specific purposes.
6593 The trust checking code now has a default behaviour: it will just
6594 check for an object with the same NID as the passed id. Functions can
6595 be provided to override either the default behaviour or the behaviour
6596 for a given id. SSL client, server and email already have functions
6597 in place for compatibility: they check the NID and also return "trusted"
6598 if the certificate is self signed.
6601 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6602 traditional format into an EVP_PKEY structure.
6605 *) Add a password callback function PEM_cb() which either prompts for
6606 a password if usr_data is NULL or otherwise assumes it is a null
6607 terminated password. Allow passwords to be passed on command line
6608 environment or config files in a few more utilities.
6611 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6612 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6613 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6614 Update documentation.
6617 *) Support for ASN1 "NULL" type. This could be handled before by using
6618 ASN1_TYPE but there wasn't any function that would try to read a NULL
6619 and produce an error if it couldn't. For compatibility we also have
6620 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6621 don't allocate anything because they don't need to.
6624 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6626 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6628 *) Rebuild of the memory allocation routines used by OpenSSL code and
6629 possibly others as well. The purpose is to make an interface that
6630 provide hooks so anyone can build a separate set of allocation and
6631 deallocation routines to be used by OpenSSL, for example memory
6632 pool implementations, or something else, which was previously hard
6633 since Malloc(), Realloc() and Free() were defined as macros having
6634 the values malloc, realloc and free, respectively (except for Win32
6635 compilations). The same is provided for memory debugging code.
6636 OpenSSL already comes with functionality to find memory leaks, but
6637 this gives people a chance to debug other memory problems.
6639 With these changes, a new set of functions and macros have appeared:
6641 CRYPTO_set_mem_debug_functions() [F]
6642 CRYPTO_get_mem_debug_functions() [F]
6643 CRYPTO_dbg_set_options() [F]
6644 CRYPTO_dbg_get_options() [F]
6645 CRYPTO_malloc_debug_init() [M]
6647 The memory debug functions are NULL by default, unless the library
6648 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6649 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6650 gives the standard debugging functions that come with OpenSSL) or
6651 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6652 provided by the library user) must be used. When the standard
6653 debugging functions are used, CRYPTO_dbg_set_options can be used to
6654 request additional information:
6655 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6656 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6658 Also, things like CRYPTO_set_mem_functions will always give the
6659 expected result (the new set of functions is used for allocation
6660 and deallocation) at all times, regardless of platform and compiler
6663 To finish it up, some functions that were never use in any other
6664 way than through macros have a new API and new semantic:
6667 CRYPTO_dbg_realloc()
6670 All macros of value have retained their old syntax.
6671 [Richard Levitte and Bodo Moeller]
6673 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6674 ordering of SMIMECapabilities wasn't in "strength order" and there
6675 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6679 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6680 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6681 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6683 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6684 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6685 functionality to handle multipart/signed properly) and a utility
6686 called 'smime' to call all this stuff. This is based on code I
6687 originally wrote for Celo who have kindly allowed it to be
6688 included in OpenSSL.
6691 *) Add variants des_set_key_checked and des_set_key_unchecked of
6692 des_set_key (aka des_key_sched). Global variable des_check_key
6693 decides which of these is called by des_set_key; this way
6694 des_check_key behaves as it always did, but applications and
6695 the library itself, which was buggy for des_check_key == 1,
6696 have a cleaner way to pick the version they need.
6699 *) New function PKCS12_newpass() which changes the password of a
6703 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6704 dynamic mix. In both cases the ids can be used as an index into the
6705 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6706 functions so they accept a list of the field values and the
6707 application doesn't need to directly manipulate the X509_TRUST
6711 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6715 *) Modify the way the V3 extension code looks up extensions. This now
6716 works in a similar way to the object code: we have some "standard"
6717 extensions in a static table which is searched with OBJ_bsearch()
6718 and the application can add dynamic ones if needed. The file
6719 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6720 updated whenever a new extension is added to the core code and kept
6721 in ext_nid order. There is a simple program 'tabtest.c' which checks
6722 this. New extensions are not added too often so this file can readily
6723 be maintained manually.
6725 There are two big advantages in doing things this way. The extensions
6726 can be looked up immediately and no longer need to be "added" using
6727 X509V3_add_standard_extensions(): this function now does nothing.
6728 [Side note: I get *lots* of email saying the extension code doesn't
6729 work because people forget to call this function]
6730 Also no dynamic allocation is done unless new extensions are added:
6731 so if we don't add custom extensions there is no need to call
6732 X509V3_EXT_cleanup().
6735 *) Modify enc utility's salting as follows: make salting the default. Add a
6736 magic header, so unsalted files fail gracefully instead of just decrypting
6737 to garbage. This is because not salting is a big security hole, so people
6738 should be discouraged from doing it.
6741 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6742 digest to be passed on the command line but it only used this
6743 parameter when signing a certificate. Modified so all relevant
6744 operations are affected by the digest parameter including the
6745 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6746 DSA key was used because it didn't fix the digest.
6749 *) Initial certificate chain verify code. Currently tests the untrusted
6750 certificates for consistency with the verify purpose (which is set
6751 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6753 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6754 this is because it will reject chains with invalid extensions whereas
6755 every previous version of OpenSSL and SSLeay made no checks at all.
6757 Trust code: checks the root CA for the relevant trust settings. Trust
6758 settings have an initial value consistent with the verify purpose: e.g.
6759 if the verify purpose is for SSL client use it expects the CA to be
6760 trusted for SSL client use. However the default value can be changed to
6761 permit custom trust settings: one example of this would be to only trust
6762 certificates from a specific "secure" set of CAs.
6764 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6765 which should be used for version portability: especially since the
6766 verify structure is likely to change more often now.
6768 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6769 to set them. If not set then assume SSL clients will verify SSL servers
6772 Two new options to the verify program: -untrusted allows a set of
6773 untrusted certificates to be passed in and -purpose which sets the
6774 intended purpose of the certificate. If a purpose is set then the
6775 new chain verify code is used to check extension consistency.
6778 *) Support for the authority information access extension.
6781 *) Modify RSA and DSA PEM read routines to transparently handle
6782 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6783 public keys in a format compatible with certificate
6784 SubjectPublicKeyInfo structures. Unfortunately there were already
6785 functions called *_PublicKey_* which used various odd formats so
6786 these are retained for compatibility: however the DSA variants were
6787 never in a public release so they have been deleted. Changed dsa/rsa
6788 utilities to handle the new format: note no releases ever handled public
6789 keys so we should be OK.
6791 The primary motivation for this change is to avoid the same fiasco
6792 that dogs private keys: there are several incompatible private key
6793 formats some of which are standard and some OpenSSL specific and
6794 require various evil hacks to allow partial transparent handling and
6795 even then it doesn't work with DER formats. Given the option anything
6796 other than PKCS#8 should be dumped: but the other formats have to
6797 stay in the name of compatibility.
6799 With public keys and the benefit of hindsight one standard format
6800 is used which works with EVP_PKEY, RSA or DSA structures: though
6801 it clearly returns an error if you try to read the wrong kind of key.
6803 Added a -pubkey option to the 'x509' utility to output the public key.
6804 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6805 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6806 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6807 that do the same as the EVP_PKEY_assign_*() except they up the
6808 reference count of the added key (they don't "swallow" the
6812 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6813 CRLs would fail if the file contained no certificates or no CRLs:
6814 added a new function to read in both types and return the number
6815 read: this means that if none are read it will be an error. The
6816 DER versions of the certificate and CRL reader would always fail
6817 because it isn't possible to mix certificates and CRLs in DER format
6818 without choking one or the other routine. Changed this to just read
6819 a certificate: this is the best we can do. Also modified the code
6820 in apps/verify.c to take notice of return codes: it was previously
6821 attempting to read in certificates from NULL pointers and ignoring
6822 any errors: this is one reason why the cert and CRL reader seemed
6823 to work. It doesn't check return codes from the default certificate
6824 routines: these may well fail if the certificates aren't installed.
