5 Changes between 0.9.8n and 0.9.8o [xx XXX xxxx]
7 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
8 access or freeing data twice (CVE-2010-0742)
9 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
11 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
12 common in certificates and some applications which only call
13 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
17 Reduce copying into .apps and .test in makevms.com
18 Don't try to use blank CA certificate in CA.com
19 Allow use of C files from original directories in maketests.com
20 [Steven M. Schweda" <sms@antinode.info>]
22 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
24 *) When rejecting SSL/TLS records due to an incorrect version number, never
25 update s->server with a new major version number. As of
26 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
27 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
28 the previous behavior could result in a read attempt at NULL when
29 receiving specific incorrect SSL/TLS records once record payload
30 protection is active. (CVE-2010-0740)
31 [Bodo Moeller, Adam Langley <agl@chromium.org>]
33 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
34 could be crashed if the relevant tables were not present (e.g. chrooted).
35 [Tomas Hoger <thoger@redhat.com>]
37 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
39 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
40 [Martin Olsson, Neel Mehta]
42 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
43 accommodate for stack sorting, always a write lock!).
46 *) On some versions of WIN32 Heap32Next is very slow. This can cause
47 excessive delays in the RAND_poll(): over a minute. As a workaround
48 include a time check in the inner Heap32Next loop too.
51 *) The code that handled flushing of data in SSL/TLS originally used the
52 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
53 the problem outlined in PR#1949. The fix suggested there however can
54 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
55 of Apache). So instead simplify the code to flush unconditionally.
56 This should be fine since flushing with no data to flush is a no op.
59 *) Handle TLS versions 2.0 and later properly and correctly use the
60 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
61 off ancient servers have a habit of sticking around for a while...
64 *) Modify compression code so it frees up structures without using the
65 ex_data callbacks. This works around a problem where some applications
66 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
67 restarting) then use compression (e.g. SSL with compression) later.
68 This results in significant per-connection memory leaks and
69 has caused some security issues including CVE-2008-1678 and
73 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
74 change when encrypting or decrypting.
77 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
78 connect and renegotiate with servers which do not support RI.
79 Until RI is more widely deployed this option is enabled by default.
82 *) Add "missing" ssl ctrls to clear options and mode.
85 *) If client attempts to renegotiate and doesn't support RI respond with
86 a no_renegotiation alert as required by RFC5746. Some renegotiating
87 TLS clients will continue a connection gracefully when they receive
88 the alert. Unfortunately OpenSSL mishandled this alert and would hang
89 waiting for a server hello which it will never receive. Now we treat a
90 received no_renegotiation alert as a fatal error. This is because
91 applications requesting a renegotiation might well expect it to succeed
92 and would have no code in place to handle the server denying it so the
93 only safe thing to do is to terminate the connection.
96 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
97 peer supports secure renegotiation and 0 otherwise. Print out peer
98 renegotiation support in s_client/s_server.
101 *) Replace the highly broken and deprecated SPKAC certification method with
102 the updated NID creation version. This should correctly handle UTF8.
105 *) Implement RFC5746. Re-enable renegotiation but require the extension
106 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
107 turns out to be a bad idea. It has been replaced by
108 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
109 SSL_CTX_set_options(). This is really not recommended unless you
110 know what you are doing.
111 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
113 *) Fixes to stateless session resumption handling. Use initial_ctx when
114 issuing and attempting to decrypt tickets in case it has changed during
115 servername handling. Use a non-zero length session ID when attempting
116 stateless session resumption: this makes it possible to determine if
117 a resumption has occurred immediately after receiving server hello
118 (several places in OpenSSL subtly assume this) instead of later in
122 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
123 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
124 fixes for a few places where the return code is not checked
126 [Julia Lawall <julia@diku.dk>]
128 *) Add --strict-warnings option to Configure script to include devteam
129 warnings in other configurations.
132 *) Add support for --libdir option and LIBDIR variable in makefiles. This
133 makes it possible to install openssl libraries in locations which
134 have names other than "lib", for example "/usr/lib64" which some
136 [Steve Henson, based on patch from Jeremy Utley]
138 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
139 X690 8.9.12 and can produce some misleading textual output of OIDs.
140 [Steve Henson, reported by Dan Kaminsky]
142 *) Delete MD2 from algorithm tables. This follows the recommendation in
143 several standards that it is not used in new applications due to
144 several cryptographic weaknesses. For binary compatibility reasons
145 the MD2 API is still compiled in by default.
148 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
152 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
153 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
155 [Guenter <lists@gknw.net>]
157 *) Fix the server certificate chain building code to use X509_verify_cert(),
158 it used to have an ad-hoc builder which was unable to cope with anything
159 other than a simple chain.
160 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
162 *) Don't check self signed certificate signatures in X509_verify_cert()
163 by default (a flag can override this): it just wastes time without
164 adding any security. As a useful side effect self signed root CAs
165 with non-FIPS digests are now usable in FIPS mode.
168 *) In dtls1_process_out_of_seq_message() the check if the current message
169 is already buffered was missing. For every new message was memory
170 allocated, allowing an attacker to perform an denial of service attack
171 with sending out of seq handshake messages until there is no memory
172 left. Additionally every future messege was buffered, even if the
173 sequence number made no sense and would be part of another handshake.
174 So only messages with sequence numbers less than 10 in advance will be
175 buffered. (CVE-2009-1378)
176 [Robin Seggelmann, discovered by Daniel Mentz]
178 *) Records are buffered if they arrive with a future epoch to be
179 processed after finishing the corresponding handshake. There is
180 currently no limitation to this buffer allowing an attacker to perform
181 a DOS attack with sending records with future epochs until there is no
182 memory left. This patch adds the pqueue_size() function to detemine
183 the size of a buffer and limits the record buffer to 100 entries.
185 [Robin Seggelmann, discovered by Daniel Mentz]
187 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
188 parent structure is freed. (CVE-2009-1379)
191 *) Handle non-blocking I/O properly in SSL_shutdown() call.
192 [Darryl Miles <darryl-mailinglists@netbauds.net>]
195 [Ilya O. <vrghost@gmail.com>]
197 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
199 *) Disable renegotiation completely - this fixes a severe security
200 problem (CVE-2009-3555) at the cost of breaking all
201 renegotiation. Renegotiation can be re-enabled by setting
202 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
203 run-time. This is really not recommended unless you know what
207 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
209 *) Don't set val to NULL when freeing up structures, it is freed up by
210 underlying code. If sizeof(void *) > sizeof(long) this can result in
211 zeroing past the valid field. (CVE-2009-0789)
212 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
214 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
215 checked correctly. This would allow some invalid signed attributes to
216 appear to verify correctly. (CVE-2009-0591)
217 [Ivan Nestlerode <inestlerode@us.ibm.com>]
219 *) Reject UniversalString and BMPString types with invalid lengths. This
220 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
221 a legal length. (CVE-2009-0590)
224 *) Set S/MIME signing as the default purpose rather than setting it
225 unconditionally. This allows applications to override it at the store
229 *) Permit restricted recursion of ASN1 strings. This is needed in practice
230 to handle some structures.
233 *) Improve efficiency of mem_gets: don't search whole buffer each time
235 [Jeremy Shapiro <jnshapir@us.ibm.com>]
237 *) New -hex option for openssl rand.
240 *) Print out UTF8String and NumericString when parsing ASN1.
243 *) Support NumericString type for name components.
246 *) Allow CC in the environment to override the automatically chosen
247 compiler. Note that nothing is done to ensure flags work with the
251 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
253 *) Properly check EVP_VerifyFinal() and similar return values
255 [Ben Laurie, Bodo Moeller, Google Security Team]
257 *) Enable TLS extensions by default.
260 *) Allow the CHIL engine to be loaded, whether the application is
261 multithreaded or not. (This does not release the developer from the
262 obligation to set up the dynamic locking callbacks.)
263 [Sander Temme <sander@temme.net>]
265 *) Use correct exit code if there is an error in dgst command.
266 [Steve Henson; problem pointed out by Roland Dirlewanger]
268 *) Tweak Configure so that you need to say "experimental-jpake" to enable
269 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
272 *) Add experimental JPAKE support, including demo authentication in
273 s_client and s_server.
276 *) Set the comparison function in v3_addr_canonize().
277 [Rob Austein <sra@hactrn.net>]
279 *) Add support for XMPP STARTTLS in s_client.
280 [Philip Paeps <philip@freebsd.org>]
282 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
283 to ensure that even with this option, only ciphersuites in the
284 server's preference list will be accepted. (Note that the option
285 applies only when resuming a session, so the earlier behavior was
286 just about the algorithm choice for symmetric cryptography.)
289 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
291 *) Fix NULL pointer dereference if a DTLS server received
292 ChangeCipherSpec as first record (CVE-2009-1386).
295 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
296 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
299 *) The fix in 0.9.8c that supposedly got rid of unsafe
300 double-checked locking was incomplete for RSA blinding,
301 addressing just one layer of what turns out to have been
302 doubly unsafe triple-checked locking.
304 So now fix this for real by retiring the MONT_HELPER macro
305 in crypto/rsa/rsa_eay.c.
307 [Bodo Moeller; problem pointed out by Marius Schilder]
309 *) Various precautionary measures:
311 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
313 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
314 (NB: This would require knowledge of the secret session ticket key
315 to exploit, in which case you'd be SOL either way.)
317 - Change bn_nist.c so that it will properly handle input BIGNUMs
318 outside the expected range.
320 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
323 [Neel Mehta, Bodo Moeller]
325 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
326 the load fails. Useful for distros.
327 [Ben Laurie and the FreeBSD team]
329 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
332 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
335 *) Expand ENGINE to support engine supplied SSL client certificate functions.
337 This work was sponsored by Logica.
340 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
341 keystores. Support for SSL/TLS client authentication too.
342 Not compiled unless enable-capieng specified to Configure.
344 This work was sponsored by Logica.
347 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
348 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
349 attribute creation routines such as certifcate requests and PKCS#12
353 Changes between 0.9.8g and 0.9.8h [28 May 2008]
355 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
356 handshake which could lead to a cilent crash as found using the
357 Codenomicon TLS test suite (CVE-2008-1672)
358 [Steve Henson, Mark Cox]
360 *) Fix double free in TLS server name extensions which could lead to
361 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
364 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
366 Clear the error queue to ensure that error entries left from
367 older function calls do not interfere with the correct operation.
368 [Lutz Jaenicke, Erik de Castro Lopo]
370 *) Remove root CA certificates of commercial CAs:
372 The OpenSSL project does not recommend any specific CA and does not
373 have any policy with respect to including or excluding any CA.
374 Therefore it does not make any sense to ship an arbitrary selection
375 of root CA certificates with the OpenSSL software.
378 *) RSA OAEP patches to fix two separate invalid memory reads.
379 The first one involves inputs when 'lzero' is greater than
380 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
381 before the beginning of from). The second one involves inputs where
382 the 'db' section contains nothing but zeroes (there is a one-byte
383 invalid read after the end of 'db').
384 [Ivan Nestlerode <inestlerode@us.ibm.com>]
386 *) Partial backport from 0.9.9-dev:
388 Introduce bn_mul_mont (dedicated Montgomery multiplication
389 procedure) as a candidate for BIGNUM assembler implementation.
390 While 0.9.9-dev uses assembler for various architectures, only
391 x86_64 is available by default here in the 0.9.8 branch, and
392 32-bit x86 is available through a compile-time setting.
394 To try the 32-bit x86 assembler implementation, use Configure
395 option "enable-montasm" (which exists only for this backport).
397 As "enable-montasm" for 32-bit x86 disclaims code stability
398 anyway, in this constellation we activate additional code
399 backported from 0.9.9-dev for further performance improvements,
400 namely BN_from_montgomery_word. (To enable this otherwise,
401 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
403 [Andy Polyakov (backport partially by Bodo Moeller)]
405 *) Add TLS session ticket callback. This allows an application to set
406 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
407 values. This is useful for key rollover for example where several key
408 sets may exist with different names.
411 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
412 This was broken until now in 0.9.8 releases, such that the only way
413 a registered ENGINE could be used (assuming it initialises
414 successfully on the host) was to explicitly set it as the default
415 for the relevant algorithms. This is in contradiction with 0.9.7
416 behaviour and the documentation. With this fix, when an ENGINE is
417 registered into a given algorithm's table of implementations, the
418 'uptodate' flag is reset so that auto-discovery will be used next
419 time a new context for that algorithm attempts to select an
421 [Ian Lister (tweaked by Geoff Thorpe)]
423 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
424 implemention in the following ways:
426 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
429 Lack of BER streaming support means one pass streaming processing is
430 only supported if data is detached: setting the streaming flag is
431 ignored for embedded content.
433 CMS support is disabled by default and must be explicitly enabled
434 with the enable-cms configuration option.
437 *) Update the GMP engine glue to do direct copies between BIGNUM and
438 mpz_t when openssl and GMP use the same limb size. Otherwise the
439 existing "conversion via a text string export" trick is still used.
440 [Paul Sheer <paulsheer@gmail.com>]
442 *) Zlib compression BIO. This is a filter BIO which compressed and
443 uncompresses any data passed through it.
446 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
447 RFC3394 compatible AES key wrapping.
450 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
451 sets string data without copying. X509_ALGOR_set0() and
452 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
453 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
454 from an X509_ATTRIBUTE structure optionally checking it occurs only
455 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
459 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
460 to get the expected BN_FLG_CONSTTIME behavior.
461 [Bodo Moeller (Google)]
465 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
466 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
467 - added some more tests to do_tests.pl
468 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
469 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
470 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
471 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
472 - various changes to netware.pl to enable gcc-cross builds on Win32
474 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
475 - various changes to fix missing prototype warnings
476 - fixed x86nasm.pl to create correct asm files for NASM COFF output
477 - added AES, WHIRLPOOL and CPUID assembler code to build files
478 - added missing AES assembler make rules to mk1mf.pl
479 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
480 [Guenter Knauf <eflash@gmx.net>]
482 *) Implement certificate status request TLS extension defined in RFC3546.
483 A client can set the appropriate parameters and receive the encoded
484 OCSP response via a callback. A server can query the supplied parameters
485 and set the encoded OCSP response in the callback. Add simplified examples
486 to s_client and s_server.
489 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
492 + Binary incompatibility of ssl_ctx_st structure
493 + DTLS interoperation with non-compliant servers
494 + Don't call get_session_cb() without proposed session
495 + Fix ia64 assembler code
496 [Andy Polyakov, Steve Henson]
498 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
500 *) DTLS Handshake overhaul. There were longstanding issues with
501 OpenSSL DTLS implementation, which were making it impossible for
502 RFC 4347 compliant client to communicate with OpenSSL server.
503 Unfortunately just fixing these incompatibilities would "cut off"
504 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
505 server keeps tolerating non RFC compliant syntax. The opposite is
506 not true, 0.9.8f client can not communicate with earlier server.
507 This update even addresses CVE-2007-4995.
510 *) Changes to avoid need for function casts in OpenSSL: some compilers
511 (gcc 4.2 and later) reject their use.
512 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
515 *) Add RFC4507 support to OpenSSL. This includes the corrections in
516 RFC4507bis. The encrypted ticket format is an encrypted encoded
517 SSL_SESSION structure, that way new session features are automatically
520 If a client application caches session in an SSL_SESSION structure
521 support is transparent because tickets are now stored in the encoded
524 The SSL_CTX structure automatically generates keys for ticket
525 protection in servers so again support should be possible
526 with no application modification.
528 If a client or server wishes to disable RFC4507 support then the option
529 SSL_OP_NO_TICKET can be set.
531 Add a TLS extension debugging callback to allow the contents of any client
532 or server extensions to be examined.
534 This work was sponsored by Google.
537 *) Add initial support for TLS extensions, specifically for the server_name
538 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
539 have new members for a host name. The SSL data structure has an
540 additional member SSL_CTX *initial_ctx so that new sessions can be
541 stored in that context to allow for session resumption, even after the
542 SSL has been switched to a new SSL_CTX in reaction to a client's
543 server_name extension.
545 New functions (subject to change):
548 SSL_get_servername_type()
551 New CTRL codes and macros (subject to change):
553 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
554 - SSL_CTX_set_tlsext_servername_callback()
555 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
556 - SSL_CTX_set_tlsext_servername_arg()
557 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
559 openssl s_client has a new '-servername ...' option.
561 openssl s_server has new options '-servername_host ...', '-cert2 ...',
562 '-key2 ...', '-servername_fatal' (subject to change). This allows
563 testing the HostName extension for a specific single host name ('-cert'
564 and '-key' remain fallbacks for handshakes without HostName
565 negotiation). If the unrecogninzed_name alert has to be sent, this by
566 default is a warning; it becomes fatal with the '-servername_fatal'
569 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
571 *) Add AES and SSE2 assembly language support to VC++ build.
574 *) Mitigate attack on final subtraction in Montgomery reduction.
577 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
578 (which previously caused an internal error).
581 *) Squeeze another 10% out of IGE mode when in != out.
584 *) AES IGE mode speedup.
585 [Dean Gaudet (Google)]
587 *) Add the Korean symmetric 128-bit cipher SEED (see
588 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
589 add SEED ciphersuites from RFC 4162:
591 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
592 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
593 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
594 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
596 To minimize changes between patchlevels in the OpenSSL 0.9.8
597 series, SEED remains excluded from compilation unless OpenSSL
598 is configured with 'enable-seed'.
601 *) Mitigate branch prediction attacks, which can be practical if a
602 single processor is shared, allowing a spy process to extract
603 information. For detailed background information, see
604 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
605 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
606 and Necessary Software Countermeasures"). The core of the change
607 are new versions BN_div_no_branch() and
608 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
609 respectively, which are slower, but avoid the security-relevant
610 conditional branches. These are automatically called by BN_div()
611 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
612 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
613 remove a conditional branch.
615 BN_FLG_CONSTTIME is the new name for the previous
616 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
617 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
618 in the exponent causes BN_mod_exp_mont() to use the alternative
619 implementation in BN_mod_exp_mont_consttime().) The old name
620 remains as a deprecated alias.
622 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
623 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
624 constant-time implementations for more than just exponentiation.
625 Here too the old name is kept as a deprecated alias.
627 BN_BLINDING_new() will now use BN_dup() for the modulus so that
628 the BN_BLINDING structure gets an independent copy of the
629 modulus. This means that the previous "BIGNUM *m" argument to
630 BN_BLINDING_new() and to BN_BLINDING_create_param() now
631 essentially becomes "const BIGNUM *m", although we can't actually
632 change this in the header file before 0.9.9. It allows
633 RSA_setup_blinding() to use BN_with_flags() on the modulus to
634 enable BN_FLG_CONSTTIME.
636 [Matthew D Wood (Intel Corp)]
638 *) In the SSL/TLS server implementation, be strict about session ID
639 context matching (which matters if an application uses a single
640 external cache for different purposes). Previously,
641 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
642 set. This did ensure strict client verification, but meant that,
643 with applications using a single external cache for quite
644 different requirements, clients could circumvent ciphersuite
645 restrictions for a given session ID context by starting a session
646 in a different context.
649 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
650 a ciphersuite string such as "DEFAULT:RSA" cannot enable
651 authentication-only ciphersuites.
654 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
655 not complete and could lead to a possible single byte overflow
656 (CVE-2007-5135) [Ben Laurie]
658 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
660 *) Since AES128 and AES256 (and similarly Camellia128 and
661 Camellia256) share a single mask bit in the logic of
662 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
663 kludge to work properly if AES128 is available and AES256 isn't
664 (or if Camellia128 is available and Camellia256 isn't).
667 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
668 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
669 When a point or a seed is encoded in a BIT STRING, we need to
670 prevent the removal of trailing zero bits to get the proper DER
671 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
672 of a NamedBitList, for which trailing 0 bits need to be removed.)
675 *) Have SSL/TLS server implementation tolerate "mismatched" record
676 protocol version while receiving ClientHello even if the
677 ClientHello is fragmented. (The server can't insist on the
678 particular protocol version it has chosen before the ServerHello
679 message has informed the client about his choice.)
682 *) Add RFC 3779 support.
683 [Rob Austein for ARIN, Ben Laurie]
685 *) Load error codes if they are not already present instead of using a
686 static variable. This allows them to be cleanly unloaded and reloaded.
687 Improve header file function name parsing.
690 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
691 or CAPABILITY handshake as required by RFCs.
694 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
696 *) Introduce limits to prevent malicious keys being able to
697 cause a denial of service. (CVE-2006-2940)
698 [Steve Henson, Bodo Moeller]
700 *) Fix ASN.1 parsing of certain invalid structures that can result
701 in a denial of service. (CVE-2006-2937) [Steve Henson]
703 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
704 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
706 *) Fix SSL client code which could crash if connecting to a
707 malicious SSLv2 server. (CVE-2006-4343)
708 [Tavis Ormandy and Will Drewry, Google Security Team]
710 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
711 match only those. Before that, "AES256-SHA" would be interpreted
712 as a pattern and match "AES128-SHA" too (since AES128-SHA got
713 the same strength classification in 0.9.7h) as we currently only
714 have a single AES bit in the ciphersuite description bitmap.
715 That change, however, also applied to ciphersuite strings such as
716 "RC4-MD5" that intentionally matched multiple ciphersuites --
717 namely, SSL 2.0 ciphersuites in addition to the more common ones
718 from SSL 3.0/TLS 1.0.
720 So we change the selection algorithm again: Naming an explicit
721 ciphersuite selects this one ciphersuite, and any other similar
722 ciphersuite (same bitmap) from *other* protocol versions.
723 Thus, "RC4-MD5" again will properly select both the SSL 2.0
724 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
726 Since SSL 2.0 does not have any ciphersuites for which the
727 128/256 bit distinction would be relevant, this works for now.
728 The proper fix will be to use different bits for AES128 and
729 AES256, which would have avoided the problems from the beginning;
730 however, bits are scarce, so we can only do this in a new release
731 (not just a patchlevel) when we can change the SSL_CIPHER
732 definition to split the single 'unsigned long mask' bitmap into
733 multiple values to extend the available space.
737 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
739 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
740 (CVE-2006-4339) [Ben Laurie and Google Security Team]
742 *) Add AES IGE and biIGE modes.
745 *) Change the Unix randomness entropy gathering to use poll() when
746 possible instead of select(), since the latter has some
747 undesirable limitations.
748 [Darryl Miles via Richard Levitte and Bodo Moeller]
750 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
751 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
752 cannot be implicitly activated as part of, e.g., the "AES" alias.
753 However, please upgrade to OpenSSL 0.9.9[-dev] for
754 non-experimental use of the ECC ciphersuites to get TLS extension
755 support, which is required for curve and point format negotiation
756 to avoid potential handshake problems.
759 *) Disable rogue ciphersuites:
761 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
762 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
763 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
765 The latter two were purportedly from
766 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
769 Also deactivate the remaining ciphersuites from
770 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
771 unofficial, and the ID has long expired.
774 *) Fix RSA blinding Heisenbug (problems sometimes occured on
775 dual-core machines) and other potential thread-safety issues.
778 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
779 versions), which is now available for royalty-free use
780 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
781 Also, add Camellia TLS ciphersuites from RFC 4132.
783 To minimize changes between patchlevels in the OpenSSL 0.9.8
784 series, Camellia remains excluded from compilation unless OpenSSL
785 is configured with 'enable-camellia'.
788 *) Disable the padding bug check when compression is in use. The padding
789 bug check assumes the first packet is of even length, this is not
790 necessarily true if compresssion is enabled and can result in false
791 positives causing handshake failure. The actual bug test is ancient
792 code so it is hoped that implementations will either have fixed it by
793 now or any which still have the bug do not support compression.
796 Changes between 0.9.8a and 0.9.8b [04 May 2006]
798 *) When applying a cipher rule check to see if string match is an explicit
799 cipher suite and only match that one cipher suite if it is.
802 *) Link in manifests for VC++ if needed.
803 [Austin Ziegler <halostatue@gmail.com>]
805 *) Update support for ECC-based TLS ciphersuites according to
806 draft-ietf-tls-ecc-12.txt with proposed changes (but without
807 TLS extensions, which are supported starting with the 0.9.9
808 branch, not in the OpenSSL 0.9.8 branch).
811 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
812 opaque EVP_CIPHER_CTX handling.
815 *) Fixes and enhancements to zlib compression code. We now only use
816 "zlib1.dll" and use the default __cdecl calling convention on Win32
817 to conform with the standards mentioned here:
818 http://www.zlib.net/DLL_FAQ.txt
819 Static zlib linking now works on Windows and the new --with-zlib-include
820 --with-zlib-lib options to Configure can be used to supply the location
821 of the headers and library. Gracefully handle case where zlib library
825 *) Several fixes and enhancements to the OID generation code. The old code
826 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
827 handle numbers larger than ULONG_MAX, truncated printing and had a
828 non standard OBJ_obj2txt() behaviour.
831 *) Add support for building of engines under engine/ as shared libraries
832 under VC++ build system.
835 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
836 Hopefully, we will not see any false combination of paths any more.
839 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
841 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
842 (part of SSL_OP_ALL). This option used to disable the
843 countermeasure against man-in-the-middle protocol-version
844 rollback in the SSL 2.0 server implementation, which is a bad
845 idea. (CVE-2005-2969)
847 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
848 for Information Security, National Institute of Advanced Industrial
849 Science and Technology [AIST], Japan)]
851 *) Add two function to clear and return the verify parameter flags.
854 *) Keep cipherlists sorted in the source instead of sorting them at
855 runtime, thus removing the need for a lock.
858 *) Avoid some small subgroup attacks in Diffie-Hellman.
859 [Nick Mathewson and Ben Laurie]
861 *) Add functions for well-known primes.
864 *) Extended Windows CE support.
865 [Satoshi Nakamura and Andy Polyakov]
867 *) Initialize SSL_METHOD structures at compile time instead of during
868 runtime, thus removing the need for a lock.
871 *) Make PKCS7_decrypt() work even if no certificate is supplied by
872 attempting to decrypt each encrypted key in turn. Add support to
876 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
878 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
881 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
884 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
885 key into the same file any more.
888 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
891 *) Add -utf8 command line and config file option to 'ca'.
892 [Stefan <stf@udoma.org]
894 *) Removed the macro des_crypt(), as it seems to conflict with some
895 libraries. Use DES_crypt().
898 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
899 involves renaming the source and generated shared-libs for
900 both. The engines will accept the corrected or legacy ids
901 ('ncipher' and '4758_cca' respectively) when binding. NB,
902 this only applies when building 'shared'.
903 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
905 *) Add attribute functions to EVP_PKEY structure. Modify
906 PKCS12_create() to recognize a CSP name attribute and
907 use it. Make -CSP option work again in pkcs12 utility.
910 *) Add new functionality to the bn blinding code:
911 - automatic re-creation of the BN_BLINDING parameters after
912 a fixed number of uses (currently 32)
913 - add new function for parameter creation
914 - introduce flags to control the update behaviour of the
915 BN_BLINDING parameters
916 - hide BN_BLINDING structure
917 Add a second BN_BLINDING slot to the RSA structure to improve
918 performance when a single RSA object is shared among several
922 *) Add support for DTLS.
923 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
925 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
926 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
929 *) Remove buggy and incompletet DH cert support from
930 ssl/ssl_rsa.c and ssl/s3_both.c
933 *) Use SHA-1 instead of MD5 as the default digest algorithm for
934 the apps/openssl applications.
937 *) Compile clean with "-Wall -Wmissing-prototypes
938 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
939 DEBUG_SAFESTACK must also be set.
942 *) Change ./Configure so that certain algorithms can be disabled by default.
943 The new counterpiece to "no-xxx" is "enable-xxx".
945 The patented RC5 and MDC2 algorithms will now be disabled unless
946 "enable-rc5" and "enable-mdc2", respectively, are specified.
948 (IDEA remains enabled despite being patented. This is because IDEA
949 is frequently required for interoperability, and there is no license
950 fee for non-commercial use. As before, "no-idea" can be used to
951 avoid this algorithm.)
955 *) Add processing of proxy certificates (see RFC 3820). This work was
956 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
957 EGEE (Enabling Grids for E-science in Europe).
960 *) RC4 performance overhaul on modern architectures/implementations, such
961 as Intel P4, IA-64 and AMD64.
964 *) New utility extract-section.pl. This can be used specify an alternative
965 section number in a pod file instead of having to treat each file as
966 a separate case in Makefile. This can be done by adding two lines to the
969 =for comment openssl_section:XXX
971 The blank line is mandatory.
975 *) New arguments -certform, -keyform and -pass for s_client and s_server
976 to allow alternative format key and certificate files and passphrase
980 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
981 update associated structures and add various utility functions.
983 Add new policy related verify parameters, include policy checking in
984 standard verify code. Enhance 'smime' application with extra parameters
985 to support policy checking and print out.
988 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
989 Nehemiah processors. These extensions support AES encryption in hardware
990 as well as RNG (though RNG support is currently disabled).
991 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
993 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
996 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
997 [Andy Polyakov and a number of other people]
999 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1000 implementation contributed by IBM.
1001 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1003 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1004 exponent rather than 'unsigned long'. There is a corresponding change to
1005 the new 'rsa_keygen' element of the RSA_METHOD structure.
1006 [Jelte Jansen, Geoff Thorpe]
1008 *) Functionality for creating the initial serial number file is now
1009 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1011 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1012 number file to 1, which is bound to cause problems. To avoid
1013 the problems while respecting compatibility between different 0.9.7
1014 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1015 CA.pl for serial number initialization. With the new release 0.9.8,
1016 we can fix the problem directly in the 'ca' utility.)
1019 *) Reduced header interdepencies by declaring more opaque objects in
1020 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1021 give fewer recursive includes, which could break lazy source code - so
1022 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1023 developers should define this symbol when building and using openssl to
1024 ensure they track the recommended behaviour, interfaces, [etc], but
1025 backwards-compatible behaviour prevails when this isn't defined.
1028 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1031 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1032 This will generate a random key of the appropriate length based on the
1033 cipher context. The EVP_CIPHER can provide its own random key generation
1034 routine to support keys of a specific form. This is used in the des and
1035 3des routines to generate a key of the correct parity. Update S/MIME
1036 code to use new functions and hence generate correct parity DES keys.
1037 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1038 valid (weak or incorrect parity).
1041 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1042 as looking them up. This is useful when the verified structure may contain
1043 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1044 present unless the new PKCS7_NO_CRL flag is asserted.
