5 Changes between 0.9.8l (?) and 0.9.8m (?) [xx XXX xxxx]
7 *) Handle TLS versions 2.0 and later properly and correctly use the
8 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9 off ancient servers have a habit of sticking around for a while...
12 *) Modify compression code so it frees up structures without using the
13 ex_data callbacks. This works around a problem where some applications
14 call CRYPTO_free_all_ex_data() before application exit (e.g. when
15 restarting) then use compression (e.g. SSL with compression) later.
16 This results in significant per-connection memory leaks and
17 has caused some security issues including CVE-2008-1678 and
21 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
22 change when encrypting or decrypting.
25 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
26 connect (but not renegotiate) with servers which do not support RI.
27 Until RI is more widely deployed this option is enabled by default.
30 *) Add "missing" ssl ctrls to clear options and mode.
33 *) If client attempts to renegotiate and doesn't support RI respond with
34 a no_renegotiation alert as required by draft-ietf-tls-renegotiation.
35 Some renegotiating TLS clients will continue a connection gracefully
36 when they receive the alert. Unfortunately OpenSSL mishandled
37 this alert and would hang waiting for a server hello which it will never
38 receive. Now we treat a received no_renegotiation alert as a fatal
39 error. This is because applications requesting a renegotiation might well
40 expect it to succeed and would have no code in place to handle the server
41 denying it so the only safe thing to do is to terminate the connection.
44 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
45 peer supports secure renegotiation and 0 otherwise. Print out peer
46 renegotiation support in s_client/s_server.
49 *) Replace the highly broken and deprecated SPKAC certification method with
50 the updated NID creation version. This should correctly handle UTF8.
53 *) Implement draft-ietf-tls-renegotiation-03. Re-enable
54 renegotiation but require the extension as needed. Unfortunately,
55 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION turns out to be a
56 bad idea. It has been replaced by
57 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
58 SSL_CTX_set_options(). This is really not recommended unless you
59 know what you are doing.
60 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
62 *) Fixes to stateless session resumption handling. Use initial_ctx when
63 issuing and attempting to decrypt tickets in case it has changed during
64 servername handling. Use a non-zero length session ID when attempting
65 stateless session resumption: this makes it possible to determine if
66 a resumption has occurred immediately after receiving server hello
67 (several places in OpenSSL subtly assume this) instead of later in
71 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
72 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
73 fixes for a few places where the return code is not checked
75 [Julia Lawall <julia@diku.dk>]
77 *) Add --strict-warnings option to Configure script to include devteam
78 warnings in other configurations.
81 *) Add support for --libdir option and LIBDIR variable in makefiles. This
82 makes it possible to install openssl libraries in locations which
83 have names other than "lib", for example "/usr/lib64" which some
85 [Steve Henson, based on patch from Jeremy Utley]
87 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
88 X690 8.9.12 and can produce some misleading textual output of OIDs.
89 [Steve Henson, reported by Dan Kaminsky]
91 *) Delete MD2 from algorithm tables. This follows the recommendation in
92 several standards that it is not used in new applications due to
93 several cryptographic weaknesses. For binary compatibility reasons
94 the MD2 API is still compiled in by default.
97 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
101 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
102 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
104 [Guenter <lists@gknw.net>]
106 *) Fix the server certificate chain building code to use X509_verify_cert(),
107 it used to have an ad-hoc builder which was unable to cope with anything
108 other than a simple chain.
109 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
111 *) Don't check self signed certificate signatures in X509_verify_cert()
112 by default (a flag can override this): it just wastes time without
113 adding any security. As a useful side effect self signed root CAs
114 with non-FIPS digests are now usable in FIPS mode.
117 *) In dtls1_process_out_of_seq_message() the check if the current message
118 is already buffered was missing. For every new message was memory
119 allocated, allowing an attacker to perform an denial of service attack
120 with sending out of seq handshake messages until there is no memory
121 left. Additionally every future messege was buffered, even if the
122 sequence number made no sense and would be part of another handshake.
123 So only messages with sequence numbers less than 10 in advance will be
124 buffered. (CVE-2009-1378)
125 [Robin Seggelmann, discovered by Daniel Mentz]
127 *) Records are buffered if they arrive with a future epoch to be
128 processed after finishing the corresponding handshake. There is
129 currently no limitation to this buffer allowing an attacker to perform
130 a DOS attack with sending records with future epochs until there is no
131 memory left. This patch adds the pqueue_size() function to detemine
132 the size of a buffer and limits the record buffer to 100 entries.
134 [Robin Seggelmann, discovered by Daniel Mentz]
136 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
137 parent structure is freed. (CVE-2009-1379)
140 *) Handle non-blocking I/O properly in SSL_shutdown() call.
141 [Darryl Miles <darryl-mailinglists@netbauds.net>]
144 [Ilya O. <vrghost@gmail.com>]
146 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
148 *) Disable renegotiation completely - this fixes a severe security
149 problem (CVE-2009-3555) at the cost of breaking all
150 renegotiation. Renegotiation can be re-enabled by setting
151 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
152 run-time. This is really not recommended unless you know what
156 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
158 *) Don't set val to NULL when freeing up structures, it is freed up by
159 underlying code. If sizeof(void *) > sizeof(long) this can result in
160 zeroing past the valid field. (CVE-2009-0789)
161 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
163 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
164 checked correctly. This would allow some invalid signed attributes to
165 appear to verify correctly. (CVE-2009-0591)
166 [Ivan Nestlerode <inestlerode@us.ibm.com>]
168 *) Reject UniversalString and BMPString types with invalid lengths. This
169 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
170 a legal length. (CVE-2009-0590)
173 *) Set S/MIME signing as the default purpose rather than setting it
174 unconditionally. This allows applications to override it at the store
178 *) Permit restricted recursion of ASN1 strings. This is needed in practice
179 to handle some structures.
182 *) Improve efficiency of mem_gets: don't search whole buffer each time
184 [Jeremy Shapiro <jnshapir@us.ibm.com>]
186 *) New -hex option for openssl rand.
189 *) Print out UTF8String and NumericString when parsing ASN1.
192 *) Support NumericString type for name components.
195 *) Allow CC in the environment to override the automatically chosen
196 compiler. Note that nothing is done to ensure flags work with the
200 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
202 *) Properly check EVP_VerifyFinal() and similar return values
204 [Ben Laurie, Bodo Moeller, Google Security Team]
206 *) Enable TLS extensions by default.
209 *) Allow the CHIL engine to be loaded, whether the application is
210 multithreaded or not. (This does not release the developer from the
211 obligation to set up the dynamic locking callbacks.)
212 [Sander Temme <sander@temme.net>]
214 *) Use correct exit code if there is an error in dgst command.
215 [Steve Henson; problem pointed out by Roland Dirlewanger]
217 *) Tweak Configure so that you need to say "experimental-jpake" to enable
218 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
221 *) Add experimental JPAKE support, including demo authentication in
222 s_client and s_server.
225 *) Set the comparison function in v3_addr_canonize().
226 [Rob Austein <sra@hactrn.net>]
228 *) Add support for XMPP STARTTLS in s_client.
229 [Philip Paeps <philip@freebsd.org>]
231 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
232 to ensure that even with this option, only ciphersuites in the
233 server's preference list will be accepted. (Note that the option
234 applies only when resuming a session, so the earlier behavior was
235 just about the algorithm choice for symmetric cryptography.)
238 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
240 *) Fix NULL pointer dereference if a DTLS server received
241 ChangeCipherSpec as first record (CVE-2009-1386).
244 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
245 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
248 *) The fix in 0.9.8c that supposedly got rid of unsafe
249 double-checked locking was incomplete for RSA blinding,
250 addressing just one layer of what turns out to have been
251 doubly unsafe triple-checked locking.
253 So now fix this for real by retiring the MONT_HELPER macro
254 in crypto/rsa/rsa_eay.c.
256 [Bodo Moeller; problem pointed out by Marius Schilder]
258 *) Various precautionary measures:
260 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
262 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
263 (NB: This would require knowledge of the secret session ticket key
264 to exploit, in which case you'd be SOL either way.)
266 - Change bn_nist.c so that it will properly handle input BIGNUMs
267 outside the expected range.
269 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
272 [Neel Mehta, Bodo Moeller]
274 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
275 the load fails. Useful for distros.
276 [Ben Laurie and the FreeBSD team]
278 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
281 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
284 *) Expand ENGINE to support engine supplied SSL client certificate functions.
286 This work was sponsored by Logica.
289 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
290 keystores. Support for SSL/TLS client authentication too.
291 Not compiled unless enable-capieng specified to Configure.
293 This work was sponsored by Logica.
296 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
297 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
298 attribute creation routines such as certifcate requests and PKCS#12
302 Changes between 0.9.8g and 0.9.8h [28 May 2008]
304 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
305 handshake which could lead to a cilent crash as found using the
306 Codenomicon TLS test suite (CVE-2008-1672)
307 [Steve Henson, Mark Cox]
309 *) Fix double free in TLS server name extensions which could lead to
310 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
313 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
315 Clear the error queue to ensure that error entries left from
316 older function calls do not interfere with the correct operation.
317 [Lutz Jaenicke, Erik de Castro Lopo]
319 *) Remove root CA certificates of commercial CAs:
321 The OpenSSL project does not recommend any specific CA and does not
322 have any policy with respect to including or excluding any CA.
323 Therefore it does not make any sense to ship an arbitrary selection
324 of root CA certificates with the OpenSSL software.
327 *) RSA OAEP patches to fix two separate invalid memory reads.
328 The first one involves inputs when 'lzero' is greater than
329 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
330 before the beginning of from). The second one involves inputs where
331 the 'db' section contains nothing but zeroes (there is a one-byte
332 invalid read after the end of 'db').
333 [Ivan Nestlerode <inestlerode@us.ibm.com>]
335 *) Partial backport from 0.9.9-dev:
337 Introduce bn_mul_mont (dedicated Montgomery multiplication
338 procedure) as a candidate for BIGNUM assembler implementation.
339 While 0.9.9-dev uses assembler for various architectures, only
340 x86_64 is available by default here in the 0.9.8 branch, and
341 32-bit x86 is available through a compile-time setting.
343 To try the 32-bit x86 assembler implementation, use Configure
344 option "enable-montasm" (which exists only for this backport).
346 As "enable-montasm" for 32-bit x86 disclaims code stability
347 anyway, in this constellation we activate additional code
348 backported from 0.9.9-dev for further performance improvements,
349 namely BN_from_montgomery_word. (To enable this otherwise,
350 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
352 [Andy Polyakov (backport partially by Bodo Moeller)]
354 *) Add TLS session ticket callback. This allows an application to set
355 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
356 values. This is useful for key rollover for example where several key
357 sets may exist with different names.
360 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
361 This was broken until now in 0.9.8 releases, such that the only way
362 a registered ENGINE could be used (assuming it initialises
363 successfully on the host) was to explicitly set it as the default
364 for the relevant algorithms. This is in contradiction with 0.9.7
365 behaviour and the documentation. With this fix, when an ENGINE is
366 registered into a given algorithm's table of implementations, the
367 'uptodate' flag is reset so that auto-discovery will be used next
368 time a new context for that algorithm attempts to select an
370 [Ian Lister (tweaked by Geoff Thorpe)]
372 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
373 implemention in the following ways:
375 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
378 Lack of BER streaming support means one pass streaming processing is
379 only supported if data is detached: setting the streaming flag is
380 ignored for embedded content.
382 CMS support is disabled by default and must be explicitly enabled
383 with the enable-cms configuration option.
386 *) Update the GMP engine glue to do direct copies between BIGNUM and
387 mpz_t when openssl and GMP use the same limb size. Otherwise the
388 existing "conversion via a text string export" trick is still used.
389 [Paul Sheer <paulsheer@gmail.com>]
391 *) Zlib compression BIO. This is a filter BIO which compressed and
392 uncompresses any data passed through it.
395 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
396 RFC3394 compatible AES key wrapping.
399 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
400 sets string data without copying. X509_ALGOR_set0() and
401 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
402 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
403 from an X509_ATTRIBUTE structure optionally checking it occurs only
404 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
408 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
409 to get the expected BN_FLG_CONSTTIME behavior.
410 [Bodo Moeller (Google)]
414 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
415 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
416 - added some more tests to do_tests.pl
417 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
418 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
419 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
420 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
421 - various changes to netware.pl to enable gcc-cross builds on Win32
423 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
424 - various changes to fix missing prototype warnings
425 - fixed x86nasm.pl to create correct asm files for NASM COFF output
426 - added AES, WHIRLPOOL and CPUID assembler code to build files
427 - added missing AES assembler make rules to mk1mf.pl
428 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
429 [Guenter Knauf <eflash@gmx.net>]
431 *) Implement certificate status request TLS extension defined in RFC3546.
432 A client can set the appropriate parameters and receive the encoded
433 OCSP response via a callback. A server can query the supplied parameters
434 and set the encoded OCSP response in the callback. Add simplified examples
435 to s_client and s_server.
438 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
441 + Binary incompatibility of ssl_ctx_st structure
442 + DTLS interoperation with non-compliant servers
443 + Don't call get_session_cb() without proposed session
444 + Fix ia64 assembler code
445 [Andy Polyakov, Steve Henson]
447 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
449 *) DTLS Handshake overhaul. There were longstanding issues with
450 OpenSSL DTLS implementation, which were making it impossible for
451 RFC 4347 compliant client to communicate with OpenSSL server.
452 Unfortunately just fixing these incompatibilities would "cut off"
453 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
454 server keeps tolerating non RFC compliant syntax. The opposite is
455 not true, 0.9.8f client can not communicate with earlier server.
456 This update even addresses CVE-2007-4995.
459 *) Changes to avoid need for function casts in OpenSSL: some compilers
460 (gcc 4.2 and later) reject their use.
461 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
464 *) Add RFC4507 support to OpenSSL. This includes the corrections in
465 RFC4507bis. The encrypted ticket format is an encrypted encoded
466 SSL_SESSION structure, that way new session features are automatically
469 If a client application caches session in an SSL_SESSION structure
470 support is transparent because tickets are now stored in the encoded
473 The SSL_CTX structure automatically generates keys for ticket
474 protection in servers so again support should be possible
475 with no application modification.
477 If a client or server wishes to disable RFC4507 support then the option
478 SSL_OP_NO_TICKET can be set.
480 Add a TLS extension debugging callback to allow the contents of any client
481 or server extensions to be examined.
483 This work was sponsored by Google.
486 *) Add initial support for TLS extensions, specifically for the server_name
487 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
488 have new members for a host name. The SSL data structure has an
489 additional member SSL_CTX *initial_ctx so that new sessions can be
490 stored in that context to allow for session resumption, even after the
491 SSL has been switched to a new SSL_CTX in reaction to a client's
492 server_name extension.
494 New functions (subject to change):
497 SSL_get_servername_type()
500 New CTRL codes and macros (subject to change):
502 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
503 - SSL_CTX_set_tlsext_servername_callback()
504 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
505 - SSL_CTX_set_tlsext_servername_arg()
506 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
508 openssl s_client has a new '-servername ...' option.
510 openssl s_server has new options '-servername_host ...', '-cert2 ...',
511 '-key2 ...', '-servername_fatal' (subject to change). This allows
512 testing the HostName extension for a specific single host name ('-cert'
513 and '-key' remain fallbacks for handshakes without HostName
514 negotiation). If the unrecogninzed_name alert has to be sent, this by
515 default is a warning; it becomes fatal with the '-servername_fatal'
518 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
520 *) Add AES and SSE2 assembly language support to VC++ build.
523 *) Mitigate attack on final subtraction in Montgomery reduction.
526 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
527 (which previously caused an internal error).
530 *) Squeeze another 10% out of IGE mode when in != out.
533 *) AES IGE mode speedup.
534 [Dean Gaudet (Google)]
536 *) Add the Korean symmetric 128-bit cipher SEED (see
537 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
538 add SEED ciphersuites from RFC 4162:
540 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
541 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
542 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
543 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
545 To minimize changes between patchlevels in the OpenSSL 0.9.8
546 series, SEED remains excluded from compilation unless OpenSSL
547 is configured with 'enable-seed'.
550 *) Mitigate branch prediction attacks, which can be practical if a
551 single processor is shared, allowing a spy process to extract
552 information. For detailed background information, see
553 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
554 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
555 and Necessary Software Countermeasures"). The core of the change
556 are new versions BN_div_no_branch() and
557 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
558 respectively, which are slower, but avoid the security-relevant
559 conditional branches. These are automatically called by BN_div()
560 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
561 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
562 remove a conditional branch.
564 BN_FLG_CONSTTIME is the new name for the previous
565 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
566 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
567 in the exponent causes BN_mod_exp_mont() to use the alternative
568 implementation in BN_mod_exp_mont_consttime().) The old name
569 remains as a deprecated alias.
571 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
572 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
573 constant-time implementations for more than just exponentiation.
574 Here too the old name is kept as a deprecated alias.
576 BN_BLINDING_new() will now use BN_dup() for the modulus so that
577 the BN_BLINDING structure gets an independent copy of the
578 modulus. This means that the previous "BIGNUM *m" argument to
579 BN_BLINDING_new() and to BN_BLINDING_create_param() now
580 essentially becomes "const BIGNUM *m", although we can't actually
581 change this in the header file before 0.9.9. It allows
582 RSA_setup_blinding() to use BN_with_flags() on the modulus to
583 enable BN_FLG_CONSTTIME.
585 [Matthew D Wood (Intel Corp)]
587 *) In the SSL/TLS server implementation, be strict about session ID
588 context matching (which matters if an application uses a single
589 external cache for different purposes). Previously,
590 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
591 set. This did ensure strict client verification, but meant that,
592 with applications using a single external cache for quite
593 different requirements, clients could circumvent ciphersuite
594 restrictions for a given session ID context by starting a session
595 in a different context.
598 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
599 a ciphersuite string such as "DEFAULT:RSA" cannot enable
600 authentication-only ciphersuites.
603 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
604 not complete and could lead to a possible single byte overflow
605 (CVE-2007-5135) [Ben Laurie]
607 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
609 *) Since AES128 and AES256 (and similarly Camellia128 and
610 Camellia256) share a single mask bit in the logic of
611 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
612 kludge to work properly if AES128 is available and AES256 isn't
613 (or if Camellia128 is available and Camellia256 isn't).
616 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
617 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
618 When a point or a seed is encoded in a BIT STRING, we need to
619 prevent the removal of trailing zero bits to get the proper DER
620 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
621 of a NamedBitList, for which trailing 0 bits need to be removed.)
624 *) Have SSL/TLS server implementation tolerate "mismatched" record
625 protocol version while receiving ClientHello even if the
626 ClientHello is fragmented. (The server can't insist on the
627 particular protocol version it has chosen before the ServerHello
628 message has informed the client about his choice.)
631 *) Add RFC 3779 support.
632 [Rob Austein for ARIN, Ben Laurie]
634 *) Load error codes if they are not already present instead of using a
635 static variable. This allows them to be cleanly unloaded and reloaded.
636 Improve header file function name parsing.
639 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
640 or CAPABILITY handshake as required by RFCs.
643 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
645 *) Introduce limits to prevent malicious keys being able to
646 cause a denial of service. (CVE-2006-2940)
647 [Steve Henson, Bodo Moeller]
649 *) Fix ASN.1 parsing of certain invalid structures that can result
650 in a denial of service. (CVE-2006-2937) [Steve Henson]
652 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
653 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
655 *) Fix SSL client code which could crash if connecting to a
656 malicious SSLv2 server. (CVE-2006-4343)
657 [Tavis Ormandy and Will Drewry, Google Security Team]
659 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
660 match only those. Before that, "AES256-SHA" would be interpreted
661 as a pattern and match "AES128-SHA" too (since AES128-SHA got
662 the same strength classification in 0.9.7h) as we currently only
663 have a single AES bit in the ciphersuite description bitmap.
664 That change, however, also applied to ciphersuite strings such as
665 "RC4-MD5" that intentionally matched multiple ciphersuites --
666 namely, SSL 2.0 ciphersuites in addition to the more common ones
667 from SSL 3.0/TLS 1.0.
669 So we change the selection algorithm again: Naming an explicit
670 ciphersuite selects this one ciphersuite, and any other similar
671 ciphersuite (same bitmap) from *other* protocol versions.
672 Thus, "RC4-MD5" again will properly select both the SSL 2.0
673 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
675 Since SSL 2.0 does not have any ciphersuites for which the
676 128/256 bit distinction would be relevant, this works for now.
677 The proper fix will be to use different bits for AES128 and
678 AES256, which would have avoided the problems from the beginning;
679 however, bits are scarce, so we can only do this in a new release
680 (not just a patchlevel) when we can change the SSL_CIPHER
681 definition to split the single 'unsigned long mask' bitmap into
682 multiple values to extend the available space.
686 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
688 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
689 (CVE-2006-4339) [Ben Laurie and Google Security Team]
691 *) Add AES IGE and biIGE modes.
694 *) Change the Unix randomness entropy gathering to use poll() when
695 possible instead of select(), since the latter has some
696 undesirable limitations.
697 [Darryl Miles via Richard Levitte and Bodo Moeller]
699 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
700 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
701 cannot be implicitly activated as part of, e.g., the "AES" alias.
702 However, please upgrade to OpenSSL 0.9.9[-dev] for
703 non-experimental use of the ECC ciphersuites to get TLS extension
704 support, which is required for curve and point format negotiation
705 to avoid potential handshake problems.
708 *) Disable rogue ciphersuites:
710 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
711 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
712 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
714 The latter two were purportedly from
715 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
718 Also deactivate the remaining ciphersuites from
719 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
720 unofficial, and the ID has long expired.
723 *) Fix RSA blinding Heisenbug (problems sometimes occured on
724 dual-core machines) and other potential thread-safety issues.
727 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
728 versions), which is now available for royalty-free use
729 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
730 Also, add Camellia TLS ciphersuites from RFC 4132.
732 To minimize changes between patchlevels in the OpenSSL 0.9.8
733 series, Camellia remains excluded from compilation unless OpenSSL
734 is configured with 'enable-camellia'.
737 *) Disable the padding bug check when compression is in use. The padding
738 bug check assumes the first packet is of even length, this is not
739 necessarily true if compresssion is enabled and can result in false
740 positives causing handshake failure. The actual bug test is ancient
741 code so it is hoped that implementations will either have fixed it by
742 now or any which still have the bug do not support compression.
745 Changes between 0.9.8a and 0.9.8b [04 May 2006]
747 *) When applying a cipher rule check to see if string match is an explicit
748 cipher suite and only match that one cipher suite if it is.
751 *) Link in manifests for VC++ if needed.
752 [Austin Ziegler <halostatue@gmail.com>]
754 *) Update support for ECC-based TLS ciphersuites according to
755 draft-ietf-tls-ecc-12.txt with proposed changes (but without
756 TLS extensions, which are supported starting with the 0.9.9
757 branch, not in the OpenSSL 0.9.8 branch).
760 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
761 opaque EVP_CIPHER_CTX handling.
764 *) Fixes and enhancements to zlib compression code. We now only use
765 "zlib1.dll" and use the default __cdecl calling convention on Win32
766 to conform with the standards mentioned here:
767 http://www.zlib.net/DLL_FAQ.txt
768 Static zlib linking now works on Windows and the new --with-zlib-include
769 --with-zlib-lib options to Configure can be used to supply the location
770 of the headers and library. Gracefully handle case where zlib library
774 *) Several fixes and enhancements to the OID generation code. The old code
775 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
776 handle numbers larger than ULONG_MAX, truncated printing and had a
777 non standard OBJ_obj2txt() behaviour.
780 *) Add support for building of engines under engine/ as shared libraries
781 under VC++ build system.
784 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
785 Hopefully, we will not see any false combination of paths any more.
788 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
790 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
791 (part of SSL_OP_ALL). This option used to disable the
792 countermeasure against man-in-the-middle protocol-version
793 rollback in the SSL 2.0 server implementation, which is a bad
794 idea. (CVE-2005-2969)
796 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
797 for Information Security, National Institute of Advanced Industrial
798 Science and Technology [AIST], Japan)]
800 *) Add two function to clear and return the verify parameter flags.
803 *) Keep cipherlists sorted in the source instead of sorting them at
804 runtime, thus removing the need for a lock.
807 *) Avoid some small subgroup attacks in Diffie-Hellman.
808 [Nick Mathewson and Ben Laurie]
810 *) Add functions for well-known primes.
813 *) Extended Windows CE support.
814 [Satoshi Nakamura and Andy Polyakov]
816 *) Initialize SSL_METHOD structures at compile time instead of during
817 runtime, thus removing the need for a lock.
820 *) Make PKCS7_decrypt() work even if no certificate is supplied by
821 attempting to decrypt each encrypted key in turn. Add support to
825 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
827 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
830 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
833 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
834 key into the same file any more.
837 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
840 *) Add -utf8 command line and config file option to 'ca'.
841 [Stefan <stf@udoma.org]
843 *) Removed the macro des_crypt(), as it seems to conflict with some
844 libraries. Use DES_crypt().
847 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
848 involves renaming the source and generated shared-libs for
849 both. The engines will accept the corrected or legacy ids
850 ('ncipher' and '4758_cca' respectively) when binding. NB,
851 this only applies when building 'shared'.
852 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
854 *) Add attribute functions to EVP_PKEY structure. Modify
855 PKCS12_create() to recognize a CSP name attribute and
856 use it. Make -CSP option work again in pkcs12 utility.
859 *) Add new functionality to the bn blinding code:
860 - automatic re-creation of the BN_BLINDING parameters after
861 a fixed number of uses (currently 32)
862 - add new function for parameter creation
863 - introduce flags to control the update behaviour of the
864 BN_BLINDING parameters
865 - hide BN_BLINDING structure
866 Add a second BN_BLINDING slot to the RSA structure to improve
867 performance when a single RSA object is shared among several
871 *) Add support for DTLS.
872 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
874 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
875 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
878 *) Remove buggy and incompletet DH cert support from
879 ssl/ssl_rsa.c and ssl/s3_both.c
882 *) Use SHA-1 instead of MD5 as the default digest algorithm for
883 the apps/openssl applications.
886 *) Compile clean with "-Wall -Wmissing-prototypes
887 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
888 DEBUG_SAFESTACK must also be set.
891 *) Change ./Configure so that certain algorithms can be disabled by default.
892 The new counterpiece to "no-xxx" is "enable-xxx".
894 The patented RC5 and MDC2 algorithms will now be disabled unless
895 "enable-rc5" and "enable-mdc2", respectively, are specified.
897 (IDEA remains enabled despite being patented. This is because IDEA
898 is frequently required for interoperability, and there is no license
899 fee for non-commercial use. As before, "no-idea" can be used to
900 avoid this algorithm.)
904 *) Add processing of proxy certificates (see RFC 3820). This work was
905 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
906 EGEE (Enabling Grids for E-science in Europe).
909 *) RC4 performance overhaul on modern architectures/implementations, such
910 as Intel P4, IA-64 and AMD64.
913 *) New utility extract-section.pl. This can be used specify an alternative
914 section number in a pod file instead of having to treat each file as
915 a separate case in Makefile. This can be done by adding two lines to the
918 =for comment openssl_section:XXX
920 The blank line is mandatory.
924 *) New arguments -certform, -keyform and -pass for s_client and s_server
925 to allow alternative format key and certificate files and passphrase
929 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
930 update associated structures and add various utility functions.
932 Add new policy related verify parameters, include policy checking in
933 standard verify code. Enhance 'smime' application with extra parameters
934 to support policy checking and print out.
937 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
938 Nehemiah processors. These extensions support AES encryption in hardware
939 as well as RNG (though RNG support is currently disabled).
940 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
942 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
945 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
946 [Andy Polyakov and a number of other people]
948 *) Improved PowerPC platform support. Most notably BIGNUM assembler
949 implementation contributed by IBM.
950 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
952 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
953 exponent rather than 'unsigned long'. There is a corresponding change to
954 the new 'rsa_keygen' element of the RSA_METHOD structure.
955 [Jelte Jansen, Geoff Thorpe]
957 *) Functionality for creating the initial serial number file is now
958 moved from CA.pl to the 'ca' utility with a new option -create_serial.
960 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
961 number file to 1, which is bound to cause problems. To avoid
962 the problems while respecting compatibility between different 0.9.7
963 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
964 CA.pl for serial number initialization. With the new release 0.9.8,
965 we can fix the problem directly in the 'ca' utility.)
968 *) Reduced header interdepencies by declaring more opaque objects in
969 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
970 give fewer recursive includes, which could break lazy source code - so
971 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
972 developers should define this symbol when building and using openssl to
973 ensure they track the recommended behaviour, interfaces, [etc], but
974 backwards-compatible behaviour prevails when this isn't defined.
977 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
980 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
981 This will generate a random key of the appropriate length based on the
982 cipher context. The EVP_CIPHER can provide its own random key generation
983 routine to support keys of a specific form. This is used in the des and
984 3des routines to generate a key of the correct parity. Update S/MIME
985 code to use new functions and hence generate correct parity DES keys.
986 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
987 valid (weak or incorrect parity).
990 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
991 as looking them up. This is useful when the verified structure may contain
992 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
993 present unless the new PKCS7_NO_CRL flag is asserted.
996 *) Extend ASN1 oid configuration module. It now additionally accepts the
999 shortName = some long name, 1.2.3.4
1002 *) Reimplemented the BN_CTX implementation. There is now no more static
1003 limitation on the number of variables it can handle nor the depth of the
1004 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1005 information can now expand as required, and rather than having a single
1006 static array of bignums, BN_CTX now uses a linked-list of such arrays
1007 allowing it to expand on demand whilst maintaining the usefulness of
1008 BN_CTX's "bundling".
1011 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1012 to allow all RSA operations to function using a single BN_CTX.
1015 *) Preliminary support for certificate policy evaluation and checking. This
1016 is initially intended to pass the tests outlined in "Conformance Testing
1017 of Relying Party Client Certificate Path Processing Logic" v1.07.
1020 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1021 remained unused and not that useful. A variety of other little bignum
1022 tweaks and fixes have also been made continuing on from the audit (see
1026 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1027 associated ASN1, EVP and SSL functions and old ASN1 macros.