6827 *) Code to support otherName option in GeneralName.
6830 *) First update to verify code. Change the verify utility
6831 so it warns if it is passed a self signed certificate:
6832 for consistency with the normal behaviour. X509_verify
6833 has been modified to it will now verify a self signed
6834 certificate if *exactly* the same certificate appears
6835 in the store: it was previously impossible to trust a
6836 single self signed certificate. This means that:
6837 openssl verify ss.pem
6838 now gives a warning about a self signed certificate but
6839 openssl verify -CAfile ss.pem ss.pem
6843 *) For servers, store verify_result in SSL_SESSION data structure
6844 (and add it to external session representation).
6845 This is needed when client certificate verifications fails,
6846 but an application-provided verification callback (set by
6847 SSL_CTX_set_cert_verify_callback) allows accepting the session
6848 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6849 but returns 1): When the session is reused, we have to set
6850 ssl->verify_result to the appropriate error code to avoid
6852 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6854 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6855 case in PKCS7_dataInit() where the signed PKCS7 structure
6856 didn't contain any existing data because it was being created.
6857 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6859 *) Add a salt to the key derivation routines in enc.c. This
6860 forms the first 8 bytes of the encrypted file. Also add a
6861 -S option to allow a salt to be input on the command line.
6864 *) New function X509_cmp(). Oddly enough there wasn't a function
6865 to compare two certificates. We do this by working out the SHA1
6866 hash and comparing that. X509_cmp() will be needed by the trust
6870 *) SSL_get1_session() is like SSL_get_session(), but increments
6871 the reference count in the SSL_SESSION returned.
6872 [Geoff Thorpe <geoff@eu.c2.net>]
6874 *) Fix for 'req': it was adding a null to request attributes.
6875 Also change the X509_LOOKUP and X509_INFO code to handle
6876 certificate auxiliary information.
6879 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6883 *) Add the possibility to add extra information to the memory leak
6884 detecting output, to form tracebacks, showing from where each
6885 allocation was originated: CRYPTO_push_info("constant string") adds
6886 the string plus current file name and line number to a per-thread
6887 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6888 is like calling CYRPTO_pop_info() until the stack is empty.
6889 Also updated memory leak detection code to be multi-thread-safe.
6892 *) Add options -text and -noout to pkcs7 utility and delete the
6893 encryption options which never did anything. Update docs.
6896 *) Add options to some of the utilities to allow the pass phrase
6897 to be included on either the command line (not recommended on
6898 OSes like Unix) or read from the environment. Update the
6899 manpages and fix a few bugs.
6902 *) Add a few manpages for some of the openssl commands.
6905 *) Fix the -revoke option in ca. It was freeing up memory twice,
6906 leaking and not finding already revoked certificates.
6909 *) Extensive changes to support certificate auxiliary information.
6910 This involves the use of X509_CERT_AUX structure and X509_AUX
6911 functions. An X509_AUX function such as PEM_read_X509_AUX()
6912 can still read in a certificate file in the usual way but it
6913 will also read in any additional "auxiliary information". By
6914 doing things this way a fair degree of compatibility can be
6915 retained: existing certificates can have this information added
6916 using the new 'x509' options.
6918 Current auxiliary information includes an "alias" and some trust
6919 settings. The trust settings will ultimately be used in enhanced
6920 certificate chain verification routines: currently a certificate
6921 can only be trusted if it is self signed and then it is trusted
6925 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6926 The problem was that one of the replacement routines had not been working
6927 since SSLeay releases. For now the offending routine has been replaced
6928 with non-optimised assembler. Even so, this now gives around 95%
6929 performance improvement for 1024 bit RSA signs.
6932 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6933 handling. Most clients have the effective key size in bits equal to
6934 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6935 A few however don't do this and instead use the size of the decrypted key
6936 to determine the RC2 key length and the AlgorithmIdentifier to determine
6937 the effective key length. In this case the effective key length can still
6938 be 40 bits but the key length can be 168 bits for example. This is fixed
6939 by manually forcing an RC2 key into the EVP_PKEY structure because the
6940 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6941 the key length and effective key length are equal.
6944 *) Add a bunch of functions that should simplify the creation of
6945 X509_NAME structures. Now you should be able to do:
6946 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6947 and have it automatically work out the correct field type and fill in
6948 the structures. The more adventurous can try:
6949 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6950 and it will (hopefully) work out the correct multibyte encoding.
6953 *) Change the 'req' utility to use the new field handling and multibyte
6954 copy routines. Before the DN field creation was handled in an ad hoc
6955 way in req, ca, and x509 which was rather broken and didn't support
6956 BMPStrings or UTF8Strings. Since some software doesn't implement
6957 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6958 using the dirstring_type option. See the new comment in the default
6959 openssl.cnf for more info.
6962 *) Make crypto/rand/md_rand.c more robust:
6963 - Assure unique random numbers after fork().
6964 - Make sure that concurrent threads access the global counter and
6965 md serializably so that we never lose entropy in them
6966 or use exactly the same state in multiple threads.
6967 Access to the large state is not always serializable because
6968 the additional locking could be a performance killer, and
6969 md should be large enough anyway.
6972 *) New file apps/app_rand.c with commonly needed functionality
6973 for handling the random seed file.
6975 Use the random seed file in some applications that previously did not:
6977 dsaparam -genkey (which also ignored its '-rand' option),
6980 x509 (when signing).
6981 Except on systems with /dev/urandom, it is crucial to have a random
6982 seed file at least for key creation, DSA signing, and for DH exchanges;
6983 for RSA signatures we could do without one.
6985 gendh and gendsa (unlike genrsa) used to read only the first byte
6986 of each file listed in the '-rand' option. The function as previously
6987 found in genrsa is now in app_rand.c and is used by all programs
6988 that support '-rand'.
6991 *) In RAND_write_file, use mode 0600 for creating files;
6992 don't just chmod when it may be too late.
6995 *) Report an error from X509_STORE_load_locations
6996 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6999 *) New function ASN1_mbstring_copy() this copies a string in either
7000 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7001 into an ASN1_STRING type. A mask of permissible types is passed
7002 and it chooses the "minimal" type to use or an error if not type
7006 *) Add function equivalents to the various macros in asn1.h. The old
7007 macros are retained with an M_ prefix. Code inside the library can
7008 use the M_ macros. External code (including the openssl utility)
7009 should *NOT* in order to be "shared library friendly".
7012 *) Add various functions that can check a certificate's extensions
7013 to see if it usable for various purposes such as SSL client,
7014 server or S/MIME and CAs of these types. This is currently
7015 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7016 verification. Also added a -purpose flag to x509 utility to
7017 print out all the purposes.
7020 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7024 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7025 for, obtain and decode and extension and obtain its critical flag.
7026 This allows all the necessary extension code to be handled in a
7027 single function call.
7030 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7031 platforms. See crypto/rc4/rc4_enc.c for further details.
7034 *) New -noout option to asn1parse. This causes no output to be produced
7035 its main use is when combined with -strparse and -out to extract data
7036 from a file (which may not be in ASN.1 format).
7039 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7040 when producing the local key id.
7041 [Richard Levitte <levitte@stacken.kth.se>]
7043 *) New option -dhparam in s_server. This allows a DH parameter file to be
7044 stated explicitly. If it is not stated then it tries the first server
7045 certificate file. The previous behaviour hard coded the filename
7049 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7050 a public key to be input or output. For example:
7051 openssl rsa -in key.pem -pubout -out pubkey.pem
7052 Also added necessary DSA public key functions to handle this.
7055 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7056 in the message. This was handled by allowing
7057 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7058 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7060 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7061 to the end of the strings whereas this didn't. This would cause problems
7062 if strings read with d2i_ASN1_bytes() were later modified.