1047 *) Extend ASN1 oid configuration module. It now additionally accepts the
1050 shortName = some long name, 1.2.3.4
1053 *) Reimplemented the BN_CTX implementation. There is now no more static
1054 limitation on the number of variables it can handle nor the depth of the
1055 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1056 information can now expand as required, and rather than having a single
1057 static array of bignums, BN_CTX now uses a linked-list of such arrays
1058 allowing it to expand on demand whilst maintaining the usefulness of
1059 BN_CTX's "bundling".
1062 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1063 to allow all RSA operations to function using a single BN_CTX.
1066 *) Preliminary support for certificate policy evaluation and checking. This
1067 is initially intended to pass the tests outlined in "Conformance Testing
1068 of Relying Party Client Certificate Path Processing Logic" v1.07.
1071 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1072 remained unused and not that useful. A variety of other little bignum
1073 tweaks and fixes have also been made continuing on from the audit (see
1077 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1078 associated ASN1, EVP and SSL functions and old ASN1 macros.
1081 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1082 and this should never fail. So the return value from the use of
1083 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1084 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1087 *) BN_CTX_get() should return zero-valued bignums, providing the same
1088 initialised value as BN_new().
1089 [Geoff Thorpe, suggested by Ulf Möller]
1091 *) Support for inhibitAnyPolicy certificate extension.
1094 *) An audit of the BIGNUM code is underway, for which debugging code is
1095 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1096 is considered valid when processing BIGNUMs, and causes execution to
1097 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1098 further steps are taken to deliberately pollute unused data in BIGNUM
1099 structures to try and expose faulty code further on. For now, openssl will
1100 (in its default mode of operation) continue to tolerate the inconsistent
1101 forms that it has tolerated in the past, but authors and packagers should
1102 consider trying openssl and their own applications when compiled with
1103 these debugging symbols defined. It will help highlight potential bugs in
1104 their own code, and will improve the test coverage for OpenSSL itself. At
1105 some point, these tighter rules will become openssl's default to improve
1106 maintainability, though the assert()s and other overheads will remain only
1107 in debugging configurations. See bn.h for more details.
1108 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1110 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1111 that can only be obtained through BN_CTX_new() (which implicitly
1112 initialises it). The presence of this function only made it possible
1113 to overwrite an existing structure (and cause memory leaks).
1116 *) Because of the callback-based approach for implementing LHASH as a
1117 template type, lh_insert() adds opaque objects to hash-tables and
1118 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1119 to clean up those corresponding objects before destroying the hash table
1120 (and losing the object pointers). So some over-zealous constifications in
1121 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1122 objects as "const" and the lh_doall[_arg] callback wrappers are not
1123 prototyped to have "const" restrictions on the object pointers they are
1124 given (and so aren't required to cast them away any more).
1127 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1128 (speed) prefers to use its own implementation. The two implementations
1129 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1130 its object type properly exposed (MS_TM) instead of casting to/from "char
1131 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1132 aren't necessarily the greatest nomenclatures - but this is what was used
1133 internally to the implementation so I've used that for now.
1136 *) Ensure that deprecated functions do not get compiled when
1137 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1138 the self-tests were still using deprecated key-generation functions so
1139 these have been updated also.
1142 *) Reorganise PKCS#7 code to separate the digest location functionality
1143 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1144 New function PKCS7_set_digest() to set the digest type for PKCS#7
1145 digestedData type. Add additional code to correctly generate the
1146 digestedData type and add support for this type in PKCS7 initialization
1150 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1151 structure of type "other".
1154 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1155 sure the loop does correctly stop and breaking ("division by zero")
1156 modulus operations are not performed. The (pre-generated) prime
1157 table crypto/bn/bn_prime.h was already correct, but it could not be
1158 re-generated on some platforms because of the "division by zero"
1159 situation in the script.
1160 [Ralf S. Engelschall]
1162 *) Update support for ECC-based TLS ciphersuites according to
1163 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1164 SHA-1 now is only used for "small" curves (where the
1165 representation of a field element takes up to 24 bytes); for
1166 larger curves, the field element resulting from ECDH is directly
1167 used as premaster secret.
1168 [Douglas Stebila (Sun Microsystems Laboratories)]
1170 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1171 curve secp160r1 to the tests.
1172 [Douglas Stebila (Sun Microsystems Laboratories)]
1174 *) Add the possibility to load symbols globally with DSO.
1175 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1177 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1178 control of the error stack.
1181 *) Add support for STORE in ENGINE.
1184 *) Add the STORE type. The intention is to provide a common interface
1185 to certificate and key stores, be they simple file-based stores, or
1186 HSM-type store, or LDAP stores, or...
1187 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1190 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1191 pass a list of arguments to any function as well as provide a way
1192 for a function to pass data back to the caller.
1195 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1196 works like BUF_strdup() but can be used to duplicate a portion of
1197 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1201 *) Add the function sk_find_ex() which works like sk_find(), but will
1202 return an index to an element even if an exact match couldn't be
1203 found. The index is guaranteed to point at the element where the
1204 searched-for key would be inserted to preserve sorting order.
1207 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1208 takes an extra flags argument for optional functionality. Currently,
1209 the following flags are defined:
1211 OBJ_BSEARCH_VALUE_ON_NOMATCH
1212 This one gets OBJ_bsearch_ex() to return a pointer to the first
1213 element where the comparing function returns a negative or zero
1216 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1217 This one gets OBJ_bsearch_ex() to return a pointer to the first
1218 element where the comparing function returns zero. This is useful
1219 if there are more than one element where the comparing function
1223 *) Make it possible to create self-signed certificates with 'openssl ca'
1224 in such a way that the self-signed certificate becomes part of the
1225 CA database and uses the same mechanisms for serial number generation
1226 as all other certificate signing. The new flag '-selfsign' enables
1227 this functionality. Adapt CA.sh and CA.pl.in.
1230 *) Add functionality to check the public key of a certificate request
1231 against a given private. This is useful to check that a certificate
1232 request can be signed by that key (self-signing).
1235 *) Make it possible to have multiple active certificates with the same
1236 subject in the CA index file. This is done only if the keyword
1237 'unique_subject' is set to 'no' in the main CA section (default
1238 if 'CA_default') of the configuration file. The value is saved
1239 with the database itself in a separate index attribute file,
1240 named like the index file with '.attr' appended to the name.
1243 *) Generate muti valued AVAs using '+' notation in config files for
1247 *) Support for nameConstraints certificate extension.
1250 *) Support for policyConstraints certificate extension.
1253 *) Support for policyMappings certificate extension.
1256 *) Make sure the default DSA_METHOD implementation only uses its
1257 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1258 and change its own handlers to be NULL so as to remove unnecessary
1259 indirection. This lets alternative implementations fallback to the
1260 default implementation more easily.
1263 *) Support for directoryName in GeneralName related extensions
1267 *) Make it possible to link applications using Makefile.shared.
1268 Make that possible even when linking against static libraries!
1271 *) Support for single pass processing for S/MIME signing. This now
1272 means that S/MIME signing can be done from a pipe, in addition
1273 cleartext signing (multipart/signed type) is effectively streaming
1274 and the signed data does not need to be all held in memory.
1276 This is done with a new flag PKCS7_STREAM. When this flag is set
1277 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1278 is done after the data is output (and digests calculated) in
1279 SMIME_write_PKCS7().
1282 *) Add full support for -rpath/-R, both in shared libraries and
1283 applications, at least on the platforms where it's known how
1287 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1288 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1289 will now compute a table of multiples of the generator that
1290 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1291 faster (notably in the case of a single point multiplication,
1292 scalar * generator).
1293 [Nils Larsch, Bodo Moeller]
1295 *) IPv6 support for certificate extensions. The various extensions
1296 which use the IP:a.b.c.d can now take IPv6 addresses using the
1297 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1301 *) Added an ENGINE that implements RSA by performing private key
1302 exponentiations with the GMP library. The conversions to and from
1303 GMP's mpz_t format aren't optimised nor are any montgomery forms
1304 cached, and on x86 it appears OpenSSL's own performance has caught up.
1305 However there are likely to be other architectures where GMP could
1306 provide a boost. This ENGINE is not built in by default, but it can be
1307 specified at Configure time and should be accompanied by the necessary
1308 linker additions, eg;
1309 ./config -DOPENSSL_USE_GMP -lgmp
1312 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1313 testing availability of engines with "-t" - the old behaviour is
1314 produced by increasing the feature's verbosity with "-tt".
1317 *) ECDSA routines: under certain error conditions uninitialized BN objects
1318 could be freed. Solution: make sure initialization is performed early
1319 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1323 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1324 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1325 software implementations. For DSA and DH, parameter generation can
1326 also be overriden by providing the appropriate method callbacks.
1329 *) Change the "progress" mechanism used in key-generation and
1330 primality testing to functions that take a new BN_GENCB pointer in
1331 place of callback/argument pairs. The new API functions have "_ex"
1332 postfixes and the older functions are reimplemented as wrappers for
1333 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1334 declarations of the old functions to help (graceful) attempts to
1335 migrate to the new functions. Also, the new key-generation API
1336 functions operate on a caller-supplied key-structure and return
1337 success/failure rather than returning a key or NULL - this is to
1338 help make "keygen" another member function of RSA_METHOD etc.
1340 Example for using the new callback interface:
1342 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1346 BN_GENCB_set(&my_cb, my_callback, my_arg);
1348 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1349 /* For the meaning of a, b in calls to my_callback(), see the
1350 * documentation of the function that calls the callback.
1351 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1352 * my_callback should return 1 if it wants BN_is_prime_ex()
1353 * to continue, or 0 to stop.
1358 *) Change the ZLIB compression method to be stateful, and make it
1359 available to TLS with the number defined in
1360 draft-ietf-tls-compression-04.txt.
1363 *) Add the ASN.1 structures and functions for CertificatePair, which
1364 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1366 CertificatePair ::= SEQUENCE {
1367 forward [0] Certificate OPTIONAL,
1368 reverse [1] Certificate OPTIONAL,
1369 -- at least one of the pair shall be present -- }
1371 Also implement the PEM functions to read and write certificate
1372 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1374 This needed to be defined, mostly for the sake of the LDAP
1375 attribute crossCertificatePair, but may prove useful elsewhere as
1379 *) Make it possible to inhibit symlinking of shared libraries in
1380 Makefile.shared, for Cygwin's sake.
1383 *) Extend the BIGNUM API by creating a function
1384 void BN_set_negative(BIGNUM *a, int neg);
1385 and a macro that behave like
1386 int BN_is_negative(const BIGNUM *a);
1388 to avoid the need to access 'a->neg' directly in applications.
1391 *) Implement fast modular reduction for pseudo-Mersenne primes
1392 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1393 EC_GROUP_new_curve_GFp() will now automatically use this
1395 [Nils Larsch <nla@trustcenter.de>]
1397 *) Add new lock type (CRYPTO_LOCK_BN).
1400 *) Change the ENGINE framework to automatically load engines
1401 dynamically from specific directories unless they could be
1402 found to already be built in or loaded. Move all the
1403 current engines except for the cryptodev one to a new
1405 The engines in engines/ are built as shared libraries if
1406 the "shared" options was given to ./Configure or ./config.
1407 Otherwise, they are inserted in libcrypto.a.
1408 /usr/local/ssl/engines is the default directory for dynamic
1409 engines, but that can be overriden at configure time through
1410 the usual use of --prefix and/or --openssldir, and at run
1411 time with the environment variable OPENSSL_ENGINES.
1412 [Geoff Thorpe and Richard Levitte]
1414 *) Add Makefile.shared, a helper makefile to build shared
1415 libraries. Addapt Makefile.org.
1418 *) Add version info to Win32 DLLs.
1419 [Peter 'Luna' Runestig" <peter@runestig.com>]
1421 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1422 can be added using this API to created arbitrary PKCS#12
1423 files while avoiding the low level API.
1425 New options to PKCS12_create(), key or cert can be NULL and
1426 will then be omitted from the output file. The encryption
1427 algorithm NIDs can be set to -1 for no encryption, the mac
1428 iteration count can be set to 0 to omit the mac.
1430 Enhance pkcs12 utility by making the -nokeys and -nocerts
1431 options work when creating a PKCS#12 file. New option -nomac
1432 to omit the mac, NONE can be set for an encryption algorithm.
1433 New code is modified to use the enhanced PKCS12_create()
1434 instead of the low level API.
1437 *) Extend ASN1 encoder to support indefinite length constructed
1438 encoding. This can output sequences tags and octet strings in
1439 this form. Modify pk7_asn1.c to support indefinite length
1440 encoding. This is experimental and needs additional code to
1441 be useful, such as an ASN1 bio and some enhanced streaming
1444 Extend template encode functionality so that tagging is passed
1445 down to the template encoder.
1448 *) Let 'openssl req' fail if an argument to '-newkey' is not
1449 recognized instead of using RSA as a default.
1452 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1453 As these are not official, they are not included in "ALL";
1454 the "ECCdraft" ciphersuite group alias can be used to select them.
1455 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1457 *) Add ECDH engine support.
1458 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1460 *) Add ECDH in new directory crypto/ecdh/.
1461 [Douglas Stebila (Sun Microsystems Laboratories)]
1463 *) Let BN_rand_range() abort with an error after 100 iterations
1464 without success (which indicates a broken PRNG).
1467 *) Change BN_mod_sqrt() so that it verifies that the input value
1468 is really the square of the return value. (Previously,
1469 BN_mod_sqrt would show GIGO behaviour.)
1472 *) Add named elliptic curves over binary fields from X9.62, SECG,
1473 and WAP/WTLS; add OIDs that were still missing.
1475 [Sheueling Chang Shantz and Douglas Stebila
1476 (Sun Microsystems Laboratories)]
1478 *) Extend the EC library for elliptic curves over binary fields
1479 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1482 EC_GF2m_simple_method
1486 EC_GROUP_new_curve_GF2m
1487 EC_GROUP_set_curve_GF2m
1488 EC_GROUP_get_curve_GF2m
1489 EC_POINT_set_affine_coordinates_GF2m
1490 EC_POINT_get_affine_coordinates_GF2m
1491 EC_POINT_set_compressed_coordinates_GF2m
1493 Point compression for binary fields is disabled by default for
1494 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1497 As binary polynomials are represented as BIGNUMs, various members
1498 of the EC_GROUP and EC_POINT data structures can be shared
1499 between the implementations for prime fields and binary fields;
1500 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1501 are essentially identical to their ..._GFp counterparts.
1502 (For simplicity, the '..._GFp' prefix has been dropped from
1503 various internal method names.)
1505 An internal 'field_div' method (similar to 'field_mul' and
1506 'field_sqr') has been added; this is used only for binary fields.
1508 [Sheueling Chang Shantz and Douglas Stebila
1509 (Sun Microsystems Laboratories)]
1511 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1512 through methods ('mul', 'precompute_mult').
1514 The generic implementations (now internally called 'ec_wNAF_mul'
1515 and 'ec_wNAF_precomputed_mult') remain the default if these
1516 methods are undefined.
1518 [Sheueling Chang Shantz and Douglas Stebila
1519 (Sun Microsystems Laboratories)]
1521 *) New function EC_GROUP_get_degree, which is defined through
1522 EC_METHOD. For curves over prime fields, this returns the bit
1523 length of the modulus.
1525 [Sheueling Chang Shantz and Douglas Stebila
1526 (Sun Microsystems Laboratories)]
1528 *) New functions EC_GROUP_dup, EC_POINT_dup.
1529 (These simply call ..._new and ..._copy).
1531 [Sheueling Chang Shantz and Douglas Stebila
1532 (Sun Microsystems Laboratories)]
1534 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1535 Polynomials are represented as BIGNUMs (where the sign bit is not
1536 used) in the following functions [macros]:
1539 BN_GF2m_sub [= BN_GF2m_add]
1540 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1541 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1542 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1544 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1545 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1546 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1547 BN_GF2m_cmp [= BN_ucmp]
1549 (Note that only the 'mod' functions are actually for fields GF(2^m).
1550 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1552 For some functions, an the irreducible polynomial defining a
1553 field can be given as an 'unsigned int[]' with strictly
1554 decreasing elements giving the indices of those bits that are set;
1555 i.e., p[] represents the polynomial
1556 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1558 p[0] > p[1] > ... > p[k] = 0.
1559 This applies to the following functions:
1564 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1565 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1567 BN_GF2m_mod_sqrt_arr
1568 BN_GF2m_mod_solve_quad_arr
1572 Conversion can be performed by the following functions:
1577 bntest.c has additional tests for binary polynomial arithmetic.
1579 Two implementations for BN_GF2m_mod_div() are available.
1580 The default algorithm simply uses BN_GF2m_mod_inv() and
1581 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1582 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1583 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1585 [Sheueling Chang Shantz and Douglas Stebila
1586 (Sun Microsystems Laboratories)]
1588 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1589 functionality is disabled at compile-time.
1590 [Douglas Stebila <douglas.stebila@sun.com>]
1592 *) Change default behaviour of 'openssl asn1parse' so that more
1593 information is visible when viewing, e.g., a certificate:
1595 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1596 mode the content of non-printable OCTET STRINGs is output in a
1597 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1598 avoid the appearance of a printable string.
1599 [Nils Larsch <nla@trustcenter.de>]
1601 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1603 EC_GROUP_set_asn1_flag()
1604 EC_GROUP_get_asn1_flag()
1605 EC_GROUP_set_point_conversion_form()
1606 EC_GROUP_get_point_conversion_form()
1607 These control ASN1 encoding details:
1608 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1609 has been set to OPENSSL_EC_NAMED_CURVE.
1610 - Points are encoded in uncompressed form by default; options for
1611 asn1_for are as for point2oct, namely
1612 POINT_CONVERSION_COMPRESSED
1613 POINT_CONVERSION_UNCOMPRESSED
1614 POINT_CONVERSION_HYBRID
1616 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1619 EC_GROUP_get0_seed()
1620 EC_GROUP_get_seed_len()
1621 This is used only for ASN1 purposes (so far).
1622 [Nils Larsch <nla@trustcenter.de>]
1624 *) Add 'field_type' member to EC_METHOD, which holds the NID
1625 of the appropriate field type OID. The new function
1626 EC_METHOD_get_field_type() returns this value.
1627 [Nils Larsch <nla@trustcenter.de>]
1632 EC_POINT_point2hex()
1633 EC_POINT_hex2point()
1634 providing useful interfaces to EC_POINT_point2oct() and
1635 EC_POINT_oct2point().
1636 [Nils Larsch <nla@trustcenter.de>]
1638 *) Change internals of the EC library so that the functions
1639 EC_GROUP_set_generator()
1640 EC_GROUP_get_generator()
1641 EC_GROUP_get_order()
1642 EC_GROUP_get_cofactor()
1643 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1644 to methods, which would lead to unnecessary code duplication when
1645 adding different types of curves.
1646 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1648 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1649 arithmetic, and such that modified wNAFs are generated
1650 (which avoid length expansion in many cases).
1653 *) Add a function EC_GROUP_check_discriminant() (defined via
1654 EC_METHOD) that verifies that the curve discriminant is non-zero.
1656 Add a function EC_GROUP_check() that makes some sanity tests
1657 on a EC_GROUP, its generator and order. This includes
1658 EC_GROUP_check_discriminant().
1659 [Nils Larsch <nla@trustcenter.de>]
1661 *) Add ECDSA in new directory crypto/ecdsa/.
1663 Add applications 'openssl ecparam' and 'openssl ecdsa'
1664 (these are based on 'openssl dsaparam' and 'openssl dsa').
1666 ECDSA support is also included in various other files across the
1667 library. Most notably,
1668 - 'openssl req' now has a '-newkey ecdsa:file' option;
1669 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1670 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1671 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1672 them suitable for ECDSA where domain parameters must be
1673 extracted before the specific public key;
1674 - ECDSA engine support has been added.
1675 [Nils Larsch <nla@trustcenter.de>]
1677 *) Include some named elliptic curves, and add OIDs from X9.62,
1678 SECG, and WAP/WTLS. Each curve can be obtained from the new
1680 EC_GROUP_new_by_curve_name(),
1681 and the list of available named curves can be obtained with
1682 EC_get_builtin_curves().
1683 Also add a 'curve_name' member to EC_GROUP objects, which can be
1685 EC_GROUP_set_curve_name()
1686 EC_GROUP_get_curve_name()
1687 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1689 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1690 was actually never needed) and in BN_mul(). The removal in BN_mul()
1691 required a small change in bn_mul_part_recursive() and the addition
1692 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1693 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1694 bn_sub_words() and bn_add_words() except they take arrays with
1698 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1700 *) Cleanse PEM buffers before freeing them since they may contain
1702 [Benjamin Bennett <ben@psc.edu>]
1704 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1705 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1706 authentication-only ciphersuites.
1709 *) Since AES128 and AES256 share a single mask bit in the logic of
1710 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1711 kludge to work properly if AES128 is available and AES256 isn't.
1714 *) Expand security boundary to match 1.1.1 module.
1717 *) Remove redundant features: hash file source, editing of test vectors
1718 modify fipsld to use external fips_premain.c signature.
1721 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1722 run algorithm test programs.
1725 *) Make algorithm test programs more tolerant of whitespace.
1728 *) Have SSL/TLS server implementation tolerate "mismatched" record
1729 protocol version while receiving ClientHello even if the
1730 ClientHello is fragmented. (The server can't insist on the
1731 particular protocol version it has chosen before the ServerHello
1732 message has informed the client about his choice.)
1735 *) Load error codes if they are not already present instead of using a
1736 static variable. This allows them to be cleanly unloaded and reloaded.
1739 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1741 *) Introduce limits to prevent malicious keys being able to
1742 cause a denial of service. (CVE-2006-2940)
1743 [Steve Henson, Bodo Moeller]
1745 *) Fix ASN.1 parsing of certain invalid structures that can result
1746 in a denial of service. (CVE-2006-2937) [Steve Henson]
1748 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1749 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1751 *) Fix SSL client code which could crash if connecting to a
1752 malicious SSLv2 server. (CVE-2006-4343)
1753 [Tavis Ormandy and Will Drewry, Google Security Team]
1755 *) Change ciphersuite string processing so that an explicit
1756 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1757 will no longer include "AES128-SHA"), and any other similar
1758 ciphersuite (same bitmap) from *other* protocol versions (so that
1759 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1760 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1761 changes from 0.9.8b and 0.9.8d.
1764 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1766 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1767 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1769 *) Change the Unix randomness entropy gathering to use poll() when
1770 possible instead of select(), since the latter has some
1771 undesirable limitations.
1772 [Darryl Miles via Richard Levitte and Bodo Moeller]
1774 *) Disable rogue ciphersuites:
1776 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1777 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1778 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1780 The latter two were purportedly from
1781 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1784 Also deactive the remaining ciphersuites from
1785 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1786 unofficial, and the ID has long expired.
1789 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1790 dual-core machines) and other potential thread-safety issues.
1793 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1795 *) Adapt fipsld and the build system to link against the validated FIPS
1796 module in FIPS mode.
1799 *) Fixes for VC++ 2005 build under Windows.
1802 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1803 from a Windows bash shell such as MSYS. It is autodetected from the
1804 "config" script when run from a VC++ environment. Modify standard VC++
1805 build to use fipscanister.o from the GNU make build.
1808 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1810 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1811 The value now differs depending on if you build for FIPS or not.
1812 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1813 safely run with a non-FIPSed libcrypto, as it may crash because of
1814 the difference induced by this change.
1817 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1819 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1820 (part of SSL_OP_ALL). This option used to disable the
1821 countermeasure against man-in-the-middle protocol-version
1822 rollback in the SSL 2.0 server implementation, which is a bad
1823 idea. (CVE-2005-2969)
1825 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1826 for Information Security, National Institute of Advanced Industrial
1827 Science and Technology [AIST], Japan)]
1829 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1830 mainly for FIPS compliance and not fully integrated at this stage.
1833 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1834 the exponentiation using a fixed-length exponent. (Otherwise,
1835 the information leaked through timing could expose the secret key
1836 after many signatures; cf. Bleichenbacher's attack on DSA with
1840 *) Make a new fixed-window mod_exp implementation the default for
1841 RSA, DSA, and DH private-key operations so that the sequence of
1842 squares and multiplies and the memory access pattern are
1843 independent of the particular secret key. This will mitigate
1844 cache-timing and potential related attacks.
1846 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1847 and this is automatically used by BN_mod_exp_mont() if the new flag
1848 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1849 will use this BN flag for private exponents unless the flag
1850 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1851 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1853 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1855 *) Change the client implementation for SSLv23_method() and
1856 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1857 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1858 (Previously, the SSL 2.0 backwards compatible Client Hello
1859 message format would be used even with SSL_OP_NO_SSLv2.)
1862 *) Add support for smime-type MIME parameter in S/MIME messages which some
1866 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1867 a threadsafe manner. Modify rsa code to use new function and add calls
1868 to dsa and dh code (which had race conditions before).
1871 *) Include the fixed error library code in the C error file definitions
1872 instead of fixing them up at runtime. This keeps the error code
1873 structures constant.
1876 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1878 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1881 *) Fixes for newer kerberos headers. NB: the casts are needed because
1882 the 'length' field is signed on one version and unsigned on another
1883 with no (?) obvious way to tell the difference, without these VC++
1884 complains. Also the "definition" of FAR (blank) is no longer included
1885 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1886 some needed definitions.
1889 *) Undo Cygwin change.
1892 *) Added support for proxy certificates according to RFC 3820.
1893 Because they may be a security thread to unaware applications,
1894 they must be explicitely allowed in run-time. See
1895 docs/HOWTO/proxy_certificates.txt for further information.
1898 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1900 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1901 server and client random values. Previously
1902 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1903 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1905 This change has negligible security impact because:
1907 1. Server and client random values still have 24 bytes of pseudo random
1910 2. Server and client random values are sent in the clear in the initial
1913 3. The master secret is derived using the premaster secret (48 bytes in
1914 size for static RSA ciphersuites) as well as client server and random
1917 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1920 [Stephen Henson, reported by UK NISCC]
1922 *) Use Windows randomness collection on Cygwin.
1925 *) Fix hang in EGD/PRNGD query when communication socket is closed
1926 prematurely by EGD/PRNGD.
1927 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1929 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1932 *) Back-port of selected performance improvements from development
1933 branch, as well as improved support for PowerPC platforms.
1936 *) Add lots of checks for memory allocation failure, error codes to indicate
1937 failure and freeing up memory if a failure occurs.
1938 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1940 *) Add new -passin argument to dgst.
1943 *) Perform some character comparisons of different types in X509_NAME_cmp:
1944 this is needed for some certificates that reencode DNs into UTF8Strings
1945 (in violation of RFC3280) and can't or wont issue name rollover
1949 *) Make an explicit check during certificate validation to see that
1950 the CA setting in each certificate on the chain is correct. As a
1951 side effect always do the following basic checks on extensions,
1952 not just when there's an associated purpose to the check:
1954 - if there is an unhandled critical extension (unless the user
1955 has chosen to ignore this fault)
1956 - if the path length has been exceeded (if one is set at all)
1957 - that certain extensions fit the associated purpose (if one has
1961 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1963 *) Avoid a race condition when CRLs are checked in a multi threaded
1964 environment. This would happen due to the reordering of the revoked
1965 entries during signature checking and serial number lookup. Now the
1966 encoding is cached and the serial number sort performed under a lock.
1967 Add new STACK function sk_is_sorted().
1970 *) Add Delta CRL to the extension code.
1973 *) Various fixes to s3_pkt.c so alerts are sent properly.
1974 [David Holmes <d.holmes@f5.com>]
1976 *) Reduce the chances of duplicate issuer name and serial numbers (in
1977 violation of RFC3280) using the OpenSSL certificate creation utilities.
1978 This is done by creating a random 64 bit value for the initial serial
1979 number when a serial number file is created or when a self signed
1980 certificate is created using 'openssl req -x509'. The initial serial
1981 number file is created using 'openssl x509 -next_serial' in CA.pl
1982 rather than being initialized to 1.
1985 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1987 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1988 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1989 [Joe Orton, Steve Henson]
1991 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1993 [Joe Orton, Steve Henson]
1995 *) Make it possible to have multiple active certificates with the same
1996 subject in the CA index file. This is done only if the keyword
1997 'unique_subject' is set to 'no' in the main CA section (default
1998 if 'CA_default') of the configuration file. The value is saved
1999 with the database itself in a separate index attribute file,
2000 named like the index file with '.attr' appended to the name.
2003 *) X509 verify fixes. Disable broken certificate workarounds when
2004 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2005 keyUsage extension present. Don't accept CRLs with unhandled critical
2006 extensions: since verify currently doesn't process CRL extensions this
2007 rejects a CRL with *any* critical extensions. Add new verify error codes
2011 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2012 A clarification of RFC2560 will require the use of OCTET STRINGs and
2013 some implementations cannot handle the current raw format. Since OpenSSL
2014 copies and compares OCSP nonces as opaque blobs without any attempt at
2015 parsing them this should not create any compatibility issues.
2018 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2019 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2020 this HMAC (and other) operations are several times slower than OpenSSL
2024 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2025 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2027 *) Use the correct content when signing type "other".
2030 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2032 *) Fix various bugs revealed by running the NISCC test suite:
2034 Stop out of bounds reads in the ASN1 code when presented with
2035 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2037 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2039 If verify callback ignores invalid public key errors don't try to check
2040 certificate signature with the NULL public key.
2044 *) New -ignore_err option in ocsp application to stop the server
2045 exiting on the first error in a request.
2048 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2049 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2053 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2054 extra data after the compression methods not only for TLS 1.0
2055 but also for SSL 3.0 (as required by the specification).
2056 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2058 *) Change X509_certificate_type() to mark the key as exported/exportable
2059 when it's 512 *bits* long, not 512 bytes.
2062 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2063 blocks during encryption.
2066 *) Various fixes to base64 BIO and non blocking I/O. On write
2067 flushes were not handled properly if the BIO retried. On read
2068 data was not being buffered properly and had various logic bugs.
2069 This also affects blocking I/O when the data being decoded is a
2073 *) Various S/MIME bugfixes and compatibility changes:
2074 output correct application/pkcs7 MIME type if
2075 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2076 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2077 of files as .eml work). Correctly handle very long lines in MIME
2081 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2083 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2084 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2085 a protocol version number mismatch like a decryption error
2086 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2089 *) Turn on RSA blinding by default in the default implementation
2090 to avoid a timing attack. Applications that don't want it can call
2091 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2092 They would be ill-advised to do so in most cases.
