5 Changes between 0.9.8x and 0.9.8y [xx XXX xxxx]
7 *) Return an error when checking OCSP signatures when key is NULL.
8 This fixes a DoS attack. (CVE-2013-0166)
11 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
12 the right response is stapled. Also change SSL_get_certificate()
13 so it returns the certificate actually sent.
14 See http://rt.openssl.org/Ticket/Display.html?id=2836.
16 [Rob Stradling <rob.stradling@comodo.com>]
18 *) Fix possible deadlock when decoding public keys.
21 Changes between 0.9.8w and 0.9.8x [10 May 2012]
23 *) Sanity check record length before skipping explicit IV in DTLS
26 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
27 fuzzing as a service testing platform.
31 *) Initialise tkeylen properly when encrypting CMS messages.
32 Thanks to Solar Designer of Openwall for reporting this issue.
35 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
37 *) The fix for CVE-2012-2110 did not take into account that the
38 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
39 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
40 rejecting negative len parameter. (CVE-2012-2131)
41 [Tomas Hoger <thoger@redhat.com>]
43 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
45 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
46 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
47 in CRYPTO_realloc_clean.
49 Thanks to Tavis Ormandy, Google Security Team, for discovering this
50 issue and to Adam Langley <agl@chromium.org> for fixing it.
52 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
54 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
56 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
57 in CMS and PKCS7 code. When RSA decryption fails use a random key for
58 content decryption and always return the same error. Note: this attack
59 needs on average 2^20 messages so it only affects automated senders. The
60 old behaviour can be reenabled in the CMS code by setting the
61 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
62 an MMA defence is not necessary.
63 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
64 this issue. (CVE-2012-0884)
67 *) Fix CVE-2011-4619: make sure we really are receiving a
68 client hello before rejecting multiple SGC restarts. Thanks to
69 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
72 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
74 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
75 Thanks to Antonio Martin, Enterprise Secure Access Research and
76 Development, Cisco Systems, Inc. for discovering this bug and
77 preparing a fix. (CVE-2012-0050)
80 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
82 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
83 of the Vaudenay padding oracle attack on CBC mode encryption
84 which enables an efficient plaintext recovery attack against
85 the OpenSSL implementation of DTLS. Their attack exploits timing
86 differences arising during decryption processing. A research
87 paper describing this attack can be found at:
88 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
89 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
90 Security Group at Royal Holloway, University of London
91 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
92 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
93 for preparing the fix. (CVE-2011-4108)
94 [Robin Seggelmann, Michael Tuexen]
96 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
97 [Ben Laurie, Kasper <ekasper@google.com>]
99 *) Clear bytes used for block padding of SSL 3.0 records.
101 [Adam Langley (Google)]
103 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
104 Kadianakis <desnacked@gmail.com> for discovering this issue and
105 Adam Langley for preparing the fix. (CVE-2011-4619)
106 [Adam Langley (Google)]
108 *) Prevent malformed RFC3779 data triggering an assertion failure.
109 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
110 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
111 [Rob Austein <sra@hactrn.net>]
113 *) Fix ssl_ciph.c set-up race.
114 [Adam Langley (Google)]
116 *) Fix spurious failures in ecdsatest.c.
117 [Emilia Käsper (Google)]
119 *) Fix the BIO_f_buffer() implementation (which was mixing different
120 interpretations of the '..._len' fields).
121 [Adam Langley (Google)]
123 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
124 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
125 threads won't reuse the same blinding coefficients.
127 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
128 lock to call BN_BLINDING_invert_ex, and avoids one use of
129 BN_BLINDING_update for each BN_BLINDING structure (previously,
130 the last update always remained unused).
131 [Emilia Käsper (Google)]
133 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
134 for multi-threaded use of ECDH.
135 [Adam Langley (Google)]
137 *) Fix x509_name_ex_d2i memory leak on bad inputs.
140 *) Add protection against ECDSA timing attacks as mentioned in the paper
141 by Billy Bob Brumley and Nicola Tuveri, see:
143 http://eprint.iacr.org/2011/232.pdf
145 [Billy Bob Brumley and Nicola Tuveri]
147 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
149 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
150 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
152 *) Fix bug in string printing code: if *any* escaping is enabled we must
153 escape the escape character (backslash) or the resulting string is
157 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
159 *) Disable code workaround for ancient and obsolete Netscape browsers
160 and servers: an attacker can use it in a ciphersuite downgrade attack.
161 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
164 *) Fixed J-PAKE implementation error, originally discovered by
165 Sebastien Martini, further info and confirmation from Stefan
166 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
169 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
171 *) Fix extension code to avoid race conditions which can result in a buffer
172 overrun vulnerability: resumed sessions must not be modified as they can
173 be shared by multiple threads. CVE-2010-3864
176 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
179 *) Don't reencode certificate when calculating signature: cache and use
180 the original encoding instead. This makes signature verification of
181 some broken encodings work correctly.
184 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
185 is also one of the inputs.
186 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
188 *) Don't repeatedly append PBE algorithms to table if they already exist.
189 Sort table on each new add. This effectively makes the table read only
190 after all algorithms are added and subsequent calls to PKCS12_pbe_add
194 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
196 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
199 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
200 access or freeing data twice (CVE-2010-0742)
201 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
203 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
204 common in certificates and some applications which only call
205 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
209 Reduce copying into .apps and .test in makevms.com
210 Don't try to use blank CA certificate in CA.com
211 Allow use of C files from original directories in maketests.com
212 [Steven M. Schweda" <sms@antinode.info>]
214 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
216 *) When rejecting SSL/TLS records due to an incorrect version number, never
217 update s->server with a new major version number. As of
218 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
219 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
220 the previous behavior could result in a read attempt at NULL when
221 receiving specific incorrect SSL/TLS records once record payload
222 protection is active. (CVE-2010-0740)
223 [Bodo Moeller, Adam Langley <agl@chromium.org>]
225 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
226 could be crashed if the relevant tables were not present (e.g. chrooted).
227 [Tomas Hoger <thoger@redhat.com>]
229 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
231 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
232 [Martin Olsson, Neel Mehta]
234 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
235 accommodate for stack sorting, always a write lock!).
238 *) On some versions of WIN32 Heap32Next is very slow. This can cause
239 excessive delays in the RAND_poll(): over a minute. As a workaround
240 include a time check in the inner Heap32Next loop too.
243 *) The code that handled flushing of data in SSL/TLS originally used the
244 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
245 the problem outlined in PR#1949. The fix suggested there however can
246 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
247 of Apache). So instead simplify the code to flush unconditionally.
248 This should be fine since flushing with no data to flush is a no op.
251 *) Handle TLS versions 2.0 and later properly and correctly use the
252 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
253 off ancient servers have a habit of sticking around for a while...
256 *) Modify compression code so it frees up structures without using the
257 ex_data callbacks. This works around a problem where some applications
258 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
259 restarting) then use compression (e.g. SSL with compression) later.
260 This results in significant per-connection memory leaks and
261 has caused some security issues including CVE-2008-1678 and
265 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
266 change when encrypting or decrypting.
269 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
270 connect and renegotiate with servers which do not support RI.
271 Until RI is more widely deployed this option is enabled by default.
274 *) Add "missing" ssl ctrls to clear options and mode.
277 *) If client attempts to renegotiate and doesn't support RI respond with
278 a no_renegotiation alert as required by RFC5746. Some renegotiating
279 TLS clients will continue a connection gracefully when they receive
280 the alert. Unfortunately OpenSSL mishandled this alert and would hang
281 waiting for a server hello which it will never receive. Now we treat a
282 received no_renegotiation alert as a fatal error. This is because
283 applications requesting a renegotiation might well expect it to succeed
284 and would have no code in place to handle the server denying it so the
285 only safe thing to do is to terminate the connection.
288 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
289 peer supports secure renegotiation and 0 otherwise. Print out peer
290 renegotiation support in s_client/s_server.
293 *) Replace the highly broken and deprecated SPKAC certification method with
294 the updated NID creation version. This should correctly handle UTF8.
297 *) Implement RFC5746. Re-enable renegotiation but require the extension
298 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
299 turns out to be a bad idea. It has been replaced by
300 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
301 SSL_CTX_set_options(). This is really not recommended unless you
302 know what you are doing.
303 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
305 *) Fixes to stateless session resumption handling. Use initial_ctx when
306 issuing and attempting to decrypt tickets in case it has changed during
307 servername handling. Use a non-zero length session ID when attempting
308 stateless session resumption: this makes it possible to determine if
309 a resumption has occurred immediately after receiving server hello
310 (several places in OpenSSL subtly assume this) instead of later in
314 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
315 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
316 fixes for a few places where the return code is not checked
318 [Julia Lawall <julia@diku.dk>]
320 *) Add --strict-warnings option to Configure script to include devteam
321 warnings in other configurations.
324 *) Add support for --libdir option and LIBDIR variable in makefiles. This
325 makes it possible to install openssl libraries in locations which
326 have names other than "lib", for example "/usr/lib64" which some
328 [Steve Henson, based on patch from Jeremy Utley]
330 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
331 X690 8.9.12 and can produce some misleading textual output of OIDs.
332 [Steve Henson, reported by Dan Kaminsky]
334 *) Delete MD2 from algorithm tables. This follows the recommendation in
335 several standards that it is not used in new applications due to
336 several cryptographic weaknesses. For binary compatibility reasons
337 the MD2 API is still compiled in by default.
340 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
344 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
345 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
347 [Guenter <lists@gknw.net>]
349 *) Fix the server certificate chain building code to use X509_verify_cert(),
350 it used to have an ad-hoc builder which was unable to cope with anything
351 other than a simple chain.
352 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
354 *) Don't check self signed certificate signatures in X509_verify_cert()
355 by default (a flag can override this): it just wastes time without
356 adding any security. As a useful side effect self signed root CAs
357 with non-FIPS digests are now usable in FIPS mode.
360 *) In dtls1_process_out_of_seq_message() the check if the current message
361 is already buffered was missing. For every new message was memory
362 allocated, allowing an attacker to perform an denial of service attack
363 with sending out of seq handshake messages until there is no memory
364 left. Additionally every future messege was buffered, even if the
365 sequence number made no sense and would be part of another handshake.
366 So only messages with sequence numbers less than 10 in advance will be
367 buffered. (CVE-2009-1378)
368 [Robin Seggelmann, discovered by Daniel Mentz]
370 *) Records are buffered if they arrive with a future epoch to be
371 processed after finishing the corresponding handshake. There is
372 currently no limitation to this buffer allowing an attacker to perform
373 a DOS attack with sending records with future epochs until there is no
374 memory left. This patch adds the pqueue_size() function to detemine
375 the size of a buffer and limits the record buffer to 100 entries.
377 [Robin Seggelmann, discovered by Daniel Mentz]
379 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
380 parent structure is freed. (CVE-2009-1379)
383 *) Handle non-blocking I/O properly in SSL_shutdown() call.
384 [Darryl Miles <darryl-mailinglists@netbauds.net>]
387 [Ilya O. <vrghost@gmail.com>]
389 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
391 *) Disable renegotiation completely - this fixes a severe security
392 problem (CVE-2009-3555) at the cost of breaking all
393 renegotiation. Renegotiation can be re-enabled by setting
394 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
395 run-time. This is really not recommended unless you know what
399 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
401 *) Don't set val to NULL when freeing up structures, it is freed up by
402 underlying code. If sizeof(void *) > sizeof(long) this can result in
403 zeroing past the valid field. (CVE-2009-0789)
404 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
406 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
407 checked correctly. This would allow some invalid signed attributes to
408 appear to verify correctly. (CVE-2009-0591)
409 [Ivan Nestlerode <inestlerode@us.ibm.com>]
411 *) Reject UniversalString and BMPString types with invalid lengths. This
412 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
413 a legal length. (CVE-2009-0590)
416 *) Set S/MIME signing as the default purpose rather than setting it
417 unconditionally. This allows applications to override it at the store
421 *) Permit restricted recursion of ASN1 strings. This is needed in practice
422 to handle some structures.
425 *) Improve efficiency of mem_gets: don't search whole buffer each time
427 [Jeremy Shapiro <jnshapir@us.ibm.com>]
429 *) New -hex option for openssl rand.
432 *) Print out UTF8String and NumericString when parsing ASN1.
435 *) Support NumericString type for name components.
438 *) Allow CC in the environment to override the automatically chosen
439 compiler. Note that nothing is done to ensure flags work with the
443 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
445 *) Properly check EVP_VerifyFinal() and similar return values
447 [Ben Laurie, Bodo Moeller, Google Security Team]
449 *) Enable TLS extensions by default.
452 *) Allow the CHIL engine to be loaded, whether the application is
453 multithreaded or not. (This does not release the developer from the
454 obligation to set up the dynamic locking callbacks.)
455 [Sander Temme <sander@temme.net>]
457 *) Use correct exit code if there is an error in dgst command.
458 [Steve Henson; problem pointed out by Roland Dirlewanger]
460 *) Tweak Configure so that you need to say "experimental-jpake" to enable
461 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
464 *) Add experimental JPAKE support, including demo authentication in
465 s_client and s_server.
468 *) Set the comparison function in v3_addr_canonize().
469 [Rob Austein <sra@hactrn.net>]
471 *) Add support for XMPP STARTTLS in s_client.
472 [Philip Paeps <philip@freebsd.org>]
474 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
475 to ensure that even with this option, only ciphersuites in the
476 server's preference list will be accepted. (Note that the option
477 applies only when resuming a session, so the earlier behavior was
478 just about the algorithm choice for symmetric cryptography.)
481 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
483 *) Fix NULL pointer dereference if a DTLS server received
484 ChangeCipherSpec as first record (CVE-2009-1386).
487 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
488 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
491 *) The fix in 0.9.8c that supposedly got rid of unsafe
492 double-checked locking was incomplete for RSA blinding,
493 addressing just one layer of what turns out to have been
494 doubly unsafe triple-checked locking.
496 So now fix this for real by retiring the MONT_HELPER macro
497 in crypto/rsa/rsa_eay.c.
499 [Bodo Moeller; problem pointed out by Marius Schilder]
501 *) Various precautionary measures:
503 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
505 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
506 (NB: This would require knowledge of the secret session ticket key
507 to exploit, in which case you'd be SOL either way.)
509 - Change bn_nist.c so that it will properly handle input BIGNUMs
510 outside the expected range.
512 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
515 [Neel Mehta, Bodo Moeller]
517 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
518 the load fails. Useful for distros.
519 [Ben Laurie and the FreeBSD team]
521 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
524 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
527 *) Expand ENGINE to support engine supplied SSL client certificate functions.
529 This work was sponsored by Logica.
532 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
533 keystores. Support for SSL/TLS client authentication too.
534 Not compiled unless enable-capieng specified to Configure.
536 This work was sponsored by Logica.
539 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
540 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
541 attribute creation routines such as certifcate requests and PKCS#12
545 Changes between 0.9.8g and 0.9.8h [28 May 2008]
547 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
548 handshake which could lead to a cilent crash as found using the
549 Codenomicon TLS test suite (CVE-2008-1672)
550 [Steve Henson, Mark Cox]
552 *) Fix double free in TLS server name extensions which could lead to
553 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
556 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
558 Clear the error queue to ensure that error entries left from
559 older function calls do not interfere with the correct operation.
560 [Lutz Jaenicke, Erik de Castro Lopo]
562 *) Remove root CA certificates of commercial CAs:
564 The OpenSSL project does not recommend any specific CA and does not
565 have any policy with respect to including or excluding any CA.
566 Therefore it does not make any sense to ship an arbitrary selection
567 of root CA certificates with the OpenSSL software.
570 *) RSA OAEP patches to fix two separate invalid memory reads.
571 The first one involves inputs when 'lzero' is greater than
572 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
573 before the beginning of from). The second one involves inputs where
574 the 'db' section contains nothing but zeroes (there is a one-byte
575 invalid read after the end of 'db').
576 [Ivan Nestlerode <inestlerode@us.ibm.com>]
578 *) Partial backport from 0.9.9-dev:
580 Introduce bn_mul_mont (dedicated Montgomery multiplication
581 procedure) as a candidate for BIGNUM assembler implementation.
582 While 0.9.9-dev uses assembler for various architectures, only
583 x86_64 is available by default here in the 0.9.8 branch, and
584 32-bit x86 is available through a compile-time setting.
586 To try the 32-bit x86 assembler implementation, use Configure
587 option "enable-montasm" (which exists only for this backport).
589 As "enable-montasm" for 32-bit x86 disclaims code stability
590 anyway, in this constellation we activate additional code
591 backported from 0.9.9-dev for further performance improvements,
592 namely BN_from_montgomery_word. (To enable this otherwise,
593 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
595 [Andy Polyakov (backport partially by Bodo Moeller)]
597 *) Add TLS session ticket callback. This allows an application to set
598 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
599 values. This is useful for key rollover for example where several key
600 sets may exist with different names.
603 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
604 This was broken until now in 0.9.8 releases, such that the only way
605 a registered ENGINE could be used (assuming it initialises
606 successfully on the host) was to explicitly set it as the default
607 for the relevant algorithms. This is in contradiction with 0.9.7
608 behaviour and the documentation. With this fix, when an ENGINE is
609 registered into a given algorithm's table of implementations, the
610 'uptodate' flag is reset so that auto-discovery will be used next
611 time a new context for that algorithm attempts to select an
613 [Ian Lister (tweaked by Geoff Thorpe)]
615 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
616 implemention in the following ways:
618 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
621 Lack of BER streaming support means one pass streaming processing is
622 only supported if data is detached: setting the streaming flag is
623 ignored for embedded content.
625 CMS support is disabled by default and must be explicitly enabled
626 with the enable-cms configuration option.
629 *) Update the GMP engine glue to do direct copies between BIGNUM and
630 mpz_t when openssl and GMP use the same limb size. Otherwise the
631 existing "conversion via a text string export" trick is still used.
632 [Paul Sheer <paulsheer@gmail.com>]
634 *) Zlib compression BIO. This is a filter BIO which compressed and
635 uncompresses any data passed through it.
638 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
639 RFC3394 compatible AES key wrapping.
642 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
643 sets string data without copying. X509_ALGOR_set0() and
644 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
645 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
646 from an X509_ATTRIBUTE structure optionally checking it occurs only
647 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
651 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
652 to get the expected BN_FLG_CONSTTIME behavior.
653 [Bodo Moeller (Google)]
657 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
658 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
659 - added some more tests to do_tests.pl
660 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
661 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
662 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
663 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
664 - various changes to netware.pl to enable gcc-cross builds on Win32
666 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
667 - various changes to fix missing prototype warnings
668 - fixed x86nasm.pl to create correct asm files for NASM COFF output
669 - added AES, WHIRLPOOL and CPUID assembler code to build files
670 - added missing AES assembler make rules to mk1mf.pl
671 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
672 [Guenter Knauf <eflash@gmx.net>]
674 *) Implement certificate status request TLS extension defined in RFC3546.
675 A client can set the appropriate parameters and receive the encoded
676 OCSP response via a callback. A server can query the supplied parameters
677 and set the encoded OCSP response in the callback. Add simplified examples
678 to s_client and s_server.
681 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
684 + Binary incompatibility of ssl_ctx_st structure
685 + DTLS interoperation with non-compliant servers
686 + Don't call get_session_cb() without proposed session
687 + Fix ia64 assembler code
688 [Andy Polyakov, Steve Henson]
690 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
692 *) DTLS Handshake overhaul. There were longstanding issues with
693 OpenSSL DTLS implementation, which were making it impossible for
694 RFC 4347 compliant client to communicate with OpenSSL server.
695 Unfortunately just fixing these incompatibilities would "cut off"
696 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
697 server keeps tolerating non RFC compliant syntax. The opposite is
698 not true, 0.9.8f client can not communicate with earlier server.
699 This update even addresses CVE-2007-4995.
702 *) Changes to avoid need for function casts in OpenSSL: some compilers
703 (gcc 4.2 and later) reject their use.
704 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
707 *) Add RFC4507 support to OpenSSL. This includes the corrections in
708 RFC4507bis. The encrypted ticket format is an encrypted encoded
709 SSL_SESSION structure, that way new session features are automatically
712 If a client application caches session in an SSL_SESSION structure
713 support is transparent because tickets are now stored in the encoded
716 The SSL_CTX structure automatically generates keys for ticket
717 protection in servers so again support should be possible
718 with no application modification.
720 If a client or server wishes to disable RFC4507 support then the option
721 SSL_OP_NO_TICKET can be set.
723 Add a TLS extension debugging callback to allow the contents of any client
724 or server extensions to be examined.
726 This work was sponsored by Google.
729 *) Add initial support for TLS extensions, specifically for the server_name
730 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
731 have new members for a host name. The SSL data structure has an
732 additional member SSL_CTX *initial_ctx so that new sessions can be
733 stored in that context to allow for session resumption, even after the
734 SSL has been switched to a new SSL_CTX in reaction to a client's
735 server_name extension.
737 New functions (subject to change):
740 SSL_get_servername_type()
743 New CTRL codes and macros (subject to change):
745 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
746 - SSL_CTX_set_tlsext_servername_callback()
747 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
748 - SSL_CTX_set_tlsext_servername_arg()
749 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
751 openssl s_client has a new '-servername ...' option.
753 openssl s_server has new options '-servername_host ...', '-cert2 ...',
754 '-key2 ...', '-servername_fatal' (subject to change). This allows
755 testing the HostName extension for a specific single host name ('-cert'
756 and '-key' remain fallbacks for handshakes without HostName
757 negotiation). If the unrecogninzed_name alert has to be sent, this by
758 default is a warning; it becomes fatal with the '-servername_fatal'
761 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
763 *) Add AES and SSE2 assembly language support to VC++ build.
766 *) Mitigate attack on final subtraction in Montgomery reduction.
769 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
770 (which previously caused an internal error).
773 *) Squeeze another 10% out of IGE mode when in != out.
776 *) AES IGE mode speedup.
777 [Dean Gaudet (Google)]
779 *) Add the Korean symmetric 128-bit cipher SEED (see
780 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
781 add SEED ciphersuites from RFC 4162:
783 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
784 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
785 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
786 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
788 To minimize changes between patchlevels in the OpenSSL 0.9.8
789 series, SEED remains excluded from compilation unless OpenSSL
790 is configured with 'enable-seed'.
793 *) Mitigate branch prediction attacks, which can be practical if a
794 single processor is shared, allowing a spy process to extract
795 information. For detailed background information, see
796 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
797 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
798 and Necessary Software Countermeasures"). The core of the change
799 are new versions BN_div_no_branch() and
800 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
801 respectively, which are slower, but avoid the security-relevant
802 conditional branches. These are automatically called by BN_div()
803 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
804 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
805 remove a conditional branch.
807 BN_FLG_CONSTTIME is the new name for the previous
808 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
809 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
810 in the exponent causes BN_mod_exp_mont() to use the alternative
811 implementation in BN_mod_exp_mont_consttime().) The old name
812 remains as a deprecated alias.
814 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
815 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
816 constant-time implementations for more than just exponentiation.
817 Here too the old name is kept as a deprecated alias.
819 BN_BLINDING_new() will now use BN_dup() for the modulus so that
820 the BN_BLINDING structure gets an independent copy of the
821 modulus. This means that the previous "BIGNUM *m" argument to
822 BN_BLINDING_new() and to BN_BLINDING_create_param() now
823 essentially becomes "const BIGNUM *m", although we can't actually
824 change this in the header file before 0.9.9. It allows
825 RSA_setup_blinding() to use BN_with_flags() on the modulus to
826 enable BN_FLG_CONSTTIME.
828 [Matthew D Wood (Intel Corp)]
830 *) In the SSL/TLS server implementation, be strict about session ID
831 context matching (which matters if an application uses a single
832 external cache for different purposes). Previously,
833 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
834 set. This did ensure strict client verification, but meant that,
835 with applications using a single external cache for quite
836 different requirements, clients could circumvent ciphersuite
837 restrictions for a given session ID context by starting a session
838 in a different context.
841 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
842 a ciphersuite string such as "DEFAULT:RSA" cannot enable
843 authentication-only ciphersuites.
846 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
847 not complete and could lead to a possible single byte overflow
848 (CVE-2007-5135) [Ben Laurie]
850 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
852 *) Since AES128 and AES256 (and similarly Camellia128 and
853 Camellia256) share a single mask bit in the logic of
854 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
855 kludge to work properly if AES128 is available and AES256 isn't
856 (or if Camellia128 is available and Camellia256 isn't).
859 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
860 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
861 When a point or a seed is encoded in a BIT STRING, we need to
862 prevent the removal of trailing zero bits to get the proper DER
863 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
864 of a NamedBitList, for which trailing 0 bits need to be removed.)
867 *) Have SSL/TLS server implementation tolerate "mismatched" record
868 protocol version while receiving ClientHello even if the
869 ClientHello is fragmented. (The server can't insist on the
870 particular protocol version it has chosen before the ServerHello
871 message has informed the client about his choice.)
874 *) Add RFC 3779 support.
875 [Rob Austein for ARIN, Ben Laurie]
877 *) Load error codes if they are not already present instead of using a
878 static variable. This allows them to be cleanly unloaded and reloaded.
879 Improve header file function name parsing.
882 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
883 or CAPABILITY handshake as required by RFCs.
886 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
888 *) Introduce limits to prevent malicious keys being able to
889 cause a denial of service. (CVE-2006-2940)
890 [Steve Henson, Bodo Moeller]
892 *) Fix ASN.1 parsing of certain invalid structures that can result
893 in a denial of service. (CVE-2006-2937) [Steve Henson]
895 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
896 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
898 *) Fix SSL client code which could crash if connecting to a
899 malicious SSLv2 server. (CVE-2006-4343)
900 [Tavis Ormandy and Will Drewry, Google Security Team]
902 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
903 match only those. Before that, "AES256-SHA" would be interpreted
904 as a pattern and match "AES128-SHA" too (since AES128-SHA got
905 the same strength classification in 0.9.7h) as we currently only
906 have a single AES bit in the ciphersuite description bitmap.
907 That change, however, also applied to ciphersuite strings such as
908 "RC4-MD5" that intentionally matched multiple ciphersuites --
909 namely, SSL 2.0 ciphersuites in addition to the more common ones
910 from SSL 3.0/TLS 1.0.
912 So we change the selection algorithm again: Naming an explicit
913 ciphersuite selects this one ciphersuite, and any other similar
914 ciphersuite (same bitmap) from *other* protocol versions.
915 Thus, "RC4-MD5" again will properly select both the SSL 2.0
916 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
918 Since SSL 2.0 does not have any ciphersuites for which the
919 128/256 bit distinction would be relevant, this works for now.
920 The proper fix will be to use different bits for AES128 and
921 AES256, which would have avoided the problems from the beginning;
922 however, bits are scarce, so we can only do this in a new release
923 (not just a patchlevel) when we can change the SSL_CIPHER
924 definition to split the single 'unsigned long mask' bitmap into
925 multiple values to extend the available space.
929 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
931 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
932 (CVE-2006-4339) [Ben Laurie and Google Security Team]
934 *) Add AES IGE and biIGE modes.
937 *) Change the Unix randomness entropy gathering to use poll() when
938 possible instead of select(), since the latter has some
939 undesirable limitations.
940 [Darryl Miles via Richard Levitte and Bodo Moeller]
942 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
943 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
944 cannot be implicitly activated as part of, e.g., the "AES" alias.
945 However, please upgrade to OpenSSL 0.9.9[-dev] for
946 non-experimental use of the ECC ciphersuites to get TLS extension
947 support, which is required for curve and point format negotiation
948 to avoid potential handshake problems.
951 *) Disable rogue ciphersuites:
953 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
954 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
955 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
957 The latter two were purportedly from
958 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
961 Also deactivate the remaining ciphersuites from
962 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
963 unofficial, and the ID has long expired.
966 *) Fix RSA blinding Heisenbug (problems sometimes occured on
967 dual-core machines) and other potential thread-safety issues.
970 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
971 versions), which is now available for royalty-free use
972 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
973 Also, add Camellia TLS ciphersuites from RFC 4132.
975 To minimize changes between patchlevels in the OpenSSL 0.9.8
976 series, Camellia remains excluded from compilation unless OpenSSL
977 is configured with 'enable-camellia'.
980 *) Disable the padding bug check when compression is in use. The padding
981 bug check assumes the first packet is of even length, this is not
982 necessarily true if compresssion is enabled and can result in false
983 positives causing handshake failure. The actual bug test is ancient
984 code so it is hoped that implementations will either have fixed it by
985 now or any which still have the bug do not support compression.
988 Changes between 0.9.8a and 0.9.8b [04 May 2006]
990 *) When applying a cipher rule check to see if string match is an explicit
991 cipher suite and only match that one cipher suite if it is.
994 *) Link in manifests for VC++ if needed.
995 [Austin Ziegler <halostatue@gmail.com>]
997 *) Update support for ECC-based TLS ciphersuites according to
998 draft-ietf-tls-ecc-12.txt with proposed changes (but without
999 TLS extensions, which are supported starting with the 0.9.9
1000 branch, not in the OpenSSL 0.9.8 branch).
1003 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1004 opaque EVP_CIPHER_CTX handling.
1007 *) Fixes and enhancements to zlib compression code. We now only use
1008 "zlib1.dll" and use the default __cdecl calling convention on Win32
1009 to conform with the standards mentioned here:
1010 http://www.zlib.net/DLL_FAQ.txt
1011 Static zlib linking now works on Windows and the new --with-zlib-include
1012 --with-zlib-lib options to Configure can be used to supply the location
1013 of the headers and library. Gracefully handle case where zlib library
1017 *) Several fixes and enhancements to the OID generation code. The old code
1018 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1019 handle numbers larger than ULONG_MAX, truncated printing and had a
1020 non standard OBJ_obj2txt() behaviour.
1023 *) Add support for building of engines under engine/ as shared libraries
1024 under VC++ build system.
1027 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1028 Hopefully, we will not see any false combination of paths any more.
1031 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1033 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1034 (part of SSL_OP_ALL). This option used to disable the
1035 countermeasure against man-in-the-middle protocol-version
1036 rollback in the SSL 2.0 server implementation, which is a bad
1037 idea. (CVE-2005-2969)
1039 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1040 for Information Security, National Institute of Advanced Industrial
1041 Science and Technology [AIST], Japan)]
1043 *) Add two function to clear and return the verify parameter flags.
1046 *) Keep cipherlists sorted in the source instead of sorting them at
1047 runtime, thus removing the need for a lock.
1050 *) Avoid some small subgroup attacks in Diffie-Hellman.
1051 [Nick Mathewson and Ben Laurie]
1053 *) Add functions for well-known primes.
1056 *) Extended Windows CE support.
1057 [Satoshi Nakamura and Andy Polyakov]
1059 *) Initialize SSL_METHOD structures at compile time instead of during
1060 runtime, thus removing the need for a lock.
1063 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1064 attempting to decrypt each encrypted key in turn. Add support to
1068 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1070 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1073 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1076 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1077 key into the same file any more.
