5 Changes between 0.9.8k and 0.9.8l [xx XXX xxxx]
7 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
11 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
12 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
14 [Guenter <lists@gknw.net>]
16 *) Fix the server certificate chain building code to use X509_verify_cert(),
17 it used to have an ad-hoc builder which was unable to cope with anything
18 other than a simple chain.
19 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
21 *) Don't check self signed certificate signatures in X509_verify_cert()
22 by default (a flag can override this): it just wastes time without
23 adding any security. As a useful side effect self signed root CAs
24 with non-FIPS digests are now usable in FIPS mode.
27 *) In dtls1_process_out_of_seq_message() the check if the current message
28 is already buffered was missing. For every new message was memory
29 allocated, allowing an attacker to perform an denial of service attack
30 with sending out of seq handshake messages until there is no memory
31 left. Additionally every future messege was buffered, even if the
32 sequence number made no sense and would be part of another handshake.
33 So only messages with sequence numbers less than 10 in advance will be
34 buffered. (CVE-2009-1378)
35 [Robin Seggelmann, discovered by Daniel Mentz]
37 *) Records are buffered if they arrive with a future epoch to be
38 processed after finishing the corresponding handshake. There is
39 currently no limitation to this buffer allowing an attacker to perform
40 a DOS attack with sending records with future epochs until there is no
41 memory left. This patch adds the pqueue_size() function to detemine
42 the size of a buffer and limits the record buffer to 100 entries.
44 [Robin Seggelmann, discovered by Daniel Mentz]
46 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
47 parent structure is freed. (CVE-2009-1379)
50 *) Handle non-blocking I/O properly in SSL_shutdown() call.
51 [Darryl Miles <darryl-mailinglists@netbauds.net>]
54 [Ilya O. <vrghost@gmail.com>]
56 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
58 *) Don't set val to NULL when freeing up structures, it is freed up by
59 underlying code. If sizeof(void *) > sizeof(long) this can result in
60 zeroing past the valid field. (CVE-2009-0789)
61 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
63 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
64 checked correctly. This would allow some invalid signed attributes to
65 appear to verify correctly. (CVE-2009-0591)
66 [Ivan Nestlerode <inestlerode@us.ibm.com>]
68 *) Reject UniversalString and BMPString types with invalid lengths. This
69 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
70 a legal length. (CVE-2009-0590)
73 *) Set S/MIME signing as the default purpose rather than setting it
74 unconditionally. This allows applications to override it at the store
78 *) Permit restricted recursion of ASN1 strings. This is needed in practice
79 to handle some structures.
82 *) Improve efficiency of mem_gets: don't search whole buffer each time
84 [Jeremy Shapiro <jnshapir@us.ibm.com>]
86 *) New -hex option for openssl rand.
89 *) Print out UTF8String and NumericString when parsing ASN1.
92 *) Support NumericString type for name components.
95 *) Allow CC in the environment to override the automatically chosen
96 compiler. Note that nothing is done to ensure flags work with the
100 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
102 *) Properly check EVP_VerifyFinal() and similar return values
104 [Ben Laurie, Bodo Moeller, Google Security Team]
106 *) Enable TLS extensions by default.
109 *) Allow the CHIL engine to be loaded, whether the application is
110 multithreaded or not. (This does not release the developer from the
111 obligation to set up the dynamic locking callbacks.)
112 [Sander Temme <sander@temme.net>]
114 *) Use correct exit code if there is an error in dgst command.
115 [Steve Henson; problem pointed out by Roland Dirlewanger]
117 *) Tweak Configure so that you need to say "experimental-jpake" to enable
118 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
121 *) Add experimental JPAKE support, including demo authentication in
122 s_client and s_server.
125 *) Set the comparison function in v3_addr_canonize().
126 [Rob Austein <sra@hactrn.net>]
128 *) Add support for XMPP STARTTLS in s_client.
129 [Philip Paeps <philip@freebsd.org>]
131 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
132 to ensure that even with this option, only ciphersuites in the
133 server's preference list will be accepted. (Note that the option
134 applies only when resuming a session, so the earlier behavior was
135 just about the algorithm choice for symmetric cryptography.)
138 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
140 *) Fix NULL pointer dereference if a DTLS server recieved
141 ChangeCipherSpec as first record (CVE-2009-1386)
142 [Bodo Moeller, discovered by Alex Lam]
144 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
145 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
148 *) The fix in 0.9.8c that supposedly got rid of unsafe
149 double-checked locking was incomplete for RSA blinding,
150 addressing just one layer of what turns out to have been
151 doubly unsafe triple-checked locking.