2093 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2095 *) Change RSA blinding code so that it works when the PRNG is not
2096 seeded (in this case, the secret RSA exponent is abused as
2097 an unpredictable seed -- if it is not unpredictable, there
2098 is no point in blinding anyway). Make RSA blinding thread-safe
2099 by remembering the creator's thread ID in rsa->blinding and
2100 having all other threads use local one-time blinding factors
2101 (this requires more computation than sharing rsa->blinding, but
2102 avoids excessive locking; and if an RSA object is not shared
2103 between threads, blinding will still be very fast).
2106 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2107 ENGINE as defaults for all supported algorithms irrespective of
2108 the 'flags' parameter. 'flags' is now honoured, so applications
2109 should make sure they are passing it correctly.
2112 *) Target "mingw" now allows native Windows code to be generated in
2113 the Cygwin environment as well as with the MinGW compiler.
2116 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2118 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2119 via timing by performing a MAC computation even if incorrrect
2120 block cipher padding has been found. This is a countermeasure
2121 against active attacks where the attacker has to distinguish
2122 between bad padding and a MAC verification error. (CVE-2003-0078)
2124 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2125 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2126 Martin Vuagnoux (EPFL, Ilion)]
2128 *) Make the no-err option work as intended. The intention with no-err
2129 is not to have the whole error stack handling routines removed from
2130 libcrypto, it's only intended to remove all the function name and
2131 reason texts, thereby removing some of the footprint that may not
2132 be interesting if those errors aren't displayed anyway.
2134 NOTE: it's still possible for any application or module to have it's
2135 own set of error texts inserted. The routines are there, just not
2136 used by default when no-err is given.
2139 *) Add support for FreeBSD on IA64.
2140 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2142 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2143 Kerberos function mit_des_cbc_cksum(). Before this change,
2144 the value returned by DES_cbc_cksum() was like the one from
2145 mit_des_cbc_cksum(), except the bytes were swapped.
2146 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2148 *) Allow an application to disable the automatic SSL chain building.
2149 Before this a rather primitive chain build was always performed in
2150 ssl3_output_cert_chain(): an application had no way to send the
2151 correct chain if the automatic operation produced an incorrect result.
2153 Now the chain builder is disabled if either:
2155 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2157 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2159 The reasoning behind this is that an application would not want the
2160 auto chain building to take place if extra chain certificates are
2161 present and it might also want a means of sending no additional
2162 certificates (for example the chain has two certificates and the
2166 *) Add the possibility to build without the ENGINE framework.
2167 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2169 *) Under Win32 gmtime() can return NULL: check return value in
2170 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2173 *) DSA routines: under certain error conditions uninitialized BN objects
2174 could be freed. Solution: make sure initialization is performed early
2175 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2176 Nils Larsch <nla@trustcenter.de> via PR#459)
2179 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2180 checked on reconnect on the client side, therefore session resumption
2181 could still fail with a "ssl session id is different" error. This
2182 behaviour is masked when SSL_OP_ALL is used due to
2183 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2184 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2185 followup to PR #377.
2188 *) IA-32 assembler support enhancements: unified ELF targets, support
2189 for SCO/Caldera platforms, fix for Cygwin shared build.
2192 *) Add support for FreeBSD on sparc64. As a consequence, support for
2193 FreeBSD on non-x86 processors is separate from x86 processors on
2194 the config script, much like the NetBSD support.
2195 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2197 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2199 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2202 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2203 code (06) was taken as the first octet of the session ID and the last
2204 octet was ignored consequently. As a result SSLv2 client side session
2205 caching could not have worked due to the session ID mismatch between
2207 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2211 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2212 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2216 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2217 seems that in spite of existing for more than a year, many application
2218 author have done nothing to provide the necessary callbacks, which
2219 means that this particular engine will not work properly anywhere.
2220 This is a very unfortunate situation which forces us, in the name
2221 of usability, to give the hw_ncipher.c a static lock, which is part
2223 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2224 appear in 0.9.8 or later. We EXPECT application authors to have
2225 dealt properly with this when 0.9.8 is released (unless we actually
2226 make such changes in the libcrypto locking code that changes will
2227 have to be made anyway).
2230 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2231 octets have been read, EOF or an error occurs. Without this change
2232 some truncated ASN1 structures will not produce an error.
2235 *) Disable Heimdal support, since it hasn't been fully implemented.
2236 Still give the possibility to force the use of Heimdal, but with
2237 warnings and a request that patches get sent to openssl-dev.
2240 *) Add the VC-CE target, introduce the WINCE sysname, and add
2241 INSTALL.WCE and appropriate conditionals to make it build.
2242 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2244 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2245 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2246 edit numbers of the version.
2247 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2249 *) Introduce safe string copy and catenation functions
2250 (BUF_strlcpy() and BUF_strlcat()).
2251 [Ben Laurie (CHATS) and Richard Levitte]
2253 *) Avoid using fixed-size buffers for one-line DNs.
2254 [Ben Laurie (CHATS)]
2256 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2257 resizing buffers containing secrets, and use where appropriate.
2258 [Ben Laurie (CHATS)]
2260 *) Avoid using fixed size buffers for configuration file location.
2261 [Ben Laurie (CHATS)]
2263 *) Avoid filename truncation for various CA files.
2264 [Ben Laurie (CHATS)]
2266 *) Use sizeof in preference to magic numbers.
2267 [Ben Laurie (CHATS)]
2269 *) Avoid filename truncation in cert requests.
2270 [Ben Laurie (CHATS)]
2272 *) Add assertions to check for (supposedly impossible) buffer
2274 [Ben Laurie (CHATS)]
2276 *) Don't cache truncated DNS entries in the local cache (this could
2277 potentially lead to a spoofing attack).
2278 [Ben Laurie (CHATS)]
2280 *) Fix various buffers to be large enough for hex/decimal
2281 representations in a platform independent manner.
2282 [Ben Laurie (CHATS)]
2284 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2285 resizing buffers containing secrets, and use where appropriate.
2286 [Ben Laurie (CHATS)]
2288 *) Add BIO_indent() to avoid much slightly worrying code to do
2290 [Ben Laurie (CHATS)]
2292 *) Convert sprintf()/BIO_puts() to BIO_printf().
2293 [Ben Laurie (CHATS)]
2295 *) buffer_gets() could terminate with the buffer only half
2297 [Ben Laurie (CHATS)]
2299 *) Add assertions to prevent user-supplied crypto functions from
2300 overflowing internal buffers by having large block sizes, etc.
2301 [Ben Laurie (CHATS)]
2303 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2305 [Ben Laurie (CHATS)]
2307 *) Eliminate unused copy of key in RC4.
2308 [Ben Laurie (CHATS)]
2310 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2311 [Ben Laurie (CHATS)]
2313 *) Fix off-by-one error in EGD path.
2314 [Ben Laurie (CHATS)]
2316 *) If RANDFILE path is too long, ignore instead of truncating.
2317 [Ben Laurie (CHATS)]
2319 *) Eliminate unused and incorrectly sized X.509 structure
2321 [Ben Laurie (CHATS)]
2323 *) Eliminate unused and dangerous function knumber().
2324 [Ben Laurie (CHATS)]
2326 *) Eliminate unused and dangerous structure, KSSL_ERR.
2327 [Ben Laurie (CHATS)]
2329 *) Protect against overlong session ID context length in an encoded
2330 session object. Since these are local, this does not appear to be
2332 [Ben Laurie (CHATS)]
2334 *) Change from security patch (see 0.9.6e below) that did not affect
2335 the 0.9.6 release series:
2337 Remote buffer overflow in SSL3 protocol - an attacker could
2338 supply an oversized master key in Kerberos-enabled versions.
2340 [Ben Laurie (CHATS)]
2342 *) Change the SSL kerb5 codes to match RFC 2712.
2345 *) Make -nameopt work fully for req and add -reqopt switch.
2346 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2348 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2349 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2351 *) Make sure tests can be performed even if the corresponding algorithms
2352 have been removed entirely. This was also the last step to make
2353 OpenSSL compilable with DJGPP under all reasonable conditions.
2354 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2356 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2357 to allow version independent disabling of normally unselected ciphers,
2358 which may be activated as a side-effect of selecting a single cipher.
2360 (E.g., cipher list string "RSA" enables ciphersuites that are left
2361 out of "ALL" because they do not provide symmetric encryption.
2362 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2363 [Lutz Jaenicke, Bodo Moeller]
2365 *) Add appropriate support for separate platform-dependent build
2366 directories. The recommended way to make a platform-dependent
2367 build directory is the following (tested on Linux), maybe with
2370 # Place yourself outside of the OpenSSL source tree. In
2371 # this example, the environment variable OPENSSL_SOURCE
2372 # is assumed to contain the absolute OpenSSL source directory.
2373 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2374 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2375 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2376 mkdir -p `dirname $F`
2377 ln -s $OPENSSL_SOURCE/$F $F
2380 To be absolutely sure not to disturb the source tree, a "make clean"
2381 is a good thing. If it isn't successfull, don't worry about it,
2382 it probably means the source directory is very clean.
2385 *) Make sure any ENGINE control commands make local copies of string
2386 pointers passed to them whenever necessary. Otherwise it is possible
2387 the caller may have overwritten (or deallocated) the original string
2388 data when a later ENGINE operation tries to use the stored values.
2389 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2391 *) Improve diagnostics in file reading and command-line digests.
2392 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2394 *) Add AES modes CFB and OFB to the object database. Correct an
2395 error in AES-CFB decryption.
2398 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2399 allows existing EVP_CIPHER_CTX structures to be reused after
2400 calling EVP_*Final(). This behaviour is used by encryption
2401 BIOs and some applications. This has the side effect that
2402 applications must explicitly clean up cipher contexts with
2403 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2406 *) Check the values of dna and dnb in bn_mul_recursive before calling
2407 bn_mul_comba (a non zero value means the a or b arrays do not contain
2408 n2 elements) and fallback to bn_mul_normal if either is not zero.
2411 *) Fix escaping of non-ASCII characters when using the -subj option
2412 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2415 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2416 form for "surname", serialNumber has no short form.
2417 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2418 therefore remove "mail" short name for "internet 7".
2419 The OID for unique identifiers in X509 certificates is
2420 x500UniqueIdentifier, not uniqueIdentifier.
2421 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2424 *) Add an "init" command to the ENGINE config module and auto initialize
2425 ENGINEs. Without any "init" command the ENGINE will be initialized
2426 after all ctrl commands have been executed on it. If init=1 the
2427 ENGINE is initailized at that point (ctrls before that point are run
2428 on the uninitialized ENGINE and after on the initialized one). If
2429 init=0 then the ENGINE will not be iniatialized at all.
2432 *) Fix the 'app_verify_callback' interface so that the user-defined
2433 argument is actually passed to the callback: In the
2434 SSL_CTX_set_cert_verify_callback() prototype, the callback
2435 declaration has been changed from
2438 int (*cb)(X509_STORE_CTX *,void *);
2439 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2440 i=s->ctx->app_verify_callback(&ctx)
2441 has been changed into
2442 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2444 To update applications using SSL_CTX_set_cert_verify_callback(),
2445 a dummy argument can be added to their callback functions.
2446 [D. K. Smetters <smetters@parc.xerox.com>]
2448 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2449 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2451 *) Add and OPENSSL_LOAD_CONF define which will cause
2452 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2453 This allows older applications to transparently support certain
2454 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2455 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2456 load the config file and OPENSSL_add_all_algorithms_conf() which will
2457 always load it have also been added.
2460 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2461 Adjust NIDs and EVP layer.
2462 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2464 *) Config modules support in openssl utility.
2466 Most commands now load modules from the config file,
2467 though in a few (such as version) this isn't done
2468 because it couldn't be used for anything.
2470 In the case of ca and req the config file used is
2471 the same as the utility itself: that is the -config
2472 command line option can be used to specify an
2476 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2477 use "openssl_conf" if filename is NULL use default openssl config file.
2480 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2481 config section name. Add a new flag to tolerate a missing config file
2482 and move code to CONF_modules_load_file().
2485 *) Support for crypto accelerator cards from Accelerated Encryption
2486 Processing, www.aep.ie. (Use engine 'aep')
2487 The support was copied from 0.9.6c [engine] and adapted/corrected
2488 to work with the new engine framework.
2489 [AEP Inc. and Richard Levitte]
2491 *) Support for SureWare crypto accelerator cards from Baltimore
2492 Technologies. (Use engine 'sureware')
2493 The support was copied from 0.9.6c [engine] and adapted
2494 to work with the new engine framework.
2497 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2498 make the newer ENGINE framework commands for the CHIL engine work.
2499 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2501 *) Make it possible to produce shared libraries on ReliantUNIX.
2502 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2504 *) Add the configuration target debug-linux-ppro.
2505 Make 'openssl rsa' use the general key loading routines
2506 implemented in apps.c, and make those routines able to
2507 handle the key format FORMAT_NETSCAPE and the variant
2509 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2511 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2512 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2514 *) Add -keyform to rsautl, and document -engine.
2515 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2517 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2518 BIO_R_NO_SUCH_FILE error code rather than the generic
2519 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2522 *) Add new functions
2524 ERR_peek_last_error_line
2525 ERR_peek_last_error_line_data.
2526 These are similar to
2529 ERR_peek_error_line_data,
2530 but report on the latest error recorded rather than the first one
2531 still in the error queue.
2532 [Ben Laurie, Bodo Moeller]
2534 *) default_algorithms option in ENGINE config module. This allows things
2536 default_algorithms = ALL
2537 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2540 *) Prelminary ENGINE config module.
2543 *) New experimental application configuration code.
2546 *) Change the AES code to follow the same name structure as all other
2547 symmetric ciphers, and behave the same way. Move everything to
2548 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2549 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2551 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2552 [Ben Laurie and Theo de Raadt]
2554 *) Add option to output public keys in req command.
2555 [Massimiliano Pala madwolf@openca.org]
2557 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2558 (up to about 10% better than before for P-192 and P-224).
2561 *) New functions/macros
2563 SSL_CTX_set_msg_callback(ctx, cb)
2564 SSL_CTX_set_msg_callback_arg(ctx, arg)
2565 SSL_set_msg_callback(ssl, cb)
2566 SSL_set_msg_callback_arg(ssl, arg)
2568 to request calling a callback function
2570 void cb(int write_p, int version, int content_type,
2571 const void *buf, size_t len, SSL *ssl, void *arg)
2573 whenever a protocol message has been completely received
2574 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2575 protocol version according to which the SSL library interprets
2576 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2577 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2578 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2579 specification (change_cipher_spec(20), alert(21), handshake(22)).
2580 'buf' and 'len' point to the actual message, 'ssl' to the
2581 SSL object, and 'arg' is the application-defined value set by
2582 SSL[_CTX]_set_msg_callback_arg().
2584 'openssl s_client' and 'openssl s_server' have new '-msg' options
2585 to enable a callback that displays all protocol messages.
2588 *) Change the shared library support so shared libraries are built as
2589 soon as the corresponding static library is finished, and thereby get
2590 openssl and the test programs linked against the shared library.
2591 This still only happens when the keyword "shard" has been given to
2592 the configuration scripts.
2594 NOTE: shared library support is still an experimental thing, and
2595 backward binary compatibility is still not guaranteed.
2596 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2598 *) Add support for Subject Information Access extension.
2599 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2601 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2602 additional bytes when new memory had to be allocated, not just
2603 when reusing an existing buffer.
2606 *) New command line and configuration option 'utf8' for the req command.
2607 This allows field values to be specified as UTF8 strings.
2610 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2611 runs for the former and machine-readable output for the latter.
2614 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2615 of the e-mail address in the DN (i.e., it will go into a certificate
2616 extension only). The new configuration file option 'email_in_dn = no'
2617 has the same effect.
2618 [Massimiliano Pala madwolf@openca.org]
2620 *) Change all functions with names starting with des_ to be starting
2621 with DES_ instead. Add wrappers that are compatible with libdes,
2622 but are named _ossl_old_des_*. Finally, add macros that map the
2623 des_* symbols to the corresponding _ossl_old_des_* if libdes
2624 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2625 desired, the des_* symbols will be mapped to DES_*, with one
2628 Since we provide two compatibility mappings, the user needs to
2629 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2630 compatibility is desired. The default (i.e., when that macro
2631 isn't defined) is OpenSSL 0.9.6c compatibility.
2633 There are also macros that enable and disable the support of old
2634 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2635 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2636 are defined, the default will apply: to support the old des routines.
2638 In either case, one must include openssl/des.h to get the correct
2639 definitions. Do not try to just include openssl/des_old.h, that
2642 NOTE: This is a major break of an old API into a new one. Software
2643 authors are encouraged to switch to the DES_ style functions. Some
2644 time in the future, des_old.h and the libdes compatibility functions
2645 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2646 default), and then completely removed.
2649 *) Test for certificates which contain unsupported critical extensions.
2650 If such a certificate is found during a verify operation it is
2651 rejected by default: this behaviour can be overridden by either
2652 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2653 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2654 X509_supported_extension() has also been added which returns 1 if a
2655 particular extension is supported.
2658 *) Modify the behaviour of EVP cipher functions in similar way to digests
2659 to retain compatibility with existing code.
2662 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2663 compatibility with existing code. In particular the 'ctx' parameter does
2664 not have to be to be initialized before the call to EVP_DigestInit() and
2665 it is tidied up after a call to EVP_DigestFinal(). New function
2666 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2667 EVP_MD_CTX_copy() changed to not require the destination to be
2668 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2669 requires the destination to be valid.
2671 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2672 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2675 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2676 so that complete 'Handshake' protocol structures are kept in memory
2677 instead of overwriting 'msg_type' and 'length' with 'body' data.
2680 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2681 [Massimo Santin via Richard Levitte]
2683 *) Major restructuring to the underlying ENGINE code. This includes
2684 reduction of linker bloat, separation of pure "ENGINE" manipulation
2685 (initialisation, etc) from functionality dealing with implementations
2686 of specific crypto iterfaces. This change also introduces integrated
2687 support for symmetric ciphers and digest implementations - so ENGINEs
2688 can now accelerate these by providing EVP_CIPHER and EVP_MD
2689 implementations of their own. This is detailed in crypto/engine/README
2690 as it couldn't be adequately described here. However, there are a few
2691 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2692 were changed in the original introduction of ENGINE code have now
2693 reverted back - the hooking from this code to ENGINE is now a good
2694 deal more passive and at run-time, operations deal directly with
2695 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2696 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2697 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2698 they were not being used by the framework as there is no concept of a
2699 BIGNUM_METHOD and they could not be generalised to the new
2700 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2701 ENGINE_cpy() has been removed as it cannot be consistently defined in
2705 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2708 *) Change mkdef.pl to sort symbols that get the same entry number,
2709 and make sure the automatically generated functions ERR_load_*
2710 become part of libeay.num as well.
2713 *) New function SSL_renegotiate_pending(). This returns true once
2714 renegotiation has been requested (either SSL_renegotiate() call
2715 or HelloRequest/ClientHello receveived from the peer) and becomes
2716 false once a handshake has been completed.
2717 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2718 sends a HelloRequest, but does not ensure that a handshake takes
2719 place. SSL_renegotiate_pending() is useful for checking if the
2720 client has followed the request.)
2723 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2724 By default, clients may request session resumption even during
2725 renegotiation (if session ID contexts permit); with this option,
2726 session resumption is possible only in the first handshake.
2728 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2729 more bits available for options that should not be part of
2730 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2733 *) Add some demos for certificate and certificate request creation.
2736 *) Make maximum certificate chain size accepted from the peer application
2737 settable (SSL*_get/set_max_cert_list()), as proposed by
2738 "Douglas E. Engert" <deengert@anl.gov>.
2741 *) Add support for shared libraries for Unixware-7
2742 (Boyd Lynn Gerber <gerberb@zenez.com>).
2745 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2746 be done prior to destruction. Use this to unload error strings from
2747 ENGINEs that load their own error strings. NB: This adds two new API
2748 functions to "get" and "set" this destroy handler in an ENGINE.
2751 *) Alter all existing ENGINE implementations (except "openssl" and
2752 "openbsd") to dynamically instantiate their own error strings. This
2753 makes them more flexible to be built both as statically-linked ENGINEs
2754 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2755 Also, add stub code to each that makes building them as self-contained
2756 shared-libraries easier (see README.ENGINE).
2759 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2760 implementations into applications that are completely implemented in
2761 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2762 commands that can be used to configure what shared-library to load and
2763 to control aspects of the way it is handled. Also, made an update to
2764 the README.ENGINE file that brings its information up-to-date and
2765 provides some information and instructions on the "dynamic" ENGINE
2766 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2769 *) Make it possible to unload ranges of ERR strings with a new
2770 "ERR_unload_strings" function.
2773 *) Add a copy() function to EVP_MD.
2776 *) Make EVP_MD routines take a context pointer instead of just the
2777 md_data void pointer.
2780 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2781 that the digest can only process a single chunk of data
2782 (typically because it is provided by a piece of
2783 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2784 is only going to provide a single chunk of data, and hence the
2785 framework needn't accumulate the data for oneshot drivers.
2788 *) As with "ERR", make it possible to replace the underlying "ex_data"
2789 functions. This change also alters the storage and management of global
2790 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2791 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2792 index counters. The API functions that use this state have been changed
2793 to take a "class_index" rather than pointers to the class's local STACK
2794 and counter, and there is now an API function to dynamically create new
2795 classes. This centralisation allows us to (a) plug a lot of the
2796 thread-safety problems that existed, and (b) makes it possible to clean
2797 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2798 such data would previously have always leaked in application code and
2799 workarounds were in place to make the memory debugging turn a blind eye
2800 to it. Application code that doesn't use this new function will still
2801 leak as before, but their memory debugging output will announce it now
2802 rather than letting it slide.
2804 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2805 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2806 has a return value to indicate success or failure.
2809 *) Make it possible to replace the underlying "ERR" functions such that the
2810 global state (2 LHASH tables and 2 locks) is only used by the "default"
2811 implementation. This change also adds two functions to "get" and "set"
2812 the implementation prior to it being automatically set the first time
2813 any other ERR function takes place. Ie. an application can call "get",
2814 pass the return value to a module it has just loaded, and that module
2815 can call its own "set" function using that value. This means the
2816 module's "ERR" operations will use (and modify) the error state in the
2817 application and not in its own statically linked copy of OpenSSL code.
2820 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2821 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2822 the operation, and provides a more encapsulated way for external code
2823 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2824 to use these functions rather than manually incrementing the counts.
2826 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2829 *) Add EVP test program.
2832 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2835 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2836 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2837 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2838 These allow a CRL to be built without having to access X509_CRL fields
2839 directly. Modify 'ca' application to use new functions.
2842 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2843 bug workarounds. Rollback attack detection is a security feature.
2844 The problem will only arise on OpenSSL servers when TLSv1 is not
2845 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2846 Software authors not wanting to support TLSv1 will have special reasons
2847 for their choice and can explicitly enable this option.
2848 [Bodo Moeller, Lutz Jaenicke]
2850 *) Rationalise EVP so it can be extended: don't include a union of
2851 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2852 (similar to those existing for EVP_CIPHER_CTX).
2857 EVP_MD_CTX_init(&md); /* new function call */
2858 EVP_DigestInit(&md, EVP_sha1());
2859 EVP_DigestUpdate(&md, in, len);
2860 EVP_DigestFinal(&md, out, NULL);
2861 EVP_MD_CTX_cleanup(&md); /* new function call */
2865 *) Make DES key schedule conform to the usual scheme, as well as
2866 correcting its structure. This means that calls to DES functions
2867 now have to pass a pointer to a des_key_schedule instead of a
2868 plain des_key_schedule (which was actually always a pointer
2871 des_key_schedule ks;
2873 des_set_key_checked(..., &ks);
2874 des_ncbc_encrypt(..., &ks, ...);
2876 (Note that a later change renames 'des_...' into 'DES_...'.)
2879 *) Initial reduction of linker bloat: the use of some functions, such as
2880 PEM causes large amounts of unused functions to be linked in due to
2881 poor organisation. For example pem_all.c contains every PEM function
2882 which has a knock on effect of linking in large amounts of (unused)
2883 ASN1 code. Grouping together similar functions and splitting unrelated
2884 functions prevents this.
2887 *) Cleanup of EVP macros.
2890 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2891 correct _ecb suffix.
2894 *) Add initial OCSP responder support to ocsp application. The
2895 revocation information is handled using the text based index
2896 use by the ca application. The responder can either handle
2897 requests generated internally, supplied in files (for example
2898 via a CGI script) or using an internal minimal server.
2901 *) Add configuration choices to get zlib compression for TLS.
2904 *) Changes to Kerberos SSL for RFC 2712 compliance:
2905 1. Implemented real KerberosWrapper, instead of just using
2906 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2907 2. Implemented optional authenticator field of KerberosWrapper.
2909 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2910 and authenticator structs; see crypto/krb5/.
2912 Generalized Kerberos calls to support multiple Kerberos libraries.
2913 [Vern Staats <staatsvr@asc.hpc.mil>,
2914 Jeffrey Altman <jaltman@columbia.edu>
2915 via Richard Levitte]
2917 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2918 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2919 values for each of the key sizes rather than having just
2920 parameters (and 'speed' generating keys each time).
2923 *) Speed up EVP routines.
2926 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2927 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2928 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2929 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2931 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2932 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2933 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2936 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2938 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2941 *) Added the OS2-EMX target.
2942 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2944 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2945 to support NCONF routines in extension code. New function CONF_set_nconf()
2946 to allow functions which take an NCONF to also handle the old LHASH
2947 structure: this means that the old CONF compatible routines can be
2948 retained (in particular wrt extensions) without having to duplicate the
2949 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2952 *) Enhance the general user interface with mechanisms for inner control
2953 and with possibilities to have yes/no kind of prompts.
2956 *) Change all calls to low level digest routines in the library and
2957 applications to use EVP. Add missing calls to HMAC_cleanup() and
2958 don't assume HMAC_CTX can be copied using memcpy().
2959 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2961 *) Add the possibility to control engines through control names but with
2962 arbitrary arguments instead of just a string.
2963 Change the key loaders to take a UI_METHOD instead of a callback
2964 function pointer. NOTE: this breaks binary compatibility with earlier
2965 versions of OpenSSL [engine].
2966 Adapt the nCipher code for these new conditions and add a card insertion
2970 *) Enhance the general user interface with mechanisms to better support
2971 dialog box interfaces, application-defined prompts, the possibility
2972 to use defaults (for example default passwords from somewhere else)
2973 and interrupts/cancellations.
2976 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2977 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2980 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2981 tidy up some unnecessarily weird code in 'sk_new()').
2982 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2984 *) Change the key loading routines for ENGINEs to use the same kind
2985 callback (pem_password_cb) as all other routines that need this
2989 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2990 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2991 than this minimum value is recommended.
2994 *) New random seeder for OpenVMS, using the system process statistics
2995 that are easily reachable.
2998 *) Windows apparently can't transparently handle global
2999 variables defined in DLLs. Initialisations such as:
3001 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3003 wont compile. This is used by the any applications that need to
3004 declare their own ASN1 modules. This was fixed by adding the option
3005 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3006 needed for static libraries under Win32.
3009 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3010 setting of purpose and trust fields. New X509_STORE trust and
3011 purpose functions and tidy up setting in other SSL functions.
3014 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3015 structure. These are inherited by X509_STORE_CTX when it is
3016 initialised. This allows various defaults to be set in the
3017 X509_STORE structure (such as flags for CRL checking and custom
3018 purpose or trust settings) for functions which only use X509_STORE_CTX
3019 internally such as S/MIME.
3021 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3022 trust settings if they are not set in X509_STORE. This allows X509_STORE
3023 purposes and trust (in S/MIME for example) to override any set by default.
3025 Add command line options for CRL checking to smime, s_client and s_server
3029 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3030 are set then the CRL is looked up in the X509_STORE structure and
3031 its validity and signature checked, then if the certificate is found
3032 in the CRL the verify fails with a revoked error.
3034 Various new CRL related callbacks added to X509_STORE_CTX structure.
3036 Command line options added to 'verify' application to support this.
3038 This needs some additional work, such as being able to handle multiple
3039 CRLs with different times, extension based lookup (rather than just
3040 by subject name) and ultimately more complete V2 CRL extension
3044 *) Add a general user interface API (crypto/ui/). This is designed
3045 to replace things like des_read_password and friends (backward
3046 compatibility functions using this new API are provided).
3047 The purpose is to remove prompting functions from the DES code
3048 section as well as provide for prompting through dialog boxes in
3049 a window system and the like.
3052 *) Add "ex_data" support to ENGINE so implementations can add state at a
3053 per-structure level rather than having to store it globally.
3056 *) Make it possible for ENGINE structures to be copied when retrieved by
3057 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3058 This causes the "original" ENGINE structure to act like a template,
3059 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3060 operational state can be localised to each ENGINE structure, despite the
3061 fact they all share the same "methods". New ENGINE structures returned in
3062 this case have no functional references and the return value is the single
3063 structural reference. This matches the single structural reference returned
3064 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3068 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3069 needs to match any other type at all we need to manually clear the
3073 *) Changes to the "openssl engine" utility to include;
3074 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3075 about an ENGINE's available control commands.
3076 - executing control commands from command line arguments using the
3077 '-pre' and '-post' switches. '-post' is only used if '-t' is
3078 specified and the ENGINE is successfully initialised. The syntax for
3079 the individual commands are colon-separated, for example;
3080 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3083 *) New dynamic control command support for ENGINEs. ENGINEs can now
3084 declare their own commands (numbers), names (strings), descriptions,
3085 and input types for run-time discovery by calling applications. A
3086 subset of these commands are implicitly classed as "executable"
3087 depending on their input type, and only these can be invoked through
3088 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3089 can be based on user input, config files, etc). The distinction is
3090 that "executable" commands cannot return anything other than a boolean
3091 result and can only support numeric or string input, whereas some
3092 discoverable commands may only be for direct use through
3093 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3094 pointers, or other custom uses. The "executable" commands are to
3095 support parameterisations of ENGINE behaviour that can be
3096 unambiguously defined by ENGINEs and used consistently across any
3097 OpenSSL-based application. Commands have been added to all the
3098 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3099 control over shared-library paths without source code alterations.