1030 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1031 and this should never fail. So the return value from the use of
1032 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1033 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1036 *) BN_CTX_get() should return zero-valued bignums, providing the same
1037 initialised value as BN_new().
1038 [Geoff Thorpe, suggested by Ulf Möller]
1040 *) Support for inhibitAnyPolicy certificate extension.
1043 *) An audit of the BIGNUM code is underway, for which debugging code is
1044 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1045 is considered valid when processing BIGNUMs, and causes execution to
1046 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1047 further steps are taken to deliberately pollute unused data in BIGNUM
1048 structures to try and expose faulty code further on. For now, openssl will
1049 (in its default mode of operation) continue to tolerate the inconsistent
1050 forms that it has tolerated in the past, but authors and packagers should
1051 consider trying openssl and their own applications when compiled with
1052 these debugging symbols defined. It will help highlight potential bugs in
1053 their own code, and will improve the test coverage for OpenSSL itself. At
1054 some point, these tighter rules will become openssl's default to improve
1055 maintainability, though the assert()s and other overheads will remain only
1056 in debugging configurations. See bn.h for more details.
1057 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1059 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1060 that can only be obtained through BN_CTX_new() (which implicitly
1061 initialises it). The presence of this function only made it possible
1062 to overwrite an existing structure (and cause memory leaks).
1065 *) Because of the callback-based approach for implementing LHASH as a
1066 template type, lh_insert() adds opaque objects to hash-tables and
1067 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1068 to clean up those corresponding objects before destroying the hash table
1069 (and losing the object pointers). So some over-zealous constifications in
1070 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1071 objects as "const" and the lh_doall[_arg] callback wrappers are not
1072 prototyped to have "const" restrictions on the object pointers they are
1073 given (and so aren't required to cast them away any more).
1076 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1077 (speed) prefers to use its own implementation. The two implementations
1078 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1079 its object type properly exposed (MS_TM) instead of casting to/from "char
1080 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1081 aren't necessarily the greatest nomenclatures - but this is what was used
1082 internally to the implementation so I've used that for now.
1085 *) Ensure that deprecated functions do not get compiled when
1086 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1087 the self-tests were still using deprecated key-generation functions so
1088 these have been updated also.
1091 *) Reorganise PKCS#7 code to separate the digest location functionality
1092 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1093 New function PKCS7_set_digest() to set the digest type for PKCS#7
1094 digestedData type. Add additional code to correctly generate the
1095 digestedData type and add support for this type in PKCS7 initialization
1099 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1100 structure of type "other".
1103 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1104 sure the loop does correctly stop and breaking ("division by zero")
1105 modulus operations are not performed. The (pre-generated) prime
1106 table crypto/bn/bn_prime.h was already correct, but it could not be
1107 re-generated on some platforms because of the "division by zero"
1108 situation in the script.
1109 [Ralf S. Engelschall]
1111 *) Update support for ECC-based TLS ciphersuites according to
1112 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1113 SHA-1 now is only used for "small" curves (where the
1114 representation of a field element takes up to 24 bytes); for
1115 larger curves, the field element resulting from ECDH is directly
1116 used as premaster secret.
1117 [Douglas Stebila (Sun Microsystems Laboratories)]
1119 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1120 curve secp160r1 to the tests.
1121 [Douglas Stebila (Sun Microsystems Laboratories)]
1123 *) Add the possibility to load symbols globally with DSO.
1124 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1126 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1127 control of the error stack.
1130 *) Add support for STORE in ENGINE.
1133 *) Add the STORE type. The intention is to provide a common interface
1134 to certificate and key stores, be they simple file-based stores, or
1135 HSM-type store, or LDAP stores, or...
1136 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1139 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1140 pass a list of arguments to any function as well as provide a way
1141 for a function to pass data back to the caller.
1144 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1145 works like BUF_strdup() but can be used to duplicate a portion of
1146 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1150 *) Add the function sk_find_ex() which works like sk_find(), but will
1151 return an index to an element even if an exact match couldn't be
1152 found. The index is guaranteed to point at the element where the
1153 searched-for key would be inserted to preserve sorting order.
1156 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1157 takes an extra flags argument for optional functionality. Currently,
1158 the following flags are defined:
1160 OBJ_BSEARCH_VALUE_ON_NOMATCH
1161 This one gets OBJ_bsearch_ex() to return a pointer to the first
1162 element where the comparing function returns a negative or zero
1165 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1166 This one gets OBJ_bsearch_ex() to return a pointer to the first
1167 element where the comparing function returns zero. This is useful
1168 if there are more than one element where the comparing function
1172 *) Make it possible to create self-signed certificates with 'openssl ca'
1173 in such a way that the self-signed certificate becomes part of the
1174 CA database and uses the same mechanisms for serial number generation
1175 as all other certificate signing. The new flag '-selfsign' enables
1176 this functionality. Adapt CA.sh and CA.pl.in.
1179 *) Add functionality to check the public key of a certificate request
1180 against a given private. This is useful to check that a certificate
1181 request can be signed by that key (self-signing).
1184 *) Make it possible to have multiple active certificates with the same
1185 subject in the CA index file. This is done only if the keyword
1186 'unique_subject' is set to 'no' in the main CA section (default
1187 if 'CA_default') of the configuration file. The value is saved
1188 with the database itself in a separate index attribute file,
1189 named like the index file with '.attr' appended to the name.
1192 *) Generate muti valued AVAs using '+' notation in config files for
1196 *) Support for nameConstraints certificate extension.
1199 *) Support for policyConstraints certificate extension.
1202 *) Support for policyMappings certificate extension.
1205 *) Make sure the default DSA_METHOD implementation only uses its
1206 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1207 and change its own handlers to be NULL so as to remove unnecessary
1208 indirection. This lets alternative implementations fallback to the
1209 default implementation more easily.
1212 *) Support for directoryName in GeneralName related extensions
1216 *) Make it possible to link applications using Makefile.shared.
1217 Make that possible even when linking against static libraries!
1220 *) Support for single pass processing for S/MIME signing. This now
1221 means that S/MIME signing can be done from a pipe, in addition
1222 cleartext signing (multipart/signed type) is effectively streaming
1223 and the signed data does not need to be all held in memory.
1225 This is done with a new flag PKCS7_STREAM. When this flag is set
1226 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1227 is done after the data is output (and digests calculated) in
1228 SMIME_write_PKCS7().
1231 *) Add full support for -rpath/-R, both in shared libraries and
1232 applications, at least on the platforms where it's known how
1236 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1237 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1238 will now compute a table of multiples of the generator that
1239 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1240 faster (notably in the case of a single point multiplication,
1241 scalar * generator).
1242 [Nils Larsch, Bodo Moeller]
1244 *) IPv6 support for certificate extensions. The various extensions
1245 which use the IP:a.b.c.d can now take IPv6 addresses using the
1246 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1250 *) Added an ENGINE that implements RSA by performing private key
1251 exponentiations with the GMP library. The conversions to and from
1252 GMP's mpz_t format aren't optimised nor are any montgomery forms
1253 cached, and on x86 it appears OpenSSL's own performance has caught up.
1254 However there are likely to be other architectures where GMP could
1255 provide a boost. This ENGINE is not built in by default, but it can be
1256 specified at Configure time and should be accompanied by the necessary
1257 linker additions, eg;
1258 ./config -DOPENSSL_USE_GMP -lgmp
1261 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1262 testing availability of engines with "-t" - the old behaviour is
1263 produced by increasing the feature's verbosity with "-tt".
1266 *) ECDSA routines: under certain error conditions uninitialized BN objects
1267 could be freed. Solution: make sure initialization is performed early
1268 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1272 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1273 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1274 software implementations. For DSA and DH, parameter generation can
1275 also be overriden by providing the appropriate method callbacks.
1278 *) Change the "progress" mechanism used in key-generation and
1279 primality testing to functions that take a new BN_GENCB pointer in
1280 place of callback/argument pairs. The new API functions have "_ex"
1281 postfixes and the older functions are reimplemented as wrappers for
1282 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1283 declarations of the old functions to help (graceful) attempts to
1284 migrate to the new functions. Also, the new key-generation API
1285 functions operate on a caller-supplied key-structure and return
1286 success/failure rather than returning a key or NULL - this is to
1287 help make "keygen" another member function of RSA_METHOD etc.
1289 Example for using the new callback interface:
1291 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1295 BN_GENCB_set(&my_cb, my_callback, my_arg);
1297 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1298 /* For the meaning of a, b in calls to my_callback(), see the
1299 * documentation of the function that calls the callback.
1300 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1301 * my_callback should return 1 if it wants BN_is_prime_ex()
1302 * to continue, or 0 to stop.
1307 *) Change the ZLIB compression method to be stateful, and make it
1308 available to TLS with the number defined in
1309 draft-ietf-tls-compression-04.txt.
1312 *) Add the ASN.1 structures and functions for CertificatePair, which
1313 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1315 CertificatePair ::= SEQUENCE {
1316 forward [0] Certificate OPTIONAL,
1317 reverse [1] Certificate OPTIONAL,
1318 -- at least one of the pair shall be present -- }
1320 Also implement the PEM functions to read and write certificate
1321 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1323 This needed to be defined, mostly for the sake of the LDAP
1324 attribute crossCertificatePair, but may prove useful elsewhere as
1328 *) Make it possible to inhibit symlinking of shared libraries in
1329 Makefile.shared, for Cygwin's sake.
1332 *) Extend the BIGNUM API by creating a function
1333 void BN_set_negative(BIGNUM *a, int neg);
1334 and a macro that behave like
1335 int BN_is_negative(const BIGNUM *a);
1337 to avoid the need to access 'a->neg' directly in applications.
1340 *) Implement fast modular reduction for pseudo-Mersenne primes
1341 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1342 EC_GROUP_new_curve_GFp() will now automatically use this
1344 [Nils Larsch <nla@trustcenter.de>]
1346 *) Add new lock type (CRYPTO_LOCK_BN).
1349 *) Change the ENGINE framework to automatically load engines
1350 dynamically from specific directories unless they could be
1351 found to already be built in or loaded. Move all the
1352 current engines except for the cryptodev one to a new
1354 The engines in engines/ are built as shared libraries if
1355 the "shared" options was given to ./Configure or ./config.
1356 Otherwise, they are inserted in libcrypto.a.
1357 /usr/local/ssl/engines is the default directory for dynamic
1358 engines, but that can be overriden at configure time through
1359 the usual use of --prefix and/or --openssldir, and at run
1360 time with the environment variable OPENSSL_ENGINES.
1361 [Geoff Thorpe and Richard Levitte]
1363 *) Add Makefile.shared, a helper makefile to build shared
1364 libraries. Addapt Makefile.org.
1367 *) Add version info to Win32 DLLs.
1368 [Peter 'Luna' Runestig" <peter@runestig.com>]
1370 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1371 can be added using this API to created arbitrary PKCS#12
1372 files while avoiding the low level API.
1374 New options to PKCS12_create(), key or cert can be NULL and
1375 will then be omitted from the output file. The encryption
1376 algorithm NIDs can be set to -1 for no encryption, the mac
1377 iteration count can be set to 0 to omit the mac.
1379 Enhance pkcs12 utility by making the -nokeys and -nocerts
1380 options work when creating a PKCS#12 file. New option -nomac
1381 to omit the mac, NONE can be set for an encryption algorithm.
1382 New code is modified to use the enhanced PKCS12_create()
1383 instead of the low level API.
1386 *) Extend ASN1 encoder to support indefinite length constructed
1387 encoding. This can output sequences tags and octet strings in
1388 this form. Modify pk7_asn1.c to support indefinite length
1389 encoding. This is experimental and needs additional code to
1390 be useful, such as an ASN1 bio and some enhanced streaming
1393 Extend template encode functionality so that tagging is passed
1394 down to the template encoder.
1397 *) Let 'openssl req' fail if an argument to '-newkey' is not
1398 recognized instead of using RSA as a default.
1401 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1402 As these are not official, they are not included in "ALL";
1403 the "ECCdraft" ciphersuite group alias can be used to select them.
1404 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1406 *) Add ECDH engine support.
1407 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1409 *) Add ECDH in new directory crypto/ecdh/.
1410 [Douglas Stebila (Sun Microsystems Laboratories)]
1412 *) Let BN_rand_range() abort with an error after 100 iterations
1413 without success (which indicates a broken PRNG).
1416 *) Change BN_mod_sqrt() so that it verifies that the input value
1417 is really the square of the return value. (Previously,
1418 BN_mod_sqrt would show GIGO behaviour.)
1421 *) Add named elliptic curves over binary fields from X9.62, SECG,
1422 and WAP/WTLS; add OIDs that were still missing.
1424 [Sheueling Chang Shantz and Douglas Stebila
1425 (Sun Microsystems Laboratories)]
1427 *) Extend the EC library for elliptic curves over binary fields
1428 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1431 EC_GF2m_simple_method
1435 EC_GROUP_new_curve_GF2m
1436 EC_GROUP_set_curve_GF2m
1437 EC_GROUP_get_curve_GF2m
1438 EC_POINT_set_affine_coordinates_GF2m
1439 EC_POINT_get_affine_coordinates_GF2m
1440 EC_POINT_set_compressed_coordinates_GF2m
1442 Point compression for binary fields is disabled by default for
1443 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1446 As binary polynomials are represented as BIGNUMs, various members
1447 of the EC_GROUP and EC_POINT data structures can be shared
1448 between the implementations for prime fields and binary fields;
1449 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1450 are essentially identical to their ..._GFp counterparts.
1451 (For simplicity, the '..._GFp' prefix has been dropped from
1452 various internal method names.)
1454 An internal 'field_div' method (similar to 'field_mul' and
1455 'field_sqr') has been added; this is used only for binary fields.
1457 [Sheueling Chang Shantz and Douglas Stebila
1458 (Sun Microsystems Laboratories)]
1460 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1461 through methods ('mul', 'precompute_mult').
1463 The generic implementations (now internally called 'ec_wNAF_mul'
1464 and 'ec_wNAF_precomputed_mult') remain the default if these
1465 methods are undefined.
1467 [Sheueling Chang Shantz and Douglas Stebila
1468 (Sun Microsystems Laboratories)]
1470 *) New function EC_GROUP_get_degree, which is defined through
1471 EC_METHOD. For curves over prime fields, this returns the bit
1472 length of the modulus.
1474 [Sheueling Chang Shantz and Douglas Stebila
1475 (Sun Microsystems Laboratories)]
1477 *) New functions EC_GROUP_dup, EC_POINT_dup.
1478 (These simply call ..._new and ..._copy).
1480 [Sheueling Chang Shantz and Douglas Stebila
1481 (Sun Microsystems Laboratories)]
1483 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1484 Polynomials are represented as BIGNUMs (where the sign bit is not
1485 used) in the following functions [macros]:
1488 BN_GF2m_sub [= BN_GF2m_add]
1489 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1490 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1491 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1493 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1494 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1495 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1496 BN_GF2m_cmp [= BN_ucmp]
1498 (Note that only the 'mod' functions are actually for fields GF(2^m).
1499 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1501 For some functions, an the irreducible polynomial defining a
1502 field can be given as an 'unsigned int[]' with strictly
1503 decreasing elements giving the indices of those bits that are set;
1504 i.e., p[] represents the polynomial
1505 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1507 p[0] > p[1] > ... > p[k] = 0.
1508 This applies to the following functions:
1513 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1514 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1516 BN_GF2m_mod_sqrt_arr
1517 BN_GF2m_mod_solve_quad_arr
1521 Conversion can be performed by the following functions:
1526 bntest.c has additional tests for binary polynomial arithmetic.
1528 Two implementations for BN_GF2m_mod_div() are available.
1529 The default algorithm simply uses BN_GF2m_mod_inv() and
1530 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1531 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1532 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1534 [Sheueling Chang Shantz and Douglas Stebila
1535 (Sun Microsystems Laboratories)]
1537 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1538 functionality is disabled at compile-time.
1539 [Douglas Stebila <douglas.stebila@sun.com>]
1541 *) Change default behaviour of 'openssl asn1parse' so that more
1542 information is visible when viewing, e.g., a certificate:
1544 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1545 mode the content of non-printable OCTET STRINGs is output in a
1546 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1547 avoid the appearance of a printable string.
1548 [Nils Larsch <nla@trustcenter.de>]
1550 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1552 EC_GROUP_set_asn1_flag()
1553 EC_GROUP_get_asn1_flag()
1554 EC_GROUP_set_point_conversion_form()
1555 EC_GROUP_get_point_conversion_form()
1556 These control ASN1 encoding details:
1557 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1558 has been set to OPENSSL_EC_NAMED_CURVE.
1559 - Points are encoded in uncompressed form by default; options for
1560 asn1_for are as for point2oct, namely
1561 POINT_CONVERSION_COMPRESSED
1562 POINT_CONVERSION_UNCOMPRESSED
1563 POINT_CONVERSION_HYBRID
1565 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1568 EC_GROUP_get0_seed()
1569 EC_GROUP_get_seed_len()
1570 This is used only for ASN1 purposes (so far).
1571 [Nils Larsch <nla@trustcenter.de>]
1573 *) Add 'field_type' member to EC_METHOD, which holds the NID
1574 of the appropriate field type OID. The new function
1575 EC_METHOD_get_field_type() returns this value.
1576 [Nils Larsch <nla@trustcenter.de>]
1581 EC_POINT_point2hex()
1582 EC_POINT_hex2point()
1583 providing useful interfaces to EC_POINT_point2oct() and
1584 EC_POINT_oct2point().
1585 [Nils Larsch <nla@trustcenter.de>]
1587 *) Change internals of the EC library so that the functions
1588 EC_GROUP_set_generator()
1589 EC_GROUP_get_generator()
1590 EC_GROUP_get_order()
1591 EC_GROUP_get_cofactor()
1592 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1593 to methods, which would lead to unnecessary code duplication when
1594 adding different types of curves.
1595 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1597 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1598 arithmetic, and such that modified wNAFs are generated
1599 (which avoid length expansion in many cases).
1602 *) Add a function EC_GROUP_check_discriminant() (defined via
1603 EC_METHOD) that verifies that the curve discriminant is non-zero.
1605 Add a function EC_GROUP_check() that makes some sanity tests
1606 on a EC_GROUP, its generator and order. This includes
1607 EC_GROUP_check_discriminant().
1608 [Nils Larsch <nla@trustcenter.de>]
1610 *) Add ECDSA in new directory crypto/ecdsa/.
1612 Add applications 'openssl ecparam' and 'openssl ecdsa'
1613 (these are based on 'openssl dsaparam' and 'openssl dsa').
1615 ECDSA support is also included in various other files across the
1616 library. Most notably,
1617 - 'openssl req' now has a '-newkey ecdsa:file' option;
1618 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1619 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1620 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1621 them suitable for ECDSA where domain parameters must be
1622 extracted before the specific public key;
1623 - ECDSA engine support has been added.
1624 [Nils Larsch <nla@trustcenter.de>]
1626 *) Include some named elliptic curves, and add OIDs from X9.62,
1627 SECG, and WAP/WTLS. Each curve can be obtained from the new
1629 EC_GROUP_new_by_curve_name(),
1630 and the list of available named curves can be obtained with
1631 EC_get_builtin_curves().
1632 Also add a 'curve_name' member to EC_GROUP objects, which can be
1634 EC_GROUP_set_curve_name()
1635 EC_GROUP_get_curve_name()
1636 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1638 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1639 was actually never needed) and in BN_mul(). The removal in BN_mul()
1640 required a small change in bn_mul_part_recursive() and the addition
1641 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1642 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1643 bn_sub_words() and bn_add_words() except they take arrays with
1647 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1649 *) Cleanse PEM buffers before freeing them since they may contain
1651 [Benjamin Bennett <ben@psc.edu>]
1653 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1654 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1655 authentication-only ciphersuites.
1658 *) Since AES128 and AES256 share a single mask bit in the logic of
1659 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1660 kludge to work properly if AES128 is available and AES256 isn't.
1663 *) Expand security boundary to match 1.1.1 module.
1666 *) Remove redundant features: hash file source, editing of test vectors
1667 modify fipsld to use external fips_premain.c signature.
1670 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1671 run algorithm test programs.
1674 *) Make algorithm test programs more tolerant of whitespace.
1677 *) Have SSL/TLS server implementation tolerate "mismatched" record
1678 protocol version while receiving ClientHello even if the
1679 ClientHello is fragmented. (The server can't insist on the
1680 particular protocol version it has chosen before the ServerHello
1681 message has informed the client about his choice.)
1684 *) Load error codes if they are not already present instead of using a
1685 static variable. This allows them to be cleanly unloaded and reloaded.
1688 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1690 *) Introduce limits to prevent malicious keys being able to
1691 cause a denial of service. (CVE-2006-2940)
1692 [Steve Henson, Bodo Moeller]
1694 *) Fix ASN.1 parsing of certain invalid structures that can result
1695 in a denial of service. (CVE-2006-2937) [Steve Henson]
1697 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1698 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1700 *) Fix SSL client code which could crash if connecting to a
1701 malicious SSLv2 server. (CVE-2006-4343)
1702 [Tavis Ormandy and Will Drewry, Google Security Team]
1704 *) Change ciphersuite string processing so that an explicit
1705 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1706 will no longer include "AES128-SHA"), and any other similar
1707 ciphersuite (same bitmap) from *other* protocol versions (so that
1708 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1709 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1710 changes from 0.9.8b and 0.9.8d.
1713 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1715 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1716 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1718 *) Change the Unix randomness entropy gathering to use poll() when
1719 possible instead of select(), since the latter has some
1720 undesirable limitations.
1721 [Darryl Miles via Richard Levitte and Bodo Moeller]
1723 *) Disable rogue ciphersuites:
1725 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1726 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1727 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1729 The latter two were purportedly from
1730 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1733 Also deactive the remaining ciphersuites from
1734 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1735 unofficial, and the ID has long expired.
1738 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1739 dual-core machines) and other potential thread-safety issues.
1742 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1744 *) Adapt fipsld and the build system to link against the validated FIPS
1745 module in FIPS mode.
1748 *) Fixes for VC++ 2005 build under Windows.
1751 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1752 from a Windows bash shell such as MSYS. It is autodetected from the
1753 "config" script when run from a VC++ environment. Modify standard VC++
1754 build to use fipscanister.o from the GNU make build.
1757 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1759 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1760 The value now differs depending on if you build for FIPS or not.
1761 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1762 safely run with a non-FIPSed libcrypto, as it may crash because of
1763 the difference induced by this change.
1766 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1768 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1769 (part of SSL_OP_ALL). This option used to disable the
1770 countermeasure against man-in-the-middle protocol-version
1771 rollback in the SSL 2.0 server implementation, which is a bad
1772 idea. (CVE-2005-2969)
1774 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1775 for Information Security, National Institute of Advanced Industrial
1776 Science and Technology [AIST], Japan)]
1778 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1779 mainly for FIPS compliance and not fully integrated at this stage.
1782 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1783 the exponentiation using a fixed-length exponent. (Otherwise,
1784 the information leaked through timing could expose the secret key
1785 after many signatures; cf. Bleichenbacher's attack on DSA with
1789 *) Make a new fixed-window mod_exp implementation the default for
1790 RSA, DSA, and DH private-key operations so that the sequence of
1791 squares and multiplies and the memory access pattern are
1792 independent of the particular secret key. This will mitigate
1793 cache-timing and potential related attacks.
1795 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1796 and this is automatically used by BN_mod_exp_mont() if the new flag
1797 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1798 will use this BN flag for private exponents unless the flag
1799 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1800 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1802 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1804 *) Change the client implementation for SSLv23_method() and
1805 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1806 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1807 (Previously, the SSL 2.0 backwards compatible Client Hello
1808 message format would be used even with SSL_OP_NO_SSLv2.)
1811 *) Add support for smime-type MIME parameter in S/MIME messages which some
1815 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1816 a threadsafe manner. Modify rsa code to use new function and add calls
1817 to dsa and dh code (which had race conditions before).
1820 *) Include the fixed error library code in the C error file definitions
1821 instead of fixing them up at runtime. This keeps the error code
1822 structures constant.
1825 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1827 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1830 *) Fixes for newer kerberos headers. NB: the casts are needed because
1831 the 'length' field is signed on one version and unsigned on another
1832 with no (?) obvious way to tell the difference, without these VC++
1833 complains. Also the "definition" of FAR (blank) is no longer included
1834 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1835 some needed definitions.
1838 *) Undo Cygwin change.
1841 *) Added support for proxy certificates according to RFC 3820.
1842 Because they may be a security thread to unaware applications,
1843 they must be explicitely allowed in run-time. See
1844 docs/HOWTO/proxy_certificates.txt for further information.
1847 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1849 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1850 server and client random values. Previously
1851 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1852 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1854 This change has negligible security impact because:
1856 1. Server and client random values still have 24 bytes of pseudo random
1859 2. Server and client random values are sent in the clear in the initial
1862 3. The master secret is derived using the premaster secret (48 bytes in
1863 size for static RSA ciphersuites) as well as client server and random
1866 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1869 [Stephen Henson, reported by UK NISCC]
1871 *) Use Windows randomness collection on Cygwin.
1874 *) Fix hang in EGD/PRNGD query when communication socket is closed
1875 prematurely by EGD/PRNGD.
1876 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1878 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1881 *) Back-port of selected performance improvements from development
1882 branch, as well as improved support for PowerPC platforms.
1885 *) Add lots of checks for memory allocation failure, error codes to indicate
1886 failure and freeing up memory if a failure occurs.
1887 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1889 *) Add new -passin argument to dgst.
1892 *) Perform some character comparisons of different types in X509_NAME_cmp:
1893 this is needed for some certificates that reencode DNs into UTF8Strings
1894 (in violation of RFC3280) and can't or wont issue name rollover
1898 *) Make an explicit check during certificate validation to see that
1899 the CA setting in each certificate on the chain is correct. As a
1900 side effect always do the following basic checks on extensions,
1901 not just when there's an associated purpose to the check:
1903 - if there is an unhandled critical extension (unless the user
1904 has chosen to ignore this fault)
1905 - if the path length has been exceeded (if one is set at all)
1906 - that certain extensions fit the associated purpose (if one has
1910 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1912 *) Avoid a race condition when CRLs are checked in a multi threaded
1913 environment. This would happen due to the reordering of the revoked
1914 entries during signature checking and serial number lookup. Now the
1915 encoding is cached and the serial number sort performed under a lock.
1916 Add new STACK function sk_is_sorted().
1919 *) Add Delta CRL to the extension code.
1922 *) Various fixes to s3_pkt.c so alerts are sent properly.
1923 [David Holmes <d.holmes@f5.com>]
1925 *) Reduce the chances of duplicate issuer name and serial numbers (in
1926 violation of RFC3280) using the OpenSSL certificate creation utilities.
1927 This is done by creating a random 64 bit value for the initial serial
1928 number when a serial number file is created or when a self signed
1929 certificate is created using 'openssl req -x509'. The initial serial
1930 number file is created using 'openssl x509 -next_serial' in CA.pl
1931 rather than being initialized to 1.
1934 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1936 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1937 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1938 [Joe Orton, Steve Henson]
1940 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1942 [Joe Orton, Steve Henson]
1944 *) Make it possible to have multiple active certificates with the same
1945 subject in the CA index file. This is done only if the keyword
1946 'unique_subject' is set to 'no' in the main CA section (default
1947 if 'CA_default') of the configuration file. The value is saved
1948 with the database itself in a separate index attribute file,
1949 named like the index file with '.attr' appended to the name.
1952 *) X509 verify fixes. Disable broken certificate workarounds when
1953 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1954 keyUsage extension present. Don't accept CRLs with unhandled critical
1955 extensions: since verify currently doesn't process CRL extensions this
1956 rejects a CRL with *any* critical extensions. Add new verify error codes
1960 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1961 A clarification of RFC2560 will require the use of OCTET STRINGs and
1962 some implementations cannot handle the current raw format. Since OpenSSL
1963 copies and compares OCSP nonces as opaque blobs without any attempt at
1964 parsing them this should not create any compatibility issues.
1967 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1968 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1969 this HMAC (and other) operations are several times slower than OpenSSL
1973 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1974 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1976 *) Use the correct content when signing type "other".
1979 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1981 *) Fix various bugs revealed by running the NISCC test suite:
1983 Stop out of bounds reads in the ASN1 code when presented with
1984 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1986 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1988 If verify callback ignores invalid public key errors don't try to check
1989 certificate signature with the NULL public key.
1993 *) New -ignore_err option in ocsp application to stop the server
1994 exiting on the first error in a request.
1997 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1998 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2002 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2003 extra data after the compression methods not only for TLS 1.0
2004 but also for SSL 3.0 (as required by the specification).
2005 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2007 *) Change X509_certificate_type() to mark the key as exported/exportable
2008 when it's 512 *bits* long, not 512 bytes.
2011 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2012 blocks during encryption.
2015 *) Various fixes to base64 BIO and non blocking I/O. On write
2016 flushes were not handled properly if the BIO retried. On read
2017 data was not being buffered properly and had various logic bugs.
2018 This also affects blocking I/O when the data being decoded is a
2022 *) Various S/MIME bugfixes and compatibility changes:
2023 output correct application/pkcs7 MIME type if
2024 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2025 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2026 of files as .eml work). Correctly handle very long lines in MIME
2030 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2032 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2033 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2034 a protocol version number mismatch like a decryption error
2035 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2038 *) Turn on RSA blinding by default in the default implementation
2039 to avoid a timing attack. Applications that don't want it can call
2040 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2041 They would be ill-advised to do so in most cases.
2042 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2044 *) Change RSA blinding code so that it works when the PRNG is not
2045 seeded (in this case, the secret RSA exponent is abused as
2046 an unpredictable seed -- if it is not unpredictable, there
2047 is no point in blinding anyway). Make RSA blinding thread-safe
2048 by remembering the creator's thread ID in rsa->blinding and
2049 having all other threads use local one-time blinding factors
2050 (this requires more computation than sharing rsa->blinding, but
2051 avoids excessive locking; and if an RSA object is not shared
2052 between threads, blinding will still be very fast).
2055 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2056 ENGINE as defaults for all supported algorithms irrespective of
2057 the 'flags' parameter. 'flags' is now honoured, so applications
2058 should make sure they are passing it correctly.