7063 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7065 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7066 data and it contains EOF it will end up returning an error. This is
7067 caused by input 46 bytes long. The cause is due to the way base64
7068 BIOs find the start of base64 encoded data. They do this by trying a
7069 trial decode on each line until they find one that works. When they
7070 do a flag is set and it starts again knowing it can pass all the
7071 data directly through the decoder. Unfortunately it doesn't reset
7072 the context it uses. This means that if EOF is reached an attempt
7073 is made to pass two EOFs through the context and this causes the
7074 resulting error. This can also cause other problems as well. As is
7075 usual with these problems it takes *ages* to find and the fix is
7076 trivial: move one line.
7077 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7079 *) Ugly workaround to get s_client and s_server working under Windows. The
7080 old code wouldn't work because it needed to select() on sockets and the
7081 tty (for keypresses and to see if data could be written). Win32 only
7082 supports select() on sockets so we select() with a 1s timeout on the
7083 sockets and then see if any characters are waiting to be read, if none
7084 are present then we retry, we also assume we can always write data to
7085 the tty. This isn't nice because the code then blocks until we've
7086 received a complete line of data and it is effectively polling the
7087 keyboard at 1s intervals: however it's quite a bit better than not
7088 working at all :-) A dedicated Windows application might handle this
7089 with an event loop for example.
7092 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7093 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7094 will be called when RSA_sign() and RSA_verify() are used. This is useful
7095 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7096 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7097 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7098 This necessitated the support of an extra signature type NID_md5_sha1
7099 for SSL signatures and modifications to the SSL library to use it instead
7100 of calling RSA_public_decrypt() and RSA_private_encrypt().
7103 *) Add new -verify -CAfile and -CApath options to the crl program, these
7104 will lookup a CRL issuers certificate and verify the signature in a
7105 similar way to the verify program. Tidy up the crl program so it
7106 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7107 less strict. It will now permit CRL extensions even if it is not
7108 a V2 CRL: this will allow it to tolerate some broken CRLs.
7111 *) Initialize all non-automatic variables each time one of the openssl
7112 sub-programs is started (this is necessary as they may be started
7113 multiple times from the "OpenSSL>" prompt).
7114 [Lennart Bang, Bodo Moeller]
7116 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7117 removing all other RSA functionality (this is what NO_RSA does). This
7118 is so (for example) those in the US can disable those operations covered
7119 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7123 *) Non-copying interface to BIO pairs.
7124 (still largely untested)
7127 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7128 ASCII string. This was handled independently in various places before.
7131 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7132 UTF8 strings a character at a time.
7135 *) Use client_version from client hello to select the protocol
7136 (s23_srvr.c) and for RSA client key exchange verification
7137 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7140 *) Add various utility functions to handle SPKACs, these were previously
7141 handled by poking round in the structure internals. Added new function
7142 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7143 print, verify and generate SPKACs. Based on an original idea from
7144 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7147 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7150 *) Allow the config file extension section to be overwritten on the
7151 command line. Based on an original idea from Massimiliano Pala
7152 <madwolf@comune.modena.it>. The new option is called -extensions
7153 and can be applied to ca, req and x509. Also -reqexts to override
7154 the request extensions in req and -crlexts to override the crl extensions
7158 *) Add new feature to the SPKAC handling in ca. Now you can include
7159 the same field multiple times by preceding it by "XXXX." for example:
7162 this is the same syntax as used in the req config file.
7165 *) Allow certificate extensions to be added to certificate requests. These
7166 are specified in a 'req_extensions' option of the req section of the
7167 config file. They can be printed out with the -text option to req but
7168 are otherwise ignored at present.
7171 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7172 data read consists of only the final block it would not decrypted because
7173 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7174 A misplaced 'break' also meant the decrypted final block might not be
7175 copied until the next read.
7178 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7179 a few extra parameters to the DH structure: these will be useful if
7180 for example we want the value of 'q' or implement X9.42 DH.
7183 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7184 provides hooks that allow the default DSA functions or functions on a
7185 "per key" basis to be replaced. This allows hardware acceleration and
7186 hardware key storage to be handled without major modification to the
7187 library. Also added low level modexp hooks and CRYPTO_EX structure and
7188 associated functions.
7191 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7192 as "read only": it can't be written to and the buffer it points to will
7193 not be freed. Reading from a read only BIO is much more efficient than
7194 a normal memory BIO. This was added because there are several times when
7195 an area of memory needs to be read from a BIO. The previous method was
7196 to create a memory BIO and write the data to it, this results in two
7197 copies of the data and an O(n^2) reading algorithm. There is a new
7198 function BIO_new_mem_buf() which creates a read only memory BIO from
7199 an area of memory. Also modified the PKCS#7 routines to use read only
7203 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7204 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7205 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7206 but a retry condition occured while trying to read the rest.
7209 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7210 NID_pkcs7_encrypted by default: this was wrong since this should almost
7211 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7212 the encrypted data type: this is a more sensible place to put it and it
7213 allows the PKCS#12 code to be tidied up that duplicated this
7217 *) Changed obj_dat.pl script so it takes its input and output files on
7218 the command line. This should avoid shell escape redirection problems
7222 *) Initial support for certificate extension requests, these are included
7223 in things like Xenroll certificate requests. Included functions to allow
7224 extensions to be obtained and added.
7227 *) -crlf option to s_client and s_server for sending newlines as
7228 CRLF (as required by many protocols).
7231 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7233 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7234 [Ralf S. Engelschall]
7236 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7237 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7239 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7243 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7244 DH parameters/keys (q is lost during that conversion, but the resulting
7245 DH parameters contain its length).
7247 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7248 much faster than DH_generate_parameters (which creates parameters
7249 where p = 2*q + 1), and also the smaller q makes DH computations
7250 much more efficient (160-bit exponentiation instead of 1024-bit
7251 exponentiation); so this provides a convenient way to support DHE
7252 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7253 utter importance to use
7254 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7256 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7257 when such DH parameters are used, because otherwise small subgroup
7258 attacks may become possible!
7261 *) Avoid memory leak in i2d_DHparams.
7264 *) Allow the -k option to be used more than once in the enc program:
7265 this allows the same encrypted message to be read by multiple recipients.
7268 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7269 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7270 it will always use the numerical form of the OID, even if it has a short
7274 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7275 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7276 otherwise bn_mod_exp was called. In the case of hardware keys for example
7277 no private key components need be present and it might store extra data
7278 in the RSA structure, which cannot be accessed from bn_mod_exp.
7279 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7280 private key operations.
7283 *) Added support for SPARC Linux.
7286 *) pem_password_cb function type incompatibly changed from
7287 typedef int pem_password_cb(char *buf, int size, int rwflag);
7289 ....(char *buf, int size, int rwflag, void *userdata);
7290 so that applications can pass data to their callbacks:
7291 The PEM[_ASN1]_{read,write}... functions and macros now take an
7292 additional void * argument, which is just handed through whenever
7293 the password callback is called.
7294 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7296 New function SSL_CTX_set_default_passwd_cb_userdata.
7298 Compatibility note: As many C implementations push function arguments
7299 onto the stack in reverse order, the new library version is likely to
7300 interoperate with programs that have been compiled with the old
7301 pem_password_cb definition (PEM_whatever takes some data that
7302 happens to be on the stack as its last argument, and the callback
7303 just ignores this garbage); but there is no guarantee whatsoever that
7306 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7307 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7308 problems not only on Windows, but also on some Unix platforms.
7309 To avoid problematic command lines, these definitions are now in an
7310 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7311 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7314 *) MIPS III/IV assembler module is reimplemented.
7317 *) More DES library cleanups: remove references to srand/rand and
7318 delete an unused file.
7321 *) Add support for the the free Netwide assembler (NASM) under Win32,
7322 since not many people have MASM (ml) and it can be hard to obtain.