3102 *) Changed all ENGINE implementations to dynamically allocate their
3103 ENGINEs rather than declaring them statically. Apart from this being
3104 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3105 this also allows the implementations to compile without using the
3106 internal engine_int.h header.
3109 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3110 'const' value. Any code that should be able to modify a RAND_METHOD
3111 should already have non-const pointers to it (ie. they should only
3112 modify their own ones).
3115 *) Made a variety of little tweaks to the ENGINE code.
3116 - "atalla" and "ubsec" string definitions were moved from header files
3117 to C code. "nuron" string definitions were placed in variables
3118 rather than hard-coded - allowing parameterisation of these values
3119 later on via ctrl() commands.
3120 - Removed unused "#if 0"'d code.
3121 - Fixed engine list iteration code so it uses ENGINE_free() to release
3122 structural references.
3123 - Constified the RAND_METHOD element of ENGINE structures.
3124 - Constified various get/set functions as appropriate and added
3125 missing functions (including a catch-all ENGINE_cpy that duplicates
3126 all ENGINE values onto a new ENGINE except reference counts/state).
3127 - Removed NULL parameter checks in get/set functions. Setting a method
3128 or function to NULL is a way of cancelling out a previously set
3129 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3130 and doesn't justify the extra error symbols and code.
3131 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3132 flags from engine_int.h to engine.h.
3133 - Changed prototypes for ENGINE handler functions (init(), finish(),
3134 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3137 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3138 to the algorithm using long division. The binary algorithm can be
3139 used only if the modulus is odd. On 32-bit systems, it is faster
3140 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3141 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3142 up to 450 bits. In 64-bit environments, the binary algorithm
3143 appears to be advantageous for much longer moduli; here we use it
3144 for moduli up to 2048 bits.
3147 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3148 could not support the combine flag in choice fields.
3151 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3152 extensions from a certificate request to the certificate.
3155 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3156 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3157 file: this allows the display of the certificate about to be
3158 signed to be customised, to allow certain fields to be included
3159 or excluded and extension details. The old system didn't display
3160 multicharacter strings properly, omitted fields not in the policy
3161 and couldn't display additional details such as extensions.
3164 *) Function EC_POINTs_mul for multiple scalar multiplication
3165 of an arbitrary number of elliptic curve points
3166 \sum scalars[i]*points[i],
3167 optionally including the generator defined for the EC_GROUP:
3168 scalar*generator + \sum scalars[i]*points[i].
3170 EC_POINT_mul is a simple wrapper function for the typical case
3171 that the point list has just one item (besides the optional
3175 *) First EC_METHODs for curves over GF(p):
3177 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3178 operations and provides various method functions that can also
3179 operate with faster implementations of modular arithmetic.
3181 EC_GFp_mont_method() reuses most functions that are part of
3182 EC_GFp_simple_method, but uses Montgomery arithmetic.
3184 [Bodo Moeller; point addition and point doubling
3185 implementation directly derived from source code provided by
3186 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3188 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3189 crypto/ec/ec_lib.c):
3191 Curves are EC_GROUP objects (with an optional group generator)
3192 based on EC_METHODs that are built into the library.
3194 Points are EC_POINT objects based on EC_GROUP objects.
3196 Most of the framework would be able to handle curves over arbitrary
3197 finite fields, but as there are no obvious types for fields other
3198 than GF(p), some functions are limited to that for now.
3201 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3202 that the file contains a complete HTTP response.
3205 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3206 change the def and num file printf format specifier from "%-40sXXX"
3207 to "%-39s XXX". The latter will always guarantee a space after the
3208 field while the former will cause them to run together if the field
3209 is 40 of more characters long.
3212 *) Constify the cipher and digest 'method' functions and structures
3213 and modify related functions to take constant EVP_MD and EVP_CIPHER
3217 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3218 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3221 *) Modify EVP_Digest*() routines so they now return values. Although the
3222 internal software routines can never fail additional hardware versions
3226 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3228 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3229 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3232 ERR_R_NESTED_ASN1_ERROR
3234 ERR_R_MISSING_ASN1_EOS
3235 were 4 .. 9, conflicting with
3236 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3238 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3239 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3241 Add new error code 'ERR_R_INTERNAL_ERROR'.
3244 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3248 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3249 sets the subject name for a new request or supersedes the
3250 subject name in a given request. Formats that can be parsed are
3251 'CN=Some Name, OU=myOU, C=IT'
3253 'CN=Some Name/OU=myOU/C=IT'.
3255 Add options '-batch' and '-verbose' to 'openssl req'.
3256 [Massimiliano Pala <madwolf@hackmasters.net>]
3258 *) Introduce the possibility to access global variables through
3259 functions on platform were that's the best way to handle exporting
3260 global variables in shared libraries. To enable this functionality,
3261 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3262 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3263 is normally done by Configure or something similar).
3265 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3266 in the source file (foo.c) like this:
3268 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3269 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3271 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3272 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3274 OPENSSL_DECLARE_GLOBAL(int,foo);
3275 #define foo OPENSSL_GLOBAL_REF(foo)
3276 OPENSSL_DECLARE_GLOBAL(double,bar);
3277 #define bar OPENSSL_GLOBAL_REF(bar)
3279 The #defines are very important, and therefore so is including the
3280 header file everywhere where the defined globals are used.
3282 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3283 of ASN.1 items, but that structure is a bit different.
3285 The largest change is in util/mkdef.pl which has been enhanced with
3286 better and easier to understand logic to choose which symbols should
3287 go into the Windows .def files as well as a number of fixes and code
3288 cleanup (among others, algorithm keywords are now sorted
3289 lexicographically to avoid constant rewrites).
3292 *) In BN_div() keep a copy of the sign of 'num' before writing the
3293 result to 'rm' because if rm==num the value will be overwritten
3294 and produce the wrong result if 'num' is negative: this caused
3295 problems with BN_mod() and BN_nnmod().
3298 *) Function OCSP_request_verify(). This checks the signature on an
3299 OCSP request and verifies the signer certificate. The signer
3300 certificate is just checked for a generic purpose and OCSP request
3304 *) Add OCSP_check_validity() function to check the validity of OCSP
3305 responses. OCSP responses are prepared in real time and may only
3306 be a few seconds old. Simply checking that the current time lies
3307 between thisUpdate and nextUpdate max reject otherwise valid responses
3308 caused by either OCSP responder or client clock inaccuracy. Instead
3309 we allow thisUpdate and nextUpdate to fall within a certain period of
3310 the current time. The age of the response can also optionally be
3311 checked. Two new options -validity_period and -status_age added to
3315 *) If signature or public key algorithm is unrecognized print out its
3316 OID rather that just UNKNOWN.
3319 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3320 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3321 ID to be generated from the issuer certificate alone which can then be
3322 passed to OCSP_id_issuer_cmp().
3325 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3326 ASN1 modules to export functions returning ASN1_ITEM pointers
3327 instead of the ASN1_ITEM structures themselves. This adds several
3328 new macros which allow the underlying ASN1 function/structure to
3329 be accessed transparently. As a result code should not use ASN1_ITEM
3330 references directly (such as &X509_it) but instead use the relevant
3331 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3332 use of the new ASN1 code on platforms where exporting structures
3333 is problematical (for example in shared libraries) but exporting
3334 functions returning pointers to structures is not.
3337 *) Add support for overriding the generation of SSL/TLS session IDs.
3338 These callbacks can be registered either in an SSL_CTX or per SSL.
3339 The purpose of this is to allow applications to control, if they wish,
3340 the arbitrary values chosen for use as session IDs, particularly as it
3341 can be useful for session caching in multiple-server environments. A
3342 command-line switch for testing this (and any client code that wishes
3343 to use such a feature) has been added to "s_server".
3344 [Geoff Thorpe, Lutz Jaenicke]
3346 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3347 of the form '#if defined(...) || defined(...) || ...' and
3348 '#if !defined(...) && !defined(...) && ...'. This also avoids
3349 the growing number of special cases it was previously handling.
3352 *) Make all configuration macros available for application by making
3353 sure they are available in opensslconf.h, by giving them names starting
3354 with "OPENSSL_" to avoid conflicts with other packages and by making
3355 sure e_os2.h will cover all platform-specific cases together with
3357 Additionally, it is now possible to define configuration/platform-
3358 specific names (called "system identities"). In the C code, these
3359 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3360 macro with the name beginning with "OPENSSL_SYS_", which is determined
3361 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3365 *) New option -set_serial to 'req' and 'x509' this allows the serial
3366 number to use to be specified on the command line. Previously self
3367 signed certificates were hard coded with serial number 0 and the
3368 CA options of 'x509' had to use a serial number in a file which was
3372 *) New options to 'ca' utility to support V2 CRL entry extensions.
3373 Currently CRL reason, invalidity date and hold instruction are
3374 supported. Add new CRL extensions to V3 code and some new objects.
3377 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3378 disable standard block padding (aka PKCS#5 padding) in the EVP
3379 API, which was previously mandatory. This means that the data is
3380 not padded in any way and so the total length much be a multiple
3381 of the block size, otherwise an error occurs.
3384 *) Initial (incomplete) OCSP SSL support.
3387 *) New function OCSP_parse_url(). This splits up a URL into its host,
3388 port and path components: primarily to parse OCSP URLs. New -url
3389 option to ocsp utility.
3392 *) New nonce behavior. The return value of OCSP_check_nonce() now
3393 reflects the various checks performed. Applications can decide
3394 whether to tolerate certain situations such as an absent nonce
3395 in a response when one was present in a request: the ocsp application
3396 just prints out a warning. New function OCSP_add1_basic_nonce()
3397 this is to allow responders to include a nonce in a response even if
3398 the request is nonce-less.
3401 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3402 skipped when using openssl x509 multiple times on a single input file,
3403 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3406 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3407 set string type: to handle setting ASN1_TIME structures. Fix ca
3408 utility to correctly initialize revocation date of CRLs.
3411 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3412 the clients preferred ciphersuites and rather use its own preferences.
3413 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3414 Internet Explorer by ensuring unchanged hash method during stepup.
3415 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3418 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3419 to aes and add a new 'exist' option to print out symbols that don't
3423 *) Additional options to ocsp utility to allow flags to be set and
3424 additional certificates supplied.
3427 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3428 OCSP client a number of certificate to only verify the response
3432 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3433 handle the new API. Currently only ECB, CBC modes supported. Add new
3436 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3437 Encryption Standard (AES) Ciphersuites for Transport Layer
3438 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3439 not enabled by default and were not part of the "ALL" ciphersuite
3440 alias because they were not yet official; they could be
3441 explicitly requested by specifying the "AESdraft" ciphersuite
3442 group alias. In the final release of OpenSSL 0.9.7, the group
3443 alias is called "AES" and is part of "ALL".)
3444 [Ben Laurie, Steve Henson, Bodo Moeller]
3446 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3447 request to response.
3450 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3451 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3452 extract information from a certificate request. OCSP_response_create()
3453 creates a response and optionally adds a basic response structure.
3454 OCSP_basic_add1_status() adds a complete single response to a basic
3455 response and returns the OCSP_SINGLERESP structure just added (to allow
3456 extensions to be included for example). OCSP_basic_add1_cert() adds a
3457 certificate to a basic response and OCSP_basic_sign() signs a basic
3458 response with various flags. New helper functions ASN1_TIME_check()
3459 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3460 (converts ASN1_TIME to GeneralizedTime).
3463 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3464 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3465 structure from a certificate. X509_pubkey_digest() digests the public_key
3466 contents: this is used in various key identifiers.
3469 *) Make sk_sort() tolerate a NULL argument.
3470 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3472 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3473 passed by the function are trusted implicitly. If any of them signed the
3474 response then it is assumed to be valid and is not verified.
3477 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3478 to data. This was previously part of the PKCS7 ASN1 code. This
3479 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3480 [Steve Henson, reported by Kenneth R. Robinette
3481 <support@securenetterm.com>]
3483 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3484 routines: without these tracing memory leaks is very painful.
3485 Fix leaks in PKCS12 and PKCS7 routines.
3488 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3489 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3490 effectively meant GeneralizedTime would never be used. Now it
3491 is initialised to -1 but X509_time_adj() now has to check the value
3492 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3493 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3494 [Steve Henson, reported by Kenneth R. Robinette
3495 <support@securenetterm.com>]
3497 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3498 result in a zero length in the ASN1_INTEGER structure which was
3499 not consistent with the structure when d2i_ASN1_INTEGER() was used
3500 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3501 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3502 where it did not print out a minus for negative ASN1_INTEGER.
3505 *) Add summary printout to ocsp utility. The various functions which
3506 convert status values to strings have been renamed to:
3507 OCSP_response_status_str(), OCSP_cert_status_str() and
3508 OCSP_crl_reason_str() and are no longer static. New options
3509 to verify nonce values and to disable verification. OCSP response
3510 printout format cleaned up.
3513 *) Add additional OCSP certificate checks. These are those specified
3514 in RFC2560. This consists of two separate checks: the CA of the
3515 certificate being checked must either be the OCSP signer certificate
3516 or the issuer of the OCSP signer certificate. In the latter case the
3517 OCSP signer certificate must contain the OCSP signing extended key
3518 usage. This check is performed by attempting to match the OCSP
3519 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3520 in the OCSP_CERTID structures of the response.
3523 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3524 and related routines. This uses the standard OpenSSL certificate
3525 verify routines to perform initial checks (just CA validity) and
3526 to obtain the certificate chain. Then additional checks will be
3527 performed on the chain. Currently the root CA is checked to see
3528 if it is explicitly trusted for OCSP signing. This is used to set
3529 a root CA as a global signing root: that is any certificate that
3530 chains to that CA is an acceptable OCSP signing certificate.
3533 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3534 extensions from a separate configuration file.
3535 As when reading extensions from the main configuration file,
3536 the '-extensions ...' option may be used for specifying the
3538 [Massimiliano Pala <madwolf@comune.modena.it>]
3540 *) New OCSP utility. Allows OCSP requests to be generated or
3541 read. The request can be sent to a responder and the output
3542 parsed, outputed or printed in text form. Not complete yet:
3543 still needs to check the OCSP response validity.
3546 *) New subcommands for 'openssl ca':
3547 'openssl ca -status <serial>' prints the status of the cert with
3548 the given serial number (according to the index file).
3549 'openssl ca -updatedb' updates the expiry status of certificates
3551 [Massimiliano Pala <madwolf@comune.modena.it>]
3553 *) New '-newreq-nodes' command option to CA.pl. This is like
3554 '-newreq', but calls 'openssl req' with the '-nodes' option
3555 so that the resulting key is not encrypted.
3556 [Damien Miller <djm@mindrot.org>]
3558 *) New configuration for the GNU Hurd.
3559 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3561 *) Initial code to implement OCSP basic response verify. This
3562 is currently incomplete. Currently just finds the signer's
3563 certificate and verifies the signature on the response.
3566 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3567 value of OPENSSLDIR. This is available via the new '-d' option
3568 to 'openssl version', and is also included in 'openssl version -a'.
3571 *) Allowing defining memory allocation callbacks that will be given
3572 file name and line number information in additional arguments
3573 (a const char* and an int). The basic functionality remains, as
3574 well as the original possibility to just replace malloc(),
3575 realloc() and free() by functions that do not know about these
3576 additional arguments. To register and find out the current
3577 settings for extended allocation functions, the following
3578 functions are provided:
3580 CRYPTO_set_mem_ex_functions
3581 CRYPTO_set_locked_mem_ex_functions
3582 CRYPTO_get_mem_ex_functions
3583 CRYPTO_get_locked_mem_ex_functions
3585 These work the same way as CRYPTO_set_mem_functions and friends.
3586 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3587 extended allocation function is enabled.
3588 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3589 a conventional allocation function is enabled.
3590 [Richard Levitte, Bodo Moeller]
3592 *) Finish off removing the remaining LHASH function pointer casts.
3593 There should no longer be any prototype-casting required when using
3594 the LHASH abstraction, and any casts that remain are "bugs". See
3595 the callback types and macros at the head of lhash.h for details
3596 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3599 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3600 If /dev/[u]random devices are not available or do not return enough
3601 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3603 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3604 /etc/entropy will be queried once each in this sequence, quering stops
3605 when enough entropy was collected without querying more sockets.
3608 *) Change the Unix RAND_poll() variant to be able to poll several
3609 random devices, as specified by DEVRANDOM, until a sufficient amount
3610 of data has been collected. We spend at most 10 ms on each file
3611 (select timeout) and read in non-blocking mode. DEVRANDOM now
3612 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3613 (previously it was just the string "/dev/urandom"), so on typical
3614 platforms the 10 ms delay will never occur.
3615 Also separate out the Unix variant to its own file, rand_unix.c.
3616 For VMS, there's a currently-empty rand_vms.c.
3619 *) Move OCSP client related routines to ocsp_cl.c. These
3620 provide utility functions which an application needing
3621 to issue a request to an OCSP responder and analyse the
3622 response will typically need: as opposed to those which an
3623 OCSP responder itself would need which will be added later.
3625 OCSP_request_sign() signs an OCSP request with an API similar
3626 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3627 response. OCSP_response_get1_basic() extracts basic response
3628 from response. OCSP_resp_find_status(): finds and extracts status
3629 information from an OCSP_CERTID structure (which will be created
3630 when the request structure is built). These are built from lower
3631 level functions which work on OCSP_SINGLERESP structures but
3632 wont normally be used unless the application wishes to examine
3633 extensions in the OCSP response for example.
3635 Replace nonce routines with a pair of functions.
3636 OCSP_request_add1_nonce() adds a nonce value and optionally
3637 generates a random value. OCSP_check_nonce() checks the
3638 validity of the nonce in an OCSP response.
3641 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3642 This doesn't copy the supplied OCSP_CERTID and avoids the
3643 need to free up the newly created id. Change return type
3644 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3645 This can then be used to add extensions to the request.
3646 Deleted OCSP_request_new(), since most of its functionality
3647 is now in OCSP_REQUEST_new() (and the case insensitive name
3648 clash) apart from the ability to set the request name which
3649 will be added elsewhere.
3652 *) Update OCSP API. Remove obsolete extensions argument from
3653 various functions. Extensions are now handled using the new
3654 OCSP extension code. New simple OCSP HTTP function which
3655 can be used to send requests and parse the response.
3658 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3659 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3660 uses the special reorder version of SET OF to sort the attributes
3661 and reorder them to match the encoded order. This resolves a long
3662 standing problem: a verify on a PKCS7 structure just after signing
3663 it used to fail because the attribute order did not match the
3664 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3665 it uses the received order. This is necessary to tolerate some broken
3666 software that does not order SET OF. This is handled by encoding
3667 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3668 to produce the required SET OF.
3671 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3672 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3673 files to get correct declarations of the ASN.1 item variables.
3676 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3677 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3678 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3679 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3680 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3681 ASN1_ITEM and no wrapper functions.
3684 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3685 replace the old function pointer based I/O routines. Change most of
3686 the *_d2i_bio() and *_d2i_fp() functions to use these.
3689 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3690 lines, recognice more "algorithms" that can be deselected, and make
3691 it complain about algorithm deselection that isn't recognised.
3694 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3695 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3696 to use new functions. Add NO_ASN1_OLD which can be set to remove
3697 some old style ASN1 functions: this can be used to determine if old
3698 code will still work when these eventually go away.
3701 *) New extension functions for OCSP structures, these follow the
3702 same conventions as certificates and CRLs.
3705 *) New function X509V3_add1_i2d(). This automatically encodes and
3706 adds an extension. Its behaviour can be customised with various
3707 flags to append, replace or delete. Various wrappers added for
3708 certifcates and CRLs.
3711 *) Fix to avoid calling the underlying ASN1 print routine when
3712 an extension cannot be parsed. Correct a typo in the
3713 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3716 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3717 entries for variables.
3720 *) Add functionality to apps/openssl.c for detecting locking
3721 problems: As the program is single-threaded, all we have
3722 to do is register a locking callback using an array for
3723 storing which locks are currently held by the program.
3726 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3727 SSL_get_ex_data_X509_STORE_idx(), which is used in
3728 ssl_verify_cert_chain() and thus can be called at any time
3729 during TLS/SSL handshakes so that thread-safety is essential.
3730 Unfortunately, the ex_data design is not at all suited
3731 for multi-threaded use, so it probably should be abolished.
3734 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3735 [Broadcom, tweaked and integrated by Geoff Thorpe]
3737 *) Move common extension printing code to new function
3738 X509V3_print_extensions(). Reorganise OCSP print routines and
3739 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3742 *) New function X509_signature_print() to remove duplication in some
3746 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3747 set (this was treated exactly the same as SET OF previously). This
3748 is used to reorder the STACK representing the structure to match the
3749 encoding. This will be used to get round a problem where a PKCS7
3750 structure which was signed could not be verified because the STACK
3751 order did not reflect the encoded order.
3754 *) Reimplement the OCSP ASN1 module using the new code.
3757 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3758 for its ASN1 operations. The old style function pointers still exist
3759 for now but they will eventually go away.
3762 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3763 completely replaces the old ASN1 functionality with a table driven
3764 encoder and decoder which interprets an ASN1_ITEM structure describing
3765 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3766 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3767 has also been converted to the new form.
3770 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3771 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3772 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3773 for negative moduli.
3776 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3777 of not touching the result's sign bit.
3780 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3784 *) Changed the LHASH code to use prototypes for callbacks, and created
3785 macros to declare and implement thin (optionally static) functions
3786 that provide type-safety and avoid function pointer casting for the
3787 type-specific callbacks.
3790 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3792 [Veers Staats <staatsvr@asc.hpc.mil>,
3793 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3795 *) Reformat the FAQ so the different questions and answers can be divided
3796 in sections depending on the subject.
3799 *) Have the zlib compression code load ZLIB.DLL dynamically under
3803 *) New function BN_mod_sqrt for computing square roots modulo a prime
3804 (using the probabilistic Tonelli-Shanks algorithm unless
3805 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3806 be handled deterministically).
3807 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3809 *) Make BN_mod_inverse faster by explicitly handling small quotients
3810 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3811 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3814 *) New function BN_kronecker.
3817 *) Fix BN_gcd so that it works on negative inputs; the result is
3818 positive unless both parameters are zero.
3819 Previously something reasonably close to an infinite loop was
3820 possible because numbers could be growing instead of shrinking
3821 in the implementation of Euclid's algorithm.
3824 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3825 sign of the number in question.
3827 Fix BN_is_word(a,w) to work correctly for w == 0.
3829 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3830 because its test if the absolute value of 'a' equals 'w'.
3831 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3832 it exists mostly for use in the implementations of BN_is_zero(),
3833 BN_is_one(), and BN_is_word().
3836 *) New function BN_swap.
3839 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3840 the exponentiation functions are more likely to produce reasonable
3841 results on negative inputs.
3844 *) Change BN_mod_mul so that the result is always non-negative.
3845 Previously, it could be negative if one of the factors was negative;
3846 I don't think anyone really wanted that behaviour.
3849 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3850 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3851 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3852 and add new functions:
3861 BN_mod_lshift1_quick
3865 These functions always generate non-negative results.
3867 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3868 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3870 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3871 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3872 be reduced modulo m.
3873 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3876 The following entry accidentily appeared in the CHANGES file
3877 distributed with OpenSSL 0.9.7. The modifications described in
3878 it do *not* apply to OpenSSL 0.9.7.
3880 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3881 was actually never needed) and in BN_mul(). The removal in BN_mul()
3882 required a small change in bn_mul_part_recursive() and the addition
3883 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3884 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3885 bn_sub_words() and bn_add_words() except they take arrays with
3890 *) In 'openssl passwd', verify passwords read from the terminal
3891 unless the '-salt' option is used (which usually means that
3892 verification would just waste user's time since the resulting
3893 hash is going to be compared with some given password hash)
3894 or the new '-noverify' option is used.
3896 This is an incompatible change, but it does not affect
3897 non-interactive use of 'openssl passwd' (passwords on the command
3898 line, '-stdin' option, '-in ...' option) and thus should not
3902 *) Remove all references to RSAref, since there's no more need for it.
3905 *) Make DSO load along a path given through an environment variable
3906 (SHLIB_PATH) with shl_load().
3909 *) Constify the ENGINE code as a result of BIGNUM constification.
3910 Also constify the RSA code and most things related to it. In a
3911 few places, most notable in the depth of the ASN.1 code, ugly
3912 casts back to non-const were required (to be solved at a later
3916 *) Make it so the openssl application has all engines loaded by default.
3919 *) Constify the BIGNUM routines a little more.
3922 *) Add the following functions:
3924 ENGINE_load_cswift()
3926 ENGINE_load_atalla()
3928 ENGINE_load_builtin_engines()
3930 That way, an application can itself choose if external engines that
3931 are built-in in OpenSSL shall ever be used or not. The benefit is
3932 that applications won't have to be linked with libdl or other dso
3933 libraries unless it's really needed.
3935 Changed 'openssl engine' to load all engines on demand.
3936 Changed the engine header files to avoid the duplication of some
3937 declarations (they differed!).
3940 *) 'openssl engine' can now list capabilities.
3943 *) Better error reporting in 'openssl engine'.
3946 *) Never call load_dh_param(NULL) in s_server.
3949 *) Add engine application. It can currently list engines by name and
3950 identity, and test if they are actually available.
3953 *) Improve RPM specification file by forcing symbolic linking and making
3954 sure the installed documentation is also owned by root.root.
3955 [Damien Miller <djm@mindrot.org>]
3957 *) Give the OpenSSL applications more possibilities to make use of
3958 keys (public as well as private) handled by engines.
3961 *) Add OCSP code that comes from CertCo.
3964 *) Add VMS support for the Rijndael code.
3967 *) Added untested support for Nuron crypto accelerator.
3970 *) Add support for external cryptographic devices. This code was
3971 previously distributed separately as the "engine" branch.
3972 [Geoff Thorpe, Richard Levitte]
3974 *) Rework the filename-translation in the DSO code. It is now possible to
3975 have far greater control over how a "name" is turned into a filename
3976 depending on the operating environment and any oddities about the
3977 different shared library filenames on each system.
3980 *) Support threads on FreeBSD-elf in Configure.
3983 *) Fix for SHA1 assembly problem with MASM: it produces
3984 warnings about corrupt line number information when assembling
3985 with debugging information. This is caused by the overlapping
3987 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3990 NCONF_get_number() has no error checking at all. As a replacement,
3991 NCONF_get_number_e() is defined (_e for "error checking") and is
3992 promoted strongly. The old NCONF_get_number is kept around for
3993 binary backward compatibility.
3994 Make it possible for methods to load from something other than a BIO,
3995 by providing a function pointer that is given a name instead of a BIO.
3996 For example, this could be used to load configuration data from an
4000 *) Fix for non blocking accept BIOs. Added new I/O special reason
4001 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4002 with non blocking I/O was not possible because no retry code was
4003 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4007 *) Added the beginnings of Rijndael support.
4010 *) Fix for bug in DirectoryString mask setting. Add support for
4011 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4012 to allow certificate printing to more controllable, additional
4013 'certopt' option to 'x509' to allow new printing options to be
4017 *) Clean old EAY MD5 hack from e_os.h.
4020 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4022 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4023 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4024 [Joe Orton, Steve Henson]
4026 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4028 *) Fix additional bug revealed by the NISCC test suite:
4030 Stop bug triggering large recursion when presented with
4031 certain ASN.1 tags (CVE-2003-0851)
4034 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4036 *) Fix various bugs revealed by running the NISCC test suite:
4038 Stop out of bounds reads in the ASN1 code when presented with
4039 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4041 If verify callback ignores invalid public key errors don't try to check
4042 certificate signature with the NULL public key.
4046 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4047 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4051 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4052 extra data after the compression methods not only for TLS 1.0
4053 but also for SSL 3.0 (as required by the specification).
4054 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4056 *) Change X509_certificate_type() to mark the key as exported/exportable
4057 when it's 512 *bits* long, not 512 bytes.
4060 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4062 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4063 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4064 a protocol version number mismatch like a decryption error
4065 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4068 *) Turn on RSA blinding by default in the default implementation
4069 to avoid a timing attack. Applications that don't want it can call
4070 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4071 They would be ill-advised to do so in most cases.
4072 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4074 *) Change RSA blinding code so that it works when the PRNG is not
4075 seeded (in this case, the secret RSA exponent is abused as
4076 an unpredictable seed -- if it is not unpredictable, there
4077 is no point in blinding anyway). Make RSA blinding thread-safe
4078 by remembering the creator's thread ID in rsa->blinding and
4079 having all other threads use local one-time blinding factors
4080 (this requires more computation than sharing rsa->blinding, but
4081 avoids excessive locking; and if an RSA object is not shared
4082 between threads, blinding will still be very fast).
4085 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4087 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4088 via timing by performing a MAC computation even if incorrrect
4089 block cipher padding has been found. This is a countermeasure
4090 against active attacks where the attacker has to distinguish
4091 between bad padding and a MAC verification error. (CVE-2003-0078)
4093 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4094 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4095 Martin Vuagnoux (EPFL, Ilion)]
4097 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4099 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4100 memory from it's contents. This is done with a counter that will
4101 place alternating values in each byte. This can be used to solve
4102 two issues: 1) the removal of calls to memset() by highly optimizing
4103 compilers, and 2) cleansing with other values than 0, since those can
4104 be read through on certain media, for example a swap space on disk.
4107 *) Bugfix: client side session caching did not work with external caching,
4108 because the session->cipher setting was not restored when reloading
4109 from the external cache. This problem was masked, when
4110 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4111 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4114 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4115 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4116 [Zeev Lieber <zeev-l@yahoo.com>]
4118 *) Undo an undocumented change introduced in 0.9.6e which caused
4119 repeated calls to OpenSSL_add_all_ciphers() and
4120 OpenSSL_add_all_digests() to be ignored, even after calling
4124 *) Change the default configuration reader to deal with last line not
4125 being properly terminated.
4128 *) Change X509_NAME_cmp() so it applies the special rules on handling
4129 DN values that are of type PrintableString, as well as RDNs of type
4130 emailAddress where the value has the type ia5String.
4131 [stefank@valicert.com via Richard Levitte]
4133 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4134 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4135 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4136 the bitwise-OR of the two for use by the majority of applications
4137 wanting this behaviour, and update the docs. The documented
4138 behaviour and actual behaviour were inconsistent and had been
4139 changing anyway, so this is more a bug-fix than a behavioural
4141 [Geoff Thorpe, diagnosed by Nadav Har'El]
4143 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4144 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4147 *) Fix initialization code race conditions in
4148 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4149 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4150 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4151 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4152 ssl2_get_cipher_by_char(),
4153 ssl3_get_cipher_by_char().
