5 Changes between 0.9.8p and 0.9.8q [xx XXX xxxx]
9 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
11 *) Fix extension code to avoid race conditions which can result in a buffer
12 overrun vulnerability: resumed sessions must not be modified as they can
13 be shared by multiple threads. CVE-2010-3864
16 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
19 *) Don't reencode certificate when calculating signature: cache and use
20 the original encoding instead. This makes signature verification of
21 some broken encodings work correctly.
24 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
25 is also one of the inputs.
26 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
28 *) Don't repeatedly append PBE algorithms to table if they already exist.
29 Sort table on each new add. This effectively makes the table read only
30 after all algorithms are added and subsequent calls to PKCS12_pbe_add
34 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
36 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
39 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
40 access or freeing data twice (CVE-2010-0742)
41 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
43 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
44 common in certificates and some applications which only call
45 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
49 Reduce copying into .apps and .test in makevms.com
50 Don't try to use blank CA certificate in CA.com
51 Allow use of C files from original directories in maketests.com
52 [Steven M. Schweda" <sms@antinode.info>]
54 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
56 *) When rejecting SSL/TLS records due to an incorrect version number, never
57 update s->server with a new major version number. As of
58 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
59 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
60 the previous behavior could result in a read attempt at NULL when
61 receiving specific incorrect SSL/TLS records once record payload
62 protection is active. (CVE-2010-0740)
63 [Bodo Moeller, Adam Langley <agl@chromium.org>]
65 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
66 could be crashed if the relevant tables were not present (e.g. chrooted).
67 [Tomas Hoger <thoger@redhat.com>]
69 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
71 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
72 [Martin Olsson, Neel Mehta]
74 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
75 accommodate for stack sorting, always a write lock!).
78 *) On some versions of WIN32 Heap32Next is very slow. This can cause
79 excessive delays in the RAND_poll(): over a minute. As a workaround
80 include a time check in the inner Heap32Next loop too.
83 *) The code that handled flushing of data in SSL/TLS originally used the
84 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
85 the problem outlined in PR#1949. The fix suggested there however can
86 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
87 of Apache). So instead simplify the code to flush unconditionally.
88 This should be fine since flushing with no data to flush is a no op.
91 *) Handle TLS versions 2.0 and later properly and correctly use the
92 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
93 off ancient servers have a habit of sticking around for a while...
96 *) Modify compression code so it frees up structures without using the
97 ex_data callbacks. This works around a problem where some applications
98 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
99 restarting) then use compression (e.g. SSL with compression) later.
100 This results in significant per-connection memory leaks and
101 has caused some security issues including CVE-2008-1678 and
105 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
106 change when encrypting or decrypting.
109 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
110 connect and renegotiate with servers which do not support RI.
111 Until RI is more widely deployed this option is enabled by default.
114 *) Add "missing" ssl ctrls to clear options and mode.
117 *) If client attempts to renegotiate and doesn't support RI respond with
118 a no_renegotiation alert as required by RFC5746. Some renegotiating
119 TLS clients will continue a connection gracefully when they receive
120 the alert. Unfortunately OpenSSL mishandled this alert and would hang
121 waiting for a server hello which it will never receive. Now we treat a
122 received no_renegotiation alert as a fatal error. This is because
123 applications requesting a renegotiation might well expect it to succeed
124 and would have no code in place to handle the server denying it so the
125 only safe thing to do is to terminate the connection.
128 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
129 peer supports secure renegotiation and 0 otherwise. Print out peer
130 renegotiation support in s_client/s_server.
133 *) Replace the highly broken and deprecated SPKAC certification method with
134 the updated NID creation version. This should correctly handle UTF8.
137 *) Implement RFC5746. Re-enable renegotiation but require the extension
138 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
139 turns out to be a bad idea. It has been replaced by
140 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
141 SSL_CTX_set_options(). This is really not recommended unless you
142 know what you are doing.
143 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
145 *) Fixes to stateless session resumption handling. Use initial_ctx when
146 issuing and attempting to decrypt tickets in case it has changed during
147 servername handling. Use a non-zero length session ID when attempting
148 stateless session resumption: this makes it possible to determine if
149 a resumption has occurred immediately after receiving server hello
150 (several places in OpenSSL subtly assume this) instead of later in
154 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
155 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
156 fixes for a few places where the return code is not checked
158 [Julia Lawall <julia@diku.dk>]
160 *) Add --strict-warnings option to Configure script to include devteam
161 warnings in other configurations.
164 *) Add support for --libdir option and LIBDIR variable in makefiles. This
165 makes it possible to install openssl libraries in locations which
166 have names other than "lib", for example "/usr/lib64" which some
168 [Steve Henson, based on patch from Jeremy Utley]
170 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
171 X690 8.9.12 and can produce some misleading textual output of OIDs.
172 [Steve Henson, reported by Dan Kaminsky]
174 *) Delete MD2 from algorithm tables. This follows the recommendation in
175 several standards that it is not used in new applications due to
176 several cryptographic weaknesses. For binary compatibility reasons
177 the MD2 API is still compiled in by default.
180 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
184 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
185 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
187 [Guenter <lists@gknw.net>]
189 *) Fix the server certificate chain building code to use X509_verify_cert(),
190 it used to have an ad-hoc builder which was unable to cope with anything
191 other than a simple chain.
192 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
194 *) Don't check self signed certificate signatures in X509_verify_cert()
195 by default (a flag can override this): it just wastes time without
196 adding any security. As a useful side effect self signed root CAs
197 with non-FIPS digests are now usable in FIPS mode.
200 *) In dtls1_process_out_of_seq_message() the check if the current message
201 is already buffered was missing. For every new message was memory
202 allocated, allowing an attacker to perform an denial of service attack
203 with sending out of seq handshake messages until there is no memory
204 left. Additionally every future messege was buffered, even if the
205 sequence number made no sense and would be part of another handshake.
206 So only messages with sequence numbers less than 10 in advance will be
207 buffered. (CVE-2009-1378)
208 [Robin Seggelmann, discovered by Daniel Mentz]
210 *) Records are buffered if they arrive with a future epoch to be
211 processed after finishing the corresponding handshake. There is
212 currently no limitation to this buffer allowing an attacker to perform
213 a DOS attack with sending records with future epochs until there is no
214 memory left. This patch adds the pqueue_size() function to detemine
215 the size of a buffer and limits the record buffer to 100 entries.
217 [Robin Seggelmann, discovered by Daniel Mentz]
219 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
220 parent structure is freed. (CVE-2009-1379)
223 *) Handle non-blocking I/O properly in SSL_shutdown() call.
224 [Darryl Miles <darryl-mailinglists@netbauds.net>]
227 [Ilya O. <vrghost@gmail.com>]
229 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
231 *) Disable renegotiation completely - this fixes a severe security
232 problem (CVE-2009-3555) at the cost of breaking all
233 renegotiation. Renegotiation can be re-enabled by setting
234 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
235 run-time. This is really not recommended unless you know what
239 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
241 *) Don't set val to NULL when freeing up structures, it is freed up by
242 underlying code. If sizeof(void *) > sizeof(long) this can result in
243 zeroing past the valid field. (CVE-2009-0789)
244 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
246 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
247 checked correctly. This would allow some invalid signed attributes to
248 appear to verify correctly. (CVE-2009-0591)
249 [Ivan Nestlerode <inestlerode@us.ibm.com>]
251 *) Reject UniversalString and BMPString types with invalid lengths. This
252 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
253 a legal length. (CVE-2009-0590)
256 *) Set S/MIME signing as the default purpose rather than setting it
257 unconditionally. This allows applications to override it at the store
261 *) Permit restricted recursion of ASN1 strings. This is needed in practice
262 to handle some structures.
265 *) Improve efficiency of mem_gets: don't search whole buffer each time
267 [Jeremy Shapiro <jnshapir@us.ibm.com>]
269 *) New -hex option for openssl rand.
272 *) Print out UTF8String and NumericString when parsing ASN1.
275 *) Support NumericString type for name components.
278 *) Allow CC in the environment to override the automatically chosen
279 compiler. Note that nothing is done to ensure flags work with the
283 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
285 *) Properly check EVP_VerifyFinal() and similar return values
287 [Ben Laurie, Bodo Moeller, Google Security Team]
289 *) Enable TLS extensions by default.
292 *) Allow the CHIL engine to be loaded, whether the application is
293 multithreaded or not. (This does not release the developer from the
294 obligation to set up the dynamic locking callbacks.)
295 [Sander Temme <sander@temme.net>]
297 *) Use correct exit code if there is an error in dgst command.
298 [Steve Henson; problem pointed out by Roland Dirlewanger]
300 *) Tweak Configure so that you need to say "experimental-jpake" to enable
301 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
304 *) Add experimental JPAKE support, including demo authentication in
305 s_client and s_server.
308 *) Set the comparison function in v3_addr_canonize().
309 [Rob Austein <sra@hactrn.net>]
311 *) Add support for XMPP STARTTLS in s_client.
312 [Philip Paeps <philip@freebsd.org>]
314 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
315 to ensure that even with this option, only ciphersuites in the
316 server's preference list will be accepted. (Note that the option
317 applies only when resuming a session, so the earlier behavior was
318 just about the algorithm choice for symmetric cryptography.)
321 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
323 *) Fix NULL pointer dereference if a DTLS server received
324 ChangeCipherSpec as first record (CVE-2009-1386).
327 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
328 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
331 *) The fix in 0.9.8c that supposedly got rid of unsafe
332 double-checked locking was incomplete for RSA blinding,
333 addressing just one layer of what turns out to have been
334 doubly unsafe triple-checked locking.
336 So now fix this for real by retiring the MONT_HELPER macro
337 in crypto/rsa/rsa_eay.c.
339 [Bodo Moeller; problem pointed out by Marius Schilder]
341 *) Various precautionary measures:
343 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
345 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
346 (NB: This would require knowledge of the secret session ticket key
347 to exploit, in which case you'd be SOL either way.)
349 - Change bn_nist.c so that it will properly handle input BIGNUMs
350 outside the expected range.
352 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
355 [Neel Mehta, Bodo Moeller]
357 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
358 the load fails. Useful for distros.
359 [Ben Laurie and the FreeBSD team]
361 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
364 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
367 *) Expand ENGINE to support engine supplied SSL client certificate functions.
369 This work was sponsored by Logica.
372 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
373 keystores. Support for SSL/TLS client authentication too.
374 Not compiled unless enable-capieng specified to Configure.
376 This work was sponsored by Logica.
379 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
380 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
381 attribute creation routines such as certifcate requests and PKCS#12
385 Changes between 0.9.8g and 0.9.8h [28 May 2008]
387 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
388 handshake which could lead to a cilent crash as found using the
389 Codenomicon TLS test suite (CVE-2008-1672)
390 [Steve Henson, Mark Cox]
392 *) Fix double free in TLS server name extensions which could lead to
393 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
396 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
398 Clear the error queue to ensure that error entries left from
399 older function calls do not interfere with the correct operation.
400 [Lutz Jaenicke, Erik de Castro Lopo]
402 *) Remove root CA certificates of commercial CAs:
404 The OpenSSL project does not recommend any specific CA and does not
405 have any policy with respect to including or excluding any CA.
406 Therefore it does not make any sense to ship an arbitrary selection
407 of root CA certificates with the OpenSSL software.
410 *) RSA OAEP patches to fix two separate invalid memory reads.
411 The first one involves inputs when 'lzero' is greater than
412 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
413 before the beginning of from). The second one involves inputs where
414 the 'db' section contains nothing but zeroes (there is a one-byte
415 invalid read after the end of 'db').
416 [Ivan Nestlerode <inestlerode@us.ibm.com>]
418 *) Partial backport from 0.9.9-dev:
420 Introduce bn_mul_mont (dedicated Montgomery multiplication
421 procedure) as a candidate for BIGNUM assembler implementation.
422 While 0.9.9-dev uses assembler for various architectures, only
423 x86_64 is available by default here in the 0.9.8 branch, and
424 32-bit x86 is available through a compile-time setting.
426 To try the 32-bit x86 assembler implementation, use Configure
427 option "enable-montasm" (which exists only for this backport).
429 As "enable-montasm" for 32-bit x86 disclaims code stability
430 anyway, in this constellation we activate additional code
431 backported from 0.9.9-dev for further performance improvements,
432 namely BN_from_montgomery_word. (To enable this otherwise,
433 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
435 [Andy Polyakov (backport partially by Bodo Moeller)]
437 *) Add TLS session ticket callback. This allows an application to set
438 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
439 values. This is useful for key rollover for example where several key
440 sets may exist with different names.
443 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
444 This was broken until now in 0.9.8 releases, such that the only way
445 a registered ENGINE could be used (assuming it initialises
446 successfully on the host) was to explicitly set it as the default
447 for the relevant algorithms. This is in contradiction with 0.9.7
448 behaviour and the documentation. With this fix, when an ENGINE is
449 registered into a given algorithm's table of implementations, the
450 'uptodate' flag is reset so that auto-discovery will be used next
451 time a new context for that algorithm attempts to select an
453 [Ian Lister (tweaked by Geoff Thorpe)]
455 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
456 implemention in the following ways:
458 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
461 Lack of BER streaming support means one pass streaming processing is
462 only supported if data is detached: setting the streaming flag is
463 ignored for embedded content.
465 CMS support is disabled by default and must be explicitly enabled
466 with the enable-cms configuration option.
469 *) Update the GMP engine glue to do direct copies between BIGNUM and
470 mpz_t when openssl and GMP use the same limb size. Otherwise the
471 existing "conversion via a text string export" trick is still used.
472 [Paul Sheer <paulsheer@gmail.com>]
474 *) Zlib compression BIO. This is a filter BIO which compressed and
475 uncompresses any data passed through it.
478 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
479 RFC3394 compatible AES key wrapping.
482 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
483 sets string data without copying. X509_ALGOR_set0() and
484 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
485 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
486 from an X509_ATTRIBUTE structure optionally checking it occurs only
487 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
491 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
492 to get the expected BN_FLG_CONSTTIME behavior.
493 [Bodo Moeller (Google)]
497 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
498 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
499 - added some more tests to do_tests.pl
500 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
501 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
502 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
503 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
504 - various changes to netware.pl to enable gcc-cross builds on Win32
506 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
507 - various changes to fix missing prototype warnings
508 - fixed x86nasm.pl to create correct asm files for NASM COFF output
509 - added AES, WHIRLPOOL and CPUID assembler code to build files
510 - added missing AES assembler make rules to mk1mf.pl
511 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
512 [Guenter Knauf <eflash@gmx.net>]
514 *) Implement certificate status request TLS extension defined in RFC3546.
515 A client can set the appropriate parameters and receive the encoded
516 OCSP response via a callback. A server can query the supplied parameters
517 and set the encoded OCSP response in the callback. Add simplified examples
518 to s_client and s_server.
521 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
524 + Binary incompatibility of ssl_ctx_st structure
525 + DTLS interoperation with non-compliant servers
526 + Don't call get_session_cb() without proposed session
527 + Fix ia64 assembler code
528 [Andy Polyakov, Steve Henson]
530 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
532 *) DTLS Handshake overhaul. There were longstanding issues with
533 OpenSSL DTLS implementation, which were making it impossible for
534 RFC 4347 compliant client to communicate with OpenSSL server.
535 Unfortunately just fixing these incompatibilities would "cut off"
536 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
537 server keeps tolerating non RFC compliant syntax. The opposite is
538 not true, 0.9.8f client can not communicate with earlier server.
539 This update even addresses CVE-2007-4995.
542 *) Changes to avoid need for function casts in OpenSSL: some compilers
543 (gcc 4.2 and later) reject their use.
544 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
547 *) Add RFC4507 support to OpenSSL. This includes the corrections in
548 RFC4507bis. The encrypted ticket format is an encrypted encoded
549 SSL_SESSION structure, that way new session features are automatically
552 If a client application caches session in an SSL_SESSION structure
553 support is transparent because tickets are now stored in the encoded
556 The SSL_CTX structure automatically generates keys for ticket
557 protection in servers so again support should be possible
558 with no application modification.
560 If a client or server wishes to disable RFC4507 support then the option
561 SSL_OP_NO_TICKET can be set.
563 Add a TLS extension debugging callback to allow the contents of any client
564 or server extensions to be examined.
566 This work was sponsored by Google.
569 *) Add initial support for TLS extensions, specifically for the server_name
570 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
571 have new members for a host name. The SSL data structure has an
572 additional member SSL_CTX *initial_ctx so that new sessions can be
573 stored in that context to allow for session resumption, even after the
574 SSL has been switched to a new SSL_CTX in reaction to a client's
575 server_name extension.
577 New functions (subject to change):
580 SSL_get_servername_type()
583 New CTRL codes and macros (subject to change):
585 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
586 - SSL_CTX_set_tlsext_servername_callback()
587 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
588 - SSL_CTX_set_tlsext_servername_arg()
589 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
591 openssl s_client has a new '-servername ...' option.
593 openssl s_server has new options '-servername_host ...', '-cert2 ...',
594 '-key2 ...', '-servername_fatal' (subject to change). This allows
595 testing the HostName extension for a specific single host name ('-cert'
596 and '-key' remain fallbacks for handshakes without HostName
597 negotiation). If the unrecogninzed_name alert has to be sent, this by
598 default is a warning; it becomes fatal with the '-servername_fatal'
601 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
603 *) Add AES and SSE2 assembly language support to VC++ build.
606 *) Mitigate attack on final subtraction in Montgomery reduction.
609 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
610 (which previously caused an internal error).
613 *) Squeeze another 10% out of IGE mode when in != out.
616 *) AES IGE mode speedup.
617 [Dean Gaudet (Google)]
619 *) Add the Korean symmetric 128-bit cipher SEED (see
620 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
621 add SEED ciphersuites from RFC 4162:
623 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
624 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
625 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
626 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
628 To minimize changes between patchlevels in the OpenSSL 0.9.8
629 series, SEED remains excluded from compilation unless OpenSSL
630 is configured with 'enable-seed'.
633 *) Mitigate branch prediction attacks, which can be practical if a
634 single processor is shared, allowing a spy process to extract
635 information. For detailed background information, see
636 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
637 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
638 and Necessary Software Countermeasures"). The core of the change
639 are new versions BN_div_no_branch() and
640 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
641 respectively, which are slower, but avoid the security-relevant
642 conditional branches. These are automatically called by BN_div()
643 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
644 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
645 remove a conditional branch.
647 BN_FLG_CONSTTIME is the new name for the previous
648 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
649 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
650 in the exponent causes BN_mod_exp_mont() to use the alternative
651 implementation in BN_mod_exp_mont_consttime().) The old name
652 remains as a deprecated alias.
654 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
655 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
656 constant-time implementations for more than just exponentiation.
657 Here too the old name is kept as a deprecated alias.
659 BN_BLINDING_new() will now use BN_dup() for the modulus so that
660 the BN_BLINDING structure gets an independent copy of the
661 modulus. This means that the previous "BIGNUM *m" argument to
662 BN_BLINDING_new() and to BN_BLINDING_create_param() now
663 essentially becomes "const BIGNUM *m", although we can't actually
664 change this in the header file before 0.9.9. It allows
665 RSA_setup_blinding() to use BN_with_flags() on the modulus to
666 enable BN_FLG_CONSTTIME.
668 [Matthew D Wood (Intel Corp)]
670 *) In the SSL/TLS server implementation, be strict about session ID
671 context matching (which matters if an application uses a single
672 external cache for different purposes). Previously,
673 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
674 set. This did ensure strict client verification, but meant that,
675 with applications using a single external cache for quite
676 different requirements, clients could circumvent ciphersuite
677 restrictions for a given session ID context by starting a session
678 in a different context.
681 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
682 a ciphersuite string such as "DEFAULT:RSA" cannot enable
683 authentication-only ciphersuites.
686 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
687 not complete and could lead to a possible single byte overflow
688 (CVE-2007-5135) [Ben Laurie]
690 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
692 *) Since AES128 and AES256 (and similarly Camellia128 and
693 Camellia256) share a single mask bit in the logic of
694 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
695 kludge to work properly if AES128 is available and AES256 isn't
696 (or if Camellia128 is available and Camellia256 isn't).
699 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
700 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
701 When a point or a seed is encoded in a BIT STRING, we need to
702 prevent the removal of trailing zero bits to get the proper DER
703 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
704 of a NamedBitList, for which trailing 0 bits need to be removed.)
707 *) Have SSL/TLS server implementation tolerate "mismatched" record
708 protocol version while receiving ClientHello even if the
709 ClientHello is fragmented. (The server can't insist on the
710 particular protocol version it has chosen before the ServerHello
711 message has informed the client about his choice.)
714 *) Add RFC 3779 support.
715 [Rob Austein for ARIN, Ben Laurie]
717 *) Load error codes if they are not already present instead of using a
718 static variable. This allows them to be cleanly unloaded and reloaded.
719 Improve header file function name parsing.
722 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
723 or CAPABILITY handshake as required by RFCs.
726 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
728 *) Introduce limits to prevent malicious keys being able to
729 cause a denial of service. (CVE-2006-2940)
730 [Steve Henson, Bodo Moeller]
732 *) Fix ASN.1 parsing of certain invalid structures that can result
733 in a denial of service. (CVE-2006-2937) [Steve Henson]
735 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
736 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
738 *) Fix SSL client code which could crash if connecting to a
739 malicious SSLv2 server. (CVE-2006-4343)
740 [Tavis Ormandy and Will Drewry, Google Security Team]
742 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
743 match only those. Before that, "AES256-SHA" would be interpreted
744 as a pattern and match "AES128-SHA" too (since AES128-SHA got
745 the same strength classification in 0.9.7h) as we currently only
746 have a single AES bit in the ciphersuite description bitmap.
747 That change, however, also applied to ciphersuite strings such as
748 "RC4-MD5" that intentionally matched multiple ciphersuites --
749 namely, SSL 2.0 ciphersuites in addition to the more common ones
750 from SSL 3.0/TLS 1.0.
752 So we change the selection algorithm again: Naming an explicit
753 ciphersuite selects this one ciphersuite, and any other similar
754 ciphersuite (same bitmap) from *other* protocol versions.
755 Thus, "RC4-MD5" again will properly select both the SSL 2.0
756 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
758 Since SSL 2.0 does not have any ciphersuites for which the
759 128/256 bit distinction would be relevant, this works for now.
760 The proper fix will be to use different bits for AES128 and
761 AES256, which would have avoided the problems from the beginning;
762 however, bits are scarce, so we can only do this in a new release
763 (not just a patchlevel) when we can change the SSL_CIPHER
764 definition to split the single 'unsigned long mask' bitmap into
765 multiple values to extend the available space.
769 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
771 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
772 (CVE-2006-4339) [Ben Laurie and Google Security Team]
774 *) Add AES IGE and biIGE modes.
777 *) Change the Unix randomness entropy gathering to use poll() when
778 possible instead of select(), since the latter has some
779 undesirable limitations.
780 [Darryl Miles via Richard Levitte and Bodo Moeller]
782 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
783 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
784 cannot be implicitly activated as part of, e.g., the "AES" alias.
785 However, please upgrade to OpenSSL 0.9.9[-dev] for
786 non-experimental use of the ECC ciphersuites to get TLS extension
787 support, which is required for curve and point format negotiation
788 to avoid potential handshake problems.
791 *) Disable rogue ciphersuites:
793 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
794 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
795 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
797 The latter two were purportedly from
798 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
801 Also deactivate the remaining ciphersuites from
802 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
803 unofficial, and the ID has long expired.
806 *) Fix RSA blinding Heisenbug (problems sometimes occured on
807 dual-core machines) and other potential thread-safety issues.
810 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
811 versions), which is now available for royalty-free use
812 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
813 Also, add Camellia TLS ciphersuites from RFC 4132.
815 To minimize changes between patchlevels in the OpenSSL 0.9.8
816 series, Camellia remains excluded from compilation unless OpenSSL
817 is configured with 'enable-camellia'.
820 *) Disable the padding bug check when compression is in use. The padding
821 bug check assumes the first packet is of even length, this is not
822 necessarily true if compresssion is enabled and can result in false
823 positives causing handshake failure. The actual bug test is ancient
824 code so it is hoped that implementations will either have fixed it by
825 now or any which still have the bug do not support compression.
828 Changes between 0.9.8a and 0.9.8b [04 May 2006]
830 *) When applying a cipher rule check to see if string match is an explicit
831 cipher suite and only match that one cipher suite if it is.
834 *) Link in manifests for VC++ if needed.
835 [Austin Ziegler <halostatue@gmail.com>]
837 *) Update support for ECC-based TLS ciphersuites according to
838 draft-ietf-tls-ecc-12.txt with proposed changes (but without
839 TLS extensions, which are supported starting with the 0.9.9
840 branch, not in the OpenSSL 0.9.8 branch).
843 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
844 opaque EVP_CIPHER_CTX handling.
847 *) Fixes and enhancements to zlib compression code. We now only use
848 "zlib1.dll" and use the default __cdecl calling convention on Win32
849 to conform with the standards mentioned here:
850 http://www.zlib.net/DLL_FAQ.txt
851 Static zlib linking now works on Windows and the new --with-zlib-include
852 --with-zlib-lib options to Configure can be used to supply the location
853 of the headers and library. Gracefully handle case where zlib library
857 *) Several fixes and enhancements to the OID generation code. The old code
858 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
859 handle numbers larger than ULONG_MAX, truncated printing and had a
860 non standard OBJ_obj2txt() behaviour.
863 *) Add support for building of engines under engine/ as shared libraries
864 under VC++ build system.
867 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
868 Hopefully, we will not see any false combination of paths any more.
871 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
873 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
874 (part of SSL_OP_ALL). This option used to disable the
875 countermeasure against man-in-the-middle protocol-version
876 rollback in the SSL 2.0 server implementation, which is a bad
877 idea. (CVE-2005-2969)
879 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
880 for Information Security, National Institute of Advanced Industrial
881 Science and Technology [AIST], Japan)]
883 *) Add two function to clear and return the verify parameter flags.
886 *) Keep cipherlists sorted in the source instead of sorting them at
887 runtime, thus removing the need for a lock.
890 *) Avoid some small subgroup attacks in Diffie-Hellman.
891 [Nick Mathewson and Ben Laurie]
893 *) Add functions for well-known primes.
896 *) Extended Windows CE support.
897 [Satoshi Nakamura and Andy Polyakov]
899 *) Initialize SSL_METHOD structures at compile time instead of during
900 runtime, thus removing the need for a lock.
903 *) Make PKCS7_decrypt() work even if no certificate is supplied by
904 attempting to decrypt each encrypted key in turn. Add support to
908 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
910 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
913 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
916 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
917 key into the same file any more.
920 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
923 *) Add -utf8 command line and config file option to 'ca'.
924 [Stefan <stf@udoma.org]
926 *) Removed the macro des_crypt(), as it seems to conflict with some
927 libraries. Use DES_crypt().
930 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
931 involves renaming the source and generated shared-libs for
932 both. The engines will accept the corrected or legacy ids
933 ('ncipher' and '4758_cca' respectively) when binding. NB,
934 this only applies when building 'shared'.
935 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
937 *) Add attribute functions to EVP_PKEY structure. Modify
938 PKCS12_create() to recognize a CSP name attribute and
939 use it. Make -CSP option work again in pkcs12 utility.
942 *) Add new functionality to the bn blinding code:
943 - automatic re-creation of the BN_BLINDING parameters after
944 a fixed number of uses (currently 32)
945 - add new function for parameter creation
946 - introduce flags to control the update behaviour of the
947 BN_BLINDING parameters
948 - hide BN_BLINDING structure
949 Add a second BN_BLINDING slot to the RSA structure to improve
950 performance when a single RSA object is shared among several
954 *) Add support for DTLS.
955 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
957 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
958 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
961 *) Remove buggy and incompletet DH cert support from
962 ssl/ssl_rsa.c and ssl/s3_both.c
965 *) Use SHA-1 instead of MD5 as the default digest algorithm for
966 the apps/openssl applications.
969 *) Compile clean with "-Wall -Wmissing-prototypes
970 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
971 DEBUG_SAFESTACK must also be set.
974 *) Change ./Configure so that certain algorithms can be disabled by default.
975 The new counterpiece to "no-xxx" is "enable-xxx".
977 The patented RC5 and MDC2 algorithms will now be disabled unless
978 "enable-rc5" and "enable-mdc2", respectively, are specified.
980 (IDEA remains enabled despite being patented. This is because IDEA
981 is frequently required for interoperability, and there is no license
982 fee for non-commercial use. As before, "no-idea" can be used to
983 avoid this algorithm.)
987 *) Add processing of proxy certificates (see RFC 3820). This work was
988 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
989 EGEE (Enabling Grids for E-science in Europe).
992 *) RC4 performance overhaul on modern architectures/implementations, such
993 as Intel P4, IA-64 and AMD64.
996 *) New utility extract-section.pl. This can be used specify an alternative
997 section number in a pod file instead of having to treat each file as
998 a separate case in Makefile. This can be done by adding two lines to the
1001 =for comment openssl_section:XXX
1003 The blank line is mandatory.
1007 *) New arguments -certform, -keyform and -pass for s_client and s_server
1008 to allow alternative format key and certificate files and passphrase
1012 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1013 update associated structures and add various utility functions.
1015 Add new policy related verify parameters, include policy checking in
1016 standard verify code. Enhance 'smime' application with extra parameters
1017 to support policy checking and print out.
1020 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1021 Nehemiah processors. These extensions support AES encryption in hardware
1022 as well as RNG (though RNG support is currently disabled).
1023 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1025 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1028 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1029 [Andy Polyakov and a number of other people]
1031 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1032 implementation contributed by IBM.
1033 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1035 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1036 exponent rather than 'unsigned long'. There is a corresponding change to
1037 the new 'rsa_keygen' element of the RSA_METHOD structure.
1038 [Jelte Jansen, Geoff Thorpe]
1040 *) Functionality for creating the initial serial number file is now
1041 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1043 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1044 number file to 1, which is bound to cause problems. To avoid
1045 the problems while respecting compatibility between different 0.9.7
1046 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1047 CA.pl for serial number initialization. With the new release 0.9.8,
1048 we can fix the problem directly in the 'ca' utility.)
1051 *) Reduced header interdepencies by declaring more opaque objects in
1052 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1053 give fewer recursive includes, which could break lazy source code - so
1054 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1055 developers should define this symbol when building and using openssl to
1056 ensure they track the recommended behaviour, interfaces, [etc], but
1057 backwards-compatible behaviour prevails when this isn't defined.