7323 This is currently experimental but it seems to work OK and pass all
7324 the tests. Check out INSTALL.W32 for info.
7327 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7328 without temporary keys kept an extra copy of the server key,
7329 and connections with temporary keys did not free everything in case
7333 *) New function RSA_check_key and new openssl rsa option -check
7334 for verifying the consistency of RSA keys.
7335 [Ulf Moeller, Bodo Moeller]
7337 *) Various changes to make Win32 compile work:
7338 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7339 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7340 comparison" warnings.
7341 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7344 *) Add a debugging option to PKCS#5 v2 key generation function: when
7345 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7346 derived keys are printed to stderr.
7349 *) Copy the flags in ASN1_STRING_dup().
7350 [Roman E. Pavlov <pre@mo.msk.ru>]
7352 *) The x509 application mishandled signing requests containing DSA
7353 keys when the signing key was also DSA and the parameters didn't match.
7355 It was supposed to omit the parameters when they matched the signing key:
7356 the verifying software was then supposed to automatically use the CA's
7357 parameters if they were absent from the end user certificate.
7359 Omitting parameters is no longer recommended. The test was also
7360 the wrong way round! This was probably due to unusual behaviour in
7361 EVP_cmp_parameters() which returns 1 if the parameters match.
7362 This meant that parameters were omitted when they *didn't* match and
7363 the certificate was useless. Certificates signed with 'ca' didn't have
7365 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7367 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7368 The interface is as follows:
7369 Applications can use
7370 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7371 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7372 "off" is now the default.
7373 The library internally uses
7374 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7375 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7376 to disable memory-checking temporarily.
7378 Some inconsistent states that previously were possible (and were
7379 even the default) are now avoided.
7381 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7382 with each memory chunk allocated; this is occasionally more helpful
7383 than just having a counter.
7385 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7387 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7391 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7392 which largely parallels "options", but is for changing API behaviour,
7393 whereas "options" are about protocol behaviour.
7394 Initial "mode" flags are:
7396 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7397 a single record has been written.
7398 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7399 retries use the same buffer location.
7400 (But all of the contents must be
7404 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7407 *) Fix problems with no-hmac etc.
7408 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7410 *) New functions RSA_get_default_method(), RSA_set_method() and
7411 RSA_get_method(). These allows replacement of RSA_METHODs without having
7412 to mess around with the internals of an RSA structure.
7415 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7416 Also really enable memory leak checks in openssl.c and in some
7418 [Chad C. Mulligan, Bodo Moeller]
7420 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7421 up the length of negative integers. This has now been simplified to just
7422 store the length when it is first determined and use it later, rather
7423 than trying to keep track of where data is copied and updating it to
7425 [Steve Henson, reported by Brien Wheeler
7426 <bwheeler@authentica-security.com>]
7428 *) Add a new function PKCS7_signatureVerify. This allows the verification
7429 of a PKCS#7 signature but with the signing certificate passed to the
7430 function itself. This contrasts with PKCS7_dataVerify which assumes the
7431 certificate is present in the PKCS#7 structure. This isn't always the
7432 case: certificates can be omitted from a PKCS#7 structure and be
7433 distributed by "out of band" means (such as a certificate database).
7436 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7437 function prototypes in pem.h, also change util/mkdef.pl to add the
7438 necessary function names.
7441 *) mk1mf.pl (used by Windows builds) did not properly read the
7442 options set by Configure in the top level Makefile, and Configure
7443 was not even able to write more than one option correctly.
7444 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7447 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7448 file to be loaded from a BIO or FILE pointer. The BIO version will
7449 for example allow memory BIOs to contain config info.
7452 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7453 Whoever hopes to achieve shared-library compatibility across versions
7454 must use this, not the compile-time macro.
7455 (Exercise 0.9.4: Which is the minimum library version required by
7457 Note: All this applies only to multi-threaded programs, others don't
7461 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7462 through a BIO pair triggered the default case, i.e.
7463 SSLerr(...,SSL_R_UNKNOWN_STATE).
7466 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7467 can use the SSL library even if none of the specific BIOs is
7471 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7472 for the encoded length.
7473 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7475 *) Add initial documentation of the X509V3 functions.
7478 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7479 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7480 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7481 secure PKCS#8 private key format with a high iteration count.
7484 *) Fix determination of Perl interpreter: A perl or perl5
7485 _directory_ in $PATH was also accepted as the interpreter.
7486 [Ralf S. Engelschall]
7488 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7489 wrong with it but it was very old and did things like calling
7490 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7494 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7495 to use the new extension code.
7498 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7499 with macros. This should make it easier to change their form, add extra
7500 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7504 *) Add to configuration table a new entry that can specify an alternative
7505 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7506 according to Mark Crispin <MRC@Panda.COM>.
7510 *) DES CBC did not update the IV. Weird.
7513 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7514 Changing the behaviour of the former might break existing programs --
7515 where IV updating is needed, des_ncbc_encrypt can be used.
7518 *) When bntest is run from "make test" it drives bc to check its
7519 calculations, as well as internally checking them. If an internal check
7520 fails, it needs to cause bc to give a non-zero result or make test carries
7521 on without noticing the failure. Fixed.
7524 *) DES library cleanups.
7527 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7528 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7529 ciphers. NOTE: although the key derivation function has been verified
7530 against some published test vectors it has not been extensively tested
7531 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7535 *) Instead of "mkdir -p", which is not fully portable, use new
7536 Perl script "util/mkdir-p.pl".
7539 *) Rewrite the way password based encryption (PBE) is handled. It used to
7540 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7541 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7542 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7543 the 'parameter' field of the AlgorithmIdentifier is passed to the
7544 underlying key generation function so it must do its own ASN1 parsing.
7545 This has also changed the EVP_PBE_CipherInit() function which now has a
7546 'parameter' argument instead of literal salt and iteration count values
7547 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7550 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7551 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7552 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7553 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7554 value was just used as a "magic string" and not used directly its
7555 value doesn't matter.
7558 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7562 *) "linux-sparc64" configuration (ultrapenguin).
7563 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7564 "linux-sparc" configuration.
7565 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7567 *) config now generates no-xxx options for missing ciphers.
7570 *) Support the EBCDIC character set (work in progress).
7571 File ebcdic.c not yet included because it has a different license.
7572 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7574 *) Support BS2000/OSD-POSIX.
7575 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7577 *) Make callbacks for key generation use void * instead of char *.
7580 *) Make S/MIME samples compile (not yet tested).
7583 *) Additional typesafe stacks.
7586 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7590 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7592 *) New configuration variant "sco5-gcc".
7594 *) Updated some demos.
7595 [Sean O Riordain, Wade Scholine]
7597 *) Add missing BIO_free at exit of pkcs12 application.
7600 *) Fix memory leak in conf.c.
7603 *) Updates for Win32 to assembler version of MD5.
7606 *) Set #! path to perl in apps/der_chop to where we found it
7607 instead of using a fixed path.
7610 *) SHA library changes for irix64-mips4-cc.
7613 *) Improvements for VMS support.
7617 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7619 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7620 This also avoids the problems with SC4.2 and unpatched SC5.
7621 [Andy Polyakov <appro@fy.chalmers.se>]
7623 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7624 These are required because of the typesafe stack would otherwise break
7625 existing code. If old code used a structure member which used to be STACK
7626 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7627 sk_num or sk_value it would produce an error because the num, data members
7628 are not present in STACK_OF. Now it just produces a warning. sk_set
7629 replaces the old method of assigning a value to sk_value
7630 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7631 that does this will no longer work (and should use sk_set instead) but
7632 this could be regarded as a "questionable" behaviour anyway.
7635 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7636 correctly handle encrypted S/MIME data.