4154 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4156 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4157 the cached sessions are flushed, as the remove_cb() might use ex_data
4158 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4159 (see [openssl.org #212]).
4160 [Geoff Thorpe, Lutz Jaenicke]
4162 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4163 length, instead of the encoding length to d2i_ASN1_OBJECT.
4166 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4168 *) [In 0.9.6g-engine release:]
4169 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4170 [Lynn Gazis <lgazis@rainbow.com>]
4172 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4174 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4175 and get fix the header length calculation.
4176 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4177 Alon Kantor <alonk@checkpoint.com> (and others),
4180 *) Use proper error handling instead of 'assertions' in buffer
4181 overflow checks added in 0.9.6e. This prevents DoS (the
4182 assertions could call abort()).
4183 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4185 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4187 *) Add various sanity checks to asn1_get_length() to reject
4188 the ASN1 length bytes if they exceed sizeof(long), will appear
4189 negative or the content length exceeds the length of the
4191 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4193 *) Fix cipher selection routines: ciphers without encryption had no flags
4194 for the cipher strength set and where therefore not handled correctly
4195 by the selection routines (PR #130).
4198 *) Fix EVP_dsa_sha macro.
4202 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4203 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4204 that was added in OpenSSL 0.9.6d.
4206 As the countermeasure turned out to be incompatible with some
4207 broken SSL implementations, the new option is part of SSL_OP_ALL.
4208 SSL_OP_ALL is usually employed when compatibility with weird SSL
4209 implementations is desired (e.g. '-bugs' option to 's_client' and
4210 's_server'), so the new option is automatically set in many
4214 *) Changes in security patch:
4216 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4217 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4218 Air Force Materiel Command, USAF, under agreement number
4221 *) Add various sanity checks to asn1_get_length() to reject
4222 the ASN1 length bytes if they exceed sizeof(long), will appear
4223 negative or the content length exceeds the length of the
4224 supplied buffer. (CVE-2002-0659)
4225 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4227 *) Assertions for various potential buffer overflows, not known to
4229 [Ben Laurie (CHATS)]
4231 *) Various temporary buffers to hold ASCII versions of integers were
4232 too small for 64 bit platforms. (CVE-2002-0655)
4233 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4235 *) Remote buffer overflow in SSL3 protocol - an attacker could
4236 supply an oversized session ID to a client. (CVE-2002-0656)
4237 [Ben Laurie (CHATS)]
4239 *) Remote buffer overflow in SSL2 protocol - an attacker could
4240 supply an oversized client master key. (CVE-2002-0656)
4241 [Ben Laurie (CHATS)]
4243 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4245 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4246 encoded as NULL) with id-dsa-with-sha1.
4247 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4249 *) Check various X509_...() return values in apps/req.c.
4250 [Nils Larsch <nla@trustcenter.de>]
4252 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4253 an end-of-file condition would erronously be flagged, when the CRLF
4254 was just at the end of a processed block. The bug was discovered when
4255 processing data through a buffering memory BIO handing the data to a
4256 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4257 <ptsekov@syntrex.com> and Nedelcho Stanev.
4260 *) Implement a countermeasure against a vulnerability recently found
4261 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4262 before application data chunks to avoid the use of known IVs
4263 with data potentially chosen by the attacker.
4266 *) Fix length checks in ssl3_get_client_hello().
4269 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4270 to prevent ssl3_read_internal() from incorrectly assuming that
4271 ssl3_read_bytes() found application data while handshake
4272 processing was enabled when in fact s->s3->in_read_app_data was
4273 merely automatically cleared during the initial handshake.
4274 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4276 *) Fix object definitions for Private and Enterprise: they were not
4277 recognized in their shortname (=lowercase) representation. Extend
4278 obj_dat.pl to issue an error when using undefined keywords instead
4279 of silently ignoring the problem (Svenning Sorensen
4280 <sss@sss.dnsalias.net>).
4283 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4284 generators, i.e. generators other than 2 and 5. (Previously, the
4285 code did not properly initialise the 'add' and 'rem' values to
4286 BN_generate_prime().)
4288 In the new general case, we do not insist that 'generator' is
4289 actually a primitive root: This requirement is rather pointless;
4290 a generator of the order-q subgroup is just as good, if not
4294 *) Map new X509 verification errors to alerts. Discovered and submitted by
4295 Tom Wu <tom@arcot.com>.
4298 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4299 returning non-zero before the data has been completely received
4300 when using non-blocking I/O.
4301 [Bodo Moeller; problem pointed out by John Hughes]
4303 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4304 [Ben Laurie, Lutz Jaenicke]
4306 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4307 Yoram Zahavi <YoramZ@gilian.com>).
4310 *) Add information about CygWin 1.3 and on, and preserve proper
4311 configuration for the versions before that.
4312 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4314 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4315 check whether we deal with a copy of a session and do not delete from
4316 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4317 <izhar@checkpoint.com>.
4320 *) Do not store session data into the internal session cache, if it
4321 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4322 flag is set). Proposed by Aslam <aslam@funk.com>.
4325 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4329 *) [In 0.9.6d-engine release:]
4330 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4331 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4333 *) Add the configuration target linux-s390x.
4334 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4336 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4337 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4338 variable as an indication that a ClientHello message has been
4339 received. As the flag value will be lost between multiple
4340 invocations of ssl3_accept when using non-blocking I/O, the
4341 function may not be aware that a handshake has actually taken
4342 place, thus preventing a new session from being added to the
4345 To avoid this problem, we now set s->new_session to 2 instead of
4346 using a local variable.
4347 [Lutz Jaenicke, Bodo Moeller]
4349 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4350 if the SSL_R_LENGTH_MISMATCH error is detected.
4351 [Geoff Thorpe, Bodo Moeller]
4353 *) New 'shared_ldflag' column in Configure platform table.
4356 *) Fix EVP_CIPHER_mode macro.
4357 ["Dan S. Camper" <dan@bti.net>]
4359 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4360 type, we must throw them away by setting rr->length to 0.
4361 [D P Chang <dpc@qualys.com>]
4363 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4365 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4366 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4367 worked incorrectly for those cases where range = 10..._2 and
4368 3*range is two bits longer than range.)
4371 *) Only add signing time to PKCS7 structures if it is not already
4375 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4376 OBJ_ld_ce should be OBJ_id_ce.
4377 Also some ip-pda OIDs in crypto/objects/objects.txt were
4378 incorrect (cf. RFC 3039).
4379 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4381 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4382 returns early because it has nothing to do.
4383 [Andy Schneider <andy.schneider@bjss.co.uk>]
4385 *) [In 0.9.6c-engine release:]
4386 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4387 [Andy Schneider <andy.schneider@bjss.co.uk>]
4389 *) [In 0.9.6c-engine release:]
4390 Add support for Cryptographic Appliance's keyserver technology.
4391 (Use engine 'keyclient')
4392 [Cryptographic Appliances and Geoff Thorpe]
4394 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4395 is called via tools/c89.sh because arguments have to be
4396 rearranged (all '-L' options must appear before the first object
4398 [Richard Shapiro <rshapiro@abinitio.com>]
4400 *) [In 0.9.6c-engine release:]
4401 Add support for Broadcom crypto accelerator cards, backported
4403 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4405 *) [In 0.9.6c-engine release:]
4406 Add support for SureWare crypto accelerator cards from
4407 Baltimore Technologies. (Use engine 'sureware')
4408 [Baltimore Technologies and Mark Cox]
4410 *) [In 0.9.6c-engine release:]
4411 Add support for crypto accelerator cards from Accelerated
4412 Encryption Processing, www.aep.ie. (Use engine 'aep')
4413 [AEP Inc. and Mark Cox]
4415 *) Add a configuration entry for gcc on UnixWare.
4416 [Gary Benson <gbenson@redhat.com>]
4418 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4419 messages are stored in a single piece (fixed-length part and
4420 variable-length part combined) and fix various bugs found on the way.
4423 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4424 instead. BIO_gethostbyname() does not know what timeouts are
4425 appropriate, so entries would stay in cache even when they have
4427 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4429 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4430 faced with a pathologically small ClientHello fragment that does
4431 not contain client_version: Instead of aborting with an error,
4432 simply choose the highest available protocol version (i.e.,
4433 TLS 1.0 unless it is disabled). In practice, ClientHello
4434 messages are never sent like this, but this change gives us
4435 strictly correct behaviour at least for TLS.
4438 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4439 never resets s->method to s->ctx->method when called from within
4440 one of the SSL handshake functions.
4441 [Bodo Moeller; problem pointed out by Niko Baric]
4443 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4444 (sent using the client's version number) if client_version is
4445 smaller than the protocol version in use. Also change
4446 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4447 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4448 the client will at least see that alert.
4451 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4455 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4456 client receives HelloRequest while in a handshake.
4457 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4459 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4460 should end in 'break', not 'goto end' which circuments various
4461 cleanups done in state SSL_ST_OK. But session related stuff
4462 must be disabled for SSL_ST_OK in the case that we just sent a
4465 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4466 before just sending a HelloRequest.
4467 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4469 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4470 reveal whether illegal block cipher padding was found or a MAC
4471 verification error occured. (Neither SSLerr() codes nor alerts
4472 are directly visible to potential attackers, but the information
4473 may leak via logfiles.)
4475 Similar changes are not required for the SSL 2.0 implementation
4476 because the number of padding bytes is sent in clear for SSL 2.0,
4477 and the extra bytes are just ignored. However ssl/s2_pkt.c
4478 failed to verify that the purported number of padding bytes is in
4482 *) Add OpenUNIX-8 support including shared libraries
4483 (Boyd Lynn Gerber <gerberb@zenez.com>).
4486 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4487 'wristwatch attack' using huge encoding parameters (cf.
4488 James H. Manger's CRYPTO 2001 paper). Note that the
4489 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4490 encoding parameters and hence was not vulnerable.
4493 *) BN_sqr() bug fix.
4494 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4496 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4497 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4498 followed by modular reduction.
4499 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4501 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4502 equivalent based on BN_pseudo_rand() instead of BN_rand().
4505 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4506 This function was broken, as the check for a new client hello message
4507 to handle SGC did not allow these large messages.
4508 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4511 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4514 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4515 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4518 *) Rework the configuration and shared library support for Tru64 Unix.
4519 The configuration part makes use of modern compiler features and
4520 still retains old compiler behavior for those that run older versions
4521 of the OS. The shared library support part includes a variant that
4522 uses the RPATH feature, and is available through the special
4523 configuration target "alpha-cc-rpath", which will never be selected
4525 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4527 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4528 with the same message size as in ssl3_get_certificate_request().
4529 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4530 messages might inadvertently be reject as too long.
4531 [Petr Lampa <lampa@fee.vutbr.cz>]
4533 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4536 *) Modified SSL library such that the verify_callback that has been set
4537 specificly for an SSL object with SSL_set_verify() is actually being
4538 used. Before the change, a verify_callback set with this function was
4539 ignored and the verify_callback() set in the SSL_CTX at the time of
4540 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4541 to allow the necessary settings.
4544 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4545 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4546 done automatically (in contradiction to the requirements of the C
4547 standard). This made problems when used from OpenSSH.
4550 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4551 dh->length and always used
4553 BN_rand_range(priv_key, dh->p).
4555 BN_rand_range() is not necessary for Diffie-Hellman, and this
4556 specific range makes Diffie-Hellman unnecessarily inefficient if
4557 dh->length (recommended exponent length) is much smaller than the
4558 length of dh->p. We could use BN_rand_range() if the order of
4559 the subgroup was stored in the DH structure, but we only have
4564 BN_rand(priv_key, l, ...)
4566 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4572 RSA_eay_public_encrypt
4573 RSA_eay_private_decrypt
4574 RSA_eay_private_encrypt (signing)
4575 RSA_eay_public_decrypt (signature verification)
4577 (default implementations for RSA_public_encrypt,
4578 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4579 always reject numbers >= n.
4582 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4583 to synchronize access to 'locking_thread'. This is necessary on
4584 systems where access to 'locking_thread' (an 'unsigned long'
4585 variable) is not atomic.
4588 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4589 *before* setting the 'crypto_lock_rand' flag. The previous code had
4590 a race condition if 0 is a valid thread ID.
4591 [Travis Vitek <vitek@roguewave.com>]
4593 *) Add support for shared libraries under Irix.
4594 [Albert Chin-A-Young <china@thewrittenword.com>]
4596 *) Add configuration option to build on Linux on both big-endian and
4598 [Ralf Baechle <ralf@uni-koblenz.de>]
4600 *) Add the possibility to create shared libraries on HP-UX.
4603 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4605 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4606 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4607 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4608 PRNG state recovery was possible based on the output of
4609 one PRNG request appropriately sized to gain knowledge on
4610 'md' followed by enough consecutive 1-byte PRNG requests
4611 to traverse all of 'state'.
4613 1. When updating 'md_local' (the current thread's copy of 'md')
4614 during PRNG output generation, hash all of the previous
4615 'md_local' value, not just the half used for PRNG output.
4617 2. Make the number of bytes from 'state' included into the hash
4618 independent from the number of PRNG bytes requested.
4620 The first measure alone would be sufficient to avoid
4621 Markku-Juhani's attack. (Actually it had never occurred
4622 to me that the half of 'md_local' used for chaining was the
4623 half from which PRNG output bytes were taken -- I had always
4624 assumed that the secret half would be used.) The second
4625 measure makes sure that additional data from 'state' is never
4626 mixed into 'md_local' in small portions; this heuristically
4627 further strengthens the PRNG.
4630 *) Fix crypto/bn/asm/mips3.s.
4633 *) When only the key is given to "enc", the IV is undefined. Print out
4634 an error message in this case.
4637 *) Handle special case when X509_NAME is empty in X509 printing routines.
4640 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4641 positive and less than q.
4644 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4645 used: it isn't thread safe and the add_lock_callback should handle
4647 [Paul Rose <Paul.Rose@bridge.com>]
4649 *) Verify that incoming data obeys the block size in
4650 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4654 [Ulf Möller, Bodo Möller]
4656 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4657 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4658 when fixing the server behaviour for backwards-compatible 'client
4659 hello' messages. (Note that the attack is impractical against
4660 SSL 3.0 and TLS 1.0 anyway because length and version checking
4661 means that the probability of guessing a valid ciphertext is
4662 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4665 Before 0.9.5, the countermeasure (hide the error by generating a
4666 random 'decryption result') did not work properly because
4667 ERR_clear_error() was missing, meaning that SSL_get_error() would
4668 detect the supposedly ignored error.
4670 Both problems are now fixed.
4673 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4674 (previously it was 1024).
4677 *) Fix for compatibility mode trust settings: ignore trust settings
4678 unless some valid trust or reject settings are present.
4681 *) Fix for blowfish EVP: its a variable length cipher.
4684 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4685 parameters in DSA public key structures and return an error in the
4686 DSA routines if parameters are absent.
4689 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4690 in the current directory if neither $RANDFILE nor $HOME was set.
4691 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4692 caused some confusion to Windows users who haven't defined $HOME.
4693 Thus RAND_file_name() is changed again: e_os.h can define a
4694 DEFAULT_HOME, which will be used if $HOME is not set.
4695 For Windows, we use "C:"; on other platforms, we still require
4696 environment variables.
4698 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4699 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4700 having multiple threads call RAND_poll() concurrently.
4703 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4704 combination of a flag and a thread ID variable.
4705 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4706 flag), *other* threads can enter ssleay_add_bytes without obeying
4707 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4708 that they do not hold after the first thread unsets add_do_not_lock).
4711 *) Change bctest again: '-x' expressions are not available in all
4715 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4717 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4718 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4720 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4721 the default extension for executables, if any. Also, make the perl
4722 scripts that use symlink() to test if it really exists and use "cp"
4723 if it doesn't. All this made OpenSSL compilable and installable in
4727 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4728 If SEQUENCE is length is indefinite just set c->slen to the total
4729 amount of data available.
4730 [Steve Henson, reported by shige@FreeBSD.org]
4731 [This change does not apply to 0.9.7.]
4733 *) Change bctest to avoid here-documents inside command substitution
4734 (workaround for FreeBSD /bin/sh bug).
4735 For compatibility with Ultrix, avoid shell functions (introduced
4736 in the bctest version that searches along $PATH).
4739 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4740 with des_encrypt() defined on some operating systems, like Solaris
4744 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4745 On the Importance of Eliminating Errors in Cryptographic
4746 Computations, J. Cryptology 14 (2001) 2, 101-119,
4747 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4750 *) MIPS assembler BIGNUM division bug fix.
4753 *) Disabled incorrect Alpha assembler code.
4756 *) Fix PKCS#7 decode routines so they correctly update the length
4757 after reading an EOC for the EXPLICIT tag.
4759 [This change does not apply to 0.9.7.]
4761 *) Fix bug in PKCS#12 key generation routines. This was triggered
4762 if a 3DES key was generated with a 0 initial byte. Include
4763 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4764 (but broken) behaviour.
4767 *) Enhance bctest to search for a working bc along $PATH and print
4769 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4771 *) Fix memory leaks in err.c: free err_data string if necessary;
4772 don't write to the wrong index in ERR_set_error_data.
4775 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4779 *) Replace rdtsc with _emit statements for VC++ version 5.
4780 [Jeremy Cooper <jeremy@baymoo.org>]
4782 *) Make it possible to reuse SSLv2 sessions.
4785 *) In copy_email() check for >= 0 as a return value for
4786 X509_NAME_get_index_by_NID() since 0 is a valid index.
4787 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4789 *) Avoid coredump with unsupported or invalid public keys by checking if
4790 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4791 PKCS7_verify() fails with non detached data.
4794 *) Don't use getenv in library functions when run as setuid/setgid.
4795 New function OPENSSL_issetugid().
4798 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4799 due to incorrect handling of multi-threading:
4801 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4803 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4805 3. Count how many times MemCheck_off() has been called so that
4806 nested use can be treated correctly. This also avoids
4807 inband-signalling in the previous code (which relied on the
4808 assumption that thread ID 0 is impossible).
4811 *) Add "-rand" option also to s_client and s_server.
4814 *) Fix CPU detection on Irix 6.x.
4815 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4816 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4818 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4821 [This change does not apply to 0.9.7.]
4823 *) Use the cached encoding of an X509_NAME structure rather than
4824 copying it. This is apparently the reason for the libsafe "errors"
4825 but the code is actually correct.
4828 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4829 Bleichenbacher's DSA attack.
4830 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4831 to be set and top=0 forces the highest bit to be set; top=-1 is new
4832 and leaves the highest bit random.
4833 [Ulf Moeller, Bodo Moeller]
4835 *) In the NCONF_...-based implementations for CONF_... queries
4836 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4837 a temporary CONF structure with the data component set to NULL
4838 (which gives segmentation faults in lh_retrieve).
4839 Instead, use NULL for the CONF pointer in CONF_get_string and
4840 CONF_get_number (which may use environment variables) and directly
4841 return NULL from CONF_get_section.
4844 *) Fix potential buffer overrun for EBCDIC.
4847 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4848 keyUsage if basicConstraints absent for a CA.
4851 *) Make SMIME_write_PKCS7() write mail header values with a format that
4852 is more generally accepted (no spaces before the semicolon), since
4853 some programs can't parse those values properly otherwise. Also make
4854 sure BIO's that break lines after each write do not create invalid
4858 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4859 macros previously used would not encode an empty SEQUENCE OF
4860 and break the signature.
4862 [This change does not apply to 0.9.7.]
4864 *) Zero the premaster secret after deriving the master secret in
4868 *) Add some EVP_add_digest_alias registrations (as found in
4869 OpenSSL_add_all_digests()) to SSL_library_init()
4870 aka OpenSSL_add_ssl_algorithms(). This provides improved
4871 compatibility with peers using X.509 certificates
4872 with unconventional AlgorithmIdentifier OIDs.
4875 *) Fix for Irix with NO_ASM.
4876 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4878 *) ./config script fixes.
4879 [Ulf Moeller, Richard Levitte]
4881 *) Fix 'openssl passwd -1'.
4884 *) Change PKCS12_key_gen_asc() so it can cope with non null
4885 terminated strings whose length is passed in the passlen
4886 parameter, for example from PEM callbacks. This was done
4887 by adding an extra length parameter to asc2uni().
4888 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4890 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4891 call failed, free the DSA structure.
4894 *) Fix to uni2asc() to cope with zero length Unicode strings.
4895 These are present in some PKCS#12 files.
4898 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4899 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4900 when writing a 32767 byte record.
4901 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4903 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4904 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4906 (RSA objects have a reference count access to which is protected
4907 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4908 so they are meant to be shared between threads.)
4909 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4910 "Reddie, Steven" <Steven.Reddie@ca.com>]
4912 *) Fix a deadlock in CRYPTO_mem_leaks().
4915 *) Use better test patterns in bntest.
4918 *) rand_win.c fix for Borland C.
4921 *) BN_rshift bugfix for n == 0.
4924 *) Add a 'bctest' script that checks for some known 'bc' bugs
4925 so that 'make test' does not abort just because 'bc' is broken.
4928 *) Store verify_result within SSL_SESSION also for client side to
4929 avoid potential security hole. (Re-used sessions on the client side
4930 always resulted in verify_result==X509_V_OK, not using the original
4931 result of the server certificate verification.)
4934 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4935 SSL3_RT_APPLICATION_DATA, return 0.
4936 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4940 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4941 releases, have been re-implemented by renaming the previous
4942 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4943 and ssl3_read_internal, respectively, and adding 'peek' parameters
4944 to them. The new ssl[23]_{read,peek} functions are calls to
4945 ssl[23]_read_internal with the 'peek' flag set appropriately.
4946 A 'peek' parameter has also been added to ssl3_read_bytes, which
4947 does the actual work for ssl3_read_internal.
4950 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4951 the method-specific "init()" handler. Also clean up ex_data after
4952 calling the method-specific "finish()" handler. Previously, this was
4953 happening the other way round.
4956 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4957 The previous value, 12, was not always sufficient for BN_mod_exp().
4960 *) Make sure that shared libraries get the internal name engine with
4961 the full version number and not just 0. This should mark the
4962 shared libraries as not backward compatible. Of course, this should
4963 be changed again when we can guarantee backward binary compatibility.
4966 *) Fix typo in get_cert_by_subject() in by_dir.c
4967 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4969 *) Rework the system to generate shared libraries:
4971 - Make note of the expected extension for the shared libraries and
4972 if there is a need for symbolic links from for example libcrypto.so.0
4973 to libcrypto.so.0.9.7. There is extended info in Configure for
4976 - Make as few rebuilds of the shared libraries as possible.
4978 - Still avoid linking the OpenSSL programs with the shared libraries.
4980 - When installing, install the shared libraries separately from the
4984 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4986 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4987 and not in SSL_clear because the latter is also used by the
4988 accept/connect functions; previously, the settings made by
4989 SSL_set_read_ahead would be lost during the handshake.
4990 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4992 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4993 Previously, it would create entries for disableed algorithms no
4997 *) Added several new manual pages for SSL_* function.
5000 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5002 *) In ssl23_get_client_hello, generate an error message when faced
5003 with an initial SSL 3.0/TLS record that is too small to contain the
5004 first two bytes of the ClientHello message, i.e. client_version.
5005 (Note that this is a pathologic case that probably has never happened
5006 in real life.) The previous approach was to use the version number
5007 from the record header as a substitute; but our protocol choice
5008 should not depend on that one because it is not authenticated
5009 by the Finished messages.
5012 *) More robust randomness gathering functions for Windows.
5013 [Jeffrey Altman <jaltman@columbia.edu>]
5015 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5016 not set then we don't setup the error code for issuer check errors
5017 to avoid possibly overwriting other errors which the callback does
5018 handle. If an application does set the flag then we assume it knows
5019 what it is doing and can handle the new informational codes
5023 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5024 a general "ANY" type, as such it should be able to decode anything
5025 including tagged types. However it didn't check the class so it would
5026 wrongly interpret tagged types in the same way as their universal
5027 counterpart and unknown types were just rejected. Changed so that the
5028 tagged and unknown types are handled in the same way as a SEQUENCE:
5029 that is the encoding is stored intact. There is also a new type
5030 "V_ASN1_OTHER" which is used when the class is not universal, in this
5031 case we have no idea what the actual type is so we just lump them all
5035 *) On VMS, stdout may very well lead to a file that is written to
5036 in a record-oriented fashion. That means that every write() will
5037 write a separate record, which will be read separately by the
5038 programs trying to read from it. This can be very confusing.
5040 The solution is to put a BIO filter in the way that will buffer
5041 text until a linefeed is reached, and then write everything a
5042 line at a time, so every record written will be an actual line,
5043 not chunks of lines and not (usually doesn't happen, but I've
5044 seen it once) several lines in one record. BIO_f_linebuffer() is
5047 Currently, it's a VMS-only method, because that's where it has
5048 been tested well enough.
5051 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5052 it can return incorrect results.
5053 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5054 but it was in 0.9.6-beta[12].)
5057 *) Disable the check for content being present when verifying detached
5058 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5059 include zero length content when signing messages.
5062 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5063 BIO_ctrl (for BIO pairs).
5066 *) Add DSO method for VMS.
5069 *) Bug fix: Montgomery multiplication could produce results with the
5073 *) Add RPM specification openssl.spec and modify it to build three
5074 packages. The default package contains applications, application
5075 documentation and run-time libraries. The devel package contains
5076 include files, static libraries and function documentation. The
5077 doc package contains the contents of the doc directory. The original
5078 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5081 *) Add a large number of documentation files for many SSL routines.
5082 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5084 *) Add a configuration entry for Sony News 4.
5085 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5087 *) Don't set the two most significant bits to one when generating a
5088 random number < q in the DSA library.
5091 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5092 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5093 the underlying transport is blocking) if a handshake took place.
5094 (The default behaviour is needed by applications such as s_client
5095 and s_server that use select() to determine when to use SSL_read;
5096 but for applications that know in advance when to expect data, it
5097 just makes things more complicated.)
5100 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5104 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5105 work better on such systems.
5106 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5108 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5109 Update PKCS12_parse() so it copies the friendlyName and the
5110 keyid to the certificates aux info.
5113 *) Fix bug in PKCS7_verify() which caused an infinite loop
5114 if there was more than one signature.
5115 [Sven Uszpelkat <su@celocom.de>]
5117 *) Major change in util/mkdef.pl to include extra information
5118 about each symbol, as well as presentig variables as well
5119 as functions. This change means that there's n more need
5120 to rebuild the .num files when some algorithms are excluded.
5123 *) Allow the verify time to be set by an application,
5124 rather than always using the current time.
5127 *) Phase 2 verify code reorganisation. The certificate
5128 verify code now looks up an issuer certificate by a
5129 number of criteria: subject name, authority key id
5130 and key usage. It also verifies self signed certificates
5131 by the same criteria. The main comparison function is
5132 X509_check_issued() which performs these checks.
5134 Lot of changes were necessary in order to support this
5135 without completely rewriting the lookup code.
5137 Authority and subject key identifier are now cached.
5139 The LHASH 'certs' is X509_STORE has now been replaced
5140 by a STACK_OF(X509_OBJECT). This is mainly because an
5141 LHASH can't store or retrieve multiple objects with
5142 the same hash value.
5144 As a result various functions (which were all internal
5145 use only) have changed to handle the new X509_STORE
5146 structure. This will break anything that messed round
5147 with X509_STORE internally.
5149 The functions X509_STORE_add_cert() now checks for an
5150 exact match, rather than just subject name.
5152 The X509_STORE API doesn't directly support the retrieval
5153 of multiple certificates matching a given criteria, however
5154 this can be worked round by performing a lookup first
5155 (which will fill the cache with candidate certificates)
5156 and then examining the cache for matches. This is probably
5157 the best we can do without throwing out X509_LOOKUP
5158 entirely (maybe later...).
5160 The X509_VERIFY_CTX structure has been enhanced considerably.
5162 All certificate lookup operations now go via a get_issuer()
5163 callback. Although this currently uses an X509_STORE it
5164 can be replaced by custom lookups. This is a simple way
5165 to bypass the X509_STORE hackery necessary to make this
5166 work and makes it possible to use more efficient techniques
5167 in future. A very simple version which uses a simple
5168 STACK for its trusted certificate store is also provided
5169 using X509_STORE_CTX_trusted_stack().
5171 The verify_cb() and verify() callbacks now have equivalents
5172 in the X509_STORE_CTX structure.
5174 X509_STORE_CTX also has a 'flags' field which can be used
5175 to customise the verify behaviour.
5178 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5179 excludes S/MIME capabilities.
5182 *) When a certificate request is read in keep a copy of the
5183 original encoding of the signed data and use it when outputing
5184 again. Signatures then use the original encoding rather than
5185 a decoded, encoded version which may cause problems if the
5186 request is improperly encoded.
5189 *) For consistency with other BIO_puts implementations, call
5190 buffer_write(b, ...) directly in buffer_puts instead of calling
5193 In BIO_puts, increment b->num_write as in BIO_write.
5194 [Peter.Sylvester@EdelWeb.fr]
5196 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5197 BN_zero, we may not return a BIGNUM with an array consisting of
5201 *) Avoid calling abort() from within the library when problems are
5202 detected, except if preprocessor symbols have been defined
5203 (such as REF_CHECK, BN_DEBUG etc.).
5206 *) New openssl application 'rsautl'. This utility can be
5207 used for low level RSA operations. DER public key
5208 BIO/fp routines also added.
5211 *) New Configure entry and patches for compiling on QNX 4.
5212 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5214 *) A demo state-machine implementation was sponsored by
5215 Nuron (http://www.nuron.com/) and is now available in
5216 demos/state_machine.
5219 *) New options added to the 'dgst' utility for signature
5220 generation and verification.
5223 *) Unrecognized PKCS#7 content types are now handled via a
5224 catch all ASN1_TYPE structure. This allows unsupported
5225 types to be stored as a "blob" and an application can
5226 encode and decode it manually.
5229 *) Fix various signed/unsigned issues to make a_strex.c
5231 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5233 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5234 length if passed a buffer. ASN1_INTEGER_to_BN failed
5235 if passed a NULL BN and its argument was negative.