2061 *) Target "mingw" now allows native Windows code to be generated in
2062 the Cygwin environment as well as with the MinGW compiler.
2065 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2067 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2068 via timing by performing a MAC computation even if incorrrect
2069 block cipher padding has been found. This is a countermeasure
2070 against active attacks where the attacker has to distinguish
2071 between bad padding and a MAC verification error. (CVE-2003-0078)
2073 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2074 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2075 Martin Vuagnoux (EPFL, Ilion)]
2077 *) Make the no-err option work as intended. The intention with no-err
2078 is not to have the whole error stack handling routines removed from
2079 libcrypto, it's only intended to remove all the function name and
2080 reason texts, thereby removing some of the footprint that may not
2081 be interesting if those errors aren't displayed anyway.
2083 NOTE: it's still possible for any application or module to have it's
2084 own set of error texts inserted. The routines are there, just not
2085 used by default when no-err is given.
2088 *) Add support for FreeBSD on IA64.
2089 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2091 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2092 Kerberos function mit_des_cbc_cksum(). Before this change,
2093 the value returned by DES_cbc_cksum() was like the one from
2094 mit_des_cbc_cksum(), except the bytes were swapped.
2095 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2097 *) Allow an application to disable the automatic SSL chain building.
2098 Before this a rather primitive chain build was always performed in
2099 ssl3_output_cert_chain(): an application had no way to send the
2100 correct chain if the automatic operation produced an incorrect result.
2102 Now the chain builder is disabled if either:
2104 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2106 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2108 The reasoning behind this is that an application would not want the
2109 auto chain building to take place if extra chain certificates are
2110 present and it might also want a means of sending no additional
2111 certificates (for example the chain has two certificates and the
2115 *) Add the possibility to build without the ENGINE framework.
2116 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2118 *) Under Win32 gmtime() can return NULL: check return value in
2119 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2122 *) DSA routines: under certain error conditions uninitialized BN objects
2123 could be freed. Solution: make sure initialization is performed early
2124 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2125 Nils Larsch <nla@trustcenter.de> via PR#459)
2128 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2129 checked on reconnect on the client side, therefore session resumption
2130 could still fail with a "ssl session id is different" error. This
2131 behaviour is masked when SSL_OP_ALL is used due to
2132 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2133 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2134 followup to PR #377.
2137 *) IA-32 assembler support enhancements: unified ELF targets, support
2138 for SCO/Caldera platforms, fix for Cygwin shared build.
2141 *) Add support for FreeBSD on sparc64. As a consequence, support for
2142 FreeBSD on non-x86 processors is separate from x86 processors on
2143 the config script, much like the NetBSD support.
2144 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2146 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2148 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2151 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2152 code (06) was taken as the first octet of the session ID and the last
2153 octet was ignored consequently. As a result SSLv2 client side session
2154 caching could not have worked due to the session ID mismatch between
2156 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2160 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2161 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2165 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2166 seems that in spite of existing for more than a year, many application
2167 author have done nothing to provide the necessary callbacks, which
2168 means that this particular engine will not work properly anywhere.
2169 This is a very unfortunate situation which forces us, in the name
2170 of usability, to give the hw_ncipher.c a static lock, which is part
2172 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2173 appear in 0.9.8 or later. We EXPECT application authors to have
2174 dealt properly with this when 0.9.8 is released (unless we actually
2175 make such changes in the libcrypto locking code that changes will
2176 have to be made anyway).
2179 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2180 octets have been read, EOF or an error occurs. Without this change
2181 some truncated ASN1 structures will not produce an error.
2184 *) Disable Heimdal support, since it hasn't been fully implemented.
2185 Still give the possibility to force the use of Heimdal, but with
2186 warnings and a request that patches get sent to openssl-dev.
2189 *) Add the VC-CE target, introduce the WINCE sysname, and add
2190 INSTALL.WCE and appropriate conditionals to make it build.
2191 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2193 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2194 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2195 edit numbers of the version.
2196 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2198 *) Introduce safe string copy and catenation functions
2199 (BUF_strlcpy() and BUF_strlcat()).
2200 [Ben Laurie (CHATS) and Richard Levitte]
2202 *) Avoid using fixed-size buffers for one-line DNs.
2203 [Ben Laurie (CHATS)]
2205 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2206 resizing buffers containing secrets, and use where appropriate.
2207 [Ben Laurie (CHATS)]
2209 *) Avoid using fixed size buffers for configuration file location.
2210 [Ben Laurie (CHATS)]
2212 *) Avoid filename truncation for various CA files.
2213 [Ben Laurie (CHATS)]
2215 *) Use sizeof in preference to magic numbers.
2216 [Ben Laurie (CHATS)]
2218 *) Avoid filename truncation in cert requests.
2219 [Ben Laurie (CHATS)]
2221 *) Add assertions to check for (supposedly impossible) buffer
2223 [Ben Laurie (CHATS)]
2225 *) Don't cache truncated DNS entries in the local cache (this could
2226 potentially lead to a spoofing attack).
2227 [Ben Laurie (CHATS)]
2229 *) Fix various buffers to be large enough for hex/decimal
2230 representations in a platform independent manner.
2231 [Ben Laurie (CHATS)]
2233 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2234 resizing buffers containing secrets, and use where appropriate.
2235 [Ben Laurie (CHATS)]
2237 *) Add BIO_indent() to avoid much slightly worrying code to do
2239 [Ben Laurie (CHATS)]
2241 *) Convert sprintf()/BIO_puts() to BIO_printf().
2242 [Ben Laurie (CHATS)]
2244 *) buffer_gets() could terminate with the buffer only half
2246 [Ben Laurie (CHATS)]
2248 *) Add assertions to prevent user-supplied crypto functions from
2249 overflowing internal buffers by having large block sizes, etc.
2250 [Ben Laurie (CHATS)]
2252 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2254 [Ben Laurie (CHATS)]
2256 *) Eliminate unused copy of key in RC4.
2257 [Ben Laurie (CHATS)]
2259 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2260 [Ben Laurie (CHATS)]
2262 *) Fix off-by-one error in EGD path.
2263 [Ben Laurie (CHATS)]
2265 *) If RANDFILE path is too long, ignore instead of truncating.
2266 [Ben Laurie (CHATS)]
2268 *) Eliminate unused and incorrectly sized X.509 structure
2270 [Ben Laurie (CHATS)]
2272 *) Eliminate unused and dangerous function knumber().
2273 [Ben Laurie (CHATS)]
2275 *) Eliminate unused and dangerous structure, KSSL_ERR.
2276 [Ben Laurie (CHATS)]
2278 *) Protect against overlong session ID context length in an encoded
2279 session object. Since these are local, this does not appear to be
2281 [Ben Laurie (CHATS)]
2283 *) Change from security patch (see 0.9.6e below) that did not affect
2284 the 0.9.6 release series:
2286 Remote buffer overflow in SSL3 protocol - an attacker could
2287 supply an oversized master key in Kerberos-enabled versions.
2289 [Ben Laurie (CHATS)]
2291 *) Change the SSL kerb5 codes to match RFC 2712.
2294 *) Make -nameopt work fully for req and add -reqopt switch.
2295 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2297 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2298 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2300 *) Make sure tests can be performed even if the corresponding algorithms
2301 have been removed entirely. This was also the last step to make
2302 OpenSSL compilable with DJGPP under all reasonable conditions.
2303 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2305 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2306 to allow version independent disabling of normally unselected ciphers,
2307 which may be activated as a side-effect of selecting a single cipher.
2309 (E.g., cipher list string "RSA" enables ciphersuites that are left
2310 out of "ALL" because they do not provide symmetric encryption.
2311 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2312 [Lutz Jaenicke, Bodo Moeller]
2314 *) Add appropriate support for separate platform-dependent build
2315 directories. The recommended way to make a platform-dependent
2316 build directory is the following (tested on Linux), maybe with
2319 # Place yourself outside of the OpenSSL source tree. In
2320 # this example, the environment variable OPENSSL_SOURCE
2321 # is assumed to contain the absolute OpenSSL source directory.
2322 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2323 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2324 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2325 mkdir -p `dirname $F`
2326 ln -s $OPENSSL_SOURCE/$F $F
2329 To be absolutely sure not to disturb the source tree, a "make clean"
2330 is a good thing. If it isn't successfull, don't worry about it,
2331 it probably means the source directory is very clean.
2334 *) Make sure any ENGINE control commands make local copies of string
2335 pointers passed to them whenever necessary. Otherwise it is possible
2336 the caller may have overwritten (or deallocated) the original string
2337 data when a later ENGINE operation tries to use the stored values.
2338 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2340 *) Improve diagnostics in file reading and command-line digests.
2341 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2343 *) Add AES modes CFB and OFB to the object database. Correct an
2344 error in AES-CFB decryption.
2347 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2348 allows existing EVP_CIPHER_CTX structures to be reused after
2349 calling EVP_*Final(). This behaviour is used by encryption
2350 BIOs and some applications. This has the side effect that
2351 applications must explicitly clean up cipher contexts with
2352 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2355 *) Check the values of dna and dnb in bn_mul_recursive before calling
2356 bn_mul_comba (a non zero value means the a or b arrays do not contain
2357 n2 elements) and fallback to bn_mul_normal if either is not zero.
2360 *) Fix escaping of non-ASCII characters when using the -subj option
2361 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2364 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2365 form for "surname", serialNumber has no short form.
2366 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2367 therefore remove "mail" short name for "internet 7".
2368 The OID for unique identifiers in X509 certificates is
2369 x500UniqueIdentifier, not uniqueIdentifier.
2370 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2373 *) Add an "init" command to the ENGINE config module and auto initialize
2374 ENGINEs. Without any "init" command the ENGINE will be initialized
2375 after all ctrl commands have been executed on it. If init=1 the
2376 ENGINE is initailized at that point (ctrls before that point are run
2377 on the uninitialized ENGINE and after on the initialized one). If
2378 init=0 then the ENGINE will not be iniatialized at all.
2381 *) Fix the 'app_verify_callback' interface so that the user-defined
2382 argument is actually passed to the callback: In the
2383 SSL_CTX_set_cert_verify_callback() prototype, the callback
2384 declaration has been changed from
2387 int (*cb)(X509_STORE_CTX *,void *);
2388 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2389 i=s->ctx->app_verify_callback(&ctx)
2390 has been changed into
2391 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2393 To update applications using SSL_CTX_set_cert_verify_callback(),
2394 a dummy argument can be added to their callback functions.
2395 [D. K. Smetters <smetters@parc.xerox.com>]
2397 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2398 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2400 *) Add and OPENSSL_LOAD_CONF define which will cause
2401 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2402 This allows older applications to transparently support certain
2403 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2404 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2405 load the config file and OPENSSL_add_all_algorithms_conf() which will
2406 always load it have also been added.
2409 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2410 Adjust NIDs and EVP layer.
2411 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2413 *) Config modules support in openssl utility.
2415 Most commands now load modules from the config file,
2416 though in a few (such as version) this isn't done
2417 because it couldn't be used for anything.
2419 In the case of ca and req the config file used is
2420 the same as the utility itself: that is the -config
2421 command line option can be used to specify an
2425 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2426 use "openssl_conf" if filename is NULL use default openssl config file.
2429 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2430 config section name. Add a new flag to tolerate a missing config file
2431 and move code to CONF_modules_load_file().
2434 *) Support for crypto accelerator cards from Accelerated Encryption
2435 Processing, www.aep.ie. (Use engine 'aep')
2436 The support was copied from 0.9.6c [engine] and adapted/corrected
2437 to work with the new engine framework.
2438 [AEP Inc. and Richard Levitte]
2440 *) Support for SureWare crypto accelerator cards from Baltimore
2441 Technologies. (Use engine 'sureware')
2442 The support was copied from 0.9.6c [engine] and adapted
2443 to work with the new engine framework.
2446 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2447 make the newer ENGINE framework commands for the CHIL engine work.
2448 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2450 *) Make it possible to produce shared libraries on ReliantUNIX.
2451 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2453 *) Add the configuration target debug-linux-ppro.
2454 Make 'openssl rsa' use the general key loading routines
2455 implemented in apps.c, and make those routines able to
2456 handle the key format FORMAT_NETSCAPE and the variant
2458 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2460 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2461 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2463 *) Add -keyform to rsautl, and document -engine.
2464 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2466 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2467 BIO_R_NO_SUCH_FILE error code rather than the generic
2468 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2471 *) Add new functions
2473 ERR_peek_last_error_line
2474 ERR_peek_last_error_line_data.
2475 These are similar to
2478 ERR_peek_error_line_data,
2479 but report on the latest error recorded rather than the first one
2480 still in the error queue.
2481 [Ben Laurie, Bodo Moeller]
2483 *) default_algorithms option in ENGINE config module. This allows things
2485 default_algorithms = ALL
2486 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2489 *) Prelminary ENGINE config module.
2492 *) New experimental application configuration code.
2495 *) Change the AES code to follow the same name structure as all other
2496 symmetric ciphers, and behave the same way. Move everything to
2497 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2498 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2500 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2501 [Ben Laurie and Theo de Raadt]
2503 *) Add option to output public keys in req command.
2504 [Massimiliano Pala madwolf@openca.org]
2506 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2507 (up to about 10% better than before for P-192 and P-224).
2510 *) New functions/macros
2512 SSL_CTX_set_msg_callback(ctx, cb)
2513 SSL_CTX_set_msg_callback_arg(ctx, arg)
2514 SSL_set_msg_callback(ssl, cb)
2515 SSL_set_msg_callback_arg(ssl, arg)
2517 to request calling a callback function
2519 void cb(int write_p, int version, int content_type,
2520 const void *buf, size_t len, SSL *ssl, void *arg)
2522 whenever a protocol message has been completely received
2523 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2524 protocol version according to which the SSL library interprets
2525 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2526 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2527 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2528 specification (change_cipher_spec(20), alert(21), handshake(22)).
2529 'buf' and 'len' point to the actual message, 'ssl' to the
2530 SSL object, and 'arg' is the application-defined value set by
2531 SSL[_CTX]_set_msg_callback_arg().
2533 'openssl s_client' and 'openssl s_server' have new '-msg' options
2534 to enable a callback that displays all protocol messages.
2537 *) Change the shared library support so shared libraries are built as
2538 soon as the corresponding static library is finished, and thereby get
2539 openssl and the test programs linked against the shared library.
2540 This still only happens when the keyword "shard" has been given to
2541 the configuration scripts.
2543 NOTE: shared library support is still an experimental thing, and
2544 backward binary compatibility is still not guaranteed.
2545 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2547 *) Add support for Subject Information Access extension.
2548 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2550 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2551 additional bytes when new memory had to be allocated, not just
2552 when reusing an existing buffer.
2555 *) New command line and configuration option 'utf8' for the req command.
2556 This allows field values to be specified as UTF8 strings.
2559 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2560 runs for the former and machine-readable output for the latter.
2563 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2564 of the e-mail address in the DN (i.e., it will go into a certificate
2565 extension only). The new configuration file option 'email_in_dn = no'
2566 has the same effect.
2567 [Massimiliano Pala madwolf@openca.org]
2569 *) Change all functions with names starting with des_ to be starting
2570 with DES_ instead. Add wrappers that are compatible with libdes,
2571 but are named _ossl_old_des_*. Finally, add macros that map the
2572 des_* symbols to the corresponding _ossl_old_des_* if libdes
2573 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2574 desired, the des_* symbols will be mapped to DES_*, with one
2577 Since we provide two compatibility mappings, the user needs to
2578 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2579 compatibility is desired. The default (i.e., when that macro
2580 isn't defined) is OpenSSL 0.9.6c compatibility.
2582 There are also macros that enable and disable the support of old
2583 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2584 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2585 are defined, the default will apply: to support the old des routines.
2587 In either case, one must include openssl/des.h to get the correct
2588 definitions. Do not try to just include openssl/des_old.h, that
2591 NOTE: This is a major break of an old API into a new one. Software
2592 authors are encouraged to switch to the DES_ style functions. Some
2593 time in the future, des_old.h and the libdes compatibility functions
2594 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2595 default), and then completely removed.
2598 *) Test for certificates which contain unsupported critical extensions.
2599 If such a certificate is found during a verify operation it is
2600 rejected by default: this behaviour can be overridden by either
2601 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2602 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2603 X509_supported_extension() has also been added which returns 1 if a
2604 particular extension is supported.
2607 *) Modify the behaviour of EVP cipher functions in similar way to digests
2608 to retain compatibility with existing code.
2611 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2612 compatibility with existing code. In particular the 'ctx' parameter does
2613 not have to be to be initialized before the call to EVP_DigestInit() and
2614 it is tidied up after a call to EVP_DigestFinal(). New function
2615 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2616 EVP_MD_CTX_copy() changed to not require the destination to be
2617 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2618 requires the destination to be valid.
2620 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2621 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2624 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2625 so that complete 'Handshake' protocol structures are kept in memory
2626 instead of overwriting 'msg_type' and 'length' with 'body' data.
2629 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2630 [Massimo Santin via Richard Levitte]
2632 *) Major restructuring to the underlying ENGINE code. This includes
2633 reduction of linker bloat, separation of pure "ENGINE" manipulation
2634 (initialisation, etc) from functionality dealing with implementations
2635 of specific crypto iterfaces. This change also introduces integrated
2636 support for symmetric ciphers and digest implementations - so ENGINEs
2637 can now accelerate these by providing EVP_CIPHER and EVP_MD
2638 implementations of their own. This is detailed in crypto/engine/README
2639 as it couldn't be adequately described here. However, there are a few
2640 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2641 were changed in the original introduction of ENGINE code have now
2642 reverted back - the hooking from this code to ENGINE is now a good
2643 deal more passive and at run-time, operations deal directly with
2644 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2645 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2646 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2647 they were not being used by the framework as there is no concept of a
2648 BIGNUM_METHOD and they could not be generalised to the new
2649 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2650 ENGINE_cpy() has been removed as it cannot be consistently defined in
2654 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2657 *) Change mkdef.pl to sort symbols that get the same entry number,
2658 and make sure the automatically generated functions ERR_load_*
2659 become part of libeay.num as well.
2662 *) New function SSL_renegotiate_pending(). This returns true once
2663 renegotiation has been requested (either SSL_renegotiate() call
2664 or HelloRequest/ClientHello receveived from the peer) and becomes
2665 false once a handshake has been completed.
2666 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2667 sends a HelloRequest, but does not ensure that a handshake takes
2668 place. SSL_renegotiate_pending() is useful for checking if the
2669 client has followed the request.)
2672 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2673 By default, clients may request session resumption even during
2674 renegotiation (if session ID contexts permit); with this option,
2675 session resumption is possible only in the first handshake.
2677 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2678 more bits available for options that should not be part of
2679 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2682 *) Add some demos for certificate and certificate request creation.
2685 *) Make maximum certificate chain size accepted from the peer application
2686 settable (SSL*_get/set_max_cert_list()), as proposed by
2687 "Douglas E. Engert" <deengert@anl.gov>.
2690 *) Add support for shared libraries for Unixware-7
2691 (Boyd Lynn Gerber <gerberb@zenez.com>).
2694 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2695 be done prior to destruction. Use this to unload error strings from
2696 ENGINEs that load their own error strings. NB: This adds two new API
2697 functions to "get" and "set" this destroy handler in an ENGINE.
2700 *) Alter all existing ENGINE implementations (except "openssl" and
2701 "openbsd") to dynamically instantiate their own error strings. This
2702 makes them more flexible to be built both as statically-linked ENGINEs
2703 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2704 Also, add stub code to each that makes building them as self-contained
2705 shared-libraries easier (see README.ENGINE).
2708 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2709 implementations into applications that are completely implemented in
2710 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2711 commands that can be used to configure what shared-library to load and
2712 to control aspects of the way it is handled. Also, made an update to
2713 the README.ENGINE file that brings its information up-to-date and
2714 provides some information and instructions on the "dynamic" ENGINE
2715 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2718 *) Make it possible to unload ranges of ERR strings with a new
2719 "ERR_unload_strings" function.
2722 *) Add a copy() function to EVP_MD.
2725 *) Make EVP_MD routines take a context pointer instead of just the
2726 md_data void pointer.
2729 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2730 that the digest can only process a single chunk of data
2731 (typically because it is provided by a piece of
2732 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2733 is only going to provide a single chunk of data, and hence the
2734 framework needn't accumulate the data for oneshot drivers.
2737 *) As with "ERR", make it possible to replace the underlying "ex_data"
2738 functions. This change also alters the storage and management of global
2739 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2740 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2741 index counters. The API functions that use this state have been changed
2742 to take a "class_index" rather than pointers to the class's local STACK
2743 and counter, and there is now an API function to dynamically create new
2744 classes. This centralisation allows us to (a) plug a lot of the
2745 thread-safety problems that existed, and (b) makes it possible to clean
2746 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2747 such data would previously have always leaked in application code and
2748 workarounds were in place to make the memory debugging turn a blind eye
2749 to it. Application code that doesn't use this new function will still
2750 leak as before, but their memory debugging output will announce it now
2751 rather than letting it slide.
2753 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2754 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2755 has a return value to indicate success or failure.
2758 *) Make it possible to replace the underlying "ERR" functions such that the
2759 global state (2 LHASH tables and 2 locks) is only used by the "default"
2760 implementation. This change also adds two functions to "get" and "set"
2761 the implementation prior to it being automatically set the first time
2762 any other ERR function takes place. Ie. an application can call "get",
2763 pass the return value to a module it has just loaded, and that module
2764 can call its own "set" function using that value. This means the
2765 module's "ERR" operations will use (and modify) the error state in the
2766 application and not in its own statically linked copy of OpenSSL code.
2769 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2770 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2771 the operation, and provides a more encapsulated way for external code
2772 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2773 to use these functions rather than manually incrementing the counts.
2775 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2778 *) Add EVP test program.
2781 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2784 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2785 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2786 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2787 These allow a CRL to be built without having to access X509_CRL fields
2788 directly. Modify 'ca' application to use new functions.
2791 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2792 bug workarounds. Rollback attack detection is a security feature.
2793 The problem will only arise on OpenSSL servers when TLSv1 is not
2794 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2795 Software authors not wanting to support TLSv1 will have special reasons
2796 for their choice and can explicitly enable this option.
2797 [Bodo Moeller, Lutz Jaenicke]
2799 *) Rationalise EVP so it can be extended: don't include a union of
2800 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2801 (similar to those existing for EVP_CIPHER_CTX).
2806 EVP_MD_CTX_init(&md); /* new function call */
2807 EVP_DigestInit(&md, EVP_sha1());
2808 EVP_DigestUpdate(&md, in, len);
2809 EVP_DigestFinal(&md, out, NULL);
2810 EVP_MD_CTX_cleanup(&md); /* new function call */
2814 *) Make DES key schedule conform to the usual scheme, as well as
2815 correcting its structure. This means that calls to DES functions
2816 now have to pass a pointer to a des_key_schedule instead of a
2817 plain des_key_schedule (which was actually always a pointer
2820 des_key_schedule ks;
2822 des_set_key_checked(..., &ks);
2823 des_ncbc_encrypt(..., &ks, ...);
2825 (Note that a later change renames 'des_...' into 'DES_...'.)
2828 *) Initial reduction of linker bloat: the use of some functions, such as
2829 PEM causes large amounts of unused functions to be linked in due to
2830 poor organisation. For example pem_all.c contains every PEM function
2831 which has a knock on effect of linking in large amounts of (unused)
2832 ASN1 code. Grouping together similar functions and splitting unrelated
2833 functions prevents this.
2836 *) Cleanup of EVP macros.
2839 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2840 correct _ecb suffix.
2843 *) Add initial OCSP responder support to ocsp application. The
2844 revocation information is handled using the text based index
2845 use by the ca application. The responder can either handle
2846 requests generated internally, supplied in files (for example
2847 via a CGI script) or using an internal minimal server.
2850 *) Add configuration choices to get zlib compression for TLS.
2853 *) Changes to Kerberos SSL for RFC 2712 compliance:
2854 1. Implemented real KerberosWrapper, instead of just using
2855 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2856 2. Implemented optional authenticator field of KerberosWrapper.
2858 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2859 and authenticator structs; see crypto/krb5/.
2861 Generalized Kerberos calls to support multiple Kerberos libraries.
2862 [Vern Staats <staatsvr@asc.hpc.mil>,
2863 Jeffrey Altman <jaltman@columbia.edu>
2864 via Richard Levitte]
2866 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2867 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2868 values for each of the key sizes rather than having just
2869 parameters (and 'speed' generating keys each time).
2872 *) Speed up EVP routines.
2875 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2876 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2877 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2878 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2880 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2881 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2882 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2885 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2887 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2890 *) Added the OS2-EMX target.
2891 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2893 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2894 to support NCONF routines in extension code. New function CONF_set_nconf()
2895 to allow functions which take an NCONF to also handle the old LHASH
2896 structure: this means that the old CONF compatible routines can be
2897 retained (in particular wrt extensions) without having to duplicate the
2898 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2901 *) Enhance the general user interface with mechanisms for inner control
2902 and with possibilities to have yes/no kind of prompts.
2905 *) Change all calls to low level digest routines in the library and
2906 applications to use EVP. Add missing calls to HMAC_cleanup() and
2907 don't assume HMAC_CTX can be copied using memcpy().
2908 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2910 *) Add the possibility to control engines through control names but with
2911 arbitrary arguments instead of just a string.
2912 Change the key loaders to take a UI_METHOD instead of a callback
2913 function pointer. NOTE: this breaks binary compatibility with earlier
2914 versions of OpenSSL [engine].
2915 Adapt the nCipher code for these new conditions and add a card insertion
2919 *) Enhance the general user interface with mechanisms to better support
2920 dialog box interfaces, application-defined prompts, the possibility
2921 to use defaults (for example default passwords from somewhere else)
2922 and interrupts/cancellations.
2925 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2926 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2929 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2930 tidy up some unnecessarily weird code in 'sk_new()').
2931 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2933 *) Change the key loading routines for ENGINEs to use the same kind
2934 callback (pem_password_cb) as all other routines that need this
2938 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2939 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2940 than this minimum value is recommended.
2943 *) New random seeder for OpenVMS, using the system process statistics
2944 that are easily reachable.
2947 *) Windows apparently can't transparently handle global
2948 variables defined in DLLs. Initialisations such as:
2950 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2952 wont compile. This is used by the any applications that need to
2953 declare their own ASN1 modules. This was fixed by adding the option
2954 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2955 needed for static libraries under Win32.
2958 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2959 setting of purpose and trust fields. New X509_STORE trust and
2960 purpose functions and tidy up setting in other SSL functions.
2963 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2964 structure. These are inherited by X509_STORE_CTX when it is
2965 initialised. This allows various defaults to be set in the
2966 X509_STORE structure (such as flags for CRL checking and custom
2967 purpose or trust settings) for functions which only use X509_STORE_CTX
2968 internally such as S/MIME.
2970 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2971 trust settings if they are not set in X509_STORE. This allows X509_STORE
2972 purposes and trust (in S/MIME for example) to override any set by default.
2974 Add command line options for CRL checking to smime, s_client and s_server
2978 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2979 are set then the CRL is looked up in the X509_STORE structure and
2980 its validity and signature checked, then if the certificate is found
2981 in the CRL the verify fails with a revoked error.
2983 Various new CRL related callbacks added to X509_STORE_CTX structure.
2985 Command line options added to 'verify' application to support this.
2987 This needs some additional work, such as being able to handle multiple
2988 CRLs with different times, extension based lookup (rather than just
2989 by subject name) and ultimately more complete V2 CRL extension
2993 *) Add a general user interface API (crypto/ui/). This is designed
2994 to replace things like des_read_password and friends (backward
2995 compatibility functions using this new API are provided).
2996 The purpose is to remove prompting functions from the DES code
2997 section as well as provide for prompting through dialog boxes in
2998 a window system and the like.
3001 *) Add "ex_data" support to ENGINE so implementations can add state at a
3002 per-structure level rather than having to store it globally.
3005 *) Make it possible for ENGINE structures to be copied when retrieved by
3006 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3007 This causes the "original" ENGINE structure to act like a template,
3008 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3009 operational state can be localised to each ENGINE structure, despite the
3010 fact they all share the same "methods". New ENGINE structures returned in
3011 this case have no functional references and the return value is the single
3012 structural reference. This matches the single structural reference returned
3013 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3017 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3018 needs to match any other type at all we need to manually clear the
3022 *) Changes to the "openssl engine" utility to include;
3023 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3024 about an ENGINE's available control commands.
3025 - executing control commands from command line arguments using the
3026 '-pre' and '-post' switches. '-post' is only used if '-t' is
3027 specified and the ENGINE is successfully initialised. The syntax for
3028 the individual commands are colon-separated, for example;
3029 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3032 *) New dynamic control command support for ENGINEs. ENGINEs can now
3033 declare their own commands (numbers), names (strings), descriptions,
3034 and input types for run-time discovery by calling applications. A
3035 subset of these commands are implicitly classed as "executable"
3036 depending on their input type, and only these can be invoked through
3037 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3038 can be based on user input, config files, etc). The distinction is
3039 that "executable" commands cannot return anything other than a boolean
3040 result and can only support numeric or string input, whereas some
3041 discoverable commands may only be for direct use through
3042 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3043 pointers, or other custom uses. The "executable" commands are to
3044 support parameterisations of ENGINE behaviour that can be
3045 unambiguously defined by ENGINEs and used consistently across any
3046 OpenSSL-based application. Commands have been added to all the
3047 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3048 control over shared-library paths without source code alterations.
3051 *) Changed all ENGINE implementations to dynamically allocate their
3052 ENGINEs rather than declaring them statically. Apart from this being
3053 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3054 this also allows the implementations to compile without using the
3055 internal engine_int.h header.
3058 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3059 'const' value. Any code that should be able to modify a RAND_METHOD
3060 should already have non-const pointers to it (ie. they should only
3061 modify their own ones).
3064 *) Made a variety of little tweaks to the ENGINE code.
3065 - "atalla" and "ubsec" string definitions were moved from header files
3066 to C code. "nuron" string definitions were placed in variables
3067 rather than hard-coded - allowing parameterisation of these values
3068 later on via ctrl() commands.
3069 - Removed unused "#if 0"'d code.