7639 *) Change type of various DES function arguments from des_cblock
7640 (which means, in function argument declarations, pointer to char)
7641 to des_cblock * (meaning pointer to array with 8 char elements),
7642 which allows the compiler to do more typechecking; it was like
7643 that back in SSLeay, but with lots of ugly casts.
7645 Introduce new type const_des_cblock.
7648 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7649 problems: find RecipientInfo structure that matches recipient certificate
7650 and initialise the ASN1 structures properly based on passed cipher.
7653 *) Belatedly make the BN tests actually check the results.
7656 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7657 to and from BNs: it was completely broken. New compilation option
7658 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7659 key elements as negative integers.
7662 *) Reorganize and speed up MD5.
7663 [Andy Polyakov <appro@fy.chalmers.se>]
7666 [Richard Levitte <richard@levitte.org>]
7668 *) New option -out to asn1parse to allow the parsed structure to be
7669 output to a file. This is most useful when combined with the -strparse
7670 option to examine the output of things like OCTET STRINGS.
7673 *) Make SSL library a little more fool-proof by not requiring any longer
7674 that SSL_set_{accept,connect}_state be called before
7675 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7676 in many applications because usually everything *appeared* to work as
7677 intended anyway -- now it really works as intended).
7680 *) Move openssl.cnf out of lib/.
7683 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7684 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7685 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7686 [Ralf S. Engelschall]
7688 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7689 handle PKCS#7 enveloped data properly.
7690 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7692 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7693 copying pointers. The cert_st handling is changed by this in
7694 various ways (and thus what used to be known as ctx->default_cert
7695 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7696 any longer when s->cert does not give us what we need).
7697 ssl_cert_instantiate becomes obsolete by this change.
7698 As soon as we've got the new code right (possibly it already is?),
7699 we have solved a couple of bugs of the earlier code where s->cert
7700 was used as if it could not have been shared with other SSL structures.
7702 Note that using the SSL API in certain dirty ways now will result
7703 in different behaviour than observed with earlier library versions:
7704 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7705 does not influence s as it used to.
7707 In order to clean up things more thoroughly, inside SSL_SESSION
7708 we don't use CERT any longer, but a new structure SESS_CERT
7709 that holds per-session data (if available); currently, this is
7710 the peer's certificate chain and, for clients, the server's certificate
7711 and temporary key. CERT holds only those values that can have
7712 meaningful defaults in an SSL_CTX.
7715 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7716 from the internal representation. Various PKCS#7 fixes: remove some
7717 evil casts and set the enc_dig_alg field properly based on the signing
7721 *) Allow PKCS#12 password to be set from the command line or the
7722 environment. Let 'ca' get its config file name from the environment
7723 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7727 *) Allow certificate policies extension to use an IA5STRING for the
7728 organization field. This is contrary to the PKIX definition but
7729 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7733 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7734 without disallowing inline assembler and the like for non-pedantic builds.
7737 *) Support Borland C++ builder.
7738 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7743 *) SHA-1 cleanups and performance enhancements.
7744 [Andy Polyakov <appro@fy.chalmers.se>]
7746 *) Sparc v8plus assembler for the bignum library.
7747 [Andy Polyakov <appro@fy.chalmers.se>]
7749 *) Accept any -xxx and +xxx compiler options in Configure.
7752 *) Update HPUX configuration.
7755 *) Add missing sk_<type>_unshift() function to safestack.h
7756 [Ralf S. Engelschall]
7758 *) New function SSL_CTX_use_certificate_chain_file that sets the
7759 "extra_cert"s in addition to the certificate. (This makes sense
7760 only for "PEM" format files, as chains as a whole are not
7764 *) Support verify_depth from the SSL API.
7765 x509_vfy.c had what can be considered an off-by-one-error:
7766 Its depth (which was not part of the external interface)
7767 was actually counting the number of certificates in a chain;
7768 now it really counts the depth.
7771 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7772 instead of X509err, which often resulted in confusing error
7773 messages since the error codes are not globally unique
7774 (e.g. an alleged error in ssl3_accept when a certificate
7775 didn't match the private key).
7777 *) New function SSL_CTX_set_session_id_context that allows to set a default
7778 value (so that you don't need SSL_set_session_id_context for each
7779 connection using the SSL_CTX).
7782 *) OAEP decoding bug fix.
7785 *) Support INSTALL_PREFIX for package builders, as proposed by
7789 *) New Configure options "threads" and "no-threads". For systems
7790 where the proper compiler options are known (currently Solaris
7791 and Linux), "threads" is the default.
7794 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7797 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7798 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7799 such as /usr/local/bin.
7802 *) "make linux-shared" to build shared libraries.
7803 [Niels Poppe <niels@netbox.org>]
7805 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7808 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7809 extension adding in x509 utility.
7812 *) Remove NOPROTO sections and error code comments.
7815 *) Partial rewrite of the DEF file generator to now parse the ANSI
7819 *) New Configure options --prefix=DIR and --openssldir=DIR.
7822 *) Complete rewrite of the error code script(s). It is all now handled
7823 by one script at the top level which handles error code gathering,
7824 header rewriting and C source file generation. It should be much better
7825 than the old method: it now uses a modified version of Ulf's parser to
7826 read the ANSI prototypes in all header files (thus the old K&R definitions
7827 aren't needed for error creation any more) and do a better job of
7828 translating function codes into names. The old 'ASN1 error code imbedded
7829 in a comment' is no longer necessary and it doesn't use .err files which
7830 have now been deleted. Also the error code call doesn't have to appear all
7831 on one line (which resulted in some large lines...).
7834 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7837 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7838 0 (which usually indicates a closed connection), but continue reading.
7841 *) Fix some race conditions.
7844 *) Add support for CRL distribution points extension. Add Certificate
7845 Policies and CRL distribution points documentation.
7848 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7851 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7852 8 of keying material. Merlin has also confirmed interop with this fix
7853 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7854 [Merlin Hughes <merlin@baltimore.ie>]
7856 *) Fix lots of warnings.
7857 [Richard Levitte <levitte@stacken.kth.se>]
7859 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7860 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7861 [Richard Levitte <levitte@stacken.kth.se>]
7863 *) Fix problems with sizeof(long) == 8.
7864 [Andy Polyakov <appro@fy.chalmers.se>]
7866 *) Change functions to ANSI C.
7869 *) Fix typos in error codes.
7870 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7872 *) Remove defunct assembler files from Configure.
7875 *) SPARC v8 assembler BIGNUM implementation.
7876 [Andy Polyakov <appro@fy.chalmers.se>]
7878 *) Support for Certificate Policies extension: both print and set.
7879 Various additions to support the r2i method this uses.
7882 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7883 return a const string when you are expecting an allocated buffer.
7886 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7887 types DirectoryString and DisplayText.
7890 *) Add code to allow r2i extensions to access the configuration database,
7891 add an LHASH database driver and add several ctx helper functions.
7894 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7895 fail when they extended the size of a BIGNUM.
7898 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7899 support typesafe stack.
7902 *) Fix typo in SSL_[gs]et_options().
7903 [Nils Frostberg <nils@medcom.se>]
7905 *) Delete various functions and files that belonged to the (now obsolete)
7906 old X509V3 handling code.
7909 *) New Configure option "rsaref".
7912 *) Don't auto-generate pem.h.
7915 *) Introduce type-safe ASN.1 SETs.
7918 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7919 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7921 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7922 that links with OpenSSL (well at least cause lots of warnings), but fear
7923 not: the conversion is trivial, and it eliminates loads of evil casts. A
7924 few STACKed things have been converted already. Feel free to convert more.
7925 In the fullness of time, I'll do away with the STACK type altogether.
7928 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7929 specified in <certfile> by updating the entry in the index.txt file.
7930 This way one no longer has to edit the index.txt file manually for
7931 revoking a certificate. The -revoke option does the gory details now.