5236 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5238 *) Modification to PKCS#7 encoding routines to output definite
5239 length encoding. Since currently the whole structures are in
5240 memory there's not real point in using indefinite length
5241 constructed encoding. However if OpenSSL is compiled with
5242 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5245 *) Added BIO_vprintf() and BIO_vsnprintf().
5248 *) Added more prefixes to parse for in the the strings written
5249 through a logging bio, to cover all the levels that are available
5250 through syslog. The prefixes are now:
5252 PANIC, EMERG, EMR => LOG_EMERG
5253 ALERT, ALR => LOG_ALERT
5254 CRIT, CRI => LOG_CRIT
5255 ERROR, ERR => LOG_ERR
5256 WARNING, WARN, WAR => LOG_WARNING
5257 NOTICE, NOTE, NOT => LOG_NOTICE
5258 INFO, INF => LOG_INFO
5259 DEBUG, DBG => LOG_DEBUG
5261 and as before, if none of those prefixes are present at the
5262 beginning of the string, LOG_ERR is chosen.
5264 On Win32, the LOG_* levels are mapped according to this:
5266 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5267 LOG_WARNING => EVENTLOG_WARNING_TYPE
5268 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5272 *) Made it possible to reconfigure with just the configuration
5273 argument "reconf" or "reconfigure". The command line arguments
5274 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5275 and are retrieved from there when reconfiguring.
5279 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5281 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5284 *) The obj_dat.pl script was messing up the sorting of object
5285 names. The reason was that it compared the quoted version
5286 of strings as a result "OCSP" > "OCSP Signing" because
5287 " > SPACE. Changed script to store unquoted versions of
5288 names and add quotes on output. It was also omitting some
5289 names from the lookup table if they were given a default
5290 value (that is if SN is missing it is given the same
5291 value as LN and vice versa), these are now added on the
5292 grounds that if an object has a name we should be able to
5293 look it up. Finally added warning output when duplicate
5294 short or long names are found.
5297 *) Changes needed for Tandem NSK.
5298 [Scott Uroff <scott@xypro.com>]
5300 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5301 RSA_padding_check_SSLv23(), special padding was never detected
5302 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5303 version rollback attacks was not effective.
5305 In s23_clnt.c, don't use special rollback-attack detection padding
5306 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5307 client; similarly, in s23_srvr.c, don't do the rollback check if
5308 SSL 2.0 is the only protocol enabled in the server.
5311 *) Make it possible to get hexdumps of unprintable data with 'openssl
5312 asn1parse'. By implication, the functions ASN1_parse_dump() and
5313 BIO_dump_indent() are added.
5316 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5317 these print out strings and name structures based on various
5318 flags including RFC2253 support and proper handling of
5319 multibyte characters. Added options to the 'x509' utility
5320 to allow the various flags to be set.
5323 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5324 Also change the functions X509_cmp_current_time() and
5325 X509_gmtime_adj() work with an ASN1_TIME structure,
5326 this will enable certificates using GeneralizedTime in validity
5327 dates to be checked.
5330 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5331 negative public key encodings) on by default,
5332 NO_NEG_PUBKEY_BUG can be set to disable it.
5335 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5336 content octets. An i2c_ASN1_OBJECT is unnecessary because
5337 the encoding can be trivially obtained from the structure.
5340 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5341 not read locks (CRYPTO_r_[un]lock).
5344 *) A first attempt at creating official support for shared
5345 libraries through configuration. I've kept it so the
5346 default is static libraries only, and the OpenSSL programs
5347 are always statically linked for now, but there are
5348 preparations for dynamic linking in place.
5349 This has been tested on Linux and Tru64.
5352 *) Randomness polling function for Win9x, as described in:
5353 Peter Gutmann, Software Generation of Practically Strong
5357 *) Fix so PRNG is seeded in req if using an already existing
5361 *) New options to smime application. -inform and -outform
5362 allow alternative formats for the S/MIME message including
5363 PEM and DER. The -content option allows the content to be
5364 specified separately. This should allow things like Netscape
5365 form signing output easier to verify.
5368 *) Fix the ASN1 encoding of tags using the 'long form'.
5371 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5372 STRING types. These convert content octets to and from the
5373 underlying type. The actual tag and length octets are
5374 already assumed to have been read in and checked. These
5375 are needed because all other string types have virtually
5376 identical handling apart from the tag. By having versions
5377 of the ASN1 functions that just operate on content octets
5378 IMPLICIT tagging can be handled properly. It also allows
5379 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5380 and ASN1_INTEGER are identical apart from the tag.
5383 *) Change the handling of OID objects as follows:
5385 - New object identifiers are inserted in objects.txt, following
5386 the syntax given in objects.README.
5387 - objects.pl is used to process obj_mac.num and create a new
5389 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5392 This is currently kind of a hack, and the perl code in objects.pl
5393 isn't very elegant, but it works as I intended. The simplest way
5394 to check that it worked correctly is to look in obj_dat.h and
5395 check the array nid_objs and make sure the objects haven't moved
5396 around (this is important!). Additions are OK, as well as
5397 consistent name changes.
5400 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5403 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5404 The given file adds to whatever has already been seeded into the
5405 random pool through the RANDFILE configuration file option or
5406 environment variable, or the default random state file.
5409 *) mkstack.pl now sorts each macro group into lexical order.
5410 Previously the output order depended on the order the files
5411 appeared in the directory, resulting in needless rewriting
5415 *) Patches to make OpenSSL compile under Win32 again. Mostly
5416 work arounds for the VC++ problem that it treats func() as
5417 func(void). Also stripped out the parts of mkdef.pl that
5418 added extra typesafe functions: these no longer exist.
5421 *) Reorganisation of the stack code. The macros are now all
5422 collected in safestack.h . Each macro is defined in terms of
5423 a "stack macro" of the form SKM_<name>(type, a, b). The
5424 DEBUG_SAFESTACK is now handled in terms of function casts,
5425 this has the advantage of retaining type safety without the
5426 use of additional functions. If DEBUG_SAFESTACK is not defined
5427 then the non typesafe macros are used instead. Also modified the
5428 mkstack.pl script to handle the new form. Needs testing to see
5429 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5430 the default if no major problems. Similar behaviour for ASN1_SET_OF
5431 and PKCS12_STACK_OF.
5434 *) When some versions of IIS use the 'NET' form of private key the
5435 key derivation algorithm is different. Normally MD5(password) is
5436 used as a 128 bit RC4 key. In the modified case
5437 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5438 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5439 as the old Netscape_RSA functions except they have an additional
5440 'sgckey' parameter which uses the modified algorithm. Also added
5441 an -sgckey command line option to the rsa utility. Thanks to
5442 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5443 algorithm to openssl-dev.
5446 *) The evp_local.h macros were using 'c.##kname' which resulted in
5447 invalid expansion on some systems (SCO 5.0.5 for example).
5448 Corrected to 'c.kname'.
5449 [Phillip Porch <root@theporch.com>]
5451 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5452 a STACK of email addresses from a certificate or request, these look
5453 in the subject name and the subject alternative name extensions and
5454 omit any duplicate addresses.
5457 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5458 This makes DSA verification about 2 % faster.
5461 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5462 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5463 plus overhead for 1024 bit moduli).
5464 This makes exponentiations about 0.5 % faster for 1024 bit
5465 exponents (as measured by "openssl speed rsa2048").
5468 *) Rename memory handling macros to avoid conflicts with other
5470 Malloc => OPENSSL_malloc
5471 Malloc_locked => OPENSSL_malloc_locked
5472 Realloc => OPENSSL_realloc
5473 Free => OPENSSL_free
5476 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5477 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5480 *) CygWin32 support.
5481 [John Jarvie <jjarvie@newsguy.com>]
5483 *) The type-safe stack code has been rejigged. It is now only compiled
5484 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5485 by default all type-specific stack functions are "#define"d back to
5486 standard stack functions. This results in more streamlined output
5487 but retains the type-safety checking possibilities of the original
5491 *) The STACK code has been cleaned up, and certain type declarations
5492 that didn't make a lot of sense have been brought in line. This has
5493 also involved a cleanup of sorts in safestack.h to more correctly
5494 map type-safe stack functions onto their plain stack counterparts.
5495 This work has also resulted in a variety of "const"ifications of
5496 lots of the code, especially "_cmp" operations which should normally
5497 be prototyped with "const" parameters anyway.
5500 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5501 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5502 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5503 where all of 'md' is used each time the PRNG is used, but 'state'
5504 is used only indexed by a cyclic counter. As entropy may not be
5505 well distributed from the beginning, 'md' is important as a
5506 chaining variable. However, the output function chains only half
5507 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5508 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5509 in all of 'state' being rewritten, with the new values depending
5510 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5513 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5514 the handshake is continued after ssl_verify_cert_chain();
5515 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5516 can lead to 'unexplainable' connection aborts later.
5517 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5519 *) Major EVP API cipher revision.
5520 Add hooks for extra EVP features. This allows various cipher
5521 parameters to be set in the EVP interface. Support added for variable
5522 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5523 setting of RC2 and RC5 parameters.
5525 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5528 Remove lots of duplicated code from the EVP library. For example *every*
5529 cipher init() function handles the 'iv' in the same way according to the
5530 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5531 for CFB and OFB modes they zero ctx->num.
5533 New functionality allows removal of S/MIME code RC2 hack.
5535 Most of the routines have the same form and so can be declared in terms
5538 By shifting this to the top level EVP_CipherInit() it can be removed from
5539 all individual ciphers. If the cipher wants to handle IVs or keys
5540 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5543 Change lots of functions like EVP_EncryptUpdate() to now return a
5544 value: although software versions of the algorithms cannot fail
5545 any installed hardware versions can.
5548 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5549 this option is set, tolerate broken clients that send the negotiated
5550 protocol version number instead of the requested protocol version
5554 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5555 i.e. non-zero for export ciphersuites, zero otherwise.
5556 Previous versions had this flag inverted, inconsistent with
5557 rsa_tmp_cb (..._TMP_RSA_CB).
5558 [Bodo Moeller; problem reported by Amit Chopra]
5560 *) Add missing DSA library text string. Work around for some IIS
5561 key files with invalid SEQUENCE encoding.
5564 *) Add a document (doc/standards.txt) that list all kinds of standards
5565 and so on that are implemented in OpenSSL.
5568 *) Enhance c_rehash script. Old version would mishandle certificates
5569 with the same subject name hash and wouldn't handle CRLs at all.
5570 Added -fingerprint option to crl utility, to support new c_rehash
5574 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5577 *) Fix for SSL server purpose checking. Server checking was
5578 rejecting certificates which had extended key usage present
5579 but no ssl client purpose.
5580 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5582 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5583 is a little unclear about how a blank password is handled.
5584 Since the password in encoded as a BMPString with terminating
5585 double NULL a zero length password would end up as just the
5586 double NULL. However no password at all is different and is
5587 handled differently in the PKCS#12 key generation code. NS
5588 treats a blank password as zero length. MSIE treats it as no
5589 password on export: but it will try both on import. We now do
5590 the same: PKCS12_parse() tries zero length and no password if
5591 the password is set to "" or NULL (NULL is now a valid password:
5592 it wasn't before) as does the pkcs12 application.
5595 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5596 perror when PEM_read_bio_X509_REQ fails, the error message must
5597 be obtained from the error queue.
5600 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5601 it in ERR_remove_state if appropriate, and change ERR_get_state
5602 accordingly to avoid race conditions (this is necessary because
5603 thread_hash is no longer constant once set).
5606 *) Bugfix for linux-elf makefile.one.
5609 *) RSA_get_default_method() will now cause a default
5610 RSA_METHOD to be chosen if one doesn't exist already.
5611 Previously this was only set during a call to RSA_new()
5612 or RSA_new_method(NULL) meaning it was possible for
5613 RSA_get_default_method() to return NULL.
5616 *) Added native name translation to the existing DSO code
5617 that will convert (if the flag to do so is set) filenames
5618 that are sufficiently small and have no path information
5619 into a canonical native form. Eg. "blah" converted to
5620 "libblah.so" or "blah.dll" etc.
5623 *) New function ERR_error_string_n(e, buf, len) which is like
5624 ERR_error_string(e, buf), but writes at most 'len' bytes
5625 including the 0 terminator. For ERR_error_string_n, 'buf'
5627 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5629 *) CONF library reworked to become more general. A new CONF
5630 configuration file reader "class" is implemented as well as a
5631 new functions (NCONF_*, for "New CONF") to handle it. The now
5632 old CONF_* functions are still there, but are reimplemented to
5633 work in terms of the new functions. Also, a set of functions
5634 to handle the internal storage of the configuration data is
5635 provided to make it easier to write new configuration file
5636 reader "classes" (I can definitely see something reading a
5637 configuration file in XML format, for example), called _CONF_*,
5638 or "the configuration storage API"...
5640 The new configuration file reading functions are:
5642 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5643 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5645 NCONF_default, NCONF_WIN32
5647 NCONF_dump_fp, NCONF_dump_bio
5649 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5650 NCONF_new creates a new CONF object. This works in the same way
5651 as other interfaces in OpenSSL, like the BIO interface.
5652 NCONF_dump_* dump the internal storage of the configuration file,
5653 which is useful for debugging. All other functions take the same
5654 arguments as the old CONF_* functions wth the exception of the
5655 first that must be a `CONF *' instead of a `LHASH *'.
5657 To make it easer to use the new classes with the old CONF_* functions,
5658 the function CONF_set_default_method is provided.
5661 *) Add '-tls1' option to 'openssl ciphers', which was already
5662 mentioned in the documentation but had not been implemented.
5663 (This option is not yet really useful because even the additional
5664 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5667 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5668 OpenSSL-based applications) load shared libraries and bind to
5669 them in a portable way.
5670 [Geoff Thorpe, with contributions from Richard Levitte]
5672 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5674 *) Make sure _lrotl and _lrotr are only used with MSVC.
5676 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5677 (the default implementation of RAND_status).
5679 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5680 to '-clrext' (= clear extensions), as intended and documented.
5681 [Bodo Moeller; inconsistency pointed out by Michael Attili
5684 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5685 was larger than the MD block size.
5686 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5688 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5689 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5690 using the passed key: if the passed key was a private key the result
5691 of X509_print(), for example, would be to print out all the private key
5695 *) des_quad_cksum() byte order bug fix.
5696 [Ulf Möller, using the problem description in krb4-0.9.7, where
5697 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5699 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5701 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5703 *) For easily testing in shell scripts whether some command
5704 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5705 returns with exit code 0 iff no command of the given name is available.
5706 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5707 the output goes to stdout and nothing is printed to stderr.
5708 Additional arguments are always ignored.
5710 Since for each cipher there is a command of the same name,
5711 the 'no-cipher' compilation switches can be tested this way.
5713 ('openssl no-XXX' is not able to detect pseudo-commands such
5714 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5717 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5720 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5721 is set; it will be thrown away anyway because each handshake creates
5723 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5724 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5725 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5726 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5729 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5730 'Q' and 'R' lose their special meanings (quit/renegotiate).
5731 This is part of what -quiet does; unlike -quiet, -ign_eof
5732 does not suppress any output.
5735 *) Add compatibility options to the purpose and trust code. The
5736 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5737 accepts a certificate or CA, this was the previous behaviour,
5738 with all the associated security issues.
5740 X509_TRUST_COMPAT is the old trust behaviour: only and
5741 automatically trust self signed roots in certificate store. A
5742 new trust setting X509_TRUST_DEFAULT is used to specify that
5743 a purpose has no associated trust setting and it should instead
5744 use the value in the default purpose.
5747 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5748 and fix a memory leak.
5751 *) In util/mkerr.pl (which implements 'make errors'), preserve
5752 reason strings from the previous version of the .c file, as
5753 the default to have only downcase letters (and digits) in
5754 automatically generated reasons codes is not always appropriate.
5757 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5758 using strerror. Previously, ERR_reason_error_string() returned
5759 library names as reason strings for SYSerr; but SYSerr is a special
5760 case where small numbers are errno values, not library numbers.
5763 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5764 converts DSA parameters into DH parameters. (When creating parameters,
5765 DSA_generate_parameters is used.)
5768 *) Include 'length' (recommended exponent length) in C code generated
5769 by 'openssl dhparam -C'.
5772 *) The second argument to set_label in perlasm was already being used
5773 so couldn't be used as a "file scope" flag. Moved to third argument
5777 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5778 instead of RAND_bytes for encryption IVs and salts.
5781 *) Include RAND_status() into RAND_METHOD instead of implementing
5782 it only for md_rand.c Otherwise replacing the PRNG by calling
5783 RAND_set_rand_method would be impossible.
5786 *) Don't let DSA_generate_key() enter an infinite loop if the random
5787 number generation fails.
5790 *) New 'rand' application for creating pseudo-random output.
5793 *) Added configuration support for Linux/IA64
5794 [Rolf Haberrecker <rolf@suse.de>]
5796 *) Assembler module support for Mingw32.
5799 *) Shared library support for HPUX (in shlib/).
5800 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5802 *) Shared library support for Solaris gcc.
5803 [Lutz Behnke <behnke@trustcenter.de>]
5805 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5807 *) PKCS7_encrypt() was adding text MIME headers twice because they
5808 were added manually and by SMIME_crlf_copy().
5811 *) In bntest.c don't call BN_rand with zero bits argument.
5812 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5814 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5815 case was implemented. This caused BN_div_recp() to fail occasionally.
5818 *) Add an optional second argument to the set_label() in the perl
5819 assembly language builder. If this argument exists and is set
5820 to 1 it signals that the assembler should use a symbol whose
5821 scope is the entire file, not just the current function. This
5822 is needed with MASM which uses the format label:: for this scope.
5823 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5825 *) Change the ASN1 types so they are typedefs by default. Before
5826 almost all types were #define'd to ASN1_STRING which was causing
5827 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5831 *) Change names of new functions to the new get1/get0 naming
5832 convention: After 'get1', the caller owns a reference count
5833 and has to call ..._free; 'get0' returns a pointer to some
5834 data structure without incrementing reference counters.
5835 (Some of the existing 'get' functions increment a reference
5836 counter, some don't.)
5837 Similarly, 'set1' and 'add1' functions increase reference
5838 counters or duplicate objects.
5841 *) Allow for the possibility of temp RSA key generation failure:
5842 the code used to assume it always worked and crashed on failure.
5845 *) Fix potential buffer overrun problem in BIO_printf().
5846 [Ulf Möller, using public domain code by Patrick Powell; problem
5847 pointed out by David Sacerdote <das33@cornell.edu>]
5849 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5850 RAND_egd() and RAND_status(). In the command line application,
5851 the EGD socket can be specified like a seed file using RANDFILE
5855 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5856 Some CAs (e.g. Verisign) distribute certificates in this form.
5859 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5860 list to exclude them. This means that no special compilation option
5861 is needed to use anonymous DH: it just needs to be included in the
5865 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5866 EVP_MD_type. The old functionality is available in a new macro called
5867 EVP_MD_md(). Change code that uses it and update docs.
5870 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5871 where the 'void *' argument is replaced by a function pointer argument.
5872 Previously 'void *' was abused to point to functions, which works on
5873 many platforms, but is not correct. As these functions are usually
5874 called by macros defined in OpenSSL header files, most source code
5875 should work without changes.
5878 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5879 sections with information on -D... compiler switches used for
5880 compiling the library so that applications can see them. To enable
5881 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5882 must be defined. E.g.,
5883 #define OPENSSL_ALGORITHM_DEFINES
5884 #include <openssl/opensslconf.h>
5885 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5886 [Richard Levitte, Ulf and Bodo Möller]
5888 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5892 *) Change the 'other' type in certificate aux info to a STACK_OF
5893 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5894 the required ASN1 format: arbitrary types determined by an OID.
5897 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5898 argument to 'req'. This is not because the function is newer or
5899 better than others it just uses the work 'NEW' in the certificate
5900 request header lines. Some software needs this.
5903 *) Reorganise password command line arguments: now passwords can be
5904 obtained from various sources. Delete the PEM_cb function and make
5905 it the default behaviour: i.e. if the callback is NULL and the
5906 usrdata argument is not NULL interpret it as a null terminated pass
5907 phrase. If usrdata and the callback are NULL then the pass phrase
5908 is prompted for as usual.
5911 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5912 the support is automatically enabled. The resulting binaries will
5913 autodetect the card and use it if present.
5914 [Ben Laurie and Compaq Inc.]
5916 *) Work around for Netscape hang bug. This sends certificate request
5917 and server done in one record. Since this is perfectly legal in the
5918 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5919 the bugs/SSLv3 entry for more info.
5922 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5925 *) Add -rand argument to smime and pkcs12 applications and read/write
5929 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5932 *) Add command line password options to the remaining applications.
5935 *) Bug fix for BN_div_recp() for numerators with an even number of
5939 *) More tests in bntest.c, and changed test_bn output.
5942 *) ./config recognizes MacOS X now.
5945 *) Bug fix for BN_div() when the first words of num and divsor are
5946 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5949 *) Add support for various broken PKCS#8 formats, and command line
5950 options to produce them.
5953 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5954 get temporary BIGNUMs from a BN_CTX.
5957 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5961 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5962 include a #define from the old name to the new. The original intent
5963 was that statically linked binaries could for example just call
5964 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5965 link with digests. This never worked becayse SSLeay_add_all_digests()
5966 and SSLeay_add_all_ciphers() were in the same source file so calling
5967 one would link with the other. They are now in separate source files.
5970 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5973 *) Use a less unusual form of the Miller-Rabin primality test (it used
5974 a binary algorithm for exponentiation integrated into the Miller-Rabin
5975 loop, our standard modexp algorithms are faster).
5978 *) Support for the EBCDIC character set completed.
5979 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5981 *) Source code cleanups: use const where appropriate, eliminate casts,
5982 use void * instead of char * in lhash.
5985 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5986 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5987 this the server could overwrite ephemeral keys that the client
5991 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5992 using 50 iterations of the Rabin-Miller test.
5994 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5995 iterations of the Rabin-Miller test as required by the appendix
5996 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5997 As BN_is_prime_fasttest includes trial division, DSA parameter
5998 generation becomes much faster.
6000 This implies a change for the callback functions in DSA_is_prime
6001 and DSA_generate_parameters: The callback function is called once
6002 for each positive witness in the Rabin-Miller test, not just
6003 occasionally in the inner loop; and the parameters to the
6004 callback function now provide an iteration count for the outer
6005 loop rather than for the current invocation of the inner loop.
6006 DSA_generate_parameters additionally can call the callback
6007 function with an 'iteration count' of -1, meaning that a
6008 candidate has passed the trial division test (when q is generated
6009 from an application-provided seed, trial division is skipped).
6012 *) New function BN_is_prime_fasttest that optionally does trial
6013 division before starting the Rabin-Miller test and has
6014 an additional BN_CTX * argument (whereas BN_is_prime always
6015 has to allocate at least one BN_CTX).
6016 'callback(1, -1, cb_arg)' is called when a number has passed the
6017 trial division stage.
6020 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6024 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6027 *) New function BN_pseudo_rand().
6030 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6031 bignum version of BN_from_montgomery() with the working code from
6032 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6036 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6037 made it impossible to use the same SSL_SESSION data structure in
6038 SSL2 clients in multiple threads.
6041 *) The return value of RAND_load_file() no longer counts bytes obtained
6042 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6043 to seed the PRNG (previously an explicit byte count was required).
6044 [Ulf Möller, Bodo Möller]
6046 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6047 used (char *) instead of (void *) and had casts all over the place.
6050 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6053 *) Retain source code compatibility for BN_prime_checks macro:
6054 BN_is_prime(..., BN_prime_checks, ...) now uses
6055 BN_prime_checks_for_size to determine the appropriate number of
6056 Rabin-Miller iterations.
6059 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6060 DH_CHECK_P_NOT_SAFE_PRIME.
6061 (Check if this is true? OpenPGP calls them "strong".)
6064 *) Merge the functionality of "dh" and "gendh" programs into a new program
6065 "dhparam". The old programs are retained for now but will handle DH keys
6066 (instead of parameters) in future.
6069 *) Make the ciphers, s_server and s_client programs check the return values
6070 when a new cipher list is set.
6073 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6074 ciphers. Before when the 56bit ciphers were enabled the sorting was
6077 The syntax for the cipher sorting has been extended to support sorting by
6078 cipher-strength (using the strength_bits hard coded in the tables).
6079 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6081 Fix a bug in the cipher-command parser: when supplying a cipher command
6082 string with an "undefined" symbol (neither command nor alphanumeric
6083 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6084 an error is flagged.
6086 Due to the strength-sorting extension, the code of the
6087 ssl_create_cipher_list() function was completely rearranged. I hope that
6088 the readability was also increased :-)
6089 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6091 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6092 for the first serial number and places 2 in the serial number file. This
6093 avoids problems when the root CA is created with serial number zero and
6094 the first user certificate has the same issuer name and serial number
6098 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6099 the new code. Add documentation for this stuff.
6102 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6103 X509_*() to X509at_*() on the grounds that they don't handle X509
6104 structures and behave in an analagous way to the X509v3 functions:
6105 they shouldn't be called directly but wrapper functions should be used
6108 So we also now have some wrapper functions that call the X509at functions
6109 when passed certificate requests. (TO DO: similar things can be done with
6110 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6111 things. Some of these need some d2i or i2d and print functionality
6112 because they handle more complex structures.)
6115 *) Add missing #ifndefs that caused missing symbols when building libssl
6116 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6117 NO_RSA in ssl/s2*.c.
6118 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6120 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6121 has a return value which indicates the quality of the random data
6122 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6123 error queue. New function RAND_pseudo_bytes() generates output that is
6124 guaranteed to be unique but not unpredictable. RAND_add is like
6125 RAND_seed, but takes an extra argument for an entropy estimate
6126 (RAND_seed always assumes full entropy).
6129 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6130 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6131 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6132 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6133 false-positive rate of at most 2^-80 for random input.
6136 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6139 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6140 in the 0.9.5 release), this returns the chain
6141 from an X509_CTX structure with a dup of the stack and all
6142 the X509 reference counts upped: so the stack will exist
6143 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6146 Also make SSL_SESSION_print() print out the verify return
6150 *) Add manpage for the pkcs12 command. Also change the default
6151 behaviour so MAC iteration counts are used unless the new
6152 -nomaciter option is used. This improves file security and
6153 only older versions of MSIE (4.0 for example) need it.
6156 *) Honor the no-xxx Configure options when creating .DEF files.
6159 *) Add PKCS#10 attributes to field table: challengePassword,
6160 unstructuredName and unstructuredAddress. These are taken from
6161 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6162 international characters are used.
6164 More changes to X509_ATTRIBUTE code: allow the setting of types
6165 based on strings. Remove the 'loc' parameter when adding
6166 attributes because these will be a SET OF encoding which is sorted
6170 *) Initial changes to the 'req' utility to allow request generation
6171 automation. This will allow an application to just generate a template
6172 file containing all the field values and have req construct the
6175 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6176 used all over the place including certificate requests and PKCS#7
6177 structures. They are currently handled manually where necessary with
6178 some primitive wrappers for PKCS#7. The new functions behave in a
6179 manner analogous to the X509 extension functions: they allow
6180 attributes to be looked up by NID and added.
6182 Later something similar to the X509V3 code would be desirable to
6183 automatically handle the encoding, decoding and printing of the
6184 more complex types. The string types like challengePassword can
6185 be handled by the string table functions.
6187 Also modified the multi byte string table handling. Now there is
6188 a 'global mask' which masks out certain types. The table itself
6189 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6190 is useful when for example there is only one permissible type
6191 (as in countryName) and using the mask might result in no valid
6195 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6196 SSL_get_peer_finished to allow applications to obtain the latest
6197 Finished messages sent to the peer or expected from the peer,
6198 respectively. (SSL_get_peer_finished is usually the Finished message
6199 actually received from the peer, otherwise the protocol will be aborted.)
6201 As the Finished message are message digests of the complete handshake
6202 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6203 be used for external authentication procedures when the authentication
6204 provided by SSL/TLS is not desired or is not enough.
6207 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6208 the host supports BWX extension and if Compaq C is present on the
6209 $PATH. Just exploiting of the BWX extension results in 20-30%
6210 performance kick for some algorithms, e.g. DES and RC4 to mention
6211 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6215 *) Add support for MS "fast SGC". This is arguably a violation of the
6216 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6217 weak crypto and after checking the certificate is SGC a second one
6218 with strong crypto. MS SGC stops the first handshake after receiving
6219 the server certificate message and sends a second client hello. Since
6220 a server will typically do all the time consuming operations before
6221 expecting any further messages from the client (server key exchange
6222 is the most expensive) there is little difference between the two.
6224 To get OpenSSL to support MS SGC we have to permit a second client
6225 hello message after we have sent server done. In addition we have to
6226 reset the MAC if we do get this second client hello.
6229 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6230 if a DER encoded private key is RSA or DSA traditional format. Changed
6231 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6232 format DER encoded private key. Newer code should use PKCS#8 format which
6233 has the key type encoded in the ASN1 structure. Added DER private key
6234 support to pkcs8 application.
6237 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6238 ciphersuites has been selected (as required by the SSL 3/TLS 1
6239 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6240 is set, we interpret this as a request to violate the specification
6241 (the worst that can happen is a handshake failure, and 'correct'
6242 behaviour would result in a handshake failure anyway).
6245 *) In SSL_CTX_add_session, take into account that there might be multiple
6246 SSL_SESSION structures with the same session ID (e.g. when two threads
6247 concurrently obtain them from an external cache).
6248 The internal cache can handle only one SSL_SESSION with a given ID,
6249 so if there's a conflict, we now throw out the old one to achieve
6253 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6254 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6255 some routines that use cipher OIDs: some ciphers do not have OIDs
6256 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6260 *) Simplify the trust setting structure and code. Now we just have
6261 two sequences of OIDs for trusted and rejected settings. These will
6262 typically have values the same as the extended key usage extension
6263 and any application specific purposes.
6265 The trust checking code now has a default behaviour: it will just
6266 check for an object with the same NID as the passed id. Functions can
6267 be provided to override either the default behaviour or the behaviour
6268 for a given id. SSL client, server and email already have functions
6269 in place for compatibility: they check the NID and also return "trusted"
6270 if the certificate is self signed.
6273 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6274 traditional format into an EVP_PKEY structure.
6277 *) Add a password callback function PEM_cb() which either prompts for
6278 a password if usr_data is NULL or otherwise assumes it is a null
6279 terminated password. Allow passwords to be passed on command line
6280 environment or config files in a few more utilities.