3070 - Fixed engine list iteration code so it uses ENGINE_free() to release
3071 structural references.
3072 - Constified the RAND_METHOD element of ENGINE structures.
3073 - Constified various get/set functions as appropriate and added
3074 missing functions (including a catch-all ENGINE_cpy that duplicates
3075 all ENGINE values onto a new ENGINE except reference counts/state).
3076 - Removed NULL parameter checks in get/set functions. Setting a method
3077 or function to NULL is a way of cancelling out a previously set
3078 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3079 and doesn't justify the extra error symbols and code.
3080 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3081 flags from engine_int.h to engine.h.
3082 - Changed prototypes for ENGINE handler functions (init(), finish(),
3083 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3086 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3087 to the algorithm using long division. The binary algorithm can be
3088 used only if the modulus is odd. On 32-bit systems, it is faster
3089 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3090 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3091 up to 450 bits. In 64-bit environments, the binary algorithm
3092 appears to be advantageous for much longer moduli; here we use it
3093 for moduli up to 2048 bits.
3096 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3097 could not support the combine flag in choice fields.
3100 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3101 extensions from a certificate request to the certificate.
3104 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3105 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3106 file: this allows the display of the certificate about to be
3107 signed to be customised, to allow certain fields to be included
3108 or excluded and extension details. The old system didn't display
3109 multicharacter strings properly, omitted fields not in the policy
3110 and couldn't display additional details such as extensions.
3113 *) Function EC_POINTs_mul for multiple scalar multiplication
3114 of an arbitrary number of elliptic curve points
3115 \sum scalars[i]*points[i],
3116 optionally including the generator defined for the EC_GROUP:
3117 scalar*generator + \sum scalars[i]*points[i].
3119 EC_POINT_mul is a simple wrapper function for the typical case
3120 that the point list has just one item (besides the optional
3124 *) First EC_METHODs for curves over GF(p):
3126 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3127 operations and provides various method functions that can also
3128 operate with faster implementations of modular arithmetic.
3130 EC_GFp_mont_method() reuses most functions that are part of
3131 EC_GFp_simple_method, but uses Montgomery arithmetic.
3133 [Bodo Moeller; point addition and point doubling
3134 implementation directly derived from source code provided by
3135 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3137 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3138 crypto/ec/ec_lib.c):
3140 Curves are EC_GROUP objects (with an optional group generator)
3141 based on EC_METHODs that are built into the library.
3143 Points are EC_POINT objects based on EC_GROUP objects.
3145 Most of the framework would be able to handle curves over arbitrary
3146 finite fields, but as there are no obvious types for fields other
3147 than GF(p), some functions are limited to that for now.
3150 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3151 that the file contains a complete HTTP response.
3154 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3155 change the def and num file printf format specifier from "%-40sXXX"
3156 to "%-39s XXX". The latter will always guarantee a space after the
3157 field while the former will cause them to run together if the field
3158 is 40 of more characters long.
3161 *) Constify the cipher and digest 'method' functions and structures
3162 and modify related functions to take constant EVP_MD and EVP_CIPHER
3166 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3167 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3170 *) Modify EVP_Digest*() routines so they now return values. Although the
3171 internal software routines can never fail additional hardware versions
3175 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3177 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3178 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3181 ERR_R_NESTED_ASN1_ERROR
3183 ERR_R_MISSING_ASN1_EOS
3184 were 4 .. 9, conflicting with
3185 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3187 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3188 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3190 Add new error code 'ERR_R_INTERNAL_ERROR'.
3193 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3197 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3198 sets the subject name for a new request or supersedes the
3199 subject name in a given request. Formats that can be parsed are
3200 'CN=Some Name, OU=myOU, C=IT'
3202 'CN=Some Name/OU=myOU/C=IT'.
3204 Add options '-batch' and '-verbose' to 'openssl req'.
3205 [Massimiliano Pala <madwolf@hackmasters.net>]
3207 *) Introduce the possibility to access global variables through
3208 functions on platform were that's the best way to handle exporting
3209 global variables in shared libraries. To enable this functionality,
3210 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3211 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3212 is normally done by Configure or something similar).
3214 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3215 in the source file (foo.c) like this:
3217 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3218 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3220 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3221 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3223 OPENSSL_DECLARE_GLOBAL(int,foo);
3224 #define foo OPENSSL_GLOBAL_REF(foo)
3225 OPENSSL_DECLARE_GLOBAL(double,bar);
3226 #define bar OPENSSL_GLOBAL_REF(bar)
3228 The #defines are very important, and therefore so is including the
3229 header file everywhere where the defined globals are used.
3231 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3232 of ASN.1 items, but that structure is a bit different.
3234 The largest change is in util/mkdef.pl which has been enhanced with
3235 better and easier to understand logic to choose which symbols should
3236 go into the Windows .def files as well as a number of fixes and code
3237 cleanup (among others, algorithm keywords are now sorted
3238 lexicographically to avoid constant rewrites).
3241 *) In BN_div() keep a copy of the sign of 'num' before writing the
3242 result to 'rm' because if rm==num the value will be overwritten
3243 and produce the wrong result if 'num' is negative: this caused
3244 problems with BN_mod() and BN_nnmod().
3247 *) Function OCSP_request_verify(). This checks the signature on an
3248 OCSP request and verifies the signer certificate. The signer
3249 certificate is just checked for a generic purpose and OCSP request
3253 *) Add OCSP_check_validity() function to check the validity of OCSP
3254 responses. OCSP responses are prepared in real time and may only
3255 be a few seconds old. Simply checking that the current time lies
3256 between thisUpdate and nextUpdate max reject otherwise valid responses
3257 caused by either OCSP responder or client clock inaccuracy. Instead
3258 we allow thisUpdate and nextUpdate to fall within a certain period of
3259 the current time. The age of the response can also optionally be
3260 checked. Two new options -validity_period and -status_age added to
3264 *) If signature or public key algorithm is unrecognized print out its
3265 OID rather that just UNKNOWN.
3268 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3269 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3270 ID to be generated from the issuer certificate alone which can then be
3271 passed to OCSP_id_issuer_cmp().
3274 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3275 ASN1 modules to export functions returning ASN1_ITEM pointers
3276 instead of the ASN1_ITEM structures themselves. This adds several
3277 new macros which allow the underlying ASN1 function/structure to
3278 be accessed transparently. As a result code should not use ASN1_ITEM
3279 references directly (such as &X509_it) but instead use the relevant
3280 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3281 use of the new ASN1 code on platforms where exporting structures
3282 is problematical (for example in shared libraries) but exporting
3283 functions returning pointers to structures is not.
3286 *) Add support for overriding the generation of SSL/TLS session IDs.
3287 These callbacks can be registered either in an SSL_CTX or per SSL.
3288 The purpose of this is to allow applications to control, if they wish,
3289 the arbitrary values chosen for use as session IDs, particularly as it
3290 can be useful for session caching in multiple-server environments. A
3291 command-line switch for testing this (and any client code that wishes
3292 to use such a feature) has been added to "s_server".
3293 [Geoff Thorpe, Lutz Jaenicke]
3295 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3296 of the form '#if defined(...) || defined(...) || ...' and
3297 '#if !defined(...) && !defined(...) && ...'. This also avoids
3298 the growing number of special cases it was previously handling.
3301 *) Make all configuration macros available for application by making
3302 sure they are available in opensslconf.h, by giving them names starting
3303 with "OPENSSL_" to avoid conflicts with other packages and by making
3304 sure e_os2.h will cover all platform-specific cases together with
3306 Additionally, it is now possible to define configuration/platform-
3307 specific names (called "system identities"). In the C code, these
3308 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3309 macro with the name beginning with "OPENSSL_SYS_", which is determined
3310 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3314 *) New option -set_serial to 'req' and 'x509' this allows the serial
3315 number to use to be specified on the command line. Previously self
3316 signed certificates were hard coded with serial number 0 and the
3317 CA options of 'x509' had to use a serial number in a file which was
3321 *) New options to 'ca' utility to support V2 CRL entry extensions.
3322 Currently CRL reason, invalidity date and hold instruction are
3323 supported. Add new CRL extensions to V3 code and some new objects.
3326 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3327 disable standard block padding (aka PKCS#5 padding) in the EVP
3328 API, which was previously mandatory. This means that the data is
3329 not padded in any way and so the total length much be a multiple
3330 of the block size, otherwise an error occurs.
3333 *) Initial (incomplete) OCSP SSL support.
3336 *) New function OCSP_parse_url(). This splits up a URL into its host,
3337 port and path components: primarily to parse OCSP URLs. New -url
3338 option to ocsp utility.
3341 *) New nonce behavior. The return value of OCSP_check_nonce() now
3342 reflects the various checks performed. Applications can decide
3343 whether to tolerate certain situations such as an absent nonce
3344 in a response when one was present in a request: the ocsp application
3345 just prints out a warning. New function OCSP_add1_basic_nonce()
3346 this is to allow responders to include a nonce in a response even if
3347 the request is nonce-less.
3350 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3351 skipped when using openssl x509 multiple times on a single input file,
3352 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3355 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3356 set string type: to handle setting ASN1_TIME structures. Fix ca
3357 utility to correctly initialize revocation date of CRLs.
3360 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3361 the clients preferred ciphersuites and rather use its own preferences.
3362 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3363 Internet Explorer by ensuring unchanged hash method during stepup.
3364 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3367 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3368 to aes and add a new 'exist' option to print out symbols that don't
3372 *) Additional options to ocsp utility to allow flags to be set and
3373 additional certificates supplied.
3376 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3377 OCSP client a number of certificate to only verify the response
3381 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3382 handle the new API. Currently only ECB, CBC modes supported. Add new
3385 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3386 Encryption Standard (AES) Ciphersuites for Transport Layer
3387 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3388 not enabled by default and were not part of the "ALL" ciphersuite
3389 alias because they were not yet official; they could be
3390 explicitly requested by specifying the "AESdraft" ciphersuite
3391 group alias. In the final release of OpenSSL 0.9.7, the group
3392 alias is called "AES" and is part of "ALL".)
3393 [Ben Laurie, Steve Henson, Bodo Moeller]
3395 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3396 request to response.
3399 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3400 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3401 extract information from a certificate request. OCSP_response_create()
3402 creates a response and optionally adds a basic response structure.
3403 OCSP_basic_add1_status() adds a complete single response to a basic
3404 response and returns the OCSP_SINGLERESP structure just added (to allow
3405 extensions to be included for example). OCSP_basic_add1_cert() adds a
3406 certificate to a basic response and OCSP_basic_sign() signs a basic
3407 response with various flags. New helper functions ASN1_TIME_check()
3408 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3409 (converts ASN1_TIME to GeneralizedTime).
3412 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3413 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3414 structure from a certificate. X509_pubkey_digest() digests the public_key
3415 contents: this is used in various key identifiers.
3418 *) Make sk_sort() tolerate a NULL argument.
3419 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3421 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3422 passed by the function are trusted implicitly. If any of them signed the
3423 response then it is assumed to be valid and is not verified.
3426 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3427 to data. This was previously part of the PKCS7 ASN1 code. This
3428 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3429 [Steve Henson, reported by Kenneth R. Robinette
3430 <support@securenetterm.com>]
3432 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3433 routines: without these tracing memory leaks is very painful.
3434 Fix leaks in PKCS12 and PKCS7 routines.
3437 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3438 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3439 effectively meant GeneralizedTime would never be used. Now it
3440 is initialised to -1 but X509_time_adj() now has to check the value
3441 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3442 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3443 [Steve Henson, reported by Kenneth R. Robinette
3444 <support@securenetterm.com>]
3446 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3447 result in a zero length in the ASN1_INTEGER structure which was
3448 not consistent with the structure when d2i_ASN1_INTEGER() was used
3449 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3450 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3451 where it did not print out a minus for negative ASN1_INTEGER.
3454 *) Add summary printout to ocsp utility. The various functions which
3455 convert status values to strings have been renamed to:
3456 OCSP_response_status_str(), OCSP_cert_status_str() and
3457 OCSP_crl_reason_str() and are no longer static. New options
3458 to verify nonce values and to disable verification. OCSP response
3459 printout format cleaned up.
3462 *) Add additional OCSP certificate checks. These are those specified
3463 in RFC2560. This consists of two separate checks: the CA of the
3464 certificate being checked must either be the OCSP signer certificate
3465 or the issuer of the OCSP signer certificate. In the latter case the
3466 OCSP signer certificate must contain the OCSP signing extended key
3467 usage. This check is performed by attempting to match the OCSP
3468 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3469 in the OCSP_CERTID structures of the response.
3472 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3473 and related routines. This uses the standard OpenSSL certificate
3474 verify routines to perform initial checks (just CA validity) and
3475 to obtain the certificate chain. Then additional checks will be
3476 performed on the chain. Currently the root CA is checked to see
3477 if it is explicitly trusted for OCSP signing. This is used to set
3478 a root CA as a global signing root: that is any certificate that
3479 chains to that CA is an acceptable OCSP signing certificate.
3482 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3483 extensions from a separate configuration file.
3484 As when reading extensions from the main configuration file,
3485 the '-extensions ...' option may be used for specifying the
3487 [Massimiliano Pala <madwolf@comune.modena.it>]
3489 *) New OCSP utility. Allows OCSP requests to be generated or
3490 read. The request can be sent to a responder and the output
3491 parsed, outputed or printed in text form. Not complete yet:
3492 still needs to check the OCSP response validity.
3495 *) New subcommands for 'openssl ca':
3496 'openssl ca -status <serial>' prints the status of the cert with
3497 the given serial number (according to the index file).
3498 'openssl ca -updatedb' updates the expiry status of certificates
3500 [Massimiliano Pala <madwolf@comune.modena.it>]
3502 *) New '-newreq-nodes' command option to CA.pl. This is like
3503 '-newreq', but calls 'openssl req' with the '-nodes' option
3504 so that the resulting key is not encrypted.
3505 [Damien Miller <djm@mindrot.org>]
3507 *) New configuration for the GNU Hurd.
3508 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3510 *) Initial code to implement OCSP basic response verify. This
3511 is currently incomplete. Currently just finds the signer's
3512 certificate and verifies the signature on the response.
3515 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3516 value of OPENSSLDIR. This is available via the new '-d' option
3517 to 'openssl version', and is also included in 'openssl version -a'.
3520 *) Allowing defining memory allocation callbacks that will be given
3521 file name and line number information in additional arguments
3522 (a const char* and an int). The basic functionality remains, as
3523 well as the original possibility to just replace malloc(),
3524 realloc() and free() by functions that do not know about these
3525 additional arguments. To register and find out the current
3526 settings for extended allocation functions, the following
3527 functions are provided:
3529 CRYPTO_set_mem_ex_functions
3530 CRYPTO_set_locked_mem_ex_functions
3531 CRYPTO_get_mem_ex_functions
3532 CRYPTO_get_locked_mem_ex_functions
3534 These work the same way as CRYPTO_set_mem_functions and friends.
3535 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3536 extended allocation function is enabled.
3537 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3538 a conventional allocation function is enabled.
3539 [Richard Levitte, Bodo Moeller]
3541 *) Finish off removing the remaining LHASH function pointer casts.
3542 There should no longer be any prototype-casting required when using
3543 the LHASH abstraction, and any casts that remain are "bugs". See
3544 the callback types and macros at the head of lhash.h for details
3545 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3548 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3549 If /dev/[u]random devices are not available or do not return enough
3550 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3552 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3553 /etc/entropy will be queried once each in this sequence, quering stops
3554 when enough entropy was collected without querying more sockets.
3557 *) Change the Unix RAND_poll() variant to be able to poll several
3558 random devices, as specified by DEVRANDOM, until a sufficient amount
3559 of data has been collected. We spend at most 10 ms on each file
3560 (select timeout) and read in non-blocking mode. DEVRANDOM now
3561 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3562 (previously it was just the string "/dev/urandom"), so on typical
3563 platforms the 10 ms delay will never occur.
3564 Also separate out the Unix variant to its own file, rand_unix.c.
3565 For VMS, there's a currently-empty rand_vms.c.
3568 *) Move OCSP client related routines to ocsp_cl.c. These
3569 provide utility functions which an application needing
3570 to issue a request to an OCSP responder and analyse the
3571 response will typically need: as opposed to those which an
3572 OCSP responder itself would need which will be added later.
3574 OCSP_request_sign() signs an OCSP request with an API similar
3575 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3576 response. OCSP_response_get1_basic() extracts basic response
3577 from response. OCSP_resp_find_status(): finds and extracts status
3578 information from an OCSP_CERTID structure (which will be created
3579 when the request structure is built). These are built from lower
3580 level functions which work on OCSP_SINGLERESP structures but
3581 wont normally be used unless the application wishes to examine
3582 extensions in the OCSP response for example.
3584 Replace nonce routines with a pair of functions.
3585 OCSP_request_add1_nonce() adds a nonce value and optionally
3586 generates a random value. OCSP_check_nonce() checks the
3587 validity of the nonce in an OCSP response.
3590 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3591 This doesn't copy the supplied OCSP_CERTID and avoids the
3592 need to free up the newly created id. Change return type
3593 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3594 This can then be used to add extensions to the request.
3595 Deleted OCSP_request_new(), since most of its functionality
3596 is now in OCSP_REQUEST_new() (and the case insensitive name
3597 clash) apart from the ability to set the request name which
3598 will be added elsewhere.
3601 *) Update OCSP API. Remove obsolete extensions argument from
3602 various functions. Extensions are now handled using the new
3603 OCSP extension code. New simple OCSP HTTP function which
3604 can be used to send requests and parse the response.
3607 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3608 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3609 uses the special reorder version of SET OF to sort the attributes
3610 and reorder them to match the encoded order. This resolves a long
3611 standing problem: a verify on a PKCS7 structure just after signing
3612 it used to fail because the attribute order did not match the
3613 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3614 it uses the received order. This is necessary to tolerate some broken
3615 software that does not order SET OF. This is handled by encoding
3616 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3617 to produce the required SET OF.
3620 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3621 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3622 files to get correct declarations of the ASN.1 item variables.
3625 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3626 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3627 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3628 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3629 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3630 ASN1_ITEM and no wrapper functions.
3633 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3634 replace the old function pointer based I/O routines. Change most of
3635 the *_d2i_bio() and *_d2i_fp() functions to use these.
3638 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3639 lines, recognice more "algorithms" that can be deselected, and make
3640 it complain about algorithm deselection that isn't recognised.
3643 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3644 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3645 to use new functions. Add NO_ASN1_OLD which can be set to remove
3646 some old style ASN1 functions: this can be used to determine if old
3647 code will still work when these eventually go away.
3650 *) New extension functions for OCSP structures, these follow the
3651 same conventions as certificates and CRLs.
3654 *) New function X509V3_add1_i2d(). This automatically encodes and
3655 adds an extension. Its behaviour can be customised with various
3656 flags to append, replace or delete. Various wrappers added for
3657 certifcates and CRLs.
3660 *) Fix to avoid calling the underlying ASN1 print routine when
3661 an extension cannot be parsed. Correct a typo in the
3662 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3665 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3666 entries for variables.
3669 *) Add functionality to apps/openssl.c for detecting locking
3670 problems: As the program is single-threaded, all we have
3671 to do is register a locking callback using an array for
3672 storing which locks are currently held by the program.
3675 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3676 SSL_get_ex_data_X509_STORE_idx(), which is used in
3677 ssl_verify_cert_chain() and thus can be called at any time
3678 during TLS/SSL handshakes so that thread-safety is essential.
3679 Unfortunately, the ex_data design is not at all suited
3680 for multi-threaded use, so it probably should be abolished.
3683 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3684 [Broadcom, tweaked and integrated by Geoff Thorpe]
3686 *) Move common extension printing code to new function
3687 X509V3_print_extensions(). Reorganise OCSP print routines and
3688 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3691 *) New function X509_signature_print() to remove duplication in some
3695 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3696 set (this was treated exactly the same as SET OF previously). This
3697 is used to reorder the STACK representing the structure to match the
3698 encoding. This will be used to get round a problem where a PKCS7
3699 structure which was signed could not be verified because the STACK
3700 order did not reflect the encoded order.
3703 *) Reimplement the OCSP ASN1 module using the new code.
3706 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3707 for its ASN1 operations. The old style function pointers still exist
3708 for now but they will eventually go away.
3711 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3712 completely replaces the old ASN1 functionality with a table driven
3713 encoder and decoder which interprets an ASN1_ITEM structure describing
3714 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3715 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3716 has also been converted to the new form.
3719 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3720 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3721 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3722 for negative moduli.
3725 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3726 of not touching the result's sign bit.
3729 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3733 *) Changed the LHASH code to use prototypes for callbacks, and created
3734 macros to declare and implement thin (optionally static) functions
3735 that provide type-safety and avoid function pointer casting for the
3736 type-specific callbacks.
3739 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3741 [Veers Staats <staatsvr@asc.hpc.mil>,
3742 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3744 *) Reformat the FAQ so the different questions and answers can be divided
3745 in sections depending on the subject.
3748 *) Have the zlib compression code load ZLIB.DLL dynamically under
3752 *) New function BN_mod_sqrt for computing square roots modulo a prime
3753 (using the probabilistic Tonelli-Shanks algorithm unless
3754 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3755 be handled deterministically).
3756 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3758 *) Make BN_mod_inverse faster by explicitly handling small quotients
3759 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3760 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3763 *) New function BN_kronecker.
3766 *) Fix BN_gcd so that it works on negative inputs; the result is
3767 positive unless both parameters are zero.
3768 Previously something reasonably close to an infinite loop was
3769 possible because numbers could be growing instead of shrinking
3770 in the implementation of Euclid's algorithm.
3773 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3774 sign of the number in question.
3776 Fix BN_is_word(a,w) to work correctly for w == 0.
3778 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3779 because its test if the absolute value of 'a' equals 'w'.
3780 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3781 it exists mostly for use in the implementations of BN_is_zero(),
3782 BN_is_one(), and BN_is_word().
3785 *) New function BN_swap.
3788 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3789 the exponentiation functions are more likely to produce reasonable
3790 results on negative inputs.
3793 *) Change BN_mod_mul so that the result is always non-negative.
3794 Previously, it could be negative if one of the factors was negative;
3795 I don't think anyone really wanted that behaviour.
3798 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3799 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3800 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3801 and add new functions:
3810 BN_mod_lshift1_quick
3814 These functions always generate non-negative results.
3816 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3817 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3819 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3820 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3821 be reduced modulo m.
3822 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3825 The following entry accidentily appeared in the CHANGES file
3826 distributed with OpenSSL 0.9.7. The modifications described in
3827 it do *not* apply to OpenSSL 0.9.7.
3829 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3830 was actually never needed) and in BN_mul(). The removal in BN_mul()
3831 required a small change in bn_mul_part_recursive() and the addition
3832 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3833 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3834 bn_sub_words() and bn_add_words() except they take arrays with
3839 *) In 'openssl passwd', verify passwords read from the terminal
3840 unless the '-salt' option is used (which usually means that
3841 verification would just waste user's time since the resulting
3842 hash is going to be compared with some given password hash)
3843 or the new '-noverify' option is used.
3845 This is an incompatible change, but it does not affect
3846 non-interactive use of 'openssl passwd' (passwords on the command
3847 line, '-stdin' option, '-in ...' option) and thus should not
3851 *) Remove all references to RSAref, since there's no more need for it.
3854 *) Make DSO load along a path given through an environment variable
3855 (SHLIB_PATH) with shl_load().
3858 *) Constify the ENGINE code as a result of BIGNUM constification.
3859 Also constify the RSA code and most things related to it. In a
3860 few places, most notable in the depth of the ASN.1 code, ugly
3861 casts back to non-const were required (to be solved at a later
3865 *) Make it so the openssl application has all engines loaded by default.
3868 *) Constify the BIGNUM routines a little more.
3871 *) Add the following functions:
3873 ENGINE_load_cswift()
3875 ENGINE_load_atalla()
3877 ENGINE_load_builtin_engines()
3879 That way, an application can itself choose if external engines that
3880 are built-in in OpenSSL shall ever be used or not. The benefit is
3881 that applications won't have to be linked with libdl or other dso
3882 libraries unless it's really needed.
3884 Changed 'openssl engine' to load all engines on demand.
3885 Changed the engine header files to avoid the duplication of some
3886 declarations (they differed!).
3889 *) 'openssl engine' can now list capabilities.
3892 *) Better error reporting in 'openssl engine'.
3895 *) Never call load_dh_param(NULL) in s_server.
3898 *) Add engine application. It can currently list engines by name and
3899 identity, and test if they are actually available.
3902 *) Improve RPM specification file by forcing symbolic linking and making
3903 sure the installed documentation is also owned by root.root.
3904 [Damien Miller <djm@mindrot.org>]
3906 *) Give the OpenSSL applications more possibilities to make use of
3907 keys (public as well as private) handled by engines.
3910 *) Add OCSP code that comes from CertCo.
3913 *) Add VMS support for the Rijndael code.
3916 *) Added untested support for Nuron crypto accelerator.
3919 *) Add support for external cryptographic devices. This code was
3920 previously distributed separately as the "engine" branch.
3921 [Geoff Thorpe, Richard Levitte]
3923 *) Rework the filename-translation in the DSO code. It is now possible to
3924 have far greater control over how a "name" is turned into a filename
3925 depending on the operating environment and any oddities about the
3926 different shared library filenames on each system.
3929 *) Support threads on FreeBSD-elf in Configure.
3932 *) Fix for SHA1 assembly problem with MASM: it produces
3933 warnings about corrupt line number information when assembling
3934 with debugging information. This is caused by the overlapping
3936 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3939 NCONF_get_number() has no error checking at all. As a replacement,
3940 NCONF_get_number_e() is defined (_e for "error checking") and is
3941 promoted strongly. The old NCONF_get_number is kept around for
3942 binary backward compatibility.
3943 Make it possible for methods to load from something other than a BIO,
3944 by providing a function pointer that is given a name instead of a BIO.
3945 For example, this could be used to load configuration data from an
3949 *) Fix for non blocking accept BIOs. Added new I/O special reason
3950 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3951 with non blocking I/O was not possible because no retry code was
3952 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3956 *) Added the beginnings of Rijndael support.
3959 *) Fix for bug in DirectoryString mask setting. Add support for
3960 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3961 to allow certificate printing to more controllable, additional
3962 'certopt' option to 'x509' to allow new printing options to be
3966 *) Clean old EAY MD5 hack from e_os.h.
3969 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3971 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3972 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3973 [Joe Orton, Steve Henson]
3975 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3977 *) Fix additional bug revealed by the NISCC test suite:
3979 Stop bug triggering large recursion when presented with
3980 certain ASN.1 tags (CVE-2003-0851)
3983 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3985 *) Fix various bugs revealed by running the NISCC test suite:
3987 Stop out of bounds reads in the ASN1 code when presented with
3988 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3990 If verify callback ignores invalid public key errors don't try to check
3991 certificate signature with the NULL public key.
3995 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3996 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4000 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4001 extra data after the compression methods not only for TLS 1.0
4002 but also for SSL 3.0 (as required by the specification).
4003 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4005 *) Change X509_certificate_type() to mark the key as exported/exportable
4006 when it's 512 *bits* long, not 512 bytes.
4009 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4011 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4012 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4013 a protocol version number mismatch like a decryption error
4014 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4017 *) Turn on RSA blinding by default in the default implementation
4018 to avoid a timing attack. Applications that don't want it can call
4019 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4020 They would be ill-advised to do so in most cases.
4021 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4023 *) Change RSA blinding code so that it works when the PRNG is not
4024 seeded (in this case, the secret RSA exponent is abused as
4025 an unpredictable seed -- if it is not unpredictable, there
4026 is no point in blinding anyway). Make RSA blinding thread-safe
4027 by remembering the creator's thread ID in rsa->blinding and
4028 having all other threads use local one-time blinding factors
4029 (this requires more computation than sharing rsa->blinding, but
4030 avoids excessive locking; and if an RSA object is not shared
4031 between threads, blinding will still be very fast).
4034 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4036 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4037 via timing by performing a MAC computation even if incorrrect
4038 block cipher padding has been found. This is a countermeasure
4039 against active attacks where the attacker has to distinguish
4040 between bad padding and a MAC verification error. (CVE-2003-0078)
4042 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4043 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4044 Martin Vuagnoux (EPFL, Ilion)]
4046 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4048 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4049 memory from it's contents. This is done with a counter that will
4050 place alternating values in each byte. This can be used to solve
4051 two issues: 1) the removal of calls to memset() by highly optimizing
4052 compilers, and 2) cleansing with other values than 0, since those can
4053 be read through on certain media, for example a swap space on disk.
4056 *) Bugfix: client side session caching did not work with external caching,
4057 because the session->cipher setting was not restored when reloading
4058 from the external cache. This problem was masked, when
4059 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4060 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4063 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4064 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4065 [Zeev Lieber <zeev-l@yahoo.com>]
4067 *) Undo an undocumented change introduced in 0.9.6e which caused
4068 repeated calls to OpenSSL_add_all_ciphers() and
4069 OpenSSL_add_all_digests() to be ignored, even after calling
4073 *) Change the default configuration reader to deal with last line not
4074 being properly terminated.
4077 *) Change X509_NAME_cmp() so it applies the special rules on handling
4078 DN values that are of type PrintableString, as well as RDNs of type
4079 emailAddress where the value has the type ia5String.
4080 [stefank@valicert.com via Richard Levitte]
4082 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4083 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4084 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4085 the bitwise-OR of the two for use by the majority of applications
4086 wanting this behaviour, and update the docs. The documented
4087 behaviour and actual behaviour were inconsistent and had been
4088 changing anyway, so this is more a bug-fix than a behavioural
4090 [Geoff Thorpe, diagnosed by Nadav Har'El]
4092 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4093 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4096 *) Fix initialization code race conditions in
4097 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4098 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4099 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4100 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4101 ssl2_get_cipher_by_char(),
4102 ssl3_get_cipher_by_char().
4103 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4105 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4106 the cached sessions are flushed, as the remove_cb() might use ex_data
4107 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4108 (see [openssl.org #212]).
4109 [Geoff Thorpe, Lutz Jaenicke]
4111 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4112 length, instead of the encoding length to d2i_ASN1_OBJECT.