1080 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1083 *) Add -utf8 command line and config file option to 'ca'.
1084 [Stefan <stf@udoma.org]
1086 *) Removed the macro des_crypt(), as it seems to conflict with some
1087 libraries. Use DES_crypt().
1090 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1091 involves renaming the source and generated shared-libs for
1092 both. The engines will accept the corrected or legacy ids
1093 ('ncipher' and '4758_cca' respectively) when binding. NB,
1094 this only applies when building 'shared'.
1095 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1097 *) Add attribute functions to EVP_PKEY structure. Modify
1098 PKCS12_create() to recognize a CSP name attribute and
1099 use it. Make -CSP option work again in pkcs12 utility.
1102 *) Add new functionality to the bn blinding code:
1103 - automatic re-creation of the BN_BLINDING parameters after
1104 a fixed number of uses (currently 32)
1105 - add new function for parameter creation
1106 - introduce flags to control the update behaviour of the
1107 BN_BLINDING parameters
1108 - hide BN_BLINDING structure
1109 Add a second BN_BLINDING slot to the RSA structure to improve
1110 performance when a single RSA object is shared among several
1114 *) Add support for DTLS.
1115 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1117 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1118 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1121 *) Remove buggy and incompletet DH cert support from
1122 ssl/ssl_rsa.c and ssl/s3_both.c
1125 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1126 the apps/openssl applications.
1129 *) Compile clean with "-Wall -Wmissing-prototypes
1130 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1131 DEBUG_SAFESTACK must also be set.
1134 *) Change ./Configure so that certain algorithms can be disabled by default.
1135 The new counterpiece to "no-xxx" is "enable-xxx".
1137 The patented RC5 and MDC2 algorithms will now be disabled unless
1138 "enable-rc5" and "enable-mdc2", respectively, are specified.
1140 (IDEA remains enabled despite being patented. This is because IDEA
1141 is frequently required for interoperability, and there is no license
1142 fee for non-commercial use. As before, "no-idea" can be used to
1143 avoid this algorithm.)
1147 *) Add processing of proxy certificates (see RFC 3820). This work was
1148 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1149 EGEE (Enabling Grids for E-science in Europe).
1152 *) RC4 performance overhaul on modern architectures/implementations, such
1153 as Intel P4, IA-64 and AMD64.
1156 *) New utility extract-section.pl. This can be used specify an alternative
1157 section number in a pod file instead of having to treat each file as
1158 a separate case in Makefile. This can be done by adding two lines to the
1161 =for comment openssl_section:XXX
1163 The blank line is mandatory.
1167 *) New arguments -certform, -keyform and -pass for s_client and s_server
1168 to allow alternative format key and certificate files and passphrase
1172 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1173 update associated structures and add various utility functions.
1175 Add new policy related verify parameters, include policy checking in
1176 standard verify code. Enhance 'smime' application with extra parameters
1177 to support policy checking and print out.
1180 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1181 Nehemiah processors. These extensions support AES encryption in hardware
1182 as well as RNG (though RNG support is currently disabled).
1183 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1185 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1188 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1189 [Andy Polyakov and a number of other people]
1191 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1192 implementation contributed by IBM.
1193 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1195 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1196 exponent rather than 'unsigned long'. There is a corresponding change to
1197 the new 'rsa_keygen' element of the RSA_METHOD structure.
1198 [Jelte Jansen, Geoff Thorpe]
1200 *) Functionality for creating the initial serial number file is now
1201 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1203 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1204 number file to 1, which is bound to cause problems. To avoid
1205 the problems while respecting compatibility between different 0.9.7
1206 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1207 CA.pl for serial number initialization. With the new release 0.9.8,
1208 we can fix the problem directly in the 'ca' utility.)
1211 *) Reduced header interdepencies by declaring more opaque objects in
1212 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1213 give fewer recursive includes, which could break lazy source code - so
1214 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1215 developers should define this symbol when building and using openssl to
1216 ensure they track the recommended behaviour, interfaces, [etc], but
1217 backwards-compatible behaviour prevails when this isn't defined.
1220 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1223 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1224 This will generate a random key of the appropriate length based on the
1225 cipher context. The EVP_CIPHER can provide its own random key generation
1226 routine to support keys of a specific form. This is used in the des and
1227 3des routines to generate a key of the correct parity. Update S/MIME
1228 code to use new functions and hence generate correct parity DES keys.
1229 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1230 valid (weak or incorrect parity).
1233 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1234 as looking them up. This is useful when the verified structure may contain
1235 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1236 present unless the new PKCS7_NO_CRL flag is asserted.
1239 *) Extend ASN1 oid configuration module. It now additionally accepts the
1242 shortName = some long name, 1.2.3.4
1245 *) Reimplemented the BN_CTX implementation. There is now no more static
1246 limitation on the number of variables it can handle nor the depth of the
1247 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1248 information can now expand as required, and rather than having a single
1249 static array of bignums, BN_CTX now uses a linked-list of such arrays
1250 allowing it to expand on demand whilst maintaining the usefulness of
1251 BN_CTX's "bundling".
1254 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1255 to allow all RSA operations to function using a single BN_CTX.
1258 *) Preliminary support for certificate policy evaluation and checking. This
1259 is initially intended to pass the tests outlined in "Conformance Testing
1260 of Relying Party Client Certificate Path Processing Logic" v1.07.
1263 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1264 remained unused and not that useful. A variety of other little bignum
1265 tweaks and fixes have also been made continuing on from the audit (see
1269 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1270 associated ASN1, EVP and SSL functions and old ASN1 macros.
1273 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1274 and this should never fail. So the return value from the use of
1275 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1276 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1279 *) BN_CTX_get() should return zero-valued bignums, providing the same
1280 initialised value as BN_new().
1281 [Geoff Thorpe, suggested by Ulf Möller]
1283 *) Support for inhibitAnyPolicy certificate extension.
1286 *) An audit of the BIGNUM code is underway, for which debugging code is
1287 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1288 is considered valid when processing BIGNUMs, and causes execution to
1289 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1290 further steps are taken to deliberately pollute unused data in BIGNUM
1291 structures to try and expose faulty code further on. For now, openssl will
1292 (in its default mode of operation) continue to tolerate the inconsistent
1293 forms that it has tolerated in the past, but authors and packagers should
1294 consider trying openssl and their own applications when compiled with
1295 these debugging symbols defined. It will help highlight potential bugs in
1296 their own code, and will improve the test coverage for OpenSSL itself. At
1297 some point, these tighter rules will become openssl's default to improve
1298 maintainability, though the assert()s and other overheads will remain only
1299 in debugging configurations. See bn.h for more details.
1300 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1302 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1303 that can only be obtained through BN_CTX_new() (which implicitly
1304 initialises it). The presence of this function only made it possible
1305 to overwrite an existing structure (and cause memory leaks).
1308 *) Because of the callback-based approach for implementing LHASH as a
1309 template type, lh_insert() adds opaque objects to hash-tables and
1310 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1311 to clean up those corresponding objects before destroying the hash table
1312 (and losing the object pointers). So some over-zealous constifications in
1313 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1314 objects as "const" and the lh_doall[_arg] callback wrappers are not
1315 prototyped to have "const" restrictions on the object pointers they are
1316 given (and so aren't required to cast them away any more).
1319 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1320 (speed) prefers to use its own implementation. The two implementations
1321 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1322 its object type properly exposed (MS_TM) instead of casting to/from "char
1323 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1324 aren't necessarily the greatest nomenclatures - but this is what was used
1325 internally to the implementation so I've used that for now.
1328 *) Ensure that deprecated functions do not get compiled when
1329 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1330 the self-tests were still using deprecated key-generation functions so
1331 these have been updated also.
1334 *) Reorganise PKCS#7 code to separate the digest location functionality
1335 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1336 New function PKCS7_set_digest() to set the digest type for PKCS#7
1337 digestedData type. Add additional code to correctly generate the
1338 digestedData type and add support for this type in PKCS7 initialization
1342 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1343 structure of type "other".
1346 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1347 sure the loop does correctly stop and breaking ("division by zero")
1348 modulus operations are not performed. The (pre-generated) prime
1349 table crypto/bn/bn_prime.h was already correct, but it could not be
1350 re-generated on some platforms because of the "division by zero"
1351 situation in the script.
1352 [Ralf S. Engelschall]
1354 *) Update support for ECC-based TLS ciphersuites according to
1355 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1356 SHA-1 now is only used for "small" curves (where the
1357 representation of a field element takes up to 24 bytes); for
1358 larger curves, the field element resulting from ECDH is directly
1359 used as premaster secret.
1360 [Douglas Stebila (Sun Microsystems Laboratories)]
1362 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1363 curve secp160r1 to the tests.
1364 [Douglas Stebila (Sun Microsystems Laboratories)]
1366 *) Add the possibility to load symbols globally with DSO.
1367 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1369 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1370 control of the error stack.
1373 *) Add support for STORE in ENGINE.
1376 *) Add the STORE type. The intention is to provide a common interface
1377 to certificate and key stores, be they simple file-based stores, or
1378 HSM-type store, or LDAP stores, or...
1379 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1382 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1383 pass a list of arguments to any function as well as provide a way
1384 for a function to pass data back to the caller.
1387 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1388 works like BUF_strdup() but can be used to duplicate a portion of
1389 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1393 *) Add the function sk_find_ex() which works like sk_find(), but will
1394 return an index to an element even if an exact match couldn't be
1395 found. The index is guaranteed to point at the element where the
1396 searched-for key would be inserted to preserve sorting order.
1399 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1400 takes an extra flags argument for optional functionality. Currently,
1401 the following flags are defined:
1403 OBJ_BSEARCH_VALUE_ON_NOMATCH
1404 This one gets OBJ_bsearch_ex() to return a pointer to the first
1405 element where the comparing function returns a negative or zero
1408 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1409 This one gets OBJ_bsearch_ex() to return a pointer to the first
1410 element where the comparing function returns zero. This is useful
1411 if there are more than one element where the comparing function
1415 *) Make it possible to create self-signed certificates with 'openssl ca'
1416 in such a way that the self-signed certificate becomes part of the
1417 CA database and uses the same mechanisms for serial number generation
1418 as all other certificate signing. The new flag '-selfsign' enables
1419 this functionality. Adapt CA.sh and CA.pl.in.
1422 *) Add functionality to check the public key of a certificate request
1423 against a given private. This is useful to check that a certificate
1424 request can be signed by that key (self-signing).
1427 *) Make it possible to have multiple active certificates with the same
1428 subject in the CA index file. This is done only if the keyword
1429 'unique_subject' is set to 'no' in the main CA section (default
1430 if 'CA_default') of the configuration file. The value is saved
1431 with the database itself in a separate index attribute file,
1432 named like the index file with '.attr' appended to the name.
1435 *) Generate muti valued AVAs using '+' notation in config files for
1439 *) Support for nameConstraints certificate extension.
1442 *) Support for policyConstraints certificate extension.
1445 *) Support for policyMappings certificate extension.
1448 *) Make sure the default DSA_METHOD implementation only uses its
1449 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1450 and change its own handlers to be NULL so as to remove unnecessary
1451 indirection. This lets alternative implementations fallback to the
1452 default implementation more easily.
1455 *) Support for directoryName in GeneralName related extensions
1459 *) Make it possible to link applications using Makefile.shared.
1460 Make that possible even when linking against static libraries!
1463 *) Support for single pass processing for S/MIME signing. This now
1464 means that S/MIME signing can be done from a pipe, in addition
1465 cleartext signing (multipart/signed type) is effectively streaming
1466 and the signed data does not need to be all held in memory.
1468 This is done with a new flag PKCS7_STREAM. When this flag is set
1469 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1470 is done after the data is output (and digests calculated) in
1471 SMIME_write_PKCS7().
1474 *) Add full support for -rpath/-R, both in shared libraries and
1475 applications, at least on the platforms where it's known how
1479 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1480 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1481 will now compute a table of multiples of the generator that
1482 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1483 faster (notably in the case of a single point multiplication,
1484 scalar * generator).
1485 [Nils Larsch, Bodo Moeller]
1487 *) IPv6 support for certificate extensions. The various extensions
1488 which use the IP:a.b.c.d can now take IPv6 addresses using the
1489 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1493 *) Added an ENGINE that implements RSA by performing private key
1494 exponentiations with the GMP library. The conversions to and from
1495 GMP's mpz_t format aren't optimised nor are any montgomery forms
1496 cached, and on x86 it appears OpenSSL's own performance has caught up.
1497 However there are likely to be other architectures where GMP could
1498 provide a boost. This ENGINE is not built in by default, but it can be
1499 specified at Configure time and should be accompanied by the necessary
1500 linker additions, eg;
1501 ./config -DOPENSSL_USE_GMP -lgmp
1504 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1505 testing availability of engines with "-t" - the old behaviour is
1506 produced by increasing the feature's verbosity with "-tt".
1509 *) ECDSA routines: under certain error conditions uninitialized BN objects
1510 could be freed. Solution: make sure initialization is performed early
1511 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1515 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1516 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1517 software implementations. For DSA and DH, parameter generation can
1518 also be overriden by providing the appropriate method callbacks.
1521 *) Change the "progress" mechanism used in key-generation and
1522 primality testing to functions that take a new BN_GENCB pointer in
1523 place of callback/argument pairs. The new API functions have "_ex"
1524 postfixes and the older functions are reimplemented as wrappers for
1525 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1526 declarations of the old functions to help (graceful) attempts to
1527 migrate to the new functions. Also, the new key-generation API
1528 functions operate on a caller-supplied key-structure and return
1529 success/failure rather than returning a key or NULL - this is to
1530 help make "keygen" another member function of RSA_METHOD etc.
1532 Example for using the new callback interface:
1534 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1538 BN_GENCB_set(&my_cb, my_callback, my_arg);
1540 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1541 /* For the meaning of a, b in calls to my_callback(), see the
1542 * documentation of the function that calls the callback.
1543 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1544 * my_callback should return 1 if it wants BN_is_prime_ex()
1545 * to continue, or 0 to stop.
1550 *) Change the ZLIB compression method to be stateful, and make it
1551 available to TLS with the number defined in
1552 draft-ietf-tls-compression-04.txt.
1555 *) Add the ASN.1 structures and functions for CertificatePair, which
1556 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1558 CertificatePair ::= SEQUENCE {
1559 forward [0] Certificate OPTIONAL,
1560 reverse [1] Certificate OPTIONAL,
1561 -- at least one of the pair shall be present -- }
1563 Also implement the PEM functions to read and write certificate
1564 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1566 This needed to be defined, mostly for the sake of the LDAP
1567 attribute crossCertificatePair, but may prove useful elsewhere as
1571 *) Make it possible to inhibit symlinking of shared libraries in
1572 Makefile.shared, for Cygwin's sake.
1575 *) Extend the BIGNUM API by creating a function
1576 void BN_set_negative(BIGNUM *a, int neg);
1577 and a macro that behave like
1578 int BN_is_negative(const BIGNUM *a);
1580 to avoid the need to access 'a->neg' directly in applications.
1583 *) Implement fast modular reduction for pseudo-Mersenne primes
1584 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1585 EC_GROUP_new_curve_GFp() will now automatically use this
1587 [Nils Larsch <nla@trustcenter.de>]
1589 *) Add new lock type (CRYPTO_LOCK_BN).
1592 *) Change the ENGINE framework to automatically load engines
1593 dynamically from specific directories unless they could be
1594 found to already be built in or loaded. Move all the
1595 current engines except for the cryptodev one to a new
1597 The engines in engines/ are built as shared libraries if
1598 the "shared" options was given to ./Configure or ./config.
1599 Otherwise, they are inserted in libcrypto.a.
1600 /usr/local/ssl/engines is the default directory for dynamic
1601 engines, but that can be overriden at configure time through
1602 the usual use of --prefix and/or --openssldir, and at run
1603 time with the environment variable OPENSSL_ENGINES.
1604 [Geoff Thorpe and Richard Levitte]
1606 *) Add Makefile.shared, a helper makefile to build shared
1607 libraries. Addapt Makefile.org.
1610 *) Add version info to Win32 DLLs.
1611 [Peter 'Luna' Runestig" <peter@runestig.com>]
1613 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1614 can be added using this API to created arbitrary PKCS#12
1615 files while avoiding the low level API.
1617 New options to PKCS12_create(), key or cert can be NULL and
1618 will then be omitted from the output file. The encryption
1619 algorithm NIDs can be set to -1 for no encryption, the mac
1620 iteration count can be set to 0 to omit the mac.
1622 Enhance pkcs12 utility by making the -nokeys and -nocerts
1623 options work when creating a PKCS#12 file. New option -nomac
1624 to omit the mac, NONE can be set for an encryption algorithm.
1625 New code is modified to use the enhanced PKCS12_create()
1626 instead of the low level API.
1629 *) Extend ASN1 encoder to support indefinite length constructed
1630 encoding. This can output sequences tags and octet strings in
1631 this form. Modify pk7_asn1.c to support indefinite length
1632 encoding. This is experimental and needs additional code to
1633 be useful, such as an ASN1 bio and some enhanced streaming
1636 Extend template encode functionality so that tagging is passed
1637 down to the template encoder.
1640 *) Let 'openssl req' fail if an argument to '-newkey' is not
1641 recognized instead of using RSA as a default.
1644 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1645 As these are not official, they are not included in "ALL";
1646 the "ECCdraft" ciphersuite group alias can be used to select them.
1647 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1649 *) Add ECDH engine support.
1650 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1652 *) Add ECDH in new directory crypto/ecdh/.
1653 [Douglas Stebila (Sun Microsystems Laboratories)]
1655 *) Let BN_rand_range() abort with an error after 100 iterations
1656 without success (which indicates a broken PRNG).
1659 *) Change BN_mod_sqrt() so that it verifies that the input value
1660 is really the square of the return value. (Previously,
1661 BN_mod_sqrt would show GIGO behaviour.)
1664 *) Add named elliptic curves over binary fields from X9.62, SECG,
1665 and WAP/WTLS; add OIDs that were still missing.
1667 [Sheueling Chang Shantz and Douglas Stebila
1668 (Sun Microsystems Laboratories)]
1670 *) Extend the EC library for elliptic curves over binary fields
1671 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1674 EC_GF2m_simple_method
1678 EC_GROUP_new_curve_GF2m
1679 EC_GROUP_set_curve_GF2m
1680 EC_GROUP_get_curve_GF2m
1681 EC_POINT_set_affine_coordinates_GF2m
1682 EC_POINT_get_affine_coordinates_GF2m
1683 EC_POINT_set_compressed_coordinates_GF2m
1685 Point compression for binary fields is disabled by default for
1686 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1689 As binary polynomials are represented as BIGNUMs, various members
1690 of the EC_GROUP and EC_POINT data structures can be shared
1691 between the implementations for prime fields and binary fields;
1692 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1693 are essentially identical to their ..._GFp counterparts.
1694 (For simplicity, the '..._GFp' prefix has been dropped from
1695 various internal method names.)
1697 An internal 'field_div' method (similar to 'field_mul' and
1698 'field_sqr') has been added; this is used only for binary fields.
1700 [Sheueling Chang Shantz and Douglas Stebila
1701 (Sun Microsystems Laboratories)]
1703 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1704 through methods ('mul', 'precompute_mult').
1706 The generic implementations (now internally called 'ec_wNAF_mul'
1707 and 'ec_wNAF_precomputed_mult') remain the default if these
1708 methods are undefined.
1710 [Sheueling Chang Shantz and Douglas Stebila
1711 (Sun Microsystems Laboratories)]
1713 *) New function EC_GROUP_get_degree, which is defined through
1714 EC_METHOD. For curves over prime fields, this returns the bit
1715 length of the modulus.
1717 [Sheueling Chang Shantz and Douglas Stebila
1718 (Sun Microsystems Laboratories)]
1720 *) New functions EC_GROUP_dup, EC_POINT_dup.
1721 (These simply call ..._new and ..._copy).
1723 [Sheueling Chang Shantz and Douglas Stebila
1724 (Sun Microsystems Laboratories)]
1726 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1727 Polynomials are represented as BIGNUMs (where the sign bit is not
1728 used) in the following functions [macros]:
1731 BN_GF2m_sub [= BN_GF2m_add]
1732 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1733 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1734 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1736 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1737 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1738 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1739 BN_GF2m_cmp [= BN_ucmp]
1741 (Note that only the 'mod' functions are actually for fields GF(2^m).
1742 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1744 For some functions, an the irreducible polynomial defining a
1745 field can be given as an 'unsigned int[]' with strictly
1746 decreasing elements giving the indices of those bits that are set;
1747 i.e., p[] represents the polynomial
1748 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1750 p[0] > p[1] > ... > p[k] = 0.
1751 This applies to the following functions:
1756 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1757 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1759 BN_GF2m_mod_sqrt_arr
1760 BN_GF2m_mod_solve_quad_arr
1764 Conversion can be performed by the following functions:
1769 bntest.c has additional tests for binary polynomial arithmetic.
1771 Two implementations for BN_GF2m_mod_div() are available.
1772 The default algorithm simply uses BN_GF2m_mod_inv() and
1773 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1774 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1775 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1777 [Sheueling Chang Shantz and Douglas Stebila
1778 (Sun Microsystems Laboratories)]
1780 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1781 functionality is disabled at compile-time.
1782 [Douglas Stebila <douglas.stebila@sun.com>]
1784 *) Change default behaviour of 'openssl asn1parse' so that more
1785 information is visible when viewing, e.g., a certificate:
1787 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1788 mode the content of non-printable OCTET STRINGs is output in a
1789 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1790 avoid the appearance of a printable string.
1791 [Nils Larsch <nla@trustcenter.de>]
1793 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1795 EC_GROUP_set_asn1_flag()
1796 EC_GROUP_get_asn1_flag()
1797 EC_GROUP_set_point_conversion_form()
1798 EC_GROUP_get_point_conversion_form()
1799 These control ASN1 encoding details:
1800 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1801 has been set to OPENSSL_EC_NAMED_CURVE.
1802 - Points are encoded in uncompressed form by default; options for
1803 asn1_for are as for point2oct, namely
1804 POINT_CONVERSION_COMPRESSED
1805 POINT_CONVERSION_UNCOMPRESSED
1806 POINT_CONVERSION_HYBRID
1808 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1811 EC_GROUP_get0_seed()
1812 EC_GROUP_get_seed_len()
1813 This is used only for ASN1 purposes (so far).
1814 [Nils Larsch <nla@trustcenter.de>]
1816 *) Add 'field_type' member to EC_METHOD, which holds the NID
1817 of the appropriate field type OID. The new function
1818 EC_METHOD_get_field_type() returns this value.
1819 [Nils Larsch <nla@trustcenter.de>]
1824 EC_POINT_point2hex()
1825 EC_POINT_hex2point()
1826 providing useful interfaces to EC_POINT_point2oct() and
1827 EC_POINT_oct2point().
1828 [Nils Larsch <nla@trustcenter.de>]
1830 *) Change internals of the EC library so that the functions
1831 EC_GROUP_set_generator()
1832 EC_GROUP_get_generator()
1833 EC_GROUP_get_order()
1834 EC_GROUP_get_cofactor()
1835 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1836 to methods, which would lead to unnecessary code duplication when
1837 adding different types of curves.
1838 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1840 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1841 arithmetic, and such that modified wNAFs are generated
1842 (which avoid length expansion in many cases).
1845 *) Add a function EC_GROUP_check_discriminant() (defined via
1846 EC_METHOD) that verifies that the curve discriminant is non-zero.
1848 Add a function EC_GROUP_check() that makes some sanity tests
1849 on a EC_GROUP, its generator and order. This includes
1850 EC_GROUP_check_discriminant().
1851 [Nils Larsch <nla@trustcenter.de>]
1853 *) Add ECDSA in new directory crypto/ecdsa/.
1855 Add applications 'openssl ecparam' and 'openssl ecdsa'
1856 (these are based on 'openssl dsaparam' and 'openssl dsa').
1858 ECDSA support is also included in various other files across the
1859 library. Most notably,
1860 - 'openssl req' now has a '-newkey ecdsa:file' option;
1861 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1862 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1863 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1864 them suitable for ECDSA where domain parameters must be
1865 extracted before the specific public key;
1866 - ECDSA engine support has been added.
1867 [Nils Larsch <nla@trustcenter.de>]
1869 *) Include some named elliptic curves, and add OIDs from X9.62,
1870 SECG, and WAP/WTLS. Each curve can be obtained from the new
1872 EC_GROUP_new_by_curve_name(),
1873 and the list of available named curves can be obtained with
1874 EC_get_builtin_curves().
1875 Also add a 'curve_name' member to EC_GROUP objects, which can be
1877 EC_GROUP_set_curve_name()
1878 EC_GROUP_get_curve_name()
1879 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1881 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1882 was actually never needed) and in BN_mul(). The removal in BN_mul()
1883 required a small change in bn_mul_part_recursive() and the addition
1884 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1885 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1886 bn_sub_words() and bn_add_words() except they take arrays with
1890 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1892 *) Cleanse PEM buffers before freeing them since they may contain
1894 [Benjamin Bennett <ben@psc.edu>]
1896 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1897 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1898 authentication-only ciphersuites.
1901 *) Since AES128 and AES256 share a single mask bit in the logic of
1902 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1903 kludge to work properly if AES128 is available and AES256 isn't.
1906 *) Expand security boundary to match 1.1.1 module.
1909 *) Remove redundant features: hash file source, editing of test vectors
1910 modify fipsld to use external fips_premain.c signature.
1913 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1914 run algorithm test programs.
1917 *) Make algorithm test programs more tolerant of whitespace.
1920 *) Have SSL/TLS server implementation tolerate "mismatched" record
1921 protocol version while receiving ClientHello even if the
1922 ClientHello is fragmented. (The server can't insist on the
1923 particular protocol version it has chosen before the ServerHello
1924 message has informed the client about his choice.)
1927 *) Load error codes if they are not already present instead of using a
1928 static variable. This allows them to be cleanly unloaded and reloaded.
1931 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1933 *) Introduce limits to prevent malicious keys being able to
1934 cause a denial of service. (CVE-2006-2940)
1935 [Steve Henson, Bodo Moeller]
1937 *) Fix ASN.1 parsing of certain invalid structures that can result
1938 in a denial of service. (CVE-2006-2937) [Steve Henson]
1940 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1941 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1943 *) Fix SSL client code which could crash if connecting to a
1944 malicious SSLv2 server. (CVE-2006-4343)
1945 [Tavis Ormandy and Will Drewry, Google Security Team]
1947 *) Change ciphersuite string processing so that an explicit
1948 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1949 will no longer include "AES128-SHA"), and any other similar
1950 ciphersuite (same bitmap) from *other* protocol versions (so that
1951 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1952 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1953 changes from 0.9.8b and 0.9.8d.
1956 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1958 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1959 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1961 *) Change the Unix randomness entropy gathering to use poll() when
1962 possible instead of select(), since the latter has some
1963 undesirable limitations.
1964 [Darryl Miles via Richard Levitte and Bodo Moeller]
1966 *) Disable rogue ciphersuites:
1968 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1969 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1970 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1972 The latter two were purportedly from
1973 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1976 Also deactive the remaining ciphersuites from
1977 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1978 unofficial, and the ID has long expired.
1981 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1982 dual-core machines) and other potential thread-safety issues.
1985 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1987 *) Adapt fipsld and the build system to link against the validated FIPS
1988 module in FIPS mode.
1991 *) Fixes for VC++ 2005 build under Windows.
1994 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1995 from a Windows bash shell such as MSYS. It is autodetected from the
1996 "config" script when run from a VC++ environment. Modify standard VC++
1997 build to use fipscanister.o from the GNU make build.
2000 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2002 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2003 The value now differs depending on if you build for FIPS or not.
2004 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2005 safely run with a non-FIPSed libcrypto, as it may crash because of
2006 the difference induced by this change.
2009 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2011 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2012 (part of SSL_OP_ALL). This option used to disable the
2013 countermeasure against man-in-the-middle protocol-version
2014 rollback in the SSL 2.0 server implementation, which is a bad
2015 idea. (CVE-2005-2969)
2017 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2018 for Information Security, National Institute of Advanced Industrial
2019 Science and Technology [AIST], Japan)]
2021 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2022 mainly for FIPS compliance and not fully integrated at this stage.
2025 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2026 the exponentiation using a fixed-length exponent. (Otherwise,
2027 the information leaked through timing could expose the secret key
2028 after many signatures; cf. Bleichenbacher's attack on DSA with
2032 *) Make a new fixed-window mod_exp implementation the default for
2033 RSA, DSA, and DH private-key operations so that the sequence of
2034 squares and multiplies and the memory access pattern are
2035 independent of the particular secret key. This will mitigate
2036 cache-timing and potential related attacks.
2038 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2039 and this is automatically used by BN_mod_exp_mont() if the new flag
2040 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2041 will use this BN flag for private exponents unless the flag
2042 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2043 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2045 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2047 *) Change the client implementation for SSLv23_method() and
2048 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2049 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2050 (Previously, the SSL 2.0 backwards compatible Client Hello
2051 message format would be used even with SSL_OP_NO_SSLv2.)
2054 *) Add support for smime-type MIME parameter in S/MIME messages which some
2058 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2059 a threadsafe manner. Modify rsa code to use new function and add calls
2060 to dsa and dh code (which had race conditions before).
2063 *) Include the fixed error library code in the C error file definitions
2064 instead of fixing them up at runtime. This keeps the error code
2065 structures constant.
2068 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2070 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2073 *) Fixes for newer kerberos headers. NB: the casts are needed because
2074 the 'length' field is signed on one version and unsigned on another
2075 with no (?) obvious way to tell the difference, without these VC++
2076 complains. Also the "definition" of FAR (blank) is no longer included
2077 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2078 some needed definitions.
2081 *) Undo Cygwin change.
2084 *) Added support for proxy certificates according to RFC 3820.
2085 Because they may be a security thread to unaware applications,
2086 they must be explicitely allowed in run-time. See
2087 docs/HOWTO/proxy_certificates.txt for further information.
2090 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2092 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2093 server and client random values. Previously
2094 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2095 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2097 This change has negligible security impact because:
2099 1. Server and client random values still have 24 bytes of pseudo random
2102 2. Server and client random values are sent in the clear in the initial
2105 3. The master secret is derived using the premaster secret (48 bytes in
2106 size for static RSA ciphersuites) as well as client server and random
2109 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2112 [Stephen Henson, reported by UK NISCC]
2114 *) Use Windows randomness collection on Cygwin.
2117 *) Fix hang in EGD/PRNGD query when communication socket is closed
2118 prematurely by EGD/PRNGD.
2119 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2121 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2124 *) Back-port of selected performance improvements from development
2125 branch, as well as improved support for PowerPC platforms.
2128 *) Add lots of checks for memory allocation failure, error codes to indicate
2129 failure and freeing up memory if a failure occurs.
2130 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2132 *) Add new -passin argument to dgst.