1060 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1063 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1064 This will generate a random key of the appropriate length based on the
1065 cipher context. The EVP_CIPHER can provide its own random key generation
1066 routine to support keys of a specific form. This is used in the des and
1067 3des routines to generate a key of the correct parity. Update S/MIME
1068 code to use new functions and hence generate correct parity DES keys.
1069 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1070 valid (weak or incorrect parity).
1073 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1074 as looking them up. This is useful when the verified structure may contain
1075 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1076 present unless the new PKCS7_NO_CRL flag is asserted.
1079 *) Extend ASN1 oid configuration module. It now additionally accepts the
1082 shortName = some long name, 1.2.3.4
1085 *) Reimplemented the BN_CTX implementation. There is now no more static
1086 limitation on the number of variables it can handle nor the depth of the
1087 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1088 information can now expand as required, and rather than having a single
1089 static array of bignums, BN_CTX now uses a linked-list of such arrays
1090 allowing it to expand on demand whilst maintaining the usefulness of
1091 BN_CTX's "bundling".
1094 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1095 to allow all RSA operations to function using a single BN_CTX.
1098 *) Preliminary support for certificate policy evaluation and checking. This
1099 is initially intended to pass the tests outlined in "Conformance Testing
1100 of Relying Party Client Certificate Path Processing Logic" v1.07.
1103 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1104 remained unused and not that useful. A variety of other little bignum
1105 tweaks and fixes have also been made continuing on from the audit (see
1109 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1110 associated ASN1, EVP and SSL functions and old ASN1 macros.
1113 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1114 and this should never fail. So the return value from the use of
1115 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1116 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1119 *) BN_CTX_get() should return zero-valued bignums, providing the same
1120 initialised value as BN_new().
1121 [Geoff Thorpe, suggested by Ulf Möller]
1123 *) Support for inhibitAnyPolicy certificate extension.
1126 *) An audit of the BIGNUM code is underway, for which debugging code is
1127 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1128 is considered valid when processing BIGNUMs, and causes execution to
1129 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1130 further steps are taken to deliberately pollute unused data in BIGNUM
1131 structures to try and expose faulty code further on. For now, openssl will
1132 (in its default mode of operation) continue to tolerate the inconsistent
1133 forms that it has tolerated in the past, but authors and packagers should
1134 consider trying openssl and their own applications when compiled with
1135 these debugging symbols defined. It will help highlight potential bugs in
1136 their own code, and will improve the test coverage for OpenSSL itself. At
1137 some point, these tighter rules will become openssl's default to improve
1138 maintainability, though the assert()s and other overheads will remain only
1139 in debugging configurations. See bn.h for more details.
1140 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1142 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1143 that can only be obtained through BN_CTX_new() (which implicitly
1144 initialises it). The presence of this function only made it possible
1145 to overwrite an existing structure (and cause memory leaks).
1148 *) Because of the callback-based approach for implementing LHASH as a
1149 template type, lh_insert() adds opaque objects to hash-tables and
1150 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1151 to clean up those corresponding objects before destroying the hash table
1152 (and losing the object pointers). So some over-zealous constifications in
1153 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1154 objects as "const" and the lh_doall[_arg] callback wrappers are not
1155 prototyped to have "const" restrictions on the object pointers they are
1156 given (and so aren't required to cast them away any more).
1159 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1160 (speed) prefers to use its own implementation. The two implementations
1161 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1162 its object type properly exposed (MS_TM) instead of casting to/from "char
1163 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1164 aren't necessarily the greatest nomenclatures - but this is what was used
1165 internally to the implementation so I've used that for now.
1168 *) Ensure that deprecated functions do not get compiled when
1169 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1170 the self-tests were still using deprecated key-generation functions so
1171 these have been updated also.
1174 *) Reorganise PKCS#7 code to separate the digest location functionality
1175 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1176 New function PKCS7_set_digest() to set the digest type for PKCS#7
1177 digestedData type. Add additional code to correctly generate the
1178 digestedData type and add support for this type in PKCS7 initialization
1182 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1183 structure of type "other".
1186 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1187 sure the loop does correctly stop and breaking ("division by zero")
1188 modulus operations are not performed. The (pre-generated) prime
1189 table crypto/bn/bn_prime.h was already correct, but it could not be
1190 re-generated on some platforms because of the "division by zero"
1191 situation in the script.
1192 [Ralf S. Engelschall]
1194 *) Update support for ECC-based TLS ciphersuites according to
1195 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1196 SHA-1 now is only used for "small" curves (where the
1197 representation of a field element takes up to 24 bytes); for
1198 larger curves, the field element resulting from ECDH is directly
1199 used as premaster secret.
1200 [Douglas Stebila (Sun Microsystems Laboratories)]
1202 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1203 curve secp160r1 to the tests.
1204 [Douglas Stebila (Sun Microsystems Laboratories)]
1206 *) Add the possibility to load symbols globally with DSO.
1207 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1209 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1210 control of the error stack.
1213 *) Add support for STORE in ENGINE.
1216 *) Add the STORE type. The intention is to provide a common interface
1217 to certificate and key stores, be they simple file-based stores, or
1218 HSM-type store, or LDAP stores, or...
1219 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1222 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1223 pass a list of arguments to any function as well as provide a way
1224 for a function to pass data back to the caller.
1227 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1228 works like BUF_strdup() but can be used to duplicate a portion of
1229 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1233 *) Add the function sk_find_ex() which works like sk_find(), but will
1234 return an index to an element even if an exact match couldn't be
1235 found. The index is guaranteed to point at the element where the
1236 searched-for key would be inserted to preserve sorting order.
1239 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1240 takes an extra flags argument for optional functionality. Currently,
1241 the following flags are defined:
1243 OBJ_BSEARCH_VALUE_ON_NOMATCH
1244 This one gets OBJ_bsearch_ex() to return a pointer to the first
1245 element where the comparing function returns a negative or zero
1248 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1249 This one gets OBJ_bsearch_ex() to return a pointer to the first
1250 element where the comparing function returns zero. This is useful
1251 if there are more than one element where the comparing function
1255 *) Make it possible to create self-signed certificates with 'openssl ca'
1256 in such a way that the self-signed certificate becomes part of the
1257 CA database and uses the same mechanisms for serial number generation
1258 as all other certificate signing. The new flag '-selfsign' enables
1259 this functionality. Adapt CA.sh and CA.pl.in.
1262 *) Add functionality to check the public key of a certificate request
1263 against a given private. This is useful to check that a certificate
1264 request can be signed by that key (self-signing).
1267 *) Make it possible to have multiple active certificates with the same
1268 subject in the CA index file. This is done only if the keyword
1269 'unique_subject' is set to 'no' in the main CA section (default
1270 if 'CA_default') of the configuration file. The value is saved
1271 with the database itself in a separate index attribute file,
1272 named like the index file with '.attr' appended to the name.
1275 *) Generate muti valued AVAs using '+' notation in config files for
1279 *) Support for nameConstraints certificate extension.
1282 *) Support for policyConstraints certificate extension.
1285 *) Support for policyMappings certificate extension.
1288 *) Make sure the default DSA_METHOD implementation only uses its
1289 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1290 and change its own handlers to be NULL so as to remove unnecessary
1291 indirection. This lets alternative implementations fallback to the
1292 default implementation more easily.
1295 *) Support for directoryName in GeneralName related extensions
1299 *) Make it possible to link applications using Makefile.shared.
1300 Make that possible even when linking against static libraries!
1303 *) Support for single pass processing for S/MIME signing. This now
1304 means that S/MIME signing can be done from a pipe, in addition
1305 cleartext signing (multipart/signed type) is effectively streaming
1306 and the signed data does not need to be all held in memory.
1308 This is done with a new flag PKCS7_STREAM. When this flag is set
1309 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1310 is done after the data is output (and digests calculated) in
1311 SMIME_write_PKCS7().
1314 *) Add full support for -rpath/-R, both in shared libraries and
1315 applications, at least on the platforms where it's known how
1319 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1320 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1321 will now compute a table of multiples of the generator that
1322 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1323 faster (notably in the case of a single point multiplication,
1324 scalar * generator).
1325 [Nils Larsch, Bodo Moeller]
1327 *) IPv6 support for certificate extensions. The various extensions
1328 which use the IP:a.b.c.d can now take IPv6 addresses using the
1329 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1333 *) Added an ENGINE that implements RSA by performing private key
1334 exponentiations with the GMP library. The conversions to and from
1335 GMP's mpz_t format aren't optimised nor are any montgomery forms
1336 cached, and on x86 it appears OpenSSL's own performance has caught up.
1337 However there are likely to be other architectures where GMP could
1338 provide a boost. This ENGINE is not built in by default, but it can be
1339 specified at Configure time and should be accompanied by the necessary
1340 linker additions, eg;
1341 ./config -DOPENSSL_USE_GMP -lgmp
1344 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1345 testing availability of engines with "-t" - the old behaviour is
1346 produced by increasing the feature's verbosity with "-tt".
1349 *) ECDSA routines: under certain error conditions uninitialized BN objects
1350 could be freed. Solution: make sure initialization is performed early
1351 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1355 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1356 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1357 software implementations. For DSA and DH, parameter generation can
1358 also be overriden by providing the appropriate method callbacks.
1361 *) Change the "progress" mechanism used in key-generation and
1362 primality testing to functions that take a new BN_GENCB pointer in
1363 place of callback/argument pairs. The new API functions have "_ex"
1364 postfixes and the older functions are reimplemented as wrappers for
1365 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1366 declarations of the old functions to help (graceful) attempts to
1367 migrate to the new functions. Also, the new key-generation API
1368 functions operate on a caller-supplied key-structure and return
1369 success/failure rather than returning a key or NULL - this is to
1370 help make "keygen" another member function of RSA_METHOD etc.
1372 Example for using the new callback interface:
1374 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1378 BN_GENCB_set(&my_cb, my_callback, my_arg);
1380 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1381 /* For the meaning of a, b in calls to my_callback(), see the
1382 * documentation of the function that calls the callback.
1383 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1384 * my_callback should return 1 if it wants BN_is_prime_ex()
1385 * to continue, or 0 to stop.
1390 *) Change the ZLIB compression method to be stateful, and make it
1391 available to TLS with the number defined in
1392 draft-ietf-tls-compression-04.txt.
1395 *) Add the ASN.1 structures and functions for CertificatePair, which
1396 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1398 CertificatePair ::= SEQUENCE {
1399 forward [0] Certificate OPTIONAL,
1400 reverse [1] Certificate OPTIONAL,
1401 -- at least one of the pair shall be present -- }
1403 Also implement the PEM functions to read and write certificate
1404 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1406 This needed to be defined, mostly for the sake of the LDAP
1407 attribute crossCertificatePair, but may prove useful elsewhere as
1411 *) Make it possible to inhibit symlinking of shared libraries in
1412 Makefile.shared, for Cygwin's sake.
1415 *) Extend the BIGNUM API by creating a function
1416 void BN_set_negative(BIGNUM *a, int neg);
1417 and a macro that behave like
1418 int BN_is_negative(const BIGNUM *a);
1420 to avoid the need to access 'a->neg' directly in applications.
1423 *) Implement fast modular reduction for pseudo-Mersenne primes
1424 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1425 EC_GROUP_new_curve_GFp() will now automatically use this
1427 [Nils Larsch <nla@trustcenter.de>]
1429 *) Add new lock type (CRYPTO_LOCK_BN).
1432 *) Change the ENGINE framework to automatically load engines
1433 dynamically from specific directories unless they could be
1434 found to already be built in or loaded. Move all the
1435 current engines except for the cryptodev one to a new
1437 The engines in engines/ are built as shared libraries if
1438 the "shared" options was given to ./Configure or ./config.
1439 Otherwise, they are inserted in libcrypto.a.
1440 /usr/local/ssl/engines is the default directory for dynamic
1441 engines, but that can be overriden at configure time through
1442 the usual use of --prefix and/or --openssldir, and at run
1443 time with the environment variable OPENSSL_ENGINES.
1444 [Geoff Thorpe and Richard Levitte]
1446 *) Add Makefile.shared, a helper makefile to build shared
1447 libraries. Addapt Makefile.org.
1450 *) Add version info to Win32 DLLs.
1451 [Peter 'Luna' Runestig" <peter@runestig.com>]
1453 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1454 can be added using this API to created arbitrary PKCS#12
1455 files while avoiding the low level API.
1457 New options to PKCS12_create(), key or cert can be NULL and
1458 will then be omitted from the output file. The encryption
1459 algorithm NIDs can be set to -1 for no encryption, the mac
1460 iteration count can be set to 0 to omit the mac.
1462 Enhance pkcs12 utility by making the -nokeys and -nocerts
1463 options work when creating a PKCS#12 file. New option -nomac
1464 to omit the mac, NONE can be set for an encryption algorithm.
1465 New code is modified to use the enhanced PKCS12_create()
1466 instead of the low level API.
1469 *) Extend ASN1 encoder to support indefinite length constructed
1470 encoding. This can output sequences tags and octet strings in
1471 this form. Modify pk7_asn1.c to support indefinite length
1472 encoding. This is experimental and needs additional code to
1473 be useful, such as an ASN1 bio and some enhanced streaming
1476 Extend template encode functionality so that tagging is passed
1477 down to the template encoder.
1480 *) Let 'openssl req' fail if an argument to '-newkey' is not
1481 recognized instead of using RSA as a default.
1484 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1485 As these are not official, they are not included in "ALL";
1486 the "ECCdraft" ciphersuite group alias can be used to select them.
1487 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1489 *) Add ECDH engine support.
1490 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1492 *) Add ECDH in new directory crypto/ecdh/.
1493 [Douglas Stebila (Sun Microsystems Laboratories)]
1495 *) Let BN_rand_range() abort with an error after 100 iterations
1496 without success (which indicates a broken PRNG).
1499 *) Change BN_mod_sqrt() so that it verifies that the input value
1500 is really the square of the return value. (Previously,
1501 BN_mod_sqrt would show GIGO behaviour.)
1504 *) Add named elliptic curves over binary fields from X9.62, SECG,
1505 and WAP/WTLS; add OIDs that were still missing.
1507 [Sheueling Chang Shantz and Douglas Stebila
1508 (Sun Microsystems Laboratories)]
1510 *) Extend the EC library for elliptic curves over binary fields
1511 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1514 EC_GF2m_simple_method
1518 EC_GROUP_new_curve_GF2m
1519 EC_GROUP_set_curve_GF2m
1520 EC_GROUP_get_curve_GF2m
1521 EC_POINT_set_affine_coordinates_GF2m
1522 EC_POINT_get_affine_coordinates_GF2m
1523 EC_POINT_set_compressed_coordinates_GF2m
1525 Point compression for binary fields is disabled by default for
1526 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1529 As binary polynomials are represented as BIGNUMs, various members
1530 of the EC_GROUP and EC_POINT data structures can be shared
1531 between the implementations for prime fields and binary fields;
1532 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1533 are essentially identical to their ..._GFp counterparts.
1534 (For simplicity, the '..._GFp' prefix has been dropped from
1535 various internal method names.)
1537 An internal 'field_div' method (similar to 'field_mul' and
1538 'field_sqr') has been added; this is used only for binary fields.
1540 [Sheueling Chang Shantz and Douglas Stebila
1541 (Sun Microsystems Laboratories)]
1543 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1544 through methods ('mul', 'precompute_mult').
1546 The generic implementations (now internally called 'ec_wNAF_mul'
1547 and 'ec_wNAF_precomputed_mult') remain the default if these
1548 methods are undefined.
1550 [Sheueling Chang Shantz and Douglas Stebila
1551 (Sun Microsystems Laboratories)]
1553 *) New function EC_GROUP_get_degree, which is defined through
1554 EC_METHOD. For curves over prime fields, this returns the bit
1555 length of the modulus.
1557 [Sheueling Chang Shantz and Douglas Stebila
1558 (Sun Microsystems Laboratories)]
1560 *) New functions EC_GROUP_dup, EC_POINT_dup.
1561 (These simply call ..._new and ..._copy).
1563 [Sheueling Chang Shantz and Douglas Stebila
1564 (Sun Microsystems Laboratories)]
1566 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1567 Polynomials are represented as BIGNUMs (where the sign bit is not
1568 used) in the following functions [macros]:
1571 BN_GF2m_sub [= BN_GF2m_add]
1572 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1573 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1574 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1576 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1577 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1578 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1579 BN_GF2m_cmp [= BN_ucmp]
1581 (Note that only the 'mod' functions are actually for fields GF(2^m).
1582 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1584 For some functions, an the irreducible polynomial defining a
1585 field can be given as an 'unsigned int[]' with strictly
1586 decreasing elements giving the indices of those bits that are set;
1587 i.e., p[] represents the polynomial
1588 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1590 p[0] > p[1] > ... > p[k] = 0.
1591 This applies to the following functions:
1596 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1597 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1599 BN_GF2m_mod_sqrt_arr
1600 BN_GF2m_mod_solve_quad_arr
1604 Conversion can be performed by the following functions:
1609 bntest.c has additional tests for binary polynomial arithmetic.
1611 Two implementations for BN_GF2m_mod_div() are available.
1612 The default algorithm simply uses BN_GF2m_mod_inv() and
1613 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1614 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1615 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1617 [Sheueling Chang Shantz and Douglas Stebila
1618 (Sun Microsystems Laboratories)]
1620 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1621 functionality is disabled at compile-time.
1622 [Douglas Stebila <douglas.stebila@sun.com>]
1624 *) Change default behaviour of 'openssl asn1parse' so that more
1625 information is visible when viewing, e.g., a certificate:
1627 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1628 mode the content of non-printable OCTET STRINGs is output in a
1629 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1630 avoid the appearance of a printable string.
1631 [Nils Larsch <nla@trustcenter.de>]
1633 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1635 EC_GROUP_set_asn1_flag()
1636 EC_GROUP_get_asn1_flag()
1637 EC_GROUP_set_point_conversion_form()
1638 EC_GROUP_get_point_conversion_form()
1639 These control ASN1 encoding details:
1640 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1641 has been set to OPENSSL_EC_NAMED_CURVE.
1642 - Points are encoded in uncompressed form by default; options for
1643 asn1_for are as for point2oct, namely
1644 POINT_CONVERSION_COMPRESSED
1645 POINT_CONVERSION_UNCOMPRESSED
1646 POINT_CONVERSION_HYBRID
1648 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1651 EC_GROUP_get0_seed()
1652 EC_GROUP_get_seed_len()
1653 This is used only for ASN1 purposes (so far).
1654 [Nils Larsch <nla@trustcenter.de>]
1656 *) Add 'field_type' member to EC_METHOD, which holds the NID
1657 of the appropriate field type OID. The new function
1658 EC_METHOD_get_field_type() returns this value.
1659 [Nils Larsch <nla@trustcenter.de>]
1664 EC_POINT_point2hex()
1665 EC_POINT_hex2point()
1666 providing useful interfaces to EC_POINT_point2oct() and
1667 EC_POINT_oct2point().
1668 [Nils Larsch <nla@trustcenter.de>]
1670 *) Change internals of the EC library so that the functions
1671 EC_GROUP_set_generator()
1672 EC_GROUP_get_generator()
1673 EC_GROUP_get_order()
1674 EC_GROUP_get_cofactor()
1675 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1676 to methods, which would lead to unnecessary code duplication when
1677 adding different types of curves.
1678 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1680 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1681 arithmetic, and such that modified wNAFs are generated
1682 (which avoid length expansion in many cases).
1685 *) Add a function EC_GROUP_check_discriminant() (defined via
1686 EC_METHOD) that verifies that the curve discriminant is non-zero.
1688 Add a function EC_GROUP_check() that makes some sanity tests
1689 on a EC_GROUP, its generator and order. This includes
1690 EC_GROUP_check_discriminant().
1691 [Nils Larsch <nla@trustcenter.de>]
1693 *) Add ECDSA in new directory crypto/ecdsa/.
1695 Add applications 'openssl ecparam' and 'openssl ecdsa'
1696 (these are based on 'openssl dsaparam' and 'openssl dsa').
1698 ECDSA support is also included in various other files across the
1699 library. Most notably,
1700 - 'openssl req' now has a '-newkey ecdsa:file' option;
1701 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1702 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1703 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1704 them suitable for ECDSA where domain parameters must be
1705 extracted before the specific public key;
1706 - ECDSA engine support has been added.
1707 [Nils Larsch <nla@trustcenter.de>]
1709 *) Include some named elliptic curves, and add OIDs from X9.62,
1710 SECG, and WAP/WTLS. Each curve can be obtained from the new
1712 EC_GROUP_new_by_curve_name(),
1713 and the list of available named curves can be obtained with
1714 EC_get_builtin_curves().
1715 Also add a 'curve_name' member to EC_GROUP objects, which can be
1717 EC_GROUP_set_curve_name()
1718 EC_GROUP_get_curve_name()
1719 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1721 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1722 was actually never needed) and in BN_mul(). The removal in BN_mul()
1723 required a small change in bn_mul_part_recursive() and the addition
1724 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1725 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1726 bn_sub_words() and bn_add_words() except they take arrays with
1730 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1732 *) Cleanse PEM buffers before freeing them since they may contain
1734 [Benjamin Bennett <ben@psc.edu>]
1736 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1737 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1738 authentication-only ciphersuites.
1741 *) Since AES128 and AES256 share a single mask bit in the logic of
1742 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1743 kludge to work properly if AES128 is available and AES256 isn't.
1746 *) Expand security boundary to match 1.1.1 module.
1749 *) Remove redundant features: hash file source, editing of test vectors
1750 modify fipsld to use external fips_premain.c signature.
1753 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1754 run algorithm test programs.
1757 *) Make algorithm test programs more tolerant of whitespace.
1760 *) Have SSL/TLS server implementation tolerate "mismatched" record
1761 protocol version while receiving ClientHello even if the
1762 ClientHello is fragmented. (The server can't insist on the
1763 particular protocol version it has chosen before the ServerHello
1764 message has informed the client about his choice.)
1767 *) Load error codes if they are not already present instead of using a
1768 static variable. This allows them to be cleanly unloaded and reloaded.
1771 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1773 *) Introduce limits to prevent malicious keys being able to
1774 cause a denial of service. (CVE-2006-2940)
1775 [Steve Henson, Bodo Moeller]
1777 *) Fix ASN.1 parsing of certain invalid structures that can result
1778 in a denial of service. (CVE-2006-2937) [Steve Henson]
1780 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1781 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1783 *) Fix SSL client code which could crash if connecting to a
1784 malicious SSLv2 server. (CVE-2006-4343)
1785 [Tavis Ormandy and Will Drewry, Google Security Team]
1787 *) Change ciphersuite string processing so that an explicit
1788 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1789 will no longer include "AES128-SHA"), and any other similar
1790 ciphersuite (same bitmap) from *other* protocol versions (so that
1791 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1792 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1793 changes from 0.9.8b and 0.9.8d.
1796 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1798 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1799 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1801 *) Change the Unix randomness entropy gathering to use poll() when
1802 possible instead of select(), since the latter has some
1803 undesirable limitations.
1804 [Darryl Miles via Richard Levitte and Bodo Moeller]
1806 *) Disable rogue ciphersuites:
1808 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1809 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1810 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1812 The latter two were purportedly from
1813 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1816 Also deactive the remaining ciphersuites from
1817 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1818 unofficial, and the ID has long expired.
1821 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1822 dual-core machines) and other potential thread-safety issues.
1825 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1827 *) Adapt fipsld and the build system to link against the validated FIPS
1828 module in FIPS mode.
1831 *) Fixes for VC++ 2005 build under Windows.
1834 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1835 from a Windows bash shell such as MSYS. It is autodetected from the
1836 "config" script when run from a VC++ environment. Modify standard VC++
1837 build to use fipscanister.o from the GNU make build.
1840 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1842 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1843 The value now differs depending on if you build for FIPS or not.
1844 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1845 safely run with a non-FIPSed libcrypto, as it may crash because of
1846 the difference induced by this change.
1849 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1851 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1852 (part of SSL_OP_ALL). This option used to disable the
1853 countermeasure against man-in-the-middle protocol-version
1854 rollback in the SSL 2.0 server implementation, which is a bad
1855 idea. (CVE-2005-2969)
1857 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1858 for Information Security, National Institute of Advanced Industrial
1859 Science and Technology [AIST], Japan)]
1861 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1862 mainly for FIPS compliance and not fully integrated at this stage.
1865 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1866 the exponentiation using a fixed-length exponent. (Otherwise,
1867 the information leaked through timing could expose the secret key
1868 after many signatures; cf. Bleichenbacher's attack on DSA with
1872 *) Make a new fixed-window mod_exp implementation the default for
1873 RSA, DSA, and DH private-key operations so that the sequence of
1874 squares and multiplies and the memory access pattern are
1875 independent of the particular secret key. This will mitigate
1876 cache-timing and potential related attacks.
1878 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1879 and this is automatically used by BN_mod_exp_mont() if the new flag
1880 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1881 will use this BN flag for private exponents unless the flag
1882 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1883 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1885 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1887 *) Change the client implementation for SSLv23_method() and
1888 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1889 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1890 (Previously, the SSL 2.0 backwards compatible Client Hello
1891 message format would be used even with SSL_OP_NO_SSLv2.)
1894 *) Add support for smime-type MIME parameter in S/MIME messages which some
1898 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1899 a threadsafe manner. Modify rsa code to use new function and add calls
1900 to dsa and dh code (which had race conditions before).
1903 *) Include the fixed error library code in the C error file definitions
1904 instead of fixing them up at runtime. This keeps the error code
1905 structures constant.
1908 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1910 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1913 *) Fixes for newer kerberos headers. NB: the casts are needed because
1914 the 'length' field is signed on one version and unsigned on another
1915 with no (?) obvious way to tell the difference, without these VC++
1916 complains. Also the "definition" of FAR (blank) is no longer included
1917 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1918 some needed definitions.
1921 *) Undo Cygwin change.
1924 *) Added support for proxy certificates according to RFC 3820.
1925 Because they may be a security thread to unaware applications,
1926 they must be explicitely allowed in run-time. See
1927 docs/HOWTO/proxy_certificates.txt for further information.
1930 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1932 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1933 server and client random values. Previously
1934 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1935 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1937 This change has negligible security impact because:
1939 1. Server and client random values still have 24 bytes of pseudo random
1942 2. Server and client random values are sent in the clear in the initial
1945 3. The master secret is derived using the premaster secret (48 bytes in
1946 size for static RSA ciphersuites) as well as client server and random
1949 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1952 [Stephen Henson, reported by UK NISCC]
1954 *) Use Windows randomness collection on Cygwin.
1957 *) Fix hang in EGD/PRNGD query when communication socket is closed
1958 prematurely by EGD/PRNGD.
1959 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1961 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1964 *) Back-port of selected performance improvements from development
1965 branch, as well as improved support for PowerPC platforms.
1968 *) Add lots of checks for memory allocation failure, error codes to indicate
1969 failure and freeing up memory if a failure occurs.
1970 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1972 *) Add new -passin argument to dgst.
1975 *) Perform some character comparisons of different types in X509_NAME_cmp:
1976 this is needed for some certificates that reencode DNs into UTF8Strings
1977 (in violation of RFC3280) and can't or wont issue name rollover
1981 *) Make an explicit check during certificate validation to see that
1982 the CA setting in each certificate on the chain is correct. As a
1983 side effect always do the following basic checks on extensions,
1984 not just when there's an associated purpose to the check:
1986 - if there is an unhandled critical extension (unless the user
1987 has chosen to ignore this fault)
1988 - if the path length has been exceeded (if one is set at all)
1989 - that certain extensions fit the associated purpose (if one has
1993 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1995 *) Avoid a race condition when CRLs are checked in a multi threaded
1996 environment. This would happen due to the reordering of the revoked
1997 entries during signature checking and serial number lookup. Now the
1998 encoding is cached and the serial number sort performed under a lock.
1999 Add new STACK function sk_is_sorted().
2002 *) Add Delta CRL to the extension code.
2005 *) Various fixes to s3_pkt.c so alerts are sent properly.
2006 [David Holmes <d.holmes@f5.com>]
2008 *) Reduce the chances of duplicate issuer name and serial numbers (in
2009 violation of RFC3280) using the OpenSSL certificate creation utilities.
2010 This is done by creating a random 64 bit value for the initial serial
2011 number when a serial number file is created or when a self signed
2012 certificate is created using 'openssl req -x509'. The initial serial
2013 number file is created using 'openssl x509 -next_serial' in CA.pl
2014 rather than being initialized to 1.
2017 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2019 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2020 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2021 [Joe Orton, Steve Henson]
2023 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2025 [Joe Orton, Steve Henson]
2027 *) Make it possible to have multiple active certificates with the same
2028 subject in the CA index file. This is done only if the keyword
2029 'unique_subject' is set to 'no' in the main CA section (default
2030 if 'CA_default') of the configuration file. The value is saved
2031 with the database itself in a separate index attribute file,
2032 named like the index file with '.attr' appended to the name.
2035 *) X509 verify fixes. Disable broken certificate workarounds when
2036 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2037 keyUsage extension present. Don't accept CRLs with unhandled critical
2038 extensions: since verify currently doesn't process CRL extensions this
2039 rejects a CRL with *any* critical extensions. Add new verify error codes
2043 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2044 A clarification of RFC2560 will require the use of OCTET STRINGs and
2045 some implementations cannot handle the current raw format. Since OpenSSL
2046 copies and compares OCSP nonces as opaque blobs without any attempt at
2047 parsing them this should not create any compatibility issues.
2050 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2051 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2052 this HMAC (and other) operations are several times slower than OpenSSL
2056 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2057 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2059 *) Use the correct content when signing type "other".
2062 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2064 *) Fix various bugs revealed by running the NISCC test suite:
2066 Stop out of bounds reads in the ASN1 code when presented with
2067 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2069 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2071 If verify callback ignores invalid public key errors don't try to check
2072 certificate signature with the NULL public key.
2076 *) New -ignore_err option in ocsp application to stop the server
2077 exiting on the first error in a request.
2080 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2081 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2085 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2086 extra data after the compression methods not only for TLS 1.0
2087 but also for SSL 3.0 (as required by the specification).
2088 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2090 *) Change X509_certificate_type() to mark the key as exported/exportable
2091 when it's 512 *bits* long, not 512 bytes.
2094 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2095 blocks during encryption.