4115 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4117 *) [In 0.9.6g-engine release:]
4118 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4119 [Lynn Gazis <lgazis@rainbow.com>]
4121 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4123 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4124 and get fix the header length calculation.
4125 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4126 Alon Kantor <alonk@checkpoint.com> (and others),
4129 *) Use proper error handling instead of 'assertions' in buffer
4130 overflow checks added in 0.9.6e. This prevents DoS (the
4131 assertions could call abort()).
4132 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4134 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4136 *) Add various sanity checks to asn1_get_length() to reject
4137 the ASN1 length bytes if they exceed sizeof(long), will appear
4138 negative or the content length exceeds the length of the
4140 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4142 *) Fix cipher selection routines: ciphers without encryption had no flags
4143 for the cipher strength set and where therefore not handled correctly
4144 by the selection routines (PR #130).
4147 *) Fix EVP_dsa_sha macro.
4151 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4152 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4153 that was added in OpenSSL 0.9.6d.
4155 As the countermeasure turned out to be incompatible with some
4156 broken SSL implementations, the new option is part of SSL_OP_ALL.
4157 SSL_OP_ALL is usually employed when compatibility with weird SSL
4158 implementations is desired (e.g. '-bugs' option to 's_client' and
4159 's_server'), so the new option is automatically set in many
4163 *) Changes in security patch:
4165 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4166 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4167 Air Force Materiel Command, USAF, under agreement number
4170 *) Add various sanity checks to asn1_get_length() to reject
4171 the ASN1 length bytes if they exceed sizeof(long), will appear
4172 negative or the content length exceeds the length of the
4173 supplied buffer. (CVE-2002-0659)
4174 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4176 *) Assertions for various potential buffer overflows, not known to
4178 [Ben Laurie (CHATS)]
4180 *) Various temporary buffers to hold ASCII versions of integers were
4181 too small for 64 bit platforms. (CVE-2002-0655)
4182 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4184 *) Remote buffer overflow in SSL3 protocol - an attacker could
4185 supply an oversized session ID to a client. (CVE-2002-0656)
4186 [Ben Laurie (CHATS)]
4188 *) Remote buffer overflow in SSL2 protocol - an attacker could
4189 supply an oversized client master key. (CVE-2002-0656)
4190 [Ben Laurie (CHATS)]
4192 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4194 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4195 encoded as NULL) with id-dsa-with-sha1.
4196 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4198 *) Check various X509_...() return values in apps/req.c.
4199 [Nils Larsch <nla@trustcenter.de>]
4201 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4202 an end-of-file condition would erronously be flagged, when the CRLF
4203 was just at the end of a processed block. The bug was discovered when
4204 processing data through a buffering memory BIO handing the data to a
4205 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4206 <ptsekov@syntrex.com> and Nedelcho Stanev.
4209 *) Implement a countermeasure against a vulnerability recently found
4210 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4211 before application data chunks to avoid the use of known IVs
4212 with data potentially chosen by the attacker.
4215 *) Fix length checks in ssl3_get_client_hello().
4218 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4219 to prevent ssl3_read_internal() from incorrectly assuming that
4220 ssl3_read_bytes() found application data while handshake
4221 processing was enabled when in fact s->s3->in_read_app_data was
4222 merely automatically cleared during the initial handshake.
4223 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4225 *) Fix object definitions for Private and Enterprise: they were not
4226 recognized in their shortname (=lowercase) representation. Extend
4227 obj_dat.pl to issue an error when using undefined keywords instead
4228 of silently ignoring the problem (Svenning Sorensen
4229 <sss@sss.dnsalias.net>).
4232 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4233 generators, i.e. generators other than 2 and 5. (Previously, the
4234 code did not properly initialise the 'add' and 'rem' values to
4235 BN_generate_prime().)
4237 In the new general case, we do not insist that 'generator' is
4238 actually a primitive root: This requirement is rather pointless;
4239 a generator of the order-q subgroup is just as good, if not
4243 *) Map new X509 verification errors to alerts. Discovered and submitted by
4244 Tom Wu <tom@arcot.com>.
4247 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4248 returning non-zero before the data has been completely received
4249 when using non-blocking I/O.
4250 [Bodo Moeller; problem pointed out by John Hughes]
4252 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4253 [Ben Laurie, Lutz Jaenicke]
4255 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4256 Yoram Zahavi <YoramZ@gilian.com>).
4259 *) Add information about CygWin 1.3 and on, and preserve proper
4260 configuration for the versions before that.
4261 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4263 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4264 check whether we deal with a copy of a session and do not delete from
4265 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4266 <izhar@checkpoint.com>.
4269 *) Do not store session data into the internal session cache, if it
4270 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4271 flag is set). Proposed by Aslam <aslam@funk.com>.
4274 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4278 *) [In 0.9.6d-engine release:]
4279 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4280 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4282 *) Add the configuration target linux-s390x.
4283 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4285 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4286 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4287 variable as an indication that a ClientHello message has been
4288 received. As the flag value will be lost between multiple
4289 invocations of ssl3_accept when using non-blocking I/O, the
4290 function may not be aware that a handshake has actually taken
4291 place, thus preventing a new session from being added to the
4294 To avoid this problem, we now set s->new_session to 2 instead of
4295 using a local variable.
4296 [Lutz Jaenicke, Bodo Moeller]
4298 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4299 if the SSL_R_LENGTH_MISMATCH error is detected.
4300 [Geoff Thorpe, Bodo Moeller]
4302 *) New 'shared_ldflag' column in Configure platform table.
4305 *) Fix EVP_CIPHER_mode macro.
4306 ["Dan S. Camper" <dan@bti.net>]
4308 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4309 type, we must throw them away by setting rr->length to 0.
4310 [D P Chang <dpc@qualys.com>]
4312 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4314 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4315 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4316 worked incorrectly for those cases where range = 10..._2 and
4317 3*range is two bits longer than range.)
4320 *) Only add signing time to PKCS7 structures if it is not already
4324 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4325 OBJ_ld_ce should be OBJ_id_ce.
4326 Also some ip-pda OIDs in crypto/objects/objects.txt were
4327 incorrect (cf. RFC 3039).
4328 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4330 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4331 returns early because it has nothing to do.
4332 [Andy Schneider <andy.schneider@bjss.co.uk>]
4334 *) [In 0.9.6c-engine release:]
4335 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4336 [Andy Schneider <andy.schneider@bjss.co.uk>]
4338 *) [In 0.9.6c-engine release:]
4339 Add support for Cryptographic Appliance's keyserver technology.
4340 (Use engine 'keyclient')
4341 [Cryptographic Appliances and Geoff Thorpe]
4343 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4344 is called via tools/c89.sh because arguments have to be
4345 rearranged (all '-L' options must appear before the first object
4347 [Richard Shapiro <rshapiro@abinitio.com>]
4349 *) [In 0.9.6c-engine release:]
4350 Add support for Broadcom crypto accelerator cards, backported
4352 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4354 *) [In 0.9.6c-engine release:]
4355 Add support for SureWare crypto accelerator cards from
4356 Baltimore Technologies. (Use engine 'sureware')
4357 [Baltimore Technologies and Mark Cox]
4359 *) [In 0.9.6c-engine release:]
4360 Add support for crypto accelerator cards from Accelerated
4361 Encryption Processing, www.aep.ie. (Use engine 'aep')
4362 [AEP Inc. and Mark Cox]
4364 *) Add a configuration entry for gcc on UnixWare.
4365 [Gary Benson <gbenson@redhat.com>]
4367 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4368 messages are stored in a single piece (fixed-length part and
4369 variable-length part combined) and fix various bugs found on the way.
4372 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4373 instead. BIO_gethostbyname() does not know what timeouts are
4374 appropriate, so entries would stay in cache even when they have
4376 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4378 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4379 faced with a pathologically small ClientHello fragment that does
4380 not contain client_version: Instead of aborting with an error,
4381 simply choose the highest available protocol version (i.e.,
4382 TLS 1.0 unless it is disabled). In practice, ClientHello
4383 messages are never sent like this, but this change gives us
4384 strictly correct behaviour at least for TLS.
4387 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4388 never resets s->method to s->ctx->method when called from within
4389 one of the SSL handshake functions.
4390 [Bodo Moeller; problem pointed out by Niko Baric]
4392 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4393 (sent using the client's version number) if client_version is
4394 smaller than the protocol version in use. Also change
4395 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4396 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4397 the client will at least see that alert.
4400 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4404 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4405 client receives HelloRequest while in a handshake.
4406 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4408 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4409 should end in 'break', not 'goto end' which circuments various
4410 cleanups done in state SSL_ST_OK. But session related stuff
4411 must be disabled for SSL_ST_OK in the case that we just sent a
4414 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4415 before just sending a HelloRequest.
4416 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4418 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4419 reveal whether illegal block cipher padding was found or a MAC
4420 verification error occured. (Neither SSLerr() codes nor alerts
4421 are directly visible to potential attackers, but the information
4422 may leak via logfiles.)
4424 Similar changes are not required for the SSL 2.0 implementation
4425 because the number of padding bytes is sent in clear for SSL 2.0,
4426 and the extra bytes are just ignored. However ssl/s2_pkt.c
4427 failed to verify that the purported number of padding bytes is in
4431 *) Add OpenUNIX-8 support including shared libraries
4432 (Boyd Lynn Gerber <gerberb@zenez.com>).
4435 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4436 'wristwatch attack' using huge encoding parameters (cf.
4437 James H. Manger's CRYPTO 2001 paper). Note that the
4438 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4439 encoding parameters and hence was not vulnerable.
4442 *) BN_sqr() bug fix.
4443 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4445 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4446 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4447 followed by modular reduction.
4448 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4450 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4451 equivalent based on BN_pseudo_rand() instead of BN_rand().
4454 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4455 This function was broken, as the check for a new client hello message
4456 to handle SGC did not allow these large messages.
4457 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4460 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4463 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4464 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4467 *) Rework the configuration and shared library support for Tru64 Unix.
4468 The configuration part makes use of modern compiler features and
4469 still retains old compiler behavior for those that run older versions
4470 of the OS. The shared library support part includes a variant that
4471 uses the RPATH feature, and is available through the special
4472 configuration target "alpha-cc-rpath", which will never be selected
4474 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4476 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4477 with the same message size as in ssl3_get_certificate_request().
4478 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4479 messages might inadvertently be reject as too long.
4480 [Petr Lampa <lampa@fee.vutbr.cz>]
4482 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4485 *) Modified SSL library such that the verify_callback that has been set
4486 specificly for an SSL object with SSL_set_verify() is actually being
4487 used. Before the change, a verify_callback set with this function was
4488 ignored and the verify_callback() set in the SSL_CTX at the time of
4489 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4490 to allow the necessary settings.
4493 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4494 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4495 done automatically (in contradiction to the requirements of the C
4496 standard). This made problems when used from OpenSSH.
4499 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4500 dh->length and always used
4502 BN_rand_range(priv_key, dh->p).
4504 BN_rand_range() is not necessary for Diffie-Hellman, and this
4505 specific range makes Diffie-Hellman unnecessarily inefficient if
4506 dh->length (recommended exponent length) is much smaller than the
4507 length of dh->p. We could use BN_rand_range() if the order of
4508 the subgroup was stored in the DH structure, but we only have
4513 BN_rand(priv_key, l, ...)
4515 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4521 RSA_eay_public_encrypt
4522 RSA_eay_private_decrypt
4523 RSA_eay_private_encrypt (signing)
4524 RSA_eay_public_decrypt (signature verification)
4526 (default implementations for RSA_public_encrypt,
4527 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4528 always reject numbers >= n.
4531 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4532 to synchronize access to 'locking_thread'. This is necessary on
4533 systems where access to 'locking_thread' (an 'unsigned long'
4534 variable) is not atomic.
4537 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4538 *before* setting the 'crypto_lock_rand' flag. The previous code had
4539 a race condition if 0 is a valid thread ID.
4540 [Travis Vitek <vitek@roguewave.com>]
4542 *) Add support for shared libraries under Irix.
4543 [Albert Chin-A-Young <china@thewrittenword.com>]
4545 *) Add configuration option to build on Linux on both big-endian and
4547 [Ralf Baechle <ralf@uni-koblenz.de>]
4549 *) Add the possibility to create shared libraries on HP-UX.
4552 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4554 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4555 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4556 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4557 PRNG state recovery was possible based on the output of
4558 one PRNG request appropriately sized to gain knowledge on
4559 'md' followed by enough consecutive 1-byte PRNG requests
4560 to traverse all of 'state'.
4562 1. When updating 'md_local' (the current thread's copy of 'md')
4563 during PRNG output generation, hash all of the previous
4564 'md_local' value, not just the half used for PRNG output.
4566 2. Make the number of bytes from 'state' included into the hash
4567 independent from the number of PRNG bytes requested.
4569 The first measure alone would be sufficient to avoid
4570 Markku-Juhani's attack. (Actually it had never occurred
4571 to me that the half of 'md_local' used for chaining was the
4572 half from which PRNG output bytes were taken -- I had always
4573 assumed that the secret half would be used.) The second
4574 measure makes sure that additional data from 'state' is never
4575 mixed into 'md_local' in small portions; this heuristically
4576 further strengthens the PRNG.
4579 *) Fix crypto/bn/asm/mips3.s.
4582 *) When only the key is given to "enc", the IV is undefined. Print out
4583 an error message in this case.
4586 *) Handle special case when X509_NAME is empty in X509 printing routines.
4589 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4590 positive and less than q.
4593 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4594 used: it isn't thread safe and the add_lock_callback should handle
4596 [Paul Rose <Paul.Rose@bridge.com>]
4598 *) Verify that incoming data obeys the block size in
4599 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4603 [Ulf Möller, Bodo Möller]
4605 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4606 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4607 when fixing the server behaviour for backwards-compatible 'client
4608 hello' messages. (Note that the attack is impractical against
4609 SSL 3.0 and TLS 1.0 anyway because length and version checking
4610 means that the probability of guessing a valid ciphertext is
4611 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4614 Before 0.9.5, the countermeasure (hide the error by generating a
4615 random 'decryption result') did not work properly because
4616 ERR_clear_error() was missing, meaning that SSL_get_error() would
4617 detect the supposedly ignored error.
4619 Both problems are now fixed.
4622 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4623 (previously it was 1024).
4626 *) Fix for compatibility mode trust settings: ignore trust settings
4627 unless some valid trust or reject settings are present.
4630 *) Fix for blowfish EVP: its a variable length cipher.
4633 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4634 parameters in DSA public key structures and return an error in the
4635 DSA routines if parameters are absent.
4638 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4639 in the current directory if neither $RANDFILE nor $HOME was set.
4640 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4641 caused some confusion to Windows users who haven't defined $HOME.
4642 Thus RAND_file_name() is changed again: e_os.h can define a
4643 DEFAULT_HOME, which will be used if $HOME is not set.
4644 For Windows, we use "C:"; on other platforms, we still require
4645 environment variables.
4647 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4648 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4649 having multiple threads call RAND_poll() concurrently.
4652 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4653 combination of a flag and a thread ID variable.
4654 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4655 flag), *other* threads can enter ssleay_add_bytes without obeying
4656 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4657 that they do not hold after the first thread unsets add_do_not_lock).
4660 *) Change bctest again: '-x' expressions are not available in all
4664 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4666 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4667 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4669 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4670 the default extension for executables, if any. Also, make the perl
4671 scripts that use symlink() to test if it really exists and use "cp"
4672 if it doesn't. All this made OpenSSL compilable and installable in
4676 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4677 If SEQUENCE is length is indefinite just set c->slen to the total
4678 amount of data available.
4679 [Steve Henson, reported by shige@FreeBSD.org]
4680 [This change does not apply to 0.9.7.]
4682 *) Change bctest to avoid here-documents inside command substitution
4683 (workaround for FreeBSD /bin/sh bug).
4684 For compatibility with Ultrix, avoid shell functions (introduced
4685 in the bctest version that searches along $PATH).
4688 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4689 with des_encrypt() defined on some operating systems, like Solaris
4693 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4694 On the Importance of Eliminating Errors in Cryptographic
4695 Computations, J. Cryptology 14 (2001) 2, 101-119,
4696 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4699 *) MIPS assembler BIGNUM division bug fix.
4702 *) Disabled incorrect Alpha assembler code.
4705 *) Fix PKCS#7 decode routines so they correctly update the length
4706 after reading an EOC for the EXPLICIT tag.
4708 [This change does not apply to 0.9.7.]
4710 *) Fix bug in PKCS#12 key generation routines. This was triggered
4711 if a 3DES key was generated with a 0 initial byte. Include
4712 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4713 (but broken) behaviour.
4716 *) Enhance bctest to search for a working bc along $PATH and print
4718 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4720 *) Fix memory leaks in err.c: free err_data string if necessary;
4721 don't write to the wrong index in ERR_set_error_data.
4724 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4728 *) Replace rdtsc with _emit statements for VC++ version 5.
4729 [Jeremy Cooper <jeremy@baymoo.org>]
4731 *) Make it possible to reuse SSLv2 sessions.
4734 *) In copy_email() check for >= 0 as a return value for
4735 X509_NAME_get_index_by_NID() since 0 is a valid index.
4736 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4738 *) Avoid coredump with unsupported or invalid public keys by checking if
4739 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4740 PKCS7_verify() fails with non detached data.
4743 *) Don't use getenv in library functions when run as setuid/setgid.
4744 New function OPENSSL_issetugid().
4747 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4748 due to incorrect handling of multi-threading:
4750 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4752 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4754 3. Count how many times MemCheck_off() has been called so that
4755 nested use can be treated correctly. This also avoids
4756 inband-signalling in the previous code (which relied on the
4757 assumption that thread ID 0 is impossible).
4760 *) Add "-rand" option also to s_client and s_server.
4763 *) Fix CPU detection on Irix 6.x.
4764 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4765 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4767 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4770 [This change does not apply to 0.9.7.]
4772 *) Use the cached encoding of an X509_NAME structure rather than
4773 copying it. This is apparently the reason for the libsafe "errors"
4774 but the code is actually correct.
4777 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4778 Bleichenbacher's DSA attack.
4779 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4780 to be set and top=0 forces the highest bit to be set; top=-1 is new
4781 and leaves the highest bit random.
4782 [Ulf Moeller, Bodo Moeller]
4784 *) In the NCONF_...-based implementations for CONF_... queries
4785 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4786 a temporary CONF structure with the data component set to NULL
4787 (which gives segmentation faults in lh_retrieve).
4788 Instead, use NULL for the CONF pointer in CONF_get_string and
4789 CONF_get_number (which may use environment variables) and directly
4790 return NULL from CONF_get_section.
4793 *) Fix potential buffer overrun for EBCDIC.
4796 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4797 keyUsage if basicConstraints absent for a CA.
4800 *) Make SMIME_write_PKCS7() write mail header values with a format that
4801 is more generally accepted (no spaces before the semicolon), since
4802 some programs can't parse those values properly otherwise. Also make
4803 sure BIO's that break lines after each write do not create invalid
4807 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4808 macros previously used would not encode an empty SEQUENCE OF
4809 and break the signature.
4811 [This change does not apply to 0.9.7.]
4813 *) Zero the premaster secret after deriving the master secret in
4817 *) Add some EVP_add_digest_alias registrations (as found in
4818 OpenSSL_add_all_digests()) to SSL_library_init()
4819 aka OpenSSL_add_ssl_algorithms(). This provides improved
4820 compatibility with peers using X.509 certificates
4821 with unconventional AlgorithmIdentifier OIDs.
4824 *) Fix for Irix with NO_ASM.
4825 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4827 *) ./config script fixes.
4828 [Ulf Moeller, Richard Levitte]
4830 *) Fix 'openssl passwd -1'.
4833 *) Change PKCS12_key_gen_asc() so it can cope with non null
4834 terminated strings whose length is passed in the passlen
4835 parameter, for example from PEM callbacks. This was done
4836 by adding an extra length parameter to asc2uni().
4837 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4839 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4840 call failed, free the DSA structure.
4843 *) Fix to uni2asc() to cope with zero length Unicode strings.
4844 These are present in some PKCS#12 files.
4847 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4848 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4849 when writing a 32767 byte record.
4850 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4852 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4853 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4855 (RSA objects have a reference count access to which is protected
4856 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4857 so they are meant to be shared between threads.)
4858 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4859 "Reddie, Steven" <Steven.Reddie@ca.com>]
4861 *) Fix a deadlock in CRYPTO_mem_leaks().
4864 *) Use better test patterns in bntest.
4867 *) rand_win.c fix for Borland C.
4870 *) BN_rshift bugfix for n == 0.
4873 *) Add a 'bctest' script that checks for some known 'bc' bugs
4874 so that 'make test' does not abort just because 'bc' is broken.
4877 *) Store verify_result within SSL_SESSION also for client side to
4878 avoid potential security hole. (Re-used sessions on the client side
4879 always resulted in verify_result==X509_V_OK, not using the original
4880 result of the server certificate verification.)
4883 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4884 SSL3_RT_APPLICATION_DATA, return 0.
4885 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4889 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4890 releases, have been re-implemented by renaming the previous
4891 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4892 and ssl3_read_internal, respectively, and adding 'peek' parameters
4893 to them. The new ssl[23]_{read,peek} functions are calls to
4894 ssl[23]_read_internal with the 'peek' flag set appropriately.
4895 A 'peek' parameter has also been added to ssl3_read_bytes, which
4896 does the actual work for ssl3_read_internal.
4899 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4900 the method-specific "init()" handler. Also clean up ex_data after
4901 calling the method-specific "finish()" handler. Previously, this was
4902 happening the other way round.
4905 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4906 The previous value, 12, was not always sufficient for BN_mod_exp().
4909 *) Make sure that shared libraries get the internal name engine with
4910 the full version number and not just 0. This should mark the
4911 shared libraries as not backward compatible. Of course, this should
4912 be changed again when we can guarantee backward binary compatibility.
4915 *) Fix typo in get_cert_by_subject() in by_dir.c
4916 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4918 *) Rework the system to generate shared libraries:
4920 - Make note of the expected extension for the shared libraries and
4921 if there is a need for symbolic links from for example libcrypto.so.0
4922 to libcrypto.so.0.9.7. There is extended info in Configure for
4925 - Make as few rebuilds of the shared libraries as possible.
4927 - Still avoid linking the OpenSSL programs with the shared libraries.
4929 - When installing, install the shared libraries separately from the
4933 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4935 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4936 and not in SSL_clear because the latter is also used by the
4937 accept/connect functions; previously, the settings made by
4938 SSL_set_read_ahead would be lost during the handshake.
4939 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4941 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4942 Previously, it would create entries for disableed algorithms no
4946 *) Added several new manual pages for SSL_* function.
4949 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4951 *) In ssl23_get_client_hello, generate an error message when faced
4952 with an initial SSL 3.0/TLS record that is too small to contain the
4953 first two bytes of the ClientHello message, i.e. client_version.
4954 (Note that this is a pathologic case that probably has never happened
4955 in real life.) The previous approach was to use the version number
4956 from the record header as a substitute; but our protocol choice
4957 should not depend on that one because it is not authenticated
4958 by the Finished messages.
4961 *) More robust randomness gathering functions for Windows.
4962 [Jeffrey Altman <jaltman@columbia.edu>]
4964 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4965 not set then we don't setup the error code for issuer check errors
4966 to avoid possibly overwriting other errors which the callback does
4967 handle. If an application does set the flag then we assume it knows
4968 what it is doing and can handle the new informational codes
4972 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4973 a general "ANY" type, as such it should be able to decode anything
4974 including tagged types. However it didn't check the class so it would
4975 wrongly interpret tagged types in the same way as their universal
4976 counterpart and unknown types were just rejected. Changed so that the
4977 tagged and unknown types are handled in the same way as a SEQUENCE:
4978 that is the encoding is stored intact. There is also a new type
4979 "V_ASN1_OTHER" which is used when the class is not universal, in this
4980 case we have no idea what the actual type is so we just lump them all
4984 *) On VMS, stdout may very well lead to a file that is written to
4985 in a record-oriented fashion. That means that every write() will
4986 write a separate record, which will be read separately by the
4987 programs trying to read from it. This can be very confusing.
4989 The solution is to put a BIO filter in the way that will buffer
4990 text until a linefeed is reached, and then write everything a
4991 line at a time, so every record written will be an actual line,
4992 not chunks of lines and not (usually doesn't happen, but I've
4993 seen it once) several lines in one record. BIO_f_linebuffer() is
4996 Currently, it's a VMS-only method, because that's where it has
4997 been tested well enough.
5000 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5001 it can return incorrect results.
5002 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5003 but it was in 0.9.6-beta[12].)
5006 *) Disable the check for content being present when verifying detached
5007 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5008 include zero length content when signing messages.
5011 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5012 BIO_ctrl (for BIO pairs).
5015 *) Add DSO method for VMS.
5018 *) Bug fix: Montgomery multiplication could produce results with the
5022 *) Add RPM specification openssl.spec and modify it to build three
5023 packages. The default package contains applications, application
5024 documentation and run-time libraries. The devel package contains
5025 include files, static libraries and function documentation. The
5026 doc package contains the contents of the doc directory. The original
5027 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5030 *) Add a large number of documentation files for many SSL routines.
5031 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5033 *) Add a configuration entry for Sony News 4.
5034 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5036 *) Don't set the two most significant bits to one when generating a
5037 random number < q in the DSA library.
5040 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5041 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5042 the underlying transport is blocking) if a handshake took place.
5043 (The default behaviour is needed by applications such as s_client
5044 and s_server that use select() to determine when to use SSL_read;
5045 but for applications that know in advance when to expect data, it
5046 just makes things more complicated.)
5049 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5053 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5054 work better on such systems.
5055 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5057 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5058 Update PKCS12_parse() so it copies the friendlyName and the
5059 keyid to the certificates aux info.
5062 *) Fix bug in PKCS7_verify() which caused an infinite loop
5063 if there was more than one signature.
5064 [Sven Uszpelkat <su@celocom.de>]
5066 *) Major change in util/mkdef.pl to include extra information
5067 about each symbol, as well as presentig variables as well
5068 as functions. This change means that there's n more need
5069 to rebuild the .num files when some algorithms are excluded.
5072 *) Allow the verify time to be set by an application,
5073 rather than always using the current time.
5076 *) Phase 2 verify code reorganisation. The certificate
5077 verify code now looks up an issuer certificate by a
5078 number of criteria: subject name, authority key id
5079 and key usage. It also verifies self signed certificates
5080 by the same criteria. The main comparison function is
5081 X509_check_issued() which performs these checks.
5083 Lot of changes were necessary in order to support this
5084 without completely rewriting the lookup code.
5086 Authority and subject key identifier are now cached.
5088 The LHASH 'certs' is X509_STORE has now been replaced
5089 by a STACK_OF(X509_OBJECT). This is mainly because an
5090 LHASH can't store or retrieve multiple objects with
5091 the same hash value.
5093 As a result various functions (which were all internal
5094 use only) have changed to handle the new X509_STORE
5095 structure. This will break anything that messed round
5096 with X509_STORE internally.
5098 The functions X509_STORE_add_cert() now checks for an
5099 exact match, rather than just subject name.
5101 The X509_STORE API doesn't directly support the retrieval
5102 of multiple certificates matching a given criteria, however
5103 this can be worked round by performing a lookup first
5104 (which will fill the cache with candidate certificates)
5105 and then examining the cache for matches. This is probably
5106 the best we can do without throwing out X509_LOOKUP
5107 entirely (maybe later...).
5109 The X509_VERIFY_CTX structure has been enhanced considerably.
5111 All certificate lookup operations now go via a get_issuer()
5112 callback. Although this currently uses an X509_STORE it
5113 can be replaced by custom lookups. This is a simple way
5114 to bypass the X509_STORE hackery necessary to make this
5115 work and makes it possible to use more efficient techniques
5116 in future. A very simple version which uses a simple
5117 STACK for its trusted certificate store is also provided
5118 using X509_STORE_CTX_trusted_stack().
5120 The verify_cb() and verify() callbacks now have equivalents
5121 in the X509_STORE_CTX structure.
5123 X509_STORE_CTX also has a 'flags' field which can be used
5124 to customise the verify behaviour.
5127 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5128 excludes S/MIME capabilities.
5131 *) When a certificate request is read in keep a copy of the
5132 original encoding of the signed data and use it when outputing
5133 again. Signatures then use the original encoding rather than
5134 a decoded, encoded version which may cause problems if the
5135 request is improperly encoded.
5138 *) For consistency with other BIO_puts implementations, call
5139 buffer_write(b, ...) directly in buffer_puts instead of calling
5142 In BIO_puts, increment b->num_write as in BIO_write.
5143 [Peter.Sylvester@EdelWeb.fr]
5145 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5146 BN_zero, we may not return a BIGNUM with an array consisting of
5150 *) Avoid calling abort() from within the library when problems are
5151 detected, except if preprocessor symbols have been defined
5152 (such as REF_CHECK, BN_DEBUG etc.).
5155 *) New openssl application 'rsautl'. This utility can be
5156 used for low level RSA operations. DER public key
5157 BIO/fp routines also added.
5160 *) New Configure entry and patches for compiling on QNX 4.
5161 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5163 *) A demo state-machine implementation was sponsored by
5164 Nuron (http://www.nuron.com/) and is now available in
5165 demos/state_machine.
5168 *) New options added to the 'dgst' utility for signature
5169 generation and verification.
5172 *) Unrecognized PKCS#7 content types are now handled via a
5173 catch all ASN1_TYPE structure. This allows unsupported
5174 types to be stored as a "blob" and an application can
5175 encode and decode it manually.
5178 *) Fix various signed/unsigned issues to make a_strex.c
5180 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5182 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5183 length if passed a buffer. ASN1_INTEGER_to_BN failed
5184 if passed a NULL BN and its argument was negative.
5185 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5187 *) Modification to PKCS#7 encoding routines to output definite
5188 length encoding. Since currently the whole structures are in
5189 memory there's not real point in using indefinite length
5190 constructed encoding. However if OpenSSL is compiled with
5191 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5194 *) Added BIO_vprintf() and BIO_vsnprintf().