2135 *) Perform some character comparisons of different types in X509_NAME_cmp:
2136 this is needed for some certificates that reencode DNs into UTF8Strings
2137 (in violation of RFC3280) and can't or wont issue name rollover
2141 *) Make an explicit check during certificate validation to see that
2142 the CA setting in each certificate on the chain is correct. As a
2143 side effect always do the following basic checks on extensions,
2144 not just when there's an associated purpose to the check:
2146 - if there is an unhandled critical extension (unless the user
2147 has chosen to ignore this fault)
2148 - if the path length has been exceeded (if one is set at all)
2149 - that certain extensions fit the associated purpose (if one has
2153 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2155 *) Avoid a race condition when CRLs are checked in a multi threaded
2156 environment. This would happen due to the reordering of the revoked
2157 entries during signature checking and serial number lookup. Now the
2158 encoding is cached and the serial number sort performed under a lock.
2159 Add new STACK function sk_is_sorted().
2162 *) Add Delta CRL to the extension code.
2165 *) Various fixes to s3_pkt.c so alerts are sent properly.
2166 [David Holmes <d.holmes@f5.com>]
2168 *) Reduce the chances of duplicate issuer name and serial numbers (in
2169 violation of RFC3280) using the OpenSSL certificate creation utilities.
2170 This is done by creating a random 64 bit value for the initial serial
2171 number when a serial number file is created or when a self signed
2172 certificate is created using 'openssl req -x509'. The initial serial
2173 number file is created using 'openssl x509 -next_serial' in CA.pl
2174 rather than being initialized to 1.
2177 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2179 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2180 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2181 [Joe Orton, Steve Henson]
2183 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2185 [Joe Orton, Steve Henson]
2187 *) Make it possible to have multiple active certificates with the same
2188 subject in the CA index file. This is done only if the keyword
2189 'unique_subject' is set to 'no' in the main CA section (default
2190 if 'CA_default') of the configuration file. The value is saved
2191 with the database itself in a separate index attribute file,
2192 named like the index file with '.attr' appended to the name.
2195 *) X509 verify fixes. Disable broken certificate workarounds when
2196 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2197 keyUsage extension present. Don't accept CRLs with unhandled critical
2198 extensions: since verify currently doesn't process CRL extensions this
2199 rejects a CRL with *any* critical extensions. Add new verify error codes
2203 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2204 A clarification of RFC2560 will require the use of OCTET STRINGs and
2205 some implementations cannot handle the current raw format. Since OpenSSL
2206 copies and compares OCSP nonces as opaque blobs without any attempt at
2207 parsing them this should not create any compatibility issues.
2210 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2211 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2212 this HMAC (and other) operations are several times slower than OpenSSL
2216 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2217 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2219 *) Use the correct content when signing type "other".
2222 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2224 *) Fix various bugs revealed by running the NISCC test suite:
2226 Stop out of bounds reads in the ASN1 code when presented with
2227 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2229 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2231 If verify callback ignores invalid public key errors don't try to check
2232 certificate signature with the NULL public key.
2236 *) New -ignore_err option in ocsp application to stop the server
2237 exiting on the first error in a request.
2240 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2241 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2245 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2246 extra data after the compression methods not only for TLS 1.0
2247 but also for SSL 3.0 (as required by the specification).
2248 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2250 *) Change X509_certificate_type() to mark the key as exported/exportable
2251 when it's 512 *bits* long, not 512 bytes.
2254 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2255 blocks during encryption.
2258 *) Various fixes to base64 BIO and non blocking I/O. On write
2259 flushes were not handled properly if the BIO retried. On read
2260 data was not being buffered properly and had various logic bugs.
2261 This also affects blocking I/O when the data being decoded is a
2265 *) Various S/MIME bugfixes and compatibility changes:
2266 output correct application/pkcs7 MIME type if
2267 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2268 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2269 of files as .eml work). Correctly handle very long lines in MIME
2273 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2275 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2276 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2277 a protocol version number mismatch like a decryption error
2278 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2281 *) Turn on RSA blinding by default in the default implementation
2282 to avoid a timing attack. Applications that don't want it can call
2283 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2284 They would be ill-advised to do so in most cases.
2285 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2287 *) Change RSA blinding code so that it works when the PRNG is not
2288 seeded (in this case, the secret RSA exponent is abused as
2289 an unpredictable seed -- if it is not unpredictable, there
2290 is no point in blinding anyway). Make RSA blinding thread-safe
2291 by remembering the creator's thread ID in rsa->blinding and
2292 having all other threads use local one-time blinding factors
2293 (this requires more computation than sharing rsa->blinding, but
2294 avoids excessive locking; and if an RSA object is not shared
2295 between threads, blinding will still be very fast).
2298 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2299 ENGINE as defaults for all supported algorithms irrespective of
2300 the 'flags' parameter. 'flags' is now honoured, so applications
2301 should make sure they are passing it correctly.
2304 *) Target "mingw" now allows native Windows code to be generated in
2305 the Cygwin environment as well as with the MinGW compiler.
2308 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2310 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2311 via timing by performing a MAC computation even if incorrrect
2312 block cipher padding has been found. This is a countermeasure
2313 against active attacks where the attacker has to distinguish
2314 between bad padding and a MAC verification error. (CVE-2003-0078)
2316 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2317 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2318 Martin Vuagnoux (EPFL, Ilion)]
2320 *) Make the no-err option work as intended. The intention with no-err
2321 is not to have the whole error stack handling routines removed from
2322 libcrypto, it's only intended to remove all the function name and
2323 reason texts, thereby removing some of the footprint that may not
2324 be interesting if those errors aren't displayed anyway.
2326 NOTE: it's still possible for any application or module to have it's
2327 own set of error texts inserted. The routines are there, just not
2328 used by default when no-err is given.
2331 *) Add support for FreeBSD on IA64.
2332 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2334 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2335 Kerberos function mit_des_cbc_cksum(). Before this change,
2336 the value returned by DES_cbc_cksum() was like the one from
2337 mit_des_cbc_cksum(), except the bytes were swapped.
2338 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2340 *) Allow an application to disable the automatic SSL chain building.
2341 Before this a rather primitive chain build was always performed in
2342 ssl3_output_cert_chain(): an application had no way to send the
2343 correct chain if the automatic operation produced an incorrect result.
2345 Now the chain builder is disabled if either:
2347 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2349 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2351 The reasoning behind this is that an application would not want the
2352 auto chain building to take place if extra chain certificates are
2353 present and it might also want a means of sending no additional
2354 certificates (for example the chain has two certificates and the
2358 *) Add the possibility to build without the ENGINE framework.
2359 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2361 *) Under Win32 gmtime() can return NULL: check return value in
2362 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2365 *) DSA routines: under certain error conditions uninitialized BN objects
2366 could be freed. Solution: make sure initialization is performed early
2367 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2368 Nils Larsch <nla@trustcenter.de> via PR#459)
2371 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2372 checked on reconnect on the client side, therefore session resumption
2373 could still fail with a "ssl session id is different" error. This
2374 behaviour is masked when SSL_OP_ALL is used due to
2375 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2376 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2377 followup to PR #377.
2380 *) IA-32 assembler support enhancements: unified ELF targets, support
2381 for SCO/Caldera platforms, fix for Cygwin shared build.
2384 *) Add support for FreeBSD on sparc64. As a consequence, support for
2385 FreeBSD on non-x86 processors is separate from x86 processors on
2386 the config script, much like the NetBSD support.
2387 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2389 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2391 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2394 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2395 code (06) was taken as the first octet of the session ID and the last
2396 octet was ignored consequently. As a result SSLv2 client side session
2397 caching could not have worked due to the session ID mismatch between
2399 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2403 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2404 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2408 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2409 seems that in spite of existing for more than a year, many application
2410 author have done nothing to provide the necessary callbacks, which
2411 means that this particular engine will not work properly anywhere.
2412 This is a very unfortunate situation which forces us, in the name
2413 of usability, to give the hw_ncipher.c a static lock, which is part
2415 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2416 appear in 0.9.8 or later. We EXPECT application authors to have
2417 dealt properly with this when 0.9.8 is released (unless we actually
2418 make such changes in the libcrypto locking code that changes will
2419 have to be made anyway).
2422 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2423 octets have been read, EOF or an error occurs. Without this change
2424 some truncated ASN1 structures will not produce an error.
2427 *) Disable Heimdal support, since it hasn't been fully implemented.
2428 Still give the possibility to force the use of Heimdal, but with
2429 warnings and a request that patches get sent to openssl-dev.
2432 *) Add the VC-CE target, introduce the WINCE sysname, and add
2433 INSTALL.WCE and appropriate conditionals to make it build.
2434 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2436 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2437 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2438 edit numbers of the version.
2439 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2441 *) Introduce safe string copy and catenation functions
2442 (BUF_strlcpy() and BUF_strlcat()).
2443 [Ben Laurie (CHATS) and Richard Levitte]
2445 *) Avoid using fixed-size buffers for one-line DNs.
2446 [Ben Laurie (CHATS)]
2448 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2449 resizing buffers containing secrets, and use where appropriate.
2450 [Ben Laurie (CHATS)]
2452 *) Avoid using fixed size buffers for configuration file location.
2453 [Ben Laurie (CHATS)]
2455 *) Avoid filename truncation for various CA files.
2456 [Ben Laurie (CHATS)]
2458 *) Use sizeof in preference to magic numbers.
2459 [Ben Laurie (CHATS)]
2461 *) Avoid filename truncation in cert requests.
2462 [Ben Laurie (CHATS)]
2464 *) Add assertions to check for (supposedly impossible) buffer
2466 [Ben Laurie (CHATS)]
2468 *) Don't cache truncated DNS entries in the local cache (this could
2469 potentially lead to a spoofing attack).
2470 [Ben Laurie (CHATS)]
2472 *) Fix various buffers to be large enough for hex/decimal
2473 representations in a platform independent manner.
2474 [Ben Laurie (CHATS)]
2476 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2477 resizing buffers containing secrets, and use where appropriate.
2478 [Ben Laurie (CHATS)]
2480 *) Add BIO_indent() to avoid much slightly worrying code to do
2482 [Ben Laurie (CHATS)]
2484 *) Convert sprintf()/BIO_puts() to BIO_printf().
2485 [Ben Laurie (CHATS)]
2487 *) buffer_gets() could terminate with the buffer only half
2489 [Ben Laurie (CHATS)]
2491 *) Add assertions to prevent user-supplied crypto functions from
2492 overflowing internal buffers by having large block sizes, etc.
2493 [Ben Laurie (CHATS)]
2495 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2497 [Ben Laurie (CHATS)]
2499 *) Eliminate unused copy of key in RC4.
2500 [Ben Laurie (CHATS)]
2502 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2503 [Ben Laurie (CHATS)]
2505 *) Fix off-by-one error in EGD path.
2506 [Ben Laurie (CHATS)]
2508 *) If RANDFILE path is too long, ignore instead of truncating.
2509 [Ben Laurie (CHATS)]
2511 *) Eliminate unused and incorrectly sized X.509 structure
2513 [Ben Laurie (CHATS)]
2515 *) Eliminate unused and dangerous function knumber().
2516 [Ben Laurie (CHATS)]
2518 *) Eliminate unused and dangerous structure, KSSL_ERR.
2519 [Ben Laurie (CHATS)]
2521 *) Protect against overlong session ID context length in an encoded
2522 session object. Since these are local, this does not appear to be
2524 [Ben Laurie (CHATS)]
2526 *) Change from security patch (see 0.9.6e below) that did not affect
2527 the 0.9.6 release series:
2529 Remote buffer overflow in SSL3 protocol - an attacker could
2530 supply an oversized master key in Kerberos-enabled versions.
2532 [Ben Laurie (CHATS)]
2534 *) Change the SSL kerb5 codes to match RFC 2712.
2537 *) Make -nameopt work fully for req and add -reqopt switch.
2538 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2540 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2541 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2543 *) Make sure tests can be performed even if the corresponding algorithms
2544 have been removed entirely. This was also the last step to make
2545 OpenSSL compilable with DJGPP under all reasonable conditions.
2546 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2548 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2549 to allow version independent disabling of normally unselected ciphers,
2550 which may be activated as a side-effect of selecting a single cipher.
2552 (E.g., cipher list string "RSA" enables ciphersuites that are left
2553 out of "ALL" because they do not provide symmetric encryption.
2554 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2555 [Lutz Jaenicke, Bodo Moeller]
2557 *) Add appropriate support for separate platform-dependent build
2558 directories. The recommended way to make a platform-dependent
2559 build directory is the following (tested on Linux), maybe with
2562 # Place yourself outside of the OpenSSL source tree. In
2563 # this example, the environment variable OPENSSL_SOURCE
2564 # is assumed to contain the absolute OpenSSL source directory.
2565 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2566 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2567 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2568 mkdir -p `dirname $F`
2569 ln -s $OPENSSL_SOURCE/$F $F
2572 To be absolutely sure not to disturb the source tree, a "make clean"
2573 is a good thing. If it isn't successfull, don't worry about it,
2574 it probably means the source directory is very clean.
2577 *) Make sure any ENGINE control commands make local copies of string
2578 pointers passed to them whenever necessary. Otherwise it is possible
2579 the caller may have overwritten (or deallocated) the original string
2580 data when a later ENGINE operation tries to use the stored values.
2581 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2583 *) Improve diagnostics in file reading and command-line digests.
2584 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2586 *) Add AES modes CFB and OFB to the object database. Correct an
2587 error in AES-CFB decryption.
2590 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2591 allows existing EVP_CIPHER_CTX structures to be reused after
2592 calling EVP_*Final(). This behaviour is used by encryption
2593 BIOs and some applications. This has the side effect that
2594 applications must explicitly clean up cipher contexts with
2595 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2598 *) Check the values of dna and dnb in bn_mul_recursive before calling
2599 bn_mul_comba (a non zero value means the a or b arrays do not contain
2600 n2 elements) and fallback to bn_mul_normal if either is not zero.
2603 *) Fix escaping of non-ASCII characters when using the -subj option
2604 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2607 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2608 form for "surname", serialNumber has no short form.
2609 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2610 therefore remove "mail" short name for "internet 7".
2611 The OID for unique identifiers in X509 certificates is
2612 x500UniqueIdentifier, not uniqueIdentifier.
2613 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2616 *) Add an "init" command to the ENGINE config module and auto initialize
2617 ENGINEs. Without any "init" command the ENGINE will be initialized
2618 after all ctrl commands have been executed on it. If init=1 the
2619 ENGINE is initailized at that point (ctrls before that point are run
2620 on the uninitialized ENGINE and after on the initialized one). If
2621 init=0 then the ENGINE will not be iniatialized at all.
2624 *) Fix the 'app_verify_callback' interface so that the user-defined
2625 argument is actually passed to the callback: In the
2626 SSL_CTX_set_cert_verify_callback() prototype, the callback
2627 declaration has been changed from
2630 int (*cb)(X509_STORE_CTX *,void *);
2631 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2632 i=s->ctx->app_verify_callback(&ctx)
2633 has been changed into
2634 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2636 To update applications using SSL_CTX_set_cert_verify_callback(),
2637 a dummy argument can be added to their callback functions.
2638 [D. K. Smetters <smetters@parc.xerox.com>]
2640 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2641 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2643 *) Add and OPENSSL_LOAD_CONF define which will cause
2644 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2645 This allows older applications to transparently support certain
2646 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2647 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2648 load the config file and OPENSSL_add_all_algorithms_conf() which will
2649 always load it have also been added.
2652 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2653 Adjust NIDs and EVP layer.
2654 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2656 *) Config modules support in openssl utility.
2658 Most commands now load modules from the config file,
2659 though in a few (such as version) this isn't done
2660 because it couldn't be used for anything.
2662 In the case of ca and req the config file used is
2663 the same as the utility itself: that is the -config
2664 command line option can be used to specify an
2668 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2669 use "openssl_conf" if filename is NULL use default openssl config file.
2672 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2673 config section name. Add a new flag to tolerate a missing config file
2674 and move code to CONF_modules_load_file().
2677 *) Support for crypto accelerator cards from Accelerated Encryption
2678 Processing, www.aep.ie. (Use engine 'aep')
2679 The support was copied from 0.9.6c [engine] and adapted/corrected
2680 to work with the new engine framework.
2681 [AEP Inc. and Richard Levitte]
2683 *) Support for SureWare crypto accelerator cards from Baltimore
2684 Technologies. (Use engine 'sureware')
2685 The support was copied from 0.9.6c [engine] and adapted
2686 to work with the new engine framework.
2689 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2690 make the newer ENGINE framework commands for the CHIL engine work.
2691 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2693 *) Make it possible to produce shared libraries on ReliantUNIX.
2694 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2696 *) Add the configuration target debug-linux-ppro.
2697 Make 'openssl rsa' use the general key loading routines
2698 implemented in apps.c, and make those routines able to
2699 handle the key format FORMAT_NETSCAPE and the variant
2701 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2703 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2704 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2706 *) Add -keyform to rsautl, and document -engine.
2707 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2709 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2710 BIO_R_NO_SUCH_FILE error code rather than the generic
2711 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2714 *) Add new functions
2716 ERR_peek_last_error_line
2717 ERR_peek_last_error_line_data.
2718 These are similar to
2721 ERR_peek_error_line_data,
2722 but report on the latest error recorded rather than the first one
2723 still in the error queue.
2724 [Ben Laurie, Bodo Moeller]
2726 *) default_algorithms option in ENGINE config module. This allows things
2728 default_algorithms = ALL
2729 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2732 *) Prelminary ENGINE config module.
2735 *) New experimental application configuration code.
2738 *) Change the AES code to follow the same name structure as all other
2739 symmetric ciphers, and behave the same way. Move everything to
2740 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2741 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2743 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2744 [Ben Laurie and Theo de Raadt]
2746 *) Add option to output public keys in req command.
2747 [Massimiliano Pala madwolf@openca.org]
2749 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2750 (up to about 10% better than before for P-192 and P-224).
2753 *) New functions/macros
2755 SSL_CTX_set_msg_callback(ctx, cb)
2756 SSL_CTX_set_msg_callback_arg(ctx, arg)
2757 SSL_set_msg_callback(ssl, cb)
2758 SSL_set_msg_callback_arg(ssl, arg)
2760 to request calling a callback function
2762 void cb(int write_p, int version, int content_type,
2763 const void *buf, size_t len, SSL *ssl, void *arg)
2765 whenever a protocol message has been completely received
2766 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2767 protocol version according to which the SSL library interprets
2768 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2769 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2770 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2771 specification (change_cipher_spec(20), alert(21), handshake(22)).
2772 'buf' and 'len' point to the actual message, 'ssl' to the
2773 SSL object, and 'arg' is the application-defined value set by
2774 SSL[_CTX]_set_msg_callback_arg().
2776 'openssl s_client' and 'openssl s_server' have new '-msg' options
2777 to enable a callback that displays all protocol messages.
2780 *) Change the shared library support so shared libraries are built as
2781 soon as the corresponding static library is finished, and thereby get
2782 openssl and the test programs linked against the shared library.
2783 This still only happens when the keyword "shard" has been given to
2784 the configuration scripts.
2786 NOTE: shared library support is still an experimental thing, and
2787 backward binary compatibility is still not guaranteed.
2788 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2790 *) Add support for Subject Information Access extension.
2791 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2793 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2794 additional bytes when new memory had to be allocated, not just
2795 when reusing an existing buffer.
2798 *) New command line and configuration option 'utf8' for the req command.
2799 This allows field values to be specified as UTF8 strings.
2802 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2803 runs for the former and machine-readable output for the latter.
2806 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2807 of the e-mail address in the DN (i.e., it will go into a certificate
2808 extension only). The new configuration file option 'email_in_dn = no'
2809 has the same effect.
2810 [Massimiliano Pala madwolf@openca.org]
2812 *) Change all functions with names starting with des_ to be starting
2813 with DES_ instead. Add wrappers that are compatible with libdes,
2814 but are named _ossl_old_des_*. Finally, add macros that map the
2815 des_* symbols to the corresponding _ossl_old_des_* if libdes
2816 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2817 desired, the des_* symbols will be mapped to DES_*, with one
2820 Since we provide two compatibility mappings, the user needs to
2821 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2822 compatibility is desired. The default (i.e., when that macro
2823 isn't defined) is OpenSSL 0.9.6c compatibility.
2825 There are also macros that enable and disable the support of old
2826 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2827 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2828 are defined, the default will apply: to support the old des routines.
2830 In either case, one must include openssl/des.h to get the correct
2831 definitions. Do not try to just include openssl/des_old.h, that
2834 NOTE: This is a major break of an old API into a new one. Software
2835 authors are encouraged to switch to the DES_ style functions. Some
2836 time in the future, des_old.h and the libdes compatibility functions
2837 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2838 default), and then completely removed.
2841 *) Test for certificates which contain unsupported critical extensions.
2842 If such a certificate is found during a verify operation it is
2843 rejected by default: this behaviour can be overridden by either
2844 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2845 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2846 X509_supported_extension() has also been added which returns 1 if a
2847 particular extension is supported.
2850 *) Modify the behaviour of EVP cipher functions in similar way to digests
2851 to retain compatibility with existing code.
2854 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2855 compatibility with existing code. In particular the 'ctx' parameter does
2856 not have to be to be initialized before the call to EVP_DigestInit() and
2857 it is tidied up after a call to EVP_DigestFinal(). New function
2858 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2859 EVP_MD_CTX_copy() changed to not require the destination to be
2860 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2861 requires the destination to be valid.
2863 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2864 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2867 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2868 so that complete 'Handshake' protocol structures are kept in memory
2869 instead of overwriting 'msg_type' and 'length' with 'body' data.
2872 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2873 [Massimo Santin via Richard Levitte]
2875 *) Major restructuring to the underlying ENGINE code. This includes
2876 reduction of linker bloat, separation of pure "ENGINE" manipulation
2877 (initialisation, etc) from functionality dealing with implementations
2878 of specific crypto iterfaces. This change also introduces integrated
2879 support for symmetric ciphers and digest implementations - so ENGINEs
2880 can now accelerate these by providing EVP_CIPHER and EVP_MD
2881 implementations of their own. This is detailed in crypto/engine/README
2882 as it couldn't be adequately described here. However, there are a few
2883 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2884 were changed in the original introduction of ENGINE code have now
2885 reverted back - the hooking from this code to ENGINE is now a good
2886 deal more passive and at run-time, operations deal directly with
2887 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2888 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2889 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2890 they were not being used by the framework as there is no concept of a
2891 BIGNUM_METHOD and they could not be generalised to the new
2892 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2893 ENGINE_cpy() has been removed as it cannot be consistently defined in
2897 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2900 *) Change mkdef.pl to sort symbols that get the same entry number,
2901 and make sure the automatically generated functions ERR_load_*
2902 become part of libeay.num as well.
2905 *) New function SSL_renegotiate_pending(). This returns true once
2906 renegotiation has been requested (either SSL_renegotiate() call
2907 or HelloRequest/ClientHello receveived from the peer) and becomes
2908 false once a handshake has been completed.
2909 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2910 sends a HelloRequest, but does not ensure that a handshake takes
2911 place. SSL_renegotiate_pending() is useful for checking if the
2912 client has followed the request.)
2915 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2916 By default, clients may request session resumption even during
2917 renegotiation (if session ID contexts permit); with this option,
2918 session resumption is possible only in the first handshake.
2920 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2921 more bits available for options that should not be part of
2922 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2925 *) Add some demos for certificate and certificate request creation.
2928 *) Make maximum certificate chain size accepted from the peer application
2929 settable (SSL*_get/set_max_cert_list()), as proposed by
2930 "Douglas E. Engert" <deengert@anl.gov>.
2933 *) Add support for shared libraries for Unixware-7
2934 (Boyd Lynn Gerber <gerberb@zenez.com>).
2937 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2938 be done prior to destruction. Use this to unload error strings from
2939 ENGINEs that load their own error strings. NB: This adds two new API
2940 functions to "get" and "set" this destroy handler in an ENGINE.
2943 *) Alter all existing ENGINE implementations (except "openssl" and
2944 "openbsd") to dynamically instantiate their own error strings. This
2945 makes them more flexible to be built both as statically-linked ENGINEs
2946 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2947 Also, add stub code to each that makes building them as self-contained
2948 shared-libraries easier (see README.ENGINE).
2951 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2952 implementations into applications that are completely implemented in
2953 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2954 commands that can be used to configure what shared-library to load and
2955 to control aspects of the way it is handled. Also, made an update to
2956 the README.ENGINE file that brings its information up-to-date and
2957 provides some information and instructions on the "dynamic" ENGINE
2958 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2961 *) Make it possible to unload ranges of ERR strings with a new
2962 "ERR_unload_strings" function.
2965 *) Add a copy() function to EVP_MD.
2968 *) Make EVP_MD routines take a context pointer instead of just the
2969 md_data void pointer.
2972 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2973 that the digest can only process a single chunk of data
2974 (typically because it is provided by a piece of
2975 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2976 is only going to provide a single chunk of data, and hence the
2977 framework needn't accumulate the data for oneshot drivers.
2980 *) As with "ERR", make it possible to replace the underlying "ex_data"
2981 functions. This change also alters the storage and management of global
2982 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2983 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2984 index counters. The API functions that use this state have been changed
2985 to take a "class_index" rather than pointers to the class's local STACK
2986 and counter, and there is now an API function to dynamically create new
2987 classes. This centralisation allows us to (a) plug a lot of the
2988 thread-safety problems that existed, and (b) makes it possible to clean
2989 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2990 such data would previously have always leaked in application code and
2991 workarounds were in place to make the memory debugging turn a blind eye
2992 to it. Application code that doesn't use this new function will still
2993 leak as before, but their memory debugging output will announce it now
2994 rather than letting it slide.
2996 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2997 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2998 has a return value to indicate success or failure.
3001 *) Make it possible to replace the underlying "ERR" functions such that the
3002 global state (2 LHASH tables and 2 locks) is only used by the "default"
3003 implementation. This change also adds two functions to "get" and "set"
3004 the implementation prior to it being automatically set the first time
3005 any other ERR function takes place. Ie. an application can call "get",
3006 pass the return value to a module it has just loaded, and that module
3007 can call its own "set" function using that value. This means the
3008 module's "ERR" operations will use (and modify) the error state in the
3009 application and not in its own statically linked copy of OpenSSL code.
3012 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3013 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3014 the operation, and provides a more encapsulated way for external code
3015 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3016 to use these functions rather than manually incrementing the counts.
3018 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3021 *) Add EVP test program.
3024 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3027 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3028 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3029 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3030 These allow a CRL to be built without having to access X509_CRL fields
3031 directly. Modify 'ca' application to use new functions.
3034 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3035 bug workarounds. Rollback attack detection is a security feature.
3036 The problem will only arise on OpenSSL servers when TLSv1 is not
3037 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3038 Software authors not wanting to support TLSv1 will have special reasons
3039 for their choice and can explicitly enable this option.
3040 [Bodo Moeller, Lutz Jaenicke]
3042 *) Rationalise EVP so it can be extended: don't include a union of
3043 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3044 (similar to those existing for EVP_CIPHER_CTX).
3049 EVP_MD_CTX_init(&md); /* new function call */
3050 EVP_DigestInit(&md, EVP_sha1());
3051 EVP_DigestUpdate(&md, in, len);
3052 EVP_DigestFinal(&md, out, NULL);
3053 EVP_MD_CTX_cleanup(&md); /* new function call */
3057 *) Make DES key schedule conform to the usual scheme, as well as
3058 correcting its structure. This means that calls to DES functions
3059 now have to pass a pointer to a des_key_schedule instead of a
3060 plain des_key_schedule (which was actually always a pointer
3063 des_key_schedule ks;
3065 des_set_key_checked(..., &ks);
3066 des_ncbc_encrypt(..., &ks, ...);
3068 (Note that a later change renames 'des_...' into 'DES_...'.)
3071 *) Initial reduction of linker bloat: the use of some functions, such as
3072 PEM causes large amounts of unused functions to be linked in due to
3073 poor organisation. For example pem_all.c contains every PEM function
3074 which has a knock on effect of linking in large amounts of (unused)
3075 ASN1 code. Grouping together similar functions and splitting unrelated
3076 functions prevents this.
3079 *) Cleanup of EVP macros.
3082 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3083 correct _ecb suffix.
3086 *) Add initial OCSP responder support to ocsp application. The
3087 revocation information is handled using the text based index
3088 use by the ca application. The responder can either handle
3089 requests generated internally, supplied in files (for example
3090 via a CGI script) or using an internal minimal server.
3093 *) Add configuration choices to get zlib compression for TLS.
3096 *) Changes to Kerberos SSL for RFC 2712 compliance:
3097 1. Implemented real KerberosWrapper, instead of just using
3098 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3099 2. Implemented optional authenticator field of KerberosWrapper.
3101 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3102 and authenticator structs; see crypto/krb5/.
3104 Generalized Kerberos calls to support multiple Kerberos libraries.
3105 [Vern Staats <staatsvr@asc.hpc.mil>,
3106 Jeffrey Altman <jaltman@columbia.edu>
3107 via Richard Levitte]
3109 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3110 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3111 values for each of the key sizes rather than having just
3112 parameters (and 'speed' generating keys each time).
3115 *) Speed up EVP routines.
3118 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3119 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3120 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3121 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3123 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3124 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3125 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3128 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3130 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3133 *) Added the OS2-EMX target.
3134 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3136 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3137 to support NCONF routines in extension code. New function CONF_set_nconf()
3138 to allow functions which take an NCONF to also handle the old LHASH
3139 structure: this means that the old CONF compatible routines can be
3140 retained (in particular wrt extensions) without having to duplicate the
3141 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3144 *) Enhance the general user interface with mechanisms for inner control
3145 and with possibilities to have yes/no kind of prompts.
3148 *) Change all calls to low level digest routines in the library and
3149 applications to use EVP. Add missing calls to HMAC_cleanup() and
3150 don't assume HMAC_CTX can be copied using memcpy().
3151 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3153 *) Add the possibility to control engines through control names but with
3154 arbitrary arguments instead of just a string.
3155 Change the key loaders to take a UI_METHOD instead of a callback
3156 function pointer. NOTE: this breaks binary compatibility with earlier
3157 versions of OpenSSL [engine].
3158 Adapt the nCipher code for these new conditions and add a card insertion
3162 *) Enhance the general user interface with mechanisms to better support
3163 dialog box interfaces, application-defined prompts, the possibility
3164 to use defaults (for example default passwords from somewhere else)
3165 and interrupts/cancellations.
3168 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3169 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3172 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3173 tidy up some unnecessarily weird code in 'sk_new()').
3174 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3176 *) Change the key loading routines for ENGINEs to use the same kind
3177 callback (pem_password_cb) as all other routines that need this
3181 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3182 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3183 than this minimum value is recommended.
3186 *) New random seeder for OpenVMS, using the system process statistics
3187 that are easily reachable.
3190 *) Windows apparently can't transparently handle global
3191 variables defined in DLLs. Initialisations such as:
3193 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3195 wont compile. This is used by the any applications that need to
3196 declare their own ASN1 modules. This was fixed by adding the option
3197 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3198 needed for static libraries under Win32.