7932 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7934 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7935 `-text' option at all and this way the `-noout -text' combination was
7936 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7937 [Ralf S. Engelschall]
7939 *) Make sure a corresponding plain text error message exists for the
7940 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7941 verify callback function determined that a certificate was revoked.
7942 [Ralf S. Engelschall]
7944 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7945 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7946 all available cipers including rc5, which was forgotten until now.
7947 In order to let the testing shell script know which algorithms
7948 are available, a new (up to now undocumented) command
7949 "openssl list-cipher-commands" is used.
7952 *) Bugfix: s_client occasionally would sleep in select() when
7953 it should have checked SSL_pending() first.
7956 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7957 the raw DSA values prior to ASN.1 encoding.
7960 *) Tweaks to Configure
7961 [Niels Poppe <niels@netbox.org>]
7963 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7967 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7970 *) New config option to avoid instructions that are illegal on the 80386.
7971 The default code is faster, but requires at least a 486.
7974 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7975 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7976 same as SSL2_VERSION anyway.
7979 *) New "-showcerts" option for s_client.
7982 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7983 application. Various cleanups and fixes.
7986 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7987 modify error routines to work internally. Add error codes and PBE init
7988 to library startup routines.
7991 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7992 packing functions to asn1 and evp. Changed function names and error
7993 codes along the way.
7996 *) PKCS12 integration: and so it begins... First of several patches to
7997 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7998 objects to objects.h
8001 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8002 and display support for Thawte strong extranet extension.
8005 *) Add LinuxPPC support.
8006 [Jeff Dubrule <igor@pobox.org>]
8008 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8009 bn_div_words in alpha.s.
8010 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8012 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8013 OAEP isn't supported when OpenSSL is built with RSAref.
8014 [Ulf Moeller <ulf@fitug.de>]
8016 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8017 so they no longer are missing under -DNOPROTO.
8018 [Soren S. Jorvang <soren@t.dk>]
8021 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8023 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8024 doesn't work when the session is reused. Coming soon!
8027 *) Fix a security hole, that allows sessions to be reused in the wrong
8028 context thus bypassing client cert protection! All software that uses
8029 client certs and session caches in multiple contexts NEEDS PATCHING to
8030 allow session reuse! A fuller solution is in the works.
8031 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8033 *) Some more source tree cleanups (removed obsolete files
8034 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8035 permission on "config" script to be executable) and a fix for the INSTALL
8037 [Ulf Moeller <ulf@fitug.de>]
8039 *) Remove some legacy and erroneous uses of malloc, free instead of
8041 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8043 *) Make rsa_oaep_test return non-zero on error.
8044 [Ulf Moeller <ulf@fitug.de>]
8046 *) Add support for native Solaris shared libraries. Configure
8047 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8048 if someone would make that last step automatic.
8049 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8051 *) ctx_size was not built with the right compiler during "make links". Fixed.
8054 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8055 except NULL ciphers". This means the default cipher list will no longer
8056 enable NULL ciphers. They need to be specifically enabled e.g. with
8057 the string "DEFAULT:eNULL".
8060 *) Fix to RSA private encryption routines: if p < q then it would
8061 occasionally produce an invalid result. This will only happen with
8062 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8065 *) Be less restrictive and allow also `perl util/perlpath.pl
8066 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8067 because this way one can also use an interpreter named `perl5' (which is
8068 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8069 installed as `perl').
8070 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8072 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8073 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8075 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8076 advapi32.lib to Win32 build and change the pem test comparision
8077 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8078 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8079 and crypto/des/ede_cbcm_enc.c.
8082 *) DES quad checksum was broken on big-endian architectures. Fixed.
8085 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8086 Win32 test batch file so it (might) work again. The Win32 test batch file
8087 is horrible: I feel ill....
8090 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8091 in e_os.h. Audit of header files to check ANSI and non ANSI
8092 sections: 10 functions were absent from non ANSI section and not exported
8093 from Windows DLLs. Fixed up libeay.num for new functions.
8096 *) Make `openssl version' output lines consistent.
8097 [Ralf S. Engelschall]
8099 *) Fix Win32 symbol export lists for BIO functions: Added
8100 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8101 to ms/libeay{16,32}.def.
8102 [Ralf S. Engelschall]
8104 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8105 fine under Unix and passes some trivial tests I've now added. But the
8106 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8107 added to make sure no one expects that this stuff really works in the
8108 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8109 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8111 [Ralf S. Engelschall]
8113 *) Fix the generation of two part addresses in perl.
8114 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8116 *) Add config entry for Linux on MIPS.
8117 [John Tobey <jtobey@channel1.com>]
8119 *) Make links whenever Configure is run, unless we are on Windoze.
8122 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8123 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8127 *) Add a useful kludge to allow package maintainers to specify compiler and
8128 other platforms details on the command line without having to patch the
8129 Configure script everytime: One now can use ``perl Configure
8130 <id>:<details>'', i.e. platform ids are allowed to have details appended
8131 to them (seperated by colons). This is treated as there would be a static
8132 pre-configured entry in Configure's %table under key <id> with value
8133 <details> and ``perl Configure <id>'' is called. So, when you want to
8134 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8135 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8136 now, which overrides the FreeBSD-elf entry on-the-fly.
8137 [Ralf S. Engelschall]
8139 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8142 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8143 on the `perl Configure ...' command line. This way one can compile
8144 OpenSSL libraries with Position Independent Code (PIC) which is needed
8145 for linking it into DSOs.
8146 [Ralf S. Engelschall]
8148 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8152 *) Cleaned up the LICENSE document: The official contact for any license
8153 questions now is the OpenSSL core team under openssl-core@openssl.org.
8154 And add a paragraph about the dual-license situation to make sure people
8155 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8156 to the OpenSSL toolkit.
8157 [Ralf S. Engelschall]
8159 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8160 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8161 Additonally cleaned up the `make links' target: Remove unnecessary
8162 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8163 to speed processing and no longer clutter the display with confusing
8164 stuff. Instead only the actually done links are displayed.
8165 [Ralf S. Engelschall]
8167 *) Permit null encryption ciphersuites, used for authentication only. It used
8168 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8169 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8173 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8174 signed attributes when verifying signatures (this would break them),
8175 the detached data encoding was wrong and public keys obtained using
8176 X509_get_pubkey() weren't freed.
8179 *) Add text documentation for the BUFFER functions. Also added a work around
8180 to a Win95 console bug. This was triggered by the password read stuff: the
8181 last character typed gets carried over to the next fread(). If you were
8182 generating a new cert request using 'req' for example then the last
8183 character of the passphrase would be CR which would then enter the first
8187 *) Added the new `Includes OpenSSL Cryptography Software' button as
8188 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8189 button and can be used by applications based on OpenSSL to show the
8190 relationship to the OpenSSL project.
8191 [Ralf S. Engelschall]
8193 *) Remove confusing variables in function signatures in files
8194 ssl/ssl_lib.c and ssl/ssl.h.
8195 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8197 *) Don't install bss_file.c under PREFIX/include/
8198 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8200 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8201 functions that return function pointers and has support for NT specific
8202 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8203 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8204 unsigned to signed types: this was killing the Win32 compile.
8207 *) Add new certificate file to stack functions,
8208 SSL_add_dir_cert_subjects_to_stack() and
8209 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8210 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8211 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8212 This means that Apache-SSL and similar packages don't have to mess around
8213 to add as many CAs as they want to the preferred list.
8216 *) Experiment with doxygen documentation. Currently only partially applied to
8218 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8219 openssl.doxy as the configuration file.
8222 *) Get rid of remaining C++-style comments which strict C compilers hate.
8223 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8225 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8226 compiled in by default: it has problems with large keys.