5197 *) Added more prefixes to parse for in the the strings written
5198 through a logging bio, to cover all the levels that are available
5199 through syslog. The prefixes are now:
5201 PANIC, EMERG, EMR => LOG_EMERG
5202 ALERT, ALR => LOG_ALERT
5203 CRIT, CRI => LOG_CRIT
5204 ERROR, ERR => LOG_ERR
5205 WARNING, WARN, WAR => LOG_WARNING
5206 NOTICE, NOTE, NOT => LOG_NOTICE
5207 INFO, INF => LOG_INFO
5208 DEBUG, DBG => LOG_DEBUG
5210 and as before, if none of those prefixes are present at the
5211 beginning of the string, LOG_ERR is chosen.
5213 On Win32, the LOG_* levels are mapped according to this:
5215 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5216 LOG_WARNING => EVENTLOG_WARNING_TYPE
5217 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5221 *) Made it possible to reconfigure with just the configuration
5222 argument "reconf" or "reconfigure". The command line arguments
5223 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5224 and are retrieved from there when reconfiguring.
5228 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5230 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5233 *) The obj_dat.pl script was messing up the sorting of object
5234 names. The reason was that it compared the quoted version
5235 of strings as a result "OCSP" > "OCSP Signing" because
5236 " > SPACE. Changed script to store unquoted versions of
5237 names and add quotes on output. It was also omitting some
5238 names from the lookup table if they were given a default
5239 value (that is if SN is missing it is given the same
5240 value as LN and vice versa), these are now added on the
5241 grounds that if an object has a name we should be able to
5242 look it up. Finally added warning output when duplicate
5243 short or long names are found.
5246 *) Changes needed for Tandem NSK.
5247 [Scott Uroff <scott@xypro.com>]
5249 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5250 RSA_padding_check_SSLv23(), special padding was never detected
5251 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5252 version rollback attacks was not effective.
5254 In s23_clnt.c, don't use special rollback-attack detection padding
5255 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5256 client; similarly, in s23_srvr.c, don't do the rollback check if
5257 SSL 2.0 is the only protocol enabled in the server.
5260 *) Make it possible to get hexdumps of unprintable data with 'openssl
5261 asn1parse'. By implication, the functions ASN1_parse_dump() and
5262 BIO_dump_indent() are added.
5265 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5266 these print out strings and name structures based on various
5267 flags including RFC2253 support and proper handling of
5268 multibyte characters. Added options to the 'x509' utility
5269 to allow the various flags to be set.
5272 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5273 Also change the functions X509_cmp_current_time() and
5274 X509_gmtime_adj() work with an ASN1_TIME structure,
5275 this will enable certificates using GeneralizedTime in validity
5276 dates to be checked.
5279 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5280 negative public key encodings) on by default,
5281 NO_NEG_PUBKEY_BUG can be set to disable it.
5284 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5285 content octets. An i2c_ASN1_OBJECT is unnecessary because
5286 the encoding can be trivially obtained from the structure.
5289 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5290 not read locks (CRYPTO_r_[un]lock).
5293 *) A first attempt at creating official support for shared
5294 libraries through configuration. I've kept it so the
5295 default is static libraries only, and the OpenSSL programs
5296 are always statically linked for now, but there are
5297 preparations for dynamic linking in place.
5298 This has been tested on Linux and Tru64.
5301 *) Randomness polling function for Win9x, as described in:
5302 Peter Gutmann, Software Generation of Practically Strong
5306 *) Fix so PRNG is seeded in req if using an already existing
5310 *) New options to smime application. -inform and -outform
5311 allow alternative formats for the S/MIME message including
5312 PEM and DER. The -content option allows the content to be
5313 specified separately. This should allow things like Netscape
5314 form signing output easier to verify.
5317 *) Fix the ASN1 encoding of tags using the 'long form'.
5320 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5321 STRING types. These convert content octets to and from the
5322 underlying type. The actual tag and length octets are
5323 already assumed to have been read in and checked. These
5324 are needed because all other string types have virtually
5325 identical handling apart from the tag. By having versions
5326 of the ASN1 functions that just operate on content octets
5327 IMPLICIT tagging can be handled properly. It also allows
5328 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5329 and ASN1_INTEGER are identical apart from the tag.
5332 *) Change the handling of OID objects as follows:
5334 - New object identifiers are inserted in objects.txt, following
5335 the syntax given in objects.README.
5336 - objects.pl is used to process obj_mac.num and create a new
5338 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5341 This is currently kind of a hack, and the perl code in objects.pl
5342 isn't very elegant, but it works as I intended. The simplest way
5343 to check that it worked correctly is to look in obj_dat.h and
5344 check the array nid_objs and make sure the objects haven't moved
5345 around (this is important!). Additions are OK, as well as
5346 consistent name changes.
5349 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5352 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5353 The given file adds to whatever has already been seeded into the
5354 random pool through the RANDFILE configuration file option or
5355 environment variable, or the default random state file.
5358 *) mkstack.pl now sorts each macro group into lexical order.
5359 Previously the output order depended on the order the files
5360 appeared in the directory, resulting in needless rewriting
5364 *) Patches to make OpenSSL compile under Win32 again. Mostly
5365 work arounds for the VC++ problem that it treats func() as
5366 func(void). Also stripped out the parts of mkdef.pl that
5367 added extra typesafe functions: these no longer exist.
5370 *) Reorganisation of the stack code. The macros are now all
5371 collected in safestack.h . Each macro is defined in terms of
5372 a "stack macro" of the form SKM_<name>(type, a, b). The
5373 DEBUG_SAFESTACK is now handled in terms of function casts,
5374 this has the advantage of retaining type safety without the
5375 use of additional functions. If DEBUG_SAFESTACK is not defined
5376 then the non typesafe macros are used instead. Also modified the
5377 mkstack.pl script to handle the new form. Needs testing to see
5378 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5379 the default if no major problems. Similar behaviour for ASN1_SET_OF
5380 and PKCS12_STACK_OF.
5383 *) When some versions of IIS use the 'NET' form of private key the
5384 key derivation algorithm is different. Normally MD5(password) is
5385 used as a 128 bit RC4 key. In the modified case
5386 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5387 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5388 as the old Netscape_RSA functions except they have an additional
5389 'sgckey' parameter which uses the modified algorithm. Also added
5390 an -sgckey command line option to the rsa utility. Thanks to
5391 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5392 algorithm to openssl-dev.
5395 *) The evp_local.h macros were using 'c.##kname' which resulted in
5396 invalid expansion on some systems (SCO 5.0.5 for example).
5397 Corrected to 'c.kname'.
5398 [Phillip Porch <root@theporch.com>]
5400 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5401 a STACK of email addresses from a certificate or request, these look
5402 in the subject name and the subject alternative name extensions and
5403 omit any duplicate addresses.
5406 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5407 This makes DSA verification about 2 % faster.
5410 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5411 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5412 plus overhead for 1024 bit moduli).
5413 This makes exponentiations about 0.5 % faster for 1024 bit
5414 exponents (as measured by "openssl speed rsa2048").
5417 *) Rename memory handling macros to avoid conflicts with other
5419 Malloc => OPENSSL_malloc
5420 Malloc_locked => OPENSSL_malloc_locked
5421 Realloc => OPENSSL_realloc
5422 Free => OPENSSL_free
5425 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5426 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5429 *) CygWin32 support.
5430 [John Jarvie <jjarvie@newsguy.com>]
5432 *) The type-safe stack code has been rejigged. It is now only compiled
5433 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5434 by default all type-specific stack functions are "#define"d back to
5435 standard stack functions. This results in more streamlined output
5436 but retains the type-safety checking possibilities of the original
5440 *) The STACK code has been cleaned up, and certain type declarations
5441 that didn't make a lot of sense have been brought in line. This has
5442 also involved a cleanup of sorts in safestack.h to more correctly
5443 map type-safe stack functions onto their plain stack counterparts.
5444 This work has also resulted in a variety of "const"ifications of
5445 lots of the code, especially "_cmp" operations which should normally
5446 be prototyped with "const" parameters anyway.
5449 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5450 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5451 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5452 where all of 'md' is used each time the PRNG is used, but 'state'
5453 is used only indexed by a cyclic counter. As entropy may not be
5454 well distributed from the beginning, 'md' is important as a
5455 chaining variable. However, the output function chains only half
5456 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5457 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5458 in all of 'state' being rewritten, with the new values depending
5459 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5462 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5463 the handshake is continued after ssl_verify_cert_chain();
5464 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5465 can lead to 'unexplainable' connection aborts later.
5466 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5468 *) Major EVP API cipher revision.
5469 Add hooks for extra EVP features. This allows various cipher
5470 parameters to be set in the EVP interface. Support added for variable
5471 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5472 setting of RC2 and RC5 parameters.
5474 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5477 Remove lots of duplicated code from the EVP library. For example *every*
5478 cipher init() function handles the 'iv' in the same way according to the
5479 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5480 for CFB and OFB modes they zero ctx->num.
5482 New functionality allows removal of S/MIME code RC2 hack.
5484 Most of the routines have the same form and so can be declared in terms
5487 By shifting this to the top level EVP_CipherInit() it can be removed from
5488 all individual ciphers. If the cipher wants to handle IVs or keys
5489 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5492 Change lots of functions like EVP_EncryptUpdate() to now return a
5493 value: although software versions of the algorithms cannot fail
5494 any installed hardware versions can.
5497 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5498 this option is set, tolerate broken clients that send the negotiated
5499 protocol version number instead of the requested protocol version
5503 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5504 i.e. non-zero for export ciphersuites, zero otherwise.
5505 Previous versions had this flag inverted, inconsistent with
5506 rsa_tmp_cb (..._TMP_RSA_CB).
5507 [Bodo Moeller; problem reported by Amit Chopra]
5509 *) Add missing DSA library text string. Work around for some IIS
5510 key files with invalid SEQUENCE encoding.
5513 *) Add a document (doc/standards.txt) that list all kinds of standards
5514 and so on that are implemented in OpenSSL.
5517 *) Enhance c_rehash script. Old version would mishandle certificates
5518 with the same subject name hash and wouldn't handle CRLs at all.
5519 Added -fingerprint option to crl utility, to support new c_rehash
5523 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5526 *) Fix for SSL server purpose checking. Server checking was
5527 rejecting certificates which had extended key usage present
5528 but no ssl client purpose.
5529 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5531 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5532 is a little unclear about how a blank password is handled.
5533 Since the password in encoded as a BMPString with terminating
5534 double NULL a zero length password would end up as just the
5535 double NULL. However no password at all is different and is
5536 handled differently in the PKCS#12 key generation code. NS
5537 treats a blank password as zero length. MSIE treats it as no
5538 password on export: but it will try both on import. We now do
5539 the same: PKCS12_parse() tries zero length and no password if
5540 the password is set to "" or NULL (NULL is now a valid password:
5541 it wasn't before) as does the pkcs12 application.
5544 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5545 perror when PEM_read_bio_X509_REQ fails, the error message must
5546 be obtained from the error queue.
5549 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5550 it in ERR_remove_state if appropriate, and change ERR_get_state
5551 accordingly to avoid race conditions (this is necessary because
5552 thread_hash is no longer constant once set).
5555 *) Bugfix for linux-elf makefile.one.
5558 *) RSA_get_default_method() will now cause a default
5559 RSA_METHOD to be chosen if one doesn't exist already.
5560 Previously this was only set during a call to RSA_new()
5561 or RSA_new_method(NULL) meaning it was possible for
5562 RSA_get_default_method() to return NULL.
5565 *) Added native name translation to the existing DSO code
5566 that will convert (if the flag to do so is set) filenames
5567 that are sufficiently small and have no path information
5568 into a canonical native form. Eg. "blah" converted to
5569 "libblah.so" or "blah.dll" etc.
5572 *) New function ERR_error_string_n(e, buf, len) which is like
5573 ERR_error_string(e, buf), but writes at most 'len' bytes
5574 including the 0 terminator. For ERR_error_string_n, 'buf'
5576 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5578 *) CONF library reworked to become more general. A new CONF
5579 configuration file reader "class" is implemented as well as a
5580 new functions (NCONF_*, for "New CONF") to handle it. The now
5581 old CONF_* functions are still there, but are reimplemented to
5582 work in terms of the new functions. Also, a set of functions
5583 to handle the internal storage of the configuration data is
5584 provided to make it easier to write new configuration file
5585 reader "classes" (I can definitely see something reading a
5586 configuration file in XML format, for example), called _CONF_*,
5587 or "the configuration storage API"...
5589 The new configuration file reading functions are:
5591 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5592 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5594 NCONF_default, NCONF_WIN32
5596 NCONF_dump_fp, NCONF_dump_bio
5598 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5599 NCONF_new creates a new CONF object. This works in the same way
5600 as other interfaces in OpenSSL, like the BIO interface.
5601 NCONF_dump_* dump the internal storage of the configuration file,
5602 which is useful for debugging. All other functions take the same
5603 arguments as the old CONF_* functions wth the exception of the
5604 first that must be a `CONF *' instead of a `LHASH *'.
5606 To make it easer to use the new classes with the old CONF_* functions,
5607 the function CONF_set_default_method is provided.
5610 *) Add '-tls1' option to 'openssl ciphers', which was already
5611 mentioned in the documentation but had not been implemented.
5612 (This option is not yet really useful because even the additional
5613 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5616 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5617 OpenSSL-based applications) load shared libraries and bind to
5618 them in a portable way.
5619 [Geoff Thorpe, with contributions from Richard Levitte]
5621 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5623 *) Make sure _lrotl and _lrotr are only used with MSVC.
5625 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5626 (the default implementation of RAND_status).
5628 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5629 to '-clrext' (= clear extensions), as intended and documented.
5630 [Bodo Moeller; inconsistency pointed out by Michael Attili
5633 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5634 was larger than the MD block size.
5635 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5637 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5638 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5639 using the passed key: if the passed key was a private key the result
5640 of X509_print(), for example, would be to print out all the private key
5644 *) des_quad_cksum() byte order bug fix.
5645 [Ulf Möller, using the problem description in krb4-0.9.7, where
5646 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5648 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5650 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5652 *) For easily testing in shell scripts whether some command
5653 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5654 returns with exit code 0 iff no command of the given name is available.
5655 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5656 the output goes to stdout and nothing is printed to stderr.
5657 Additional arguments are always ignored.
5659 Since for each cipher there is a command of the same name,
5660 the 'no-cipher' compilation switches can be tested this way.
5662 ('openssl no-XXX' is not able to detect pseudo-commands such
5663 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5666 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5669 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5670 is set; it will be thrown away anyway because each handshake creates
5672 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5673 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5674 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5675 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5678 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5679 'Q' and 'R' lose their special meanings (quit/renegotiate).
5680 This is part of what -quiet does; unlike -quiet, -ign_eof
5681 does not suppress any output.
5684 *) Add compatibility options to the purpose and trust code. The
5685 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5686 accepts a certificate or CA, this was the previous behaviour,
5687 with all the associated security issues.
5689 X509_TRUST_COMPAT is the old trust behaviour: only and
5690 automatically trust self signed roots in certificate store. A
5691 new trust setting X509_TRUST_DEFAULT is used to specify that
5692 a purpose has no associated trust setting and it should instead
5693 use the value in the default purpose.
5696 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5697 and fix a memory leak.
5700 *) In util/mkerr.pl (which implements 'make errors'), preserve
5701 reason strings from the previous version of the .c file, as
5702 the default to have only downcase letters (and digits) in
5703 automatically generated reasons codes is not always appropriate.
5706 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5707 using strerror. Previously, ERR_reason_error_string() returned
5708 library names as reason strings for SYSerr; but SYSerr is a special
5709 case where small numbers are errno values, not library numbers.
5712 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5713 converts DSA parameters into DH parameters. (When creating parameters,
5714 DSA_generate_parameters is used.)
5717 *) Include 'length' (recommended exponent length) in C code generated
5718 by 'openssl dhparam -C'.
5721 *) The second argument to set_label in perlasm was already being used
5722 so couldn't be used as a "file scope" flag. Moved to third argument
5726 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5727 instead of RAND_bytes for encryption IVs and salts.
5730 *) Include RAND_status() into RAND_METHOD instead of implementing
5731 it only for md_rand.c Otherwise replacing the PRNG by calling
5732 RAND_set_rand_method would be impossible.
5735 *) Don't let DSA_generate_key() enter an infinite loop if the random
5736 number generation fails.
5739 *) New 'rand' application for creating pseudo-random output.
5742 *) Added configuration support for Linux/IA64
5743 [Rolf Haberrecker <rolf@suse.de>]
5745 *) Assembler module support for Mingw32.
5748 *) Shared library support for HPUX (in shlib/).
5749 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5751 *) Shared library support for Solaris gcc.
5752 [Lutz Behnke <behnke@trustcenter.de>]
5754 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5756 *) PKCS7_encrypt() was adding text MIME headers twice because they
5757 were added manually and by SMIME_crlf_copy().
5760 *) In bntest.c don't call BN_rand with zero bits argument.
5761 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5763 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5764 case was implemented. This caused BN_div_recp() to fail occasionally.
5767 *) Add an optional second argument to the set_label() in the perl
5768 assembly language builder. If this argument exists and is set
5769 to 1 it signals that the assembler should use a symbol whose
5770 scope is the entire file, not just the current function. This
5771 is needed with MASM which uses the format label:: for this scope.
5772 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5774 *) Change the ASN1 types so they are typedefs by default. Before
5775 almost all types were #define'd to ASN1_STRING which was causing
5776 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5780 *) Change names of new functions to the new get1/get0 naming
5781 convention: After 'get1', the caller owns a reference count
5782 and has to call ..._free; 'get0' returns a pointer to some
5783 data structure without incrementing reference counters.
5784 (Some of the existing 'get' functions increment a reference
5785 counter, some don't.)
5786 Similarly, 'set1' and 'add1' functions increase reference
5787 counters or duplicate objects.
5790 *) Allow for the possibility of temp RSA key generation failure:
5791 the code used to assume it always worked and crashed on failure.
5794 *) Fix potential buffer overrun problem in BIO_printf().
5795 [Ulf Möller, using public domain code by Patrick Powell; problem
5796 pointed out by David Sacerdote <das33@cornell.edu>]
5798 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5799 RAND_egd() and RAND_status(). In the command line application,
5800 the EGD socket can be specified like a seed file using RANDFILE
5804 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5805 Some CAs (e.g. Verisign) distribute certificates in this form.
5808 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5809 list to exclude them. This means that no special compilation option
5810 is needed to use anonymous DH: it just needs to be included in the
5814 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5815 EVP_MD_type. The old functionality is available in a new macro called
5816 EVP_MD_md(). Change code that uses it and update docs.
5819 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5820 where the 'void *' argument is replaced by a function pointer argument.
5821 Previously 'void *' was abused to point to functions, which works on
5822 many platforms, but is not correct. As these functions are usually
5823 called by macros defined in OpenSSL header files, most source code
5824 should work without changes.
5827 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5828 sections with information on -D... compiler switches used for
5829 compiling the library so that applications can see them. To enable
5830 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5831 must be defined. E.g.,
5832 #define OPENSSL_ALGORITHM_DEFINES
5833 #include <openssl/opensslconf.h>
5834 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5835 [Richard Levitte, Ulf and Bodo Möller]
5837 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5841 *) Change the 'other' type in certificate aux info to a STACK_OF
5842 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5843 the required ASN1 format: arbitrary types determined by an OID.
5846 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5847 argument to 'req'. This is not because the function is newer or
5848 better than others it just uses the work 'NEW' in the certificate
5849 request header lines. Some software needs this.
5852 *) Reorganise password command line arguments: now passwords can be
5853 obtained from various sources. Delete the PEM_cb function and make
5854 it the default behaviour: i.e. if the callback is NULL and the
5855 usrdata argument is not NULL interpret it as a null terminated pass
5856 phrase. If usrdata and the callback are NULL then the pass phrase
5857 is prompted for as usual.
5860 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5861 the support is automatically enabled. The resulting binaries will
5862 autodetect the card and use it if present.
5863 [Ben Laurie and Compaq Inc.]
5865 *) Work around for Netscape hang bug. This sends certificate request
5866 and server done in one record. Since this is perfectly legal in the
5867 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5868 the bugs/SSLv3 entry for more info.
5871 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5874 *) Add -rand argument to smime and pkcs12 applications and read/write
5878 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5881 *) Add command line password options to the remaining applications.
5884 *) Bug fix for BN_div_recp() for numerators with an even number of
5888 *) More tests in bntest.c, and changed test_bn output.
5891 *) ./config recognizes MacOS X now.
5894 *) Bug fix for BN_div() when the first words of num and divsor are
5895 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5898 *) Add support for various broken PKCS#8 formats, and command line
5899 options to produce them.
5902 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5903 get temporary BIGNUMs from a BN_CTX.
5906 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5910 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5911 include a #define from the old name to the new. The original intent
5912 was that statically linked binaries could for example just call
5913 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5914 link with digests. This never worked becayse SSLeay_add_all_digests()
5915 and SSLeay_add_all_ciphers() were in the same source file so calling
5916 one would link with the other. They are now in separate source files.
5919 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5922 *) Use a less unusual form of the Miller-Rabin primality test (it used
5923 a binary algorithm for exponentiation integrated into the Miller-Rabin
5924 loop, our standard modexp algorithms are faster).
5927 *) Support for the EBCDIC character set completed.
5928 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5930 *) Source code cleanups: use const where appropriate, eliminate casts,
5931 use void * instead of char * in lhash.
5934 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5935 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5936 this the server could overwrite ephemeral keys that the client
5940 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5941 using 50 iterations of the Rabin-Miller test.
5943 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5944 iterations of the Rabin-Miller test as required by the appendix
5945 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5946 As BN_is_prime_fasttest includes trial division, DSA parameter
5947 generation becomes much faster.
5949 This implies a change for the callback functions in DSA_is_prime
5950 and DSA_generate_parameters: The callback function is called once
5951 for each positive witness in the Rabin-Miller test, not just
5952 occasionally in the inner loop; and the parameters to the
5953 callback function now provide an iteration count for the outer
5954 loop rather than for the current invocation of the inner loop.
5955 DSA_generate_parameters additionally can call the callback
5956 function with an 'iteration count' of -1, meaning that a
5957 candidate has passed the trial division test (when q is generated
5958 from an application-provided seed, trial division is skipped).
5961 *) New function BN_is_prime_fasttest that optionally does trial
5962 division before starting the Rabin-Miller test and has
5963 an additional BN_CTX * argument (whereas BN_is_prime always
5964 has to allocate at least one BN_CTX).
5965 'callback(1, -1, cb_arg)' is called when a number has passed the
5966 trial division stage.
5969 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5973 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5976 *) New function BN_pseudo_rand().
5979 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5980 bignum version of BN_from_montgomery() with the working code from
5981 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5985 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5986 made it impossible to use the same SSL_SESSION data structure in
5987 SSL2 clients in multiple threads.
5990 *) The return value of RAND_load_file() no longer counts bytes obtained
5991 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5992 to seed the PRNG (previously an explicit byte count was required).
5993 [Ulf Möller, Bodo Möller]
5995 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5996 used (char *) instead of (void *) and had casts all over the place.
5999 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6002 *) Retain source code compatibility for BN_prime_checks macro:
6003 BN_is_prime(..., BN_prime_checks, ...) now uses
6004 BN_prime_checks_for_size to determine the appropriate number of
6005 Rabin-Miller iterations.
6008 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6009 DH_CHECK_P_NOT_SAFE_PRIME.
6010 (Check if this is true? OpenPGP calls them "strong".)
6013 *) Merge the functionality of "dh" and "gendh" programs into a new program
6014 "dhparam". The old programs are retained for now but will handle DH keys
6015 (instead of parameters) in future.
6018 *) Make the ciphers, s_server and s_client programs check the return values
6019 when a new cipher list is set.
6022 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6023 ciphers. Before when the 56bit ciphers were enabled the sorting was
6026 The syntax for the cipher sorting has been extended to support sorting by
6027 cipher-strength (using the strength_bits hard coded in the tables).
6028 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6030 Fix a bug in the cipher-command parser: when supplying a cipher command
6031 string with an "undefined" symbol (neither command nor alphanumeric
6032 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6033 an error is flagged.
6035 Due to the strength-sorting extension, the code of the
6036 ssl_create_cipher_list() function was completely rearranged. I hope that
6037 the readability was also increased :-)
6038 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6040 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6041 for the first serial number and places 2 in the serial number file. This
6042 avoids problems when the root CA is created with serial number zero and
6043 the first user certificate has the same issuer name and serial number
6047 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6048 the new code. Add documentation for this stuff.
6051 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6052 X509_*() to X509at_*() on the grounds that they don't handle X509
6053 structures and behave in an analagous way to the X509v3 functions:
6054 they shouldn't be called directly but wrapper functions should be used
6057 So we also now have some wrapper functions that call the X509at functions
6058 when passed certificate requests. (TO DO: similar things can be done with
6059 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6060 things. Some of these need some d2i or i2d and print functionality
6061 because they handle more complex structures.)
6064 *) Add missing #ifndefs that caused missing symbols when building libssl
6065 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6066 NO_RSA in ssl/s2*.c.
6067 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6069 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6070 has a return value which indicates the quality of the random data
6071 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6072 error queue. New function RAND_pseudo_bytes() generates output that is
6073 guaranteed to be unique but not unpredictable. RAND_add is like
6074 RAND_seed, but takes an extra argument for an entropy estimate
6075 (RAND_seed always assumes full entropy).
6078 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6079 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6080 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6081 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6082 false-positive rate of at most 2^-80 for random input.
6085 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6088 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6089 in the 0.9.5 release), this returns the chain
6090 from an X509_CTX structure with a dup of the stack and all
6091 the X509 reference counts upped: so the stack will exist
6092 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6095 Also make SSL_SESSION_print() print out the verify return
6099 *) Add manpage for the pkcs12 command. Also change the default
6100 behaviour so MAC iteration counts are used unless the new
6101 -nomaciter option is used. This improves file security and
6102 only older versions of MSIE (4.0 for example) need it.
6105 *) Honor the no-xxx Configure options when creating .DEF files.
6108 *) Add PKCS#10 attributes to field table: challengePassword,
6109 unstructuredName and unstructuredAddress. These are taken from
6110 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6111 international characters are used.
6113 More changes to X509_ATTRIBUTE code: allow the setting of types
6114 based on strings. Remove the 'loc' parameter when adding
6115 attributes because these will be a SET OF encoding which is sorted
6119 *) Initial changes to the 'req' utility to allow request generation
6120 automation. This will allow an application to just generate a template
6121 file containing all the field values and have req construct the
6124 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6125 used all over the place including certificate requests and PKCS#7
6126 structures. They are currently handled manually where necessary with
6127 some primitive wrappers for PKCS#7. The new functions behave in a
6128 manner analogous to the X509 extension functions: they allow
6129 attributes to be looked up by NID and added.
6131 Later something similar to the X509V3 code would be desirable to
6132 automatically handle the encoding, decoding and printing of the
6133 more complex types. The string types like challengePassword can
6134 be handled by the string table functions.
6136 Also modified the multi byte string table handling. Now there is
6137 a 'global mask' which masks out certain types. The table itself
6138 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6139 is useful when for example there is only one permissible type
6140 (as in countryName) and using the mask might result in no valid
6144 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6145 SSL_get_peer_finished to allow applications to obtain the latest
6146 Finished messages sent to the peer or expected from the peer,
6147 respectively. (SSL_get_peer_finished is usually the Finished message
6148 actually received from the peer, otherwise the protocol will be aborted.)
6150 As the Finished message are message digests of the complete handshake
6151 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6152 be used for external authentication procedures when the authentication
6153 provided by SSL/TLS is not desired or is not enough.
6156 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6157 the host supports BWX extension and if Compaq C is present on the
6158 $PATH. Just exploiting of the BWX extension results in 20-30%
6159 performance kick for some algorithms, e.g. DES and RC4 to mention
6160 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6164 *) Add support for MS "fast SGC". This is arguably a violation of the
6165 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6166 weak crypto and after checking the certificate is SGC a second one
6167 with strong crypto. MS SGC stops the first handshake after receiving
6168 the server certificate message and sends a second client hello. Since
6169 a server will typically do all the time consuming operations before
6170 expecting any further messages from the client (server key exchange
6171 is the most expensive) there is little difference between the two.
6173 To get OpenSSL to support MS SGC we have to permit a second client
6174 hello message after we have sent server done. In addition we have to
6175 reset the MAC if we do get this second client hello.
6178 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6179 if a DER encoded private key is RSA or DSA traditional format. Changed
6180 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6181 format DER encoded private key. Newer code should use PKCS#8 format which
6182 has the key type encoded in the ASN1 structure. Added DER private key
6183 support to pkcs8 application.
6186 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6187 ciphersuites has been selected (as required by the SSL 3/TLS 1
6188 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6189 is set, we interpret this as a request to violate the specification
6190 (the worst that can happen is a handshake failure, and 'correct'
6191 behaviour would result in a handshake failure anyway).
6194 *) In SSL_CTX_add_session, take into account that there might be multiple
6195 SSL_SESSION structures with the same session ID (e.g. when two threads
6196 concurrently obtain them from an external cache).
6197 The internal cache can handle only one SSL_SESSION with a given ID,
6198 so if there's a conflict, we now throw out the old one to achieve
6202 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6203 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6204 some routines that use cipher OIDs: some ciphers do not have OIDs
6205 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6209 *) Simplify the trust setting structure and code. Now we just have
6210 two sequences of OIDs for trusted and rejected settings. These will
6211 typically have values the same as the extended key usage extension
6212 and any application specific purposes.
6214 The trust checking code now has a default behaviour: it will just
6215 check for an object with the same NID as the passed id. Functions can
6216 be provided to override either the default behaviour or the behaviour
6217 for a given id. SSL client, server and email already have functions
6218 in place for compatibility: they check the NID and also return "trusted"
6219 if the certificate is self signed.
6222 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6223 traditional format into an EVP_PKEY structure.
6226 *) Add a password callback function PEM_cb() which either prompts for
6227 a password if usr_data is NULL or otherwise assumes it is a null
6228 terminated password. Allow passwords to be passed on command line
6229 environment or config files in a few more utilities.
6232 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6233 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6234 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6235 Update documentation.
6238 *) Support for ASN1 "NULL" type. This could be handled before by using
6239 ASN1_TYPE but there wasn't any function that would try to read a NULL
6240 and produce an error if it couldn't. For compatibility we also have
6241 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6242 don't allocate anything because they don't need to.