153 So now fix this for real by retiring the MONT_HELPER macro
154 in crypto/rsa/rsa_eay.c.
156 [Bodo Moeller; problem pointed out by Marius Schilder]
158 *) Various precautionary measures:
160 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
162 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
163 (NB: This would require knowledge of the secret session ticket key
164 to exploit, in which case you'd be SOL either way.)
166 - Change bn_nist.c so that it will properly handle input BIGNUMs
167 outside the expected range.
169 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
172 [Neel Mehta, Bodo Moeller]
174 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
175 the load fails. Useful for distros.
176 [Ben Laurie and the FreeBSD team]
178 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
181 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
184 *) Expand ENGINE to support engine supplied SSL client certificate functions.
186 This work was sponsored by Logica.
189 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
190 keystores. Support for SSL/TLS client authentication too.
191 Not compiled unless enable-capieng specified to Configure.
193 This work was sponsored by Logica.
196 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
197 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
198 attribute creation routines such as certifcate requests and PKCS#12
202 Changes between 0.9.8g and 0.9.8h [28 May 2008]
204 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
205 handshake which could lead to a cilent crash as found using the
206 Codenomicon TLS test suite (CVE-2008-1672)
207 [Steve Henson, Mark Cox]
209 *) Fix double free in TLS server name extensions which could lead to
210 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
213 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
215 Clear the error queue to ensure that error entries left from
216 older function calls do not interfere with the correct operation.
217 [Lutz Jaenicke, Erik de Castro Lopo]
219 *) Remove root CA certificates of commercial CAs:
221 The OpenSSL project does not recommend any specific CA and does not
222 have any policy with respect to including or excluding any CA.
223 Therefore it does not make any sense to ship an arbitrary selection
224 of root CA certificates with the OpenSSL software.
227 *) RSA OAEP patches to fix two separate invalid memory reads.
228 The first one involves inputs when 'lzero' is greater than
229 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
230 before the beginning of from). The second one involves inputs where
231 the 'db' section contains nothing but zeroes (there is a one-byte
232 invalid read after the end of 'db').
233 [Ivan Nestlerode <inestlerode@us.ibm.com>]
235 *) Partial backport from 0.9.9-dev:
237 Introduce bn_mul_mont (dedicated Montgomery multiplication
238 procedure) as a candidate for BIGNUM assembler implementation.
239 While 0.9.9-dev uses assembler for various architectures, only
240 x86_64 is available by default here in the 0.9.8 branch, and
241 32-bit x86 is available through a compile-time setting.
243 To try the 32-bit x86 assembler implementation, use Configure
244 option "enable-montasm" (which exists only for this backport).
246 As "enable-montasm" for 32-bit x86 disclaims code stability
247 anyway, in this constellation we activate additional code
248 backported from 0.9.9-dev for further performance improvements,
249 namely BN_from_montgomery_word. (To enable this otherwise,
250 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
252 [Andy Polyakov (backport partially by Bodo Moeller)]
254 *) Add TLS session ticket callback. This allows an application to set
255 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
256 values. This is useful for key rollover for example where several key
257 sets may exist with different names.
260 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
261 This was broken until now in 0.9.8 releases, such that the only way
262 a registered ENGINE could be used (assuming it initialises
263 successfully on the host) was to explicitly set it as the default
264 for the relevant algorithms. This is in contradiction with 0.9.7
265 behaviour and the documentation. With this fix, when an ENGINE is
266 registered into a given algorithm's table of implementations, the
267 'uptodate' flag is reset so that auto-discovery will be used next
268 time a new context for that algorithm attempts to select an
270 [Ian Lister (tweaked by Geoff Thorpe)]
272 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
273 implemention in the following ways:
275 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
278 Lack of BER streaming support means one pass streaming processing is
279 only supported if data is detached: setting the streaming flag is
280 ignored for embedded content.
282 CMS support is disabled by default and must be explicitly enabled
283 with the enable-cms configuration option.
286 *) Update the GMP engine glue to do direct copies between BIGNUM and
287 mpz_t when openssl and GMP use the same limb size. Otherwise the
288 existing "conversion via a text string export" trick is still used.
289 [Paul Sheer <paulsheer@gmail.com>]
291 *) Zlib compression BIO. This is a filter BIO which compressed and
292 uncompresses any data passed through it.