3201 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3202 setting of purpose and trust fields. New X509_STORE trust and
3203 purpose functions and tidy up setting in other SSL functions.
3206 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3207 structure. These are inherited by X509_STORE_CTX when it is
3208 initialised. This allows various defaults to be set in the
3209 X509_STORE structure (such as flags for CRL checking and custom
3210 purpose or trust settings) for functions which only use X509_STORE_CTX
3211 internally such as S/MIME.
3213 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3214 trust settings if they are not set in X509_STORE. This allows X509_STORE
3215 purposes and trust (in S/MIME for example) to override any set by default.
3217 Add command line options for CRL checking to smime, s_client and s_server
3221 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3222 are set then the CRL is looked up in the X509_STORE structure and
3223 its validity and signature checked, then if the certificate is found
3224 in the CRL the verify fails with a revoked error.
3226 Various new CRL related callbacks added to X509_STORE_CTX structure.
3228 Command line options added to 'verify' application to support this.
3230 This needs some additional work, such as being able to handle multiple
3231 CRLs with different times, extension based lookup (rather than just
3232 by subject name) and ultimately more complete V2 CRL extension
3236 *) Add a general user interface API (crypto/ui/). This is designed
3237 to replace things like des_read_password and friends (backward
3238 compatibility functions using this new API are provided).
3239 The purpose is to remove prompting functions from the DES code
3240 section as well as provide for prompting through dialog boxes in
3241 a window system and the like.
3244 *) Add "ex_data" support to ENGINE so implementations can add state at a
3245 per-structure level rather than having to store it globally.
3248 *) Make it possible for ENGINE structures to be copied when retrieved by
3249 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3250 This causes the "original" ENGINE structure to act like a template,
3251 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3252 operational state can be localised to each ENGINE structure, despite the
3253 fact they all share the same "methods". New ENGINE structures returned in
3254 this case have no functional references and the return value is the single
3255 structural reference. This matches the single structural reference returned
3256 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3260 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3261 needs to match any other type at all we need to manually clear the
3265 *) Changes to the "openssl engine" utility to include;
3266 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3267 about an ENGINE's available control commands.
3268 - executing control commands from command line arguments using the
3269 '-pre' and '-post' switches. '-post' is only used if '-t' is
3270 specified and the ENGINE is successfully initialised. The syntax for
3271 the individual commands are colon-separated, for example;
3272 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3275 *) New dynamic control command support for ENGINEs. ENGINEs can now
3276 declare their own commands (numbers), names (strings), descriptions,
3277 and input types for run-time discovery by calling applications. A
3278 subset of these commands are implicitly classed as "executable"
3279 depending on their input type, and only these can be invoked through
3280 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3281 can be based on user input, config files, etc). The distinction is
3282 that "executable" commands cannot return anything other than a boolean
3283 result and can only support numeric or string input, whereas some
3284 discoverable commands may only be for direct use through
3285 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3286 pointers, or other custom uses. The "executable" commands are to
3287 support parameterisations of ENGINE behaviour that can be
3288 unambiguously defined by ENGINEs and used consistently across any
3289 OpenSSL-based application. Commands have been added to all the
3290 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3291 control over shared-library paths without source code alterations.
3294 *) Changed all ENGINE implementations to dynamically allocate their
3295 ENGINEs rather than declaring them statically. Apart from this being
3296 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3297 this also allows the implementations to compile without using the
3298 internal engine_int.h header.
3301 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3302 'const' value. Any code that should be able to modify a RAND_METHOD
3303 should already have non-const pointers to it (ie. they should only
3304 modify their own ones).
3307 *) Made a variety of little tweaks to the ENGINE code.
3308 - "atalla" and "ubsec" string definitions were moved from header files
3309 to C code. "nuron" string definitions were placed in variables
3310 rather than hard-coded - allowing parameterisation of these values
3311 later on via ctrl() commands.
3312 - Removed unused "#if 0"'d code.
3313 - Fixed engine list iteration code so it uses ENGINE_free() to release
3314 structural references.
3315 - Constified the RAND_METHOD element of ENGINE structures.
3316 - Constified various get/set functions as appropriate and added
3317 missing functions (including a catch-all ENGINE_cpy that duplicates
3318 all ENGINE values onto a new ENGINE except reference counts/state).
3319 - Removed NULL parameter checks in get/set functions. Setting a method
3320 or function to NULL is a way of cancelling out a previously set
3321 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3322 and doesn't justify the extra error symbols and code.
3323 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3324 flags from engine_int.h to engine.h.
3325 - Changed prototypes for ENGINE handler functions (init(), finish(),
3326 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3329 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3330 to the algorithm using long division. The binary algorithm can be
3331 used only if the modulus is odd. On 32-bit systems, it is faster
3332 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3333 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3334 up to 450 bits. In 64-bit environments, the binary algorithm
3335 appears to be advantageous for much longer moduli; here we use it
3336 for moduli up to 2048 bits.
3339 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3340 could not support the combine flag in choice fields.
3343 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3344 extensions from a certificate request to the certificate.
3347 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3348 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3349 file: this allows the display of the certificate about to be
3350 signed to be customised, to allow certain fields to be included
3351 or excluded and extension details. The old system didn't display
3352 multicharacter strings properly, omitted fields not in the policy
3353 and couldn't display additional details such as extensions.
3356 *) Function EC_POINTs_mul for multiple scalar multiplication
3357 of an arbitrary number of elliptic curve points
3358 \sum scalars[i]*points[i],
3359 optionally including the generator defined for the EC_GROUP:
3360 scalar*generator + \sum scalars[i]*points[i].
3362 EC_POINT_mul is a simple wrapper function for the typical case
3363 that the point list has just one item (besides the optional
3367 *) First EC_METHODs for curves over GF(p):
3369 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3370 operations and provides various method functions that can also
3371 operate with faster implementations of modular arithmetic.
3373 EC_GFp_mont_method() reuses most functions that are part of
3374 EC_GFp_simple_method, but uses Montgomery arithmetic.
3376 [Bodo Moeller; point addition and point doubling
3377 implementation directly derived from source code provided by
3378 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3380 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3381 crypto/ec/ec_lib.c):
3383 Curves are EC_GROUP objects (with an optional group generator)
3384 based on EC_METHODs that are built into the library.
3386 Points are EC_POINT objects based on EC_GROUP objects.
3388 Most of the framework would be able to handle curves over arbitrary
3389 finite fields, but as there are no obvious types for fields other
3390 than GF(p), some functions are limited to that for now.
3393 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3394 that the file contains a complete HTTP response.
3397 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3398 change the def and num file printf format specifier from "%-40sXXX"
3399 to "%-39s XXX". The latter will always guarantee a space after the
3400 field while the former will cause them to run together if the field
3401 is 40 of more characters long.
3404 *) Constify the cipher and digest 'method' functions and structures
3405 and modify related functions to take constant EVP_MD and EVP_CIPHER
3409 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3410 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3413 *) Modify EVP_Digest*() routines so they now return values. Although the
3414 internal software routines can never fail additional hardware versions
3418 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3420 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3421 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3424 ERR_R_NESTED_ASN1_ERROR
3426 ERR_R_MISSING_ASN1_EOS
3427 were 4 .. 9, conflicting with
3428 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3430 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3431 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3433 Add new error code 'ERR_R_INTERNAL_ERROR'.
3436 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3440 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3441 sets the subject name for a new request or supersedes the
3442 subject name in a given request. Formats that can be parsed are
3443 'CN=Some Name, OU=myOU, C=IT'
3445 'CN=Some Name/OU=myOU/C=IT'.
3447 Add options '-batch' and '-verbose' to 'openssl req'.
3448 [Massimiliano Pala <madwolf@hackmasters.net>]
3450 *) Introduce the possibility to access global variables through
3451 functions on platform were that's the best way to handle exporting
3452 global variables in shared libraries. To enable this functionality,
3453 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3454 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3455 is normally done by Configure or something similar).
3457 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3458 in the source file (foo.c) like this:
3460 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3461 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3463 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3464 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3466 OPENSSL_DECLARE_GLOBAL(int,foo);
3467 #define foo OPENSSL_GLOBAL_REF(foo)
3468 OPENSSL_DECLARE_GLOBAL(double,bar);
3469 #define bar OPENSSL_GLOBAL_REF(bar)
3471 The #defines are very important, and therefore so is including the
3472 header file everywhere where the defined globals are used.
3474 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3475 of ASN.1 items, but that structure is a bit different.
3477 The largest change is in util/mkdef.pl which has been enhanced with
3478 better and easier to understand logic to choose which symbols should
3479 go into the Windows .def files as well as a number of fixes and code
3480 cleanup (among others, algorithm keywords are now sorted
3481 lexicographically to avoid constant rewrites).
3484 *) In BN_div() keep a copy of the sign of 'num' before writing the
3485 result to 'rm' because if rm==num the value will be overwritten
3486 and produce the wrong result if 'num' is negative: this caused
3487 problems with BN_mod() and BN_nnmod().
3490 *) Function OCSP_request_verify(). This checks the signature on an
3491 OCSP request and verifies the signer certificate. The signer
3492 certificate is just checked for a generic purpose and OCSP request
3496 *) Add OCSP_check_validity() function to check the validity of OCSP
3497 responses. OCSP responses are prepared in real time and may only
3498 be a few seconds old. Simply checking that the current time lies
3499 between thisUpdate and nextUpdate max reject otherwise valid responses
3500 caused by either OCSP responder or client clock inaccuracy. Instead
3501 we allow thisUpdate and nextUpdate to fall within a certain period of
3502 the current time. The age of the response can also optionally be
3503 checked. Two new options -validity_period and -status_age added to
3507 *) If signature or public key algorithm is unrecognized print out its
3508 OID rather that just UNKNOWN.
3511 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3512 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3513 ID to be generated from the issuer certificate alone which can then be
3514 passed to OCSP_id_issuer_cmp().
3517 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3518 ASN1 modules to export functions returning ASN1_ITEM pointers
3519 instead of the ASN1_ITEM structures themselves. This adds several
3520 new macros which allow the underlying ASN1 function/structure to
3521 be accessed transparently. As a result code should not use ASN1_ITEM
3522 references directly (such as &X509_it) but instead use the relevant
3523 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3524 use of the new ASN1 code on platforms where exporting structures
3525 is problematical (for example in shared libraries) but exporting
3526 functions returning pointers to structures is not.
3529 *) Add support for overriding the generation of SSL/TLS session IDs.
3530 These callbacks can be registered either in an SSL_CTX or per SSL.
3531 The purpose of this is to allow applications to control, if they wish,
3532 the arbitrary values chosen for use as session IDs, particularly as it
3533 can be useful for session caching in multiple-server environments. A
3534 command-line switch for testing this (and any client code that wishes
3535 to use such a feature) has been added to "s_server".
3536 [Geoff Thorpe, Lutz Jaenicke]
3538 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3539 of the form '#if defined(...) || defined(...) || ...' and
3540 '#if !defined(...) && !defined(...) && ...'. This also avoids
3541 the growing number of special cases it was previously handling.
3544 *) Make all configuration macros available for application by making
3545 sure they are available in opensslconf.h, by giving them names starting
3546 with "OPENSSL_" to avoid conflicts with other packages and by making
3547 sure e_os2.h will cover all platform-specific cases together with
3549 Additionally, it is now possible to define configuration/platform-
3550 specific names (called "system identities"). In the C code, these
3551 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3552 macro with the name beginning with "OPENSSL_SYS_", which is determined
3553 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3557 *) New option -set_serial to 'req' and 'x509' this allows the serial
3558 number to use to be specified on the command line. Previously self
3559 signed certificates were hard coded with serial number 0 and the
3560 CA options of 'x509' had to use a serial number in a file which was
3564 *) New options to 'ca' utility to support V2 CRL entry extensions.
3565 Currently CRL reason, invalidity date and hold instruction are
3566 supported. Add new CRL extensions to V3 code and some new objects.
3569 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3570 disable standard block padding (aka PKCS#5 padding) in the EVP
3571 API, which was previously mandatory. This means that the data is
3572 not padded in any way and so the total length much be a multiple
3573 of the block size, otherwise an error occurs.
3576 *) Initial (incomplete) OCSP SSL support.
3579 *) New function OCSP_parse_url(). This splits up a URL into its host,
3580 port and path components: primarily to parse OCSP URLs. New -url
3581 option to ocsp utility.
3584 *) New nonce behavior. The return value of OCSP_check_nonce() now
3585 reflects the various checks performed. Applications can decide
3586 whether to tolerate certain situations such as an absent nonce
3587 in a response when one was present in a request: the ocsp application
3588 just prints out a warning. New function OCSP_add1_basic_nonce()
3589 this is to allow responders to include a nonce in a response even if
3590 the request is nonce-less.
3593 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3594 skipped when using openssl x509 multiple times on a single input file,
3595 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3598 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3599 set string type: to handle setting ASN1_TIME structures. Fix ca
3600 utility to correctly initialize revocation date of CRLs.
3603 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3604 the clients preferred ciphersuites and rather use its own preferences.
3605 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3606 Internet Explorer by ensuring unchanged hash method during stepup.
3607 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3610 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3611 to aes and add a new 'exist' option to print out symbols that don't
3615 *) Additional options to ocsp utility to allow flags to be set and
3616 additional certificates supplied.
3619 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3620 OCSP client a number of certificate to only verify the response
3624 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3625 handle the new API. Currently only ECB, CBC modes supported. Add new
3628 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3629 Encryption Standard (AES) Ciphersuites for Transport Layer
3630 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3631 not enabled by default and were not part of the "ALL" ciphersuite
3632 alias because they were not yet official; they could be
3633 explicitly requested by specifying the "AESdraft" ciphersuite
3634 group alias. In the final release of OpenSSL 0.9.7, the group
3635 alias is called "AES" and is part of "ALL".)
3636 [Ben Laurie, Steve Henson, Bodo Moeller]
3638 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3639 request to response.
3642 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3643 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3644 extract information from a certificate request. OCSP_response_create()
3645 creates a response and optionally adds a basic response structure.
3646 OCSP_basic_add1_status() adds a complete single response to a basic
3647 response and returns the OCSP_SINGLERESP structure just added (to allow
3648 extensions to be included for example). OCSP_basic_add1_cert() adds a
3649 certificate to a basic response and OCSP_basic_sign() signs a basic
3650 response with various flags. New helper functions ASN1_TIME_check()
3651 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3652 (converts ASN1_TIME to GeneralizedTime).
3655 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3656 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3657 structure from a certificate. X509_pubkey_digest() digests the public_key
3658 contents: this is used in various key identifiers.
3661 *) Make sk_sort() tolerate a NULL argument.
3662 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3664 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3665 passed by the function are trusted implicitly. If any of them signed the
3666 response then it is assumed to be valid and is not verified.
3669 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3670 to data. This was previously part of the PKCS7 ASN1 code. This
3671 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3672 [Steve Henson, reported by Kenneth R. Robinette
3673 <support@securenetterm.com>]
3675 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3676 routines: without these tracing memory leaks is very painful.
3677 Fix leaks in PKCS12 and PKCS7 routines.
3680 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3681 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3682 effectively meant GeneralizedTime would never be used. Now it
3683 is initialised to -1 but X509_time_adj() now has to check the value
3684 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3685 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3686 [Steve Henson, reported by Kenneth R. Robinette
3687 <support@securenetterm.com>]
3689 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3690 result in a zero length in the ASN1_INTEGER structure which was
3691 not consistent with the structure when d2i_ASN1_INTEGER() was used
3692 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3693 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3694 where it did not print out a minus for negative ASN1_INTEGER.
3697 *) Add summary printout to ocsp utility. The various functions which
3698 convert status values to strings have been renamed to:
3699 OCSP_response_status_str(), OCSP_cert_status_str() and
3700 OCSP_crl_reason_str() and are no longer static. New options
3701 to verify nonce values and to disable verification. OCSP response
3702 printout format cleaned up.
3705 *) Add additional OCSP certificate checks. These are those specified
3706 in RFC2560. This consists of two separate checks: the CA of the
3707 certificate being checked must either be the OCSP signer certificate
3708 or the issuer of the OCSP signer certificate. In the latter case the
3709 OCSP signer certificate must contain the OCSP signing extended key
3710 usage. This check is performed by attempting to match the OCSP
3711 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3712 in the OCSP_CERTID structures of the response.
3715 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3716 and related routines. This uses the standard OpenSSL certificate
3717 verify routines to perform initial checks (just CA validity) and
3718 to obtain the certificate chain. Then additional checks will be
3719 performed on the chain. Currently the root CA is checked to see
3720 if it is explicitly trusted for OCSP signing. This is used to set
3721 a root CA as a global signing root: that is any certificate that
3722 chains to that CA is an acceptable OCSP signing certificate.
3725 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3726 extensions from a separate configuration file.
3727 As when reading extensions from the main configuration file,
3728 the '-extensions ...' option may be used for specifying the
3730 [Massimiliano Pala <madwolf@comune.modena.it>]
3732 *) New OCSP utility. Allows OCSP requests to be generated or
3733 read. The request can be sent to a responder and the output
3734 parsed, outputed or printed in text form. Not complete yet:
3735 still needs to check the OCSP response validity.
3738 *) New subcommands for 'openssl ca':
3739 'openssl ca -status <serial>' prints the status of the cert with
3740 the given serial number (according to the index file).
3741 'openssl ca -updatedb' updates the expiry status of certificates
3743 [Massimiliano Pala <madwolf@comune.modena.it>]
3745 *) New '-newreq-nodes' command option to CA.pl. This is like
3746 '-newreq', but calls 'openssl req' with the '-nodes' option
3747 so that the resulting key is not encrypted.
3748 [Damien Miller <djm@mindrot.org>]
3750 *) New configuration for the GNU Hurd.
3751 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3753 *) Initial code to implement OCSP basic response verify. This
3754 is currently incomplete. Currently just finds the signer's
3755 certificate and verifies the signature on the response.
3758 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3759 value of OPENSSLDIR. This is available via the new '-d' option
3760 to 'openssl version', and is also included in 'openssl version -a'.
3763 *) Allowing defining memory allocation callbacks that will be given
3764 file name and line number information in additional arguments
3765 (a const char* and an int). The basic functionality remains, as
3766 well as the original possibility to just replace malloc(),
3767 realloc() and free() by functions that do not know about these
3768 additional arguments. To register and find out the current
3769 settings for extended allocation functions, the following
3770 functions are provided:
3772 CRYPTO_set_mem_ex_functions
3773 CRYPTO_set_locked_mem_ex_functions
3774 CRYPTO_get_mem_ex_functions
3775 CRYPTO_get_locked_mem_ex_functions
3777 These work the same way as CRYPTO_set_mem_functions and friends.
3778 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3779 extended allocation function is enabled.
3780 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3781 a conventional allocation function is enabled.
3782 [Richard Levitte, Bodo Moeller]
3784 *) Finish off removing the remaining LHASH function pointer casts.
3785 There should no longer be any prototype-casting required when using
3786 the LHASH abstraction, and any casts that remain are "bugs". See
3787 the callback types and macros at the head of lhash.h for details
3788 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3791 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3792 If /dev/[u]random devices are not available or do not return enough
3793 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3795 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3796 /etc/entropy will be queried once each in this sequence, quering stops
3797 when enough entropy was collected without querying more sockets.
3800 *) Change the Unix RAND_poll() variant to be able to poll several
3801 random devices, as specified by DEVRANDOM, until a sufficient amount
3802 of data has been collected. We spend at most 10 ms on each file
3803 (select timeout) and read in non-blocking mode. DEVRANDOM now
3804 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3805 (previously it was just the string "/dev/urandom"), so on typical
3806 platforms the 10 ms delay will never occur.
3807 Also separate out the Unix variant to its own file, rand_unix.c.
3808 For VMS, there's a currently-empty rand_vms.c.
3811 *) Move OCSP client related routines to ocsp_cl.c. These
3812 provide utility functions which an application needing
3813 to issue a request to an OCSP responder and analyse the
3814 response will typically need: as opposed to those which an
3815 OCSP responder itself would need which will be added later.
3817 OCSP_request_sign() signs an OCSP request with an API similar
3818 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3819 response. OCSP_response_get1_basic() extracts basic response
3820 from response. OCSP_resp_find_status(): finds and extracts status
3821 information from an OCSP_CERTID structure (which will be created
3822 when the request structure is built). These are built from lower
3823 level functions which work on OCSP_SINGLERESP structures but
3824 wont normally be used unless the application wishes to examine
3825 extensions in the OCSP response for example.
3827 Replace nonce routines with a pair of functions.
3828 OCSP_request_add1_nonce() adds a nonce value and optionally
3829 generates a random value. OCSP_check_nonce() checks the
3830 validity of the nonce in an OCSP response.
3833 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3834 This doesn't copy the supplied OCSP_CERTID and avoids the
3835 need to free up the newly created id. Change return type
3836 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3837 This can then be used to add extensions to the request.
3838 Deleted OCSP_request_new(), since most of its functionality
3839 is now in OCSP_REQUEST_new() (and the case insensitive name
3840 clash) apart from the ability to set the request name which
3841 will be added elsewhere.
3844 *) Update OCSP API. Remove obsolete extensions argument from
3845 various functions. Extensions are now handled using the new
3846 OCSP extension code. New simple OCSP HTTP function which
3847 can be used to send requests and parse the response.
3850 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3851 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3852 uses the special reorder version of SET OF to sort the attributes
3853 and reorder them to match the encoded order. This resolves a long
3854 standing problem: a verify on a PKCS7 structure just after signing
3855 it used to fail because the attribute order did not match the
3856 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3857 it uses the received order. This is necessary to tolerate some broken
3858 software that does not order SET OF. This is handled by encoding
3859 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3860 to produce the required SET OF.
3863 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3864 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3865 files to get correct declarations of the ASN.1 item variables.
3868 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3869 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3870 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3871 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3872 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3873 ASN1_ITEM and no wrapper functions.
3876 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3877 replace the old function pointer based I/O routines. Change most of
3878 the *_d2i_bio() and *_d2i_fp() functions to use these.
3881 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3882 lines, recognice more "algorithms" that can be deselected, and make
3883 it complain about algorithm deselection that isn't recognised.
3886 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3887 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3888 to use new functions. Add NO_ASN1_OLD which can be set to remove
3889 some old style ASN1 functions: this can be used to determine if old
3890 code will still work when these eventually go away.
3893 *) New extension functions for OCSP structures, these follow the
3894 same conventions as certificates and CRLs.
3897 *) New function X509V3_add1_i2d(). This automatically encodes and
3898 adds an extension. Its behaviour can be customised with various
3899 flags to append, replace or delete. Various wrappers added for
3900 certifcates and CRLs.
3903 *) Fix to avoid calling the underlying ASN1 print routine when
3904 an extension cannot be parsed. Correct a typo in the
3905 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3908 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3909 entries for variables.
3912 *) Add functionality to apps/openssl.c for detecting locking
3913 problems: As the program is single-threaded, all we have
3914 to do is register a locking callback using an array for
3915 storing which locks are currently held by the program.
3918 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3919 SSL_get_ex_data_X509_STORE_idx(), which is used in
3920 ssl_verify_cert_chain() and thus can be called at any time
3921 during TLS/SSL handshakes so that thread-safety is essential.
3922 Unfortunately, the ex_data design is not at all suited
3923 for multi-threaded use, so it probably should be abolished.
3926 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3927 [Broadcom, tweaked and integrated by Geoff Thorpe]
3929 *) Move common extension printing code to new function
3930 X509V3_print_extensions(). Reorganise OCSP print routines and
3931 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3934 *) New function X509_signature_print() to remove duplication in some
3938 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3939 set (this was treated exactly the same as SET OF previously). This
3940 is used to reorder the STACK representing the structure to match the
3941 encoding. This will be used to get round a problem where a PKCS7
3942 structure which was signed could not be verified because the STACK
3943 order did not reflect the encoded order.
3946 *) Reimplement the OCSP ASN1 module using the new code.
3949 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3950 for its ASN1 operations. The old style function pointers still exist
3951 for now but they will eventually go away.
3954 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3955 completely replaces the old ASN1 functionality with a table driven
3956 encoder and decoder which interprets an ASN1_ITEM structure describing
3957 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3958 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3959 has also been converted to the new form.
3962 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3963 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3964 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3965 for negative moduli.
3968 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3969 of not touching the result's sign bit.
3972 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3976 *) Changed the LHASH code to use prototypes for callbacks, and created
3977 macros to declare and implement thin (optionally static) functions
3978 that provide type-safety and avoid function pointer casting for the
3979 type-specific callbacks.
3982 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3984 [Veers Staats <staatsvr@asc.hpc.mil>,
3985 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3987 *) Reformat the FAQ so the different questions and answers can be divided
3988 in sections depending on the subject.
3991 *) Have the zlib compression code load ZLIB.DLL dynamically under
3995 *) New function BN_mod_sqrt for computing square roots modulo a prime
3996 (using the probabilistic Tonelli-Shanks algorithm unless
3997 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3998 be handled deterministically).
3999 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4001 *) Make BN_mod_inverse faster by explicitly handling small quotients
4002 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4003 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4006 *) New function BN_kronecker.
4009 *) Fix BN_gcd so that it works on negative inputs; the result is
4010 positive unless both parameters are zero.
4011 Previously something reasonably close to an infinite loop was
4012 possible because numbers could be growing instead of shrinking
4013 in the implementation of Euclid's algorithm.
4016 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4017 sign of the number in question.
4019 Fix BN_is_word(a,w) to work correctly for w == 0.
4021 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4022 because its test if the absolute value of 'a' equals 'w'.
4023 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4024 it exists mostly for use in the implementations of BN_is_zero(),
4025 BN_is_one(), and BN_is_word().
4028 *) New function BN_swap.
4031 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4032 the exponentiation functions are more likely to produce reasonable
4033 results on negative inputs.
4036 *) Change BN_mod_mul so that the result is always non-negative.
4037 Previously, it could be negative if one of the factors was negative;
4038 I don't think anyone really wanted that behaviour.
4041 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4042 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4043 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4044 and add new functions:
4053 BN_mod_lshift1_quick
4057 These functions always generate non-negative results.
4059 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4060 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4062 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4063 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4064 be reduced modulo m.
4065 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4068 The following entry accidentily appeared in the CHANGES file
4069 distributed with OpenSSL 0.9.7. The modifications described in
4070 it do *not* apply to OpenSSL 0.9.7.
4072 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4073 was actually never needed) and in BN_mul(). The removal in BN_mul()
4074 required a small change in bn_mul_part_recursive() and the addition
4075 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4076 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4077 bn_sub_words() and bn_add_words() except they take arrays with
4082 *) In 'openssl passwd', verify passwords read from the terminal
4083 unless the '-salt' option is used (which usually means that
4084 verification would just waste user's time since the resulting
4085 hash is going to be compared with some given password hash)
4086 or the new '-noverify' option is used.
4088 This is an incompatible change, but it does not affect
4089 non-interactive use of 'openssl passwd' (passwords on the command
4090 line, '-stdin' option, '-in ...' option) and thus should not
4094 *) Remove all references to RSAref, since there's no more need for it.
4097 *) Make DSO load along a path given through an environment variable
4098 (SHLIB_PATH) with shl_load().
4101 *) Constify the ENGINE code as a result of BIGNUM constification.
4102 Also constify the RSA code and most things related to it. In a
4103 few places, most notable in the depth of the ASN.1 code, ugly
4104 casts back to non-const were required (to be solved at a later
4108 *) Make it so the openssl application has all engines loaded by default.
4111 *) Constify the BIGNUM routines a little more.
4114 *) Add the following functions:
4116 ENGINE_load_cswift()
4118 ENGINE_load_atalla()
4120 ENGINE_load_builtin_engines()
4122 That way, an application can itself choose if external engines that
4123 are built-in in OpenSSL shall ever be used or not. The benefit is
4124 that applications won't have to be linked with libdl or other dso
4125 libraries unless it's really needed.
4127 Changed 'openssl engine' to load all engines on demand.
4128 Changed the engine header files to avoid the duplication of some
4129 declarations (they differed!).
4132 *) 'openssl engine' can now list capabilities.
4135 *) Better error reporting in 'openssl engine'.
4138 *) Never call load_dh_param(NULL) in s_server.
4141 *) Add engine application. It can currently list engines by name and
4142 identity, and test if they are actually available.
4145 *) Improve RPM specification file by forcing symbolic linking and making
4146 sure the installed documentation is also owned by root.root.
4147 [Damien Miller <djm@mindrot.org>]
4149 *) Give the OpenSSL applications more possibilities to make use of
4150 keys (public as well as private) handled by engines.
4153 *) Add OCSP code that comes from CertCo.
4156 *) Add VMS support for the Rijndael code.
4159 *) Added untested support for Nuron crypto accelerator.
4162 *) Add support for external cryptographic devices. This code was
4163 previously distributed separately as the "engine" branch.
4164 [Geoff Thorpe, Richard Levitte]
4166 *) Rework the filename-translation in the DSO code. It is now possible to
4167 have far greater control over how a "name" is turned into a filename
4168 depending on the operating environment and any oddities about the
4169 different shared library filenames on each system.
4172 *) Support threads on FreeBSD-elf in Configure.
4175 *) Fix for SHA1 assembly problem with MASM: it produces
4176 warnings about corrupt line number information when assembling
4177 with debugging information. This is caused by the overlapping
4179 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4182 NCONF_get_number() has no error checking at all. As a replacement,
4183 NCONF_get_number_e() is defined (_e for "error checking") and is
4184 promoted strongly. The old NCONF_get_number is kept around for
4185 binary backward compatibility.
4186 Make it possible for methods to load from something other than a BIO,
4187 by providing a function pointer that is given a name instead of a BIO.
4188 For example, this could be used to load configuration data from an
4192 *) Fix for non blocking accept BIOs. Added new I/O special reason
4193 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4194 with non blocking I/O was not possible because no retry code was
4195 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4199 *) Added the beginnings of Rijndael support.
4202 *) Fix for bug in DirectoryString mask setting. Add support for
4203 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4204 to allow certificate printing to more controllable, additional
4205 'certopt' option to 'x509' to allow new printing options to be
4209 *) Clean old EAY MD5 hack from e_os.h.
4212 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4214 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4215 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4216 [Joe Orton, Steve Henson]
4218 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4220 *) Fix additional bug revealed by the NISCC test suite:
4222 Stop bug triggering large recursion when presented with
4223 certain ASN.1 tags (CVE-2003-0851)
4226 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4228 *) Fix various bugs revealed by running the NISCC test suite:
4230 Stop out of bounds reads in the ASN1 code when presented with
4231 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4233 If verify callback ignores invalid public key errors don't try to check
4234 certificate signature with the NULL public key.
4238 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4239 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4243 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4244 extra data after the compression methods not only for TLS 1.0
4245 but also for SSL 3.0 (as required by the specification).
4246 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4248 *) Change X509_certificate_type() to mark the key as exported/exportable
4249 when it's 512 *bits* long, not 512 bytes.