8229 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8230 DH private keys and/or callback functions which directly correspond to
8231 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8232 is needed for applications which have to configure certificates on a
8233 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8235 For the RSA certificate situation is makes no difference, but
8236 for the DSA certificate situation this fixes the "no shared cipher"
8237 problem where the OpenSSL cipher selection procedure failed because the
8238 temporary keys were not overtaken from the context and the API provided
8239 no way to reconfigure them.
8240 The new functions now let applications reconfigure the stuff and they
8241 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8242 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8243 non-public-API function ssl_cert_instantiate() is used as a helper
8244 function and also to reduce code redundancy inside ssl_rsa.c.
8245 [Ralf S. Engelschall]
8247 *) Move s_server -dcert and -dkey options out of the undocumented feature
8248 area because they are useful for the DSA situation and should be
8249 recognized by the users.
8250 [Ralf S. Engelschall]
8252 *) Fix the cipher decision scheme for export ciphers: the export bits are
8253 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8254 SSL_EXP_MASK. So, the original variable has to be used instead of the
8255 already masked variable.
8256 [Richard Levitte <levitte@stacken.kth.se>]
8258 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8259 [Richard Levitte <levitte@stacken.kth.se>]
8261 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8262 from `int' to `unsigned int' because it's a length and initialized by
8263 EVP_DigestFinal() which expects an `unsigned int *'.
8264 [Richard Levitte <levitte@stacken.kth.se>]
8266 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8267 script. Instead use the usual Shell->Perl transition trick.
8268 [Ralf S. Engelschall]
8270 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8271 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8272 -noout -modulus' as it's already the case for `openssl rsa -noout
8273 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8274 currently the public key is printed (a decision which was already done by
8275 `openssl dsa -modulus' in the past) which serves a similar purpose.
8276 Additionally the NO_RSA no longer completely removes the whole -modulus
8277 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8279 [Ralf S. Engelschall]
8281 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8282 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8283 [Arne Ansper <arne@ats.cyber.ee>]
8285 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8286 to be added. Now both 'req' and 'ca' can use new objects defined in the
8290 *) Add cool BIO that does syslog (or event log on NT).
8291 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8293 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8294 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8295 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8296 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8299 *) Add preliminary config info for new extension code.
8302 *) Make RSA_NO_PADDING really use no padding.
8303 [Ulf Moeller <ulf@fitug.de>]
8305 *) Generate errors when private/public key check is done.
8308 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8309 for some CRL extensions and new objects added.
8312 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8313 key usage extension and fuller support for authority key id.
8316 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8317 padding method for RSA, which is recommended for new applications in PKCS
8318 #1 v2.0 (RFC 2437, October 1998).
8319 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8320 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8321 against Bleichbacher's attack on RSA.
8322 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8325 *) Updates to the new SSL compression code
8326 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8328 *) Fix so that the version number in the master secret, when passed
8329 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8330 (because the server will not accept higher), that the version number
8331 is 0x03,0x01, not 0x03,0x00
8332 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8334 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8335 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8336 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8339 *) Support for RAW extensions where an arbitrary extension can be
8340 created by including its DER encoding. See apps/openssl.cnf for
8344 *) Make sure latest Perl versions don't interpret some generated C array
8345 code as Perl array code in the crypto/err/err_genc.pl script.
8346 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8348 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8349 not many people have the assembler. Various Win32 compilation fixes and
8350 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8354 *) Modify configure script 'Configure' to automatically create crypto/date.h
8355 file under Win32 and also build pem.h from pem.org. New script
8356 util/mkfiles.pl to create the MINFO file on environments that can't do a
8357 'make files': perl util/mkfiles.pl >MINFO should work.
8360 *) Major rework of DES function declarations, in the pursuit of correctness
8361 and purity. As a result, many evil casts evaporated, and some weirdness,
8362 too. You may find this causes warnings in your code. Zapping your evil
8363 casts will probably fix them. Mostly.
8366 *) Fix for a typo in asn1.h. Bug fix to object creation script
8367 obj_dat.pl. It considered a zero in an object definition to mean
8368 "end of object": none of the objects in objects.h have any zeros
8369 so it wasn't spotted.
8370 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8372 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8373 Masking (CBCM). In the absence of test vectors, the best I have been able
8374 to do is check that the decrypt undoes the encrypt, so far. Send me test
8375 vectors if you have them.
8378 *) Correct calculation of key length for export ciphers (too much space was
8379 allocated for null ciphers). This has not been tested!
8382 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8383 message is now correct (it understands "crypto" and "ssl" on its
8384 command line). There is also now an "update" option. This will update
8385 the util/ssleay.num and util/libeay.num files with any new functions.
8387 perl util/mkdef.pl crypto ssl update
8388 it will update them.
8391 *) Overhauled the Perl interface (perl/*):
8392 - ported BN stuff to OpenSSL's different BN library
8393 - made the perl/ source tree CVS-aware
8394 - renamed the package from SSLeay to OpenSSL (the files still contain
8395 their history because I've copied them in the repository)
8396 - removed obsolete files (the test scripts will be replaced
8397 by better Test::Harness variants in the future)
8398 [Ralf S. Engelschall]
8400 *) First cut for a very conservative source tree cleanup:
8401 1. merge various obsolete readme texts into doc/ssleay.txt
8402 where we collect the old documents and readme texts.
8403 2. remove the first part of files where I'm already sure that we no
8404 longer need them because of three reasons: either they are just temporary
8405 files which were left by Eric or they are preserved original files where
8406 I've verified that the diff is also available in the CVS via "cvs diff
8407 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8408 the crypto/md/ stuff).
8409 [Ralf S. Engelschall]
8411 *) More extension code. Incomplete support for subject and issuer alt
8412 name, issuer and authority key id. Change the i2v function parameters
8413 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8414 what that's for :-) Fix to ASN1 macro which messed up
8415 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8418 *) Preliminary support for ENUMERATED type. This is largely copied from the
8422 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8423 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8425 *) Make sure `make rehash' target really finds the `openssl' program.
8426 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8428 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8429 like to hear about it if this slows down other processors.
8432 *) Add CygWin32 platform information to Configure script.
8433 [Alan Batie <batie@aahz.jf.intel.com>]
8435 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8436 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8438 *) New program nseq to manipulate netscape certificate sequences
8441 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8445 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8446 but the BN code had some problems that would cause failures when
8447 doing certificate verification and some other functions.
8448 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8450 *) Add ASN1 and PEM code to support netscape certificate sequences.
8453 *) Add ASN1 and PEM code to support netscape certificate sequences.
8456 *) Add several PKIX and private extended key usage OIDs.
8459 *) Modify the 'ca' program to handle the new extension code. Modify
8460 openssl.cnf for new extension format, add comments.
8463 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8464 and add a sample to openssl.cnf so req -x509 now adds appropriate
8468 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8469 error code, add initial support to X509_print() and x509 application.
8472 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8473 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8474 stuff is currently isolated and isn't even compiled yet.
8477 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8478 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8479 Removed the versions check from X509 routines when loading extensions:
8480 this allows certain broken certificates that don't set the version
8481 properly to be processed.
8484 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8485 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8486 can still be regenerated with "make depend".
8489 *) Spelling mistake in C version of CAST-128.
8490 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8492 *) Changes to the error generation code. The perl script err-code.pl
8493 now reads in the old error codes and retains the old numbers, only
8494 adding new ones if necessary. It also only changes the .err files if new
8495 codes are added. The makefiles have been modified to only insert errors
8496 when needed (to avoid needlessly modifying header files). This is done
8497 by only inserting errors if the .err file is newer than the auto generated
8498 C file. To rebuild all the error codes from scratch (the old behaviour)
8499 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8500 or delete all the .err files.
8503 *) CAST-128 was incorrectly implemented for short keys. The C version has
8504 been fixed, but is untested. The assembler versions are also fixed, but
8505 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8506 to regenerate it if needed.
8507 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8508 Hagino <itojun@kame.net>]
8510 *) File was opened incorrectly in randfile.c.