2098 *) Various fixes to base64 BIO and non blocking I/O. On write
2099 flushes were not handled properly if the BIO retried. On read
2100 data was not being buffered properly and had various logic bugs.
2101 This also affects blocking I/O when the data being decoded is a
2105 *) Various S/MIME bugfixes and compatibility changes:
2106 output correct application/pkcs7 MIME type if
2107 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2108 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2109 of files as .eml work). Correctly handle very long lines in MIME
2113 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2115 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2116 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2117 a protocol version number mismatch like a decryption error
2118 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2121 *) Turn on RSA blinding by default in the default implementation
2122 to avoid a timing attack. Applications that don't want it can call
2123 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2124 They would be ill-advised to do so in most cases.
2125 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2127 *) Change RSA blinding code so that it works when the PRNG is not
2128 seeded (in this case, the secret RSA exponent is abused as
2129 an unpredictable seed -- if it is not unpredictable, there
2130 is no point in blinding anyway). Make RSA blinding thread-safe
2131 by remembering the creator's thread ID in rsa->blinding and
2132 having all other threads use local one-time blinding factors
2133 (this requires more computation than sharing rsa->blinding, but
2134 avoids excessive locking; and if an RSA object is not shared
2135 between threads, blinding will still be very fast).
2138 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2139 ENGINE as defaults for all supported algorithms irrespective of
2140 the 'flags' parameter. 'flags' is now honoured, so applications
2141 should make sure they are passing it correctly.
2144 *) Target "mingw" now allows native Windows code to be generated in
2145 the Cygwin environment as well as with the MinGW compiler.
2148 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2150 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2151 via timing by performing a MAC computation even if incorrrect
2152 block cipher padding has been found. This is a countermeasure
2153 against active attacks where the attacker has to distinguish
2154 between bad padding and a MAC verification error. (CVE-2003-0078)
2156 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2157 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2158 Martin Vuagnoux (EPFL, Ilion)]
2160 *) Make the no-err option work as intended. The intention with no-err
2161 is not to have the whole error stack handling routines removed from
2162 libcrypto, it's only intended to remove all the function name and
2163 reason texts, thereby removing some of the footprint that may not
2164 be interesting if those errors aren't displayed anyway.
2166 NOTE: it's still possible for any application or module to have it's
2167 own set of error texts inserted. The routines are there, just not
2168 used by default when no-err is given.
2171 *) Add support for FreeBSD on IA64.
2172 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2174 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2175 Kerberos function mit_des_cbc_cksum(). Before this change,
2176 the value returned by DES_cbc_cksum() was like the one from
2177 mit_des_cbc_cksum(), except the bytes were swapped.
2178 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2180 *) Allow an application to disable the automatic SSL chain building.
2181 Before this a rather primitive chain build was always performed in
2182 ssl3_output_cert_chain(): an application had no way to send the
2183 correct chain if the automatic operation produced an incorrect result.
2185 Now the chain builder is disabled if either:
2187 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2189 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2191 The reasoning behind this is that an application would not want the
2192 auto chain building to take place if extra chain certificates are
2193 present and it might also want a means of sending no additional
2194 certificates (for example the chain has two certificates and the
2198 *) Add the possibility to build without the ENGINE framework.
2199 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2201 *) Under Win32 gmtime() can return NULL: check return value in
2202 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2205 *) DSA routines: under certain error conditions uninitialized BN objects
2206 could be freed. Solution: make sure initialization is performed early
2207 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2208 Nils Larsch <nla@trustcenter.de> via PR#459)
2211 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2212 checked on reconnect on the client side, therefore session resumption
2213 could still fail with a "ssl session id is different" error. This
2214 behaviour is masked when SSL_OP_ALL is used due to
2215 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2216 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2217 followup to PR #377.
2220 *) IA-32 assembler support enhancements: unified ELF targets, support
2221 for SCO/Caldera platforms, fix for Cygwin shared build.
2224 *) Add support for FreeBSD on sparc64. As a consequence, support for
2225 FreeBSD on non-x86 processors is separate from x86 processors on
2226 the config script, much like the NetBSD support.
2227 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2229 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2231 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2234 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2235 code (06) was taken as the first octet of the session ID and the last
2236 octet was ignored consequently. As a result SSLv2 client side session
2237 caching could not have worked due to the session ID mismatch between
2239 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2243 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2244 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2248 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2249 seems that in spite of existing for more than a year, many application
2250 author have done nothing to provide the necessary callbacks, which
2251 means that this particular engine will not work properly anywhere.
2252 This is a very unfortunate situation which forces us, in the name
2253 of usability, to give the hw_ncipher.c a static lock, which is part
2255 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2256 appear in 0.9.8 or later. We EXPECT application authors to have
2257 dealt properly with this when 0.9.8 is released (unless we actually
2258 make such changes in the libcrypto locking code that changes will
2259 have to be made anyway).
2262 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2263 octets have been read, EOF or an error occurs. Without this change
2264 some truncated ASN1 structures will not produce an error.
2267 *) Disable Heimdal support, since it hasn't been fully implemented.
2268 Still give the possibility to force the use of Heimdal, but with
2269 warnings and a request that patches get sent to openssl-dev.
2272 *) Add the VC-CE target, introduce the WINCE sysname, and add
2273 INSTALL.WCE and appropriate conditionals to make it build.
2274 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2276 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2277 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2278 edit numbers of the version.
2279 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2281 *) Introduce safe string copy and catenation functions
2282 (BUF_strlcpy() and BUF_strlcat()).
2283 [Ben Laurie (CHATS) and Richard Levitte]
2285 *) Avoid using fixed-size buffers for one-line DNs.
2286 [Ben Laurie (CHATS)]
2288 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2289 resizing buffers containing secrets, and use where appropriate.
2290 [Ben Laurie (CHATS)]
2292 *) Avoid using fixed size buffers for configuration file location.
2293 [Ben Laurie (CHATS)]
2295 *) Avoid filename truncation for various CA files.
2296 [Ben Laurie (CHATS)]
2298 *) Use sizeof in preference to magic numbers.
2299 [Ben Laurie (CHATS)]
2301 *) Avoid filename truncation in cert requests.
2302 [Ben Laurie (CHATS)]
2304 *) Add assertions to check for (supposedly impossible) buffer
2306 [Ben Laurie (CHATS)]
2308 *) Don't cache truncated DNS entries in the local cache (this could
2309 potentially lead to a spoofing attack).
2310 [Ben Laurie (CHATS)]
2312 *) Fix various buffers to be large enough for hex/decimal
2313 representations in a platform independent manner.
2314 [Ben Laurie (CHATS)]
2316 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2317 resizing buffers containing secrets, and use where appropriate.
2318 [Ben Laurie (CHATS)]
2320 *) Add BIO_indent() to avoid much slightly worrying code to do
2322 [Ben Laurie (CHATS)]
2324 *) Convert sprintf()/BIO_puts() to BIO_printf().
2325 [Ben Laurie (CHATS)]
2327 *) buffer_gets() could terminate with the buffer only half
2329 [Ben Laurie (CHATS)]
2331 *) Add assertions to prevent user-supplied crypto functions from
2332 overflowing internal buffers by having large block sizes, etc.
2333 [Ben Laurie (CHATS)]
2335 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2337 [Ben Laurie (CHATS)]
2339 *) Eliminate unused copy of key in RC4.
2340 [Ben Laurie (CHATS)]
2342 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2343 [Ben Laurie (CHATS)]
2345 *) Fix off-by-one error in EGD path.
2346 [Ben Laurie (CHATS)]
2348 *) If RANDFILE path is too long, ignore instead of truncating.
2349 [Ben Laurie (CHATS)]
2351 *) Eliminate unused and incorrectly sized X.509 structure
2353 [Ben Laurie (CHATS)]
2355 *) Eliminate unused and dangerous function knumber().
2356 [Ben Laurie (CHATS)]
2358 *) Eliminate unused and dangerous structure, KSSL_ERR.
2359 [Ben Laurie (CHATS)]
2361 *) Protect against overlong session ID context length in an encoded
2362 session object. Since these are local, this does not appear to be
2364 [Ben Laurie (CHATS)]
2366 *) Change from security patch (see 0.9.6e below) that did not affect
2367 the 0.9.6 release series:
2369 Remote buffer overflow in SSL3 protocol - an attacker could
2370 supply an oversized master key in Kerberos-enabled versions.
2372 [Ben Laurie (CHATS)]
2374 *) Change the SSL kerb5 codes to match RFC 2712.
2377 *) Make -nameopt work fully for req and add -reqopt switch.
2378 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2380 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2381 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2383 *) Make sure tests can be performed even if the corresponding algorithms
2384 have been removed entirely. This was also the last step to make
2385 OpenSSL compilable with DJGPP under all reasonable conditions.
2386 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2388 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2389 to allow version independent disabling of normally unselected ciphers,
2390 which may be activated as a side-effect of selecting a single cipher.
2392 (E.g., cipher list string "RSA" enables ciphersuites that are left
2393 out of "ALL" because they do not provide symmetric encryption.
2394 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2395 [Lutz Jaenicke, Bodo Moeller]
2397 *) Add appropriate support for separate platform-dependent build
2398 directories. The recommended way to make a platform-dependent
2399 build directory is the following (tested on Linux), maybe with
2402 # Place yourself outside of the OpenSSL source tree. In
2403 # this example, the environment variable OPENSSL_SOURCE
2404 # is assumed to contain the absolute OpenSSL source directory.
2405 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2406 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2407 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2408 mkdir -p `dirname $F`
2409 ln -s $OPENSSL_SOURCE/$F $F
2412 To be absolutely sure not to disturb the source tree, a "make clean"
2413 is a good thing. If it isn't successfull, don't worry about it,
2414 it probably means the source directory is very clean.
2417 *) Make sure any ENGINE control commands make local copies of string
2418 pointers passed to them whenever necessary. Otherwise it is possible
2419 the caller may have overwritten (or deallocated) the original string
2420 data when a later ENGINE operation tries to use the stored values.
2421 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2423 *) Improve diagnostics in file reading and command-line digests.
2424 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2426 *) Add AES modes CFB and OFB to the object database. Correct an
2427 error in AES-CFB decryption.
2430 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2431 allows existing EVP_CIPHER_CTX structures to be reused after
2432 calling EVP_*Final(). This behaviour is used by encryption
2433 BIOs and some applications. This has the side effect that
2434 applications must explicitly clean up cipher contexts with
2435 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2438 *) Check the values of dna and dnb in bn_mul_recursive before calling
2439 bn_mul_comba (a non zero value means the a or b arrays do not contain
2440 n2 elements) and fallback to bn_mul_normal if either is not zero.
2443 *) Fix escaping of non-ASCII characters when using the -subj option
2444 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2447 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2448 form for "surname", serialNumber has no short form.
2449 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2450 therefore remove "mail" short name for "internet 7".
2451 The OID for unique identifiers in X509 certificates is
2452 x500UniqueIdentifier, not uniqueIdentifier.
2453 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2456 *) Add an "init" command to the ENGINE config module and auto initialize
2457 ENGINEs. Without any "init" command the ENGINE will be initialized
2458 after all ctrl commands have been executed on it. If init=1 the
2459 ENGINE is initailized at that point (ctrls before that point are run
2460 on the uninitialized ENGINE and after on the initialized one). If
2461 init=0 then the ENGINE will not be iniatialized at all.
2464 *) Fix the 'app_verify_callback' interface so that the user-defined
2465 argument is actually passed to the callback: In the
2466 SSL_CTX_set_cert_verify_callback() prototype, the callback
2467 declaration has been changed from
2470 int (*cb)(X509_STORE_CTX *,void *);
2471 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2472 i=s->ctx->app_verify_callback(&ctx)
2473 has been changed into
2474 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2476 To update applications using SSL_CTX_set_cert_verify_callback(),
2477 a dummy argument can be added to their callback functions.
2478 [D. K. Smetters <smetters@parc.xerox.com>]
2480 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2481 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2483 *) Add and OPENSSL_LOAD_CONF define which will cause
2484 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2485 This allows older applications to transparently support certain
2486 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2487 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2488 load the config file and OPENSSL_add_all_algorithms_conf() which will
2489 always load it have also been added.
2492 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2493 Adjust NIDs and EVP layer.
2494 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2496 *) Config modules support in openssl utility.
2498 Most commands now load modules from the config file,
2499 though in a few (such as version) this isn't done
2500 because it couldn't be used for anything.
2502 In the case of ca and req the config file used is
2503 the same as the utility itself: that is the -config
2504 command line option can be used to specify an
2508 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2509 use "openssl_conf" if filename is NULL use default openssl config file.
2512 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2513 config section name. Add a new flag to tolerate a missing config file
2514 and move code to CONF_modules_load_file().
2517 *) Support for crypto accelerator cards from Accelerated Encryption
2518 Processing, www.aep.ie. (Use engine 'aep')
2519 The support was copied from 0.9.6c [engine] and adapted/corrected
2520 to work with the new engine framework.
2521 [AEP Inc. and Richard Levitte]
2523 *) Support for SureWare crypto accelerator cards from Baltimore
2524 Technologies. (Use engine 'sureware')
2525 The support was copied from 0.9.6c [engine] and adapted
2526 to work with the new engine framework.
2529 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2530 make the newer ENGINE framework commands for the CHIL engine work.
2531 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2533 *) Make it possible to produce shared libraries on ReliantUNIX.
2534 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2536 *) Add the configuration target debug-linux-ppro.
2537 Make 'openssl rsa' use the general key loading routines
2538 implemented in apps.c, and make those routines able to
2539 handle the key format FORMAT_NETSCAPE and the variant
2541 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2543 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2544 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2546 *) Add -keyform to rsautl, and document -engine.
2547 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2549 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2550 BIO_R_NO_SUCH_FILE error code rather than the generic
2551 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2554 *) Add new functions
2556 ERR_peek_last_error_line
2557 ERR_peek_last_error_line_data.
2558 These are similar to
2561 ERR_peek_error_line_data,
2562 but report on the latest error recorded rather than the first one
2563 still in the error queue.
2564 [Ben Laurie, Bodo Moeller]
2566 *) default_algorithms option in ENGINE config module. This allows things
2568 default_algorithms = ALL
2569 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2572 *) Prelminary ENGINE config module.
2575 *) New experimental application configuration code.
2578 *) Change the AES code to follow the same name structure as all other
2579 symmetric ciphers, and behave the same way. Move everything to
2580 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2581 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2583 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2584 [Ben Laurie and Theo de Raadt]
2586 *) Add option to output public keys in req command.
2587 [Massimiliano Pala madwolf@openca.org]
2589 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2590 (up to about 10% better than before for P-192 and P-224).
2593 *) New functions/macros
2595 SSL_CTX_set_msg_callback(ctx, cb)
2596 SSL_CTX_set_msg_callback_arg(ctx, arg)
2597 SSL_set_msg_callback(ssl, cb)
2598 SSL_set_msg_callback_arg(ssl, arg)
2600 to request calling a callback function
2602 void cb(int write_p, int version, int content_type,
2603 const void *buf, size_t len, SSL *ssl, void *arg)
2605 whenever a protocol message has been completely received
2606 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2607 protocol version according to which the SSL library interprets
2608 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2609 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2610 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2611 specification (change_cipher_spec(20), alert(21), handshake(22)).
2612 'buf' and 'len' point to the actual message, 'ssl' to the
2613 SSL object, and 'arg' is the application-defined value set by
2614 SSL[_CTX]_set_msg_callback_arg().
2616 'openssl s_client' and 'openssl s_server' have new '-msg' options
2617 to enable a callback that displays all protocol messages.
2620 *) Change the shared library support so shared libraries are built as
2621 soon as the corresponding static library is finished, and thereby get
2622 openssl and the test programs linked against the shared library.
2623 This still only happens when the keyword "shard" has been given to
2624 the configuration scripts.
2626 NOTE: shared library support is still an experimental thing, and
2627 backward binary compatibility is still not guaranteed.
2628 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2630 *) Add support for Subject Information Access extension.
2631 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2633 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2634 additional bytes when new memory had to be allocated, not just
2635 when reusing an existing buffer.
2638 *) New command line and configuration option 'utf8' for the req command.
2639 This allows field values to be specified as UTF8 strings.
2642 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2643 runs for the former and machine-readable output for the latter.
2646 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2647 of the e-mail address in the DN (i.e., it will go into a certificate
2648 extension only). The new configuration file option 'email_in_dn = no'
2649 has the same effect.
2650 [Massimiliano Pala madwolf@openca.org]
2652 *) Change all functions with names starting with des_ to be starting
2653 with DES_ instead. Add wrappers that are compatible with libdes,
2654 but are named _ossl_old_des_*. Finally, add macros that map the
2655 des_* symbols to the corresponding _ossl_old_des_* if libdes
2656 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2657 desired, the des_* symbols will be mapped to DES_*, with one
2660 Since we provide two compatibility mappings, the user needs to
2661 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2662 compatibility is desired. The default (i.e., when that macro
2663 isn't defined) is OpenSSL 0.9.6c compatibility.
2665 There are also macros that enable and disable the support of old
2666 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2667 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2668 are defined, the default will apply: to support the old des routines.
2670 In either case, one must include openssl/des.h to get the correct
2671 definitions. Do not try to just include openssl/des_old.h, that
2674 NOTE: This is a major break of an old API into a new one. Software
2675 authors are encouraged to switch to the DES_ style functions. Some
2676 time in the future, des_old.h and the libdes compatibility functions
2677 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2678 default), and then completely removed.
2681 *) Test for certificates which contain unsupported critical extensions.
2682 If such a certificate is found during a verify operation it is
2683 rejected by default: this behaviour can be overridden by either
2684 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2685 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2686 X509_supported_extension() has also been added which returns 1 if a
2687 particular extension is supported.
2690 *) Modify the behaviour of EVP cipher functions in similar way to digests
2691 to retain compatibility with existing code.
2694 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2695 compatibility with existing code. In particular the 'ctx' parameter does
2696 not have to be to be initialized before the call to EVP_DigestInit() and
2697 it is tidied up after a call to EVP_DigestFinal(). New function
2698 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2699 EVP_MD_CTX_copy() changed to not require the destination to be
2700 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2701 requires the destination to be valid.
2703 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2704 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2707 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2708 so that complete 'Handshake' protocol structures are kept in memory
2709 instead of overwriting 'msg_type' and 'length' with 'body' data.
2712 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2713 [Massimo Santin via Richard Levitte]
2715 *) Major restructuring to the underlying ENGINE code. This includes
2716 reduction of linker bloat, separation of pure "ENGINE" manipulation
2717 (initialisation, etc) from functionality dealing with implementations
2718 of specific crypto iterfaces. This change also introduces integrated
2719 support for symmetric ciphers and digest implementations - so ENGINEs
2720 can now accelerate these by providing EVP_CIPHER and EVP_MD
2721 implementations of their own. This is detailed in crypto/engine/README
2722 as it couldn't be adequately described here. However, there are a few
2723 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2724 were changed in the original introduction of ENGINE code have now
2725 reverted back - the hooking from this code to ENGINE is now a good
2726 deal more passive and at run-time, operations deal directly with
2727 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2728 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2729 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2730 they were not being used by the framework as there is no concept of a
2731 BIGNUM_METHOD and they could not be generalised to the new
2732 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2733 ENGINE_cpy() has been removed as it cannot be consistently defined in
2737 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2740 *) Change mkdef.pl to sort symbols that get the same entry number,
2741 and make sure the automatically generated functions ERR_load_*
2742 become part of libeay.num as well.
2745 *) New function SSL_renegotiate_pending(). This returns true once
2746 renegotiation has been requested (either SSL_renegotiate() call
2747 or HelloRequest/ClientHello receveived from the peer) and becomes
2748 false once a handshake has been completed.
2749 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2750 sends a HelloRequest, but does not ensure that a handshake takes
2751 place. SSL_renegotiate_pending() is useful for checking if the
2752 client has followed the request.)
2755 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2756 By default, clients may request session resumption even during
2757 renegotiation (if session ID contexts permit); with this option,
2758 session resumption is possible only in the first handshake.
2760 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2761 more bits available for options that should not be part of
2762 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2765 *) Add some demos for certificate and certificate request creation.
2768 *) Make maximum certificate chain size accepted from the peer application
2769 settable (SSL*_get/set_max_cert_list()), as proposed by
2770 "Douglas E. Engert" <deengert@anl.gov>.
2773 *) Add support for shared libraries for Unixware-7
2774 (Boyd Lynn Gerber <gerberb@zenez.com>).
2777 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2778 be done prior to destruction. Use this to unload error strings from
2779 ENGINEs that load their own error strings. NB: This adds two new API
2780 functions to "get" and "set" this destroy handler in an ENGINE.
2783 *) Alter all existing ENGINE implementations (except "openssl" and
2784 "openbsd") to dynamically instantiate their own error strings. This
2785 makes them more flexible to be built both as statically-linked ENGINEs
2786 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2787 Also, add stub code to each that makes building them as self-contained
2788 shared-libraries easier (see README.ENGINE).
2791 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2792 implementations into applications that are completely implemented in
2793 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2794 commands that can be used to configure what shared-library to load and
2795 to control aspects of the way it is handled. Also, made an update to
2796 the README.ENGINE file that brings its information up-to-date and
2797 provides some information and instructions on the "dynamic" ENGINE
2798 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2801 *) Make it possible to unload ranges of ERR strings with a new
2802 "ERR_unload_strings" function.
2805 *) Add a copy() function to EVP_MD.
2808 *) Make EVP_MD routines take a context pointer instead of just the
2809 md_data void pointer.
2812 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2813 that the digest can only process a single chunk of data
2814 (typically because it is provided by a piece of
2815 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2816 is only going to provide a single chunk of data, and hence the
2817 framework needn't accumulate the data for oneshot drivers.
2820 *) As with "ERR", make it possible to replace the underlying "ex_data"
2821 functions. This change also alters the storage and management of global
2822 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2823 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2824 index counters. The API functions that use this state have been changed
2825 to take a "class_index" rather than pointers to the class's local STACK
2826 and counter, and there is now an API function to dynamically create new
2827 classes. This centralisation allows us to (a) plug a lot of the
2828 thread-safety problems that existed, and (b) makes it possible to clean
2829 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2830 such data would previously have always leaked in application code and
2831 workarounds were in place to make the memory debugging turn a blind eye
2832 to it. Application code that doesn't use this new function will still
2833 leak as before, but their memory debugging output will announce it now
2834 rather than letting it slide.
2836 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2837 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2838 has a return value to indicate success or failure.
2841 *) Make it possible to replace the underlying "ERR" functions such that the
2842 global state (2 LHASH tables and 2 locks) is only used by the "default"
2843 implementation. This change also adds two functions to "get" and "set"
2844 the implementation prior to it being automatically set the first time
2845 any other ERR function takes place. Ie. an application can call "get",
2846 pass the return value to a module it has just loaded, and that module
2847 can call its own "set" function using that value. This means the
2848 module's "ERR" operations will use (and modify) the error state in the
2849 application and not in its own statically linked copy of OpenSSL code.
2852 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2853 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2854 the operation, and provides a more encapsulated way for external code
2855 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2856 to use these functions rather than manually incrementing the counts.
2858 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2861 *) Add EVP test program.
2864 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2867 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2868 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2869 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2870 These allow a CRL to be built without having to access X509_CRL fields
2871 directly. Modify 'ca' application to use new functions.
2874 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2875 bug workarounds. Rollback attack detection is a security feature.
2876 The problem will only arise on OpenSSL servers when TLSv1 is not
2877 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2878 Software authors not wanting to support TLSv1 will have special reasons
2879 for their choice and can explicitly enable this option.
2880 [Bodo Moeller, Lutz Jaenicke]
2882 *) Rationalise EVP so it can be extended: don't include a union of
2883 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2884 (similar to those existing for EVP_CIPHER_CTX).
2889 EVP_MD_CTX_init(&md); /* new function call */
2890 EVP_DigestInit(&md, EVP_sha1());
2891 EVP_DigestUpdate(&md, in, len);
2892 EVP_DigestFinal(&md, out, NULL);
2893 EVP_MD_CTX_cleanup(&md); /* new function call */
2897 *) Make DES key schedule conform to the usual scheme, as well as
2898 correcting its structure. This means that calls to DES functions
2899 now have to pass a pointer to a des_key_schedule instead of a
2900 plain des_key_schedule (which was actually always a pointer
2903 des_key_schedule ks;
2905 des_set_key_checked(..., &ks);
2906 des_ncbc_encrypt(..., &ks, ...);
2908 (Note that a later change renames 'des_...' into 'DES_...'.)
2911 *) Initial reduction of linker bloat: the use of some functions, such as
2912 PEM causes large amounts of unused functions to be linked in due to
2913 poor organisation. For example pem_all.c contains every PEM function
2914 which has a knock on effect of linking in large amounts of (unused)
2915 ASN1 code. Grouping together similar functions and splitting unrelated
2916 functions prevents this.
2919 *) Cleanup of EVP macros.
2922 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2923 correct _ecb suffix.
2926 *) Add initial OCSP responder support to ocsp application. The
2927 revocation information is handled using the text based index
2928 use by the ca application. The responder can either handle
2929 requests generated internally, supplied in files (for example
2930 via a CGI script) or using an internal minimal server.
2933 *) Add configuration choices to get zlib compression for TLS.
2936 *) Changes to Kerberos SSL for RFC 2712 compliance:
2937 1. Implemented real KerberosWrapper, instead of just using
2938 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2939 2. Implemented optional authenticator field of KerberosWrapper.
2941 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2942 and authenticator structs; see crypto/krb5/.
2944 Generalized Kerberos calls to support multiple Kerberos libraries.
2945 [Vern Staats <staatsvr@asc.hpc.mil>,
2946 Jeffrey Altman <jaltman@columbia.edu>
2947 via Richard Levitte]
2949 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2950 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2951 values for each of the key sizes rather than having just
2952 parameters (and 'speed' generating keys each time).
2955 *) Speed up EVP routines.
2958 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2959 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2960 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2961 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2963 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2964 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2965 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2968 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2970 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2973 *) Added the OS2-EMX target.
2974 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2976 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2977 to support NCONF routines in extension code. New function CONF_set_nconf()
2978 to allow functions which take an NCONF to also handle the old LHASH
2979 structure: this means that the old CONF compatible routines can be
2980 retained (in particular wrt extensions) without having to duplicate the
2981 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2984 *) Enhance the general user interface with mechanisms for inner control
2985 and with possibilities to have yes/no kind of prompts.
2988 *) Change all calls to low level digest routines in the library and
2989 applications to use EVP. Add missing calls to HMAC_cleanup() and
2990 don't assume HMAC_CTX can be copied using memcpy().
2991 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2993 *) Add the possibility to control engines through control names but with
2994 arbitrary arguments instead of just a string.
2995 Change the key loaders to take a UI_METHOD instead of a callback
2996 function pointer. NOTE: this breaks binary compatibility with earlier
2997 versions of OpenSSL [engine].
2998 Adapt the nCipher code for these new conditions and add a card insertion
3002 *) Enhance the general user interface with mechanisms to better support
3003 dialog box interfaces, application-defined prompts, the possibility
3004 to use defaults (for example default passwords from somewhere else)
3005 and interrupts/cancellations.
3008 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3009 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3012 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3013 tidy up some unnecessarily weird code in 'sk_new()').
3014 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3016 *) Change the key loading routines for ENGINEs to use the same kind
3017 callback (pem_password_cb) as all other routines that need this
3021 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3022 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3023 than this minimum value is recommended.
3026 *) New random seeder for OpenVMS, using the system process statistics
3027 that are easily reachable.
3030 *) Windows apparently can't transparently handle global
3031 variables defined in DLLs. Initialisations such as:
3033 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3035 wont compile. This is used by the any applications that need to
3036 declare their own ASN1 modules. This was fixed by adding the option
3037 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3038 needed for static libraries under Win32.
3041 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3042 setting of purpose and trust fields. New X509_STORE trust and
3043 purpose functions and tidy up setting in other SSL functions.
3046 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3047 structure. These are inherited by X509_STORE_CTX when it is
3048 initialised. This allows various defaults to be set in the
3049 X509_STORE structure (such as flags for CRL checking and custom
3050 purpose or trust settings) for functions which only use X509_STORE_CTX
3051 internally such as S/MIME.
3053 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3054 trust settings if they are not set in X509_STORE. This allows X509_STORE
3055 purposes and trust (in S/MIME for example) to override any set by default.
3057 Add command line options for CRL checking to smime, s_client and s_server
3061 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3062 are set then the CRL is looked up in the X509_STORE structure and
3063 its validity and signature checked, then if the certificate is found
3064 in the CRL the verify fails with a revoked error.
3066 Various new CRL related callbacks added to X509_STORE_CTX structure.
3068 Command line options added to 'verify' application to support this.
3070 This needs some additional work, such as being able to handle multiple
3071 CRLs with different times, extension based lookup (rather than just
3072 by subject name) and ultimately more complete V2 CRL extension
3076 *) Add a general user interface API (crypto/ui/). This is designed
3077 to replace things like des_read_password and friends (backward
3078 compatibility functions using this new API are provided).
3079 The purpose is to remove prompting functions from the DES code
3080 section as well as provide for prompting through dialog boxes in
3081 a window system and the like.
3084 *) Add "ex_data" support to ENGINE so implementations can add state at a
3085 per-structure level rather than having to store it globally.
3088 *) Make it possible for ENGINE structures to be copied when retrieved by
3089 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3090 This causes the "original" ENGINE structure to act like a template,
3091 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3092 operational state can be localised to each ENGINE structure, despite the
3093 fact they all share the same "methods". New ENGINE structures returned in
3094 this case have no functional references and the return value is the single
3095 structural reference. This matches the single structural reference returned
3096 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3100 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3101 needs to match any other type at all we need to manually clear the
3105 *) Changes to the "openssl engine" utility to include;
3106 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3107 about an ENGINE's available control commands.
3108 - executing control commands from command line arguments using the
3109 '-pre' and '-post' switches. '-post' is only used if '-t' is
3110 specified and the ENGINE is successfully initialised. The syntax for
3111 the individual commands are colon-separated, for example;
3112 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3115 *) New dynamic control command support for ENGINEs. ENGINEs can now
3116 declare their own commands (numbers), names (strings), descriptions,
3117 and input types for run-time discovery by calling applications. A
3118 subset of these commands are implicitly classed as "executable"
3119 depending on their input type, and only these can be invoked through
3120 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3121 can be based on user input, config files, etc). The distinction is
3122 that "executable" commands cannot return anything other than a boolean
3123 result and can only support numeric or string input, whereas some
3124 discoverable commands may only be for direct use through
3125 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3126 pointers, or other custom uses. The "executable" commands are to
3127 support parameterisations of ENGINE behaviour that can be
3128 unambiguously defined by ENGINEs and used consistently across any
3129 OpenSSL-based application. Commands have been added to all the
3130 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3131 control over shared-library paths without source code alterations.