4252 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4254 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4255 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4256 a protocol version number mismatch like a decryption error
4257 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4260 *) Turn on RSA blinding by default in the default implementation
4261 to avoid a timing attack. Applications that don't want it can call
4262 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4263 They would be ill-advised to do so in most cases.
4264 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4266 *) Change RSA blinding code so that it works when the PRNG is not
4267 seeded (in this case, the secret RSA exponent is abused as
4268 an unpredictable seed -- if it is not unpredictable, there
4269 is no point in blinding anyway). Make RSA blinding thread-safe
4270 by remembering the creator's thread ID in rsa->blinding and
4271 having all other threads use local one-time blinding factors
4272 (this requires more computation than sharing rsa->blinding, but
4273 avoids excessive locking; and if an RSA object is not shared
4274 between threads, blinding will still be very fast).
4277 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4279 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4280 via timing by performing a MAC computation even if incorrrect
4281 block cipher padding has been found. This is a countermeasure
4282 against active attacks where the attacker has to distinguish
4283 between bad padding and a MAC verification error. (CVE-2003-0078)
4285 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4286 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4287 Martin Vuagnoux (EPFL, Ilion)]
4289 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4291 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4292 memory from it's contents. This is done with a counter that will
4293 place alternating values in each byte. This can be used to solve
4294 two issues: 1) the removal of calls to memset() by highly optimizing
4295 compilers, and 2) cleansing with other values than 0, since those can
4296 be read through on certain media, for example a swap space on disk.
4299 *) Bugfix: client side session caching did not work with external caching,
4300 because the session->cipher setting was not restored when reloading
4301 from the external cache. This problem was masked, when
4302 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4303 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4306 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4307 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4308 [Zeev Lieber <zeev-l@yahoo.com>]
4310 *) Undo an undocumented change introduced in 0.9.6e which caused
4311 repeated calls to OpenSSL_add_all_ciphers() and
4312 OpenSSL_add_all_digests() to be ignored, even after calling
4316 *) Change the default configuration reader to deal with last line not
4317 being properly terminated.
4320 *) Change X509_NAME_cmp() so it applies the special rules on handling
4321 DN values that are of type PrintableString, as well as RDNs of type
4322 emailAddress where the value has the type ia5String.
4323 [stefank@valicert.com via Richard Levitte]
4325 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4326 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4327 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4328 the bitwise-OR of the two for use by the majority of applications
4329 wanting this behaviour, and update the docs. The documented
4330 behaviour and actual behaviour were inconsistent and had been
4331 changing anyway, so this is more a bug-fix than a behavioural
4333 [Geoff Thorpe, diagnosed by Nadav Har'El]
4335 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4336 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4339 *) Fix initialization code race conditions in
4340 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4341 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4342 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4343 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4344 ssl2_get_cipher_by_char(),
4345 ssl3_get_cipher_by_char().
4346 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4348 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4349 the cached sessions are flushed, as the remove_cb() might use ex_data
4350 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4351 (see [openssl.org #212]).
4352 [Geoff Thorpe, Lutz Jaenicke]
4354 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4355 length, instead of the encoding length to d2i_ASN1_OBJECT.
4358 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4360 *) [In 0.9.6g-engine release:]
4361 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4362 [Lynn Gazis <lgazis@rainbow.com>]
4364 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4366 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4367 and get fix the header length calculation.
4368 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4369 Alon Kantor <alonk@checkpoint.com> (and others),
4372 *) Use proper error handling instead of 'assertions' in buffer
4373 overflow checks added in 0.9.6e. This prevents DoS (the
4374 assertions could call abort()).
4375 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4377 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4379 *) Add various sanity checks to asn1_get_length() to reject
4380 the ASN1 length bytes if they exceed sizeof(long), will appear
4381 negative or the content length exceeds the length of the
4383 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4385 *) Fix cipher selection routines: ciphers without encryption had no flags
4386 for the cipher strength set and where therefore not handled correctly
4387 by the selection routines (PR #130).
4390 *) Fix EVP_dsa_sha macro.
4394 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4395 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4396 that was added in OpenSSL 0.9.6d.
4398 As the countermeasure turned out to be incompatible with some
4399 broken SSL implementations, the new option is part of SSL_OP_ALL.
4400 SSL_OP_ALL is usually employed when compatibility with weird SSL
4401 implementations is desired (e.g. '-bugs' option to 's_client' and
4402 's_server'), so the new option is automatically set in many
4406 *) Changes in security patch:
4408 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4409 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4410 Air Force Materiel Command, USAF, under agreement number
4413 *) Add various sanity checks to asn1_get_length() to reject
4414 the ASN1 length bytes if they exceed sizeof(long), will appear
4415 negative or the content length exceeds the length of the
4416 supplied buffer. (CVE-2002-0659)
4417 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4419 *) Assertions for various potential buffer overflows, not known to
4421 [Ben Laurie (CHATS)]
4423 *) Various temporary buffers to hold ASCII versions of integers were
4424 too small for 64 bit platforms. (CVE-2002-0655)
4425 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4427 *) Remote buffer overflow in SSL3 protocol - an attacker could
4428 supply an oversized session ID to a client. (CVE-2002-0656)
4429 [Ben Laurie (CHATS)]
4431 *) Remote buffer overflow in SSL2 protocol - an attacker could
4432 supply an oversized client master key. (CVE-2002-0656)
4433 [Ben Laurie (CHATS)]
4435 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4437 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4438 encoded as NULL) with id-dsa-with-sha1.
4439 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4441 *) Check various X509_...() return values in apps/req.c.
4442 [Nils Larsch <nla@trustcenter.de>]
4444 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4445 an end-of-file condition would erronously be flagged, when the CRLF
4446 was just at the end of a processed block. The bug was discovered when
4447 processing data through a buffering memory BIO handing the data to a
4448 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4449 <ptsekov@syntrex.com> and Nedelcho Stanev.
4452 *) Implement a countermeasure against a vulnerability recently found
4453 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4454 before application data chunks to avoid the use of known IVs
4455 with data potentially chosen by the attacker.
4458 *) Fix length checks in ssl3_get_client_hello().
4461 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4462 to prevent ssl3_read_internal() from incorrectly assuming that
4463 ssl3_read_bytes() found application data while handshake
4464 processing was enabled when in fact s->s3->in_read_app_data was
4465 merely automatically cleared during the initial handshake.
4466 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4468 *) Fix object definitions for Private and Enterprise: they were not
4469 recognized in their shortname (=lowercase) representation. Extend
4470 obj_dat.pl to issue an error when using undefined keywords instead
4471 of silently ignoring the problem (Svenning Sorensen
4472 <sss@sss.dnsalias.net>).
4475 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4476 generators, i.e. generators other than 2 and 5. (Previously, the
4477 code did not properly initialise the 'add' and 'rem' values to
4478 BN_generate_prime().)
4480 In the new general case, we do not insist that 'generator' is
4481 actually a primitive root: This requirement is rather pointless;
4482 a generator of the order-q subgroup is just as good, if not
4486 *) Map new X509 verification errors to alerts. Discovered and submitted by
4487 Tom Wu <tom@arcot.com>.
4490 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4491 returning non-zero before the data has been completely received
4492 when using non-blocking I/O.
4493 [Bodo Moeller; problem pointed out by John Hughes]
4495 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4496 [Ben Laurie, Lutz Jaenicke]
4498 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4499 Yoram Zahavi <YoramZ@gilian.com>).
4502 *) Add information about CygWin 1.3 and on, and preserve proper
4503 configuration for the versions before that.
4504 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4506 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4507 check whether we deal with a copy of a session and do not delete from
4508 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4509 <izhar@checkpoint.com>.
4512 *) Do not store session data into the internal session cache, if it
4513 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4514 flag is set). Proposed by Aslam <aslam@funk.com>.
4517 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4521 *) [In 0.9.6d-engine release:]
4522 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4523 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4525 *) Add the configuration target linux-s390x.
4526 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4528 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4529 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4530 variable as an indication that a ClientHello message has been
4531 received. As the flag value will be lost between multiple
4532 invocations of ssl3_accept when using non-blocking I/O, the
4533 function may not be aware that a handshake has actually taken
4534 place, thus preventing a new session from being added to the
4537 To avoid this problem, we now set s->new_session to 2 instead of
4538 using a local variable.
4539 [Lutz Jaenicke, Bodo Moeller]
4541 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4542 if the SSL_R_LENGTH_MISMATCH error is detected.
4543 [Geoff Thorpe, Bodo Moeller]
4545 *) New 'shared_ldflag' column in Configure platform table.
4548 *) Fix EVP_CIPHER_mode macro.
4549 ["Dan S. Camper" <dan@bti.net>]
4551 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4552 type, we must throw them away by setting rr->length to 0.
4553 [D P Chang <dpc@qualys.com>]
4555 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4557 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4558 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4559 worked incorrectly for those cases where range = 10..._2 and
4560 3*range is two bits longer than range.)
4563 *) Only add signing time to PKCS7 structures if it is not already
4567 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4568 OBJ_ld_ce should be OBJ_id_ce.
4569 Also some ip-pda OIDs in crypto/objects/objects.txt were
4570 incorrect (cf. RFC 3039).
4571 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4573 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4574 returns early because it has nothing to do.
4575 [Andy Schneider <andy.schneider@bjss.co.uk>]
4577 *) [In 0.9.6c-engine release:]
4578 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4579 [Andy Schneider <andy.schneider@bjss.co.uk>]
4581 *) [In 0.9.6c-engine release:]
4582 Add support for Cryptographic Appliance's keyserver technology.
4583 (Use engine 'keyclient')
4584 [Cryptographic Appliances and Geoff Thorpe]
4586 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4587 is called via tools/c89.sh because arguments have to be
4588 rearranged (all '-L' options must appear before the first object
4590 [Richard Shapiro <rshapiro@abinitio.com>]
4592 *) [In 0.9.6c-engine release:]
4593 Add support for Broadcom crypto accelerator cards, backported
4595 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4597 *) [In 0.9.6c-engine release:]
4598 Add support for SureWare crypto accelerator cards from
4599 Baltimore Technologies. (Use engine 'sureware')
4600 [Baltimore Technologies and Mark Cox]
4602 *) [In 0.9.6c-engine release:]
4603 Add support for crypto accelerator cards from Accelerated
4604 Encryption Processing, www.aep.ie. (Use engine 'aep')
4605 [AEP Inc. and Mark Cox]
4607 *) Add a configuration entry for gcc on UnixWare.
4608 [Gary Benson <gbenson@redhat.com>]
4610 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4611 messages are stored in a single piece (fixed-length part and
4612 variable-length part combined) and fix various bugs found on the way.
4615 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4616 instead. BIO_gethostbyname() does not know what timeouts are
4617 appropriate, so entries would stay in cache even when they have
4619 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4621 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4622 faced with a pathologically small ClientHello fragment that does
4623 not contain client_version: Instead of aborting with an error,
4624 simply choose the highest available protocol version (i.e.,
4625 TLS 1.0 unless it is disabled). In practice, ClientHello
4626 messages are never sent like this, but this change gives us
4627 strictly correct behaviour at least for TLS.
4630 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4631 never resets s->method to s->ctx->method when called from within
4632 one of the SSL handshake functions.
4633 [Bodo Moeller; problem pointed out by Niko Baric]
4635 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4636 (sent using the client's version number) if client_version is
4637 smaller than the protocol version in use. Also change
4638 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4639 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4640 the client will at least see that alert.
4643 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4647 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4648 client receives HelloRequest while in a handshake.
4649 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4651 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4652 should end in 'break', not 'goto end' which circuments various
4653 cleanups done in state SSL_ST_OK. But session related stuff
4654 must be disabled for SSL_ST_OK in the case that we just sent a
4657 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4658 before just sending a HelloRequest.
4659 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4661 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4662 reveal whether illegal block cipher padding was found or a MAC
4663 verification error occured. (Neither SSLerr() codes nor alerts
4664 are directly visible to potential attackers, but the information
4665 may leak via logfiles.)
4667 Similar changes are not required for the SSL 2.0 implementation
4668 because the number of padding bytes is sent in clear for SSL 2.0,
4669 and the extra bytes are just ignored. However ssl/s2_pkt.c
4670 failed to verify that the purported number of padding bytes is in
4674 *) Add OpenUNIX-8 support including shared libraries
4675 (Boyd Lynn Gerber <gerberb@zenez.com>).
4678 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4679 'wristwatch attack' using huge encoding parameters (cf.
4680 James H. Manger's CRYPTO 2001 paper). Note that the
4681 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4682 encoding parameters and hence was not vulnerable.
4685 *) BN_sqr() bug fix.
4686 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4688 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4689 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4690 followed by modular reduction.
4691 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4693 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4694 equivalent based on BN_pseudo_rand() instead of BN_rand().
4697 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4698 This function was broken, as the check for a new client hello message
4699 to handle SGC did not allow these large messages.
4700 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4703 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4706 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4707 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4710 *) Rework the configuration and shared library support for Tru64 Unix.
4711 The configuration part makes use of modern compiler features and
4712 still retains old compiler behavior for those that run older versions
4713 of the OS. The shared library support part includes a variant that
4714 uses the RPATH feature, and is available through the special
4715 configuration target "alpha-cc-rpath", which will never be selected
4717 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4719 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4720 with the same message size as in ssl3_get_certificate_request().
4721 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4722 messages might inadvertently be reject as too long.
4723 [Petr Lampa <lampa@fee.vutbr.cz>]
4725 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4728 *) Modified SSL library such that the verify_callback that has been set
4729 specificly for an SSL object with SSL_set_verify() is actually being
4730 used. Before the change, a verify_callback set with this function was
4731 ignored and the verify_callback() set in the SSL_CTX at the time of
4732 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4733 to allow the necessary settings.
4736 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4737 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4738 done automatically (in contradiction to the requirements of the C
4739 standard). This made problems when used from OpenSSH.
4742 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4743 dh->length and always used
4745 BN_rand_range(priv_key, dh->p).
4747 BN_rand_range() is not necessary for Diffie-Hellman, and this
4748 specific range makes Diffie-Hellman unnecessarily inefficient if
4749 dh->length (recommended exponent length) is much smaller than the
4750 length of dh->p. We could use BN_rand_range() if the order of
4751 the subgroup was stored in the DH structure, but we only have
4756 BN_rand(priv_key, l, ...)
4758 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4764 RSA_eay_public_encrypt
4765 RSA_eay_private_decrypt
4766 RSA_eay_private_encrypt (signing)
4767 RSA_eay_public_decrypt (signature verification)
4769 (default implementations for RSA_public_encrypt,
4770 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4771 always reject numbers >= n.
4774 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4775 to synchronize access to 'locking_thread'. This is necessary on
4776 systems where access to 'locking_thread' (an 'unsigned long'
4777 variable) is not atomic.
4780 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4781 *before* setting the 'crypto_lock_rand' flag. The previous code had
4782 a race condition if 0 is a valid thread ID.
4783 [Travis Vitek <vitek@roguewave.com>]
4785 *) Add support for shared libraries under Irix.
4786 [Albert Chin-A-Young <china@thewrittenword.com>]
4788 *) Add configuration option to build on Linux on both big-endian and
4790 [Ralf Baechle <ralf@uni-koblenz.de>]
4792 *) Add the possibility to create shared libraries on HP-UX.
4795 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4797 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4798 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4799 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4800 PRNG state recovery was possible based on the output of
4801 one PRNG request appropriately sized to gain knowledge on
4802 'md' followed by enough consecutive 1-byte PRNG requests
4803 to traverse all of 'state'.
4805 1. When updating 'md_local' (the current thread's copy of 'md')
4806 during PRNG output generation, hash all of the previous
4807 'md_local' value, not just the half used for PRNG output.
4809 2. Make the number of bytes from 'state' included into the hash
4810 independent from the number of PRNG bytes requested.
4812 The first measure alone would be sufficient to avoid
4813 Markku-Juhani's attack. (Actually it had never occurred
4814 to me that the half of 'md_local' used for chaining was the
4815 half from which PRNG output bytes were taken -- I had always
4816 assumed that the secret half would be used.) The second
4817 measure makes sure that additional data from 'state' is never
4818 mixed into 'md_local' in small portions; this heuristically
4819 further strengthens the PRNG.
4822 *) Fix crypto/bn/asm/mips3.s.
4825 *) When only the key is given to "enc", the IV is undefined. Print out
4826 an error message in this case.
4829 *) Handle special case when X509_NAME is empty in X509 printing routines.
4832 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4833 positive and less than q.
4836 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4837 used: it isn't thread safe and the add_lock_callback should handle
4839 [Paul Rose <Paul.Rose@bridge.com>]
4841 *) Verify that incoming data obeys the block size in
4842 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4846 [Ulf Möller, Bodo Möller]
4848 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4849 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4850 when fixing the server behaviour for backwards-compatible 'client
4851 hello' messages. (Note that the attack is impractical against
4852 SSL 3.0 and TLS 1.0 anyway because length and version checking
4853 means that the probability of guessing a valid ciphertext is
4854 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4857 Before 0.9.5, the countermeasure (hide the error by generating a
4858 random 'decryption result') did not work properly because
4859 ERR_clear_error() was missing, meaning that SSL_get_error() would
4860 detect the supposedly ignored error.
4862 Both problems are now fixed.
4865 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4866 (previously it was 1024).
4869 *) Fix for compatibility mode trust settings: ignore trust settings
4870 unless some valid trust or reject settings are present.
4873 *) Fix for blowfish EVP: its a variable length cipher.
4876 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4877 parameters in DSA public key structures and return an error in the
4878 DSA routines if parameters are absent.
4881 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4882 in the current directory if neither $RANDFILE nor $HOME was set.
4883 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4884 caused some confusion to Windows users who haven't defined $HOME.
4885 Thus RAND_file_name() is changed again: e_os.h can define a
4886 DEFAULT_HOME, which will be used if $HOME is not set.
4887 For Windows, we use "C:"; on other platforms, we still require
4888 environment variables.
4890 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4891 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4892 having multiple threads call RAND_poll() concurrently.
4895 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4896 combination of a flag and a thread ID variable.
4897 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4898 flag), *other* threads can enter ssleay_add_bytes without obeying
4899 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4900 that they do not hold after the first thread unsets add_do_not_lock).
4903 *) Change bctest again: '-x' expressions are not available in all
4907 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4909 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4910 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4912 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4913 the default extension for executables, if any. Also, make the perl
4914 scripts that use symlink() to test if it really exists and use "cp"
4915 if it doesn't. All this made OpenSSL compilable and installable in
4919 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4920 If SEQUENCE is length is indefinite just set c->slen to the total
4921 amount of data available.
4922 [Steve Henson, reported by shige@FreeBSD.org]
4923 [This change does not apply to 0.9.7.]
4925 *) Change bctest to avoid here-documents inside command substitution
4926 (workaround for FreeBSD /bin/sh bug).
4927 For compatibility with Ultrix, avoid shell functions (introduced
4928 in the bctest version that searches along $PATH).
4931 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4932 with des_encrypt() defined on some operating systems, like Solaris
4936 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4937 On the Importance of Eliminating Errors in Cryptographic
4938 Computations, J. Cryptology 14 (2001) 2, 101-119,
4939 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4942 *) MIPS assembler BIGNUM division bug fix.
4945 *) Disabled incorrect Alpha assembler code.
4948 *) Fix PKCS#7 decode routines so they correctly update the length
4949 after reading an EOC for the EXPLICIT tag.
4951 [This change does not apply to 0.9.7.]
4953 *) Fix bug in PKCS#12 key generation routines. This was triggered
4954 if a 3DES key was generated with a 0 initial byte. Include
4955 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4956 (but broken) behaviour.
4959 *) Enhance bctest to search for a working bc along $PATH and print
4961 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4963 *) Fix memory leaks in err.c: free err_data string if necessary;
4964 don't write to the wrong index in ERR_set_error_data.
4967 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4971 *) Replace rdtsc with _emit statements for VC++ version 5.
4972 [Jeremy Cooper <jeremy@baymoo.org>]
4974 *) Make it possible to reuse SSLv2 sessions.
4977 *) In copy_email() check for >= 0 as a return value for
4978 X509_NAME_get_index_by_NID() since 0 is a valid index.
4979 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4981 *) Avoid coredump with unsupported or invalid public keys by checking if
4982 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4983 PKCS7_verify() fails with non detached data.
4986 *) Don't use getenv in library functions when run as setuid/setgid.
4987 New function OPENSSL_issetugid().
4990 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4991 due to incorrect handling of multi-threading:
4993 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4995 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4997 3. Count how many times MemCheck_off() has been called so that
4998 nested use can be treated correctly. This also avoids
4999 inband-signalling in the previous code (which relied on the
5000 assumption that thread ID 0 is impossible).
5003 *) Add "-rand" option also to s_client and s_server.
5006 *) Fix CPU detection on Irix 6.x.
5007 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5008 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5010 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5013 [This change does not apply to 0.9.7.]
5015 *) Use the cached encoding of an X509_NAME structure rather than
5016 copying it. This is apparently the reason for the libsafe "errors"
5017 but the code is actually correct.
5020 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5021 Bleichenbacher's DSA attack.
5022 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5023 to be set and top=0 forces the highest bit to be set; top=-1 is new
5024 and leaves the highest bit random.
5025 [Ulf Moeller, Bodo Moeller]
5027 *) In the NCONF_...-based implementations for CONF_... queries
5028 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5029 a temporary CONF structure with the data component set to NULL
5030 (which gives segmentation faults in lh_retrieve).
5031 Instead, use NULL for the CONF pointer in CONF_get_string and
5032 CONF_get_number (which may use environment variables) and directly
5033 return NULL from CONF_get_section.
5036 *) Fix potential buffer overrun for EBCDIC.
5039 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5040 keyUsage if basicConstraints absent for a CA.
5043 *) Make SMIME_write_PKCS7() write mail header values with a format that
5044 is more generally accepted (no spaces before the semicolon), since
5045 some programs can't parse those values properly otherwise. Also make
5046 sure BIO's that break lines after each write do not create invalid
5050 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5051 macros previously used would not encode an empty SEQUENCE OF
5052 and break the signature.
5054 [This change does not apply to 0.9.7.]
5056 *) Zero the premaster secret after deriving the master secret in
5060 *) Add some EVP_add_digest_alias registrations (as found in
5061 OpenSSL_add_all_digests()) to SSL_library_init()
5062 aka OpenSSL_add_ssl_algorithms(). This provides improved
5063 compatibility with peers using X.509 certificates
5064 with unconventional AlgorithmIdentifier OIDs.
5067 *) Fix for Irix with NO_ASM.
5068 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5070 *) ./config script fixes.
5071 [Ulf Moeller, Richard Levitte]
5073 *) Fix 'openssl passwd -1'.
5076 *) Change PKCS12_key_gen_asc() so it can cope with non null
5077 terminated strings whose length is passed in the passlen
5078 parameter, for example from PEM callbacks. This was done
5079 by adding an extra length parameter to asc2uni().
5080 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5082 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5083 call failed, free the DSA structure.
5086 *) Fix to uni2asc() to cope with zero length Unicode strings.
5087 These are present in some PKCS#12 files.
5090 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5091 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5092 when writing a 32767 byte record.
5093 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5095 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5096 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5098 (RSA objects have a reference count access to which is protected
5099 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5100 so they are meant to be shared between threads.)
5101 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5102 "Reddie, Steven" <Steven.Reddie@ca.com>]
5104 *) Fix a deadlock in CRYPTO_mem_leaks().
5107 *) Use better test patterns in bntest.
5110 *) rand_win.c fix for Borland C.
5113 *) BN_rshift bugfix for n == 0.
5116 *) Add a 'bctest' script that checks for some known 'bc' bugs
5117 so that 'make test' does not abort just because 'bc' is broken.
5120 *) Store verify_result within SSL_SESSION also for client side to
5121 avoid potential security hole. (Re-used sessions on the client side
5122 always resulted in verify_result==X509_V_OK, not using the original
5123 result of the server certificate verification.)
5126 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5127 SSL3_RT_APPLICATION_DATA, return 0.
5128 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5132 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5133 releases, have been re-implemented by renaming the previous
5134 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5135 and ssl3_read_internal, respectively, and adding 'peek' parameters
5136 to them. The new ssl[23]_{read,peek} functions are calls to
5137 ssl[23]_read_internal with the 'peek' flag set appropriately.
5138 A 'peek' parameter has also been added to ssl3_read_bytes, which
5139 does the actual work for ssl3_read_internal.
5142 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5143 the method-specific "init()" handler. Also clean up ex_data after
5144 calling the method-specific "finish()" handler. Previously, this was
5145 happening the other way round.
5148 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5149 The previous value, 12, was not always sufficient for BN_mod_exp().
5152 *) Make sure that shared libraries get the internal name engine with
5153 the full version number and not just 0. This should mark the
5154 shared libraries as not backward compatible. Of course, this should
5155 be changed again when we can guarantee backward binary compatibility.
5158 *) Fix typo in get_cert_by_subject() in by_dir.c
5159 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5161 *) Rework the system to generate shared libraries:
5163 - Make note of the expected extension for the shared libraries and
5164 if there is a need for symbolic links from for example libcrypto.so.0
5165 to libcrypto.so.0.9.7. There is extended info in Configure for
5168 - Make as few rebuilds of the shared libraries as possible.
5170 - Still avoid linking the OpenSSL programs with the shared libraries.
5172 - When installing, install the shared libraries separately from the
5176 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5178 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5179 and not in SSL_clear because the latter is also used by the
5180 accept/connect functions; previously, the settings made by
5181 SSL_set_read_ahead would be lost during the handshake.
5182 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5184 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5185 Previously, it would create entries for disableed algorithms no
5189 *) Added several new manual pages for SSL_* function.
5192 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5194 *) In ssl23_get_client_hello, generate an error message when faced
5195 with an initial SSL 3.0/TLS record that is too small to contain the
5196 first two bytes of the ClientHello message, i.e. client_version.
5197 (Note that this is a pathologic case that probably has never happened
5198 in real life.) The previous approach was to use the version number
5199 from the record header as a substitute; but our protocol choice
5200 should not depend on that one because it is not authenticated
5201 by the Finished messages.
5204 *) More robust randomness gathering functions for Windows.
5205 [Jeffrey Altman <jaltman@columbia.edu>]
5207 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5208 not set then we don't setup the error code for issuer check errors
5209 to avoid possibly overwriting other errors which the callback does
5210 handle. If an application does set the flag then we assume it knows
5211 what it is doing and can handle the new informational codes
5215 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5216 a general "ANY" type, as such it should be able to decode anything
5217 including tagged types. However it didn't check the class so it would
5218 wrongly interpret tagged types in the same way as their universal
5219 counterpart and unknown types were just rejected. Changed so that the
5220 tagged and unknown types are handled in the same way as a SEQUENCE:
5221 that is the encoding is stored intact. There is also a new type
5222 "V_ASN1_OTHER" which is used when the class is not universal, in this
5223 case we have no idea what the actual type is so we just lump them all
5227 *) On VMS, stdout may very well lead to a file that is written to
5228 in a record-oriented fashion. That means that every write() will
5229 write a separate record, which will be read separately by the
5230 programs trying to read from it. This can be very confusing.
5232 The solution is to put a BIO filter in the way that will buffer
5233 text until a linefeed is reached, and then write everything a
5234 line at a time, so every record written will be an actual line,
5235 not chunks of lines and not (usually doesn't happen, but I've
5236 seen it once) several lines in one record. BIO_f_linebuffer() is
5239 Currently, it's a VMS-only method, because that's where it has
5240 been tested well enough.
5243 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5244 it can return incorrect results.
5245 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5246 but it was in 0.9.6-beta[12].)
5249 *) Disable the check for content being present when verifying detached
5250 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5251 include zero length content when signing messages.
5254 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5255 BIO_ctrl (for BIO pairs).
5258 *) Add DSO method for VMS.
5261 *) Bug fix: Montgomery multiplication could produce results with the
5265 *) Add RPM specification openssl.spec and modify it to build three
5266 packages. The default package contains applications, application
5267 documentation and run-time libraries. The devel package contains
5268 include files, static libraries and function documentation. The
5269 doc package contains the contents of the doc directory. The original
5270 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5273 *) Add a large number of documentation files for many SSL routines.
5274 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5276 *) Add a configuration entry for Sony News 4.
5277 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5279 *) Don't set the two most significant bits to one when generating a
5280 random number < q in the DSA library.
5283 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5284 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5285 the underlying transport is blocking) if a handshake took place.
5286 (The default behaviour is needed by applications such as s_client
5287 and s_server that use select() to determine when to use SSL_read;
5288 but for applications that know in advance when to expect data, it
5289 just makes things more complicated.)
5292 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5296 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5297 work better on such systems.
5298 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5300 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5301 Update PKCS12_parse() so it copies the friendlyName and the
5302 keyid to the certificates aux info.
5305 *) Fix bug in PKCS7_verify() which caused an infinite loop
5306 if there was more than one signature.
5307 [Sven Uszpelkat <su@celocom.de>]
5309 *) Major change in util/mkdef.pl to include extra information
5310 about each symbol, as well as presentig variables as well
5311 as functions. This change means that there's n more need
5312 to rebuild the .num files when some algorithms are excluded.
5315 *) Allow the verify time to be set by an application,
5316 rather than always using the current time.
5319 *) Phase 2 verify code reorganisation. The certificate
5320 verify code now looks up an issuer certificate by a
5321 number of criteria: subject name, authority key id
5322 and key usage. It also verifies self signed certificates
5323 by the same criteria. The main comparison function is
5324 X509_check_issued() which performs these checks.
5326 Lot of changes were necessary in order to support this
5327 without completely rewriting the lookup code.
5329 Authority and subject key identifier are now cached.
5331 The LHASH 'certs' is X509_STORE has now been replaced
5332 by a STACK_OF(X509_OBJECT). This is mainly because an
5333 LHASH can't store or retrieve multiple objects with
5334 the same hash value.
5336 As a result various functions (which were all internal
5337 use only) have changed to handle the new X509_STORE
5338 structure. This will break anything that messed round
5339 with X509_STORE internally.
5341 The functions X509_STORE_add_cert() now checks for an
5342 exact match, rather than just subject name.
5344 The X509_STORE API doesn't directly support the retrieval
5345 of multiple certificates matching a given criteria, however
5346 this can be worked round by performing a lookup first
5347 (which will fill the cache with candidate certificates)
5348 and then examining the cache for matches. This is probably
5349 the best we can do without throwing out X509_LOOKUP
5350 entirely (maybe later...).
5352 The X509_VERIFY_CTX structure has been enhanced considerably.
5354 All certificate lookup operations now go via a get_issuer()
5355 callback. Although this currently uses an X509_STORE it
5356 can be replaced by custom lookups. This is a simple way
5357 to bypass the X509_STORE hackery necessary to make this
5358 work and makes it possible to use more efficient techniques
5359 in future. A very simple version which uses a simple
5360 STACK for its trusted certificate store is also provided
5361 using X509_STORE_CTX_trusted_stack().