6245 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6247 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6249 *) Rebuild of the memory allocation routines used by OpenSSL code and
6250 possibly others as well. The purpose is to make an interface that
6251 provide hooks so anyone can build a separate set of allocation and
6252 deallocation routines to be used by OpenSSL, for example memory
6253 pool implementations, or something else, which was previously hard
6254 since Malloc(), Realloc() and Free() were defined as macros having
6255 the values malloc, realloc and free, respectively (except for Win32
6256 compilations). The same is provided for memory debugging code.
6257 OpenSSL already comes with functionality to find memory leaks, but
6258 this gives people a chance to debug other memory problems.
6260 With these changes, a new set of functions and macros have appeared:
6262 CRYPTO_set_mem_debug_functions() [F]
6263 CRYPTO_get_mem_debug_functions() [F]
6264 CRYPTO_dbg_set_options() [F]
6265 CRYPTO_dbg_get_options() [F]
6266 CRYPTO_malloc_debug_init() [M]
6268 The memory debug functions are NULL by default, unless the library
6269 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6270 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6271 gives the standard debugging functions that come with OpenSSL) or
6272 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6273 provided by the library user) must be used. When the standard
6274 debugging functions are used, CRYPTO_dbg_set_options can be used to
6275 request additional information:
6276 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6277 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6279 Also, things like CRYPTO_set_mem_functions will always give the
6280 expected result (the new set of functions is used for allocation
6281 and deallocation) at all times, regardless of platform and compiler
6284 To finish it up, some functions that were never use in any other
6285 way than through macros have a new API and new semantic:
6288 CRYPTO_dbg_realloc()
6291 All macros of value have retained their old syntax.
6292 [Richard Levitte and Bodo Moeller]
6294 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6295 ordering of SMIMECapabilities wasn't in "strength order" and there
6296 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6300 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6301 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6302 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6304 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6305 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6306 functionality to handle multipart/signed properly) and a utility
6307 called 'smime' to call all this stuff. This is based on code I
6308 originally wrote for Celo who have kindly allowed it to be
6309 included in OpenSSL.
6312 *) Add variants des_set_key_checked and des_set_key_unchecked of
6313 des_set_key (aka des_key_sched). Global variable des_check_key
6314 decides which of these is called by des_set_key; this way
6315 des_check_key behaves as it always did, but applications and
6316 the library itself, which was buggy for des_check_key == 1,
6317 have a cleaner way to pick the version they need.
6320 *) New function PKCS12_newpass() which changes the password of a
6324 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6325 dynamic mix. In both cases the ids can be used as an index into the
6326 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6327 functions so they accept a list of the field values and the
6328 application doesn't need to directly manipulate the X509_TRUST
6332 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6336 *) Modify the way the V3 extension code looks up extensions. This now
6337 works in a similar way to the object code: we have some "standard"
6338 extensions in a static table which is searched with OBJ_bsearch()
6339 and the application can add dynamic ones if needed. The file
6340 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6341 updated whenever a new extension is added to the core code and kept
6342 in ext_nid order. There is a simple program 'tabtest.c' which checks
6343 this. New extensions are not added too often so this file can readily
6344 be maintained manually.
6346 There are two big advantages in doing things this way. The extensions
6347 can be looked up immediately and no longer need to be "added" using
6348 X509V3_add_standard_extensions(): this function now does nothing.
6349 [Side note: I get *lots* of email saying the extension code doesn't
6350 work because people forget to call this function]
6351 Also no dynamic allocation is done unless new extensions are added:
6352 so if we don't add custom extensions there is no need to call
6353 X509V3_EXT_cleanup().
6356 *) Modify enc utility's salting as follows: make salting the default. Add a
6357 magic header, so unsalted files fail gracefully instead of just decrypting
6358 to garbage. This is because not salting is a big security hole, so people
6359 should be discouraged from doing it.
6362 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6363 digest to be passed on the command line but it only used this
6364 parameter when signing a certificate. Modified so all relevant
6365 operations are affected by the digest parameter including the
6366 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6367 DSA key was used because it didn't fix the digest.
6370 *) Initial certificate chain verify code. Currently tests the untrusted
6371 certificates for consistency with the verify purpose (which is set
6372 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6374 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6375 this is because it will reject chains with invalid extensions whereas
6376 every previous version of OpenSSL and SSLeay made no checks at all.
6378 Trust code: checks the root CA for the relevant trust settings. Trust
6379 settings have an initial value consistent with the verify purpose: e.g.
6380 if the verify purpose is for SSL client use it expects the CA to be
6381 trusted for SSL client use. However the default value can be changed to
6382 permit custom trust settings: one example of this would be to only trust
6383 certificates from a specific "secure" set of CAs.
6385 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6386 which should be used for version portability: especially since the
6387 verify structure is likely to change more often now.
6389 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6390 to set them. If not set then assume SSL clients will verify SSL servers
6393 Two new options to the verify program: -untrusted allows a set of
6394 untrusted certificates to be passed in and -purpose which sets the
6395 intended purpose of the certificate. If a purpose is set then the
6396 new chain verify code is used to check extension consistency.
6399 *) Support for the authority information access extension.
6402 *) Modify RSA and DSA PEM read routines to transparently handle
6403 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6404 public keys in a format compatible with certificate
6405 SubjectPublicKeyInfo structures. Unfortunately there were already
6406 functions called *_PublicKey_* which used various odd formats so
6407 these are retained for compatibility: however the DSA variants were
6408 never in a public release so they have been deleted. Changed dsa/rsa
6409 utilities to handle the new format: note no releases ever handled public
6410 keys so we should be OK.
6412 The primary motivation for this change is to avoid the same fiasco
6413 that dogs private keys: there are several incompatible private key
6414 formats some of which are standard and some OpenSSL specific and
6415 require various evil hacks to allow partial transparent handling and
6416 even then it doesn't work with DER formats. Given the option anything
6417 other than PKCS#8 should be dumped: but the other formats have to
6418 stay in the name of compatibility.
6420 With public keys and the benefit of hindsight one standard format
6421 is used which works with EVP_PKEY, RSA or DSA structures: though
6422 it clearly returns an error if you try to read the wrong kind of key.
6424 Added a -pubkey option to the 'x509' utility to output the public key.
6425 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6426 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6427 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6428 that do the same as the EVP_PKEY_assign_*() except they up the
6429 reference count of the added key (they don't "swallow" the
6433 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6434 CRLs would fail if the file contained no certificates or no CRLs:
6435 added a new function to read in both types and return the number
6436 read: this means that if none are read it will be an error. The
6437 DER versions of the certificate and CRL reader would always fail
6438 because it isn't possible to mix certificates and CRLs in DER format
6439 without choking one or the other routine. Changed this to just read
6440 a certificate: this is the best we can do. Also modified the code
6441 in apps/verify.c to take notice of return codes: it was previously
6442 attempting to read in certificates from NULL pointers and ignoring
6443 any errors: this is one reason why the cert and CRL reader seemed
6444 to work. It doesn't check return codes from the default certificate
6445 routines: these may well fail if the certificates aren't installed.
6448 *) Code to support otherName option in GeneralName.
6451 *) First update to verify code. Change the verify utility
6452 so it warns if it is passed a self signed certificate:
6453 for consistency with the normal behaviour. X509_verify
6454 has been modified to it will now verify a self signed
6455 certificate if *exactly* the same certificate appears
6456 in the store: it was previously impossible to trust a
6457 single self signed certificate. This means that:
6458 openssl verify ss.pem
6459 now gives a warning about a self signed certificate but
6460 openssl verify -CAfile ss.pem ss.pem
6464 *) For servers, store verify_result in SSL_SESSION data structure
6465 (and add it to external session representation).
6466 This is needed when client certificate verifications fails,
6467 but an application-provided verification callback (set by
6468 SSL_CTX_set_cert_verify_callback) allows accepting the session
6469 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6470 but returns 1): When the session is reused, we have to set
6471 ssl->verify_result to the appropriate error code to avoid
6473 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6475 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6476 case in PKCS7_dataInit() where the signed PKCS7 structure
6477 didn't contain any existing data because it was being created.
6478 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6480 *) Add a salt to the key derivation routines in enc.c. This
6481 forms the first 8 bytes of the encrypted file. Also add a
6482 -S option to allow a salt to be input on the command line.
6485 *) New function X509_cmp(). Oddly enough there wasn't a function
6486 to compare two certificates. We do this by working out the SHA1
6487 hash and comparing that. X509_cmp() will be needed by the trust
6491 *) SSL_get1_session() is like SSL_get_session(), but increments
6492 the reference count in the SSL_SESSION returned.
6493 [Geoff Thorpe <geoff@eu.c2.net>]
6495 *) Fix for 'req': it was adding a null to request attributes.
6496 Also change the X509_LOOKUP and X509_INFO code to handle
6497 certificate auxiliary information.
6500 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6504 *) Add the possibility to add extra information to the memory leak
6505 detecting output, to form tracebacks, showing from where each
6506 allocation was originated: CRYPTO_push_info("constant string") adds
6507 the string plus current file name and line number to a per-thread
6508 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6509 is like calling CYRPTO_pop_info() until the stack is empty.
6510 Also updated memory leak detection code to be multi-thread-safe.
6513 *) Add options -text and -noout to pkcs7 utility and delete the
6514 encryption options which never did anything. Update docs.
6517 *) Add options to some of the utilities to allow the pass phrase
6518 to be included on either the command line (not recommended on
6519 OSes like Unix) or read from the environment. Update the
6520 manpages and fix a few bugs.
6523 *) Add a few manpages for some of the openssl commands.
6526 *) Fix the -revoke option in ca. It was freeing up memory twice,
6527 leaking and not finding already revoked certificates.
6530 *) Extensive changes to support certificate auxiliary information.
6531 This involves the use of X509_CERT_AUX structure and X509_AUX
6532 functions. An X509_AUX function such as PEM_read_X509_AUX()
6533 can still read in a certificate file in the usual way but it
6534 will also read in any additional "auxiliary information". By
6535 doing things this way a fair degree of compatibility can be
6536 retained: existing certificates can have this information added
6537 using the new 'x509' options.
6539 Current auxiliary information includes an "alias" and some trust
6540 settings. The trust settings will ultimately be used in enhanced
6541 certificate chain verification routines: currently a certificate
6542 can only be trusted if it is self signed and then it is trusted
6546 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6547 The problem was that one of the replacement routines had not been working
6548 since SSLeay releases. For now the offending routine has been replaced
6549 with non-optimised assembler. Even so, this now gives around 95%
6550 performance improvement for 1024 bit RSA signs.
6553 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6554 handling. Most clients have the effective key size in bits equal to
6555 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6556 A few however don't do this and instead use the size of the decrypted key
6557 to determine the RC2 key length and the AlgorithmIdentifier to determine
6558 the effective key length. In this case the effective key length can still
6559 be 40 bits but the key length can be 168 bits for example. This is fixed
6560 by manually forcing an RC2 key into the EVP_PKEY structure because the
6561 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6562 the key length and effective key length are equal.
6565 *) Add a bunch of functions that should simplify the creation of
6566 X509_NAME structures. Now you should be able to do:
6567 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6568 and have it automatically work out the correct field type and fill in
6569 the structures. The more adventurous can try:
6570 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6571 and it will (hopefully) work out the correct multibyte encoding.
6574 *) Change the 'req' utility to use the new field handling and multibyte
6575 copy routines. Before the DN field creation was handled in an ad hoc
6576 way in req, ca, and x509 which was rather broken and didn't support
6577 BMPStrings or UTF8Strings. Since some software doesn't implement
6578 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6579 using the dirstring_type option. See the new comment in the default
6580 openssl.cnf for more info.
6583 *) Make crypto/rand/md_rand.c more robust:
6584 - Assure unique random numbers after fork().
6585 - Make sure that concurrent threads access the global counter and
6586 md serializably so that we never lose entropy in them
6587 or use exactly the same state in multiple threads.
6588 Access to the large state is not always serializable because
6589 the additional locking could be a performance killer, and
6590 md should be large enough anyway.
6593 *) New file apps/app_rand.c with commonly needed functionality
6594 for handling the random seed file.
6596 Use the random seed file in some applications that previously did not:
6598 dsaparam -genkey (which also ignored its '-rand' option),
6601 x509 (when signing).
6602 Except on systems with /dev/urandom, it is crucial to have a random
6603 seed file at least for key creation, DSA signing, and for DH exchanges;
6604 for RSA signatures we could do without one.
6606 gendh and gendsa (unlike genrsa) used to read only the first byte
6607 of each file listed in the '-rand' option. The function as previously
6608 found in genrsa is now in app_rand.c and is used by all programs
6609 that support '-rand'.
6612 *) In RAND_write_file, use mode 0600 for creating files;
6613 don't just chmod when it may be too late.
6616 *) Report an error from X509_STORE_load_locations
6617 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6620 *) New function ASN1_mbstring_copy() this copies a string in either
6621 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6622 into an ASN1_STRING type. A mask of permissible types is passed
6623 and it chooses the "minimal" type to use or an error if not type
6627 *) Add function equivalents to the various macros in asn1.h. The old
6628 macros are retained with an M_ prefix. Code inside the library can
6629 use the M_ macros. External code (including the openssl utility)
6630 should *NOT* in order to be "shared library friendly".
6633 *) Add various functions that can check a certificate's extensions
6634 to see if it usable for various purposes such as SSL client,
6635 server or S/MIME and CAs of these types. This is currently
6636 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6637 verification. Also added a -purpose flag to x509 utility to
6638 print out all the purposes.
6641 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6645 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6646 for, obtain and decode and extension and obtain its critical flag.
6647 This allows all the necessary extension code to be handled in a
6648 single function call.
6651 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6652 platforms. See crypto/rc4/rc4_enc.c for further details.
6655 *) New -noout option to asn1parse. This causes no output to be produced
6656 its main use is when combined with -strparse and -out to extract data
6657 from a file (which may not be in ASN.1 format).
6660 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6661 when producing the local key id.
6662 [Richard Levitte <levitte@stacken.kth.se>]
6664 *) New option -dhparam in s_server. This allows a DH parameter file to be
6665 stated explicitly. If it is not stated then it tries the first server
6666 certificate file. The previous behaviour hard coded the filename
6670 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6671 a public key to be input or output. For example:
6672 openssl rsa -in key.pem -pubout -out pubkey.pem
6673 Also added necessary DSA public key functions to handle this.
6676 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6677 in the message. This was handled by allowing
6678 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6679 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6681 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6682 to the end of the strings whereas this didn't. This would cause problems
6683 if strings read with d2i_ASN1_bytes() were later modified.
6684 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6686 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6687 data and it contains EOF it will end up returning an error. This is
6688 caused by input 46 bytes long. The cause is due to the way base64
6689 BIOs find the start of base64 encoded data. They do this by trying a
6690 trial decode on each line until they find one that works. When they
6691 do a flag is set and it starts again knowing it can pass all the
6692 data directly through the decoder. Unfortunately it doesn't reset
6693 the context it uses. This means that if EOF is reached an attempt
6694 is made to pass two EOFs through the context and this causes the
6695 resulting error. This can also cause other problems as well. As is
6696 usual with these problems it takes *ages* to find and the fix is
6697 trivial: move one line.
6698 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6700 *) Ugly workaround to get s_client and s_server working under Windows. The
6701 old code wouldn't work because it needed to select() on sockets and the
6702 tty (for keypresses and to see if data could be written). Win32 only
6703 supports select() on sockets so we select() with a 1s timeout on the
6704 sockets and then see if any characters are waiting to be read, if none
6705 are present then we retry, we also assume we can always write data to
6706 the tty. This isn't nice because the code then blocks until we've
6707 received a complete line of data and it is effectively polling the
6708 keyboard at 1s intervals: however it's quite a bit better than not
6709 working at all :-) A dedicated Windows application might handle this
6710 with an event loop for example.
6713 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6714 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6715 will be called when RSA_sign() and RSA_verify() are used. This is useful
6716 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6717 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6718 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6719 This necessitated the support of an extra signature type NID_md5_sha1
6720 for SSL signatures and modifications to the SSL library to use it instead
6721 of calling RSA_public_decrypt() and RSA_private_encrypt().
6724 *) Add new -verify -CAfile and -CApath options to the crl program, these
6725 will lookup a CRL issuers certificate and verify the signature in a
6726 similar way to the verify program. Tidy up the crl program so it
6727 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6728 less strict. It will now permit CRL extensions even if it is not
6729 a V2 CRL: this will allow it to tolerate some broken CRLs.
6732 *) Initialize all non-automatic variables each time one of the openssl
6733 sub-programs is started (this is necessary as they may be started
6734 multiple times from the "OpenSSL>" prompt).
6735 [Lennart Bang, Bodo Moeller]
6737 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6738 removing all other RSA functionality (this is what NO_RSA does). This
6739 is so (for example) those in the US can disable those operations covered
6740 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6744 *) Non-copying interface to BIO pairs.
6745 (still largely untested)
6748 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6749 ASCII string. This was handled independently in various places before.
6752 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6753 UTF8 strings a character at a time.
6756 *) Use client_version from client hello to select the protocol
6757 (s23_srvr.c) and for RSA client key exchange verification
6758 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6761 *) Add various utility functions to handle SPKACs, these were previously
6762 handled by poking round in the structure internals. Added new function
6763 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6764 print, verify and generate SPKACs. Based on an original idea from
6765 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6768 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6771 *) Allow the config file extension section to be overwritten on the
6772 command line. Based on an original idea from Massimiliano Pala
6773 <madwolf@comune.modena.it>. The new option is called -extensions
6774 and can be applied to ca, req and x509. Also -reqexts to override
6775 the request extensions in req and -crlexts to override the crl extensions
6779 *) Add new feature to the SPKAC handling in ca. Now you can include
6780 the same field multiple times by preceding it by "XXXX." for example:
6783 this is the same syntax as used in the req config file.
6786 *) Allow certificate extensions to be added to certificate requests. These
6787 are specified in a 'req_extensions' option of the req section of the
6788 config file. They can be printed out with the -text option to req but
6789 are otherwise ignored at present.
6792 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6793 data read consists of only the final block it would not decrypted because
6794 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6795 A misplaced 'break' also meant the decrypted final block might not be
6796 copied until the next read.
6799 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6800 a few extra parameters to the DH structure: these will be useful if
6801 for example we want the value of 'q' or implement X9.42 DH.
6804 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6805 provides hooks that allow the default DSA functions or functions on a
6806 "per key" basis to be replaced. This allows hardware acceleration and
6807 hardware key storage to be handled without major modification to the
6808 library. Also added low level modexp hooks and CRYPTO_EX structure and
6809 associated functions.
6812 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6813 as "read only": it can't be written to and the buffer it points to will
6814 not be freed. Reading from a read only BIO is much more efficient than
6815 a normal memory BIO. This was added because there are several times when
6816 an area of memory needs to be read from a BIO. The previous method was
6817 to create a memory BIO and write the data to it, this results in two
6818 copies of the data and an O(n^2) reading algorithm. There is a new
6819 function BIO_new_mem_buf() which creates a read only memory BIO from
6820 an area of memory. Also modified the PKCS#7 routines to use read only
6824 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6825 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6826 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6827 but a retry condition occured while trying to read the rest.
6830 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6831 NID_pkcs7_encrypted by default: this was wrong since this should almost
6832 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6833 the encrypted data type: this is a more sensible place to put it and it
6834 allows the PKCS#12 code to be tidied up that duplicated this
6838 *) Changed obj_dat.pl script so it takes its input and output files on
6839 the command line. This should avoid shell escape redirection problems
6843 *) Initial support for certificate extension requests, these are included
6844 in things like Xenroll certificate requests. Included functions to allow
6845 extensions to be obtained and added.
6848 *) -crlf option to s_client and s_server for sending newlines as
6849 CRLF (as required by many protocols).
6852 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6854 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6855 [Ralf S. Engelschall]
6857 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6858 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6860 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6864 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6865 DH parameters/keys (q is lost during that conversion, but the resulting
6866 DH parameters contain its length).
6868 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6869 much faster than DH_generate_parameters (which creates parameters
6870 where p = 2*q + 1), and also the smaller q makes DH computations
6871 much more efficient (160-bit exponentiation instead of 1024-bit
6872 exponentiation); so this provides a convenient way to support DHE
6873 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6874 utter importance to use
6875 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6877 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6878 when such DH parameters are used, because otherwise small subgroup
6879 attacks may become possible!
6882 *) Avoid memory leak in i2d_DHparams.
6885 *) Allow the -k option to be used more than once in the enc program:
6886 this allows the same encrypted message to be read by multiple recipients.
6889 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6890 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6891 it will always use the numerical form of the OID, even if it has a short
6895 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6896 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6897 otherwise bn_mod_exp was called. In the case of hardware keys for example
6898 no private key components need be present and it might store extra data
6899 in the RSA structure, which cannot be accessed from bn_mod_exp.
6900 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6901 private key operations.
6904 *) Added support for SPARC Linux.
6907 *) pem_password_cb function type incompatibly changed from
6908 typedef int pem_password_cb(char *buf, int size, int rwflag);
6910 ....(char *buf, int size, int rwflag, void *userdata);
6911 so that applications can pass data to their callbacks:
6912 The PEM[_ASN1]_{read,write}... functions and macros now take an
6913 additional void * argument, which is just handed through whenever
6914 the password callback is called.
6915 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6917 New function SSL_CTX_set_default_passwd_cb_userdata.
6919 Compatibility note: As many C implementations push function arguments
6920 onto the stack in reverse order, the new library version is likely to
6921 interoperate with programs that have been compiled with the old
6922 pem_password_cb definition (PEM_whatever takes some data that
6923 happens to be on the stack as its last argument, and the callback
6924 just ignores this garbage); but there is no guarantee whatsoever that
6927 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6928 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6929 problems not only on Windows, but also on some Unix platforms.
6930 To avoid problematic command lines, these definitions are now in an
6931 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6932 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6935 *) MIPS III/IV assembler module is reimplemented.
6938 *) More DES library cleanups: remove references to srand/rand and
6939 delete an unused file.
6942 *) Add support for the the free Netwide assembler (NASM) under Win32,
6943 since not many people have MASM (ml) and it can be hard to obtain.
6944 This is currently experimental but it seems to work OK and pass all
6945 the tests. Check out INSTALL.W32 for info.
6948 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6949 without temporary keys kept an extra copy of the server key,
6950 and connections with temporary keys did not free everything in case
6954 *) New function RSA_check_key and new openssl rsa option -check
6955 for verifying the consistency of RSA keys.
6956 [Ulf Moeller, Bodo Moeller]
6958 *) Various changes to make Win32 compile work:
6959 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6960 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6961 comparison" warnings.
6962 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6965 *) Add a debugging option to PKCS#5 v2 key generation function: when
6966 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6967 derived keys are printed to stderr.
6970 *) Copy the flags in ASN1_STRING_dup().
6971 [Roman E. Pavlov <pre@mo.msk.ru>]
6973 *) The x509 application mishandled signing requests containing DSA
6974 keys when the signing key was also DSA and the parameters didn't match.
6976 It was supposed to omit the parameters when they matched the signing key:
6977 the verifying software was then supposed to automatically use the CA's
6978 parameters if they were absent from the end user certificate.
6980 Omitting parameters is no longer recommended. The test was also
6981 the wrong way round! This was probably due to unusual behaviour in
6982 EVP_cmp_parameters() which returns 1 if the parameters match.
6983 This meant that parameters were omitted when they *didn't* match and
6984 the certificate was useless. Certificates signed with 'ca' didn't have
6986 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6988 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6989 The interface is as follows:
6990 Applications can use
6991 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6992 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6993 "off" is now the default.
6994 The library internally uses
6995 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6996 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6997 to disable memory-checking temporarily.
6999 Some inconsistent states that previously were possible (and were
7000 even the default) are now avoided.
7002 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7003 with each memory chunk allocated; this is occasionally more helpful
7004 than just having a counter.
7006 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7008 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7012 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7013 which largely parallels "options", but is for changing API behaviour,
7014 whereas "options" are about protocol behaviour.
7015 Initial "mode" flags are:
7017 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7018 a single record has been written.
7019 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7020 retries use the same buffer location.
7021 (But all of the contents must be
7025 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7028 *) Fix problems with no-hmac etc.
7029 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7031 *) New functions RSA_get_default_method(), RSA_set_method() and
7032 RSA_get_method(). These allows replacement of RSA_METHODs without having
7033 to mess around with the internals of an RSA structure.
7036 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7037 Also really enable memory leak checks in openssl.c and in some
7039 [Chad C. Mulligan, Bodo Moeller]
7041 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7042 up the length of negative integers. This has now been simplified to just
7043 store the length when it is first determined and use it later, rather
7044 than trying to keep track of where data is copied and updating it to
7046 [Steve Henson, reported by Brien Wheeler
7047 <bwheeler@authentica-security.com>]
7049 *) Add a new function PKCS7_signatureVerify. This allows the verification
7050 of a PKCS#7 signature but with the signing certificate passed to the
7051 function itself. This contrasts with PKCS7_dataVerify which assumes the
7052 certificate is present in the PKCS#7 structure. This isn't always the
7053 case: certificates can be omitted from a PKCS#7 structure and be
7054 distributed by "out of band" means (such as a certificate database).
7057 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7058 function prototypes in pem.h, also change util/mkdef.pl to add the
7059 necessary function names.
7062 *) mk1mf.pl (used by Windows builds) did not properly read the
7063 options set by Configure in the top level Makefile, and Configure
7064 was not even able to write more than one option correctly.
7065 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7068 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7069 file to be loaded from a BIO or FILE pointer. The BIO version will
7070 for example allow memory BIOs to contain config info.
7073 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7074 Whoever hopes to achieve shared-library compatibility across versions
7075 must use this, not the compile-time macro.
7076 (Exercise 0.9.4: Which is the minimum library version required by
7078 Note: All this applies only to multi-threaded programs, others don't
7082 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7083 through a BIO pair triggered the default case, i.e.
7084 SSLerr(...,SSL_R_UNKNOWN_STATE).
7087 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7088 can use the SSL library even if none of the specific BIOs is
7092 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7093 for the encoded length.
7094 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7096 *) Add initial documentation of the X509V3 functions.
7099 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7100 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7101 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7102 secure PKCS#8 private key format with a high iteration count.
7105 *) Fix determination of Perl interpreter: A perl or perl5
7106 _directory_ in $PATH was also accepted as the interpreter.
7107 [Ralf S. Engelschall]
7109 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7110 wrong with it but it was very old and did things like calling
7111 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7115 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7116 to use the new extension code.
7119 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7120 with macros. This should make it easier to change their form, add extra
7121 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7125 *) Add to configuration table a new entry that can specify an alternative
7126 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7127 according to Mark Crispin <MRC@Panda.COM>.
7131 *) DES CBC did not update the IV. Weird.
7134 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7135 Changing the behaviour of the former might break existing programs --
7136 where IV updating is needed, des_ncbc_encrypt can be used.
7139 *) When bntest is run from "make test" it drives bc to check its
7140 calculations, as well as internally checking them. If an internal check
7141 fails, it needs to cause bc to give a non-zero result or make test carries
7142 on without noticing the failure. Fixed.
7145 *) DES library cleanups.
7148 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7149 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7150 ciphers. NOTE: although the key derivation function has been verified
7151 against some published test vectors it has not been extensively tested
7152 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7156 *) Instead of "mkdir -p", which is not fully portable, use new
7157 Perl script "util/mkdir-p.pl".
7160 *) Rewrite the way password based encryption (PBE) is handled. It used to
7161 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7162 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7163 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7164 the 'parameter' field of the AlgorithmIdentifier is passed to the
7165 underlying key generation function so it must do its own ASN1 parsing.
7166 This has also changed the EVP_PBE_CipherInit() function which now has a
7167 'parameter' argument instead of literal salt and iteration count values
7168 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7171 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7172 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7173 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7174 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7175 value was just used as a "magic string" and not used directly its
7176 value doesn't matter.
7179 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7183 *) "linux-sparc64" configuration (ultrapenguin).
7184 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7185 "linux-sparc" configuration.
7186 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7188 *) config now generates no-xxx options for missing ciphers.
7191 *) Support the EBCDIC character set (work in progress).
7192 File ebcdic.c not yet included because it has a different license.
7193 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7195 *) Support BS2000/OSD-POSIX.
7196 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7198 *) Make callbacks for key generation use void * instead of char *.
7201 *) Make S/MIME samples compile (not yet tested).
7204 *) Additional typesafe stacks.
7207 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7211 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7213 *) New configuration variant "sco5-gcc".
7215 *) Updated some demos.
7216 [Sean O Riordain, Wade Scholine]
7218 *) Add missing BIO_free at exit of pkcs12 application.
7221 *) Fix memory leak in conf.c.
7224 *) Updates for Win32 to assembler version of MD5.
7227 *) Set #! path to perl in apps/der_chop to where we found it
7228 instead of using a fixed path.
7231 *) SHA library changes for irix64-mips4-cc.
7234 *) Improvements for VMS support.
7238 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7240 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7241 This also avoids the problems with SC4.2 and unpatched SC5.
7242 [Andy Polyakov <appro@fy.chalmers.se>]
7244 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7245 These are required because of the typesafe stack would otherwise break
7246 existing code. If old code used a structure member which used to be STACK
7247 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7248 sk_num or sk_value it would produce an error because the num, data members
7249 are not present in STACK_OF. Now it just produces a warning. sk_set
7250 replaces the old method of assigning a value to sk_value
7251 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7252 that does this will no longer work (and should use sk_set instead) but
7253 this could be regarded as a "questionable" behaviour anyway.
7256 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7257 correctly handle encrypted S/MIME data.
7260 *) Change type of various DES function arguments from des_cblock
7261 (which means, in function argument declarations, pointer to char)
7262 to des_cblock * (meaning pointer to array with 8 char elements),
7263 which allows the compiler to do more typechecking; it was like
7264 that back in SSLeay, but with lots of ugly casts.
7266 Introduce new type const_des_cblock.
7269 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7270 problems: find RecipientInfo structure that matches recipient certificate
7271 and initialise the ASN1 structures properly based on passed cipher.