3134 *) Changed all ENGINE implementations to dynamically allocate their
3135 ENGINEs rather than declaring them statically. Apart from this being
3136 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3137 this also allows the implementations to compile without using the
3138 internal engine_int.h header.
3141 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3142 'const' value. Any code that should be able to modify a RAND_METHOD
3143 should already have non-const pointers to it (ie. they should only
3144 modify their own ones).
3147 *) Made a variety of little tweaks to the ENGINE code.
3148 - "atalla" and "ubsec" string definitions were moved from header files
3149 to C code. "nuron" string definitions were placed in variables
3150 rather than hard-coded - allowing parameterisation of these values
3151 later on via ctrl() commands.
3152 - Removed unused "#if 0"'d code.
3153 - Fixed engine list iteration code so it uses ENGINE_free() to release
3154 structural references.
3155 - Constified the RAND_METHOD element of ENGINE structures.
3156 - Constified various get/set functions as appropriate and added
3157 missing functions (including a catch-all ENGINE_cpy that duplicates
3158 all ENGINE values onto a new ENGINE except reference counts/state).
3159 - Removed NULL parameter checks in get/set functions. Setting a method
3160 or function to NULL is a way of cancelling out a previously set
3161 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3162 and doesn't justify the extra error symbols and code.
3163 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3164 flags from engine_int.h to engine.h.
3165 - Changed prototypes for ENGINE handler functions (init(), finish(),
3166 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3169 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3170 to the algorithm using long division. The binary algorithm can be
3171 used only if the modulus is odd. On 32-bit systems, it is faster
3172 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3173 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3174 up to 450 bits. In 64-bit environments, the binary algorithm
3175 appears to be advantageous for much longer moduli; here we use it
3176 for moduli up to 2048 bits.
3179 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3180 could not support the combine flag in choice fields.
3183 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3184 extensions from a certificate request to the certificate.
3187 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3188 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3189 file: this allows the display of the certificate about to be
3190 signed to be customised, to allow certain fields to be included
3191 or excluded and extension details. The old system didn't display
3192 multicharacter strings properly, omitted fields not in the policy
3193 and couldn't display additional details such as extensions.
3196 *) Function EC_POINTs_mul for multiple scalar multiplication
3197 of an arbitrary number of elliptic curve points
3198 \sum scalars[i]*points[i],
3199 optionally including the generator defined for the EC_GROUP:
3200 scalar*generator + \sum scalars[i]*points[i].
3202 EC_POINT_mul is a simple wrapper function for the typical case
3203 that the point list has just one item (besides the optional
3207 *) First EC_METHODs for curves over GF(p):
3209 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3210 operations and provides various method functions that can also
3211 operate with faster implementations of modular arithmetic.
3213 EC_GFp_mont_method() reuses most functions that are part of
3214 EC_GFp_simple_method, but uses Montgomery arithmetic.
3216 [Bodo Moeller; point addition and point doubling
3217 implementation directly derived from source code provided by
3218 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3220 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3221 crypto/ec/ec_lib.c):
3223 Curves are EC_GROUP objects (with an optional group generator)
3224 based on EC_METHODs that are built into the library.
3226 Points are EC_POINT objects based on EC_GROUP objects.
3228 Most of the framework would be able to handle curves over arbitrary
3229 finite fields, but as there are no obvious types for fields other
3230 than GF(p), some functions are limited to that for now.
3233 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3234 that the file contains a complete HTTP response.
3237 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3238 change the def and num file printf format specifier from "%-40sXXX"
3239 to "%-39s XXX". The latter will always guarantee a space after the
3240 field while the former will cause them to run together if the field
3241 is 40 of more characters long.
3244 *) Constify the cipher and digest 'method' functions and structures
3245 and modify related functions to take constant EVP_MD and EVP_CIPHER
3249 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3250 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3253 *) Modify EVP_Digest*() routines so they now return values. Although the
3254 internal software routines can never fail additional hardware versions
3258 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3260 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3261 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3264 ERR_R_NESTED_ASN1_ERROR
3266 ERR_R_MISSING_ASN1_EOS
3267 were 4 .. 9, conflicting with
3268 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3270 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3271 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3273 Add new error code 'ERR_R_INTERNAL_ERROR'.
3276 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3280 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3281 sets the subject name for a new request or supersedes the
3282 subject name in a given request. Formats that can be parsed are
3283 'CN=Some Name, OU=myOU, C=IT'
3285 'CN=Some Name/OU=myOU/C=IT'.
3287 Add options '-batch' and '-verbose' to 'openssl req'.
3288 [Massimiliano Pala <madwolf@hackmasters.net>]
3290 *) Introduce the possibility to access global variables through
3291 functions on platform were that's the best way to handle exporting
3292 global variables in shared libraries. To enable this functionality,
3293 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3294 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3295 is normally done by Configure or something similar).
3297 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3298 in the source file (foo.c) like this:
3300 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3301 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3303 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3304 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3306 OPENSSL_DECLARE_GLOBAL(int,foo);
3307 #define foo OPENSSL_GLOBAL_REF(foo)
3308 OPENSSL_DECLARE_GLOBAL(double,bar);
3309 #define bar OPENSSL_GLOBAL_REF(bar)
3311 The #defines are very important, and therefore so is including the
3312 header file everywhere where the defined globals are used.
3314 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3315 of ASN.1 items, but that structure is a bit different.
3317 The largest change is in util/mkdef.pl which has been enhanced with
3318 better and easier to understand logic to choose which symbols should
3319 go into the Windows .def files as well as a number of fixes and code
3320 cleanup (among others, algorithm keywords are now sorted
3321 lexicographically to avoid constant rewrites).
3324 *) In BN_div() keep a copy of the sign of 'num' before writing the
3325 result to 'rm' because if rm==num the value will be overwritten
3326 and produce the wrong result if 'num' is negative: this caused
3327 problems with BN_mod() and BN_nnmod().
3330 *) Function OCSP_request_verify(). This checks the signature on an
3331 OCSP request and verifies the signer certificate. The signer
3332 certificate is just checked for a generic purpose and OCSP request
3336 *) Add OCSP_check_validity() function to check the validity of OCSP
3337 responses. OCSP responses are prepared in real time and may only
3338 be a few seconds old. Simply checking that the current time lies
3339 between thisUpdate and nextUpdate max reject otherwise valid responses
3340 caused by either OCSP responder or client clock inaccuracy. Instead
3341 we allow thisUpdate and nextUpdate to fall within a certain period of
3342 the current time. The age of the response can also optionally be
3343 checked. Two new options -validity_period and -status_age added to
3347 *) If signature or public key algorithm is unrecognized print out its
3348 OID rather that just UNKNOWN.
3351 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3352 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3353 ID to be generated from the issuer certificate alone which can then be
3354 passed to OCSP_id_issuer_cmp().
3357 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3358 ASN1 modules to export functions returning ASN1_ITEM pointers
3359 instead of the ASN1_ITEM structures themselves. This adds several
3360 new macros which allow the underlying ASN1 function/structure to
3361 be accessed transparently. As a result code should not use ASN1_ITEM
3362 references directly (such as &X509_it) but instead use the relevant
3363 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3364 use of the new ASN1 code on platforms where exporting structures
3365 is problematical (for example in shared libraries) but exporting
3366 functions returning pointers to structures is not.
3369 *) Add support for overriding the generation of SSL/TLS session IDs.
3370 These callbacks can be registered either in an SSL_CTX or per SSL.
3371 The purpose of this is to allow applications to control, if they wish,
3372 the arbitrary values chosen for use as session IDs, particularly as it
3373 can be useful for session caching in multiple-server environments. A
3374 command-line switch for testing this (and any client code that wishes
3375 to use such a feature) has been added to "s_server".
3376 [Geoff Thorpe, Lutz Jaenicke]
3378 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3379 of the form '#if defined(...) || defined(...) || ...' and
3380 '#if !defined(...) && !defined(...) && ...'. This also avoids
3381 the growing number of special cases it was previously handling.
3384 *) Make all configuration macros available for application by making
3385 sure they are available in opensslconf.h, by giving them names starting
3386 with "OPENSSL_" to avoid conflicts with other packages and by making
3387 sure e_os2.h will cover all platform-specific cases together with
3389 Additionally, it is now possible to define configuration/platform-
3390 specific names (called "system identities"). In the C code, these
3391 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3392 macro with the name beginning with "OPENSSL_SYS_", which is determined
3393 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3397 *) New option -set_serial to 'req' and 'x509' this allows the serial
3398 number to use to be specified on the command line. Previously self
3399 signed certificates were hard coded with serial number 0 and the
3400 CA options of 'x509' had to use a serial number in a file which was
3404 *) New options to 'ca' utility to support V2 CRL entry extensions.
3405 Currently CRL reason, invalidity date and hold instruction are
3406 supported. Add new CRL extensions to V3 code and some new objects.
3409 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3410 disable standard block padding (aka PKCS#5 padding) in the EVP
3411 API, which was previously mandatory. This means that the data is
3412 not padded in any way and so the total length much be a multiple
3413 of the block size, otherwise an error occurs.
3416 *) Initial (incomplete) OCSP SSL support.
3419 *) New function OCSP_parse_url(). This splits up a URL into its host,
3420 port and path components: primarily to parse OCSP URLs. New -url
3421 option to ocsp utility.
3424 *) New nonce behavior. The return value of OCSP_check_nonce() now
3425 reflects the various checks performed. Applications can decide
3426 whether to tolerate certain situations such as an absent nonce
3427 in a response when one was present in a request: the ocsp application
3428 just prints out a warning. New function OCSP_add1_basic_nonce()
3429 this is to allow responders to include a nonce in a response even if
3430 the request is nonce-less.
3433 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3434 skipped when using openssl x509 multiple times on a single input file,
3435 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3438 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3439 set string type: to handle setting ASN1_TIME structures. Fix ca
3440 utility to correctly initialize revocation date of CRLs.
3443 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3444 the clients preferred ciphersuites and rather use its own preferences.
3445 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3446 Internet Explorer by ensuring unchanged hash method during stepup.
3447 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3450 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3451 to aes and add a new 'exist' option to print out symbols that don't
3455 *) Additional options to ocsp utility to allow flags to be set and
3456 additional certificates supplied.
3459 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3460 OCSP client a number of certificate to only verify the response
3464 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3465 handle the new API. Currently only ECB, CBC modes supported. Add new
3468 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3469 Encryption Standard (AES) Ciphersuites for Transport Layer
3470 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3471 not enabled by default and were not part of the "ALL" ciphersuite
3472 alias because they were not yet official; they could be
3473 explicitly requested by specifying the "AESdraft" ciphersuite
3474 group alias. In the final release of OpenSSL 0.9.7, the group
3475 alias is called "AES" and is part of "ALL".)
3476 [Ben Laurie, Steve Henson, Bodo Moeller]
3478 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3479 request to response.
3482 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3483 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3484 extract information from a certificate request. OCSP_response_create()
3485 creates a response and optionally adds a basic response structure.
3486 OCSP_basic_add1_status() adds a complete single response to a basic
3487 response and returns the OCSP_SINGLERESP structure just added (to allow
3488 extensions to be included for example). OCSP_basic_add1_cert() adds a
3489 certificate to a basic response and OCSP_basic_sign() signs a basic
3490 response with various flags. New helper functions ASN1_TIME_check()
3491 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3492 (converts ASN1_TIME to GeneralizedTime).
3495 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3496 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3497 structure from a certificate. X509_pubkey_digest() digests the public_key
3498 contents: this is used in various key identifiers.
3501 *) Make sk_sort() tolerate a NULL argument.
3502 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3504 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3505 passed by the function are trusted implicitly. If any of them signed the
3506 response then it is assumed to be valid and is not verified.
3509 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3510 to data. This was previously part of the PKCS7 ASN1 code. This
3511 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3512 [Steve Henson, reported by Kenneth R. Robinette
3513 <support@securenetterm.com>]
3515 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3516 routines: without these tracing memory leaks is very painful.
3517 Fix leaks in PKCS12 and PKCS7 routines.
3520 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3521 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3522 effectively meant GeneralizedTime would never be used. Now it
3523 is initialised to -1 but X509_time_adj() now has to check the value
3524 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3525 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3526 [Steve Henson, reported by Kenneth R. Robinette
3527 <support@securenetterm.com>]
3529 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3530 result in a zero length in the ASN1_INTEGER structure which was
3531 not consistent with the structure when d2i_ASN1_INTEGER() was used
3532 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3533 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3534 where it did not print out a minus for negative ASN1_INTEGER.
3537 *) Add summary printout to ocsp utility. The various functions which
3538 convert status values to strings have been renamed to:
3539 OCSP_response_status_str(), OCSP_cert_status_str() and
3540 OCSP_crl_reason_str() and are no longer static. New options
3541 to verify nonce values and to disable verification. OCSP response
3542 printout format cleaned up.
3545 *) Add additional OCSP certificate checks. These are those specified
3546 in RFC2560. This consists of two separate checks: the CA of the
3547 certificate being checked must either be the OCSP signer certificate
3548 or the issuer of the OCSP signer certificate. In the latter case the
3549 OCSP signer certificate must contain the OCSP signing extended key
3550 usage. This check is performed by attempting to match the OCSP
3551 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3552 in the OCSP_CERTID structures of the response.
3555 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3556 and related routines. This uses the standard OpenSSL certificate
3557 verify routines to perform initial checks (just CA validity) and
3558 to obtain the certificate chain. Then additional checks will be
3559 performed on the chain. Currently the root CA is checked to see
3560 if it is explicitly trusted for OCSP signing. This is used to set
3561 a root CA as a global signing root: that is any certificate that
3562 chains to that CA is an acceptable OCSP signing certificate.
3565 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3566 extensions from a separate configuration file.
3567 As when reading extensions from the main configuration file,
3568 the '-extensions ...' option may be used for specifying the
3570 [Massimiliano Pala <madwolf@comune.modena.it>]
3572 *) New OCSP utility. Allows OCSP requests to be generated or
3573 read. The request can be sent to a responder and the output
3574 parsed, outputed or printed in text form. Not complete yet:
3575 still needs to check the OCSP response validity.
3578 *) New subcommands for 'openssl ca':
3579 'openssl ca -status <serial>' prints the status of the cert with
3580 the given serial number (according to the index file).
3581 'openssl ca -updatedb' updates the expiry status of certificates
3583 [Massimiliano Pala <madwolf@comune.modena.it>]
3585 *) New '-newreq-nodes' command option to CA.pl. This is like
3586 '-newreq', but calls 'openssl req' with the '-nodes' option
3587 so that the resulting key is not encrypted.
3588 [Damien Miller <djm@mindrot.org>]
3590 *) New configuration for the GNU Hurd.
3591 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3593 *) Initial code to implement OCSP basic response verify. This
3594 is currently incomplete. Currently just finds the signer's
3595 certificate and verifies the signature on the response.
3598 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3599 value of OPENSSLDIR. This is available via the new '-d' option
3600 to 'openssl version', and is also included in 'openssl version -a'.
3603 *) Allowing defining memory allocation callbacks that will be given
3604 file name and line number information in additional arguments
3605 (a const char* and an int). The basic functionality remains, as
3606 well as the original possibility to just replace malloc(),
3607 realloc() and free() by functions that do not know about these
3608 additional arguments. To register and find out the current
3609 settings for extended allocation functions, the following
3610 functions are provided:
3612 CRYPTO_set_mem_ex_functions
3613 CRYPTO_set_locked_mem_ex_functions
3614 CRYPTO_get_mem_ex_functions
3615 CRYPTO_get_locked_mem_ex_functions
3617 These work the same way as CRYPTO_set_mem_functions and friends.
3618 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3619 extended allocation function is enabled.
3620 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3621 a conventional allocation function is enabled.
3622 [Richard Levitte, Bodo Moeller]
3624 *) Finish off removing the remaining LHASH function pointer casts.
3625 There should no longer be any prototype-casting required when using
3626 the LHASH abstraction, and any casts that remain are "bugs". See
3627 the callback types and macros at the head of lhash.h for details
3628 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3631 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3632 If /dev/[u]random devices are not available or do not return enough
3633 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3635 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3636 /etc/entropy will be queried once each in this sequence, quering stops
3637 when enough entropy was collected without querying more sockets.
3640 *) Change the Unix RAND_poll() variant to be able to poll several
3641 random devices, as specified by DEVRANDOM, until a sufficient amount
3642 of data has been collected. We spend at most 10 ms on each file
3643 (select timeout) and read in non-blocking mode. DEVRANDOM now
3644 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3645 (previously it was just the string "/dev/urandom"), so on typical
3646 platforms the 10 ms delay will never occur.
3647 Also separate out the Unix variant to its own file, rand_unix.c.
3648 For VMS, there's a currently-empty rand_vms.c.
3651 *) Move OCSP client related routines to ocsp_cl.c. These
3652 provide utility functions which an application needing
3653 to issue a request to an OCSP responder and analyse the
3654 response will typically need: as opposed to those which an
3655 OCSP responder itself would need which will be added later.
3657 OCSP_request_sign() signs an OCSP request with an API similar
3658 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3659 response. OCSP_response_get1_basic() extracts basic response
3660 from response. OCSP_resp_find_status(): finds and extracts status
3661 information from an OCSP_CERTID structure (which will be created
3662 when the request structure is built). These are built from lower
3663 level functions which work on OCSP_SINGLERESP structures but
3664 wont normally be used unless the application wishes to examine
3665 extensions in the OCSP response for example.
3667 Replace nonce routines with a pair of functions.
3668 OCSP_request_add1_nonce() adds a nonce value and optionally
3669 generates a random value. OCSP_check_nonce() checks the
3670 validity of the nonce in an OCSP response.
3673 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3674 This doesn't copy the supplied OCSP_CERTID and avoids the
3675 need to free up the newly created id. Change return type
3676 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3677 This can then be used to add extensions to the request.
3678 Deleted OCSP_request_new(), since most of its functionality
3679 is now in OCSP_REQUEST_new() (and the case insensitive name
3680 clash) apart from the ability to set the request name which
3681 will be added elsewhere.
3684 *) Update OCSP API. Remove obsolete extensions argument from
3685 various functions. Extensions are now handled using the new
3686 OCSP extension code. New simple OCSP HTTP function which
3687 can be used to send requests and parse the response.
3690 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3691 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3692 uses the special reorder version of SET OF to sort the attributes
3693 and reorder them to match the encoded order. This resolves a long
3694 standing problem: a verify on a PKCS7 structure just after signing
3695 it used to fail because the attribute order did not match the
3696 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3697 it uses the received order. This is necessary to tolerate some broken
3698 software that does not order SET OF. This is handled by encoding
3699 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3700 to produce the required SET OF.
3703 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3704 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3705 files to get correct declarations of the ASN.1 item variables.
3708 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3709 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3710 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3711 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3712 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3713 ASN1_ITEM and no wrapper functions.
3716 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3717 replace the old function pointer based I/O routines. Change most of
3718 the *_d2i_bio() and *_d2i_fp() functions to use these.
3721 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3722 lines, recognice more "algorithms" that can be deselected, and make
3723 it complain about algorithm deselection that isn't recognised.
3726 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3727 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3728 to use new functions. Add NO_ASN1_OLD which can be set to remove
3729 some old style ASN1 functions: this can be used to determine if old
3730 code will still work when these eventually go away.
3733 *) New extension functions for OCSP structures, these follow the
3734 same conventions as certificates and CRLs.
3737 *) New function X509V3_add1_i2d(). This automatically encodes and
3738 adds an extension. Its behaviour can be customised with various
3739 flags to append, replace or delete. Various wrappers added for
3740 certifcates and CRLs.
3743 *) Fix to avoid calling the underlying ASN1 print routine when
3744 an extension cannot be parsed. Correct a typo in the
3745 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3748 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3749 entries for variables.
3752 *) Add functionality to apps/openssl.c for detecting locking
3753 problems: As the program is single-threaded, all we have
3754 to do is register a locking callback using an array for
3755 storing which locks are currently held by the program.
3758 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3759 SSL_get_ex_data_X509_STORE_idx(), which is used in
3760 ssl_verify_cert_chain() and thus can be called at any time
3761 during TLS/SSL handshakes so that thread-safety is essential.
3762 Unfortunately, the ex_data design is not at all suited
3763 for multi-threaded use, so it probably should be abolished.
3766 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3767 [Broadcom, tweaked and integrated by Geoff Thorpe]
3769 *) Move common extension printing code to new function
3770 X509V3_print_extensions(). Reorganise OCSP print routines and
3771 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3774 *) New function X509_signature_print() to remove duplication in some
3778 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3779 set (this was treated exactly the same as SET OF previously). This
3780 is used to reorder the STACK representing the structure to match the
3781 encoding. This will be used to get round a problem where a PKCS7
3782 structure which was signed could not be verified because the STACK
3783 order did not reflect the encoded order.
3786 *) Reimplement the OCSP ASN1 module using the new code.
3789 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3790 for its ASN1 operations. The old style function pointers still exist
3791 for now but they will eventually go away.
3794 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3795 completely replaces the old ASN1 functionality with a table driven
3796 encoder and decoder which interprets an ASN1_ITEM structure describing
3797 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3798 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3799 has also been converted to the new form.
3802 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3803 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3804 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3805 for negative moduli.
3808 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3809 of not touching the result's sign bit.
3812 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3816 *) Changed the LHASH code to use prototypes for callbacks, and created
3817 macros to declare and implement thin (optionally static) functions
3818 that provide type-safety and avoid function pointer casting for the
3819 type-specific callbacks.
3822 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3824 [Veers Staats <staatsvr@asc.hpc.mil>,
3825 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3827 *) Reformat the FAQ so the different questions and answers can be divided
3828 in sections depending on the subject.
3831 *) Have the zlib compression code load ZLIB.DLL dynamically under
3835 *) New function BN_mod_sqrt for computing square roots modulo a prime
3836 (using the probabilistic Tonelli-Shanks algorithm unless
3837 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3838 be handled deterministically).
3839 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3841 *) Make BN_mod_inverse faster by explicitly handling small quotients
3842 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3843 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3846 *) New function BN_kronecker.
3849 *) Fix BN_gcd so that it works on negative inputs; the result is
3850 positive unless both parameters are zero.
3851 Previously something reasonably close to an infinite loop was
3852 possible because numbers could be growing instead of shrinking
3853 in the implementation of Euclid's algorithm.
3856 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3857 sign of the number in question.
3859 Fix BN_is_word(a,w) to work correctly for w == 0.
3861 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3862 because its test if the absolute value of 'a' equals 'w'.
3863 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3864 it exists mostly for use in the implementations of BN_is_zero(),
3865 BN_is_one(), and BN_is_word().
3868 *) New function BN_swap.
3871 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3872 the exponentiation functions are more likely to produce reasonable
3873 results on negative inputs.
3876 *) Change BN_mod_mul so that the result is always non-negative.
3877 Previously, it could be negative if one of the factors was negative;
3878 I don't think anyone really wanted that behaviour.
3881 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3882 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3883 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3884 and add new functions:
3893 BN_mod_lshift1_quick
3897 These functions always generate non-negative results.
3899 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3900 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3902 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3903 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3904 be reduced modulo m.
3905 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3908 The following entry accidentily appeared in the CHANGES file
3909 distributed with OpenSSL 0.9.7. The modifications described in
3910 it do *not* apply to OpenSSL 0.9.7.
3912 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3913 was actually never needed) and in BN_mul(). The removal in BN_mul()
3914 required a small change in bn_mul_part_recursive() and the addition
3915 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3916 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3917 bn_sub_words() and bn_add_words() except they take arrays with
3922 *) In 'openssl passwd', verify passwords read from the terminal
3923 unless the '-salt' option is used (which usually means that
3924 verification would just waste user's time since the resulting
3925 hash is going to be compared with some given password hash)
3926 or the new '-noverify' option is used.
3928 This is an incompatible change, but it does not affect
3929 non-interactive use of 'openssl passwd' (passwords on the command
3930 line, '-stdin' option, '-in ...' option) and thus should not
3934 *) Remove all references to RSAref, since there's no more need for it.
3937 *) Make DSO load along a path given through an environment variable
3938 (SHLIB_PATH) with shl_load().
3941 *) Constify the ENGINE code as a result of BIGNUM constification.
3942 Also constify the RSA code and most things related to it. In a
3943 few places, most notable in the depth of the ASN.1 code, ugly
3944 casts back to non-const were required (to be solved at a later
3948 *) Make it so the openssl application has all engines loaded by default.
3951 *) Constify the BIGNUM routines a little more.
3954 *) Add the following functions:
3956 ENGINE_load_cswift()
3958 ENGINE_load_atalla()
3960 ENGINE_load_builtin_engines()
3962 That way, an application can itself choose if external engines that
3963 are built-in in OpenSSL shall ever be used or not. The benefit is
3964 that applications won't have to be linked with libdl or other dso
3965 libraries unless it's really needed.
3967 Changed 'openssl engine' to load all engines on demand.
3968 Changed the engine header files to avoid the duplication of some
3969 declarations (they differed!).
3972 *) 'openssl engine' can now list capabilities.
3975 *) Better error reporting in 'openssl engine'.
3978 *) Never call load_dh_param(NULL) in s_server.
3981 *) Add engine application. It can currently list engines by name and
3982 identity, and test if they are actually available.
3985 *) Improve RPM specification file by forcing symbolic linking and making
3986 sure the installed documentation is also owned by root.root.
3987 [Damien Miller <djm@mindrot.org>]
3989 *) Give the OpenSSL applications more possibilities to make use of
3990 keys (public as well as private) handled by engines.
3993 *) Add OCSP code that comes from CertCo.
3996 *) Add VMS support for the Rijndael code.
3999 *) Added untested support for Nuron crypto accelerator.
4002 *) Add support for external cryptographic devices. This code was
4003 previously distributed separately as the "engine" branch.
4004 [Geoff Thorpe, Richard Levitte]
4006 *) Rework the filename-translation in the DSO code. It is now possible to
4007 have far greater control over how a "name" is turned into a filename
4008 depending on the operating environment and any oddities about the
4009 different shared library filenames on each system.
4012 *) Support threads on FreeBSD-elf in Configure.
4015 *) Fix for SHA1 assembly problem with MASM: it produces
4016 warnings about corrupt line number information when assembling
4017 with debugging information. This is caused by the overlapping
4019 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4022 NCONF_get_number() has no error checking at all. As a replacement,
4023 NCONF_get_number_e() is defined (_e for "error checking") and is
4024 promoted strongly. The old NCONF_get_number is kept around for
4025 binary backward compatibility.
4026 Make it possible for methods to load from something other than a BIO,
4027 by providing a function pointer that is given a name instead of a BIO.
4028 For example, this could be used to load configuration data from an
4032 *) Fix for non blocking accept BIOs. Added new I/O special reason
4033 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4034 with non blocking I/O was not possible because no retry code was
4035 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4039 *) Added the beginnings of Rijndael support.
4042 *) Fix for bug in DirectoryString mask setting. Add support for
4043 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4044 to allow certificate printing to more controllable, additional
4045 'certopt' option to 'x509' to allow new printing options to be
4049 *) Clean old EAY MD5 hack from e_os.h.
4052 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4054 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4055 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4056 [Joe Orton, Steve Henson]
4058 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4060 *) Fix additional bug revealed by the NISCC test suite:
4062 Stop bug triggering large recursion when presented with
4063 certain ASN.1 tags (CVE-2003-0851)
4066 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4068 *) Fix various bugs revealed by running the NISCC test suite:
4070 Stop out of bounds reads in the ASN1 code when presented with
4071 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4073 If verify callback ignores invalid public key errors don't try to check
4074 certificate signature with the NULL public key.
4078 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4079 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4083 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4084 extra data after the compression methods not only for TLS 1.0
4085 but also for SSL 3.0 (as required by the specification).
4086 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4088 *) Change X509_certificate_type() to mark the key as exported/exportable
4089 when it's 512 *bits* long, not 512 bytes.
4092 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4094 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4095 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4096 a protocol version number mismatch like a decryption error
4097 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4100 *) Turn on RSA blinding by default in the default implementation
4101 to avoid a timing attack. Applications that don't want it can call
4102 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4103 They would be ill-advised to do so in most cases.
4104 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4106 *) Change RSA blinding code so that it works when the PRNG is not
4107 seeded (in this case, the secret RSA exponent is abused as
4108 an unpredictable seed -- if it is not unpredictable, there
4109 is no point in blinding anyway). Make RSA blinding thread-safe
4110 by remembering the creator's thread ID in rsa->blinding and
4111 having all other threads use local one-time blinding factors
4112 (this requires more computation than sharing rsa->blinding, but
4113 avoids excessive locking; and if an RSA object is not shared
4114 between threads, blinding will still be very fast).
4117 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4119 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4120 via timing by performing a MAC computation even if incorrrect
4121 block cipher padding has been found. This is a countermeasure
4122 against active attacks where the attacker has to distinguish
4123 between bad padding and a MAC verification error. (CVE-2003-0078)
4125 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4126 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4127 Martin Vuagnoux (EPFL, Ilion)]
4129 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4131 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4132 memory from it's contents. This is done with a counter that will
4133 place alternating values in each byte. This can be used to solve
4134 two issues: 1) the removal of calls to memset() by highly optimizing
4135 compilers, and 2) cleansing with other values than 0, since those can
4136 be read through on certain media, for example a swap space on disk.
4139 *) Bugfix: client side session caching did not work with external caching,
4140 because the session->cipher setting was not restored when reloading
4141 from the external cache. This problem was masked, when
4142 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4143 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4146 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4147 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4148 [Zeev Lieber <zeev-l@yahoo.com>]
4150 *) Undo an undocumented change introduced in 0.9.6e which caused
4151 repeated calls to OpenSSL_add_all_ciphers() and
4152 OpenSSL_add_all_digests() to be ignored, even after calling
4156 *) Change the default configuration reader to deal with last line not
4157 being properly terminated.
4160 *) Change X509_NAME_cmp() so it applies the special rules on handling
4161 DN values that are of type PrintableString, as well as RDNs of type
4162 emailAddress where the value has the type ia5String.