6283 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6284 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6285 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6286 Update documentation.
6289 *) Support for ASN1 "NULL" type. This could be handled before by using
6290 ASN1_TYPE but there wasn't any function that would try to read a NULL
6291 and produce an error if it couldn't. For compatibility we also have
6292 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6293 don't allocate anything because they don't need to.
6296 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6298 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6300 *) Rebuild of the memory allocation routines used by OpenSSL code and
6301 possibly others as well. The purpose is to make an interface that
6302 provide hooks so anyone can build a separate set of allocation and
6303 deallocation routines to be used by OpenSSL, for example memory
6304 pool implementations, or something else, which was previously hard
6305 since Malloc(), Realloc() and Free() were defined as macros having
6306 the values malloc, realloc and free, respectively (except for Win32
6307 compilations). The same is provided for memory debugging code.
6308 OpenSSL already comes with functionality to find memory leaks, but
6309 this gives people a chance to debug other memory problems.
6311 With these changes, a new set of functions and macros have appeared:
6313 CRYPTO_set_mem_debug_functions() [F]
6314 CRYPTO_get_mem_debug_functions() [F]
6315 CRYPTO_dbg_set_options() [F]
6316 CRYPTO_dbg_get_options() [F]
6317 CRYPTO_malloc_debug_init() [M]
6319 The memory debug functions are NULL by default, unless the library
6320 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6321 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6322 gives the standard debugging functions that come with OpenSSL) or
6323 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6324 provided by the library user) must be used. When the standard
6325 debugging functions are used, CRYPTO_dbg_set_options can be used to
6326 request additional information:
6327 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6328 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6330 Also, things like CRYPTO_set_mem_functions will always give the
6331 expected result (the new set of functions is used for allocation
6332 and deallocation) at all times, regardless of platform and compiler
6335 To finish it up, some functions that were never use in any other
6336 way than through macros have a new API and new semantic:
6339 CRYPTO_dbg_realloc()
6342 All macros of value have retained their old syntax.
6343 [Richard Levitte and Bodo Moeller]
6345 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6346 ordering of SMIMECapabilities wasn't in "strength order" and there
6347 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6351 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6352 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6353 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6355 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6356 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6357 functionality to handle multipart/signed properly) and a utility
6358 called 'smime' to call all this stuff. This is based on code I
6359 originally wrote for Celo who have kindly allowed it to be
6360 included in OpenSSL.
6363 *) Add variants des_set_key_checked and des_set_key_unchecked of
6364 des_set_key (aka des_key_sched). Global variable des_check_key
6365 decides which of these is called by des_set_key; this way
6366 des_check_key behaves as it always did, but applications and
6367 the library itself, which was buggy for des_check_key == 1,
6368 have a cleaner way to pick the version they need.
6371 *) New function PKCS12_newpass() which changes the password of a
6375 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6376 dynamic mix. In both cases the ids can be used as an index into the
6377 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6378 functions so they accept a list of the field values and the
6379 application doesn't need to directly manipulate the X509_TRUST
6383 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6387 *) Modify the way the V3 extension code looks up extensions. This now
6388 works in a similar way to the object code: we have some "standard"
6389 extensions in a static table which is searched with OBJ_bsearch()
6390 and the application can add dynamic ones if needed. The file
6391 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6392 updated whenever a new extension is added to the core code and kept
6393 in ext_nid order. There is a simple program 'tabtest.c' which checks
6394 this. New extensions are not added too often so this file can readily
6395 be maintained manually.
6397 There are two big advantages in doing things this way. The extensions
6398 can be looked up immediately and no longer need to be "added" using
6399 X509V3_add_standard_extensions(): this function now does nothing.
6400 [Side note: I get *lots* of email saying the extension code doesn't
6401 work because people forget to call this function]
6402 Also no dynamic allocation is done unless new extensions are added:
6403 so if we don't add custom extensions there is no need to call
6404 X509V3_EXT_cleanup().
6407 *) Modify enc utility's salting as follows: make salting the default. Add a
6408 magic header, so unsalted files fail gracefully instead of just decrypting
6409 to garbage. This is because not salting is a big security hole, so people
6410 should be discouraged from doing it.
6413 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6414 digest to be passed on the command line but it only used this
6415 parameter when signing a certificate. Modified so all relevant
6416 operations are affected by the digest parameter including the
6417 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6418 DSA key was used because it didn't fix the digest.
6421 *) Initial certificate chain verify code. Currently tests the untrusted
6422 certificates for consistency with the verify purpose (which is set
6423 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6425 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6426 this is because it will reject chains with invalid extensions whereas
6427 every previous version of OpenSSL and SSLeay made no checks at all.
6429 Trust code: checks the root CA for the relevant trust settings. Trust
6430 settings have an initial value consistent with the verify purpose: e.g.
6431 if the verify purpose is for SSL client use it expects the CA to be
6432 trusted for SSL client use. However the default value can be changed to
6433 permit custom trust settings: one example of this would be to only trust
6434 certificates from a specific "secure" set of CAs.
6436 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6437 which should be used for version portability: especially since the
6438 verify structure is likely to change more often now.
6440 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6441 to set them. If not set then assume SSL clients will verify SSL servers
6444 Two new options to the verify program: -untrusted allows a set of
6445 untrusted certificates to be passed in and -purpose which sets the
6446 intended purpose of the certificate. If a purpose is set then the
6447 new chain verify code is used to check extension consistency.
6450 *) Support for the authority information access extension.
6453 *) Modify RSA and DSA PEM read routines to transparently handle
6454 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6455 public keys in a format compatible with certificate
6456 SubjectPublicKeyInfo structures. Unfortunately there were already
6457 functions called *_PublicKey_* which used various odd formats so
6458 these are retained for compatibility: however the DSA variants were
6459 never in a public release so they have been deleted. Changed dsa/rsa
6460 utilities to handle the new format: note no releases ever handled public
6461 keys so we should be OK.
6463 The primary motivation for this change is to avoid the same fiasco
6464 that dogs private keys: there are several incompatible private key
6465 formats some of which are standard and some OpenSSL specific and
6466 require various evil hacks to allow partial transparent handling and
6467 even then it doesn't work with DER formats. Given the option anything
6468 other than PKCS#8 should be dumped: but the other formats have to
6469 stay in the name of compatibility.
6471 With public keys and the benefit of hindsight one standard format
6472 is used which works with EVP_PKEY, RSA or DSA structures: though
6473 it clearly returns an error if you try to read the wrong kind of key.
6475 Added a -pubkey option to the 'x509' utility to output the public key.
6476 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6477 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6478 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6479 that do the same as the EVP_PKEY_assign_*() except they up the
6480 reference count of the added key (they don't "swallow" the
6484 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6485 CRLs would fail if the file contained no certificates or no CRLs:
6486 added a new function to read in both types and return the number
6487 read: this means that if none are read it will be an error. The
6488 DER versions of the certificate and CRL reader would always fail
6489 because it isn't possible to mix certificates and CRLs in DER format
6490 without choking one or the other routine. Changed this to just read
6491 a certificate: this is the best we can do. Also modified the code
6492 in apps/verify.c to take notice of return codes: it was previously
6493 attempting to read in certificates from NULL pointers and ignoring
6494 any errors: this is one reason why the cert and CRL reader seemed
6495 to work. It doesn't check return codes from the default certificate
6496 routines: these may well fail if the certificates aren't installed.
6499 *) Code to support otherName option in GeneralName.
6502 *) First update to verify code. Change the verify utility
6503 so it warns if it is passed a self signed certificate:
6504 for consistency with the normal behaviour. X509_verify
6505 has been modified to it will now verify a self signed
6506 certificate if *exactly* the same certificate appears
6507 in the store: it was previously impossible to trust a
6508 single self signed certificate. This means that:
6509 openssl verify ss.pem
6510 now gives a warning about a self signed certificate but
6511 openssl verify -CAfile ss.pem ss.pem
6515 *) For servers, store verify_result in SSL_SESSION data structure
6516 (and add it to external session representation).
6517 This is needed when client certificate verifications fails,
6518 but an application-provided verification callback (set by
6519 SSL_CTX_set_cert_verify_callback) allows accepting the session
6520 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6521 but returns 1): When the session is reused, we have to set
6522 ssl->verify_result to the appropriate error code to avoid
6524 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6526 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6527 case in PKCS7_dataInit() where the signed PKCS7 structure
6528 didn't contain any existing data because it was being created.
6529 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6531 *) Add a salt to the key derivation routines in enc.c. This
6532 forms the first 8 bytes of the encrypted file. Also add a
6533 -S option to allow a salt to be input on the command line.
6536 *) New function X509_cmp(). Oddly enough there wasn't a function
6537 to compare two certificates. We do this by working out the SHA1
6538 hash and comparing that. X509_cmp() will be needed by the trust
6542 *) SSL_get1_session() is like SSL_get_session(), but increments
6543 the reference count in the SSL_SESSION returned.
6544 [Geoff Thorpe <geoff@eu.c2.net>]
6546 *) Fix for 'req': it was adding a null to request attributes.
6547 Also change the X509_LOOKUP and X509_INFO code to handle
6548 certificate auxiliary information.
6551 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6555 *) Add the possibility to add extra information to the memory leak
6556 detecting output, to form tracebacks, showing from where each
6557 allocation was originated: CRYPTO_push_info("constant string") adds
6558 the string plus current file name and line number to a per-thread
6559 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6560 is like calling CYRPTO_pop_info() until the stack is empty.
6561 Also updated memory leak detection code to be multi-thread-safe.
6564 *) Add options -text and -noout to pkcs7 utility and delete the
6565 encryption options which never did anything. Update docs.
6568 *) Add options to some of the utilities to allow the pass phrase
6569 to be included on either the command line (not recommended on
6570 OSes like Unix) or read from the environment. Update the
6571 manpages and fix a few bugs.
6574 *) Add a few manpages for some of the openssl commands.
6577 *) Fix the -revoke option in ca. It was freeing up memory twice,
6578 leaking and not finding already revoked certificates.
6581 *) Extensive changes to support certificate auxiliary information.
6582 This involves the use of X509_CERT_AUX structure and X509_AUX
6583 functions. An X509_AUX function such as PEM_read_X509_AUX()
6584 can still read in a certificate file in the usual way but it
6585 will also read in any additional "auxiliary information". By
6586 doing things this way a fair degree of compatibility can be
6587 retained: existing certificates can have this information added
6588 using the new 'x509' options.
6590 Current auxiliary information includes an "alias" and some trust
6591 settings. The trust settings will ultimately be used in enhanced
6592 certificate chain verification routines: currently a certificate
6593 can only be trusted if it is self signed and then it is trusted
6597 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6598 The problem was that one of the replacement routines had not been working
6599 since SSLeay releases. For now the offending routine has been replaced
6600 with non-optimised assembler. Even so, this now gives around 95%
6601 performance improvement for 1024 bit RSA signs.
6604 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6605 handling. Most clients have the effective key size in bits equal to
6606 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6607 A few however don't do this and instead use the size of the decrypted key
6608 to determine the RC2 key length and the AlgorithmIdentifier to determine
6609 the effective key length. In this case the effective key length can still
6610 be 40 bits but the key length can be 168 bits for example. This is fixed
6611 by manually forcing an RC2 key into the EVP_PKEY structure because the
6612 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6613 the key length and effective key length are equal.
6616 *) Add a bunch of functions that should simplify the creation of
6617 X509_NAME structures. Now you should be able to do:
6618 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6619 and have it automatically work out the correct field type and fill in
6620 the structures. The more adventurous can try:
6621 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6622 and it will (hopefully) work out the correct multibyte encoding.
6625 *) Change the 'req' utility to use the new field handling and multibyte
6626 copy routines. Before the DN field creation was handled in an ad hoc
6627 way in req, ca, and x509 which was rather broken and didn't support
6628 BMPStrings or UTF8Strings. Since some software doesn't implement
6629 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6630 using the dirstring_type option. See the new comment in the default
6631 openssl.cnf for more info.
6634 *) Make crypto/rand/md_rand.c more robust:
6635 - Assure unique random numbers after fork().
6636 - Make sure that concurrent threads access the global counter and
6637 md serializably so that we never lose entropy in them
6638 or use exactly the same state in multiple threads.
6639 Access to the large state is not always serializable because
6640 the additional locking could be a performance killer, and
6641 md should be large enough anyway.
6644 *) New file apps/app_rand.c with commonly needed functionality
6645 for handling the random seed file.
6647 Use the random seed file in some applications that previously did not:
6649 dsaparam -genkey (which also ignored its '-rand' option),
6652 x509 (when signing).
6653 Except on systems with /dev/urandom, it is crucial to have a random
6654 seed file at least for key creation, DSA signing, and for DH exchanges;
6655 for RSA signatures we could do without one.
6657 gendh and gendsa (unlike genrsa) used to read only the first byte
6658 of each file listed in the '-rand' option. The function as previously
6659 found in genrsa is now in app_rand.c and is used by all programs
6660 that support '-rand'.
6663 *) In RAND_write_file, use mode 0600 for creating files;
6664 don't just chmod when it may be too late.
6667 *) Report an error from X509_STORE_load_locations
6668 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6671 *) New function ASN1_mbstring_copy() this copies a string in either
6672 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6673 into an ASN1_STRING type. A mask of permissible types is passed
6674 and it chooses the "minimal" type to use or an error if not type
6678 *) Add function equivalents to the various macros in asn1.h. The old
6679 macros are retained with an M_ prefix. Code inside the library can
6680 use the M_ macros. External code (including the openssl utility)
6681 should *NOT* in order to be "shared library friendly".
6684 *) Add various functions that can check a certificate's extensions
6685 to see if it usable for various purposes such as SSL client,
6686 server or S/MIME and CAs of these types. This is currently
6687 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6688 verification. Also added a -purpose flag to x509 utility to
6689 print out all the purposes.
6692 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6696 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6697 for, obtain and decode and extension and obtain its critical flag.
6698 This allows all the necessary extension code to be handled in a
6699 single function call.
6702 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6703 platforms. See crypto/rc4/rc4_enc.c for further details.
6706 *) New -noout option to asn1parse. This causes no output to be produced
6707 its main use is when combined with -strparse and -out to extract data
6708 from a file (which may not be in ASN.1 format).
6711 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6712 when producing the local key id.
6713 [Richard Levitte <levitte@stacken.kth.se>]
6715 *) New option -dhparam in s_server. This allows a DH parameter file to be
6716 stated explicitly. If it is not stated then it tries the first server
6717 certificate file. The previous behaviour hard coded the filename
6721 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6722 a public key to be input or output. For example:
6723 openssl rsa -in key.pem -pubout -out pubkey.pem
6724 Also added necessary DSA public key functions to handle this.
6727 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6728 in the message. This was handled by allowing
6729 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6730 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6732 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6733 to the end of the strings whereas this didn't. This would cause problems
6734 if strings read with d2i_ASN1_bytes() were later modified.
6735 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6737 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6738 data and it contains EOF it will end up returning an error. This is
6739 caused by input 46 bytes long. The cause is due to the way base64
6740 BIOs find the start of base64 encoded data. They do this by trying a
6741 trial decode on each line until they find one that works. When they
6742 do a flag is set and it starts again knowing it can pass all the
6743 data directly through the decoder. Unfortunately it doesn't reset
6744 the context it uses. This means that if EOF is reached an attempt
6745 is made to pass two EOFs through the context and this causes the
6746 resulting error. This can also cause other problems as well. As is
6747 usual with these problems it takes *ages* to find and the fix is
6748 trivial: move one line.
6749 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6751 *) Ugly workaround to get s_client and s_server working under Windows. The
6752 old code wouldn't work because it needed to select() on sockets and the
6753 tty (for keypresses and to see if data could be written). Win32 only
6754 supports select() on sockets so we select() with a 1s timeout on the
6755 sockets and then see if any characters are waiting to be read, if none
6756 are present then we retry, we also assume we can always write data to
6757 the tty. This isn't nice because the code then blocks until we've
6758 received a complete line of data and it is effectively polling the
6759 keyboard at 1s intervals: however it's quite a bit better than not
6760 working at all :-) A dedicated Windows application might handle this
6761 with an event loop for example.
6764 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6765 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6766 will be called when RSA_sign() and RSA_verify() are used. This is useful
6767 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6768 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6769 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6770 This necessitated the support of an extra signature type NID_md5_sha1
6771 for SSL signatures and modifications to the SSL library to use it instead
6772 of calling RSA_public_decrypt() and RSA_private_encrypt().
6775 *) Add new -verify -CAfile and -CApath options to the crl program, these
6776 will lookup a CRL issuers certificate and verify the signature in a
6777 similar way to the verify program. Tidy up the crl program so it
6778 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6779 less strict. It will now permit CRL extensions even if it is not
6780 a V2 CRL: this will allow it to tolerate some broken CRLs.
6783 *) Initialize all non-automatic variables each time one of the openssl
6784 sub-programs is started (this is necessary as they may be started
6785 multiple times from the "OpenSSL>" prompt).
6786 [Lennart Bang, Bodo Moeller]
6788 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6789 removing all other RSA functionality (this is what NO_RSA does). This
6790 is so (for example) those in the US can disable those operations covered
6791 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6795 *) Non-copying interface to BIO pairs.
6796 (still largely untested)
6799 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6800 ASCII string. This was handled independently in various places before.
6803 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6804 UTF8 strings a character at a time.
6807 *) Use client_version from client hello to select the protocol
6808 (s23_srvr.c) and for RSA client key exchange verification
6809 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6812 *) Add various utility functions to handle SPKACs, these were previously
6813 handled by poking round in the structure internals. Added new function
6814 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6815 print, verify and generate SPKACs. Based on an original idea from
6816 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6819 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6822 *) Allow the config file extension section to be overwritten on the
6823 command line. Based on an original idea from Massimiliano Pala
6824 <madwolf@comune.modena.it>. The new option is called -extensions
6825 and can be applied to ca, req and x509. Also -reqexts to override
6826 the request extensions in req and -crlexts to override the crl extensions
6830 *) Add new feature to the SPKAC handling in ca. Now you can include
6831 the same field multiple times by preceding it by "XXXX." for example:
6834 this is the same syntax as used in the req config file.
6837 *) Allow certificate extensions to be added to certificate requests. These
6838 are specified in a 'req_extensions' option of the req section of the
6839 config file. They can be printed out with the -text option to req but
6840 are otherwise ignored at present.
6843 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6844 data read consists of only the final block it would not decrypted because
6845 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6846 A misplaced 'break' also meant the decrypted final block might not be
6847 copied until the next read.
6850 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6851 a few extra parameters to the DH structure: these will be useful if
6852 for example we want the value of 'q' or implement X9.42 DH.
6855 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6856 provides hooks that allow the default DSA functions or functions on a
6857 "per key" basis to be replaced. This allows hardware acceleration and
6858 hardware key storage to be handled without major modification to the
6859 library. Also added low level modexp hooks and CRYPTO_EX structure and
6860 associated functions.
6863 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6864 as "read only": it can't be written to and the buffer it points to will
6865 not be freed. Reading from a read only BIO is much more efficient than
6866 a normal memory BIO. This was added because there are several times when
6867 an area of memory needs to be read from a BIO. The previous method was
6868 to create a memory BIO and write the data to it, this results in two
6869 copies of the data and an O(n^2) reading algorithm. There is a new
6870 function BIO_new_mem_buf() which creates a read only memory BIO from
6871 an area of memory. Also modified the PKCS#7 routines to use read only
6875 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6876 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6877 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6878 but a retry condition occured while trying to read the rest.
6881 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6882 NID_pkcs7_encrypted by default: this was wrong since this should almost
6883 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6884 the encrypted data type: this is a more sensible place to put it and it
6885 allows the PKCS#12 code to be tidied up that duplicated this
6889 *) Changed obj_dat.pl script so it takes its input and output files on
6890 the command line. This should avoid shell escape redirection problems
6894 *) Initial support for certificate extension requests, these are included
6895 in things like Xenroll certificate requests. Included functions to allow
6896 extensions to be obtained and added.
6899 *) -crlf option to s_client and s_server for sending newlines as
6900 CRLF (as required by many protocols).
6903 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6905 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6906 [Ralf S. Engelschall]
6908 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6909 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6911 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6915 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6916 DH parameters/keys (q is lost during that conversion, but the resulting
6917 DH parameters contain its length).
6919 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6920 much faster than DH_generate_parameters (which creates parameters
6921 where p = 2*q + 1), and also the smaller q makes DH computations
6922 much more efficient (160-bit exponentiation instead of 1024-bit
6923 exponentiation); so this provides a convenient way to support DHE
6924 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6925 utter importance to use
6926 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6928 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6929 when such DH parameters are used, because otherwise small subgroup
6930 attacks may become possible!
6933 *) Avoid memory leak in i2d_DHparams.
6936 *) Allow the -k option to be used more than once in the enc program:
6937 this allows the same encrypted message to be read by multiple recipients.
6940 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6941 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6942 it will always use the numerical form of the OID, even if it has a short
6946 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6947 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6948 otherwise bn_mod_exp was called. In the case of hardware keys for example
6949 no private key components need be present and it might store extra data
6950 in the RSA structure, which cannot be accessed from bn_mod_exp.
6951 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6952 private key operations.
6955 *) Added support for SPARC Linux.
6958 *) pem_password_cb function type incompatibly changed from
6959 typedef int pem_password_cb(char *buf, int size, int rwflag);
6961 ....(char *buf, int size, int rwflag, void *userdata);
6962 so that applications can pass data to their callbacks:
6963 The PEM[_ASN1]_{read,write}... functions and macros now take an
6964 additional void * argument, which is just handed through whenever
6965 the password callback is called.
6966 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6968 New function SSL_CTX_set_default_passwd_cb_userdata.
6970 Compatibility note: As many C implementations push function arguments
6971 onto the stack in reverse order, the new library version is likely to
6972 interoperate with programs that have been compiled with the old
6973 pem_password_cb definition (PEM_whatever takes some data that
6974 happens to be on the stack as its last argument, and the callback
6975 just ignores this garbage); but there is no guarantee whatsoever that
6978 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6979 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6980 problems not only on Windows, but also on some Unix platforms.
6981 To avoid problematic command lines, these definitions are now in an
6982 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6983 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6986 *) MIPS III/IV assembler module is reimplemented.
6989 *) More DES library cleanups: remove references to srand/rand and
6990 delete an unused file.
6993 *) Add support for the the free Netwide assembler (NASM) under Win32,
6994 since not many people have MASM (ml) and it can be hard to obtain.
6995 This is currently experimental but it seems to work OK and pass all
6996 the tests. Check out INSTALL.W32 for info.
6999 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7000 without temporary keys kept an extra copy of the server key,
7001 and connections with temporary keys did not free everything in case
7005 *) New function RSA_check_key and new openssl rsa option -check
7006 for verifying the consistency of RSA keys.
7007 [Ulf Moeller, Bodo Moeller]
7009 *) Various changes to make Win32 compile work:
7010 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7011 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7012 comparison" warnings.
7013 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7016 *) Add a debugging option to PKCS#5 v2 key generation function: when
7017 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7018 derived keys are printed to stderr.
7021 *) Copy the flags in ASN1_STRING_dup().
7022 [Roman E. Pavlov <pre@mo.msk.ru>]
7024 *) The x509 application mishandled signing requests containing DSA
7025 keys when the signing key was also DSA and the parameters didn't match.
7027 It was supposed to omit the parameters when they matched the signing key:
7028 the verifying software was then supposed to automatically use the CA's
7029 parameters if they were absent from the end user certificate.
7031 Omitting parameters is no longer recommended. The test was also
7032 the wrong way round! This was probably due to unusual behaviour in
7033 EVP_cmp_parameters() which returns 1 if the parameters match.
7034 This meant that parameters were omitted when they *didn't* match and
7035 the certificate was useless. Certificates signed with 'ca' didn't have
7037 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7039 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7040 The interface is as follows:
7041 Applications can use
7042 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7043 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7044 "off" is now the default.
7045 The library internally uses
7046 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7047 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7048 to disable memory-checking temporarily.
7050 Some inconsistent states that previously were possible (and were
7051 even the default) are now avoided.
7053 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7054 with each memory chunk allocated; this is occasionally more helpful
7055 than just having a counter.
7057 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7059 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7063 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7064 which largely parallels "options", but is for changing API behaviour,
7065 whereas "options" are about protocol behaviour.
7066 Initial "mode" flags are:
7068 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7069 a single record has been written.
7070 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7071 retries use the same buffer location.
7072 (But all of the contents must be
7076 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7079 *) Fix problems with no-hmac etc.
7080 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7082 *) New functions RSA_get_default_method(), RSA_set_method() and
7083 RSA_get_method(). These allows replacement of RSA_METHODs without having
7084 to mess around with the internals of an RSA structure.
7087 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7088 Also really enable memory leak checks in openssl.c and in some
7090 [Chad C. Mulligan, Bodo Moeller]
7092 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7093 up the length of negative integers. This has now been simplified to just
7094 store the length when it is first determined and use it later, rather
7095 than trying to keep track of where data is copied and updating it to
7097 [Steve Henson, reported by Brien Wheeler
7098 <bwheeler@authentica-security.com>]
7100 *) Add a new function PKCS7_signatureVerify. This allows the verification
7101 of a PKCS#7 signature but with the signing certificate passed to the
7102 function itself. This contrasts with PKCS7_dataVerify which assumes the
7103 certificate is present in the PKCS#7 structure. This isn't always the
7104 case: certificates can be omitted from a PKCS#7 structure and be
7105 distributed by "out of band" means (such as a certificate database).
7108 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7109 function prototypes in pem.h, also change util/mkdef.pl to add the
7110 necessary function names.
7113 *) mk1mf.pl (used by Windows builds) did not properly read the
7114 options set by Configure in the top level Makefile, and Configure
7115 was not even able to write more than one option correctly.
7116 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7119 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7120 file to be loaded from a BIO or FILE pointer. The BIO version will
7121 for example allow memory BIOs to contain config info.
7124 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7125 Whoever hopes to achieve shared-library compatibility across versions
7126 must use this, not the compile-time macro.
7127 (Exercise 0.9.4: Which is the minimum library version required by
7129 Note: All this applies only to multi-threaded programs, others don't
7133 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7134 through a BIO pair triggered the default case, i.e.
7135 SSLerr(...,SSL_R_UNKNOWN_STATE).
7138 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7139 can use the SSL library even if none of the specific BIOs is
7143 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7144 for the encoded length.
7145 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7147 *) Add initial documentation of the X509V3 functions.
7150 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7151 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7152 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7153 secure PKCS#8 private key format with a high iteration count.
7156 *) Fix determination of Perl interpreter: A perl or perl5
7157 _directory_ in $PATH was also accepted as the interpreter.
7158 [Ralf S. Engelschall]
7160 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7161 wrong with it but it was very old and did things like calling
7162 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7166 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7167 to use the new extension code.
7170 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7171 with macros. This should make it easier to change their form, add extra
7172 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7176 *) Add to configuration table a new entry that can specify an alternative
7177 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7178 according to Mark Crispin <MRC@Panda.COM>.
7182 *) DES CBC did not update the IV. Weird.
7185 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7186 Changing the behaviour of the former might break existing programs --
7187 where IV updating is needed, des_ncbc_encrypt can be used.
7190 *) When bntest is run from "make test" it drives bc to check its
7191 calculations, as well as internally checking them. If an internal check
7192 fails, it needs to cause bc to give a non-zero result or make test carries
7193 on without noticing the failure. Fixed.
7196 *) DES library cleanups.
7199 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7200 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7201 ciphers. NOTE: although the key derivation function has been verified
7202 against some published test vectors it has not been extensively tested
7203 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7207 *) Instead of "mkdir -p", which is not fully portable, use new
7208 Perl script "util/mkdir-p.pl".
7211 *) Rewrite the way password based encryption (PBE) is handled. It used to
7212 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7213 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7214 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7215 the 'parameter' field of the AlgorithmIdentifier is passed to the
7216 underlying key generation function so it must do its own ASN1 parsing.
7217 This has also changed the EVP_PBE_CipherInit() function which now has a
7218 'parameter' argument instead of literal salt and iteration count values
7219 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7222 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7223 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7224 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7225 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7226 value was just used as a "magic string" and not used directly its
7227 value doesn't matter.
7230 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7234 *) "linux-sparc64" configuration (ultrapenguin).
7235 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7236 "linux-sparc" configuration.
7237 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7239 *) config now generates no-xxx options for missing ciphers.
7242 *) Support the EBCDIC character set (work in progress).
7243 File ebcdic.c not yet included because it has a different license.
7244 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7246 *) Support BS2000/OSD-POSIX.
7247 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7249 *) Make callbacks for key generation use void * instead of char *.
7252 *) Make S/MIME samples compile (not yet tested).
7255 *) Additional typesafe stacks.
7258 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7262 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7264 *) New configuration variant "sco5-gcc".
7266 *) Updated some demos.
7267 [Sean O Riordain, Wade Scholine]
7269 *) Add missing BIO_free at exit of pkcs12 application.
7272 *) Fix memory leak in conf.c.
7275 *) Updates for Win32 to assembler version of MD5.
7278 *) Set #! path to perl in apps/der_chop to where we found it
7279 instead of using a fixed path.
7282 *) SHA library changes for irix64-mips4-cc.
7285 *) Improvements for VMS support.
7289 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7291 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7292 This also avoids the problems with SC4.2 and unpatched SC5.
7293 [Andy Polyakov <appro@fy.chalmers.se>]
7295 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7296 These are required because of the typesafe stack would otherwise break
7297 existing code. If old code used a structure member which used to be STACK
7298 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7299 sk_num or sk_value it would produce an error because the num, data members
7300 are not present in STACK_OF. Now it just produces a warning. sk_set
7301 replaces the old method of assigning a value to sk_value
7302 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7303 that does this will no longer work (and should use sk_set instead) but
7304 this could be regarded as a "questionable" behaviour anyway.
7307 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7308 correctly handle encrypted S/MIME data.
7311 *) Change type of various DES function arguments from des_cblock
7312 (which means, in function argument declarations, pointer to char)
7313 to des_cblock * (meaning pointer to array with 8 char elements),
7314 which allows the compiler to do more typechecking; it was like
7315 that back in SSLeay, but with lots of ugly casts.