5363 The verify_cb() and verify() callbacks now have equivalents
5364 in the X509_STORE_CTX structure.
5366 X509_STORE_CTX also has a 'flags' field which can be used
5367 to customise the verify behaviour.
5370 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5371 excludes S/MIME capabilities.
5374 *) When a certificate request is read in keep a copy of the
5375 original encoding of the signed data and use it when outputing
5376 again. Signatures then use the original encoding rather than
5377 a decoded, encoded version which may cause problems if the
5378 request is improperly encoded.
5381 *) For consistency with other BIO_puts implementations, call
5382 buffer_write(b, ...) directly in buffer_puts instead of calling
5385 In BIO_puts, increment b->num_write as in BIO_write.
5386 [Peter.Sylvester@EdelWeb.fr]
5388 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5389 BN_zero, we may not return a BIGNUM with an array consisting of
5393 *) Avoid calling abort() from within the library when problems are
5394 detected, except if preprocessor symbols have been defined
5395 (such as REF_CHECK, BN_DEBUG etc.).
5398 *) New openssl application 'rsautl'. This utility can be
5399 used for low level RSA operations. DER public key
5400 BIO/fp routines also added.
5403 *) New Configure entry and patches for compiling on QNX 4.
5404 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5406 *) A demo state-machine implementation was sponsored by
5407 Nuron (http://www.nuron.com/) and is now available in
5408 demos/state_machine.
5411 *) New options added to the 'dgst' utility for signature
5412 generation and verification.
5415 *) Unrecognized PKCS#7 content types are now handled via a
5416 catch all ASN1_TYPE structure. This allows unsupported
5417 types to be stored as a "blob" and an application can
5418 encode and decode it manually.
5421 *) Fix various signed/unsigned issues to make a_strex.c
5423 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5425 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5426 length if passed a buffer. ASN1_INTEGER_to_BN failed
5427 if passed a NULL BN and its argument was negative.
5428 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5430 *) Modification to PKCS#7 encoding routines to output definite
5431 length encoding. Since currently the whole structures are in
5432 memory there's not real point in using indefinite length
5433 constructed encoding. However if OpenSSL is compiled with
5434 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5437 *) Added BIO_vprintf() and BIO_vsnprintf().
5440 *) Added more prefixes to parse for in the the strings written
5441 through a logging bio, to cover all the levels that are available
5442 through syslog. The prefixes are now:
5444 PANIC, EMERG, EMR => LOG_EMERG
5445 ALERT, ALR => LOG_ALERT
5446 CRIT, CRI => LOG_CRIT
5447 ERROR, ERR => LOG_ERR
5448 WARNING, WARN, WAR => LOG_WARNING
5449 NOTICE, NOTE, NOT => LOG_NOTICE
5450 INFO, INF => LOG_INFO
5451 DEBUG, DBG => LOG_DEBUG
5453 and as before, if none of those prefixes are present at the
5454 beginning of the string, LOG_ERR is chosen.
5456 On Win32, the LOG_* levels are mapped according to this:
5458 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5459 LOG_WARNING => EVENTLOG_WARNING_TYPE
5460 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5464 *) Made it possible to reconfigure with just the configuration
5465 argument "reconf" or "reconfigure". The command line arguments
5466 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5467 and are retrieved from there when reconfiguring.
5471 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5473 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5476 *) The obj_dat.pl script was messing up the sorting of object
5477 names. The reason was that it compared the quoted version
5478 of strings as a result "OCSP" > "OCSP Signing" because
5479 " > SPACE. Changed script to store unquoted versions of
5480 names and add quotes on output. It was also omitting some
5481 names from the lookup table if they were given a default
5482 value (that is if SN is missing it is given the same
5483 value as LN and vice versa), these are now added on the
5484 grounds that if an object has a name we should be able to
5485 look it up. Finally added warning output when duplicate
5486 short or long names are found.
5489 *) Changes needed for Tandem NSK.
5490 [Scott Uroff <scott@xypro.com>]
5492 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5493 RSA_padding_check_SSLv23(), special padding was never detected
5494 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5495 version rollback attacks was not effective.
5497 In s23_clnt.c, don't use special rollback-attack detection padding
5498 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5499 client; similarly, in s23_srvr.c, don't do the rollback check if
5500 SSL 2.0 is the only protocol enabled in the server.
5503 *) Make it possible to get hexdumps of unprintable data with 'openssl
5504 asn1parse'. By implication, the functions ASN1_parse_dump() and
5505 BIO_dump_indent() are added.
5508 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5509 these print out strings and name structures based on various
5510 flags including RFC2253 support and proper handling of
5511 multibyte characters. Added options to the 'x509' utility
5512 to allow the various flags to be set.
5515 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5516 Also change the functions X509_cmp_current_time() and
5517 X509_gmtime_adj() work with an ASN1_TIME structure,
5518 this will enable certificates using GeneralizedTime in validity
5519 dates to be checked.
5522 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5523 negative public key encodings) on by default,
5524 NO_NEG_PUBKEY_BUG can be set to disable it.
5527 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5528 content octets. An i2c_ASN1_OBJECT is unnecessary because
5529 the encoding can be trivially obtained from the structure.
5532 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5533 not read locks (CRYPTO_r_[un]lock).
5536 *) A first attempt at creating official support for shared
5537 libraries through configuration. I've kept it so the
5538 default is static libraries only, and the OpenSSL programs
5539 are always statically linked for now, but there are
5540 preparations for dynamic linking in place.
5541 This has been tested on Linux and Tru64.
5544 *) Randomness polling function for Win9x, as described in:
5545 Peter Gutmann, Software Generation of Practically Strong
5549 *) Fix so PRNG is seeded in req if using an already existing
5553 *) New options to smime application. -inform and -outform
5554 allow alternative formats for the S/MIME message including
5555 PEM and DER. The -content option allows the content to be
5556 specified separately. This should allow things like Netscape
5557 form signing output easier to verify.
5560 *) Fix the ASN1 encoding of tags using the 'long form'.
5563 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5564 STRING types. These convert content octets to and from the
5565 underlying type. The actual tag and length octets are
5566 already assumed to have been read in and checked. These
5567 are needed because all other string types have virtually
5568 identical handling apart from the tag. By having versions
5569 of the ASN1 functions that just operate on content octets
5570 IMPLICIT tagging can be handled properly. It also allows
5571 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5572 and ASN1_INTEGER are identical apart from the tag.
5575 *) Change the handling of OID objects as follows:
5577 - New object identifiers are inserted in objects.txt, following
5578 the syntax given in objects.README.
5579 - objects.pl is used to process obj_mac.num and create a new
5581 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5584 This is currently kind of a hack, and the perl code in objects.pl
5585 isn't very elegant, but it works as I intended. The simplest way
5586 to check that it worked correctly is to look in obj_dat.h and
5587 check the array nid_objs and make sure the objects haven't moved
5588 around (this is important!). Additions are OK, as well as
5589 consistent name changes.
5592 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5595 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5596 The given file adds to whatever has already been seeded into the
5597 random pool through the RANDFILE configuration file option or
5598 environment variable, or the default random state file.
5601 *) mkstack.pl now sorts each macro group into lexical order.
5602 Previously the output order depended on the order the files
5603 appeared in the directory, resulting in needless rewriting
5607 *) Patches to make OpenSSL compile under Win32 again. Mostly
5608 work arounds for the VC++ problem that it treats func() as
5609 func(void). Also stripped out the parts of mkdef.pl that
5610 added extra typesafe functions: these no longer exist.
5613 *) Reorganisation of the stack code. The macros are now all
5614 collected in safestack.h . Each macro is defined in terms of
5615 a "stack macro" of the form SKM_<name>(type, a, b). The
5616 DEBUG_SAFESTACK is now handled in terms of function casts,
5617 this has the advantage of retaining type safety without the
5618 use of additional functions. If DEBUG_SAFESTACK is not defined
5619 then the non typesafe macros are used instead. Also modified the
5620 mkstack.pl script to handle the new form. Needs testing to see
5621 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5622 the default if no major problems. Similar behaviour for ASN1_SET_OF
5623 and PKCS12_STACK_OF.
5626 *) When some versions of IIS use the 'NET' form of private key the
5627 key derivation algorithm is different. Normally MD5(password) is
5628 used as a 128 bit RC4 key. In the modified case
5629 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5630 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5631 as the old Netscape_RSA functions except they have an additional
5632 'sgckey' parameter which uses the modified algorithm. Also added
5633 an -sgckey command line option to the rsa utility. Thanks to
5634 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5635 algorithm to openssl-dev.
5638 *) The evp_local.h macros were using 'c.##kname' which resulted in
5639 invalid expansion on some systems (SCO 5.0.5 for example).
5640 Corrected to 'c.kname'.
5641 [Phillip Porch <root@theporch.com>]
5643 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5644 a STACK of email addresses from a certificate or request, these look
5645 in the subject name and the subject alternative name extensions and
5646 omit any duplicate addresses.
5649 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5650 This makes DSA verification about 2 % faster.
5653 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5654 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5655 plus overhead for 1024 bit moduli).
5656 This makes exponentiations about 0.5 % faster for 1024 bit
5657 exponents (as measured by "openssl speed rsa2048").
5660 *) Rename memory handling macros to avoid conflicts with other
5662 Malloc => OPENSSL_malloc
5663 Malloc_locked => OPENSSL_malloc_locked
5664 Realloc => OPENSSL_realloc
5665 Free => OPENSSL_free
5668 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5669 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5672 *) CygWin32 support.
5673 [John Jarvie <jjarvie@newsguy.com>]
5675 *) The type-safe stack code has been rejigged. It is now only compiled
5676 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5677 by default all type-specific stack functions are "#define"d back to
5678 standard stack functions. This results in more streamlined output
5679 but retains the type-safety checking possibilities of the original
5683 *) The STACK code has been cleaned up, and certain type declarations
5684 that didn't make a lot of sense have been brought in line. This has
5685 also involved a cleanup of sorts in safestack.h to more correctly
5686 map type-safe stack functions onto their plain stack counterparts.
5687 This work has also resulted in a variety of "const"ifications of
5688 lots of the code, especially "_cmp" operations which should normally
5689 be prototyped with "const" parameters anyway.
5692 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5693 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5694 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5695 where all of 'md' is used each time the PRNG is used, but 'state'
5696 is used only indexed by a cyclic counter. As entropy may not be
5697 well distributed from the beginning, 'md' is important as a
5698 chaining variable. However, the output function chains only half
5699 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5700 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5701 in all of 'state' being rewritten, with the new values depending
5702 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5705 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5706 the handshake is continued after ssl_verify_cert_chain();
5707 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5708 can lead to 'unexplainable' connection aborts later.
5709 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5711 *) Major EVP API cipher revision.
5712 Add hooks for extra EVP features. This allows various cipher
5713 parameters to be set in the EVP interface. Support added for variable
5714 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5715 setting of RC2 and RC5 parameters.
5717 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5720 Remove lots of duplicated code from the EVP library. For example *every*
5721 cipher init() function handles the 'iv' in the same way according to the
5722 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5723 for CFB and OFB modes they zero ctx->num.
5725 New functionality allows removal of S/MIME code RC2 hack.
5727 Most of the routines have the same form and so can be declared in terms
5730 By shifting this to the top level EVP_CipherInit() it can be removed from
5731 all individual ciphers. If the cipher wants to handle IVs or keys
5732 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5735 Change lots of functions like EVP_EncryptUpdate() to now return a
5736 value: although software versions of the algorithms cannot fail
5737 any installed hardware versions can.
5740 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5741 this option is set, tolerate broken clients that send the negotiated
5742 protocol version number instead of the requested protocol version
5746 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5747 i.e. non-zero for export ciphersuites, zero otherwise.
5748 Previous versions had this flag inverted, inconsistent with
5749 rsa_tmp_cb (..._TMP_RSA_CB).
5750 [Bodo Moeller; problem reported by Amit Chopra]
5752 *) Add missing DSA library text string. Work around for some IIS
5753 key files with invalid SEQUENCE encoding.
5756 *) Add a document (doc/standards.txt) that list all kinds of standards
5757 and so on that are implemented in OpenSSL.
5760 *) Enhance c_rehash script. Old version would mishandle certificates
5761 with the same subject name hash and wouldn't handle CRLs at all.
5762 Added -fingerprint option to crl utility, to support new c_rehash
5766 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5769 *) Fix for SSL server purpose checking. Server checking was
5770 rejecting certificates which had extended key usage present
5771 but no ssl client purpose.
5772 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5774 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5775 is a little unclear about how a blank password is handled.
5776 Since the password in encoded as a BMPString with terminating
5777 double NULL a zero length password would end up as just the
5778 double NULL. However no password at all is different and is
5779 handled differently in the PKCS#12 key generation code. NS
5780 treats a blank password as zero length. MSIE treats it as no
5781 password on export: but it will try both on import. We now do
5782 the same: PKCS12_parse() tries zero length and no password if
5783 the password is set to "" or NULL (NULL is now a valid password:
5784 it wasn't before) as does the pkcs12 application.
5787 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5788 perror when PEM_read_bio_X509_REQ fails, the error message must
5789 be obtained from the error queue.
5792 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5793 it in ERR_remove_state if appropriate, and change ERR_get_state
5794 accordingly to avoid race conditions (this is necessary because
5795 thread_hash is no longer constant once set).
5798 *) Bugfix for linux-elf makefile.one.
5801 *) RSA_get_default_method() will now cause a default
5802 RSA_METHOD to be chosen if one doesn't exist already.
5803 Previously this was only set during a call to RSA_new()
5804 or RSA_new_method(NULL) meaning it was possible for
5805 RSA_get_default_method() to return NULL.
5808 *) Added native name translation to the existing DSO code
5809 that will convert (if the flag to do so is set) filenames
5810 that are sufficiently small and have no path information
5811 into a canonical native form. Eg. "blah" converted to
5812 "libblah.so" or "blah.dll" etc.
5815 *) New function ERR_error_string_n(e, buf, len) which is like
5816 ERR_error_string(e, buf), but writes at most 'len' bytes
5817 including the 0 terminator. For ERR_error_string_n, 'buf'
5819 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5821 *) CONF library reworked to become more general. A new CONF
5822 configuration file reader "class" is implemented as well as a
5823 new functions (NCONF_*, for "New CONF") to handle it. The now
5824 old CONF_* functions are still there, but are reimplemented to
5825 work in terms of the new functions. Also, a set of functions
5826 to handle the internal storage of the configuration data is
5827 provided to make it easier to write new configuration file
5828 reader "classes" (I can definitely see something reading a
5829 configuration file in XML format, for example), called _CONF_*,
5830 or "the configuration storage API"...
5832 The new configuration file reading functions are:
5834 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5835 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5837 NCONF_default, NCONF_WIN32
5839 NCONF_dump_fp, NCONF_dump_bio
5841 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5842 NCONF_new creates a new CONF object. This works in the same way
5843 as other interfaces in OpenSSL, like the BIO interface.
5844 NCONF_dump_* dump the internal storage of the configuration file,
5845 which is useful for debugging. All other functions take the same
5846 arguments as the old CONF_* functions wth the exception of the
5847 first that must be a `CONF *' instead of a `LHASH *'.
5849 To make it easer to use the new classes with the old CONF_* functions,
5850 the function CONF_set_default_method is provided.
5853 *) Add '-tls1' option to 'openssl ciphers', which was already
5854 mentioned in the documentation but had not been implemented.
5855 (This option is not yet really useful because even the additional
5856 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5859 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5860 OpenSSL-based applications) load shared libraries and bind to
5861 them in a portable way.
5862 [Geoff Thorpe, with contributions from Richard Levitte]
5864 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5866 *) Make sure _lrotl and _lrotr are only used with MSVC.
5868 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5869 (the default implementation of RAND_status).
5871 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5872 to '-clrext' (= clear extensions), as intended and documented.
5873 [Bodo Moeller; inconsistency pointed out by Michael Attili
5876 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5877 was larger than the MD block size.
5878 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5880 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5881 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5882 using the passed key: if the passed key was a private key the result
5883 of X509_print(), for example, would be to print out all the private key
5887 *) des_quad_cksum() byte order bug fix.
5888 [Ulf Möller, using the problem description in krb4-0.9.7, where
5889 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5891 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5893 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5895 *) For easily testing in shell scripts whether some command
5896 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5897 returns with exit code 0 iff no command of the given name is available.
5898 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5899 the output goes to stdout and nothing is printed to stderr.
5900 Additional arguments are always ignored.
5902 Since for each cipher there is a command of the same name,
5903 the 'no-cipher' compilation switches can be tested this way.
5905 ('openssl no-XXX' is not able to detect pseudo-commands such
5906 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5909 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5912 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5913 is set; it will be thrown away anyway because each handshake creates
5915 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5916 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5917 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5918 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5921 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5922 'Q' and 'R' lose their special meanings (quit/renegotiate).
5923 This is part of what -quiet does; unlike -quiet, -ign_eof
5924 does not suppress any output.
5927 *) Add compatibility options to the purpose and trust code. The
5928 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5929 accepts a certificate or CA, this was the previous behaviour,
5930 with all the associated security issues.
5932 X509_TRUST_COMPAT is the old trust behaviour: only and
5933 automatically trust self signed roots in certificate store. A
5934 new trust setting X509_TRUST_DEFAULT is used to specify that
5935 a purpose has no associated trust setting and it should instead
5936 use the value in the default purpose.
5939 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5940 and fix a memory leak.
5943 *) In util/mkerr.pl (which implements 'make errors'), preserve
5944 reason strings from the previous version of the .c file, as
5945 the default to have only downcase letters (and digits) in
5946 automatically generated reasons codes is not always appropriate.
5949 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5950 using strerror. Previously, ERR_reason_error_string() returned
5951 library names as reason strings for SYSerr; but SYSerr is a special
5952 case where small numbers are errno values, not library numbers.
5955 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5956 converts DSA parameters into DH parameters. (When creating parameters,
5957 DSA_generate_parameters is used.)
5960 *) Include 'length' (recommended exponent length) in C code generated
5961 by 'openssl dhparam -C'.
5964 *) The second argument to set_label in perlasm was already being used
5965 so couldn't be used as a "file scope" flag. Moved to third argument
5969 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5970 instead of RAND_bytes for encryption IVs and salts.
5973 *) Include RAND_status() into RAND_METHOD instead of implementing
5974 it only for md_rand.c Otherwise replacing the PRNG by calling
5975 RAND_set_rand_method would be impossible.
5978 *) Don't let DSA_generate_key() enter an infinite loop if the random
5979 number generation fails.
5982 *) New 'rand' application for creating pseudo-random output.
5985 *) Added configuration support for Linux/IA64
5986 [Rolf Haberrecker <rolf@suse.de>]
5988 *) Assembler module support for Mingw32.
5991 *) Shared library support for HPUX (in shlib/).
5992 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5994 *) Shared library support for Solaris gcc.
5995 [Lutz Behnke <behnke@trustcenter.de>]
5997 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5999 *) PKCS7_encrypt() was adding text MIME headers twice because they
6000 were added manually and by SMIME_crlf_copy().
6003 *) In bntest.c don't call BN_rand with zero bits argument.
6004 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6006 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6007 case was implemented. This caused BN_div_recp() to fail occasionally.
6010 *) Add an optional second argument to the set_label() in the perl
6011 assembly language builder. If this argument exists and is set
6012 to 1 it signals that the assembler should use a symbol whose
6013 scope is the entire file, not just the current function. This
6014 is needed with MASM which uses the format label:: for this scope.
6015 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6017 *) Change the ASN1 types so they are typedefs by default. Before
6018 almost all types were #define'd to ASN1_STRING which was causing
6019 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6023 *) Change names of new functions to the new get1/get0 naming
6024 convention: After 'get1', the caller owns a reference count
6025 and has to call ..._free; 'get0' returns a pointer to some
6026 data structure without incrementing reference counters.
6027 (Some of the existing 'get' functions increment a reference
6028 counter, some don't.)
6029 Similarly, 'set1' and 'add1' functions increase reference
6030 counters or duplicate objects.
6033 *) Allow for the possibility of temp RSA key generation failure:
6034 the code used to assume it always worked and crashed on failure.
6037 *) Fix potential buffer overrun problem in BIO_printf().
6038 [Ulf Möller, using public domain code by Patrick Powell; problem
6039 pointed out by David Sacerdote <das33@cornell.edu>]
6041 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6042 RAND_egd() and RAND_status(). In the command line application,
6043 the EGD socket can be specified like a seed file using RANDFILE
6047 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6048 Some CAs (e.g. Verisign) distribute certificates in this form.
6051 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6052 list to exclude them. This means that no special compilation option
6053 is needed to use anonymous DH: it just needs to be included in the
6057 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6058 EVP_MD_type. The old functionality is available in a new macro called
6059 EVP_MD_md(). Change code that uses it and update docs.
6062 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6063 where the 'void *' argument is replaced by a function pointer argument.
6064 Previously 'void *' was abused to point to functions, which works on
6065 many platforms, but is not correct. As these functions are usually
6066 called by macros defined in OpenSSL header files, most source code
6067 should work without changes.
6070 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6071 sections with information on -D... compiler switches used for
6072 compiling the library so that applications can see them. To enable
6073 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6074 must be defined. E.g.,
6075 #define OPENSSL_ALGORITHM_DEFINES
6076 #include <openssl/opensslconf.h>
6077 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6078 [Richard Levitte, Ulf and Bodo Möller]
6080 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6084 *) Change the 'other' type in certificate aux info to a STACK_OF
6085 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6086 the required ASN1 format: arbitrary types determined by an OID.
6089 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6090 argument to 'req'. This is not because the function is newer or
6091 better than others it just uses the work 'NEW' in the certificate
6092 request header lines. Some software needs this.
6095 *) Reorganise password command line arguments: now passwords can be
6096 obtained from various sources. Delete the PEM_cb function and make
6097 it the default behaviour: i.e. if the callback is NULL and the
6098 usrdata argument is not NULL interpret it as a null terminated pass
6099 phrase. If usrdata and the callback are NULL then the pass phrase
6100 is prompted for as usual.
6103 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6104 the support is automatically enabled. The resulting binaries will
6105 autodetect the card and use it if present.
6106 [Ben Laurie and Compaq Inc.]
6108 *) Work around for Netscape hang bug. This sends certificate request
6109 and server done in one record. Since this is perfectly legal in the
6110 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6111 the bugs/SSLv3 entry for more info.
6114 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6117 *) Add -rand argument to smime and pkcs12 applications and read/write
6121 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6124 *) Add command line password options to the remaining applications.
6127 *) Bug fix for BN_div_recp() for numerators with an even number of
6131 *) More tests in bntest.c, and changed test_bn output.
6134 *) ./config recognizes MacOS X now.
6137 *) Bug fix for BN_div() when the first words of num and divsor are
6138 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6141 *) Add support for various broken PKCS#8 formats, and command line
6142 options to produce them.
6145 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6146 get temporary BIGNUMs from a BN_CTX.
6149 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6153 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6154 include a #define from the old name to the new. The original intent
6155 was that statically linked binaries could for example just call
6156 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6157 link with digests. This never worked becayse SSLeay_add_all_digests()
6158 and SSLeay_add_all_ciphers() were in the same source file so calling
6159 one would link with the other. They are now in separate source files.
6162 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6165 *) Use a less unusual form of the Miller-Rabin primality test (it used
6166 a binary algorithm for exponentiation integrated into the Miller-Rabin
6167 loop, our standard modexp algorithms are faster).
6170 *) Support for the EBCDIC character set completed.
6171 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6173 *) Source code cleanups: use const where appropriate, eliminate casts,
6174 use void * instead of char * in lhash.
6177 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6178 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6179 this the server could overwrite ephemeral keys that the client
6183 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6184 using 50 iterations of the Rabin-Miller test.
6186 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6187 iterations of the Rabin-Miller test as required by the appendix
6188 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6189 As BN_is_prime_fasttest includes trial division, DSA parameter
6190 generation becomes much faster.
6192 This implies a change for the callback functions in DSA_is_prime
6193 and DSA_generate_parameters: The callback function is called once
6194 for each positive witness in the Rabin-Miller test, not just
6195 occasionally in the inner loop; and the parameters to the
6196 callback function now provide an iteration count for the outer
6197 loop rather than for the current invocation of the inner loop.
6198 DSA_generate_parameters additionally can call the callback
6199 function with an 'iteration count' of -1, meaning that a
6200 candidate has passed the trial division test (when q is generated
6201 from an application-provided seed, trial division is skipped).
6204 *) New function BN_is_prime_fasttest that optionally does trial
6205 division before starting the Rabin-Miller test and has
6206 an additional BN_CTX * argument (whereas BN_is_prime always
6207 has to allocate at least one BN_CTX).
6208 'callback(1, -1, cb_arg)' is called when a number has passed the
6209 trial division stage.
6212 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6216 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6219 *) New function BN_pseudo_rand().
6222 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6223 bignum version of BN_from_montgomery() with the working code from
6224 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6228 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6229 made it impossible to use the same SSL_SESSION data structure in
6230 SSL2 clients in multiple threads.
6233 *) The return value of RAND_load_file() no longer counts bytes obtained
6234 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6235 to seed the PRNG (previously an explicit byte count was required).
6236 [Ulf Möller, Bodo Möller]
6238 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6239 used (char *) instead of (void *) and had casts all over the place.
6242 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6245 *) Retain source code compatibility for BN_prime_checks macro:
6246 BN_is_prime(..., BN_prime_checks, ...) now uses
6247 BN_prime_checks_for_size to determine the appropriate number of
6248 Rabin-Miller iterations.
6251 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6252 DH_CHECK_P_NOT_SAFE_PRIME.
6253 (Check if this is true? OpenPGP calls them "strong".)
6256 *) Merge the functionality of "dh" and "gendh" programs into a new program
6257 "dhparam". The old programs are retained for now but will handle DH keys
6258 (instead of parameters) in future.
6261 *) Make the ciphers, s_server and s_client programs check the return values
6262 when a new cipher list is set.
6265 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6266 ciphers. Before when the 56bit ciphers were enabled the sorting was
6269 The syntax for the cipher sorting has been extended to support sorting by
6270 cipher-strength (using the strength_bits hard coded in the tables).
6271 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6273 Fix a bug in the cipher-command parser: when supplying a cipher command
6274 string with an "undefined" symbol (neither command nor alphanumeric
6275 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6276 an error is flagged.
6278 Due to the strength-sorting extension, the code of the
6279 ssl_create_cipher_list() function was completely rearranged. I hope that
6280 the readability was also increased :-)
6281 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6283 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6284 for the first serial number and places 2 in the serial number file. This
6285 avoids problems when the root CA is created with serial number zero and
6286 the first user certificate has the same issuer name and serial number
6290 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6291 the new code. Add documentation for this stuff.
6294 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6295 X509_*() to X509at_*() on the grounds that they don't handle X509
6296 structures and behave in an analagous way to the X509v3 functions:
6297 they shouldn't be called directly but wrapper functions should be used
6300 So we also now have some wrapper functions that call the X509at functions
6301 when passed certificate requests. (TO DO: similar things can be done with
6302 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6303 things. Some of these need some d2i or i2d and print functionality
6304 because they handle more complex structures.)
6307 *) Add missing #ifndefs that caused missing symbols when building libssl
6308 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6309 NO_RSA in ssl/s2*.c.
6310 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6312 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6313 has a return value which indicates the quality of the random data
6314 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6315 error queue. New function RAND_pseudo_bytes() generates output that is
6316 guaranteed to be unique but not unpredictable. RAND_add is like
6317 RAND_seed, but takes an extra argument for an entropy estimate
6318 (RAND_seed always assumes full entropy).
6321 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6322 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6323 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6324 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6325 false-positive rate of at most 2^-80 for random input.
6328 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6331 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6332 in the 0.9.5 release), this returns the chain
6333 from an X509_CTX structure with a dup of the stack and all
6334 the X509 reference counts upped: so the stack will exist
6335 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6338 Also make SSL_SESSION_print() print out the verify return
6342 *) Add manpage for the pkcs12 command. Also change the default
6343 behaviour so MAC iteration counts are used unless the new
6344 -nomaciter option is used. This improves file security and
6345 only older versions of MSIE (4.0 for example) need it.
6348 *) Honor the no-xxx Configure options when creating .DEF files.
6351 *) Add PKCS#10 attributes to field table: challengePassword,
6352 unstructuredName and unstructuredAddress. These are taken from
6353 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6354 international characters are used.
6356 More changes to X509_ATTRIBUTE code: allow the setting of types
6357 based on strings. Remove the 'loc' parameter when adding
6358 attributes because these will be a SET OF encoding which is sorted
6362 *) Initial changes to the 'req' utility to allow request generation
6363 automation. This will allow an application to just generate a template
6364 file containing all the field values and have req construct the
6367 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6368 used all over the place including certificate requests and PKCS#7
6369 structures. They are currently handled manually where necessary with
6370 some primitive wrappers for PKCS#7. The new functions behave in a
6371 manner analogous to the X509 extension functions: they allow
6372 attributes to be looked up by NID and added.
6374 Later something similar to the X509V3 code would be desirable to
6375 automatically handle the encoding, decoding and printing of the
6376 more complex types. The string types like challengePassword can
6377 be handled by the string table functions.
6379 Also modified the multi byte string table handling. Now there is
6380 a 'global mask' which masks out certain types. The table itself
6381 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6382 is useful when for example there is only one permissible type
6383 (as in countryName) and using the mask might result in no valid
6387 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6388 SSL_get_peer_finished to allow applications to obtain the latest
6389 Finished messages sent to the peer or expected from the peer,
6390 respectively. (SSL_get_peer_finished is usually the Finished message
6391 actually received from the peer, otherwise the protocol will be aborted.)
6393 As the Finished message are message digests of the complete handshake
6394 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6395 be used for external authentication procedures when the authentication
6396 provided by SSL/TLS is not desired or is not enough.
6399 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6400 the host supports BWX extension and if Compaq C is present on the
6401 $PATH. Just exploiting of the BWX extension results in 20-30%
6402 performance kick for some algorithms, e.g. DES and RC4 to mention
6403 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6407 *) Add support for MS "fast SGC". This is arguably a violation of the
6408 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6409 weak crypto and after checking the certificate is SGC a second one
6410 with strong crypto. MS SGC stops the first handshake after receiving
6411 the server certificate message and sends a second client hello. Since
6412 a server will typically do all the time consuming operations before
6413 expecting any further messages from the client (server key exchange
6414 is the most expensive) there is little difference between the two.
6416 To get OpenSSL to support MS SGC we have to permit a second client
6417 hello message after we have sent server done. In addition we have to
6418 reset the MAC if we do get this second client hello.
6421 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6422 if a DER encoded private key is RSA or DSA traditional format. Changed
6423 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6424 format DER encoded private key. Newer code should use PKCS#8 format which
6425 has the key type encoded in the ASN1 structure. Added DER private key
6426 support to pkcs8 application.
6429 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6430 ciphersuites has been selected (as required by the SSL 3/TLS 1
6431 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6432 is set, we interpret this as a request to violate the specification
6433 (the worst that can happen is a handshake failure, and 'correct'
6434 behaviour would result in a handshake failure anyway).