4163 [stefank@valicert.com via Richard Levitte]
4165 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4166 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4167 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4168 the bitwise-OR of the two for use by the majority of applications
4169 wanting this behaviour, and update the docs. The documented
4170 behaviour and actual behaviour were inconsistent and had been
4171 changing anyway, so this is more a bug-fix than a behavioural
4173 [Geoff Thorpe, diagnosed by Nadav Har'El]
4175 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4176 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4179 *) Fix initialization code race conditions in
4180 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4181 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4182 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4183 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4184 ssl2_get_cipher_by_char(),
4185 ssl3_get_cipher_by_char().
4186 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4188 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4189 the cached sessions are flushed, as the remove_cb() might use ex_data
4190 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4191 (see [openssl.org #212]).
4192 [Geoff Thorpe, Lutz Jaenicke]
4194 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4195 length, instead of the encoding length to d2i_ASN1_OBJECT.
4198 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4200 *) [In 0.9.6g-engine release:]
4201 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4202 [Lynn Gazis <lgazis@rainbow.com>]
4204 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4206 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4207 and get fix the header length calculation.
4208 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4209 Alon Kantor <alonk@checkpoint.com> (and others),
4212 *) Use proper error handling instead of 'assertions' in buffer
4213 overflow checks added in 0.9.6e. This prevents DoS (the
4214 assertions could call abort()).
4215 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4217 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4219 *) Add various sanity checks to asn1_get_length() to reject
4220 the ASN1 length bytes if they exceed sizeof(long), will appear
4221 negative or the content length exceeds the length of the
4223 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4225 *) Fix cipher selection routines: ciphers without encryption had no flags
4226 for the cipher strength set and where therefore not handled correctly
4227 by the selection routines (PR #130).
4230 *) Fix EVP_dsa_sha macro.
4234 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4235 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4236 that was added in OpenSSL 0.9.6d.
4238 As the countermeasure turned out to be incompatible with some
4239 broken SSL implementations, the new option is part of SSL_OP_ALL.
4240 SSL_OP_ALL is usually employed when compatibility with weird SSL
4241 implementations is desired (e.g. '-bugs' option to 's_client' and
4242 's_server'), so the new option is automatically set in many
4246 *) Changes in security patch:
4248 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4249 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4250 Air Force Materiel Command, USAF, under agreement number
4253 *) Add various sanity checks to asn1_get_length() to reject
4254 the ASN1 length bytes if they exceed sizeof(long), will appear
4255 negative or the content length exceeds the length of the
4256 supplied buffer. (CVE-2002-0659)
4257 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4259 *) Assertions for various potential buffer overflows, not known to
4261 [Ben Laurie (CHATS)]
4263 *) Various temporary buffers to hold ASCII versions of integers were
4264 too small for 64 bit platforms. (CVE-2002-0655)
4265 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4267 *) Remote buffer overflow in SSL3 protocol - an attacker could
4268 supply an oversized session ID to a client. (CVE-2002-0656)
4269 [Ben Laurie (CHATS)]
4271 *) Remote buffer overflow in SSL2 protocol - an attacker could
4272 supply an oversized client master key. (CVE-2002-0656)
4273 [Ben Laurie (CHATS)]
4275 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4277 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4278 encoded as NULL) with id-dsa-with-sha1.
4279 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4281 *) Check various X509_...() return values in apps/req.c.
4282 [Nils Larsch <nla@trustcenter.de>]
4284 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4285 an end-of-file condition would erronously be flagged, when the CRLF
4286 was just at the end of a processed block. The bug was discovered when
4287 processing data through a buffering memory BIO handing the data to a
4288 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4289 <ptsekov@syntrex.com> and Nedelcho Stanev.
4292 *) Implement a countermeasure against a vulnerability recently found
4293 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4294 before application data chunks to avoid the use of known IVs
4295 with data potentially chosen by the attacker.
4298 *) Fix length checks in ssl3_get_client_hello().
4301 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4302 to prevent ssl3_read_internal() from incorrectly assuming that
4303 ssl3_read_bytes() found application data while handshake
4304 processing was enabled when in fact s->s3->in_read_app_data was
4305 merely automatically cleared during the initial handshake.
4306 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4308 *) Fix object definitions for Private and Enterprise: they were not
4309 recognized in their shortname (=lowercase) representation. Extend
4310 obj_dat.pl to issue an error when using undefined keywords instead
4311 of silently ignoring the problem (Svenning Sorensen
4312 <sss@sss.dnsalias.net>).
4315 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4316 generators, i.e. generators other than 2 and 5. (Previously, the
4317 code did not properly initialise the 'add' and 'rem' values to
4318 BN_generate_prime().)
4320 In the new general case, we do not insist that 'generator' is
4321 actually a primitive root: This requirement is rather pointless;
4322 a generator of the order-q subgroup is just as good, if not
4326 *) Map new X509 verification errors to alerts. Discovered and submitted by
4327 Tom Wu <tom@arcot.com>.
4330 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4331 returning non-zero before the data has been completely received
4332 when using non-blocking I/O.
4333 [Bodo Moeller; problem pointed out by John Hughes]
4335 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4336 [Ben Laurie, Lutz Jaenicke]
4338 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4339 Yoram Zahavi <YoramZ@gilian.com>).
4342 *) Add information about CygWin 1.3 and on, and preserve proper
4343 configuration for the versions before that.
4344 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4346 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4347 check whether we deal with a copy of a session and do not delete from
4348 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4349 <izhar@checkpoint.com>.
4352 *) Do not store session data into the internal session cache, if it
4353 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4354 flag is set). Proposed by Aslam <aslam@funk.com>.
4357 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4361 *) [In 0.9.6d-engine release:]
4362 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4363 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4365 *) Add the configuration target linux-s390x.
4366 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4368 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4369 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4370 variable as an indication that a ClientHello message has been
4371 received. As the flag value will be lost between multiple
4372 invocations of ssl3_accept when using non-blocking I/O, the
4373 function may not be aware that a handshake has actually taken
4374 place, thus preventing a new session from being added to the
4377 To avoid this problem, we now set s->new_session to 2 instead of
4378 using a local variable.
4379 [Lutz Jaenicke, Bodo Moeller]
4381 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4382 if the SSL_R_LENGTH_MISMATCH error is detected.
4383 [Geoff Thorpe, Bodo Moeller]
4385 *) New 'shared_ldflag' column in Configure platform table.
4388 *) Fix EVP_CIPHER_mode macro.
4389 ["Dan S. Camper" <dan@bti.net>]
4391 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4392 type, we must throw them away by setting rr->length to 0.
4393 [D P Chang <dpc@qualys.com>]
4395 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4397 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4398 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4399 worked incorrectly for those cases where range = 10..._2 and
4400 3*range is two bits longer than range.)
4403 *) Only add signing time to PKCS7 structures if it is not already
4407 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4408 OBJ_ld_ce should be OBJ_id_ce.
4409 Also some ip-pda OIDs in crypto/objects/objects.txt were
4410 incorrect (cf. RFC 3039).
4411 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4413 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4414 returns early because it has nothing to do.
4415 [Andy Schneider <andy.schneider@bjss.co.uk>]
4417 *) [In 0.9.6c-engine release:]
4418 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4419 [Andy Schneider <andy.schneider@bjss.co.uk>]
4421 *) [In 0.9.6c-engine release:]
4422 Add support for Cryptographic Appliance's keyserver technology.
4423 (Use engine 'keyclient')
4424 [Cryptographic Appliances and Geoff Thorpe]
4426 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4427 is called via tools/c89.sh because arguments have to be
4428 rearranged (all '-L' options must appear before the first object
4430 [Richard Shapiro <rshapiro@abinitio.com>]
4432 *) [In 0.9.6c-engine release:]
4433 Add support for Broadcom crypto accelerator cards, backported
4435 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4437 *) [In 0.9.6c-engine release:]
4438 Add support for SureWare crypto accelerator cards from
4439 Baltimore Technologies. (Use engine 'sureware')
4440 [Baltimore Technologies and Mark Cox]
4442 *) [In 0.9.6c-engine release:]
4443 Add support for crypto accelerator cards from Accelerated
4444 Encryption Processing, www.aep.ie. (Use engine 'aep')
4445 [AEP Inc. and Mark Cox]
4447 *) Add a configuration entry for gcc on UnixWare.
4448 [Gary Benson <gbenson@redhat.com>]
4450 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4451 messages are stored in a single piece (fixed-length part and
4452 variable-length part combined) and fix various bugs found on the way.
4455 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4456 instead. BIO_gethostbyname() does not know what timeouts are
4457 appropriate, so entries would stay in cache even when they have
4459 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4461 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4462 faced with a pathologically small ClientHello fragment that does
4463 not contain client_version: Instead of aborting with an error,
4464 simply choose the highest available protocol version (i.e.,
4465 TLS 1.0 unless it is disabled). In practice, ClientHello
4466 messages are never sent like this, but this change gives us
4467 strictly correct behaviour at least for TLS.
4470 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4471 never resets s->method to s->ctx->method when called from within
4472 one of the SSL handshake functions.
4473 [Bodo Moeller; problem pointed out by Niko Baric]
4475 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4476 (sent using the client's version number) if client_version is
4477 smaller than the protocol version in use. Also change
4478 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4479 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4480 the client will at least see that alert.
4483 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4487 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4488 client receives HelloRequest while in a handshake.
4489 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4491 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4492 should end in 'break', not 'goto end' which circuments various
4493 cleanups done in state SSL_ST_OK. But session related stuff
4494 must be disabled for SSL_ST_OK in the case that we just sent a
4497 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4498 before just sending a HelloRequest.
4499 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4501 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4502 reveal whether illegal block cipher padding was found or a MAC
4503 verification error occured. (Neither SSLerr() codes nor alerts
4504 are directly visible to potential attackers, but the information
4505 may leak via logfiles.)
4507 Similar changes are not required for the SSL 2.0 implementation
4508 because the number of padding bytes is sent in clear for SSL 2.0,
4509 and the extra bytes are just ignored. However ssl/s2_pkt.c
4510 failed to verify that the purported number of padding bytes is in
4514 *) Add OpenUNIX-8 support including shared libraries
4515 (Boyd Lynn Gerber <gerberb@zenez.com>).
4518 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4519 'wristwatch attack' using huge encoding parameters (cf.
4520 James H. Manger's CRYPTO 2001 paper). Note that the
4521 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4522 encoding parameters and hence was not vulnerable.
4525 *) BN_sqr() bug fix.
4526 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4528 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4529 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4530 followed by modular reduction.
4531 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4533 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4534 equivalent based on BN_pseudo_rand() instead of BN_rand().
4537 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4538 This function was broken, as the check for a new client hello message
4539 to handle SGC did not allow these large messages.
4540 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4543 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4546 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4547 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4550 *) Rework the configuration and shared library support for Tru64 Unix.
4551 The configuration part makes use of modern compiler features and
4552 still retains old compiler behavior for those that run older versions
4553 of the OS. The shared library support part includes a variant that
4554 uses the RPATH feature, and is available through the special
4555 configuration target "alpha-cc-rpath", which will never be selected
4557 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4559 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4560 with the same message size as in ssl3_get_certificate_request().
4561 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4562 messages might inadvertently be reject as too long.
4563 [Petr Lampa <lampa@fee.vutbr.cz>]
4565 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4568 *) Modified SSL library such that the verify_callback that has been set
4569 specificly for an SSL object with SSL_set_verify() is actually being
4570 used. Before the change, a verify_callback set with this function was
4571 ignored and the verify_callback() set in the SSL_CTX at the time of
4572 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4573 to allow the necessary settings.
4576 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4577 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4578 done automatically (in contradiction to the requirements of the C
4579 standard). This made problems when used from OpenSSH.
4582 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4583 dh->length and always used
4585 BN_rand_range(priv_key, dh->p).
4587 BN_rand_range() is not necessary for Diffie-Hellman, and this
4588 specific range makes Diffie-Hellman unnecessarily inefficient if
4589 dh->length (recommended exponent length) is much smaller than the
4590 length of dh->p. We could use BN_rand_range() if the order of
4591 the subgroup was stored in the DH structure, but we only have
4596 BN_rand(priv_key, l, ...)
4598 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4604 RSA_eay_public_encrypt
4605 RSA_eay_private_decrypt
4606 RSA_eay_private_encrypt (signing)
4607 RSA_eay_public_decrypt (signature verification)
4609 (default implementations for RSA_public_encrypt,
4610 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4611 always reject numbers >= n.
4614 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4615 to synchronize access to 'locking_thread'. This is necessary on
4616 systems where access to 'locking_thread' (an 'unsigned long'
4617 variable) is not atomic.
4620 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4621 *before* setting the 'crypto_lock_rand' flag. The previous code had
4622 a race condition if 0 is a valid thread ID.
4623 [Travis Vitek <vitek@roguewave.com>]
4625 *) Add support for shared libraries under Irix.
4626 [Albert Chin-A-Young <china@thewrittenword.com>]
4628 *) Add configuration option to build on Linux on both big-endian and
4630 [Ralf Baechle <ralf@uni-koblenz.de>]
4632 *) Add the possibility to create shared libraries on HP-UX.
4635 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4637 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4638 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4639 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4640 PRNG state recovery was possible based on the output of
4641 one PRNG request appropriately sized to gain knowledge on
4642 'md' followed by enough consecutive 1-byte PRNG requests
4643 to traverse all of 'state'.
4645 1. When updating 'md_local' (the current thread's copy of 'md')
4646 during PRNG output generation, hash all of the previous
4647 'md_local' value, not just the half used for PRNG output.
4649 2. Make the number of bytes from 'state' included into the hash
4650 independent from the number of PRNG bytes requested.
4652 The first measure alone would be sufficient to avoid
4653 Markku-Juhani's attack. (Actually it had never occurred
4654 to me that the half of 'md_local' used for chaining was the
4655 half from which PRNG output bytes were taken -- I had always
4656 assumed that the secret half would be used.) The second
4657 measure makes sure that additional data from 'state' is never
4658 mixed into 'md_local' in small portions; this heuristically
4659 further strengthens the PRNG.
4662 *) Fix crypto/bn/asm/mips3.s.
4665 *) When only the key is given to "enc", the IV is undefined. Print out
4666 an error message in this case.
4669 *) Handle special case when X509_NAME is empty in X509 printing routines.
4672 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4673 positive and less than q.
4676 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4677 used: it isn't thread safe and the add_lock_callback should handle
4679 [Paul Rose <Paul.Rose@bridge.com>]
4681 *) Verify that incoming data obeys the block size in
4682 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4686 [Ulf Möller, Bodo Möller]
4688 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4689 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4690 when fixing the server behaviour for backwards-compatible 'client
4691 hello' messages. (Note that the attack is impractical against
4692 SSL 3.0 and TLS 1.0 anyway because length and version checking
4693 means that the probability of guessing a valid ciphertext is
4694 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4697 Before 0.9.5, the countermeasure (hide the error by generating a
4698 random 'decryption result') did not work properly because
4699 ERR_clear_error() was missing, meaning that SSL_get_error() would
4700 detect the supposedly ignored error.
4702 Both problems are now fixed.
4705 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4706 (previously it was 1024).
4709 *) Fix for compatibility mode trust settings: ignore trust settings
4710 unless some valid trust or reject settings are present.
4713 *) Fix for blowfish EVP: its a variable length cipher.
4716 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4717 parameters in DSA public key structures and return an error in the
4718 DSA routines if parameters are absent.
4721 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4722 in the current directory if neither $RANDFILE nor $HOME was set.
4723 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4724 caused some confusion to Windows users who haven't defined $HOME.
4725 Thus RAND_file_name() is changed again: e_os.h can define a
4726 DEFAULT_HOME, which will be used if $HOME is not set.
4727 For Windows, we use "C:"; on other platforms, we still require
4728 environment variables.
4730 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4731 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4732 having multiple threads call RAND_poll() concurrently.
4735 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4736 combination of a flag and a thread ID variable.
4737 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4738 flag), *other* threads can enter ssleay_add_bytes without obeying
4739 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4740 that they do not hold after the first thread unsets add_do_not_lock).
4743 *) Change bctest again: '-x' expressions are not available in all
4747 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4749 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4750 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4752 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4753 the default extension for executables, if any. Also, make the perl
4754 scripts that use symlink() to test if it really exists and use "cp"
4755 if it doesn't. All this made OpenSSL compilable and installable in
4759 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4760 If SEQUENCE is length is indefinite just set c->slen to the total
4761 amount of data available.
4762 [Steve Henson, reported by shige@FreeBSD.org]
4763 [This change does not apply to 0.9.7.]
4765 *) Change bctest to avoid here-documents inside command substitution
4766 (workaround for FreeBSD /bin/sh bug).
4767 For compatibility with Ultrix, avoid shell functions (introduced
4768 in the bctest version that searches along $PATH).
4771 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4772 with des_encrypt() defined on some operating systems, like Solaris
4776 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4777 On the Importance of Eliminating Errors in Cryptographic
4778 Computations, J. Cryptology 14 (2001) 2, 101-119,
4779 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4782 *) MIPS assembler BIGNUM division bug fix.
4785 *) Disabled incorrect Alpha assembler code.
4788 *) Fix PKCS#7 decode routines so they correctly update the length
4789 after reading an EOC for the EXPLICIT tag.
4791 [This change does not apply to 0.9.7.]
4793 *) Fix bug in PKCS#12 key generation routines. This was triggered
4794 if a 3DES key was generated with a 0 initial byte. Include
4795 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4796 (but broken) behaviour.
4799 *) Enhance bctest to search for a working bc along $PATH and print
4801 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4803 *) Fix memory leaks in err.c: free err_data string if necessary;
4804 don't write to the wrong index in ERR_set_error_data.
4807 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4811 *) Replace rdtsc with _emit statements for VC++ version 5.
4812 [Jeremy Cooper <jeremy@baymoo.org>]
4814 *) Make it possible to reuse SSLv2 sessions.
4817 *) In copy_email() check for >= 0 as a return value for
4818 X509_NAME_get_index_by_NID() since 0 is a valid index.
4819 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4821 *) Avoid coredump with unsupported or invalid public keys by checking if
4822 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4823 PKCS7_verify() fails with non detached data.
4826 *) Don't use getenv in library functions when run as setuid/setgid.
4827 New function OPENSSL_issetugid().
4830 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4831 due to incorrect handling of multi-threading:
4833 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4835 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4837 3. Count how many times MemCheck_off() has been called so that
4838 nested use can be treated correctly. This also avoids
4839 inband-signalling in the previous code (which relied on the
4840 assumption that thread ID 0 is impossible).
4843 *) Add "-rand" option also to s_client and s_server.
4846 *) Fix CPU detection on Irix 6.x.
4847 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4848 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4850 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4853 [This change does not apply to 0.9.7.]
4855 *) Use the cached encoding of an X509_NAME structure rather than
4856 copying it. This is apparently the reason for the libsafe "errors"
4857 but the code is actually correct.
4860 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4861 Bleichenbacher's DSA attack.
4862 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4863 to be set and top=0 forces the highest bit to be set; top=-1 is new
4864 and leaves the highest bit random.
4865 [Ulf Moeller, Bodo Moeller]
4867 *) In the NCONF_...-based implementations for CONF_... queries
4868 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4869 a temporary CONF structure with the data component set to NULL
4870 (which gives segmentation faults in lh_retrieve).
4871 Instead, use NULL for the CONF pointer in CONF_get_string and
4872 CONF_get_number (which may use environment variables) and directly
4873 return NULL from CONF_get_section.
4876 *) Fix potential buffer overrun for EBCDIC.
4879 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4880 keyUsage if basicConstraints absent for a CA.
4883 *) Make SMIME_write_PKCS7() write mail header values with a format that
4884 is more generally accepted (no spaces before the semicolon), since
4885 some programs can't parse those values properly otherwise. Also make
4886 sure BIO's that break lines after each write do not create invalid
4890 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4891 macros previously used would not encode an empty SEQUENCE OF
4892 and break the signature.
4894 [This change does not apply to 0.9.7.]
4896 *) Zero the premaster secret after deriving the master secret in
4900 *) Add some EVP_add_digest_alias registrations (as found in
4901 OpenSSL_add_all_digests()) to SSL_library_init()
4902 aka OpenSSL_add_ssl_algorithms(). This provides improved
4903 compatibility with peers using X.509 certificates
4904 with unconventional AlgorithmIdentifier OIDs.
4907 *) Fix for Irix with NO_ASM.
4908 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4910 *) ./config script fixes.
4911 [Ulf Moeller, Richard Levitte]
4913 *) Fix 'openssl passwd -1'.
4916 *) Change PKCS12_key_gen_asc() so it can cope with non null
4917 terminated strings whose length is passed in the passlen
4918 parameter, for example from PEM callbacks. This was done
4919 by adding an extra length parameter to asc2uni().
4920 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4922 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4923 call failed, free the DSA structure.
4926 *) Fix to uni2asc() to cope with zero length Unicode strings.
4927 These are present in some PKCS#12 files.
4930 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4931 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4932 when writing a 32767 byte record.
4933 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4935 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4936 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4938 (RSA objects have a reference count access to which is protected
4939 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4940 so they are meant to be shared between threads.)
4941 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4942 "Reddie, Steven" <Steven.Reddie@ca.com>]
4944 *) Fix a deadlock in CRYPTO_mem_leaks().
4947 *) Use better test patterns in bntest.
4950 *) rand_win.c fix for Borland C.
4953 *) BN_rshift bugfix for n == 0.
4956 *) Add a 'bctest' script that checks for some known 'bc' bugs
4957 so that 'make test' does not abort just because 'bc' is broken.
4960 *) Store verify_result within SSL_SESSION also for client side to
4961 avoid potential security hole. (Re-used sessions on the client side
4962 always resulted in verify_result==X509_V_OK, not using the original
4963 result of the server certificate verification.)
4966 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4967 SSL3_RT_APPLICATION_DATA, return 0.
4968 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4972 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4973 releases, have been re-implemented by renaming the previous
4974 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4975 and ssl3_read_internal, respectively, and adding 'peek' parameters
4976 to them. The new ssl[23]_{read,peek} functions are calls to
4977 ssl[23]_read_internal with the 'peek' flag set appropriately.
4978 A 'peek' parameter has also been added to ssl3_read_bytes, which
4979 does the actual work for ssl3_read_internal.
4982 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4983 the method-specific "init()" handler. Also clean up ex_data after
4984 calling the method-specific "finish()" handler. Previously, this was
4985 happening the other way round.
4988 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4989 The previous value, 12, was not always sufficient for BN_mod_exp().
4992 *) Make sure that shared libraries get the internal name engine with
4993 the full version number and not just 0. This should mark the
4994 shared libraries as not backward compatible. Of course, this should
4995 be changed again when we can guarantee backward binary compatibility.
4998 *) Fix typo in get_cert_by_subject() in by_dir.c
4999 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5001 *) Rework the system to generate shared libraries:
5003 - Make note of the expected extension for the shared libraries and
5004 if there is a need for symbolic links from for example libcrypto.so.0
5005 to libcrypto.so.0.9.7. There is extended info in Configure for
5008 - Make as few rebuilds of the shared libraries as possible.
5010 - Still avoid linking the OpenSSL programs with the shared libraries.
5012 - When installing, install the shared libraries separately from the
5016 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5018 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5019 and not in SSL_clear because the latter is also used by the
5020 accept/connect functions; previously, the settings made by
5021 SSL_set_read_ahead would be lost during the handshake.
5022 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5024 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5025 Previously, it would create entries for disableed algorithms no
5029 *) Added several new manual pages for SSL_* function.
5032 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5034 *) In ssl23_get_client_hello, generate an error message when faced
5035 with an initial SSL 3.0/TLS record that is too small to contain the
5036 first two bytes of the ClientHello message, i.e. client_version.
5037 (Note that this is a pathologic case that probably has never happened
5038 in real life.) The previous approach was to use the version number
5039 from the record header as a substitute; but our protocol choice
5040 should not depend on that one because it is not authenticated
5041 by the Finished messages.
5044 *) More robust randomness gathering functions for Windows.
5045 [Jeffrey Altman <jaltman@columbia.edu>]
5047 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5048 not set then we don't setup the error code for issuer check errors
5049 to avoid possibly overwriting other errors which the callback does
5050 handle. If an application does set the flag then we assume it knows
5051 what it is doing and can handle the new informational codes
5055 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5056 a general "ANY" type, as such it should be able to decode anything
5057 including tagged types. However it didn't check the class so it would
5058 wrongly interpret tagged types in the same way as their universal
5059 counterpart and unknown types were just rejected. Changed so that the
5060 tagged and unknown types are handled in the same way as a SEQUENCE:
5061 that is the encoding is stored intact. There is also a new type
5062 "V_ASN1_OTHER" which is used when the class is not universal, in this
5063 case we have no idea what the actual type is so we just lump them all
5067 *) On VMS, stdout may very well lead to a file that is written to
5068 in a record-oriented fashion. That means that every write() will
5069 write a separate record, which will be read separately by the
5070 programs trying to read from it. This can be very confusing.
5072 The solution is to put a BIO filter in the way that will buffer
5073 text until a linefeed is reached, and then write everything a
5074 line at a time, so every record written will be an actual line,
5075 not chunks of lines and not (usually doesn't happen, but I've
5076 seen it once) several lines in one record. BIO_f_linebuffer() is
5079 Currently, it's a VMS-only method, because that's where it has
5080 been tested well enough.
5083 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5084 it can return incorrect results.
5085 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5086 but it was in 0.9.6-beta[12].)
5089 *) Disable the check for content being present when verifying detached
5090 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5091 include zero length content when signing messages.
5094 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5095 BIO_ctrl (for BIO pairs).
5098 *) Add DSO method for VMS.
5101 *) Bug fix: Montgomery multiplication could produce results with the
5105 *) Add RPM specification openssl.spec and modify it to build three
5106 packages. The default package contains applications, application
5107 documentation and run-time libraries. The devel package contains
5108 include files, static libraries and function documentation. The
5109 doc package contains the contents of the doc directory. The original
5110 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5113 *) Add a large number of documentation files for many SSL routines.
5114 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5116 *) Add a configuration entry for Sony News 4.
5117 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5119 *) Don't set the two most significant bits to one when generating a
5120 random number < q in the DSA library.
5123 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5124 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5125 the underlying transport is blocking) if a handshake took place.
5126 (The default behaviour is needed by applications such as s_client
5127 and s_server that use select() to determine when to use SSL_read;
5128 but for applications that know in advance when to expect data, it
5129 just makes things more complicated.)
5132 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5136 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5137 work better on such systems.
5138 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5140 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5141 Update PKCS12_parse() so it copies the friendlyName and the
5142 keyid to the certificates aux info.
5145 *) Fix bug in PKCS7_verify() which caused an infinite loop
5146 if there was more than one signature.
5147 [Sven Uszpelkat <su@celocom.de>]
5149 *) Major change in util/mkdef.pl to include extra information
5150 about each symbol, as well as presentig variables as well
5151 as functions. This change means that there's n more need
5152 to rebuild the .num files when some algorithms are excluded.
5155 *) Allow the verify time to be set by an application,
5156 rather than always using the current time.
5159 *) Phase 2 verify code reorganisation. The certificate
5160 verify code now looks up an issuer certificate by a
5161 number of criteria: subject name, authority key id
5162 and key usage. It also verifies self signed certificates
5163 by the same criteria. The main comparison function is
5164 X509_check_issued() which performs these checks.
5166 Lot of changes were necessary in order to support this
5167 without completely rewriting the lookup code.
5169 Authority and subject key identifier are now cached.
5171 The LHASH 'certs' is X509_STORE has now been replaced
5172 by a STACK_OF(X509_OBJECT). This is mainly because an
5173 LHASH can't store or retrieve multiple objects with
5174 the same hash value.
5176 As a result various functions (which were all internal
5177 use only) have changed to handle the new X509_STORE
5178 structure. This will break anything that messed round
5179 with X509_STORE internally.
5181 The functions X509_STORE_add_cert() now checks for an
5182 exact match, rather than just subject name.
5184 The X509_STORE API doesn't directly support the retrieval
5185 of multiple certificates matching a given criteria, however
5186 this can be worked round by performing a lookup first
5187 (which will fill the cache with candidate certificates)
5188 and then examining the cache for matches. This is probably
5189 the best we can do without throwing out X509_LOOKUP
5190 entirely (maybe later...).
5192 The X509_VERIFY_CTX structure has been enhanced considerably.
5194 All certificate lookup operations now go via a get_issuer()
5195 callback. Although this currently uses an X509_STORE it
5196 can be replaced by custom lookups. This is a simple way
5197 to bypass the X509_STORE hackery necessary to make this
5198 work and makes it possible to use more efficient techniques
5199 in future. A very simple version which uses a simple
5200 STACK for its trusted certificate store is also provided
5201 using X509_STORE_CTX_trusted_stack().
5203 The verify_cb() and verify() callbacks now have equivalents
5204 in the X509_STORE_CTX structure.
5206 X509_STORE_CTX also has a 'flags' field which can be used
5207 to customise the verify behaviour.
5210 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5211 excludes S/MIME capabilities.
5214 *) When a certificate request is read in keep a copy of the
5215 original encoding of the signed data and use it when outputing
5216 again. Signatures then use the original encoding rather than
5217 a decoded, encoded version which may cause problems if the
5218 request is improperly encoded.
5221 *) For consistency with other BIO_puts implementations, call
5222 buffer_write(b, ...) directly in buffer_puts instead of calling
5225 In BIO_puts, increment b->num_write as in BIO_write.
5226 [Peter.Sylvester@EdelWeb.fr]
5228 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5229 BN_zero, we may not return a BIGNUM with an array consisting of
5233 *) Avoid calling abort() from within the library when problems are
5234 detected, except if preprocessor symbols have been defined
5235 (such as REF_CHECK, BN_DEBUG etc.).
5238 *) New openssl application 'rsautl'. This utility can be
5239 used for low level RSA operations. DER public key
5240 BIO/fp routines also added.
5243 *) New Configure entry and patches for compiling on QNX 4.
5244 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5246 *) A demo state-machine implementation was sponsored by
5247 Nuron (http://www.nuron.com/) and is now available in
5248 demos/state_machine.
5251 *) New options added to the 'dgst' utility for signature
5252 generation and verification.
5255 *) Unrecognized PKCS#7 content types are now handled via a
5256 catch all ASN1_TYPE structure. This allows unsupported
5257 types to be stored as a "blob" and an application can
5258 encode and decode it manually.