8511 [Ulf Möller <ulf@fitug.de>]
8513 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8514 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8515 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8516 al: it's just almost always a UTCTime. Note this patch adds new error
8517 codes so do a "make errors" if there are problems.
8520 *) Correct Linux 1 recognition in config.
8521 [Ulf Möller <ulf@fitug.de>]
8523 *) Remove pointless MD5 hash when using DSA keys in ca.
8524 [Anonymous <nobody@replay.com>]
8526 *) Generate an error if given an empty string as a cert directory. Also
8527 generate an error if handed NULL (previously returned 0 to indicate an
8528 error, but didn't set one).
8529 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8531 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8534 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8535 parameters. This was causing a warning which killed off the Win32 compile.
8538 *) Remove C++ style comments from crypto/bn/bn_local.h.
8539 [Neil Costigan <neil.costigan@celocom.com>]
8541 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8542 based on a text string, looking up short and long names and finally
8543 "dot" format. The "dot" format stuff didn't work. Added new function
8544 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8545 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8546 OID is not part of the table.
8549 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8550 X509_LOOKUP_by_alias().
8553 *) Sort openssl functions by name.
8556 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8557 encryption from sample DSA keys (in case anyone is interested the password
8561 *) Make _all_ *_free functions accept a NULL pointer.
8562 [Frans Heymans <fheymans@isaserver.be>]
8564 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8566 [Anonymous <nobody@replay.com>]
8568 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8569 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8571 *) Don't blow it for numeric -newkey arguments to apps/req.
8572 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8574 *) Temp key "for export" tests were wrong in s3_srvr.c.
8575 [Anonymous <nobody@replay.com>]
8577 *) Add prototype for temp key callback functions
8578 SSL_CTX_set_tmp_{rsa,dh}_callback().
8581 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8582 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8585 *) X509_name_add_entry() freed the wrong thing after an error.
8586 [Arne Ansper <arne@ats.cyber.ee>]
8588 *) rsa_eay.c would attempt to free a NULL context.
8589 [Arne Ansper <arne@ats.cyber.ee>]
8591 *) BIO_s_socket() had a broken should_retry() on Windoze.
8592 [Arne Ansper <arne@ats.cyber.ee>]
8594 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8595 [Arne Ansper <arne@ats.cyber.ee>]
8597 *) Make sure the already existing X509_STORE->depth variable is initialized
8598 in X509_STORE_new(), but document the fact that this variable is still
8599 unused in the certificate verification process.
8600 [Ralf S. Engelschall]
8602 *) Fix the various library and apps files to free up pkeys obtained from
8603 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8606 *) Fix reference counting in X509_PUBKEY_get(). This makes
8607 demos/maurice/example2.c work, amongst others, probably.
8608 [Steve Henson and Ben Laurie]
8610 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8611 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8612 are no longer created. This way we have a single and consistent command
8613 line interface `openssl <command>', similar to `cvs <command>'.
8614 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8616 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8617 BIT STRING wrapper always have zero unused bits.
8620 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8623 *) Make the top-level INSTALL documentation easier to understand.
8626 *) Makefiles updated to exit if an error occurs in a sub-directory
8627 make (including if user presses ^C) [Paul Sutton]
8629 *) Make Montgomery context stuff explicit in RSA data structure.
8632 *) Fix build order of pem and err to allow for generated pem.h.
8635 *) Fix renumbering bug in X509_NAME_delete_entry().
8638 *) Enhanced the err-ins.pl script so it makes the error library number
8639 global and can add a library name. This is needed for external ASN1 and
8640 other error libraries.
8643 *) Fixed sk_insert which never worked properly.
8646 *) Fix ASN1 macros so they can handle indefinite length construted
8647 EXPLICIT tags. Some non standard certificates use these: they can now
8651 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8652 into a single doc/ssleay.txt bundle. This way the information is still
8653 preserved but no longer messes up this directory. Now it's new room for
8654 the new set of documenation files.
8655 [Ralf S. Engelschall]
8657 *) SETs were incorrectly DER encoded. This was a major pain, because they
8658 shared code with SEQUENCEs, which aren't coded the same. This means that
8659 almost everything to do with SETs or SEQUENCEs has either changed name or
8660 number of arguments.
8661 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8663 *) Fix test data to work with the above.
8666 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8667 was already fixed by Eric for 0.9.1 it seems.
8668 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8670 *) Autodetect FreeBSD3.
8673 *) Fix various bugs in Configure. This affects the following platforms:
8677 unixware-2.0-pentium
8681 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8682 before they are needed.
8685 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8689 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8691 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8692 changed SSLeay to OpenSSL in version strings.
8693 [Ralf S. Engelschall]
8695 *) Some fixups to the top-level documents.
8698 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8699 because the symlink to include/ was missing.
8700 [Ralf S. Engelschall]
8702 *) Incorporated the popular no-RSA/DSA-only patches
8703 which allow to compile a RSA-free SSLeay.
8704 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8706 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8707 when "ssleay" is still not found.
8708 [Ralf S. Engelschall]
8710 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8711 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8713 *) Updated the README file.
8714 [Ralf S. Engelschall]
8716 *) Added various .cvsignore files in the CVS repository subdirs
8717 to make a "cvs update" really silent.
8718 [Ralf S. Engelschall]
8720 *) Recompiled the error-definition header files and added
8721 missing symbols to the Win32 linker tables.
8722 [Ralf S. Engelschall]
8724 *) Cleaned up the top-level documents;
8725 o new files: CHANGES and LICENSE
8726 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8727 o merged COPYRIGHT into LICENSE
8728 o removed obsolete TODO file
8729 o renamed MICROSOFT to INSTALL.W32
8730 [Ralf S. Engelschall]
8732 *) Removed dummy files from the 0.9.1b source tree:
8733 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8734 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8735 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8736 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8737 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8738 [Ralf S. Engelschall]
8740 *) Added various platform portability fixes.
8743 *) The Genesis of the OpenSSL rpject:
8744 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8745 Young and Tim J. Hudson created while they were working for C2Net until
8747 [The OpenSSL Project]
8750 Changes between 0.9.0b and 0.9.1b [not released]
8752 *) Updated a few CA certificates under certs/
8755 *) Changed some BIGNUM api stuff.
8758 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8759 DGUX x86, Linux Alpha, etc.
8762 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8763 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8767 *) Add -strparse option to asn1pars program which parses nested
8769 [Dr Stephen Henson <shenson@bigfoot.com>]
8771 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8774 *) DSA fix for "ca" program.
8777 *) Added "-genkey" option to "dsaparam" program.
8780 *) Added RIPE MD160 (rmd160) message digest.
8783 *) Added -a (all) option to "ssleay version" command.
8786 *) Added PLATFORM define which is the id given to Configure.
8789 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8792 *) Extended the ASN.1 parser routines.
8795 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8798 *) Added a BN_CTX to the BN library.
8801 *) Fixed the weak key values in DES library
8804 *) Changed API in EVP library for cipher aliases.
8807 *) Added support for RC2/64bit cipher.
8810 *) Converted the lhash library to the crypto/mem.c functions.
8813 *) Added more recognized ASN.1 object ids.
8816 *) Added more RSA padding checks for SSL/TLS.
8819 *) Added BIO proxy/filter functionality.
8822 *) Added extra_certs to SSL_CTX which can be used
8823 send extra CA certificates to the client in the CA cert chain sending
8824 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8827 *) Now Fortezza is denied in the authentication phase because
8828 this is key exchange mechanism is not supported by SSLeay at all.
8831 *) Additional PKCS1 checks.
8834 *) Support the string "TLSv1" for all TLS v1 ciphers.
8837 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8838 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8841 *) Fixed a few memory leaks.
8844 *) Fixed various code and comment typos.
8847 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8848 bytes sent in the client random.
8849 [Edward Bishop <ebishop@spyglass.com>]