7274 *) Belatedly make the BN tests actually check the results.
7277 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7278 to and from BNs: it was completely broken. New compilation option
7279 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7280 key elements as negative integers.
7283 *) Reorganize and speed up MD5.
7284 [Andy Polyakov <appro@fy.chalmers.se>]
7287 [Richard Levitte <richard@levitte.org>]
7289 *) New option -out to asn1parse to allow the parsed structure to be
7290 output to a file. This is most useful when combined with the -strparse
7291 option to examine the output of things like OCTET STRINGS.
7294 *) Make SSL library a little more fool-proof by not requiring any longer
7295 that SSL_set_{accept,connect}_state be called before
7296 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7297 in many applications because usually everything *appeared* to work as
7298 intended anyway -- now it really works as intended).
7301 *) Move openssl.cnf out of lib/.
7304 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7305 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7306 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7307 [Ralf S. Engelschall]
7309 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7310 handle PKCS#7 enveloped data properly.
7311 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7313 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7314 copying pointers. The cert_st handling is changed by this in
7315 various ways (and thus what used to be known as ctx->default_cert
7316 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7317 any longer when s->cert does not give us what we need).
7318 ssl_cert_instantiate becomes obsolete by this change.
7319 As soon as we've got the new code right (possibly it already is?),
7320 we have solved a couple of bugs of the earlier code where s->cert
7321 was used as if it could not have been shared with other SSL structures.
7323 Note that using the SSL API in certain dirty ways now will result
7324 in different behaviour than observed with earlier library versions:
7325 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7326 does not influence s as it used to.
7328 In order to clean up things more thoroughly, inside SSL_SESSION
7329 we don't use CERT any longer, but a new structure SESS_CERT
7330 that holds per-session data (if available); currently, this is
7331 the peer's certificate chain and, for clients, the server's certificate
7332 and temporary key. CERT holds only those values that can have
7333 meaningful defaults in an SSL_CTX.
7336 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7337 from the internal representation. Various PKCS#7 fixes: remove some
7338 evil casts and set the enc_dig_alg field properly based on the signing
7342 *) Allow PKCS#12 password to be set from the command line or the
7343 environment. Let 'ca' get its config file name from the environment
7344 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7348 *) Allow certificate policies extension to use an IA5STRING for the
7349 organization field. This is contrary to the PKIX definition but
7350 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7354 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7355 without disallowing inline assembler and the like for non-pedantic builds.
7358 *) Support Borland C++ builder.
7359 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7364 *) SHA-1 cleanups and performance enhancements.
7365 [Andy Polyakov <appro@fy.chalmers.se>]
7367 *) Sparc v8plus assembler for the bignum library.
7368 [Andy Polyakov <appro@fy.chalmers.se>]
7370 *) Accept any -xxx and +xxx compiler options in Configure.
7373 *) Update HPUX configuration.
7376 *) Add missing sk_<type>_unshift() function to safestack.h
7377 [Ralf S. Engelschall]
7379 *) New function SSL_CTX_use_certificate_chain_file that sets the
7380 "extra_cert"s in addition to the certificate. (This makes sense
7381 only for "PEM" format files, as chains as a whole are not
7385 *) Support verify_depth from the SSL API.
7386 x509_vfy.c had what can be considered an off-by-one-error:
7387 Its depth (which was not part of the external interface)
7388 was actually counting the number of certificates in a chain;
7389 now it really counts the depth.
7392 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7393 instead of X509err, which often resulted in confusing error
7394 messages since the error codes are not globally unique
7395 (e.g. an alleged error in ssl3_accept when a certificate
7396 didn't match the private key).
7398 *) New function SSL_CTX_set_session_id_context that allows to set a default
7399 value (so that you don't need SSL_set_session_id_context for each
7400 connection using the SSL_CTX).
7403 *) OAEP decoding bug fix.
7406 *) Support INSTALL_PREFIX for package builders, as proposed by
7410 *) New Configure options "threads" and "no-threads". For systems
7411 where the proper compiler options are known (currently Solaris
7412 and Linux), "threads" is the default.
7415 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7418 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7419 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7420 such as /usr/local/bin.
7423 *) "make linux-shared" to build shared libraries.
7424 [Niels Poppe <niels@netbox.org>]
7426 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7429 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7430 extension adding in x509 utility.
7433 *) Remove NOPROTO sections and error code comments.
7436 *) Partial rewrite of the DEF file generator to now parse the ANSI
7440 *) New Configure options --prefix=DIR and --openssldir=DIR.
7443 *) Complete rewrite of the error code script(s). It is all now handled
7444 by one script at the top level which handles error code gathering,
7445 header rewriting and C source file generation. It should be much better
7446 than the old method: it now uses a modified version of Ulf's parser to
7447 read the ANSI prototypes in all header files (thus the old K&R definitions
7448 aren't needed for error creation any more) and do a better job of
7449 translating function codes into names. The old 'ASN1 error code imbedded
7450 in a comment' is no longer necessary and it doesn't use .err files which
7451 have now been deleted. Also the error code call doesn't have to appear all
7452 on one line (which resulted in some large lines...).
7455 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7458 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7459 0 (which usually indicates a closed connection), but continue reading.
7462 *) Fix some race conditions.
7465 *) Add support for CRL distribution points extension. Add Certificate
7466 Policies and CRL distribution points documentation.
7469 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7472 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7473 8 of keying material. Merlin has also confirmed interop with this fix
7474 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7475 [Merlin Hughes <merlin@baltimore.ie>]
7477 *) Fix lots of warnings.
7478 [Richard Levitte <levitte@stacken.kth.se>]
7480 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7481 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7482 [Richard Levitte <levitte@stacken.kth.se>]
7484 *) Fix problems with sizeof(long) == 8.
7485 [Andy Polyakov <appro@fy.chalmers.se>]
7487 *) Change functions to ANSI C.
7490 *) Fix typos in error codes.
7491 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7493 *) Remove defunct assembler files from Configure.
7496 *) SPARC v8 assembler BIGNUM implementation.
7497 [Andy Polyakov <appro@fy.chalmers.se>]
7499 *) Support for Certificate Policies extension: both print and set.
7500 Various additions to support the r2i method this uses.
7503 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7504 return a const string when you are expecting an allocated buffer.
7507 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7508 types DirectoryString and DisplayText.
7511 *) Add code to allow r2i extensions to access the configuration database,
7512 add an LHASH database driver and add several ctx helper functions.
7515 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7516 fail when they extended the size of a BIGNUM.
7519 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7520 support typesafe stack.
7523 *) Fix typo in SSL_[gs]et_options().
7524 [Nils Frostberg <nils@medcom.se>]
7526 *) Delete various functions and files that belonged to the (now obsolete)
7527 old X509V3 handling code.
7530 *) New Configure option "rsaref".
7533 *) Don't auto-generate pem.h.
7536 *) Introduce type-safe ASN.1 SETs.
7539 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7540 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7542 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7543 that links with OpenSSL (well at least cause lots of warnings), but fear
7544 not: the conversion is trivial, and it eliminates loads of evil casts. A
7545 few STACKed things have been converted already. Feel free to convert more.
7546 In the fullness of time, I'll do away with the STACK type altogether.
7549 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7550 specified in <certfile> by updating the entry in the index.txt file.
7551 This way one no longer has to edit the index.txt file manually for
7552 revoking a certificate. The -revoke option does the gory details now.
7553 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7555 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7556 `-text' option at all and this way the `-noout -text' combination was
7557 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7558 [Ralf S. Engelschall]
7560 *) Make sure a corresponding plain text error message exists for the
7561 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7562 verify callback function determined that a certificate was revoked.
7563 [Ralf S. Engelschall]
7565 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7566 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7567 all available cipers including rc5, which was forgotten until now.
7568 In order to let the testing shell script know which algorithms
7569 are available, a new (up to now undocumented) command
7570 "openssl list-cipher-commands" is used.
7573 *) Bugfix: s_client occasionally would sleep in select() when
7574 it should have checked SSL_pending() first.
7577 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7578 the raw DSA values prior to ASN.1 encoding.
7581 *) Tweaks to Configure
7582 [Niels Poppe <niels@netbox.org>]
7584 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7588 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7591 *) New config option to avoid instructions that are illegal on the 80386.
7592 The default code is faster, but requires at least a 486.
7595 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7596 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7597 same as SSL2_VERSION anyway.
7600 *) New "-showcerts" option for s_client.
7603 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7604 application. Various cleanups and fixes.
7607 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7608 modify error routines to work internally. Add error codes and PBE init
7609 to library startup routines.
7612 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7613 packing functions to asn1 and evp. Changed function names and error
7614 codes along the way.
7617 *) PKCS12 integration: and so it begins... First of several patches to
7618 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7619 objects to objects.h
7622 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7623 and display support for Thawte strong extranet extension.
7626 *) Add LinuxPPC support.
7627 [Jeff Dubrule <igor@pobox.org>]
7629 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7630 bn_div_words in alpha.s.
7631 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7633 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7634 OAEP isn't supported when OpenSSL is built with RSAref.
7635 [Ulf Moeller <ulf@fitug.de>]
7637 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7638 so they no longer are missing under -DNOPROTO.
7639 [Soren S. Jorvang <soren@t.dk>]
7642 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7644 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7645 doesn't work when the session is reused. Coming soon!
7648 *) Fix a security hole, that allows sessions to be reused in the wrong
7649 context thus bypassing client cert protection! All software that uses
7650 client certs and session caches in multiple contexts NEEDS PATCHING to
7651 allow session reuse! A fuller solution is in the works.
7652 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7654 *) Some more source tree cleanups (removed obsolete files
7655 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7656 permission on "config" script to be executable) and a fix for the INSTALL
7658 [Ulf Moeller <ulf@fitug.de>]
7660 *) Remove some legacy and erroneous uses of malloc, free instead of
7662 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7664 *) Make rsa_oaep_test return non-zero on error.
7665 [Ulf Moeller <ulf@fitug.de>]
7667 *) Add support for native Solaris shared libraries. Configure
7668 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7669 if someone would make that last step automatic.
7670 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7672 *) ctx_size was not built with the right compiler during "make links". Fixed.
7675 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7676 except NULL ciphers". This means the default cipher list will no longer
7677 enable NULL ciphers. They need to be specifically enabled e.g. with
7678 the string "DEFAULT:eNULL".
7681 *) Fix to RSA private encryption routines: if p < q then it would
7682 occasionally produce an invalid result. This will only happen with
7683 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7686 *) Be less restrictive and allow also `perl util/perlpath.pl
7687 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7688 because this way one can also use an interpreter named `perl5' (which is
7689 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7690 installed as `perl').
7691 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7693 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7694 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7696 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7697 advapi32.lib to Win32 build and change the pem test comparision
7698 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7699 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7700 and crypto/des/ede_cbcm_enc.c.
7703 *) DES quad checksum was broken on big-endian architectures. Fixed.
7706 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7707 Win32 test batch file so it (might) work again. The Win32 test batch file
7708 is horrible: I feel ill....
7711 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7712 in e_os.h. Audit of header files to check ANSI and non ANSI
7713 sections: 10 functions were absent from non ANSI section and not exported
7714 from Windows DLLs. Fixed up libeay.num for new functions.
7717 *) Make `openssl version' output lines consistent.
7718 [Ralf S. Engelschall]
7720 *) Fix Win32 symbol export lists for BIO functions: Added
7721 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7722 to ms/libeay{16,32}.def.
7723 [Ralf S. Engelschall]
7725 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7726 fine under Unix and passes some trivial tests I've now added. But the
7727 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7728 added to make sure no one expects that this stuff really works in the
7729 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7730 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7732 [Ralf S. Engelschall]
7734 *) Fix the generation of two part addresses in perl.
7735 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7737 *) Add config entry for Linux on MIPS.
7738 [John Tobey <jtobey@channel1.com>]
7740 *) Make links whenever Configure is run, unless we are on Windoze.
7743 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7744 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7748 *) Add a useful kludge to allow package maintainers to specify compiler and
7749 other platforms details on the command line without having to patch the
7750 Configure script everytime: One now can use ``perl Configure
7751 <id>:<details>'', i.e. platform ids are allowed to have details appended
7752 to them (seperated by colons). This is treated as there would be a static
7753 pre-configured entry in Configure's %table under key <id> with value
7754 <details> and ``perl Configure <id>'' is called. So, when you want to
7755 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7756 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7757 now, which overrides the FreeBSD-elf entry on-the-fly.
7758 [Ralf S. Engelschall]
7760 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7763 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7764 on the `perl Configure ...' command line. This way one can compile
7765 OpenSSL libraries with Position Independent Code (PIC) which is needed
7766 for linking it into DSOs.
7767 [Ralf S. Engelschall]
7769 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7773 *) Cleaned up the LICENSE document: The official contact for any license
7774 questions now is the OpenSSL core team under openssl-core@openssl.org.
7775 And add a paragraph about the dual-license situation to make sure people
7776 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7777 to the OpenSSL toolkit.
7778 [Ralf S. Engelschall]
7780 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7781 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7782 Additonally cleaned up the `make links' target: Remove unnecessary
7783 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7784 to speed processing and no longer clutter the display with confusing
7785 stuff. Instead only the actually done links are displayed.
7786 [Ralf S. Engelschall]
7788 *) Permit null encryption ciphersuites, used for authentication only. It used
7789 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7790 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7794 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7795 signed attributes when verifying signatures (this would break them),
7796 the detached data encoding was wrong and public keys obtained using
7797 X509_get_pubkey() weren't freed.
7800 *) Add text documentation for the BUFFER functions. Also added a work around
7801 to a Win95 console bug. This was triggered by the password read stuff: the
7802 last character typed gets carried over to the next fread(). If you were
7803 generating a new cert request using 'req' for example then the last
7804 character of the passphrase would be CR which would then enter the first
7808 *) Added the new `Includes OpenSSL Cryptography Software' button as
7809 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7810 button and can be used by applications based on OpenSSL to show the
7811 relationship to the OpenSSL project.
7812 [Ralf S. Engelschall]
7814 *) Remove confusing variables in function signatures in files
7815 ssl/ssl_lib.c and ssl/ssl.h.
7816 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7818 *) Don't install bss_file.c under PREFIX/include/
7819 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7821 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7822 functions that return function pointers and has support for NT specific
7823 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7824 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7825 unsigned to signed types: this was killing the Win32 compile.
7828 *) Add new certificate file to stack functions,
7829 SSL_add_dir_cert_subjects_to_stack() and
7830 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7831 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7832 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7833 This means that Apache-SSL and similar packages don't have to mess around
7834 to add as many CAs as they want to the preferred list.
7837 *) Experiment with doxygen documentation. Currently only partially applied to
7839 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7840 openssl.doxy as the configuration file.
7843 *) Get rid of remaining C++-style comments which strict C compilers hate.
7844 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7846 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7847 compiled in by default: it has problems with large keys.
7850 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7851 DH private keys and/or callback functions which directly correspond to
7852 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7853 is needed for applications which have to configure certificates on a
7854 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7856 For the RSA certificate situation is makes no difference, but
7857 for the DSA certificate situation this fixes the "no shared cipher"
7858 problem where the OpenSSL cipher selection procedure failed because the
7859 temporary keys were not overtaken from the context and the API provided
7860 no way to reconfigure them.
7861 The new functions now let applications reconfigure the stuff and they
7862 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7863 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7864 non-public-API function ssl_cert_instantiate() is used as a helper
7865 function and also to reduce code redundancy inside ssl_rsa.c.
7866 [Ralf S. Engelschall]
7868 *) Move s_server -dcert and -dkey options out of the undocumented feature
7869 area because they are useful for the DSA situation and should be
7870 recognized by the users.
7871 [Ralf S. Engelschall]
7873 *) Fix the cipher decision scheme for export ciphers: the export bits are
7874 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7875 SSL_EXP_MASK. So, the original variable has to be used instead of the
7876 already masked variable.
7877 [Richard Levitte <levitte@stacken.kth.se>]
7879 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7880 [Richard Levitte <levitte@stacken.kth.se>]
7882 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7883 from `int' to `unsigned int' because it's a length and initialized by
7884 EVP_DigestFinal() which expects an `unsigned int *'.
7885 [Richard Levitte <levitte@stacken.kth.se>]
7887 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7888 script. Instead use the usual Shell->Perl transition trick.
7889 [Ralf S. Engelschall]
7891 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7892 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7893 -noout -modulus' as it's already the case for `openssl rsa -noout
7894 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7895 currently the public key is printed (a decision which was already done by
7896 `openssl dsa -modulus' in the past) which serves a similar purpose.
7897 Additionally the NO_RSA no longer completely removes the whole -modulus
7898 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7900 [Ralf S. Engelschall]
7902 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7903 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7904 [Arne Ansper <arne@ats.cyber.ee>]
7906 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7907 to be added. Now both 'req' and 'ca' can use new objects defined in the
7911 *) Add cool BIO that does syslog (or event log on NT).
7912 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7914 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7915 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7916 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7917 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7920 *) Add preliminary config info for new extension code.
7923 *) Make RSA_NO_PADDING really use no padding.
7924 [Ulf Moeller <ulf@fitug.de>]
7926 *) Generate errors when private/public key check is done.
7929 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7930 for some CRL extensions and new objects added.
7933 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7934 key usage extension and fuller support for authority key id.
7937 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7938 padding method for RSA, which is recommended for new applications in PKCS
7939 #1 v2.0 (RFC 2437, October 1998).
7940 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7941 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7942 against Bleichbacher's attack on RSA.
7943 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7946 *) Updates to the new SSL compression code
7947 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7949 *) Fix so that the version number in the master secret, when passed
7950 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7951 (because the server will not accept higher), that the version number
7952 is 0x03,0x01, not 0x03,0x00
7953 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7955 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7956 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7957 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7960 *) Support for RAW extensions where an arbitrary extension can be
7961 created by including its DER encoding. See apps/openssl.cnf for
7965 *) Make sure latest Perl versions don't interpret some generated C array
7966 code as Perl array code in the crypto/err/err_genc.pl script.
7967 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7969 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7970 not many people have the assembler. Various Win32 compilation fixes and
7971 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7975 *) Modify configure script 'Configure' to automatically create crypto/date.h
7976 file under Win32 and also build pem.h from pem.org. New script
7977 util/mkfiles.pl to create the MINFO file on environments that can't do a
7978 'make files': perl util/mkfiles.pl >MINFO should work.
7981 *) Major rework of DES function declarations, in the pursuit of correctness
7982 and purity. As a result, many evil casts evaporated, and some weirdness,
7983 too. You may find this causes warnings in your code. Zapping your evil
7984 casts will probably fix them. Mostly.
7987 *) Fix for a typo in asn1.h. Bug fix to object creation script
7988 obj_dat.pl. It considered a zero in an object definition to mean
7989 "end of object": none of the objects in objects.h have any zeros
7990 so it wasn't spotted.
7991 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7993 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7994 Masking (CBCM). In the absence of test vectors, the best I have been able
7995 to do is check that the decrypt undoes the encrypt, so far. Send me test
7996 vectors if you have them.
7999 *) Correct calculation of key length for export ciphers (too much space was
8000 allocated for null ciphers). This has not been tested!
8003 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8004 message is now correct (it understands "crypto" and "ssl" on its
8005 command line). There is also now an "update" option. This will update
8006 the util/ssleay.num and util/libeay.num files with any new functions.
8008 perl util/mkdef.pl crypto ssl update
8009 it will update them.
8012 *) Overhauled the Perl interface (perl/*):
8013 - ported BN stuff to OpenSSL's different BN library
8014 - made the perl/ source tree CVS-aware
8015 - renamed the package from SSLeay to OpenSSL (the files still contain
8016 their history because I've copied them in the repository)
8017 - removed obsolete files (the test scripts will be replaced
8018 by better Test::Harness variants in the future)
8019 [Ralf S. Engelschall]
8021 *) First cut for a very conservative source tree cleanup:
8022 1. merge various obsolete readme texts into doc/ssleay.txt
8023 where we collect the old documents and readme texts.
8024 2. remove the first part of files where I'm already sure that we no
8025 longer need them because of three reasons: either they are just temporary
8026 files which were left by Eric or they are preserved original files where
8027 I've verified that the diff is also available in the CVS via "cvs diff
8028 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8029 the crypto/md/ stuff).
8030 [Ralf S. Engelschall]
8032 *) More extension code. Incomplete support for subject and issuer alt
8033 name, issuer and authority key id. Change the i2v function parameters
8034 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8035 what that's for :-) Fix to ASN1 macro which messed up
8036 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8039 *) Preliminary support for ENUMERATED type. This is largely copied from the
8043 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8044 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8046 *) Make sure `make rehash' target really finds the `openssl' program.
8047 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8049 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8050 like to hear about it if this slows down other processors.
8053 *) Add CygWin32 platform information to Configure script.
8054 [Alan Batie <batie@aahz.jf.intel.com>]
8056 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8057 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8059 *) New program nseq to manipulate netscape certificate sequences
8062 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8066 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8067 but the BN code had some problems that would cause failures when
8068 doing certificate verification and some other functions.
8069 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8071 *) Add ASN1 and PEM code to support netscape certificate sequences.
8074 *) Add ASN1 and PEM code to support netscape certificate sequences.
8077 *) Add several PKIX and private extended key usage OIDs.
8080 *) Modify the 'ca' program to handle the new extension code. Modify
8081 openssl.cnf for new extension format, add comments.
8084 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8085 and add a sample to openssl.cnf so req -x509 now adds appropriate
8089 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8090 error code, add initial support to X509_print() and x509 application.
8093 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8094 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8095 stuff is currently isolated and isn't even compiled yet.
8098 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8099 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8100 Removed the versions check from X509 routines when loading extensions:
8101 this allows certain broken certificates that don't set the version
8102 properly to be processed.
8105 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8106 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8107 can still be regenerated with "make depend".
8110 *) Spelling mistake in C version of CAST-128.
8111 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8113 *) Changes to the error generation code. The perl script err-code.pl
8114 now reads in the old error codes and retains the old numbers, only
8115 adding new ones if necessary. It also only changes the .err files if new
8116 codes are added. The makefiles have been modified to only insert errors
8117 when needed (to avoid needlessly modifying header files). This is done
8118 by only inserting errors if the .err file is newer than the auto generated
8119 C file. To rebuild all the error codes from scratch (the old behaviour)
8120 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8121 or delete all the .err files.
8124 *) CAST-128 was incorrectly implemented for short keys. The C version has
8125 been fixed, but is untested. The assembler versions are also fixed, but
8126 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8127 to regenerate it if needed.
8128 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8129 Hagino <itojun@kame.net>]
8131 *) File was opened incorrectly in randfile.c.
8132 [Ulf Möller <ulf@fitug.de>]
8134 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8135 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8136 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8137 al: it's just almost always a UTCTime. Note this patch adds new error
8138 codes so do a "make errors" if there are problems.
8141 *) Correct Linux 1 recognition in config.
8142 [Ulf Möller <ulf@fitug.de>]
8144 *) Remove pointless MD5 hash when using DSA keys in ca.
8145 [Anonymous <nobody@replay.com>]
8147 *) Generate an error if given an empty string as a cert directory. Also
8148 generate an error if handed NULL (previously returned 0 to indicate an
8149 error, but didn't set one).
8150 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8152 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8155 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8156 parameters. This was causing a warning which killed off the Win32 compile.
8159 *) Remove C++ style comments from crypto/bn/bn_local.h.
8160 [Neil Costigan <neil.costigan@celocom.com>]
8162 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8163 based on a text string, looking up short and long names and finally
8164 "dot" format. The "dot" format stuff didn't work. Added new function
8165 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8166 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8167 OID is not part of the table.
8170 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8171 X509_LOOKUP_by_alias().
8174 *) Sort openssl functions by name.
8177 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8178 encryption from sample DSA keys (in case anyone is interested the password
8182 *) Make _all_ *_free functions accept a NULL pointer.
8183 [Frans Heymans <fheymans@isaserver.be>]
8185 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8187 [Anonymous <nobody@replay.com>]
8189 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8190 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8192 *) Don't blow it for numeric -newkey arguments to apps/req.
8193 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8195 *) Temp key "for export" tests were wrong in s3_srvr.c.
8196 [Anonymous <nobody@replay.com>]
8198 *) Add prototype for temp key callback functions
8199 SSL_CTX_set_tmp_{rsa,dh}_callback().
8202 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8203 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8206 *) X509_name_add_entry() freed the wrong thing after an error.
8207 [Arne Ansper <arne@ats.cyber.ee>]
8209 *) rsa_eay.c would attempt to free a NULL context.
8210 [Arne Ansper <arne@ats.cyber.ee>]
8212 *) BIO_s_socket() had a broken should_retry() on Windoze.
8213 [Arne Ansper <arne@ats.cyber.ee>]
8215 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8216 [Arne Ansper <arne@ats.cyber.ee>]
8218 *) Make sure the already existing X509_STORE->depth variable is initialized
8219 in X509_STORE_new(), but document the fact that this variable is still
8220 unused in the certificate verification process.
8221 [Ralf S. Engelschall]
8223 *) Fix the various library and apps files to free up pkeys obtained from
8224 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8227 *) Fix reference counting in X509_PUBKEY_get(). This makes
8228 demos/maurice/example2.c work, amongst others, probably.
8229 [Steve Henson and Ben Laurie]
8231 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8232 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8233 are no longer created. This way we have a single and consistent command
8234 line interface `openssl <command>', similar to `cvs <command>'.
8235 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8237 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8238 BIT STRING wrapper always have zero unused bits.
8241 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8244 *) Make the top-level INSTALL documentation easier to understand.
8247 *) Makefiles updated to exit if an error occurs in a sub-directory
8248 make (including if user presses ^C) [Paul Sutton]
8250 *) Make Montgomery context stuff explicit in RSA data structure.
8253 *) Fix build order of pem and err to allow for generated pem.h.
8256 *) Fix renumbering bug in X509_NAME_delete_entry().
8259 *) Enhanced the err-ins.pl script so it makes the error library number
8260 global and can add a library name. This is needed for external ASN1 and
8261 other error libraries.
8264 *) Fixed sk_insert which never worked properly.
8267 *) Fix ASN1 macros so they can handle indefinite length construted
8268 EXPLICIT tags. Some non standard certificates use these: they can now
8272 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8273 into a single doc/ssleay.txt bundle. This way the information is still
8274 preserved but no longer messes up this directory. Now it's new room for
8275 the new set of documenation files.
8276 [Ralf S. Engelschall]
8278 *) SETs were incorrectly DER encoded. This was a major pain, because they
8279 shared code with SEQUENCEs, which aren't coded the same. This means that
8280 almost everything to do with SETs or SEQUENCEs has either changed name or
8281 number of arguments.
8282 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8284 *) Fix test data to work with the above.
8287 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8288 was already fixed by Eric for 0.9.1 it seems.
8289 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8291 *) Autodetect FreeBSD3.
8294 *) Fix various bugs in Configure. This affects the following platforms:
8298 unixware-2.0-pentium
8302 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8303 before they are needed.
8306 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8310 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8312 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8313 changed SSLeay to OpenSSL in version strings.
8314 [Ralf S. Engelschall]
8316 *) Some fixups to the top-level documents.
8319 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8320 because the symlink to include/ was missing.
8321 [Ralf S. Engelschall]
8323 *) Incorporated the popular no-RSA/DSA-only patches
8324 which allow to compile a RSA-free SSLeay.
8325 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8327 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8328 when "ssleay" is still not found.
8329 [Ralf S. Engelschall]
8331 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8332 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8334 *) Updated the README file.
8335 [Ralf S. Engelschall]
8337 *) Added various .cvsignore files in the CVS repository subdirs
8338 to make a "cvs update" really silent.
8339 [Ralf S. Engelschall]
8341 *) Recompiled the error-definition header files and added
8342 missing symbols to the Win32 linker tables.
8343 [Ralf S. Engelschall]
8345 *) Cleaned up the top-level documents;
8346 o new files: CHANGES and LICENSE
8347 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8348 o merged COPYRIGHT into LICENSE
8349 o removed obsolete TODO file
8350 o renamed MICROSOFT to INSTALL.W32
8351 [Ralf S. Engelschall]
8353 *) Removed dummy files from the 0.9.1b source tree:
8354 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8355 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8356 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8357 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8358 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8359 [Ralf S. Engelschall]
8361 *) Added various platform portability fixes.
8364 *) The Genesis of the OpenSSL rpject:
8365 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8366 Young and Tim J. Hudson created while they were working for C2Net until
8368 [The OpenSSL Project]
8371 Changes between 0.9.0b and 0.9.1b [not released]
8373 *) Updated a few CA certificates under certs/
8376 *) Changed some BIGNUM api stuff.
8379 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8380 DGUX x86, Linux Alpha, etc.
8383 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8384 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8388 *) Add -strparse option to asn1pars program which parses nested
8390 [Dr Stephen Henson <shenson@bigfoot.com>]
8392 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8395 *) DSA fix for "ca" program.
8398 *) Added "-genkey" option to "dsaparam" program.
8401 *) Added RIPE MD160 (rmd160) message digest.
8404 *) Added -a (all) option to "ssleay version" command.
8407 *) Added PLATFORM define which is the id given to Configure.
8410 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8413 *) Extended the ASN.1 parser routines.
8416 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8419 *) Added a BN_CTX to the BN library.
8422 *) Fixed the weak key values in DES library
8425 *) Changed API in EVP library for cipher aliases.
8428 *) Added support for RC2/64bit cipher.
8431 *) Converted the lhash library to the crypto/mem.c functions.
8434 *) Added more recognized ASN.1 object ids.
8437 *) Added more RSA padding checks for SSL/TLS.
8440 *) Added BIO proxy/filter functionality.
8443 *) Added extra_certs to SSL_CTX which can be used
8444 send extra CA certificates to the client in the CA cert chain sending
8445 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8448 *) Now Fortezza is denied in the authentication phase because
8449 this is key exchange mechanism is not supported by SSLeay at all.
8452 *) Additional PKCS1 checks.
8455 *) Support the string "TLSv1" for all TLS v1 ciphers.
8458 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8459 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8462 *) Fixed a few memory leaks.
8465 *) Fixed various code and comment typos.
8468 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8469 bytes sent in the client random.
8470 [Edward Bishop <ebishop@spyglass.com>]