5261 *) Fix various signed/unsigned issues to make a_strex.c
5263 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5265 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5266 length if passed a buffer. ASN1_INTEGER_to_BN failed
5267 if passed a NULL BN and its argument was negative.
5268 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5270 *) Modification to PKCS#7 encoding routines to output definite
5271 length encoding. Since currently the whole structures are in
5272 memory there's not real point in using indefinite length
5273 constructed encoding. However if OpenSSL is compiled with
5274 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5277 *) Added BIO_vprintf() and BIO_vsnprintf().
5280 *) Added more prefixes to parse for in the the strings written
5281 through a logging bio, to cover all the levels that are available
5282 through syslog. The prefixes are now:
5284 PANIC, EMERG, EMR => LOG_EMERG
5285 ALERT, ALR => LOG_ALERT
5286 CRIT, CRI => LOG_CRIT
5287 ERROR, ERR => LOG_ERR
5288 WARNING, WARN, WAR => LOG_WARNING
5289 NOTICE, NOTE, NOT => LOG_NOTICE
5290 INFO, INF => LOG_INFO
5291 DEBUG, DBG => LOG_DEBUG
5293 and as before, if none of those prefixes are present at the
5294 beginning of the string, LOG_ERR is chosen.
5296 On Win32, the LOG_* levels are mapped according to this:
5298 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5299 LOG_WARNING => EVENTLOG_WARNING_TYPE
5300 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5304 *) Made it possible to reconfigure with just the configuration
5305 argument "reconf" or "reconfigure". The command line arguments
5306 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5307 and are retrieved from there when reconfiguring.
5311 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5313 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5316 *) The obj_dat.pl script was messing up the sorting of object
5317 names. The reason was that it compared the quoted version
5318 of strings as a result "OCSP" > "OCSP Signing" because
5319 " > SPACE. Changed script to store unquoted versions of
5320 names and add quotes on output. It was also omitting some
5321 names from the lookup table if they were given a default
5322 value (that is if SN is missing it is given the same
5323 value as LN and vice versa), these are now added on the
5324 grounds that if an object has a name we should be able to
5325 look it up. Finally added warning output when duplicate
5326 short or long names are found.
5329 *) Changes needed for Tandem NSK.
5330 [Scott Uroff <scott@xypro.com>]
5332 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5333 RSA_padding_check_SSLv23(), special padding was never detected
5334 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5335 version rollback attacks was not effective.
5337 In s23_clnt.c, don't use special rollback-attack detection padding
5338 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5339 client; similarly, in s23_srvr.c, don't do the rollback check if
5340 SSL 2.0 is the only protocol enabled in the server.
5343 *) Make it possible to get hexdumps of unprintable data with 'openssl
5344 asn1parse'. By implication, the functions ASN1_parse_dump() and
5345 BIO_dump_indent() are added.
5348 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5349 these print out strings and name structures based on various
5350 flags including RFC2253 support and proper handling of
5351 multibyte characters. Added options to the 'x509' utility
5352 to allow the various flags to be set.
5355 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5356 Also change the functions X509_cmp_current_time() and
5357 X509_gmtime_adj() work with an ASN1_TIME structure,
5358 this will enable certificates using GeneralizedTime in validity
5359 dates to be checked.
5362 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5363 negative public key encodings) on by default,
5364 NO_NEG_PUBKEY_BUG can be set to disable it.
5367 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5368 content octets. An i2c_ASN1_OBJECT is unnecessary because
5369 the encoding can be trivially obtained from the structure.
5372 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5373 not read locks (CRYPTO_r_[un]lock).
5376 *) A first attempt at creating official support for shared
5377 libraries through configuration. I've kept it so the
5378 default is static libraries only, and the OpenSSL programs
5379 are always statically linked for now, but there are
5380 preparations for dynamic linking in place.
5381 This has been tested on Linux and Tru64.
5384 *) Randomness polling function for Win9x, as described in:
5385 Peter Gutmann, Software Generation of Practically Strong
5389 *) Fix so PRNG is seeded in req if using an already existing
5393 *) New options to smime application. -inform and -outform
5394 allow alternative formats for the S/MIME message including
5395 PEM and DER. The -content option allows the content to be
5396 specified separately. This should allow things like Netscape
5397 form signing output easier to verify.
5400 *) Fix the ASN1 encoding of tags using the 'long form'.
5403 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5404 STRING types. These convert content octets to and from the
5405 underlying type. The actual tag and length octets are
5406 already assumed to have been read in and checked. These
5407 are needed because all other string types have virtually
5408 identical handling apart from the tag. By having versions
5409 of the ASN1 functions that just operate on content octets
5410 IMPLICIT tagging can be handled properly. It also allows
5411 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5412 and ASN1_INTEGER are identical apart from the tag.
5415 *) Change the handling of OID objects as follows:
5417 - New object identifiers are inserted in objects.txt, following
5418 the syntax given in objects.README.
5419 - objects.pl is used to process obj_mac.num and create a new
5421 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5424 This is currently kind of a hack, and the perl code in objects.pl
5425 isn't very elegant, but it works as I intended. The simplest way
5426 to check that it worked correctly is to look in obj_dat.h and
5427 check the array nid_objs and make sure the objects haven't moved
5428 around (this is important!). Additions are OK, as well as
5429 consistent name changes.
5432 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5435 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5436 The given file adds to whatever has already been seeded into the
5437 random pool through the RANDFILE configuration file option or
5438 environment variable, or the default random state file.
5441 *) mkstack.pl now sorts each macro group into lexical order.
5442 Previously the output order depended on the order the files
5443 appeared in the directory, resulting in needless rewriting
5447 *) Patches to make OpenSSL compile under Win32 again. Mostly
5448 work arounds for the VC++ problem that it treats func() as
5449 func(void). Also stripped out the parts of mkdef.pl that
5450 added extra typesafe functions: these no longer exist.
5453 *) Reorganisation of the stack code. The macros are now all
5454 collected in safestack.h . Each macro is defined in terms of
5455 a "stack macro" of the form SKM_<name>(type, a, b). The
5456 DEBUG_SAFESTACK is now handled in terms of function casts,
5457 this has the advantage of retaining type safety without the
5458 use of additional functions. If DEBUG_SAFESTACK is not defined
5459 then the non typesafe macros are used instead. Also modified the
5460 mkstack.pl script to handle the new form. Needs testing to see
5461 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5462 the default if no major problems. Similar behaviour for ASN1_SET_OF
5463 and PKCS12_STACK_OF.
5466 *) When some versions of IIS use the 'NET' form of private key the
5467 key derivation algorithm is different. Normally MD5(password) is
5468 used as a 128 bit RC4 key. In the modified case
5469 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5470 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5471 as the old Netscape_RSA functions except they have an additional
5472 'sgckey' parameter which uses the modified algorithm. Also added
5473 an -sgckey command line option to the rsa utility. Thanks to
5474 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5475 algorithm to openssl-dev.
5478 *) The evp_local.h macros were using 'c.##kname' which resulted in
5479 invalid expansion on some systems (SCO 5.0.5 for example).
5480 Corrected to 'c.kname'.
5481 [Phillip Porch <root@theporch.com>]
5483 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5484 a STACK of email addresses from a certificate or request, these look
5485 in the subject name and the subject alternative name extensions and
5486 omit any duplicate addresses.
5489 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5490 This makes DSA verification about 2 % faster.
5493 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5494 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5495 plus overhead for 1024 bit moduli).
5496 This makes exponentiations about 0.5 % faster for 1024 bit
5497 exponents (as measured by "openssl speed rsa2048").
5500 *) Rename memory handling macros to avoid conflicts with other
5502 Malloc => OPENSSL_malloc
5503 Malloc_locked => OPENSSL_malloc_locked
5504 Realloc => OPENSSL_realloc
5505 Free => OPENSSL_free
5508 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5509 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5512 *) CygWin32 support.
5513 [John Jarvie <jjarvie@newsguy.com>]
5515 *) The type-safe stack code has been rejigged. It is now only compiled
5516 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5517 by default all type-specific stack functions are "#define"d back to
5518 standard stack functions. This results in more streamlined output
5519 but retains the type-safety checking possibilities of the original
5523 *) The STACK code has been cleaned up, and certain type declarations
5524 that didn't make a lot of sense have been brought in line. This has
5525 also involved a cleanup of sorts in safestack.h to more correctly
5526 map type-safe stack functions onto their plain stack counterparts.
5527 This work has also resulted in a variety of "const"ifications of
5528 lots of the code, especially "_cmp" operations which should normally
5529 be prototyped with "const" parameters anyway.
5532 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5533 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5534 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5535 where all of 'md' is used each time the PRNG is used, but 'state'
5536 is used only indexed by a cyclic counter. As entropy may not be
5537 well distributed from the beginning, 'md' is important as a
5538 chaining variable. However, the output function chains only half
5539 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5540 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5541 in all of 'state' being rewritten, with the new values depending
5542 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5545 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5546 the handshake is continued after ssl_verify_cert_chain();
5547 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5548 can lead to 'unexplainable' connection aborts later.
5549 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5551 *) Major EVP API cipher revision.
5552 Add hooks for extra EVP features. This allows various cipher
5553 parameters to be set in the EVP interface. Support added for variable
5554 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5555 setting of RC2 and RC5 parameters.
5557 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5560 Remove lots of duplicated code from the EVP library. For example *every*
5561 cipher init() function handles the 'iv' in the same way according to the
5562 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5563 for CFB and OFB modes they zero ctx->num.
5565 New functionality allows removal of S/MIME code RC2 hack.
5567 Most of the routines have the same form and so can be declared in terms
5570 By shifting this to the top level EVP_CipherInit() it can be removed from
5571 all individual ciphers. If the cipher wants to handle IVs or keys
5572 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5575 Change lots of functions like EVP_EncryptUpdate() to now return a
5576 value: although software versions of the algorithms cannot fail
5577 any installed hardware versions can.
5580 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5581 this option is set, tolerate broken clients that send the negotiated
5582 protocol version number instead of the requested protocol version
5586 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5587 i.e. non-zero for export ciphersuites, zero otherwise.
5588 Previous versions had this flag inverted, inconsistent with
5589 rsa_tmp_cb (..._TMP_RSA_CB).
5590 [Bodo Moeller; problem reported by Amit Chopra]
5592 *) Add missing DSA library text string. Work around for some IIS
5593 key files with invalid SEQUENCE encoding.
5596 *) Add a document (doc/standards.txt) that list all kinds of standards
5597 and so on that are implemented in OpenSSL.
5600 *) Enhance c_rehash script. Old version would mishandle certificates
5601 with the same subject name hash and wouldn't handle CRLs at all.
5602 Added -fingerprint option to crl utility, to support new c_rehash
5606 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5609 *) Fix for SSL server purpose checking. Server checking was
5610 rejecting certificates which had extended key usage present
5611 but no ssl client purpose.
5612 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5614 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5615 is a little unclear about how a blank password is handled.
5616 Since the password in encoded as a BMPString with terminating
5617 double NULL a zero length password would end up as just the
5618 double NULL. However no password at all is different and is
5619 handled differently in the PKCS#12 key generation code. NS
5620 treats a blank password as zero length. MSIE treats it as no
5621 password on export: but it will try both on import. We now do
5622 the same: PKCS12_parse() tries zero length and no password if
5623 the password is set to "" or NULL (NULL is now a valid password:
5624 it wasn't before) as does the pkcs12 application.
5627 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5628 perror when PEM_read_bio_X509_REQ fails, the error message must
5629 be obtained from the error queue.
5632 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5633 it in ERR_remove_state if appropriate, and change ERR_get_state
5634 accordingly to avoid race conditions (this is necessary because
5635 thread_hash is no longer constant once set).
5638 *) Bugfix for linux-elf makefile.one.
5641 *) RSA_get_default_method() will now cause a default
5642 RSA_METHOD to be chosen if one doesn't exist already.
5643 Previously this was only set during a call to RSA_new()
5644 or RSA_new_method(NULL) meaning it was possible for
5645 RSA_get_default_method() to return NULL.
5648 *) Added native name translation to the existing DSO code
5649 that will convert (if the flag to do so is set) filenames
5650 that are sufficiently small and have no path information
5651 into a canonical native form. Eg. "blah" converted to
5652 "libblah.so" or "blah.dll" etc.
5655 *) New function ERR_error_string_n(e, buf, len) which is like
5656 ERR_error_string(e, buf), but writes at most 'len' bytes
5657 including the 0 terminator. For ERR_error_string_n, 'buf'
5659 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5661 *) CONF library reworked to become more general. A new CONF
5662 configuration file reader "class" is implemented as well as a
5663 new functions (NCONF_*, for "New CONF") to handle it. The now
5664 old CONF_* functions are still there, but are reimplemented to
5665 work in terms of the new functions. Also, a set of functions
5666 to handle the internal storage of the configuration data is
5667 provided to make it easier to write new configuration file
5668 reader "classes" (I can definitely see something reading a
5669 configuration file in XML format, for example), called _CONF_*,
5670 or "the configuration storage API"...
5672 The new configuration file reading functions are:
5674 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5675 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5677 NCONF_default, NCONF_WIN32
5679 NCONF_dump_fp, NCONF_dump_bio
5681 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5682 NCONF_new creates a new CONF object. This works in the same way
5683 as other interfaces in OpenSSL, like the BIO interface.
5684 NCONF_dump_* dump the internal storage of the configuration file,
5685 which is useful for debugging. All other functions take the same
5686 arguments as the old CONF_* functions wth the exception of the
5687 first that must be a `CONF *' instead of a `LHASH *'.
5689 To make it easer to use the new classes with the old CONF_* functions,
5690 the function CONF_set_default_method is provided.
5693 *) Add '-tls1' option to 'openssl ciphers', which was already
5694 mentioned in the documentation but had not been implemented.
5695 (This option is not yet really useful because even the additional
5696 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5699 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5700 OpenSSL-based applications) load shared libraries and bind to
5701 them in a portable way.
5702 [Geoff Thorpe, with contributions from Richard Levitte]
5704 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5706 *) Make sure _lrotl and _lrotr are only used with MSVC.
5708 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5709 (the default implementation of RAND_status).
5711 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5712 to '-clrext' (= clear extensions), as intended and documented.
5713 [Bodo Moeller; inconsistency pointed out by Michael Attili
5716 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5717 was larger than the MD block size.
5718 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5720 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5721 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5722 using the passed key: if the passed key was a private key the result
5723 of X509_print(), for example, would be to print out all the private key
5727 *) des_quad_cksum() byte order bug fix.
5728 [Ulf Möller, using the problem description in krb4-0.9.7, where
5729 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5731 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5733 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5735 *) For easily testing in shell scripts whether some command
5736 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5737 returns with exit code 0 iff no command of the given name is available.
5738 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5739 the output goes to stdout and nothing is printed to stderr.
5740 Additional arguments are always ignored.
5742 Since for each cipher there is a command of the same name,
5743 the 'no-cipher' compilation switches can be tested this way.
5745 ('openssl no-XXX' is not able to detect pseudo-commands such
5746 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5749 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5752 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5753 is set; it will be thrown away anyway because each handshake creates
5755 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5756 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5757 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5758 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5761 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5762 'Q' and 'R' lose their special meanings (quit/renegotiate).
5763 This is part of what -quiet does; unlike -quiet, -ign_eof
5764 does not suppress any output.
5767 *) Add compatibility options to the purpose and trust code. The
5768 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5769 accepts a certificate or CA, this was the previous behaviour,
5770 with all the associated security issues.
5772 X509_TRUST_COMPAT is the old trust behaviour: only and
5773 automatically trust self signed roots in certificate store. A
5774 new trust setting X509_TRUST_DEFAULT is used to specify that
5775 a purpose has no associated trust setting and it should instead
5776 use the value in the default purpose.
5779 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5780 and fix a memory leak.
5783 *) In util/mkerr.pl (which implements 'make errors'), preserve
5784 reason strings from the previous version of the .c file, as
5785 the default to have only downcase letters (and digits) in
5786 automatically generated reasons codes is not always appropriate.
5789 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5790 using strerror. Previously, ERR_reason_error_string() returned
5791 library names as reason strings for SYSerr; but SYSerr is a special
5792 case where small numbers are errno values, not library numbers.
5795 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5796 converts DSA parameters into DH parameters. (When creating parameters,
5797 DSA_generate_parameters is used.)
5800 *) Include 'length' (recommended exponent length) in C code generated
5801 by 'openssl dhparam -C'.
5804 *) The second argument to set_label in perlasm was already being used
5805 so couldn't be used as a "file scope" flag. Moved to third argument
5809 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5810 instead of RAND_bytes for encryption IVs and salts.
5813 *) Include RAND_status() into RAND_METHOD instead of implementing
5814 it only for md_rand.c Otherwise replacing the PRNG by calling
5815 RAND_set_rand_method would be impossible.
5818 *) Don't let DSA_generate_key() enter an infinite loop if the random
5819 number generation fails.
5822 *) New 'rand' application for creating pseudo-random output.
5825 *) Added configuration support for Linux/IA64
5826 [Rolf Haberrecker <rolf@suse.de>]
5828 *) Assembler module support for Mingw32.
5831 *) Shared library support for HPUX (in shlib/).
5832 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5834 *) Shared library support for Solaris gcc.
5835 [Lutz Behnke <behnke@trustcenter.de>]
5837 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5839 *) PKCS7_encrypt() was adding text MIME headers twice because they
5840 were added manually and by SMIME_crlf_copy().
5843 *) In bntest.c don't call BN_rand with zero bits argument.
5844 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5846 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5847 case was implemented. This caused BN_div_recp() to fail occasionally.
5850 *) Add an optional second argument to the set_label() in the perl
5851 assembly language builder. If this argument exists and is set
5852 to 1 it signals that the assembler should use a symbol whose
5853 scope is the entire file, not just the current function. This
5854 is needed with MASM which uses the format label:: for this scope.
5855 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5857 *) Change the ASN1 types so they are typedefs by default. Before
5858 almost all types were #define'd to ASN1_STRING which was causing
5859 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5863 *) Change names of new functions to the new get1/get0 naming
5864 convention: After 'get1', the caller owns a reference count
5865 and has to call ..._free; 'get0' returns a pointer to some
5866 data structure without incrementing reference counters.
5867 (Some of the existing 'get' functions increment a reference
5868 counter, some don't.)
5869 Similarly, 'set1' and 'add1' functions increase reference
5870 counters or duplicate objects.
5873 *) Allow for the possibility of temp RSA key generation failure:
5874 the code used to assume it always worked and crashed on failure.
5877 *) Fix potential buffer overrun problem in BIO_printf().
5878 [Ulf Möller, using public domain code by Patrick Powell; problem
5879 pointed out by David Sacerdote <das33@cornell.edu>]
5881 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5882 RAND_egd() and RAND_status(). In the command line application,
5883 the EGD socket can be specified like a seed file using RANDFILE
5887 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5888 Some CAs (e.g. Verisign) distribute certificates in this form.
5891 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5892 list to exclude them. This means that no special compilation option
5893 is needed to use anonymous DH: it just needs to be included in the
5897 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5898 EVP_MD_type. The old functionality is available in a new macro called
5899 EVP_MD_md(). Change code that uses it and update docs.
5902 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5903 where the 'void *' argument is replaced by a function pointer argument.
5904 Previously 'void *' was abused to point to functions, which works on
5905 many platforms, but is not correct. As these functions are usually
5906 called by macros defined in OpenSSL header files, most source code
5907 should work without changes.
5910 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5911 sections with information on -D... compiler switches used for
5912 compiling the library so that applications can see them. To enable
5913 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5914 must be defined. E.g.,
5915 #define OPENSSL_ALGORITHM_DEFINES
5916 #include <openssl/opensslconf.h>
5917 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5918 [Richard Levitte, Ulf and Bodo Möller]
5920 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5924 *) Change the 'other' type in certificate aux info to a STACK_OF
5925 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5926 the required ASN1 format: arbitrary types determined by an OID.
5929 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5930 argument to 'req'. This is not because the function is newer or
5931 better than others it just uses the work 'NEW' in the certificate
5932 request header lines. Some software needs this.
5935 *) Reorganise password command line arguments: now passwords can be
5936 obtained from various sources. Delete the PEM_cb function and make
5937 it the default behaviour: i.e. if the callback is NULL and the
5938 usrdata argument is not NULL interpret it as a null terminated pass
5939 phrase. If usrdata and the callback are NULL then the pass phrase
5940 is prompted for as usual.
5943 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5944 the support is automatically enabled. The resulting binaries will
5945 autodetect the card and use it if present.
5946 [Ben Laurie and Compaq Inc.]
5948 *) Work around for Netscape hang bug. This sends certificate request
5949 and server done in one record. Since this is perfectly legal in the
5950 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5951 the bugs/SSLv3 entry for more info.
5954 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5957 *) Add -rand argument to smime and pkcs12 applications and read/write
5961 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5964 *) Add command line password options to the remaining applications.
5967 *) Bug fix for BN_div_recp() for numerators with an even number of
5971 *) More tests in bntest.c, and changed test_bn output.
5974 *) ./config recognizes MacOS X now.
5977 *) Bug fix for BN_div() when the first words of num and divsor are
5978 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5981 *) Add support for various broken PKCS#8 formats, and command line
5982 options to produce them.
5985 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5986 get temporary BIGNUMs from a BN_CTX.
5989 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5993 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5994 include a #define from the old name to the new. The original intent
5995 was that statically linked binaries could for example just call
5996 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5997 link with digests. This never worked becayse SSLeay_add_all_digests()
5998 and SSLeay_add_all_ciphers() were in the same source file so calling
5999 one would link with the other. They are now in separate source files.
6002 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6005 *) Use a less unusual form of the Miller-Rabin primality test (it used
6006 a binary algorithm for exponentiation integrated into the Miller-Rabin
6007 loop, our standard modexp algorithms are faster).
6010 *) Support for the EBCDIC character set completed.
6011 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6013 *) Source code cleanups: use const where appropriate, eliminate casts,
6014 use void * instead of char * in lhash.
6017 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6018 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6019 this the server could overwrite ephemeral keys that the client
6023 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6024 using 50 iterations of the Rabin-Miller test.
6026 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6027 iterations of the Rabin-Miller test as required by the appendix
6028 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6029 As BN_is_prime_fasttest includes trial division, DSA parameter
6030 generation becomes much faster.
6032 This implies a change for the callback functions in DSA_is_prime
6033 and DSA_generate_parameters: The callback function is called once
6034 for each positive witness in the Rabin-Miller test, not just
6035 occasionally in the inner loop; and the parameters to the
6036 callback function now provide an iteration count for the outer
6037 loop rather than for the current invocation of the inner loop.
6038 DSA_generate_parameters additionally can call the callback
6039 function with an 'iteration count' of -1, meaning that a
6040 candidate has passed the trial division test (when q is generated
6041 from an application-provided seed, trial division is skipped).
6044 *) New function BN_is_prime_fasttest that optionally does trial
6045 division before starting the Rabin-Miller test and has
6046 an additional BN_CTX * argument (whereas BN_is_prime always
6047 has to allocate at least one BN_CTX).
6048 'callback(1, -1, cb_arg)' is called when a number has passed the
6049 trial division stage.
6052 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6056 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6059 *) New function BN_pseudo_rand().
6062 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6063 bignum version of BN_from_montgomery() with the working code from
6064 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6068 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6069 made it impossible to use the same SSL_SESSION data structure in
6070 SSL2 clients in multiple threads.
6073 *) The return value of RAND_load_file() no longer counts bytes obtained
6074 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6075 to seed the PRNG (previously an explicit byte count was required).
6076 [Ulf Möller, Bodo Möller]
6078 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6079 used (char *) instead of (void *) and had casts all over the place.
6082 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6085 *) Retain source code compatibility for BN_prime_checks macro:
6086 BN_is_prime(..., BN_prime_checks, ...) now uses
6087 BN_prime_checks_for_size to determine the appropriate number of
6088 Rabin-Miller iterations.
6091 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6092 DH_CHECK_P_NOT_SAFE_PRIME.
6093 (Check if this is true? OpenPGP calls them "strong".)
6096 *) Merge the functionality of "dh" and "gendh" programs into a new program
6097 "dhparam". The old programs are retained for now but will handle DH keys
6098 (instead of parameters) in future.
6101 *) Make the ciphers, s_server and s_client programs check the return values
6102 when a new cipher list is set.
6105 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6106 ciphers. Before when the 56bit ciphers were enabled the sorting was
6109 The syntax for the cipher sorting has been extended to support sorting by
6110 cipher-strength (using the strength_bits hard coded in the tables).
6111 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6113 Fix a bug in the cipher-command parser: when supplying a cipher command
6114 string with an "undefined" symbol (neither command nor alphanumeric
6115 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6116 an error is flagged.
6118 Due to the strength-sorting extension, the code of the
6119 ssl_create_cipher_list() function was completely rearranged. I hope that
6120 the readability was also increased :-)
6121 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6123 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6124 for the first serial number and places 2 in the serial number file. This
6125 avoids problems when the root CA is created with serial number zero and
6126 the first user certificate has the same issuer name and serial number
6130 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6131 the new code. Add documentation for this stuff.
6134 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6135 X509_*() to X509at_*() on the grounds that they don't handle X509
6136 structures and behave in an analagous way to the X509v3 functions:
6137 they shouldn't be called directly but wrapper functions should be used
6140 So we also now have some wrapper functions that call the X509at functions
6141 when passed certificate requests. (TO DO: similar things can be done with
6142 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6143 things. Some of these need some d2i or i2d and print functionality
6144 because they handle more complex structures.)
6147 *) Add missing #ifndefs that caused missing symbols when building libssl
6148 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6149 NO_RSA in ssl/s2*.c.
6150 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6152 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6153 has a return value which indicates the quality of the random data
6154 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6155 error queue. New function RAND_pseudo_bytes() generates output that is
6156 guaranteed to be unique but not unpredictable. RAND_add is like
6157 RAND_seed, but takes an extra argument for an entropy estimate
6158 (RAND_seed always assumes full entropy).
6161 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6162 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6163 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6164 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6165 false-positive rate of at most 2^-80 for random input.
6168 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6171 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6172 in the 0.9.5 release), this returns the chain
6173 from an X509_CTX structure with a dup of the stack and all
6174 the X509 reference counts upped: so the stack will exist
6175 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6178 Also make SSL_SESSION_print() print out the verify return
6182 *) Add manpage for the pkcs12 command. Also change the default
6183 behaviour so MAC iteration counts are used unless the new
6184 -nomaciter option is used. This improves file security and
6185 only older versions of MSIE (4.0 for example) need it.
6188 *) Honor the no-xxx Configure options when creating .DEF files.
6191 *) Add PKCS#10 attributes to field table: challengePassword,
6192 unstructuredName and unstructuredAddress. These are taken from
6193 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6194 international characters are used.
6196 More changes to X509_ATTRIBUTE code: allow the setting of types
6197 based on strings. Remove the 'loc' parameter when adding
6198 attributes because these will be a SET OF encoding which is sorted
6202 *) Initial changes to the 'req' utility to allow request generation
6203 automation. This will allow an application to just generate a template
6204 file containing all the field values and have req construct the
6207 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6208 used all over the place including certificate requests and PKCS#7
6209 structures. They are currently handled manually where necessary with
6210 some primitive wrappers for PKCS#7. The new functions behave in a
6211 manner analogous to the X509 extension functions: they allow
6212 attributes to be looked up by NID and added.
6214 Later something similar to the X509V3 code would be desirable to
6215 automatically handle the encoding, decoding and printing of the
6216 more complex types. The string types like challengePassword can
6217 be handled by the string table functions.
6219 Also modified the multi byte string table handling. Now there is
6220 a 'global mask' which masks out certain types. The table itself
6221 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6222 is useful when for example there is only one permissible type
6223 (as in countryName) and using the mask might result in no valid
6227 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6228 SSL_get_peer_finished to allow applications to obtain the latest
6229 Finished messages sent to the peer or expected from the peer,
6230 respectively. (SSL_get_peer_finished is usually the Finished message
6231 actually received from the peer, otherwise the protocol will be aborted.)
6233 As the Finished message are message digests of the complete handshake
6234 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6235 be used for external authentication procedures when the authentication
6236 provided by SSL/TLS is not desired or is not enough.
6239 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6240 the host supports BWX extension and if Compaq C is present on the
6241 $PATH. Just exploiting of the BWX extension results in 20-30%
6242 performance kick for some algorithms, e.g. DES and RC4 to mention
6243 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6247 *) Add support for MS "fast SGC". This is arguably a violation of the
6248 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6249 weak crypto and after checking the certificate is SGC a second one
6250 with strong crypto. MS SGC stops the first handshake after receiving
6251 the server certificate message and sends a second client hello. Since
6252 a server will typically do all the time consuming operations before
6253 expecting any further messages from the client (server key exchange
6254 is the most expensive) there is little difference between the two.
6256 To get OpenSSL to support MS SGC we have to permit a second client
6257 hello message after we have sent server done. In addition we have to
6258 reset the MAC if we do get this second client hello.
6261 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6262 if a DER encoded private key is RSA or DSA traditional format. Changed
6263 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6264 format DER encoded private key. Newer code should use PKCS#8 format which
6265 has the key type encoded in the ASN1 structure. Added DER private key
6266 support to pkcs8 application.
6269 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6270 ciphersuites has been selected (as required by the SSL 3/TLS 1
6271 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6272 is set, we interpret this as a request to violate the specification
6273 (the worst that can happen is a handshake failure, and 'correct'
6274 behaviour would result in a handshake failure anyway).
6277 *) In SSL_CTX_add_session, take into account that there might be multiple
6278 SSL_SESSION structures with the same session ID (e.g. when two threads
6279 concurrently obtain them from an external cache).
6280 The internal cache can handle only one SSL_SESSION with a given ID,
6281 so if there's a conflict, we now throw out the old one to achieve
6285 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6286 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6287 some routines that use cipher OIDs: some ciphers do not have OIDs
6288 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6292 *) Simplify the trust setting structure and code. Now we just have
6293 two sequences of OIDs for trusted and rejected settings. These will
6294 typically have values the same as the extended key usage extension
6295 and any application specific purposes.
6297 The trust checking code now has a default behaviour: it will just
6298 check for an object with the same NID as the passed id. Functions can
6299 be provided to override either the default behaviour or the behaviour
6300 for a given id. SSL client, server and email already have functions
6301 in place for compatibility: they check the NID and also return "trusted"
6302 if the certificate is self signed.
6305 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6306 traditional format into an EVP_PKEY structure.