7317 Introduce new type const_des_cblock.
7320 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7321 problems: find RecipientInfo structure that matches recipient certificate
7322 and initialise the ASN1 structures properly based on passed cipher.
7325 *) Belatedly make the BN tests actually check the results.
7328 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7329 to and from BNs: it was completely broken. New compilation option
7330 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7331 key elements as negative integers.
7334 *) Reorganize and speed up MD5.
7335 [Andy Polyakov <appro@fy.chalmers.se>]
7338 [Richard Levitte <richard@levitte.org>]
7340 *) New option -out to asn1parse to allow the parsed structure to be
7341 output to a file. This is most useful when combined with the -strparse
7342 option to examine the output of things like OCTET STRINGS.
7345 *) Make SSL library a little more fool-proof by not requiring any longer
7346 that SSL_set_{accept,connect}_state be called before
7347 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7348 in many applications because usually everything *appeared* to work as
7349 intended anyway -- now it really works as intended).
7352 *) Move openssl.cnf out of lib/.
7355 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7356 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7357 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7358 [Ralf S. Engelschall]
7360 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7361 handle PKCS#7 enveloped data properly.
7362 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7364 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7365 copying pointers. The cert_st handling is changed by this in
7366 various ways (and thus what used to be known as ctx->default_cert
7367 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7368 any longer when s->cert does not give us what we need).
7369 ssl_cert_instantiate becomes obsolete by this change.
7370 As soon as we've got the new code right (possibly it already is?),
7371 we have solved a couple of bugs of the earlier code where s->cert
7372 was used as if it could not have been shared with other SSL structures.
7374 Note that using the SSL API in certain dirty ways now will result
7375 in different behaviour than observed with earlier library versions:
7376 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7377 does not influence s as it used to.
7379 In order to clean up things more thoroughly, inside SSL_SESSION
7380 we don't use CERT any longer, but a new structure SESS_CERT
7381 that holds per-session data (if available); currently, this is
7382 the peer's certificate chain and, for clients, the server's certificate
7383 and temporary key. CERT holds only those values that can have
7384 meaningful defaults in an SSL_CTX.
7387 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7388 from the internal representation. Various PKCS#7 fixes: remove some
7389 evil casts and set the enc_dig_alg field properly based on the signing
7393 *) Allow PKCS#12 password to be set from the command line or the
7394 environment. Let 'ca' get its config file name from the environment
7395 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7399 *) Allow certificate policies extension to use an IA5STRING for the
7400 organization field. This is contrary to the PKIX definition but
7401 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7405 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7406 without disallowing inline assembler and the like for non-pedantic builds.
7409 *) Support Borland C++ builder.
7410 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7415 *) SHA-1 cleanups and performance enhancements.
7416 [Andy Polyakov <appro@fy.chalmers.se>]
7418 *) Sparc v8plus assembler for the bignum library.
7419 [Andy Polyakov <appro@fy.chalmers.se>]
7421 *) Accept any -xxx and +xxx compiler options in Configure.
7424 *) Update HPUX configuration.
7427 *) Add missing sk_<type>_unshift() function to safestack.h
7428 [Ralf S. Engelschall]
7430 *) New function SSL_CTX_use_certificate_chain_file that sets the
7431 "extra_cert"s in addition to the certificate. (This makes sense
7432 only for "PEM" format files, as chains as a whole are not
7436 *) Support verify_depth from the SSL API.
7437 x509_vfy.c had what can be considered an off-by-one-error:
7438 Its depth (which was not part of the external interface)
7439 was actually counting the number of certificates in a chain;
7440 now it really counts the depth.
7443 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7444 instead of X509err, which often resulted in confusing error
7445 messages since the error codes are not globally unique
7446 (e.g. an alleged error in ssl3_accept when a certificate
7447 didn't match the private key).
7449 *) New function SSL_CTX_set_session_id_context that allows to set a default
7450 value (so that you don't need SSL_set_session_id_context for each
7451 connection using the SSL_CTX).
7454 *) OAEP decoding bug fix.
7457 *) Support INSTALL_PREFIX for package builders, as proposed by
7461 *) New Configure options "threads" and "no-threads". For systems
7462 where the proper compiler options are known (currently Solaris
7463 and Linux), "threads" is the default.
7466 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7469 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7470 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7471 such as /usr/local/bin.
7474 *) "make linux-shared" to build shared libraries.
7475 [Niels Poppe <niels@netbox.org>]
7477 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7480 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7481 extension adding in x509 utility.
7484 *) Remove NOPROTO sections and error code comments.
7487 *) Partial rewrite of the DEF file generator to now parse the ANSI
7491 *) New Configure options --prefix=DIR and --openssldir=DIR.
7494 *) Complete rewrite of the error code script(s). It is all now handled
7495 by one script at the top level which handles error code gathering,
7496 header rewriting and C source file generation. It should be much better
7497 than the old method: it now uses a modified version of Ulf's parser to
7498 read the ANSI prototypes in all header files (thus the old K&R definitions
7499 aren't needed for error creation any more) and do a better job of
7500 translating function codes into names. The old 'ASN1 error code imbedded
7501 in a comment' is no longer necessary and it doesn't use .err files which
7502 have now been deleted. Also the error code call doesn't have to appear all
7503 on one line (which resulted in some large lines...).
7506 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7509 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7510 0 (which usually indicates a closed connection), but continue reading.
7513 *) Fix some race conditions.
7516 *) Add support for CRL distribution points extension. Add Certificate
7517 Policies and CRL distribution points documentation.
7520 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7523 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7524 8 of keying material. Merlin has also confirmed interop with this fix
7525 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7526 [Merlin Hughes <merlin@baltimore.ie>]
7528 *) Fix lots of warnings.
7529 [Richard Levitte <levitte@stacken.kth.se>]
7531 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7532 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7533 [Richard Levitte <levitte@stacken.kth.se>]
7535 *) Fix problems with sizeof(long) == 8.
7536 [Andy Polyakov <appro@fy.chalmers.se>]
7538 *) Change functions to ANSI C.
7541 *) Fix typos in error codes.
7542 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7544 *) Remove defunct assembler files from Configure.
7547 *) SPARC v8 assembler BIGNUM implementation.
7548 [Andy Polyakov <appro@fy.chalmers.se>]
7550 *) Support for Certificate Policies extension: both print and set.
7551 Various additions to support the r2i method this uses.
7554 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7555 return a const string when you are expecting an allocated buffer.
7558 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7559 types DirectoryString and DisplayText.
7562 *) Add code to allow r2i extensions to access the configuration database,
7563 add an LHASH database driver and add several ctx helper functions.
7566 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7567 fail when they extended the size of a BIGNUM.
7570 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7571 support typesafe stack.
7574 *) Fix typo in SSL_[gs]et_options().
7575 [Nils Frostberg <nils@medcom.se>]
7577 *) Delete various functions and files that belonged to the (now obsolete)
7578 old X509V3 handling code.
7581 *) New Configure option "rsaref".
7584 *) Don't auto-generate pem.h.
7587 *) Introduce type-safe ASN.1 SETs.
7590 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7591 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7593 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7594 that links with OpenSSL (well at least cause lots of warnings), but fear
7595 not: the conversion is trivial, and it eliminates loads of evil casts. A
7596 few STACKed things have been converted already. Feel free to convert more.
7597 In the fullness of time, I'll do away with the STACK type altogether.
7600 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7601 specified in <certfile> by updating the entry in the index.txt file.
7602 This way one no longer has to edit the index.txt file manually for
7603 revoking a certificate. The -revoke option does the gory details now.
7604 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7606 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7607 `-text' option at all and this way the `-noout -text' combination was
7608 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7609 [Ralf S. Engelschall]
7611 *) Make sure a corresponding plain text error message exists for the
7612 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7613 verify callback function determined that a certificate was revoked.
7614 [Ralf S. Engelschall]
7616 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7617 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7618 all available cipers including rc5, which was forgotten until now.
7619 In order to let the testing shell script know which algorithms
7620 are available, a new (up to now undocumented) command
7621 "openssl list-cipher-commands" is used.
7624 *) Bugfix: s_client occasionally would sleep in select() when
7625 it should have checked SSL_pending() first.
7628 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7629 the raw DSA values prior to ASN.1 encoding.
7632 *) Tweaks to Configure
7633 [Niels Poppe <niels@netbox.org>]
7635 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7639 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7642 *) New config option to avoid instructions that are illegal on the 80386.
7643 The default code is faster, but requires at least a 486.
7646 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7647 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7648 same as SSL2_VERSION anyway.
7651 *) New "-showcerts" option for s_client.
7654 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7655 application. Various cleanups and fixes.
7658 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7659 modify error routines to work internally. Add error codes and PBE init
7660 to library startup routines.
7663 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7664 packing functions to asn1 and evp. Changed function names and error
7665 codes along the way.
7668 *) PKCS12 integration: and so it begins... First of several patches to
7669 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7670 objects to objects.h
7673 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7674 and display support for Thawte strong extranet extension.
7677 *) Add LinuxPPC support.
7678 [Jeff Dubrule <igor@pobox.org>]
7680 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7681 bn_div_words in alpha.s.
7682 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7684 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7685 OAEP isn't supported when OpenSSL is built with RSAref.
7686 [Ulf Moeller <ulf@fitug.de>]
7688 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7689 so they no longer are missing under -DNOPROTO.
7690 [Soren S. Jorvang <soren@t.dk>]
7693 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7695 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7696 doesn't work when the session is reused. Coming soon!
7699 *) Fix a security hole, that allows sessions to be reused in the wrong
7700 context thus bypassing client cert protection! All software that uses
7701 client certs and session caches in multiple contexts NEEDS PATCHING to
7702 allow session reuse! A fuller solution is in the works.
7703 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7705 *) Some more source tree cleanups (removed obsolete files
7706 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7707 permission on "config" script to be executable) and a fix for the INSTALL
7709 [Ulf Moeller <ulf@fitug.de>]
7711 *) Remove some legacy and erroneous uses of malloc, free instead of
7713 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7715 *) Make rsa_oaep_test return non-zero on error.
7716 [Ulf Moeller <ulf@fitug.de>]
7718 *) Add support for native Solaris shared libraries. Configure
7719 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7720 if someone would make that last step automatic.
7721 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7723 *) ctx_size was not built with the right compiler during "make links". Fixed.
7726 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7727 except NULL ciphers". This means the default cipher list will no longer
7728 enable NULL ciphers. They need to be specifically enabled e.g. with
7729 the string "DEFAULT:eNULL".
7732 *) Fix to RSA private encryption routines: if p < q then it would
7733 occasionally produce an invalid result. This will only happen with
7734 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7737 *) Be less restrictive and allow also `perl util/perlpath.pl
7738 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7739 because this way one can also use an interpreter named `perl5' (which is
7740 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7741 installed as `perl').
7742 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7744 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7745 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7747 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7748 advapi32.lib to Win32 build and change the pem test comparision
7749 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7750 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7751 and crypto/des/ede_cbcm_enc.c.
7754 *) DES quad checksum was broken on big-endian architectures. Fixed.
7757 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7758 Win32 test batch file so it (might) work again. The Win32 test batch file
7759 is horrible: I feel ill....
7762 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7763 in e_os.h. Audit of header files to check ANSI and non ANSI
7764 sections: 10 functions were absent from non ANSI section and not exported
7765 from Windows DLLs. Fixed up libeay.num for new functions.
7768 *) Make `openssl version' output lines consistent.
7769 [Ralf S. Engelschall]
7771 *) Fix Win32 symbol export lists for BIO functions: Added
7772 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7773 to ms/libeay{16,32}.def.
7774 [Ralf S. Engelschall]
7776 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7777 fine under Unix and passes some trivial tests I've now added. But the
7778 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7779 added to make sure no one expects that this stuff really works in the
7780 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7781 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7783 [Ralf S. Engelschall]
7785 *) Fix the generation of two part addresses in perl.
7786 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7788 *) Add config entry for Linux on MIPS.
7789 [John Tobey <jtobey@channel1.com>]
7791 *) Make links whenever Configure is run, unless we are on Windoze.
7794 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7795 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7799 *) Add a useful kludge to allow package maintainers to specify compiler and
7800 other platforms details on the command line without having to patch the
7801 Configure script everytime: One now can use ``perl Configure
7802 <id>:<details>'', i.e. platform ids are allowed to have details appended
7803 to them (seperated by colons). This is treated as there would be a static
7804 pre-configured entry in Configure's %table under key <id> with value
7805 <details> and ``perl Configure <id>'' is called. So, when you want to
7806 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7807 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7808 now, which overrides the FreeBSD-elf entry on-the-fly.
7809 [Ralf S. Engelschall]
7811 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7814 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7815 on the `perl Configure ...' command line. This way one can compile
7816 OpenSSL libraries with Position Independent Code (PIC) which is needed
7817 for linking it into DSOs.
7818 [Ralf S. Engelschall]
7820 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7824 *) Cleaned up the LICENSE document: The official contact for any license
7825 questions now is the OpenSSL core team under openssl-core@openssl.org.
7826 And add a paragraph about the dual-license situation to make sure people
7827 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7828 to the OpenSSL toolkit.
7829 [Ralf S. Engelschall]
7831 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7832 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7833 Additonally cleaned up the `make links' target: Remove unnecessary
7834 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7835 to speed processing and no longer clutter the display with confusing
7836 stuff. Instead only the actually done links are displayed.
7837 [Ralf S. Engelschall]
7839 *) Permit null encryption ciphersuites, used for authentication only. It used
7840 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7841 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7845 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7846 signed attributes when verifying signatures (this would break them),
7847 the detached data encoding was wrong and public keys obtained using
7848 X509_get_pubkey() weren't freed.
7851 *) Add text documentation for the BUFFER functions. Also added a work around
7852 to a Win95 console bug. This was triggered by the password read stuff: the
7853 last character typed gets carried over to the next fread(). If you were
7854 generating a new cert request using 'req' for example then the last
7855 character of the passphrase would be CR which would then enter the first
7859 *) Added the new `Includes OpenSSL Cryptography Software' button as
7860 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7861 button and can be used by applications based on OpenSSL to show the
7862 relationship to the OpenSSL project.
7863 [Ralf S. Engelschall]
7865 *) Remove confusing variables in function signatures in files
7866 ssl/ssl_lib.c and ssl/ssl.h.
7867 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7869 *) Don't install bss_file.c under PREFIX/include/
7870 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7872 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7873 functions that return function pointers and has support for NT specific
7874 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7875 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7876 unsigned to signed types: this was killing the Win32 compile.
7879 *) Add new certificate file to stack functions,
7880 SSL_add_dir_cert_subjects_to_stack() and
7881 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7882 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7883 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7884 This means that Apache-SSL and similar packages don't have to mess around
7885 to add as many CAs as they want to the preferred list.
7888 *) Experiment with doxygen documentation. Currently only partially applied to
7890 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7891 openssl.doxy as the configuration file.
7894 *) Get rid of remaining C++-style comments which strict C compilers hate.
7895 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7897 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7898 compiled in by default: it has problems with large keys.
7901 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7902 DH private keys and/or callback functions which directly correspond to
7903 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7904 is needed for applications which have to configure certificates on a
7905 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7907 For the RSA certificate situation is makes no difference, but
7908 for the DSA certificate situation this fixes the "no shared cipher"
7909 problem where the OpenSSL cipher selection procedure failed because the
7910 temporary keys were not overtaken from the context and the API provided
7911 no way to reconfigure them.
7912 The new functions now let applications reconfigure the stuff and they
7913 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7914 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7915 non-public-API function ssl_cert_instantiate() is used as a helper
7916 function and also to reduce code redundancy inside ssl_rsa.c.
7917 [Ralf S. Engelschall]
7919 *) Move s_server -dcert and -dkey options out of the undocumented feature
7920 area because they are useful for the DSA situation and should be
7921 recognized by the users.
7922 [Ralf S. Engelschall]
7924 *) Fix the cipher decision scheme for export ciphers: the export bits are
7925 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7926 SSL_EXP_MASK. So, the original variable has to be used instead of the
7927 already masked variable.
7928 [Richard Levitte <levitte@stacken.kth.se>]
7930 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7931 [Richard Levitte <levitte@stacken.kth.se>]
7933 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7934 from `int' to `unsigned int' because it's a length and initialized by
7935 EVP_DigestFinal() which expects an `unsigned int *'.
7936 [Richard Levitte <levitte@stacken.kth.se>]
7938 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7939 script. Instead use the usual Shell->Perl transition trick.
7940 [Ralf S. Engelschall]
7942 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7943 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7944 -noout -modulus' as it's already the case for `openssl rsa -noout
7945 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7946 currently the public key is printed (a decision which was already done by
7947 `openssl dsa -modulus' in the past) which serves a similar purpose.
7948 Additionally the NO_RSA no longer completely removes the whole -modulus
7949 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7951 [Ralf S. Engelschall]
7953 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7954 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7955 [Arne Ansper <arne@ats.cyber.ee>]
7957 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7958 to be added. Now both 'req' and 'ca' can use new objects defined in the
7962 *) Add cool BIO that does syslog (or event log on NT).
7963 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7965 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7966 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7967 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7968 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7971 *) Add preliminary config info for new extension code.
7974 *) Make RSA_NO_PADDING really use no padding.
7975 [Ulf Moeller <ulf@fitug.de>]
7977 *) Generate errors when private/public key check is done.
7980 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7981 for some CRL extensions and new objects added.
7984 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7985 key usage extension and fuller support for authority key id.
7988 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7989 padding method for RSA, which is recommended for new applications in PKCS
7990 #1 v2.0 (RFC 2437, October 1998).
7991 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7992 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7993 against Bleichbacher's attack on RSA.
7994 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7997 *) Updates to the new SSL compression code
7998 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8000 *) Fix so that the version number in the master secret, when passed
8001 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8002 (because the server will not accept higher), that the version number
8003 is 0x03,0x01, not 0x03,0x00
8004 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8006 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8007 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8008 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8011 *) Support for RAW extensions where an arbitrary extension can be
8012 created by including its DER encoding. See apps/openssl.cnf for
8016 *) Make sure latest Perl versions don't interpret some generated C array
8017 code as Perl array code in the crypto/err/err_genc.pl script.
8018 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8020 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8021 not many people have the assembler. Various Win32 compilation fixes and
8022 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8026 *) Modify configure script 'Configure' to automatically create crypto/date.h
8027 file under Win32 and also build pem.h from pem.org. New script
8028 util/mkfiles.pl to create the MINFO file on environments that can't do a
8029 'make files': perl util/mkfiles.pl >MINFO should work.
8032 *) Major rework of DES function declarations, in the pursuit of correctness
8033 and purity. As a result, many evil casts evaporated, and some weirdness,
8034 too. You may find this causes warnings in your code. Zapping your evil
8035 casts will probably fix them. Mostly.
8038 *) Fix for a typo in asn1.h. Bug fix to object creation script
8039 obj_dat.pl. It considered a zero in an object definition to mean
8040 "end of object": none of the objects in objects.h have any zeros
8041 so it wasn't spotted.
8042 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8044 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8045 Masking (CBCM). In the absence of test vectors, the best I have been able
8046 to do is check that the decrypt undoes the encrypt, so far. Send me test
8047 vectors if you have them.
8050 *) Correct calculation of key length for export ciphers (too much space was
8051 allocated for null ciphers). This has not been tested!
8054 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8055 message is now correct (it understands "crypto" and "ssl" on its
8056 command line). There is also now an "update" option. This will update
8057 the util/ssleay.num and util/libeay.num files with any new functions.
8059 perl util/mkdef.pl crypto ssl update
8060 it will update them.
8063 *) Overhauled the Perl interface (perl/*):
8064 - ported BN stuff to OpenSSL's different BN library
8065 - made the perl/ source tree CVS-aware
8066 - renamed the package from SSLeay to OpenSSL (the files still contain
8067 their history because I've copied them in the repository)
8068 - removed obsolete files (the test scripts will be replaced
8069 by better Test::Harness variants in the future)
8070 [Ralf S. Engelschall]
8072 *) First cut for a very conservative source tree cleanup:
8073 1. merge various obsolete readme texts into doc/ssleay.txt
8074 where we collect the old documents and readme texts.
8075 2. remove the first part of files where I'm already sure that we no
8076 longer need them because of three reasons: either they are just temporary
8077 files which were left by Eric or they are preserved original files where
8078 I've verified that the diff is also available in the CVS via "cvs diff
8079 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8080 the crypto/md/ stuff).
8081 [Ralf S. Engelschall]
8083 *) More extension code. Incomplete support for subject and issuer alt
8084 name, issuer and authority key id. Change the i2v function parameters
8085 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8086 what that's for :-) Fix to ASN1 macro which messed up
8087 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8090 *) Preliminary support for ENUMERATED type. This is largely copied from the
8094 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8095 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8097 *) Make sure `make rehash' target really finds the `openssl' program.
8098 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8100 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8101 like to hear about it if this slows down other processors.
8104 *) Add CygWin32 platform information to Configure script.
8105 [Alan Batie <batie@aahz.jf.intel.com>]
8107 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8108 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8110 *) New program nseq to manipulate netscape certificate sequences
8113 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8117 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8118 but the BN code had some problems that would cause failures when
8119 doing certificate verification and some other functions.
8120 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8122 *) Add ASN1 and PEM code to support netscape certificate sequences.
8125 *) Add ASN1 and PEM code to support netscape certificate sequences.
8128 *) Add several PKIX and private extended key usage OIDs.
8131 *) Modify the 'ca' program to handle the new extension code. Modify
8132 openssl.cnf for new extension format, add comments.
8135 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8136 and add a sample to openssl.cnf so req -x509 now adds appropriate
8140 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8141 error code, add initial support to X509_print() and x509 application.
8144 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8145 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8146 stuff is currently isolated and isn't even compiled yet.
8149 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8150 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8151 Removed the versions check from X509 routines when loading extensions:
8152 this allows certain broken certificates that don't set the version
8153 properly to be processed.
8156 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8157 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8158 can still be regenerated with "make depend".
8161 *) Spelling mistake in C version of CAST-128.
8162 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8164 *) Changes to the error generation code. The perl script err-code.pl
8165 now reads in the old error codes and retains the old numbers, only
8166 adding new ones if necessary. It also only changes the .err files if new
8167 codes are added. The makefiles have been modified to only insert errors
8168 when needed (to avoid needlessly modifying header files). This is done
8169 by only inserting errors if the .err file is newer than the auto generated
8170 C file. To rebuild all the error codes from scratch (the old behaviour)
8171 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8172 or delete all the .err files.
8175 *) CAST-128 was incorrectly implemented for short keys. The C version has
8176 been fixed, but is untested. The assembler versions are also fixed, but
8177 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8178 to regenerate it if needed.
8179 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8180 Hagino <itojun@kame.net>]
8182 *) File was opened incorrectly in randfile.c.
8183 [Ulf Möller <ulf@fitug.de>]
8185 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8186 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8187 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8188 al: it's just almost always a UTCTime. Note this patch adds new error
8189 codes so do a "make errors" if there are problems.
8192 *) Correct Linux 1 recognition in config.
8193 [Ulf Möller <ulf@fitug.de>]
8195 *) Remove pointless MD5 hash when using DSA keys in ca.
8196 [Anonymous <nobody@replay.com>]
8198 *) Generate an error if given an empty string as a cert directory. Also
8199 generate an error if handed NULL (previously returned 0 to indicate an
8200 error, but didn't set one).
8201 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8203 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8206 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8207 parameters. This was causing a warning which killed off the Win32 compile.
8210 *) Remove C++ style comments from crypto/bn/bn_local.h.
8211 [Neil Costigan <neil.costigan@celocom.com>]
8213 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8214 based on a text string, looking up short and long names and finally
8215 "dot" format. The "dot" format stuff didn't work. Added new function
8216 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8217 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8218 OID is not part of the table.
8221 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8222 X509_LOOKUP_by_alias().
8225 *) Sort openssl functions by name.
8228 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8229 encryption from sample DSA keys (in case anyone is interested the password
8233 *) Make _all_ *_free functions accept a NULL pointer.
8234 [Frans Heymans <fheymans@isaserver.be>]
8236 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8238 [Anonymous <nobody@replay.com>]
8240 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8241 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8243 *) Don't blow it for numeric -newkey arguments to apps/req.
8244 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8246 *) Temp key "for export" tests were wrong in s3_srvr.c.
8247 [Anonymous <nobody@replay.com>]
8249 *) Add prototype for temp key callback functions
8250 SSL_CTX_set_tmp_{rsa,dh}_callback().
8253 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8254 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8257 *) X509_name_add_entry() freed the wrong thing after an error.
8258 [Arne Ansper <arne@ats.cyber.ee>]
8260 *) rsa_eay.c would attempt to free a NULL context.
8261 [Arne Ansper <arne@ats.cyber.ee>]
8263 *) BIO_s_socket() had a broken should_retry() on Windoze.
8264 [Arne Ansper <arne@ats.cyber.ee>]
8266 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8267 [Arne Ansper <arne@ats.cyber.ee>]
8269 *) Make sure the already existing X509_STORE->depth variable is initialized
8270 in X509_STORE_new(), but document the fact that this variable is still
8271 unused in the certificate verification process.
8272 [Ralf S. Engelschall]
8274 *) Fix the various library and apps files to free up pkeys obtained from
8275 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8278 *) Fix reference counting in X509_PUBKEY_get(). This makes
8279 demos/maurice/example2.c work, amongst others, probably.
8280 [Steve Henson and Ben Laurie]
8282 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8283 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8284 are no longer created. This way we have a single and consistent command
8285 line interface `openssl <command>', similar to `cvs <command>'.
8286 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8288 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8289 BIT STRING wrapper always have zero unused bits.
8292 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8295 *) Make the top-level INSTALL documentation easier to understand.
8298 *) Makefiles updated to exit if an error occurs in a sub-directory
8299 make (including if user presses ^C) [Paul Sutton]
8301 *) Make Montgomery context stuff explicit in RSA data structure.
8304 *) Fix build order of pem and err to allow for generated pem.h.
8307 *) Fix renumbering bug in X509_NAME_delete_entry().
8310 *) Enhanced the err-ins.pl script so it makes the error library number
8311 global and can add a library name. This is needed for external ASN1 and
8312 other error libraries.
8315 *) Fixed sk_insert which never worked properly.
8318 *) Fix ASN1 macros so they can handle indefinite length construted
8319 EXPLICIT tags. Some non standard certificates use these: they can now
8323 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8324 into a single doc/ssleay.txt bundle. This way the information is still
8325 preserved but no longer messes up this directory. Now it's new room for
8326 the new set of documenation files.
8327 [Ralf S. Engelschall]
8329 *) SETs were incorrectly DER encoded. This was a major pain, because they
8330 shared code with SEQUENCEs, which aren't coded the same. This means that
8331 almost everything to do with SETs or SEQUENCEs has either changed name or
8332 number of arguments.
8333 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8335 *) Fix test data to work with the above.
8338 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8339 was already fixed by Eric for 0.9.1 it seems.
8340 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8342 *) Autodetect FreeBSD3.
8345 *) Fix various bugs in Configure. This affects the following platforms:
8349 unixware-2.0-pentium
8353 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8354 before they are needed.
8357 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8361 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8363 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8364 changed SSLeay to OpenSSL in version strings.
8365 [Ralf S. Engelschall]
8367 *) Some fixups to the top-level documents.
8370 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8371 because the symlink to include/ was missing.
8372 [Ralf S. Engelschall]
8374 *) Incorporated the popular no-RSA/DSA-only patches
8375 which allow to compile a RSA-free SSLeay.
8376 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8378 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8379 when "ssleay" is still not found.
8380 [Ralf S. Engelschall]
8382 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8383 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8385 *) Updated the README file.
8386 [Ralf S. Engelschall]
8388 *) Added various .cvsignore files in the CVS repository subdirs
8389 to make a "cvs update" really silent.
8390 [Ralf S. Engelschall]
8392 *) Recompiled the error-definition header files and added
8393 missing symbols to the Win32 linker tables.
8394 [Ralf S. Engelschall]
8396 *) Cleaned up the top-level documents;
8397 o new files: CHANGES and LICENSE
8398 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8399 o merged COPYRIGHT into LICENSE
8400 o removed obsolete TODO file
8401 o renamed MICROSOFT to INSTALL.W32
8402 [Ralf S. Engelschall]
8404 *) Removed dummy files from the 0.9.1b source tree:
8405 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8406 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8407 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8408 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8409 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8410 [Ralf S. Engelschall]
8412 *) Added various platform portability fixes.
8415 *) The Genesis of the OpenSSL rpject:
8416 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8417 Young and Tim J. Hudson created while they were working for C2Net until
8419 [The OpenSSL Project]
8422 Changes between 0.9.0b and 0.9.1b [not released]
8424 *) Updated a few CA certificates under certs/
8427 *) Changed some BIGNUM api stuff.
8430 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8431 DGUX x86, Linux Alpha, etc.
8434 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8435 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8439 *) Add -strparse option to asn1pars program which parses nested
8441 [Dr Stephen Henson <shenson@bigfoot.com>]
8443 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8446 *) DSA fix for "ca" program.
8449 *) Added "-genkey" option to "dsaparam" program.
8452 *) Added RIPE MD160 (rmd160) message digest.
8455 *) Added -a (all) option to "ssleay version" command.
8458 *) Added PLATFORM define which is the id given to Configure.
8461 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8464 *) Extended the ASN.1 parser routines.
8467 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8470 *) Added a BN_CTX to the BN library.
8473 *) Fixed the weak key values in DES library
8476 *) Changed API in EVP library for cipher aliases.
8479 *) Added support for RC2/64bit cipher.
8482 *) Converted the lhash library to the crypto/mem.c functions.
8485 *) Added more recognized ASN.1 object ids.
8488 *) Added more RSA padding checks for SSL/TLS.
8491 *) Added BIO proxy/filter functionality.
8494 *) Added extra_certs to SSL_CTX which can be used
8495 send extra CA certificates to the client in the CA cert chain sending
8496 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8499 *) Now Fortezza is denied in the authentication phase because
8500 this is key exchange mechanism is not supported by SSLeay at all.
8503 *) Additional PKCS1 checks.
8506 *) Support the string "TLSv1" for all TLS v1 ciphers.
8509 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8510 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8513 *) Fixed a few memory leaks.
8516 *) Fixed various code and comment typos.
8519 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8520 bytes sent in the client random.
8521 [Edward Bishop <ebishop@spyglass.com>]
projects / thirdparty / openssl.git / blob