6437 *) In SSL_CTX_add_session, take into account that there might be multiple
6438 SSL_SESSION structures with the same session ID (e.g. when two threads
6439 concurrently obtain them from an external cache).
6440 The internal cache can handle only one SSL_SESSION with a given ID,
6441 so if there's a conflict, we now throw out the old one to achieve
6445 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6446 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6447 some routines that use cipher OIDs: some ciphers do not have OIDs
6448 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6452 *) Simplify the trust setting structure and code. Now we just have
6453 two sequences of OIDs for trusted and rejected settings. These will
6454 typically have values the same as the extended key usage extension
6455 and any application specific purposes.
6457 The trust checking code now has a default behaviour: it will just
6458 check for an object with the same NID as the passed id. Functions can
6459 be provided to override either the default behaviour or the behaviour
6460 for a given id. SSL client, server and email already have functions
6461 in place for compatibility: they check the NID and also return "trusted"
6462 if the certificate is self signed.
6465 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6466 traditional format into an EVP_PKEY structure.
6469 *) Add a password callback function PEM_cb() which either prompts for
6470 a password if usr_data is NULL or otherwise assumes it is a null
6471 terminated password. Allow passwords to be passed on command line
6472 environment or config files in a few more utilities.
6475 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6476 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6477 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6478 Update documentation.
6481 *) Support for ASN1 "NULL" type. This could be handled before by using
6482 ASN1_TYPE but there wasn't any function that would try to read a NULL
6483 and produce an error if it couldn't. For compatibility we also have
6484 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6485 don't allocate anything because they don't need to.
6488 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6490 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6492 *) Rebuild of the memory allocation routines used by OpenSSL code and
6493 possibly others as well. The purpose is to make an interface that
6494 provide hooks so anyone can build a separate set of allocation and
6495 deallocation routines to be used by OpenSSL, for example memory
6496 pool implementations, or something else, which was previously hard
6497 since Malloc(), Realloc() and Free() were defined as macros having
6498 the values malloc, realloc and free, respectively (except for Win32
6499 compilations). The same is provided for memory debugging code.
6500 OpenSSL already comes with functionality to find memory leaks, but
6501 this gives people a chance to debug other memory problems.
6503 With these changes, a new set of functions and macros have appeared:
6505 CRYPTO_set_mem_debug_functions() [F]
6506 CRYPTO_get_mem_debug_functions() [F]
6507 CRYPTO_dbg_set_options() [F]
6508 CRYPTO_dbg_get_options() [F]
6509 CRYPTO_malloc_debug_init() [M]
6511 The memory debug functions are NULL by default, unless the library
6512 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6513 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6514 gives the standard debugging functions that come with OpenSSL) or
6515 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6516 provided by the library user) must be used. When the standard
6517 debugging functions are used, CRYPTO_dbg_set_options can be used to
6518 request additional information:
6519 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6520 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6522 Also, things like CRYPTO_set_mem_functions will always give the
6523 expected result (the new set of functions is used for allocation
6524 and deallocation) at all times, regardless of platform and compiler
6527 To finish it up, some functions that were never use in any other
6528 way than through macros have a new API and new semantic:
6531 CRYPTO_dbg_realloc()
6534 All macros of value have retained their old syntax.
6535 [Richard Levitte and Bodo Moeller]
6537 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6538 ordering of SMIMECapabilities wasn't in "strength order" and there
6539 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6543 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6544 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6545 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6547 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6548 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6549 functionality to handle multipart/signed properly) and a utility
6550 called 'smime' to call all this stuff. This is based on code I
6551 originally wrote for Celo who have kindly allowed it to be
6552 included in OpenSSL.
6555 *) Add variants des_set_key_checked and des_set_key_unchecked of
6556 des_set_key (aka des_key_sched). Global variable des_check_key
6557 decides which of these is called by des_set_key; this way
6558 des_check_key behaves as it always did, but applications and
6559 the library itself, which was buggy for des_check_key == 1,
6560 have a cleaner way to pick the version they need.
6563 *) New function PKCS12_newpass() which changes the password of a
6567 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6568 dynamic mix. In both cases the ids can be used as an index into the
6569 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6570 functions so they accept a list of the field values and the
6571 application doesn't need to directly manipulate the X509_TRUST
6575 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6579 *) Modify the way the V3 extension code looks up extensions. This now
6580 works in a similar way to the object code: we have some "standard"
6581 extensions in a static table which is searched with OBJ_bsearch()
6582 and the application can add dynamic ones if needed. The file
6583 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6584 updated whenever a new extension is added to the core code and kept
6585 in ext_nid order. There is a simple program 'tabtest.c' which checks
6586 this. New extensions are not added too often so this file can readily
6587 be maintained manually.
6589 There are two big advantages in doing things this way. The extensions
6590 can be looked up immediately and no longer need to be "added" using
6591 X509V3_add_standard_extensions(): this function now does nothing.
6592 [Side note: I get *lots* of email saying the extension code doesn't
6593 work because people forget to call this function]
6594 Also no dynamic allocation is done unless new extensions are added:
6595 so if we don't add custom extensions there is no need to call
6596 X509V3_EXT_cleanup().
6599 *) Modify enc utility's salting as follows: make salting the default. Add a
6600 magic header, so unsalted files fail gracefully instead of just decrypting
6601 to garbage. This is because not salting is a big security hole, so people
6602 should be discouraged from doing it.
6605 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6606 digest to be passed on the command line but it only used this
6607 parameter when signing a certificate. Modified so all relevant
6608 operations are affected by the digest parameter including the
6609 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6610 DSA key was used because it didn't fix the digest.
6613 *) Initial certificate chain verify code. Currently tests the untrusted
6614 certificates for consistency with the verify purpose (which is set
6615 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6617 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6618 this is because it will reject chains with invalid extensions whereas
6619 every previous version of OpenSSL and SSLeay made no checks at all.
6621 Trust code: checks the root CA for the relevant trust settings. Trust
6622 settings have an initial value consistent with the verify purpose: e.g.
6623 if the verify purpose is for SSL client use it expects the CA to be
6624 trusted for SSL client use. However the default value can be changed to
6625 permit custom trust settings: one example of this would be to only trust
6626 certificates from a specific "secure" set of CAs.
6628 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6629 which should be used for version portability: especially since the
6630 verify structure is likely to change more often now.
6632 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6633 to set them. If not set then assume SSL clients will verify SSL servers
6636 Two new options to the verify program: -untrusted allows a set of
6637 untrusted certificates to be passed in and -purpose which sets the
6638 intended purpose of the certificate. If a purpose is set then the
6639 new chain verify code is used to check extension consistency.
6642 *) Support for the authority information access extension.
6645 *) Modify RSA and DSA PEM read routines to transparently handle
6646 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6647 public keys in a format compatible with certificate
6648 SubjectPublicKeyInfo structures. Unfortunately there were already
6649 functions called *_PublicKey_* which used various odd formats so
6650 these are retained for compatibility: however the DSA variants were
6651 never in a public release so they have been deleted. Changed dsa/rsa
6652 utilities to handle the new format: note no releases ever handled public
6653 keys so we should be OK.
6655 The primary motivation for this change is to avoid the same fiasco
6656 that dogs private keys: there are several incompatible private key
6657 formats some of which are standard and some OpenSSL specific and
6658 require various evil hacks to allow partial transparent handling and
6659 even then it doesn't work with DER formats. Given the option anything
6660 other than PKCS#8 should be dumped: but the other formats have to
6661 stay in the name of compatibility.
6663 With public keys and the benefit of hindsight one standard format
6664 is used which works with EVP_PKEY, RSA or DSA structures: though
6665 it clearly returns an error if you try to read the wrong kind of key.
6667 Added a -pubkey option to the 'x509' utility to output the public key.
6668 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6669 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6670 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6671 that do the same as the EVP_PKEY_assign_*() except they up the
6672 reference count of the added key (they don't "swallow" the
6676 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6677 CRLs would fail if the file contained no certificates or no CRLs:
6678 added a new function to read in both types and return the number
6679 read: this means that if none are read it will be an error. The
6680 DER versions of the certificate and CRL reader would always fail
6681 because it isn't possible to mix certificates and CRLs in DER format
6682 without choking one or the other routine. Changed this to just read
6683 a certificate: this is the best we can do. Also modified the code
6684 in apps/verify.c to take notice of return codes: it was previously
6685 attempting to read in certificates from NULL pointers and ignoring
6686 any errors: this is one reason why the cert and CRL reader seemed
6687 to work. It doesn't check return codes from the default certificate
6688 routines: these may well fail if the certificates aren't installed.
6691 *) Code to support otherName option in GeneralName.
6694 *) First update to verify code. Change the verify utility
6695 so it warns if it is passed a self signed certificate:
6696 for consistency with the normal behaviour. X509_verify
6697 has been modified to it will now verify a self signed
6698 certificate if *exactly* the same certificate appears
6699 in the store: it was previously impossible to trust a
6700 single self signed certificate. This means that:
6701 openssl verify ss.pem
6702 now gives a warning about a self signed certificate but
6703 openssl verify -CAfile ss.pem ss.pem
6707 *) For servers, store verify_result in SSL_SESSION data structure
6708 (and add it to external session representation).
6709 This is needed when client certificate verifications fails,
6710 but an application-provided verification callback (set by
6711 SSL_CTX_set_cert_verify_callback) allows accepting the session
6712 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6713 but returns 1): When the session is reused, we have to set
6714 ssl->verify_result to the appropriate error code to avoid
6716 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6718 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6719 case in PKCS7_dataInit() where the signed PKCS7 structure
6720 didn't contain any existing data because it was being created.
6721 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6723 *) Add a salt to the key derivation routines in enc.c. This
6724 forms the first 8 bytes of the encrypted file. Also add a
6725 -S option to allow a salt to be input on the command line.
6728 *) New function X509_cmp(). Oddly enough there wasn't a function
6729 to compare two certificates. We do this by working out the SHA1
6730 hash and comparing that. X509_cmp() will be needed by the trust
6734 *) SSL_get1_session() is like SSL_get_session(), but increments
6735 the reference count in the SSL_SESSION returned.
6736 [Geoff Thorpe <geoff@eu.c2.net>]
6738 *) Fix for 'req': it was adding a null to request attributes.
6739 Also change the X509_LOOKUP and X509_INFO code to handle
6740 certificate auxiliary information.
6743 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6747 *) Add the possibility to add extra information to the memory leak
6748 detecting output, to form tracebacks, showing from where each
6749 allocation was originated: CRYPTO_push_info("constant string") adds
6750 the string plus current file name and line number to a per-thread
6751 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6752 is like calling CYRPTO_pop_info() until the stack is empty.
6753 Also updated memory leak detection code to be multi-thread-safe.
6756 *) Add options -text and -noout to pkcs7 utility and delete the
6757 encryption options which never did anything. Update docs.
6760 *) Add options to some of the utilities to allow the pass phrase
6761 to be included on either the command line (not recommended on
6762 OSes like Unix) or read from the environment. Update the
6763 manpages and fix a few bugs.
6766 *) Add a few manpages for some of the openssl commands.
6769 *) Fix the -revoke option in ca. It was freeing up memory twice,
6770 leaking and not finding already revoked certificates.
6773 *) Extensive changes to support certificate auxiliary information.
6774 This involves the use of X509_CERT_AUX structure and X509_AUX
6775 functions. An X509_AUX function such as PEM_read_X509_AUX()
6776 can still read in a certificate file in the usual way but it
6777 will also read in any additional "auxiliary information". By
6778 doing things this way a fair degree of compatibility can be
6779 retained: existing certificates can have this information added
6780 using the new 'x509' options.
6782 Current auxiliary information includes an "alias" and some trust
6783 settings. The trust settings will ultimately be used in enhanced
6784 certificate chain verification routines: currently a certificate
6785 can only be trusted if it is self signed and then it is trusted
6789 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6790 The problem was that one of the replacement routines had not been working
6791 since SSLeay releases. For now the offending routine has been replaced
6792 with non-optimised assembler. Even so, this now gives around 95%
6793 performance improvement for 1024 bit RSA signs.
6796 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6797 handling. Most clients have the effective key size in bits equal to
6798 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6799 A few however don't do this and instead use the size of the decrypted key
6800 to determine the RC2 key length and the AlgorithmIdentifier to determine
6801 the effective key length. In this case the effective key length can still
6802 be 40 bits but the key length can be 168 bits for example. This is fixed
6803 by manually forcing an RC2 key into the EVP_PKEY structure because the
6804 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6805 the key length and effective key length are equal.
6808 *) Add a bunch of functions that should simplify the creation of
6809 X509_NAME structures. Now you should be able to do:
6810 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6811 and have it automatically work out the correct field type and fill in
6812 the structures. The more adventurous can try:
6813 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6814 and it will (hopefully) work out the correct multibyte encoding.
6817 *) Change the 'req' utility to use the new field handling and multibyte
6818 copy routines. Before the DN field creation was handled in an ad hoc
6819 way in req, ca, and x509 which was rather broken and didn't support
6820 BMPStrings or UTF8Strings. Since some software doesn't implement
6821 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6822 using the dirstring_type option. See the new comment in the default
6823 openssl.cnf for more info.
6826 *) Make crypto/rand/md_rand.c more robust:
6827 - Assure unique random numbers after fork().
6828 - Make sure that concurrent threads access the global counter and
6829 md serializably so that we never lose entropy in them
6830 or use exactly the same state in multiple threads.
6831 Access to the large state is not always serializable because
6832 the additional locking could be a performance killer, and
6833 md should be large enough anyway.
6836 *) New file apps/app_rand.c with commonly needed functionality
6837 for handling the random seed file.
6839 Use the random seed file in some applications that previously did not:
6841 dsaparam -genkey (which also ignored its '-rand' option),
6844 x509 (when signing).
6845 Except on systems with /dev/urandom, it is crucial to have a random
6846 seed file at least for key creation, DSA signing, and for DH exchanges;
6847 for RSA signatures we could do without one.
6849 gendh and gendsa (unlike genrsa) used to read only the first byte
6850 of each file listed in the '-rand' option. The function as previously
6851 found in genrsa is now in app_rand.c and is used by all programs
6852 that support '-rand'.
6855 *) In RAND_write_file, use mode 0600 for creating files;
6856 don't just chmod when it may be too late.
6859 *) Report an error from X509_STORE_load_locations
6860 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6863 *) New function ASN1_mbstring_copy() this copies a string in either
6864 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6865 into an ASN1_STRING type. A mask of permissible types is passed
6866 and it chooses the "minimal" type to use or an error if not type
6870 *) Add function equivalents to the various macros in asn1.h. The old
6871 macros are retained with an M_ prefix. Code inside the library can
6872 use the M_ macros. External code (including the openssl utility)
6873 should *NOT* in order to be "shared library friendly".
6876 *) Add various functions that can check a certificate's extensions
6877 to see if it usable for various purposes such as SSL client,
6878 server or S/MIME and CAs of these types. This is currently
6879 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6880 verification. Also added a -purpose flag to x509 utility to
6881 print out all the purposes.
6884 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6888 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6889 for, obtain and decode and extension and obtain its critical flag.
6890 This allows all the necessary extension code to be handled in a
6891 single function call.
6894 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6895 platforms. See crypto/rc4/rc4_enc.c for further details.
6898 *) New -noout option to asn1parse. This causes no output to be produced
6899 its main use is when combined with -strparse and -out to extract data
6900 from a file (which may not be in ASN.1 format).
6903 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6904 when producing the local key id.
6905 [Richard Levitte <levitte@stacken.kth.se>]
6907 *) New option -dhparam in s_server. This allows a DH parameter file to be
6908 stated explicitly. If it is not stated then it tries the first server
6909 certificate file. The previous behaviour hard coded the filename
6913 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6914 a public key to be input or output. For example:
6915 openssl rsa -in key.pem -pubout -out pubkey.pem
6916 Also added necessary DSA public key functions to handle this.
6919 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6920 in the message. This was handled by allowing
6921 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6922 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6924 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6925 to the end of the strings whereas this didn't. This would cause problems
6926 if strings read with d2i_ASN1_bytes() were later modified.
6927 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6929 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6930 data and it contains EOF it will end up returning an error. This is
6931 caused by input 46 bytes long. The cause is due to the way base64
6932 BIOs find the start of base64 encoded data. They do this by trying a
6933 trial decode on each line until they find one that works. When they
6934 do a flag is set and it starts again knowing it can pass all the
6935 data directly through the decoder. Unfortunately it doesn't reset
6936 the context it uses. This means that if EOF is reached an attempt
6937 is made to pass two EOFs through the context and this causes the
6938 resulting error. This can also cause other problems as well. As is
6939 usual with these problems it takes *ages* to find and the fix is
6940 trivial: move one line.
6941 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6943 *) Ugly workaround to get s_client and s_server working under Windows. The
6944 old code wouldn't work because it needed to select() on sockets and the
6945 tty (for keypresses and to see if data could be written). Win32 only
6946 supports select() on sockets so we select() with a 1s timeout on the
6947 sockets and then see if any characters are waiting to be read, if none
6948 are present then we retry, we also assume we can always write data to
6949 the tty. This isn't nice because the code then blocks until we've
6950 received a complete line of data and it is effectively polling the
6951 keyboard at 1s intervals: however it's quite a bit better than not
6952 working at all :-) A dedicated Windows application might handle this
6953 with an event loop for example.
6956 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6957 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6958 will be called when RSA_sign() and RSA_verify() are used. This is useful
6959 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6960 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6961 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6962 This necessitated the support of an extra signature type NID_md5_sha1
6963 for SSL signatures and modifications to the SSL library to use it instead
6964 of calling RSA_public_decrypt() and RSA_private_encrypt().
6967 *) Add new -verify -CAfile and -CApath options to the crl program, these
6968 will lookup a CRL issuers certificate and verify the signature in a
6969 similar way to the verify program. Tidy up the crl program so it
6970 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6971 less strict. It will now permit CRL extensions even if it is not
6972 a V2 CRL: this will allow it to tolerate some broken CRLs.
6975 *) Initialize all non-automatic variables each time one of the openssl
6976 sub-programs is started (this is necessary as they may be started
6977 multiple times from the "OpenSSL>" prompt).
6978 [Lennart Bang, Bodo Moeller]
6980 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6981 removing all other RSA functionality (this is what NO_RSA does). This
6982 is so (for example) those in the US can disable those operations covered
6983 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6987 *) Non-copying interface to BIO pairs.
6988 (still largely untested)
6991 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6992 ASCII string. This was handled independently in various places before.
6995 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6996 UTF8 strings a character at a time.
6999 *) Use client_version from client hello to select the protocol
7000 (s23_srvr.c) and for RSA client key exchange verification
7001 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7004 *) Add various utility functions to handle SPKACs, these were previously
7005 handled by poking round in the structure internals. Added new function
7006 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7007 print, verify and generate SPKACs. Based on an original idea from
7008 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7011 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7014 *) Allow the config file extension section to be overwritten on the
7015 command line. Based on an original idea from Massimiliano Pala
7016 <madwolf@comune.modena.it>. The new option is called -extensions
7017 and can be applied to ca, req and x509. Also -reqexts to override
7018 the request extensions in req and -crlexts to override the crl extensions
7022 *) Add new feature to the SPKAC handling in ca. Now you can include
7023 the same field multiple times by preceding it by "XXXX." for example:
7026 this is the same syntax as used in the req config file.
7029 *) Allow certificate extensions to be added to certificate requests. These
7030 are specified in a 'req_extensions' option of the req section of the
7031 config file. They can be printed out with the -text option to req but
7032 are otherwise ignored at present.
7035 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7036 data read consists of only the final block it would not decrypted because
7037 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7038 A misplaced 'break' also meant the decrypted final block might not be
7039 copied until the next read.
7042 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7043 a few extra parameters to the DH structure: these will be useful if
7044 for example we want the value of 'q' or implement X9.42 DH.
7047 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7048 provides hooks that allow the default DSA functions or functions on a
7049 "per key" basis to be replaced. This allows hardware acceleration and
7050 hardware key storage to be handled without major modification to the
7051 library. Also added low level modexp hooks and CRYPTO_EX structure and
7052 associated functions.
7055 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7056 as "read only": it can't be written to and the buffer it points to will
7057 not be freed. Reading from a read only BIO is much more efficient than
7058 a normal memory BIO. This was added because there are several times when
7059 an area of memory needs to be read from a BIO. The previous method was
7060 to create a memory BIO and write the data to it, this results in two
7061 copies of the data and an O(n^2) reading algorithm. There is a new
7062 function BIO_new_mem_buf() which creates a read only memory BIO from
7063 an area of memory. Also modified the PKCS#7 routines to use read only
7067 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7068 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7069 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7070 but a retry condition occured while trying to read the rest.
7073 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7074 NID_pkcs7_encrypted by default: this was wrong since this should almost
7075 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7076 the encrypted data type: this is a more sensible place to put it and it
7077 allows the PKCS#12 code to be tidied up that duplicated this
7081 *) Changed obj_dat.pl script so it takes its input and output files on
7082 the command line. This should avoid shell escape redirection problems
7086 *) Initial support for certificate extension requests, these are included
7087 in things like Xenroll certificate requests. Included functions to allow
7088 extensions to be obtained and added.
7091 *) -crlf option to s_client and s_server for sending newlines as
7092 CRLF (as required by many protocols).
7095 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7097 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7098 [Ralf S. Engelschall]
7100 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7101 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7103 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7107 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7108 DH parameters/keys (q is lost during that conversion, but the resulting
7109 DH parameters contain its length).
7111 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7112 much faster than DH_generate_parameters (which creates parameters
7113 where p = 2*q + 1), and also the smaller q makes DH computations
7114 much more efficient (160-bit exponentiation instead of 1024-bit
7115 exponentiation); so this provides a convenient way to support DHE
7116 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7117 utter importance to use
7118 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7120 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7121 when such DH parameters are used, because otherwise small subgroup
7122 attacks may become possible!
7125 *) Avoid memory leak in i2d_DHparams.
7128 *) Allow the -k option to be used more than once in the enc program:
7129 this allows the same encrypted message to be read by multiple recipients.
7132 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7133 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7134 it will always use the numerical form of the OID, even if it has a short
7138 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7139 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7140 otherwise bn_mod_exp was called. In the case of hardware keys for example
7141 no private key components need be present and it might store extra data
7142 in the RSA structure, which cannot be accessed from bn_mod_exp.
7143 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7144 private key operations.
7147 *) Added support for SPARC Linux.
7150 *) pem_password_cb function type incompatibly changed from
7151 typedef int pem_password_cb(char *buf, int size, int rwflag);
7153 ....(char *buf, int size, int rwflag, void *userdata);
7154 so that applications can pass data to their callbacks:
7155 The PEM[_ASN1]_{read,write}... functions and macros now take an
7156 additional void * argument, which is just handed through whenever
7157 the password callback is called.
7158 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7160 New function SSL_CTX_set_default_passwd_cb_userdata.
7162 Compatibility note: As many C implementations push function arguments
7163 onto the stack in reverse order, the new library version is likely to
7164 interoperate with programs that have been compiled with the old
7165 pem_password_cb definition (PEM_whatever takes some data that
7166 happens to be on the stack as its last argument, and the callback
7167 just ignores this garbage); but there is no guarantee whatsoever that
7170 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7171 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7172 problems not only on Windows, but also on some Unix platforms.
7173 To avoid problematic command lines, these definitions are now in an
7174 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7175 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7178 *) MIPS III/IV assembler module is reimplemented.
7181 *) More DES library cleanups: remove references to srand/rand and
7182 delete an unused file.
7185 *) Add support for the the free Netwide assembler (NASM) under Win32,
7186 since not many people have MASM (ml) and it can be hard to obtain.
7187 This is currently experimental but it seems to work OK and pass all
7188 the tests. Check out INSTALL.W32 for info.
7191 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7192 without temporary keys kept an extra copy of the server key,
7193 and connections with temporary keys did not free everything in case
7197 *) New function RSA_check_key and new openssl rsa option -check
7198 for verifying the consistency of RSA keys.
7199 [Ulf Moeller, Bodo Moeller]
7201 *) Various changes to make Win32 compile work:
7202 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7203 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7204 comparison" warnings.
7205 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7208 *) Add a debugging option to PKCS#5 v2 key generation function: when
7209 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7210 derived keys are printed to stderr.
7213 *) Copy the flags in ASN1_STRING_dup().
7214 [Roman E. Pavlov <pre@mo.msk.ru>]
7216 *) The x509 application mishandled signing requests containing DSA
7217 keys when the signing key was also DSA and the parameters didn't match.
7219 It was supposed to omit the parameters when they matched the signing key:
7220 the verifying software was then supposed to automatically use the CA's
7221 parameters if they were absent from the end user certificate.
7223 Omitting parameters is no longer recommended. The test was also
7224 the wrong way round! This was probably due to unusual behaviour in
7225 EVP_cmp_parameters() which returns 1 if the parameters match.
7226 This meant that parameters were omitted when they *didn't* match and
7227 the certificate was useless. Certificates signed with 'ca' didn't have
7229 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7231 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7232 The interface is as follows:
7233 Applications can use
7234 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7235 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7236 "off" is now the default.
7237 The library internally uses
7238 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7239 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7240 to disable memory-checking temporarily.
7242 Some inconsistent states that previously were possible (and were
7243 even the default) are now avoided.
7245 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7246 with each memory chunk allocated; this is occasionally more helpful
7247 than just having a counter.
7249 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7251 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7255 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7256 which largely parallels "options", but is for changing API behaviour,
7257 whereas "options" are about protocol behaviour.
7258 Initial "mode" flags are:
7260 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7261 a single record has been written.
7262 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7263 retries use the same buffer location.
7264 (But all of the contents must be
7268 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7271 *) Fix problems with no-hmac etc.
7272 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7274 *) New functions RSA_get_default_method(), RSA_set_method() and
7275 RSA_get_method(). These allows replacement of RSA_METHODs without having
7276 to mess around with the internals of an RSA structure.
7279 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7280 Also really enable memory leak checks in openssl.c and in some
7282 [Chad C. Mulligan, Bodo Moeller]
7284 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7285 up the length of negative integers. This has now been simplified to just
7286 store the length when it is first determined and use it later, rather
7287 than trying to keep track of where data is copied and updating it to
7289 [Steve Henson, reported by Brien Wheeler
7290 <bwheeler@authentica-security.com>]
7292 *) Add a new function PKCS7_signatureVerify. This allows the verification
7293 of a PKCS#7 signature but with the signing certificate passed to the
7294 function itself. This contrasts with PKCS7_dataVerify which assumes the
7295 certificate is present in the PKCS#7 structure. This isn't always the
7296 case: certificates can be omitted from a PKCS#7 structure and be
7297 distributed by "out of band" means (such as a certificate database).
7300 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7301 function prototypes in pem.h, also change util/mkdef.pl to add the
7302 necessary function names.
7305 *) mk1mf.pl (used by Windows builds) did not properly read the
7306 options set by Configure in the top level Makefile, and Configure
7307 was not even able to write more than one option correctly.
7308 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7311 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7312 file to be loaded from a BIO or FILE pointer. The BIO version will
7313 for example allow memory BIOs to contain config info.
7316 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7317 Whoever hopes to achieve shared-library compatibility across versions
7318 must use this, not the compile-time macro.
7319 (Exercise 0.9.4: Which is the minimum library version required by
7321 Note: All this applies only to multi-threaded programs, others don't
7325 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7326 through a BIO pair triggered the default case, i.e.
7327 SSLerr(...,SSL_R_UNKNOWN_STATE).
7330 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7331 can use the SSL library even if none of the specific BIOs is
7335 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7336 for the encoded length.
7337 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7339 *) Add initial documentation of the X509V3 functions.
7342 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7343 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7344 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7345 secure PKCS#8 private key format with a high iteration count.
7348 *) Fix determination of Perl interpreter: A perl or perl5
7349 _directory_ in $PATH was also accepted as the interpreter.
7350 [Ralf S. Engelschall]
7352 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7353 wrong with it but it was very old and did things like calling
7354 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7358 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7359 to use the new extension code.
7362 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7363 with macros. This should make it easier to change their form, add extra
7364 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7368 *) Add to configuration table a new entry that can specify an alternative
7369 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7370 according to Mark Crispin <MRC@Panda.COM>.
7374 *) DES CBC did not update the IV. Weird.
7377 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7378 Changing the behaviour of the former might break existing programs --
7379 where IV updating is needed, des_ncbc_encrypt can be used.
7382 *) When bntest is run from "make test" it drives bc to check its
7383 calculations, as well as internally checking them. If an internal check
7384 fails, it needs to cause bc to give a non-zero result or make test carries
7385 on without noticing the failure. Fixed.
7388 *) DES library cleanups.
7391 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7392 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7393 ciphers. NOTE: although the key derivation function has been verified
7394 against some published test vectors it has not been extensively tested
7395 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7399 *) Instead of "mkdir -p", which is not fully portable, use new
7400 Perl script "util/mkdir-p.pl".
7403 *) Rewrite the way password based encryption (PBE) is handled. It used to
7404 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7405 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7406 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7407 the 'parameter' field of the AlgorithmIdentifier is passed to the
7408 underlying key generation function so it must do its own ASN1 parsing.
7409 This has also changed the EVP_PBE_CipherInit() function which now has a
7410 'parameter' argument instead of literal salt and iteration count values
7411 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7414 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7415 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7416 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7417 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7418 value was just used as a "magic string" and not used directly its
7419 value doesn't matter.
7422 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7426 *) "linux-sparc64" configuration (ultrapenguin).
7427 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7428 "linux-sparc" configuration.
7429 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7431 *) config now generates no-xxx options for missing ciphers.
7434 *) Support the EBCDIC character set (work in progress).
7435 File ebcdic.c not yet included because it has a different license.
7436 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7438 *) Support BS2000/OSD-POSIX.
7439 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7441 *) Make callbacks for key generation use void * instead of char *.
7444 *) Make S/MIME samples compile (not yet tested).
7447 *) Additional typesafe stacks.
7450 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7454 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7456 *) New configuration variant "sco5-gcc".
7458 *) Updated some demos.
7459 [Sean O Riordain, Wade Scholine]
7461 *) Add missing BIO_free at exit of pkcs12 application.
7464 *) Fix memory leak in conf.c.
7467 *) Updates for Win32 to assembler version of MD5.
7470 *) Set #! path to perl in apps/der_chop to where we found it
7471 instead of using a fixed path.
7474 *) SHA library changes for irix64-mips4-cc.
7477 *) Improvements for VMS support.
7481 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7483 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7484 This also avoids the problems with SC4.2 and unpatched SC5.
7485 [Andy Polyakov <appro@fy.chalmers.se>]
7487 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7488 These are required because of the typesafe stack would otherwise break
7489 existing code. If old code used a structure member which used to be STACK
7490 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7491 sk_num or sk_value it would produce an error because the num, data members
7492 are not present in STACK_OF. Now it just produces a warning. sk_set
7493 replaces the old method of assigning a value to sk_value
7494 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7495 that does this will no longer work (and should use sk_set instead) but
7496 this could be regarded as a "questionable" behaviour anyway.