6309 *) Add a password callback function PEM_cb() which either prompts for
6310 a password if usr_data is NULL or otherwise assumes it is a null
6311 terminated password. Allow passwords to be passed on command line
6312 environment or config files in a few more utilities.
6315 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6316 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6317 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6318 Update documentation.
6321 *) Support for ASN1 "NULL" type. This could be handled before by using
6322 ASN1_TYPE but there wasn't any function that would try to read a NULL
6323 and produce an error if it couldn't. For compatibility we also have
6324 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6325 don't allocate anything because they don't need to.
6328 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6330 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6332 *) Rebuild of the memory allocation routines used by OpenSSL code and
6333 possibly others as well. The purpose is to make an interface that
6334 provide hooks so anyone can build a separate set of allocation and
6335 deallocation routines to be used by OpenSSL, for example memory
6336 pool implementations, or something else, which was previously hard
6337 since Malloc(), Realloc() and Free() were defined as macros having
6338 the values malloc, realloc and free, respectively (except for Win32
6339 compilations). The same is provided for memory debugging code.
6340 OpenSSL already comes with functionality to find memory leaks, but
6341 this gives people a chance to debug other memory problems.
6343 With these changes, a new set of functions and macros have appeared:
6345 CRYPTO_set_mem_debug_functions() [F]
6346 CRYPTO_get_mem_debug_functions() [F]
6347 CRYPTO_dbg_set_options() [F]
6348 CRYPTO_dbg_get_options() [F]
6349 CRYPTO_malloc_debug_init() [M]
6351 The memory debug functions are NULL by default, unless the library
6352 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6353 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6354 gives the standard debugging functions that come with OpenSSL) or
6355 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6356 provided by the library user) must be used. When the standard
6357 debugging functions are used, CRYPTO_dbg_set_options can be used to
6358 request additional information:
6359 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6360 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6362 Also, things like CRYPTO_set_mem_functions will always give the
6363 expected result (the new set of functions is used for allocation
6364 and deallocation) at all times, regardless of platform and compiler
6367 To finish it up, some functions that were never use in any other
6368 way than through macros have a new API and new semantic:
6371 CRYPTO_dbg_realloc()
6374 All macros of value have retained their old syntax.
6375 [Richard Levitte and Bodo Moeller]
6377 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6378 ordering of SMIMECapabilities wasn't in "strength order" and there
6379 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6383 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6384 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6385 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6387 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6388 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6389 functionality to handle multipart/signed properly) and a utility
6390 called 'smime' to call all this stuff. This is based on code I
6391 originally wrote for Celo who have kindly allowed it to be
6392 included in OpenSSL.
6395 *) Add variants des_set_key_checked and des_set_key_unchecked of
6396 des_set_key (aka des_key_sched). Global variable des_check_key
6397 decides which of these is called by des_set_key; this way
6398 des_check_key behaves as it always did, but applications and
6399 the library itself, which was buggy for des_check_key == 1,
6400 have a cleaner way to pick the version they need.
6403 *) New function PKCS12_newpass() which changes the password of a
6407 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6408 dynamic mix. In both cases the ids can be used as an index into the
6409 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6410 functions so they accept a list of the field values and the
6411 application doesn't need to directly manipulate the X509_TRUST
6415 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6419 *) Modify the way the V3 extension code looks up extensions. This now
6420 works in a similar way to the object code: we have some "standard"
6421 extensions in a static table which is searched with OBJ_bsearch()
6422 and the application can add dynamic ones if needed. The file
6423 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6424 updated whenever a new extension is added to the core code and kept
6425 in ext_nid order. There is a simple program 'tabtest.c' which checks
6426 this. New extensions are not added too often so this file can readily
6427 be maintained manually.
6429 There are two big advantages in doing things this way. The extensions
6430 can be looked up immediately and no longer need to be "added" using
6431 X509V3_add_standard_extensions(): this function now does nothing.
6432 [Side note: I get *lots* of email saying the extension code doesn't
6433 work because people forget to call this function]
6434 Also no dynamic allocation is done unless new extensions are added:
6435 so if we don't add custom extensions there is no need to call
6436 X509V3_EXT_cleanup().
6439 *) Modify enc utility's salting as follows: make salting the default. Add a
6440 magic header, so unsalted files fail gracefully instead of just decrypting
6441 to garbage. This is because not salting is a big security hole, so people
6442 should be discouraged from doing it.
6445 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6446 digest to be passed on the command line but it only used this
6447 parameter when signing a certificate. Modified so all relevant
6448 operations are affected by the digest parameter including the
6449 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6450 DSA key was used because it didn't fix the digest.
6453 *) Initial certificate chain verify code. Currently tests the untrusted
6454 certificates for consistency with the verify purpose (which is set
6455 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6457 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6458 this is because it will reject chains with invalid extensions whereas
6459 every previous version of OpenSSL and SSLeay made no checks at all.
6461 Trust code: checks the root CA for the relevant trust settings. Trust
6462 settings have an initial value consistent with the verify purpose: e.g.
6463 if the verify purpose is for SSL client use it expects the CA to be
6464 trusted for SSL client use. However the default value can be changed to
6465 permit custom trust settings: one example of this would be to only trust
6466 certificates from a specific "secure" set of CAs.
6468 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6469 which should be used for version portability: especially since the
6470 verify structure is likely to change more often now.
6472 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6473 to set them. If not set then assume SSL clients will verify SSL servers
6476 Two new options to the verify program: -untrusted allows a set of
6477 untrusted certificates to be passed in and -purpose which sets the
6478 intended purpose of the certificate. If a purpose is set then the
6479 new chain verify code is used to check extension consistency.
6482 *) Support for the authority information access extension.
6485 *) Modify RSA and DSA PEM read routines to transparently handle
6486 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6487 public keys in a format compatible with certificate
6488 SubjectPublicKeyInfo structures. Unfortunately there were already
6489 functions called *_PublicKey_* which used various odd formats so
6490 these are retained for compatibility: however the DSA variants were
6491 never in a public release so they have been deleted. Changed dsa/rsa
6492 utilities to handle the new format: note no releases ever handled public
6493 keys so we should be OK.
6495 The primary motivation for this change is to avoid the same fiasco
6496 that dogs private keys: there are several incompatible private key
6497 formats some of which are standard and some OpenSSL specific and
6498 require various evil hacks to allow partial transparent handling and
6499 even then it doesn't work with DER formats. Given the option anything
6500 other than PKCS#8 should be dumped: but the other formats have to
6501 stay in the name of compatibility.
6503 With public keys and the benefit of hindsight one standard format
6504 is used which works with EVP_PKEY, RSA or DSA structures: though
6505 it clearly returns an error if you try to read the wrong kind of key.
6507 Added a -pubkey option to the 'x509' utility to output the public key.
6508 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6509 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6510 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6511 that do the same as the EVP_PKEY_assign_*() except they up the
6512 reference count of the added key (they don't "swallow" the
6516 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6517 CRLs would fail if the file contained no certificates or no CRLs:
6518 added a new function to read in both types and return the number
6519 read: this means that if none are read it will be an error. The
6520 DER versions of the certificate and CRL reader would always fail
6521 because it isn't possible to mix certificates and CRLs in DER format
6522 without choking one or the other routine. Changed this to just read
6523 a certificate: this is the best we can do. Also modified the code
6524 in apps/verify.c to take notice of return codes: it was previously
6525 attempting to read in certificates from NULL pointers and ignoring
6526 any errors: this is one reason why the cert and CRL reader seemed
6527 to work. It doesn't check return codes from the default certificate
6528 routines: these may well fail if the certificates aren't installed.
6531 *) Code to support otherName option in GeneralName.
6534 *) First update to verify code. Change the verify utility
6535 so it warns if it is passed a self signed certificate:
6536 for consistency with the normal behaviour. X509_verify
6537 has been modified to it will now verify a self signed
6538 certificate if *exactly* the same certificate appears
6539 in the store: it was previously impossible to trust a
6540 single self signed certificate. This means that:
6541 openssl verify ss.pem
6542 now gives a warning about a self signed certificate but
6543 openssl verify -CAfile ss.pem ss.pem
6547 *) For servers, store verify_result in SSL_SESSION data structure
6548 (and add it to external session representation).
6549 This is needed when client certificate verifications fails,
6550 but an application-provided verification callback (set by
6551 SSL_CTX_set_cert_verify_callback) allows accepting the session
6552 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6553 but returns 1): When the session is reused, we have to set
6554 ssl->verify_result to the appropriate error code to avoid
6556 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6558 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6559 case in PKCS7_dataInit() where the signed PKCS7 structure
6560 didn't contain any existing data because it was being created.
6561 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6563 *) Add a salt to the key derivation routines in enc.c. This
6564 forms the first 8 bytes of the encrypted file. Also add a
6565 -S option to allow a salt to be input on the command line.
6568 *) New function X509_cmp(). Oddly enough there wasn't a function
6569 to compare two certificates. We do this by working out the SHA1
6570 hash and comparing that. X509_cmp() will be needed by the trust
6574 *) SSL_get1_session() is like SSL_get_session(), but increments
6575 the reference count in the SSL_SESSION returned.
6576 [Geoff Thorpe <geoff@eu.c2.net>]
6578 *) Fix for 'req': it was adding a null to request attributes.
6579 Also change the X509_LOOKUP and X509_INFO code to handle
6580 certificate auxiliary information.
6583 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6587 *) Add the possibility to add extra information to the memory leak
6588 detecting output, to form tracebacks, showing from where each
6589 allocation was originated: CRYPTO_push_info("constant string") adds
6590 the string plus current file name and line number to a per-thread
6591 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6592 is like calling CYRPTO_pop_info() until the stack is empty.
6593 Also updated memory leak detection code to be multi-thread-safe.
6596 *) Add options -text and -noout to pkcs7 utility and delete the
6597 encryption options which never did anything. Update docs.
6600 *) Add options to some of the utilities to allow the pass phrase
6601 to be included on either the command line (not recommended on
6602 OSes like Unix) or read from the environment. Update the
6603 manpages and fix a few bugs.
6606 *) Add a few manpages for some of the openssl commands.
6609 *) Fix the -revoke option in ca. It was freeing up memory twice,
6610 leaking and not finding already revoked certificates.
6613 *) Extensive changes to support certificate auxiliary information.
6614 This involves the use of X509_CERT_AUX structure and X509_AUX
6615 functions. An X509_AUX function such as PEM_read_X509_AUX()
6616 can still read in a certificate file in the usual way but it
6617 will also read in any additional "auxiliary information". By
6618 doing things this way a fair degree of compatibility can be
6619 retained: existing certificates can have this information added
6620 using the new 'x509' options.
6622 Current auxiliary information includes an "alias" and some trust
6623 settings. The trust settings will ultimately be used in enhanced
6624 certificate chain verification routines: currently a certificate
6625 can only be trusted if it is self signed and then it is trusted
6629 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6630 The problem was that one of the replacement routines had not been working
6631 since SSLeay releases. For now the offending routine has been replaced
6632 with non-optimised assembler. Even so, this now gives around 95%
6633 performance improvement for 1024 bit RSA signs.
6636 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6637 handling. Most clients have the effective key size in bits equal to
6638 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6639 A few however don't do this and instead use the size of the decrypted key
6640 to determine the RC2 key length and the AlgorithmIdentifier to determine
6641 the effective key length. In this case the effective key length can still
6642 be 40 bits but the key length can be 168 bits for example. This is fixed
6643 by manually forcing an RC2 key into the EVP_PKEY structure because the
6644 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6645 the key length and effective key length are equal.
6648 *) Add a bunch of functions that should simplify the creation of
6649 X509_NAME structures. Now you should be able to do:
6650 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6651 and have it automatically work out the correct field type and fill in
6652 the structures. The more adventurous can try:
6653 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6654 and it will (hopefully) work out the correct multibyte encoding.
6657 *) Change the 'req' utility to use the new field handling and multibyte
6658 copy routines. Before the DN field creation was handled in an ad hoc
6659 way in req, ca, and x509 which was rather broken and didn't support
6660 BMPStrings or UTF8Strings. Since some software doesn't implement
6661 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6662 using the dirstring_type option. See the new comment in the default
6663 openssl.cnf for more info.
6666 *) Make crypto/rand/md_rand.c more robust:
6667 - Assure unique random numbers after fork().
6668 - Make sure that concurrent threads access the global counter and
6669 md serializably so that we never lose entropy in them
6670 or use exactly the same state in multiple threads.
6671 Access to the large state is not always serializable because
6672 the additional locking could be a performance killer, and
6673 md should be large enough anyway.
6676 *) New file apps/app_rand.c with commonly needed functionality
6677 for handling the random seed file.
6679 Use the random seed file in some applications that previously did not:
6681 dsaparam -genkey (which also ignored its '-rand' option),
6684 x509 (when signing).
6685 Except on systems with /dev/urandom, it is crucial to have a random
6686 seed file at least for key creation, DSA signing, and for DH exchanges;
6687 for RSA signatures we could do without one.
6689 gendh and gendsa (unlike genrsa) used to read only the first byte
6690 of each file listed in the '-rand' option. The function as previously
6691 found in genrsa is now in app_rand.c and is used by all programs
6692 that support '-rand'.
6695 *) In RAND_write_file, use mode 0600 for creating files;
6696 don't just chmod when it may be too late.
6699 *) Report an error from X509_STORE_load_locations
6700 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6703 *) New function ASN1_mbstring_copy() this copies a string in either
6704 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6705 into an ASN1_STRING type. A mask of permissible types is passed
6706 and it chooses the "minimal" type to use or an error if not type
6710 *) Add function equivalents to the various macros in asn1.h. The old
6711 macros are retained with an M_ prefix. Code inside the library can
6712 use the M_ macros. External code (including the openssl utility)
6713 should *NOT* in order to be "shared library friendly".
6716 *) Add various functions that can check a certificate's extensions
6717 to see if it usable for various purposes such as SSL client,
6718 server or S/MIME and CAs of these types. This is currently
6719 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6720 verification. Also added a -purpose flag to x509 utility to
6721 print out all the purposes.
6724 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6728 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6729 for, obtain and decode and extension and obtain its critical flag.
6730 This allows all the necessary extension code to be handled in a
6731 single function call.
6734 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6735 platforms. See crypto/rc4/rc4_enc.c for further details.
6738 *) New -noout option to asn1parse. This causes no output to be produced
6739 its main use is when combined with -strparse and -out to extract data
6740 from a file (which may not be in ASN.1 format).
6743 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6744 when producing the local key id.
6745 [Richard Levitte <levitte@stacken.kth.se>]
6747 *) New option -dhparam in s_server. This allows a DH parameter file to be
6748 stated explicitly. If it is not stated then it tries the first server
6749 certificate file. The previous behaviour hard coded the filename
6753 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6754 a public key to be input or output. For example:
6755 openssl rsa -in key.pem -pubout -out pubkey.pem
6756 Also added necessary DSA public key functions to handle this.
6759 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6760 in the message. This was handled by allowing
6761 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6762 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6764 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6765 to the end of the strings whereas this didn't. This would cause problems
6766 if strings read with d2i_ASN1_bytes() were later modified.
6767 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6769 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6770 data and it contains EOF it will end up returning an error. This is
6771 caused by input 46 bytes long. The cause is due to the way base64
6772 BIOs find the start of base64 encoded data. They do this by trying a
6773 trial decode on each line until they find one that works. When they
6774 do a flag is set and it starts again knowing it can pass all the
6775 data directly through the decoder. Unfortunately it doesn't reset
6776 the context it uses. This means that if EOF is reached an attempt
6777 is made to pass two EOFs through the context and this causes the
6778 resulting error. This can also cause other problems as well. As is
6779 usual with these problems it takes *ages* to find and the fix is
6780 trivial: move one line.
6781 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6783 *) Ugly workaround to get s_client and s_server working under Windows. The
6784 old code wouldn't work because it needed to select() on sockets and the
6785 tty (for keypresses and to see if data could be written). Win32 only
6786 supports select() on sockets so we select() with a 1s timeout on the
6787 sockets and then see if any characters are waiting to be read, if none
6788 are present then we retry, we also assume we can always write data to
6789 the tty. This isn't nice because the code then blocks until we've
6790 received a complete line of data and it is effectively polling the
6791 keyboard at 1s intervals: however it's quite a bit better than not
6792 working at all :-) A dedicated Windows application might handle this
6793 with an event loop for example.
6796 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6797 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6798 will be called when RSA_sign() and RSA_verify() are used. This is useful
6799 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6800 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6801 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6802 This necessitated the support of an extra signature type NID_md5_sha1
6803 for SSL signatures and modifications to the SSL library to use it instead
6804 of calling RSA_public_decrypt() and RSA_private_encrypt().
6807 *) Add new -verify -CAfile and -CApath options to the crl program, these
6808 will lookup a CRL issuers certificate and verify the signature in a
6809 similar way to the verify program. Tidy up the crl program so it
6810 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6811 less strict. It will now permit CRL extensions even if it is not
6812 a V2 CRL: this will allow it to tolerate some broken CRLs.
6815 *) Initialize all non-automatic variables each time one of the openssl
6816 sub-programs is started (this is necessary as they may be started
6817 multiple times from the "OpenSSL>" prompt).
6818 [Lennart Bang, Bodo Moeller]
6820 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6821 removing all other RSA functionality (this is what NO_RSA does). This
6822 is so (for example) those in the US can disable those operations covered
6823 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6827 *) Non-copying interface to BIO pairs.
6828 (still largely untested)
6831 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6832 ASCII string. This was handled independently in various places before.
6835 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6836 UTF8 strings a character at a time.
6839 *) Use client_version from client hello to select the protocol
6840 (s23_srvr.c) and for RSA client key exchange verification
6841 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6844 *) Add various utility functions to handle SPKACs, these were previously
6845 handled by poking round in the structure internals. Added new function
6846 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6847 print, verify and generate SPKACs. Based on an original idea from
6848 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6851 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6854 *) Allow the config file extension section to be overwritten on the
6855 command line. Based on an original idea from Massimiliano Pala
6856 <madwolf@comune.modena.it>. The new option is called -extensions
6857 and can be applied to ca, req and x509. Also -reqexts to override
6858 the request extensions in req and -crlexts to override the crl extensions
6862 *) Add new feature to the SPKAC handling in ca. Now you can include
6863 the same field multiple times by preceding it by "XXXX." for example:
6866 this is the same syntax as used in the req config file.
6869 *) Allow certificate extensions to be added to certificate requests. These
6870 are specified in a 'req_extensions' option of the req section of the
6871 config file. They can be printed out with the -text option to req but
6872 are otherwise ignored at present.
6875 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6876 data read consists of only the final block it would not decrypted because
6877 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6878 A misplaced 'break' also meant the decrypted final block might not be
6879 copied until the next read.
6882 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6883 a few extra parameters to the DH structure: these will be useful if
6884 for example we want the value of 'q' or implement X9.42 DH.
6887 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6888 provides hooks that allow the default DSA functions or functions on a
6889 "per key" basis to be replaced. This allows hardware acceleration and
6890 hardware key storage to be handled without major modification to the
6891 library. Also added low level modexp hooks and CRYPTO_EX structure and
6892 associated functions.
6895 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6896 as "read only": it can't be written to and the buffer it points to will
6897 not be freed. Reading from a read only BIO is much more efficient than
6898 a normal memory BIO. This was added because there are several times when
6899 an area of memory needs to be read from a BIO. The previous method was
6900 to create a memory BIO and write the data to it, this results in two
6901 copies of the data and an O(n^2) reading algorithm. There is a new
6902 function BIO_new_mem_buf() which creates a read only memory BIO from
6903 an area of memory. Also modified the PKCS#7 routines to use read only
6907 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6908 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6909 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6910 but a retry condition occured while trying to read the rest.
6913 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6914 NID_pkcs7_encrypted by default: this was wrong since this should almost
6915 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6916 the encrypted data type: this is a more sensible place to put it and it
6917 allows the PKCS#12 code to be tidied up that duplicated this
6921 *) Changed obj_dat.pl script so it takes its input and output files on
6922 the command line. This should avoid shell escape redirection problems
6926 *) Initial support for certificate extension requests, these are included
6927 in things like Xenroll certificate requests. Included functions to allow
6928 extensions to be obtained and added.
6931 *) -crlf option to s_client and s_server for sending newlines as
6932 CRLF (as required by many protocols).
6935 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6937 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6938 [Ralf S. Engelschall]
6940 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6941 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6943 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6947 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6948 DH parameters/keys (q is lost during that conversion, but the resulting
6949 DH parameters contain its length).
6951 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6952 much faster than DH_generate_parameters (which creates parameters
6953 where p = 2*q + 1), and also the smaller q makes DH computations
6954 much more efficient (160-bit exponentiation instead of 1024-bit
6955 exponentiation); so this provides a convenient way to support DHE
6956 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6957 utter importance to use
6958 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6960 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6961 when such DH parameters are used, because otherwise small subgroup
6962 attacks may become possible!
6965 *) Avoid memory leak in i2d_DHparams.
6968 *) Allow the -k option to be used more than once in the enc program:
6969 this allows the same encrypted message to be read by multiple recipients.
6972 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6973 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6974 it will always use the numerical form of the OID, even if it has a short
6978 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6979 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6980 otherwise bn_mod_exp was called. In the case of hardware keys for example
6981 no private key components need be present and it might store extra data
6982 in the RSA structure, which cannot be accessed from bn_mod_exp.
6983 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6984 private key operations.
6987 *) Added support for SPARC Linux.
6990 *) pem_password_cb function type incompatibly changed from
6991 typedef int pem_password_cb(char *buf, int size, int rwflag);
6993 ....(char *buf, int size, int rwflag, void *userdata);
6994 so that applications can pass data to their callbacks:
6995 The PEM[_ASN1]_{read,write}... functions and macros now take an
6996 additional void * argument, which is just handed through whenever
6997 the password callback is called.
6998 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7000 New function SSL_CTX_set_default_passwd_cb_userdata.
7002 Compatibility note: As many C implementations push function arguments
7003 onto the stack in reverse order, the new library version is likely to
7004 interoperate with programs that have been compiled with the old
7005 pem_password_cb definition (PEM_whatever takes some data that
7006 happens to be on the stack as its last argument, and the callback
7007 just ignores this garbage); but there is no guarantee whatsoever that
7010 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7011 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7012 problems not only on Windows, but also on some Unix platforms.
7013 To avoid problematic command lines, these definitions are now in an
7014 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7015 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7018 *) MIPS III/IV assembler module is reimplemented.
7021 *) More DES library cleanups: remove references to srand/rand and
7022 delete an unused file.
7025 *) Add support for the the free Netwide assembler (NASM) under Win32,
7026 since not many people have MASM (ml) and it can be hard to obtain.
7027 This is currently experimental but it seems to work OK and pass all
7028 the tests. Check out INSTALL.W32 for info.
7031 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7032 without temporary keys kept an extra copy of the server key,
7033 and connections with temporary keys did not free everything in case
7037 *) New function RSA_check_key and new openssl rsa option -check
7038 for verifying the consistency of RSA keys.
7039 [Ulf Moeller, Bodo Moeller]
7041 *) Various changes to make Win32 compile work:
7042 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7043 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7044 comparison" warnings.
7045 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7048 *) Add a debugging option to PKCS#5 v2 key generation function: when
7049 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7050 derived keys are printed to stderr.
7053 *) Copy the flags in ASN1_STRING_dup().
7054 [Roman E. Pavlov <pre@mo.msk.ru>]
7056 *) The x509 application mishandled signing requests containing DSA
7057 keys when the signing key was also DSA and the parameters didn't match.
7059 It was supposed to omit the parameters when they matched the signing key:
7060 the verifying software was then supposed to automatically use the CA's
7061 parameters if they were absent from the end user certificate.
7063 Omitting parameters is no longer recommended. The test was also
7064 the wrong way round! This was probably due to unusual behaviour in
7065 EVP_cmp_parameters() which returns 1 if the parameters match.
7066 This meant that parameters were omitted when they *didn't* match and
7067 the certificate was useless. Certificates signed with 'ca' didn't have
7069 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7071 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7072 The interface is as follows:
7073 Applications can use
7074 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7075 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7076 "off" is now the default.
7077 The library internally uses
7078 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7079 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7080 to disable memory-checking temporarily.
7082 Some inconsistent states that previously were possible (and were
7083 even the default) are now avoided.
7085 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7086 with each memory chunk allocated; this is occasionally more helpful
7087 than just having a counter.
7089 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7091 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7095 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7096 which largely parallels "options", but is for changing API behaviour,
7097 whereas "options" are about protocol behaviour.
7098 Initial "mode" flags are:
7100 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7101 a single record has been written.
7102 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7103 retries use the same buffer location.
7104 (But all of the contents must be
7108 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7111 *) Fix problems with no-hmac etc.
7112 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7114 *) New functions RSA_get_default_method(), RSA_set_method() and
7115 RSA_get_method(). These allows replacement of RSA_METHODs without having
7116 to mess around with the internals of an RSA structure.
7119 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7120 Also really enable memory leak checks in openssl.c and in some
7122 [Chad C. Mulligan, Bodo Moeller]
7124 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7125 up the length of negative integers. This has now been simplified to just
7126 store the length when it is first determined and use it later, rather
7127 than trying to keep track of where data is copied and updating it to
7129 [Steve Henson, reported by Brien Wheeler
7130 <bwheeler@authentica-security.com>]
7132 *) Add a new function PKCS7_signatureVerify. This allows the verification
7133 of a PKCS#7 signature but with the signing certificate passed to the
7134 function itself. This contrasts with PKCS7_dataVerify which assumes the
7135 certificate is present in the PKCS#7 structure. This isn't always the
7136 case: certificates can be omitted from a PKCS#7 structure and be
7137 distributed by "out of band" means (such as a certificate database).
7140 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7141 function prototypes in pem.h, also change util/mkdef.pl to add the
7142 necessary function names.
7145 *) mk1mf.pl (used by Windows builds) did not properly read the
7146 options set by Configure in the top level Makefile, and Configure
7147 was not even able to write more than one option correctly.
7148 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7151 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7152 file to be loaded from a BIO or FILE pointer. The BIO version will
7153 for example allow memory BIOs to contain config info.
7156 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7157 Whoever hopes to achieve shared-library compatibility across versions
7158 must use this, not the compile-time macro.
7159 (Exercise 0.9.4: Which is the minimum library version required by
7161 Note: All this applies only to multi-threaded programs, others don't
7165 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7166 through a BIO pair triggered the default case, i.e.
7167 SSLerr(...,SSL_R_UNKNOWN_STATE).
7170 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7171 can use the SSL library even if none of the specific BIOs is
7175 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7176 for the encoded length.
7177 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7179 *) Add initial documentation of the X509V3 functions.
7182 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7183 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7184 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7185 secure PKCS#8 private key format with a high iteration count.
7188 *) Fix determination of Perl interpreter: A perl or perl5
7189 _directory_ in $PATH was also accepted as the interpreter.
7190 [Ralf S. Engelschall]
7192 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7193 wrong with it but it was very old and did things like calling
7194 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7198 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7199 to use the new extension code.
7202 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7203 with macros. This should make it easier to change their form, add extra
7204 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7208 *) Add to configuration table a new entry that can specify an alternative
7209 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7210 according to Mark Crispin <MRC@Panda.COM>.
7214 *) DES CBC did not update the IV. Weird.
7217 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7218 Changing the behaviour of the former might break existing programs --
7219 where IV updating is needed, des_ncbc_encrypt can be used.
7222 *) When bntest is run from "make test" it drives bc to check its
7223 calculations, as well as internally checking them. If an internal check
7224 fails, it needs to cause bc to give a non-zero result or make test carries
7225 on without noticing the failure. Fixed.
7228 *) DES library cleanups.
7231 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7232 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7233 ciphers. NOTE: although the key derivation function has been verified
7234 against some published test vectors it has not been extensively tested
7235 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7239 *) Instead of "mkdir -p", which is not fully portable, use new
7240 Perl script "util/mkdir-p.pl".
7243 *) Rewrite the way password based encryption (PBE) is handled. It used to
7244 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7245 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7246 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7247 the 'parameter' field of the AlgorithmIdentifier is passed to the
7248 underlying key generation function so it must do its own ASN1 parsing.
7249 This has also changed the EVP_PBE_CipherInit() function which now has a
7250 'parameter' argument instead of literal salt and iteration count values
7251 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7254 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7255 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7256 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7257 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7258 value was just used as a "magic string" and not used directly its
7259 value doesn't matter.
7262 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7266 *) "linux-sparc64" configuration (ultrapenguin).
7267 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7268 "linux-sparc" configuration.
7269 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7271 *) config now generates no-xxx options for missing ciphers.
7274 *) Support the EBCDIC character set (work in progress).
7275 File ebcdic.c not yet included because it has a different license.
7276 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7278 *) Support BS2000/OSD-POSIX.
7279 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7281 *) Make callbacks for key generation use void * instead of char *.
7284 *) Make S/MIME samples compile (not yet tested).
7287 *) Additional typesafe stacks.
7290 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7294 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7296 *) New configuration variant "sco5-gcc".
7298 *) Updated some demos.
7299 [Sean O Riordain, Wade Scholine]
7301 *) Add missing BIO_free at exit of pkcs12 application.
7304 *) Fix memory leak in conf.c.
7307 *) Updates for Win32 to assembler version of MD5.
7310 *) Set #! path to perl in apps/der_chop to where we found it
7311 instead of using a fixed path.
7314 *) SHA library changes for irix64-mips4-cc.
7317 *) Improvements for VMS support.
7321 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7323 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7324 This also avoids the problems with SC4.2 and unpatched SC5.
7325 [Andy Polyakov <appro@fy.chalmers.se>]
7327 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7328 These are required because of the typesafe stack would otherwise break
7329 existing code. If old code used a structure member which used to be STACK
7330 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7331 sk_num or sk_value it would produce an error because the num, data members
7332 are not present in STACK_OF. Now it just produces a warning. sk_set
7333 replaces the old method of assigning a value to sk_value
7334 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7335 that does this will no longer work (and should use sk_set instead) but
7336 this could be regarded as a "questionable" behaviour anyway.
7339 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7340 correctly handle encrypted S/MIME data.
7343 *) Change type of various DES function arguments from des_cblock
7344 (which means, in function argument declarations, pointer to char)
7345 to des_cblock * (meaning pointer to array with 8 char elements),
7346 which allows the compiler to do more typechecking; it was like
7347 that back in SSLeay, but with lots of ugly casts.
7349 Introduce new type const_des_cblock.
7352 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7353 problems: find RecipientInfo structure that matches recipient certificate
7354 and initialise the ASN1 structures properly based on passed cipher.
7357 *) Belatedly make the BN tests actually check the results.
7360 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7361 to and from BNs: it was completely broken. New compilation option
7362 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7363 key elements as negative integers.
7366 *) Reorganize and speed up MD5.
7367 [Andy Polyakov <appro@fy.chalmers.se>]
7370 [Richard Levitte <richard@levitte.org>]
7372 *) New option -out to asn1parse to allow the parsed structure to be
7373 output to a file. This is most useful when combined with the -strparse
7374 option to examine the output of things like OCTET STRINGS.
7377 *) Make SSL library a little more fool-proof by not requiring any longer
7378 that SSL_set_{accept,connect}_state be called before
7379 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7380 in many applications because usually everything *appeared* to work as
7381 intended anyway -- now it really works as intended).
7384 *) Move openssl.cnf out of lib/.
7387 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7388 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7389 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7390 [Ralf S. Engelschall]
7392 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7393 handle PKCS#7 enveloped data properly.
7394 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7396 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7397 copying pointers. The cert_st handling is changed by this in
7398 various ways (and thus what used to be known as ctx->default_cert
7399 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7400 any longer when s->cert does not give us what we need).
7401 ssl_cert_instantiate becomes obsolete by this change.
7402 As soon as we've got the new code right (possibly it already is?),
7403 we have solved a couple of bugs of the earlier code where s->cert
7404 was used as if it could not have been shared with other SSL structures.
7406 Note that using the SSL API in certain dirty ways now will result
7407 in different behaviour than observed with earlier library versions:
7408 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7409 does not influence s as it used to.
7411 In order to clean up things more thoroughly, inside SSL_SESSION
7412 we don't use CERT any longer, but a new structure SESS_CERT
7413 that holds per-session data (if available); currently, this is
7414 the peer's certificate chain and, for clients, the server's certificate
7415 and temporary key. CERT holds only those values that can have
7416 meaningful defaults in an SSL_CTX.
7419 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7420 from the internal representation. Various PKCS#7 fixes: remove some
7421 evil casts and set the enc_dig_alg field properly based on the signing
7425 *) Allow PKCS#12 password to be set from the command line or the
7426 environment. Let 'ca' get its config file name from the environment
7427 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7431 *) Allow certificate policies extension to use an IA5STRING for the
7432 organization field. This is contrary to the PKIX definition but
7433 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7437 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7438 without disallowing inline assembler and the like for non-pedantic builds.
7441 *) Support Borland C++ builder.
7442 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7447 *) SHA-1 cleanups and performance enhancements.
7448 [Andy Polyakov <appro@fy.chalmers.se>]
7450 *) Sparc v8plus assembler for the bignum library.
7451 [Andy Polyakov <appro@fy.chalmers.se>]
7453 *) Accept any -xxx and +xxx compiler options in Configure.
7456 *) Update HPUX configuration.
7459 *) Add missing sk_<type>_unshift() function to safestack.h
7460 [Ralf S. Engelschall]
7462 *) New function SSL_CTX_use_certificate_chain_file that sets the
7463 "extra_cert"s in addition to the certificate. (This makes sense
7464 only for "PEM" format files, as chains as a whole are not
7468 *) Support verify_depth from the SSL API.
7469 x509_vfy.c had what can be considered an off-by-one-error:
7470 Its depth (which was not part of the external interface)
7471 was actually counting the number of certificates in a chain;
7472 now it really counts the depth.
7475 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7476 instead of X509err, which often resulted in confusing error
7477 messages since the error codes are not globally unique
7478 (e.g. an alleged error in ssl3_accept when a certificate
7479 didn't match the private key).
7481 *) New function SSL_CTX_set_session_id_context that allows to set a default
7482 value (so that you don't need SSL_set_session_id_context for each
7483 connection using the SSL_CTX).
7486 *) OAEP decoding bug fix.
7489 *) Support INSTALL_PREFIX for package builders, as proposed by
7493 *) New Configure options "threads" and "no-threads". For systems
7494 where the proper compiler options are known (currently Solaris
7495 and Linux), "threads" is the default.
7498 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7501 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7502 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7503 such as /usr/local/bin.
7506 *) "make linux-shared" to build shared libraries.
7507 [Niels Poppe <niels@netbox.org>]
7509 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7512 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7513 extension adding in x509 utility.
7516 *) Remove NOPROTO sections and error code comments.
7519 *) Partial rewrite of the DEF file generator to now parse the ANSI
7523 *) New Configure options --prefix=DIR and --openssldir=DIR.
7526 *) Complete rewrite of the error code script(s). It is all now handled
7527 by one script at the top level which handles error code gathering,
7528 header rewriting and C source file generation. It should be much better
7529 than the old method: it now uses a modified version of Ulf's parser to
7530 read the ANSI prototypes in all header files (thus the old K&R definitions
7531 aren't needed for error creation any more) and do a better job of
7532 translating function codes into names. The old 'ASN1 error code imbedded
7533 in a comment' is no longer necessary and it doesn't use .err files which
7534 have now been deleted. Also the error code call doesn't have to appear all
7535 on one line (which resulted in some large lines...).
7538 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7541 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7542 0 (which usually indicates a closed connection), but continue reading.
7545 *) Fix some race conditions.
7548 *) Add support for CRL distribution points extension. Add Certificate
7549 Policies and CRL distribution points documentation.
7552 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7555 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7556 8 of keying material. Merlin has also confirmed interop with this fix
7557 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7558 [Merlin Hughes <merlin@baltimore.ie>]
7560 *) Fix lots of warnings.
7561 [Richard Levitte <levitte@stacken.kth.se>]
7563 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7564 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7565 [Richard Levitte <levitte@stacken.kth.se>]
7567 *) Fix problems with sizeof(long) == 8.
7568 [Andy Polyakov <appro@fy.chalmers.se>]
7570 *) Change functions to ANSI C.
7573 *) Fix typos in error codes.
7574 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7576 *) Remove defunct assembler files from Configure.
7579 *) SPARC v8 assembler BIGNUM implementation.
7580 [Andy Polyakov <appro@fy.chalmers.se>]
7582 *) Support for Certificate Policies extension: both print and set.
7583 Various additions to support the r2i method this uses.
7586 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7587 return a const string when you are expecting an allocated buffer.
7590 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7591 types DirectoryString and DisplayText.
7594 *) Add code to allow r2i extensions to access the configuration database,
7595 add an LHASH database driver and add several ctx helper functions.
7598 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7599 fail when they extended the size of a BIGNUM.
7602 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7603 support typesafe stack.
7606 *) Fix typo in SSL_[gs]et_options().
7607 [Nils Frostberg <nils@medcom.se>]
7609 *) Delete various functions and files that belonged to the (now obsolete)
7610 old X509V3 handling code.
7613 *) New Configure option "rsaref".
7616 *) Don't auto-generate pem.h.
7619 *) Introduce type-safe ASN.1 SETs.
7622 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7623 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7625 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7626 that links with OpenSSL (well at least cause lots of warnings), but fear
7627 not: the conversion is trivial, and it eliminates loads of evil casts. A
7628 few STACKed things have been converted already. Feel free to convert more.
7629 In the fullness of time, I'll do away with the STACK type altogether.
7632 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7633 specified in <certfile> by updating the entry in the index.txt file.
7634 This way one no longer has to edit the index.txt file manually for
7635 revoking a certificate. The -revoke option does the gory details now.
7636 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7638 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7639 `-text' option at all and this way the `-noout -text' combination was
7640 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7641 [Ralf S. Engelschall]
7643 *) Make sure a corresponding plain text error message exists for the
7644 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7645 verify callback function determined that a certificate was revoked.
7646 [Ralf S. Engelschall]
7648 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7649 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7650 all available cipers including rc5, which was forgotten until now.
7651 In order to let the testing shell script know which algorithms
7652 are available, a new (up to now undocumented) command
7653 "openssl list-cipher-commands" is used.
7656 *) Bugfix: s_client occasionally would sleep in select() when
7657 it should have checked SSL_pending() first.
7660 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7661 the raw DSA values prior to ASN.1 encoding.
7664 *) Tweaks to Configure
7665 [Niels Poppe <niels@netbox.org>]
7667 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7671 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7674 *) New config option to avoid instructions that are illegal on the 80386.
7675 The default code is faster, but requires at least a 486.
7678 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7679 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7680 same as SSL2_VERSION anyway.
7683 *) New "-showcerts" option for s_client.
7686 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7687 application. Various cleanups and fixes.
7690 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7691 modify error routines to work internally. Add error codes and PBE init
7692 to library startup routines.
7695 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7696 packing functions to asn1 and evp. Changed function names and error
7697 codes along the way.
7700 *) PKCS12 integration: and so it begins... First of several patches to
7701 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7702 objects to objects.h
7705 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7706 and display support for Thawte strong extranet extension.
7709 *) Add LinuxPPC support.
7710 [Jeff Dubrule <igor@pobox.org>]
7712 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7713 bn_div_words in alpha.s.
7714 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7716 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7717 OAEP isn't supported when OpenSSL is built with RSAref.
7718 [Ulf Moeller <ulf@fitug.de>]
7720 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7721 so they no longer are missing under -DNOPROTO.
7722 [Soren S. Jorvang <soren@t.dk>]
7725 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7727 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7728 doesn't work when the session is reused. Coming soon!
7731 *) Fix a security hole, that allows sessions to be reused in the wrong
7732 context thus bypassing client cert protection! All software that uses
7733 client certs and session caches in multiple contexts NEEDS PATCHING to
7734 allow session reuse! A fuller solution is in the works.
7735 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7737 *) Some more source tree cleanups (removed obsolete files
7738 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7739 permission on "config" script to be executable) and a fix for the INSTALL
7741 [Ulf Moeller <ulf@fitug.de>]
7743 *) Remove some legacy and erroneous uses of malloc, free instead of
7745 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7747 *) Make rsa_oaep_test return non-zero on error.
7748 [Ulf Moeller <ulf@fitug.de>]
7750 *) Add support for native Solaris shared libraries. Configure
7751 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7752 if someone would make that last step automatic.
7753 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7755 *) ctx_size was not built with the right compiler during "make links". Fixed.
7758 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7759 except NULL ciphers". This means the default cipher list will no longer
7760 enable NULL ciphers. They need to be specifically enabled e.g. with
7761 the string "DEFAULT:eNULL".
7764 *) Fix to RSA private encryption routines: if p < q then it would
7765 occasionally produce an invalid result. This will only happen with
7766 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7769 *) Be less restrictive and allow also `perl util/perlpath.pl
7770 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7771 because this way one can also use an interpreter named `perl5' (which is
7772 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7773 installed as `perl').
7774 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7776 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7777 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7779 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7780 advapi32.lib to Win32 build and change the pem test comparision
7781 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7782 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7783 and crypto/des/ede_cbcm_enc.c.
7786 *) DES quad checksum was broken on big-endian architectures. Fixed.
7789 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7790 Win32 test batch file so it (might) work again. The Win32 test batch file
7791 is horrible: I feel ill....
7794 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7795 in e_os.h. Audit of header files to check ANSI and non ANSI
7796 sections: 10 functions were absent from non ANSI section and not exported
7797 from Windows DLLs. Fixed up libeay.num for new functions.
7800 *) Make `openssl version' output lines consistent.
7801 [Ralf S. Engelschall]
7803 *) Fix Win32 symbol export lists for BIO functions: Added
7804 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7805 to ms/libeay{16,32}.def.
7806 [Ralf S. Engelschall]
7808 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7809 fine under Unix and passes some trivial tests I've now added. But the
7810 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7811 added to make sure no one expects that this stuff really works in the
7812 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7813 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7815 [Ralf S. Engelschall]
7817 *) Fix the generation of two part addresses in perl.
7818 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7820 *) Add config entry for Linux on MIPS.
7821 [John Tobey <jtobey@channel1.com>]
7823 *) Make links whenever Configure is run, unless we are on Windoze.
7826 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7827 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7831 *) Add a useful kludge to allow package maintainers to specify compiler and
7832 other platforms details on the command line without having to patch the
7833 Configure script everytime: One now can use ``perl Configure
7834 <id>:<details>'', i.e. platform ids are allowed to have details appended
7835 to them (seperated by colons). This is treated as there would be a static
7836 pre-configured entry in Configure's %table under key <id> with value
7837 <details> and ``perl Configure <id>'' is called. So, when you want to
7838 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7839 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7840 now, which overrides the FreeBSD-elf entry on-the-fly.
7841 [Ralf S. Engelschall]
7843 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7846 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7847 on the `perl Configure ...' command line. This way one can compile
7848 OpenSSL libraries with Position Independent Code (PIC) which is needed
7849 for linking it into DSOs.
7850 [Ralf S. Engelschall]
7852 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7856 *) Cleaned up the LICENSE document: The official contact for any license
7857 questions now is the OpenSSL core team under openssl-core@openssl.org.
7858 And add a paragraph about the dual-license situation to make sure people
7859 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7860 to the OpenSSL toolkit.
7861 [Ralf S. Engelschall]
7863 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7864 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7865 Additonally cleaned up the `make links' target: Remove unnecessary
7866 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7867 to speed processing and no longer clutter the display with confusing
7868 stuff. Instead only the actually done links are displayed.
7869 [Ralf S. Engelschall]
7871 *) Permit null encryption ciphersuites, used for authentication only. It used
7872 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7873 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7877 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7878 signed attributes when verifying signatures (this would break them),
7879 the detached data encoding was wrong and public keys obtained using
7880 X509_get_pubkey() weren't freed.
7883 *) Add text documentation for the BUFFER functions. Also added a work around
7884 to a Win95 console bug. This was triggered by the password read stuff: the
7885 last character typed gets carried over to the next fread(). If you were
7886 generating a new cert request using 'req' for example then the last
7887 character of the passphrase would be CR which would then enter the first
7891 *) Added the new `Includes OpenSSL Cryptography Software' button as
7892 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7893 button and can be used by applications based on OpenSSL to show the
7894 relationship to the OpenSSL project.
7895 [Ralf S. Engelschall]
7897 *) Remove confusing variables in function signatures in files
7898 ssl/ssl_lib.c and ssl/ssl.h.
7899 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7901 *) Don't install bss_file.c under PREFIX/include/
7902 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7904 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7905 functions that return function pointers and has support for NT specific
7906 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7907 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7908 unsigned to signed types: this was killing the Win32 compile.
7911 *) Add new certificate file to stack functions,
7912 SSL_add_dir_cert_subjects_to_stack() and
7913 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7914 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7915 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7916 This means that Apache-SSL and similar packages don't have to mess around
7917 to add as many CAs as they want to the preferred list.
7920 *) Experiment with doxygen documentation. Currently only partially applied to
7922 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7923 openssl.doxy as the configuration file.
7926 *) Get rid of remaining C++-style comments which strict C compilers hate.
7927 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7929 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7930 compiled in by default: it has problems with large keys.
7933 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7934 DH private keys and/or callback functions which directly correspond to
7935 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7936 is needed for applications which have to configure certificates on a
7937 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7939 For the RSA certificate situation is makes no difference, but
7940 for the DSA certificate situation this fixes the "no shared cipher"
7941 problem where the OpenSSL cipher selection procedure failed because the
7942 temporary keys were not overtaken from the context and the API provided
7943 no way to reconfigure them.
7944 The new functions now let applications reconfigure the stuff and they
7945 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7946 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7947 non-public-API function ssl_cert_instantiate() is used as a helper
7948 function and also to reduce code redundancy inside ssl_rsa.c.
7949 [Ralf S. Engelschall]
7951 *) Move s_server -dcert and -dkey options out of the undocumented feature
7952 area because they are useful for the DSA situation and should be
7953 recognized by the users.
7954 [Ralf S. Engelschall]
7956 *) Fix the cipher decision scheme for export ciphers: the export bits are
7957 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7958 SSL_EXP_MASK. So, the original variable has to be used instead of the
7959 already masked variable.
7960 [Richard Levitte <levitte@stacken.kth.se>]
7962 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7963 [Richard Levitte <levitte@stacken.kth.se>]
7965 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7966 from `int' to `unsigned int' because it's a length and initialized by
7967 EVP_DigestFinal() which expects an `unsigned int *'.
7968 [Richard Levitte <levitte@stacken.kth.se>]
7970 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7971 script. Instead use the usual Shell->Perl transition trick.
7972 [Ralf S. Engelschall]
7974 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7975 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7976 -noout -modulus' as it's already the case for `openssl rsa -noout
7977 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7978 currently the public key is printed (a decision which was already done by
7979 `openssl dsa -modulus' in the past) which serves a similar purpose.
7980 Additionally the NO_RSA no longer completely removes the whole -modulus
7981 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7983 [Ralf S. Engelschall]
7985 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7986 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7987 [Arne Ansper <arne@ats.cyber.ee>]
7989 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7990 to be added. Now both 'req' and 'ca' can use new objects defined in the
7994 *) Add cool BIO that does syslog (or event log on NT).
7995 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7997 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7998 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7999 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8000 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8003 *) Add preliminary config info for new extension code.
8006 *) Make RSA_NO_PADDING really use no padding.
8007 [Ulf Moeller <ulf@fitug.de>]
8009 *) Generate errors when private/public key check is done.
8012 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8013 for some CRL extensions and new objects added.
8016 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8017 key usage extension and fuller support for authority key id.
8020 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8021 padding method for RSA, which is recommended for new applications in PKCS
8022 #1 v2.0 (RFC 2437, October 1998).
8023 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8024 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8025 against Bleichbacher's attack on RSA.
8026 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8029 *) Updates to the new SSL compression code
8030 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8032 *) Fix so that the version number in the master secret, when passed
8033 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8034 (because the server will not accept higher), that the version number
8035 is 0x03,0x01, not 0x03,0x00
8036 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8038 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8039 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8040 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8043 *) Support for RAW extensions where an arbitrary extension can be
8044 created by including its DER encoding. See apps/openssl.cnf for
8048 *) Make sure latest Perl versions don't interpret some generated C array
8049 code as Perl array code in the crypto/err/err_genc.pl script.
8050 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8052 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8053 not many people have the assembler. Various Win32 compilation fixes and
8054 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8058 *) Modify configure script 'Configure' to automatically create crypto/date.h
8059 file under Win32 and also build pem.h from pem.org. New script
8060 util/mkfiles.pl to create the MINFO file on environments that can't do a
8061 'make files': perl util/mkfiles.pl >MINFO should work.
8064 *) Major rework of DES function declarations, in the pursuit of correctness
8065 and purity. As a result, many evil casts evaporated, and some weirdness,
8066 too. You may find this causes warnings in your code. Zapping your evil
8067 casts will probably fix them. Mostly.
8070 *) Fix for a typo in asn1.h. Bug fix to object creation script
8071 obj_dat.pl. It considered a zero in an object definition to mean
8072 "end of object": none of the objects in objects.h have any zeros
8073 so it wasn't spotted.
8074 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8076 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8077 Masking (CBCM). In the absence of test vectors, the best I have been able
8078 to do is check that the decrypt undoes the encrypt, so far. Send me test
8079 vectors if you have them.
8082 *) Correct calculation of key length for export ciphers (too much space was
8083 allocated for null ciphers). This has not been tested!
8086 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8087 message is now correct (it understands "crypto" and "ssl" on its
8088 command line). There is also now an "update" option. This will update
8089 the util/ssleay.num and util/libeay.num files with any new functions.
8091 perl util/mkdef.pl crypto ssl update
8092 it will update them.
8095 *) Overhauled the Perl interface (perl/*):
8096 - ported BN stuff to OpenSSL's different BN library
8097 - made the perl/ source tree CVS-aware
8098 - renamed the package from SSLeay to OpenSSL (the files still contain
8099 their history because I've copied them in the repository)
8100 - removed obsolete files (the test scripts will be replaced
8101 by better Test::Harness variants in the future)
8102 [Ralf S. Engelschall]
8104 *) First cut for a very conservative source tree cleanup:
8105 1. merge various obsolete readme texts into doc/ssleay.txt
8106 where we collect the old documents and readme texts.
8107 2. remove the first part of files where I'm already sure that we no
8108 longer need them because of three reasons: either they are just temporary
8109 files which were left by Eric or they are preserved original files where
8110 I've verified that the diff is also available in the CVS via "cvs diff
8111 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8112 the crypto/md/ stuff).
8113 [Ralf S. Engelschall]
8115 *) More extension code. Incomplete support for subject and issuer alt
8116 name, issuer and authority key id. Change the i2v function parameters
8117 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8118 what that's for :-) Fix to ASN1 macro which messed up
8119 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8122 *) Preliminary support for ENUMERATED type. This is largely copied from the
8126 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8127 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8129 *) Make sure `make rehash' target really finds the `openssl' program.
8130 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8132 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8133 like to hear about it if this slows down other processors.
8136 *) Add CygWin32 platform information to Configure script.
8137 [Alan Batie <batie@aahz.jf.intel.com>]
8139 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8140 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8142 *) New program nseq to manipulate netscape certificate sequences
8145 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8149 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8150 but the BN code had some problems that would cause failures when
8151 doing certificate verification and some other functions.
8152 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8154 *) Add ASN1 and PEM code to support netscape certificate sequences.
8157 *) Add ASN1 and PEM code to support netscape certificate sequences.
8160 *) Add several PKIX and private extended key usage OIDs.
8163 *) Modify the 'ca' program to handle the new extension code. Modify
8164 openssl.cnf for new extension format, add comments.
8167 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8168 and add a sample to openssl.cnf so req -x509 now adds appropriate
8172 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8173 error code, add initial support to X509_print() and x509 application.
8176 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8177 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8178 stuff is currently isolated and isn't even compiled yet.
8181 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8182 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8183 Removed the versions check from X509 routines when loading extensions:
8184 this allows certain broken certificates that don't set the version
8185 properly to be processed.
8188 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8189 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8190 can still be regenerated with "make depend".
8193 *) Spelling mistake in C version of CAST-128.
8194 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8196 *) Changes to the error generation code. The perl script err-code.pl
8197 now reads in the old error codes and retains the old numbers, only
8198 adding new ones if necessary. It also only changes the .err files if new
8199 codes are added. The makefiles have been modified to only insert errors
8200 when needed (to avoid needlessly modifying header files). This is done
8201 by only inserting errors if the .err file is newer than the auto generated
8202 C file. To rebuild all the error codes from scratch (the old behaviour)
8203 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8204 or delete all the .err files.
8207 *) CAST-128 was incorrectly implemented for short keys. The C version has
8208 been fixed, but is untested. The assembler versions are also fixed, but
8209 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8210 to regenerate it if needed.
8211 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8212 Hagino <itojun@kame.net>]
8214 *) File was opened incorrectly in randfile.c.
8215 [Ulf Möller <ulf@fitug.de>]
8217 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8218 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8219 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8220 al: it's just almost always a UTCTime. Note this patch adds new error
8221 codes so do a "make errors" if there are problems.
8224 *) Correct Linux 1 recognition in config.
8225 [Ulf Möller <ulf@fitug.de>]
8227 *) Remove pointless MD5 hash when using DSA keys in ca.
8228 [Anonymous <nobody@replay.com>]
8230 *) Generate an error if given an empty string as a cert directory. Also
8231 generate an error if handed NULL (previously returned 0 to indicate an
8232 error, but didn't set one).
8233 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8235 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8238 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8239 parameters. This was causing a warning which killed off the Win32 compile.
8242 *) Remove C++ style comments from crypto/bn/bn_local.h.
8243 [Neil Costigan <neil.costigan@celocom.com>]
8245 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8246 based on a text string, looking up short and long names and finally
8247 "dot" format. The "dot" format stuff didn't work. Added new function
8248 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8249 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8250 OID is not part of the table.
8253 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8254 X509_LOOKUP_by_alias().
8257 *) Sort openssl functions by name.
8260 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8261 encryption from sample DSA keys (in case anyone is interested the password
8265 *) Make _all_ *_free functions accept a NULL pointer.
8266 [Frans Heymans <fheymans@isaserver.be>]
8268 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8270 [Anonymous <nobody@replay.com>]
8272 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8273 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8275 *) Don't blow it for numeric -newkey arguments to apps/req.
8276 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8278 *) Temp key "for export" tests were wrong in s3_srvr.c.
8279 [Anonymous <nobody@replay.com>]
8281 *) Add prototype for temp key callback functions
8282 SSL_CTX_set_tmp_{rsa,dh}_callback().
8285 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8286 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8289 *) X509_name_add_entry() freed the wrong thing after an error.
8290 [Arne Ansper <arne@ats.cyber.ee>]
8292 *) rsa_eay.c would attempt to free a NULL context.
8293 [Arne Ansper <arne@ats.cyber.ee>]
8295 *) BIO_s_socket() had a broken should_retry() on Windoze.
8296 [Arne Ansper <arne@ats.cyber.ee>]
8298 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8299 [Arne Ansper <arne@ats.cyber.ee>]
8301 *) Make sure the already existing X509_STORE->depth variable is initialized
8302 in X509_STORE_new(), but document the fact that this variable is still
8303 unused in the certificate verification process.
8304 [Ralf S. Engelschall]
8306 *) Fix the various library and apps files to free up pkeys obtained from
8307 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8310 *) Fix reference counting in X509_PUBKEY_get(). This makes
8311 demos/maurice/example2.c work, amongst others, probably.
8312 [Steve Henson and Ben Laurie]
8314 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8315 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8316 are no longer created. This way we have a single and consistent command
8317 line interface `openssl <command>', similar to `cvs <command>'.
8318 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8320 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8321 BIT STRING wrapper always have zero unused bits.
8324 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8327 *) Make the top-level INSTALL documentation easier to understand.
8330 *) Makefiles updated to exit if an error occurs in a sub-directory
8331 make (including if user presses ^C) [Paul Sutton]
8333 *) Make Montgomery context stuff explicit in RSA data structure.
8336 *) Fix build order of pem and err to allow for generated pem.h.
8339 *) Fix renumbering bug in X509_NAME_delete_entry().
8342 *) Enhanced the err-ins.pl script so it makes the error library number
8343 global and can add a library name. This is needed for external ASN1 and
8344 other error libraries.
8347 *) Fixed sk_insert which never worked properly.
8350 *) Fix ASN1 macros so they can handle indefinite length construted
8351 EXPLICIT tags. Some non standard certificates use these: they can now
8355 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8356 into a single doc/ssleay.txt bundle. This way the information is still
8357 preserved but no longer messes up this directory. Now it's new room for
8358 the new set of documenation files.
8359 [Ralf S. Engelschall]
8361 *) SETs were incorrectly DER encoded. This was a major pain, because they
8362 shared code with SEQUENCEs, which aren't coded the same. This means that
8363 almost everything to do with SETs or SEQUENCEs has either changed name or
8364 number of arguments.
8365 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8367 *) Fix test data to work with the above.
8370 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8371 was already fixed by Eric for 0.9.1 it seems.
8372 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8374 *) Autodetect FreeBSD3.
8377 *) Fix various bugs in Configure. This affects the following platforms:
8381 unixware-2.0-pentium
8385 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8386 before they are needed.
8389 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8393 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8395 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8396 changed SSLeay to OpenSSL in version strings.
8397 [Ralf S. Engelschall]
8399 *) Some fixups to the top-level documents.
8402 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8403 because the symlink to include/ was missing.
8404 [Ralf S. Engelschall]
8406 *) Incorporated the popular no-RSA/DSA-only patches
8407 which allow to compile a RSA-free SSLeay.
8408 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8410 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8411 when "ssleay" is still not found.
8412 [Ralf S. Engelschall]
8414 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8415 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8417 *) Updated the README file.
8418 [Ralf S. Engelschall]
8420 *) Added various .cvsignore files in the CVS repository subdirs
8421 to make a "cvs update" really silent.
8422 [Ralf S. Engelschall]
8424 *) Recompiled the error-definition header files and added
8425 missing symbols to the Win32 linker tables.
8426 [Ralf S. Engelschall]
8428 *) Cleaned up the top-level documents;
8429 o new files: CHANGES and LICENSE
8430 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8431 o merged COPYRIGHT into LICENSE
8432 o removed obsolete TODO file
8433 o renamed MICROSOFT to INSTALL.W32
8434 [Ralf S. Engelschall]
8436 *) Removed dummy files from the 0.9.1b source tree:
8437 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8438 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8439 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8440 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8441 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8442 [Ralf S. Engelschall]
8444 *) Added various platform portability fixes.
8447 *) The Genesis of the OpenSSL rpject:
8448 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8449 Young and Tim J. Hudson created while they were working for C2Net until
8451 [The OpenSSL Project]
8454 Changes between 0.9.0b and 0.9.1b [not released]
8456 *) Updated a few CA certificates under certs/
8459 *) Changed some BIGNUM api stuff.
8462 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8463 DGUX x86, Linux Alpha, etc.
8466 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8467 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8471 *) Add -strparse option to asn1pars program which parses nested
8473 [Dr Stephen Henson <shenson@bigfoot.com>]
8475 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8478 *) DSA fix for "ca" program.
8481 *) Added "-genkey" option to "dsaparam" program.
8484 *) Added RIPE MD160 (rmd160) message digest.
8487 *) Added -a (all) option to "ssleay version" command.
8490 *) Added PLATFORM define which is the id given to Configure.
8493 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8496 *) Extended the ASN.1 parser routines.
8499 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8502 *) Added a BN_CTX to the BN library.
8505 *) Fixed the weak key values in DES library
8508 *) Changed API in EVP library for cipher aliases.
8511 *) Added support for RC2/64bit cipher.
8514 *) Converted the lhash library to the crypto/mem.c functions.
8517 *) Added more recognized ASN.1 object ids.
8520 *) Added more RSA padding checks for SSL/TLS.
8523 *) Added BIO proxy/filter functionality.
8526 *) Added extra_certs to SSL_CTX which can be used
8527 send extra CA certificates to the client in the CA cert chain sending
8528 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8531 *) Now Fortezza is denied in the authentication phase because
8532 this is key exchange mechanism is not supported by SSLeay at all.
8535 *) Additional PKCS1 checks.
8538 *) Support the string "TLSv1" for all TLS v1 ciphers.
8541 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8542 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8545 *) Fixed a few memory leaks.
8548 *) Fixed various code and comment typos.
8551 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8552 bytes sent in the client random.
8553 [Edward Bishop <ebishop@spyglass.com>]
projects / oweals / openssl.git / blob