7499 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7500 correctly handle encrypted S/MIME data.
7503 *) Change type of various DES function arguments from des_cblock
7504 (which means, in function argument declarations, pointer to char)
7505 to des_cblock * (meaning pointer to array with 8 char elements),
7506 which allows the compiler to do more typechecking; it was like
7507 that back in SSLeay, but with lots of ugly casts.
7509 Introduce new type const_des_cblock.
7512 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7513 problems: find RecipientInfo structure that matches recipient certificate
7514 and initialise the ASN1 structures properly based on passed cipher.
7517 *) Belatedly make the BN tests actually check the results.
7520 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7521 to and from BNs: it was completely broken. New compilation option
7522 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7523 key elements as negative integers.
7526 *) Reorganize and speed up MD5.
7527 [Andy Polyakov <appro@fy.chalmers.se>]
7530 [Richard Levitte <richard@levitte.org>]
7532 *) New option -out to asn1parse to allow the parsed structure to be
7533 output to a file. This is most useful when combined with the -strparse
7534 option to examine the output of things like OCTET STRINGS.
7537 *) Make SSL library a little more fool-proof by not requiring any longer
7538 that SSL_set_{accept,connect}_state be called before
7539 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7540 in many applications because usually everything *appeared* to work as
7541 intended anyway -- now it really works as intended).
7544 *) Move openssl.cnf out of lib/.
7547 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7548 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7549 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7550 [Ralf S. Engelschall]
7552 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7553 handle PKCS#7 enveloped data properly.
7554 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7556 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7557 copying pointers. The cert_st handling is changed by this in
7558 various ways (and thus what used to be known as ctx->default_cert
7559 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7560 any longer when s->cert does not give us what we need).
7561 ssl_cert_instantiate becomes obsolete by this change.
7562 As soon as we've got the new code right (possibly it already is?),
7563 we have solved a couple of bugs of the earlier code where s->cert
7564 was used as if it could not have been shared with other SSL structures.
7566 Note that using the SSL API in certain dirty ways now will result
7567 in different behaviour than observed with earlier library versions:
7568 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7569 does not influence s as it used to.
7571 In order to clean up things more thoroughly, inside SSL_SESSION
7572 we don't use CERT any longer, but a new structure SESS_CERT
7573 that holds per-session data (if available); currently, this is
7574 the peer's certificate chain and, for clients, the server's certificate
7575 and temporary key. CERT holds only those values that can have
7576 meaningful defaults in an SSL_CTX.
7579 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7580 from the internal representation. Various PKCS#7 fixes: remove some
7581 evil casts and set the enc_dig_alg field properly based on the signing
7585 *) Allow PKCS#12 password to be set from the command line or the
7586 environment. Let 'ca' get its config file name from the environment
7587 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7591 *) Allow certificate policies extension to use an IA5STRING for the
7592 organization field. This is contrary to the PKIX definition but
7593 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7597 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7598 without disallowing inline assembler and the like for non-pedantic builds.
7601 *) Support Borland C++ builder.
7602 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7607 *) SHA-1 cleanups and performance enhancements.
7608 [Andy Polyakov <appro@fy.chalmers.se>]
7610 *) Sparc v8plus assembler for the bignum library.
7611 [Andy Polyakov <appro@fy.chalmers.se>]
7613 *) Accept any -xxx and +xxx compiler options in Configure.
7616 *) Update HPUX configuration.
7619 *) Add missing sk_<type>_unshift() function to safestack.h
7620 [Ralf S. Engelschall]
7622 *) New function SSL_CTX_use_certificate_chain_file that sets the
7623 "extra_cert"s in addition to the certificate. (This makes sense
7624 only for "PEM" format files, as chains as a whole are not
7628 *) Support verify_depth from the SSL API.
7629 x509_vfy.c had what can be considered an off-by-one-error:
7630 Its depth (which was not part of the external interface)
7631 was actually counting the number of certificates in a chain;
7632 now it really counts the depth.
7635 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7636 instead of X509err, which often resulted in confusing error
7637 messages since the error codes are not globally unique
7638 (e.g. an alleged error in ssl3_accept when a certificate
7639 didn't match the private key).
7641 *) New function SSL_CTX_set_session_id_context that allows to set a default
7642 value (so that you don't need SSL_set_session_id_context for each
7643 connection using the SSL_CTX).
7646 *) OAEP decoding bug fix.
7649 *) Support INSTALL_PREFIX for package builders, as proposed by
7653 *) New Configure options "threads" and "no-threads". For systems
7654 where the proper compiler options are known (currently Solaris
7655 and Linux), "threads" is the default.
7658 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7661 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7662 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7663 such as /usr/local/bin.
7666 *) "make linux-shared" to build shared libraries.
7667 [Niels Poppe <niels@netbox.org>]
7669 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7672 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7673 extension adding in x509 utility.
7676 *) Remove NOPROTO sections and error code comments.
7679 *) Partial rewrite of the DEF file generator to now parse the ANSI
7683 *) New Configure options --prefix=DIR and --openssldir=DIR.
7686 *) Complete rewrite of the error code script(s). It is all now handled
7687 by one script at the top level which handles error code gathering,
7688 header rewriting and C source file generation. It should be much better
7689 than the old method: it now uses a modified version of Ulf's parser to
7690 read the ANSI prototypes in all header files (thus the old K&R definitions
7691 aren't needed for error creation any more) and do a better job of
7692 translating function codes into names. The old 'ASN1 error code imbedded
7693 in a comment' is no longer necessary and it doesn't use .err files which
7694 have now been deleted. Also the error code call doesn't have to appear all
7695 on one line (which resulted in some large lines...).
7698 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7701 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7702 0 (which usually indicates a closed connection), but continue reading.
7705 *) Fix some race conditions.
7708 *) Add support for CRL distribution points extension. Add Certificate
7709 Policies and CRL distribution points documentation.
7712 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7715 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7716 8 of keying material. Merlin has also confirmed interop with this fix
7717 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7718 [Merlin Hughes <merlin@baltimore.ie>]
7720 *) Fix lots of warnings.
7721 [Richard Levitte <levitte@stacken.kth.se>]
7723 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7724 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7725 [Richard Levitte <levitte@stacken.kth.se>]
7727 *) Fix problems with sizeof(long) == 8.
7728 [Andy Polyakov <appro@fy.chalmers.se>]
7730 *) Change functions to ANSI C.
7733 *) Fix typos in error codes.
7734 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7736 *) Remove defunct assembler files from Configure.
7739 *) SPARC v8 assembler BIGNUM implementation.
7740 [Andy Polyakov <appro@fy.chalmers.se>]
7742 *) Support for Certificate Policies extension: both print and set.
7743 Various additions to support the r2i method this uses.
7746 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7747 return a const string when you are expecting an allocated buffer.
7750 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7751 types DirectoryString and DisplayText.
7754 *) Add code to allow r2i extensions to access the configuration database,
7755 add an LHASH database driver and add several ctx helper functions.
7758 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7759 fail when they extended the size of a BIGNUM.
7762 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7763 support typesafe stack.
7766 *) Fix typo in SSL_[gs]et_options().
7767 [Nils Frostberg <nils@medcom.se>]
7769 *) Delete various functions and files that belonged to the (now obsolete)
7770 old X509V3 handling code.
7773 *) New Configure option "rsaref".
7776 *) Don't auto-generate pem.h.
7779 *) Introduce type-safe ASN.1 SETs.
7782 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7783 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7785 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7786 that links with OpenSSL (well at least cause lots of warnings), but fear
7787 not: the conversion is trivial, and it eliminates loads of evil casts. A
7788 few STACKed things have been converted already. Feel free to convert more.
7789 In the fullness of time, I'll do away with the STACK type altogether.
7792 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7793 specified in <certfile> by updating the entry in the index.txt file.
7794 This way one no longer has to edit the index.txt file manually for
7795 revoking a certificate. The -revoke option does the gory details now.
7796 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7798 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7799 `-text' option at all and this way the `-noout -text' combination was
7800 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7801 [Ralf S. Engelschall]
7803 *) Make sure a corresponding plain text error message exists for the
7804 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7805 verify callback function determined that a certificate was revoked.
7806 [Ralf S. Engelschall]
7808 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7809 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7810 all available cipers including rc5, which was forgotten until now.
7811 In order to let the testing shell script know which algorithms
7812 are available, a new (up to now undocumented) command
7813 "openssl list-cipher-commands" is used.
7816 *) Bugfix: s_client occasionally would sleep in select() when
7817 it should have checked SSL_pending() first.
7820 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7821 the raw DSA values prior to ASN.1 encoding.
7824 *) Tweaks to Configure
7825 [Niels Poppe <niels@netbox.org>]
7827 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7831 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7834 *) New config option to avoid instructions that are illegal on the 80386.
7835 The default code is faster, but requires at least a 486.
7838 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7839 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7840 same as SSL2_VERSION anyway.
7843 *) New "-showcerts" option for s_client.
7846 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7847 application. Various cleanups and fixes.
7850 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7851 modify error routines to work internally. Add error codes and PBE init
7852 to library startup routines.
7855 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7856 packing functions to asn1 and evp. Changed function names and error
7857 codes along the way.
7860 *) PKCS12 integration: and so it begins... First of several patches to
7861 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7862 objects to objects.h
7865 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7866 and display support for Thawte strong extranet extension.
7869 *) Add LinuxPPC support.
7870 [Jeff Dubrule <igor@pobox.org>]
7872 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7873 bn_div_words in alpha.s.
7874 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7876 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7877 OAEP isn't supported when OpenSSL is built with RSAref.
7878 [Ulf Moeller <ulf@fitug.de>]
7880 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7881 so they no longer are missing under -DNOPROTO.
7882 [Soren S. Jorvang <soren@t.dk>]
7885 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7887 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7888 doesn't work when the session is reused. Coming soon!
7891 *) Fix a security hole, that allows sessions to be reused in the wrong
7892 context thus bypassing client cert protection! All software that uses
7893 client certs and session caches in multiple contexts NEEDS PATCHING to
7894 allow session reuse! A fuller solution is in the works.
7895 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7897 *) Some more source tree cleanups (removed obsolete files
7898 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7899 permission on "config" script to be executable) and a fix for the INSTALL
7901 [Ulf Moeller <ulf@fitug.de>]
7903 *) Remove some legacy and erroneous uses of malloc, free instead of
7905 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7907 *) Make rsa_oaep_test return non-zero on error.
7908 [Ulf Moeller <ulf@fitug.de>]
7910 *) Add support for native Solaris shared libraries. Configure
7911 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7912 if someone would make that last step automatic.
7913 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7915 *) ctx_size was not built with the right compiler during "make links". Fixed.
7918 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7919 except NULL ciphers". This means the default cipher list will no longer
7920 enable NULL ciphers. They need to be specifically enabled e.g. with
7921 the string "DEFAULT:eNULL".
7924 *) Fix to RSA private encryption routines: if p < q then it would
7925 occasionally produce an invalid result. This will only happen with
7926 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7929 *) Be less restrictive and allow also `perl util/perlpath.pl
7930 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7931 because this way one can also use an interpreter named `perl5' (which is
7932 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7933 installed as `perl').
7934 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7936 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7937 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7939 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7940 advapi32.lib to Win32 build and change the pem test comparision
7941 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7942 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7943 and crypto/des/ede_cbcm_enc.c.
7946 *) DES quad checksum was broken on big-endian architectures. Fixed.
7949 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7950 Win32 test batch file so it (might) work again. The Win32 test batch file
7951 is horrible: I feel ill....
7954 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7955 in e_os.h. Audit of header files to check ANSI and non ANSI
7956 sections: 10 functions were absent from non ANSI section and not exported
7957 from Windows DLLs. Fixed up libeay.num for new functions.
7960 *) Make `openssl version' output lines consistent.
7961 [Ralf S. Engelschall]
7963 *) Fix Win32 symbol export lists for BIO functions: Added
7964 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7965 to ms/libeay{16,32}.def.
7966 [Ralf S. Engelschall]
7968 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7969 fine under Unix and passes some trivial tests I've now added. But the
7970 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7971 added to make sure no one expects that this stuff really works in the
7972 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7973 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7975 [Ralf S. Engelschall]
7977 *) Fix the generation of two part addresses in perl.
7978 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7980 *) Add config entry for Linux on MIPS.
7981 [John Tobey <jtobey@channel1.com>]
7983 *) Make links whenever Configure is run, unless we are on Windoze.
7986 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7987 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7991 *) Add a useful kludge to allow package maintainers to specify compiler and
7992 other platforms details on the command line without having to patch the
7993 Configure script everytime: One now can use ``perl Configure
7994 <id>:<details>'', i.e. platform ids are allowed to have details appended
7995 to them (seperated by colons). This is treated as there would be a static
7996 pre-configured entry in Configure's %table under key <id> with value
7997 <details> and ``perl Configure <id>'' is called. So, when you want to
7998 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7999 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8000 now, which overrides the FreeBSD-elf entry on-the-fly.
8001 [Ralf S. Engelschall]
8003 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8006 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8007 on the `perl Configure ...' command line. This way one can compile
8008 OpenSSL libraries with Position Independent Code (PIC) which is needed
8009 for linking it into DSOs.
8010 [Ralf S. Engelschall]
8012 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8016 *) Cleaned up the LICENSE document: The official contact for any license
8017 questions now is the OpenSSL core team under openssl-core@openssl.org.
8018 And add a paragraph about the dual-license situation to make sure people
8019 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8020 to the OpenSSL toolkit.
8021 [Ralf S. Engelschall]
8023 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8024 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8025 Additonally cleaned up the `make links' target: Remove unnecessary
8026 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8027 to speed processing and no longer clutter the display with confusing
8028 stuff. Instead only the actually done links are displayed.
8029 [Ralf S. Engelschall]
8031 *) Permit null encryption ciphersuites, used for authentication only. It used
8032 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8033 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8037 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8038 signed attributes when verifying signatures (this would break them),
8039 the detached data encoding was wrong and public keys obtained using
8040 X509_get_pubkey() weren't freed.
8043 *) Add text documentation for the BUFFER functions. Also added a work around
8044 to a Win95 console bug. This was triggered by the password read stuff: the
8045 last character typed gets carried over to the next fread(). If you were
8046 generating a new cert request using 'req' for example then the last
8047 character of the passphrase would be CR which would then enter the first
8051 *) Added the new `Includes OpenSSL Cryptography Software' button as
8052 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8053 button and can be used by applications based on OpenSSL to show the
8054 relationship to the OpenSSL project.
8055 [Ralf S. Engelschall]
8057 *) Remove confusing variables in function signatures in files
8058 ssl/ssl_lib.c and ssl/ssl.h.
8059 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8061 *) Don't install bss_file.c under PREFIX/include/
8062 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8064 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8065 functions that return function pointers and has support for NT specific
8066 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8067 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8068 unsigned to signed types: this was killing the Win32 compile.
8071 *) Add new certificate file to stack functions,
8072 SSL_add_dir_cert_subjects_to_stack() and
8073 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8074 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8075 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8076 This means that Apache-SSL and similar packages don't have to mess around
8077 to add as many CAs as they want to the preferred list.
8080 *) Experiment with doxygen documentation. Currently only partially applied to
8082 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8083 openssl.doxy as the configuration file.
8086 *) Get rid of remaining C++-style comments which strict C compilers hate.
8087 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8089 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8090 compiled in by default: it has problems with large keys.
8093 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8094 DH private keys and/or callback functions which directly correspond to
8095 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8096 is needed for applications which have to configure certificates on a
8097 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8099 For the RSA certificate situation is makes no difference, but
8100 for the DSA certificate situation this fixes the "no shared cipher"
8101 problem where the OpenSSL cipher selection procedure failed because the
8102 temporary keys were not overtaken from the context and the API provided
8103 no way to reconfigure them.
8104 The new functions now let applications reconfigure the stuff and they
8105 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8106 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8107 non-public-API function ssl_cert_instantiate() is used as a helper
8108 function and also to reduce code redundancy inside ssl_rsa.c.
8109 [Ralf S. Engelschall]
8111 *) Move s_server -dcert and -dkey options out of the undocumented feature
8112 area because they are useful for the DSA situation and should be
8113 recognized by the users.
8114 [Ralf S. Engelschall]
8116 *) Fix the cipher decision scheme for export ciphers: the export bits are
8117 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8118 SSL_EXP_MASK. So, the original variable has to be used instead of the
8119 already masked variable.
8120 [Richard Levitte <levitte@stacken.kth.se>]
8122 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8123 [Richard Levitte <levitte@stacken.kth.se>]
8125 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8126 from `int' to `unsigned int' because it's a length and initialized by
8127 EVP_DigestFinal() which expects an `unsigned int *'.
8128 [Richard Levitte <levitte@stacken.kth.se>]
8130 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8131 script. Instead use the usual Shell->Perl transition trick.
8132 [Ralf S. Engelschall]
8134 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8135 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8136 -noout -modulus' as it's already the case for `openssl rsa -noout
8137 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8138 currently the public key is printed (a decision which was already done by
8139 `openssl dsa -modulus' in the past) which serves a similar purpose.
8140 Additionally the NO_RSA no longer completely removes the whole -modulus
8141 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8143 [Ralf S. Engelschall]
8145 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8146 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8147 [Arne Ansper <arne@ats.cyber.ee>]
8149 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8150 to be added. Now both 'req' and 'ca' can use new objects defined in the
8154 *) Add cool BIO that does syslog (or event log on NT).
8155 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8157 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8158 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8159 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8160 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8163 *) Add preliminary config info for new extension code.
8166 *) Make RSA_NO_PADDING really use no padding.
8167 [Ulf Moeller <ulf@fitug.de>]
8169 *) Generate errors when private/public key check is done.
8172 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8173 for some CRL extensions and new objects added.
8176 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8177 key usage extension and fuller support for authority key id.
8180 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8181 padding method for RSA, which is recommended for new applications in PKCS
8182 #1 v2.0 (RFC 2437, October 1998).
8183 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8184 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8185 against Bleichbacher's attack on RSA.
8186 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8189 *) Updates to the new SSL compression code
8190 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8192 *) Fix so that the version number in the master secret, when passed
8193 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8194 (because the server will not accept higher), that the version number
8195 is 0x03,0x01, not 0x03,0x00
8196 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8198 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8199 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8200 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8203 *) Support for RAW extensions where an arbitrary extension can be
8204 created by including its DER encoding. See apps/openssl.cnf for
8208 *) Make sure latest Perl versions don't interpret some generated C array
8209 code as Perl array code in the crypto/err/err_genc.pl script.
8210 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8212 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8213 not many people have the assembler. Various Win32 compilation fixes and
8214 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8218 *) Modify configure script 'Configure' to automatically create crypto/date.h
8219 file under Win32 and also build pem.h from pem.org. New script
8220 util/mkfiles.pl to create the MINFO file on environments that can't do a
8221 'make files': perl util/mkfiles.pl >MINFO should work.
8224 *) Major rework of DES function declarations, in the pursuit of correctness
8225 and purity. As a result, many evil casts evaporated, and some weirdness,
8226 too. You may find this causes warnings in your code. Zapping your evil
8227 casts will probably fix them. Mostly.
8230 *) Fix for a typo in asn1.h. Bug fix to object creation script
8231 obj_dat.pl. It considered a zero in an object definition to mean
8232 "end of object": none of the objects in objects.h have any zeros
8233 so it wasn't spotted.
8234 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8236 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8237 Masking (CBCM). In the absence of test vectors, the best I have been able
8238 to do is check that the decrypt undoes the encrypt, so far. Send me test
8239 vectors if you have them.
8242 *) Correct calculation of key length for export ciphers (too much space was
8243 allocated for null ciphers). This has not been tested!
8246 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8247 message is now correct (it understands "crypto" and "ssl" on its
8248 command line). There is also now an "update" option. This will update
8249 the util/ssleay.num and util/libeay.num files with any new functions.
8251 perl util/mkdef.pl crypto ssl update
8252 it will update them.
8255 *) Overhauled the Perl interface (perl/*):
8256 - ported BN stuff to OpenSSL's different BN library
8257 - made the perl/ source tree CVS-aware
8258 - renamed the package from SSLeay to OpenSSL (the files still contain
8259 their history because I've copied them in the repository)
8260 - removed obsolete files (the test scripts will be replaced
8261 by better Test::Harness variants in the future)
8262 [Ralf S. Engelschall]
8264 *) First cut for a very conservative source tree cleanup:
8265 1. merge various obsolete readme texts into doc/ssleay.txt
8266 where we collect the old documents and readme texts.
8267 2. remove the first part of files where I'm already sure that we no
8268 longer need them because of three reasons: either they are just temporary
8269 files which were left by Eric or they are preserved original files where
8270 I've verified that the diff is also available in the CVS via "cvs diff
8271 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8272 the crypto/md/ stuff).
8273 [Ralf S. Engelschall]
8275 *) More extension code. Incomplete support for subject and issuer alt
8276 name, issuer and authority key id. Change the i2v function parameters
8277 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8278 what that's for :-) Fix to ASN1 macro which messed up
8279 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8282 *) Preliminary support for ENUMERATED type. This is largely copied from the
8286 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8287 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8289 *) Make sure `make rehash' target really finds the `openssl' program.
8290 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8292 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8293 like to hear about it if this slows down other processors.
8296 *) Add CygWin32 platform information to Configure script.
8297 [Alan Batie <batie@aahz.jf.intel.com>]
8299 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8300 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8302 *) New program nseq to manipulate netscape certificate sequences
8305 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8309 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8310 but the BN code had some problems that would cause failures when
8311 doing certificate verification and some other functions.
8312 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8314 *) Add ASN1 and PEM code to support netscape certificate sequences.
8317 *) Add ASN1 and PEM code to support netscape certificate sequences.
8320 *) Add several PKIX and private extended key usage OIDs.
8323 *) Modify the 'ca' program to handle the new extension code. Modify
8324 openssl.cnf for new extension format, add comments.
8327 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8328 and add a sample to openssl.cnf so req -x509 now adds appropriate
8332 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8333 error code, add initial support to X509_print() and x509 application.
8336 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8337 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8338 stuff is currently isolated and isn't even compiled yet.
8341 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8342 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8343 Removed the versions check from X509 routines when loading extensions:
8344 this allows certain broken certificates that don't set the version
8345 properly to be processed.
8348 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8349 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8350 can still be regenerated with "make depend".
8353 *) Spelling mistake in C version of CAST-128.
8354 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8356 *) Changes to the error generation code. The perl script err-code.pl
8357 now reads in the old error codes and retains the old numbers, only
8358 adding new ones if necessary. It also only changes the .err files if new
8359 codes are added. The makefiles have been modified to only insert errors
8360 when needed (to avoid needlessly modifying header files). This is done
8361 by only inserting errors if the .err file is newer than the auto generated
8362 C file. To rebuild all the error codes from scratch (the old behaviour)
8363 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8364 or delete all the .err files.
8367 *) CAST-128 was incorrectly implemented for short keys. The C version has
8368 been fixed, but is untested. The assembler versions are also fixed, but
8369 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8370 to regenerate it if needed.
8371 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8372 Hagino <itojun@kame.net>]
8374 *) File was opened incorrectly in randfile.c.
8375 [Ulf Möller <ulf@fitug.de>]
8377 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8378 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8379 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8380 al: it's just almost always a UTCTime. Note this patch adds new error
8381 codes so do a "make errors" if there are problems.
8384 *) Correct Linux 1 recognition in config.
8385 [Ulf Möller <ulf@fitug.de>]
8387 *) Remove pointless MD5 hash when using DSA keys in ca.
8388 [Anonymous <nobody@replay.com>]
8390 *) Generate an error if given an empty string as a cert directory. Also
8391 generate an error if handed NULL (previously returned 0 to indicate an
8392 error, but didn't set one).
8393 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8395 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8398 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8399 parameters. This was causing a warning which killed off the Win32 compile.
8402 *) Remove C++ style comments from crypto/bn/bn_local.h.
8403 [Neil Costigan <neil.costigan@celocom.com>]
8405 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8406 based on a text string, looking up short and long names and finally
8407 "dot" format. The "dot" format stuff didn't work. Added new function
8408 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8409 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8410 OID is not part of the table.
8413 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8414 X509_LOOKUP_by_alias().
8417 *) Sort openssl functions by name.
8420 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8421 encryption from sample DSA keys (in case anyone is interested the password
8425 *) Make _all_ *_free functions accept a NULL pointer.
8426 [Frans Heymans <fheymans@isaserver.be>]
8428 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8430 [Anonymous <nobody@replay.com>]
8432 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8433 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8435 *) Don't blow it for numeric -newkey arguments to apps/req.
8436 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8438 *) Temp key "for export" tests were wrong in s3_srvr.c.
8439 [Anonymous <nobody@replay.com>]
8441 *) Add prototype for temp key callback functions
8442 SSL_CTX_set_tmp_{rsa,dh}_callback().
8445 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8446 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8449 *) X509_name_add_entry() freed the wrong thing after an error.
8450 [Arne Ansper <arne@ats.cyber.ee>]
8452 *) rsa_eay.c would attempt to free a NULL context.
8453 [Arne Ansper <arne@ats.cyber.ee>]
8455 *) BIO_s_socket() had a broken should_retry() on Windoze.
8456 [Arne Ansper <arne@ats.cyber.ee>]
8458 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8459 [Arne Ansper <arne@ats.cyber.ee>]
8461 *) Make sure the already existing X509_STORE->depth variable is initialized
8462 in X509_STORE_new(), but document the fact that this variable is still
8463 unused in the certificate verification process.
8464 [Ralf S. Engelschall]
8466 *) Fix the various library and apps files to free up pkeys obtained from
8467 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8470 *) Fix reference counting in X509_PUBKEY_get(). This makes
8471 demos/maurice/example2.c work, amongst others, probably.
8472 [Steve Henson and Ben Laurie]
8474 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8475 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8476 are no longer created. This way we have a single and consistent command
8477 line interface `openssl <command>', similar to `cvs <command>'.
8478 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8480 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8481 BIT STRING wrapper always have zero unused bits.
8484 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8487 *) Make the top-level INSTALL documentation easier to understand.
8490 *) Makefiles updated to exit if an error occurs in a sub-directory
8491 make (including if user presses ^C) [Paul Sutton]
8493 *) Make Montgomery context stuff explicit in RSA data structure.
8496 *) Fix build order of pem and err to allow for generated pem.h.
8499 *) Fix renumbering bug in X509_NAME_delete_entry().
8502 *) Enhanced the err-ins.pl script so it makes the error library number
8503 global and can add a library name. This is needed for external ASN1 and
8504 other error libraries.
8507 *) Fixed sk_insert which never worked properly.
8510 *) Fix ASN1 macros so they can handle indefinite length construted
8511 EXPLICIT tags. Some non standard certificates use these: they can now
8515 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8516 into a single doc/ssleay.txt bundle. This way the information is still
8517 preserved but no longer messes up this directory. Now it's new room for
8518 the new set of documenation files.
8519 [Ralf S. Engelschall]
8521 *) SETs were incorrectly DER encoded. This was a major pain, because they
8522 shared code with SEQUENCEs, which aren't coded the same. This means that
8523 almost everything to do with SETs or SEQUENCEs has either changed name or
8524 number of arguments.
8525 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8527 *) Fix test data to work with the above.
8530 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8531 was already fixed by Eric for 0.9.1 it seems.
8532 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8534 *) Autodetect FreeBSD3.
8537 *) Fix various bugs in Configure. This affects the following platforms:
8541 unixware-2.0-pentium
8545 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8546 before they are needed.
8549 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8553 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8555 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8556 changed SSLeay to OpenSSL in version strings.
8557 [Ralf S. Engelschall]
8559 *) Some fixups to the top-level documents.
8562 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8563 because the symlink to include/ was missing.
8564 [Ralf S. Engelschall]
8566 *) Incorporated the popular no-RSA/DSA-only patches
8567 which allow to compile a RSA-free SSLeay.
8568 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8570 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8571 when "ssleay" is still not found.
8572 [Ralf S. Engelschall]
8574 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8575 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8577 *) Updated the README file.
8578 [Ralf S. Engelschall]
8580 *) Added various .cvsignore files in the CVS repository subdirs
8581 to make a "cvs update" really silent.
8582 [Ralf S. Engelschall]
8584 *) Recompiled the error-definition header files and added
8585 missing symbols to the Win32 linker tables.
8586 [Ralf S. Engelschall]
8588 *) Cleaned up the top-level documents;
8589 o new files: CHANGES and LICENSE
8590 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8591 o merged COPYRIGHT into LICENSE
8592 o removed obsolete TODO file
8593 o renamed MICROSOFT to INSTALL.W32
8594 [Ralf S. Engelschall]
8596 *) Removed dummy files from the 0.9.1b source tree:
8597 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8598 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8599 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8600 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8601 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8602 [Ralf S. Engelschall]
8604 *) Added various platform portability fixes.
8607 *) The Genesis of the OpenSSL rpject:
8608 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8609 Young and Tim J. Hudson created while they were working for C2Net until
8611 [The OpenSSL Project]
8614 Changes between 0.9.0b and 0.9.1b [not released]
8616 *) Updated a few CA certificates under certs/
8619 *) Changed some BIGNUM api stuff.
8622 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8623 DGUX x86, Linux Alpha, etc.
8626 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8627 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8631 *) Add -strparse option to asn1pars program which parses nested
8633 [Dr Stephen Henson <shenson@bigfoot.com>]
8635 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8638 *) DSA fix for "ca" program.
8641 *) Added "-genkey" option to "dsaparam" program.
8644 *) Added RIPE MD160 (rmd160) message digest.
8647 *) Added -a (all) option to "ssleay version" command.
8650 *) Added PLATFORM define which is the id given to Configure.
8653 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8656 *) Extended the ASN.1 parser routines.
8659 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8662 *) Added a BN_CTX to the BN library.
8665 *) Fixed the weak key values in DES library
8668 *) Changed API in EVP library for cipher aliases.
8671 *) Added support for RC2/64bit cipher.
8674 *) Converted the lhash library to the crypto/mem.c functions.
8677 *) Added more recognized ASN.1 object ids.
8680 *) Added more RSA padding checks for SSL/TLS.
8683 *) Added BIO proxy/filter functionality.
8686 *) Added extra_certs to SSL_CTX which can be used
8687 send extra CA certificates to the client in the CA cert chain sending
8688 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8691 *) Now Fortezza is denied in the authentication phase because
8692 this is key exchange mechanism is not supported by SSLeay at all.
8695 *) Additional PKCS1 checks.
8698 *) Support the string "TLSv1" for all TLS v1 ciphers.
8701 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8702 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8705 *) Fixed a few memory leaks.
8708 *) Fixed various code and comment typos.
8711 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8712 bytes sent in the client random.
8713 [Edward Bishop <ebishop@spyglass.com>]
projects / thirdparty / openssl.git / blob