295 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
296 RFC3394 compatible AES key wrapping.
299 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
300 sets string data without copying. X509_ALGOR_set0() and
301 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
302 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
303 from an X509_ATTRIBUTE structure optionally checking it occurs only
304 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
308 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
309 to get the expected BN_FLG_CONSTTIME behavior.
310 [Bodo Moeller (Google)]
314 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
315 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
316 - added some more tests to do_tests.pl
317 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
318 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
319 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
320 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
321 - various changes to netware.pl to enable gcc-cross builds on Win32
323 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
324 - various changes to fix missing prototype warnings
325 - fixed x86nasm.pl to create correct asm files for NASM COFF output
326 - added AES, WHIRLPOOL and CPUID assembler code to build files
327 - added missing AES assembler make rules to mk1mf.pl
328 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
329 [Guenter Knauf <eflash@gmx.net>]
331 *) Implement certificate status request TLS extension defined in RFC3546.
332 A client can set the appropriate parameters and receive the encoded
333 OCSP response via a callback. A server can query the supplied parameters
334 and set the encoded OCSP response in the callback. Add simplified examples
335 to s_client and s_server.
338 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
341 + Binary incompatibility of ssl_ctx_st structure
342 + DTLS interoperation with non-compliant servers
343 + Don't call get_session_cb() without proposed session
344 + Fix ia64 assembler code
345 [Andy Polyakov, Steve Henson]
347 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
349 *) DTLS Handshake overhaul. There were longstanding issues with
350 OpenSSL DTLS implementation, which were making it impossible for
351 RFC 4347 compliant client to communicate with OpenSSL server.
352 Unfortunately just fixing these incompatibilities would "cut off"
353 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
354 server keeps tolerating non RFC compliant syntax. The opposite is
355 not true, 0.9.8f client can not communicate with earlier server.
356 This update even addresses CVE-2007-4995.
359 *) Changes to avoid need for function casts in OpenSSL: some compilers
360 (gcc 4.2 and later) reject their use.
361 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
364 *) Add RFC4507 support to OpenSSL. This includes the corrections in
365 RFC4507bis. The encrypted ticket format is an encrypted encoded
366 SSL_SESSION structure, that way new session features are automatically
369 If a client application caches session in an SSL_SESSION structure
370 support is transparent because tickets are now stored in the encoded
373 The SSL_CTX structure automatically generates keys for ticket
374 protection in servers so again support should be possible
375 with no application modification.
377 If a client or server wishes to disable RFC4507 support then the option
378 SSL_OP_NO_TICKET can be set.
380 Add a TLS extension debugging callback to allow the contents of any client
381 or server extensions to be examined.
383 This work was sponsored by Google.
386 *) Add initial support for TLS extensions, specifically for the server_name
387 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
388 have new members for a host name. The SSL data structure has an
389 additional member SSL_CTX *initial_ctx so that new sessions can be
390 stored in that context to allow for session resumption, even after the
391 SSL has been switched to a new SSL_CTX in reaction to a client's
392 server_name extension.
394 New functions (subject to change):
397 SSL_get_servername_type()
400 New CTRL codes and macros (subject to change):
402 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
403 - SSL_CTX_set_tlsext_servername_callback()
404 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
405 - SSL_CTX_set_tlsext_servername_arg()
406 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
408 openssl s_client has a new '-servername ...' option.
410 openssl s_server has new options '-servername_host ...', '-cert2 ...',
411 '-key2 ...', '-servername_fatal' (subject to change). This allows
412 testing the HostName extension for a specific single host name ('-cert'
413 and '-key' remain fallbacks for handshakes without HostName
414 negotiation). If the unrecogninzed_name alert has to be sent, this by
415 default is a warning; it becomes fatal with the '-servername_fatal'
418 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
420 *) Add AES and SSE2 assembly language support to VC++ build.
423 *) Mitigate attack on final subtraction in Montgomery reduction.
426 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
427 (which previously caused an internal error).
430 *) Squeeze another 10% out of IGE mode when in != out.
433 *) AES IGE mode speedup.
434 [Dean Gaudet (Google)]
436 *) Add the Korean symmetric 128-bit cipher SEED (see
437 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
438 add SEED ciphersuites from RFC 4162:
440 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
441 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
442 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
443 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
445 To minimize changes between patchlevels in the OpenSSL 0.9.8
446 series, SEED remains excluded from compilation unless OpenSSL
447 is configured with 'enable-seed'.
450 *) Mitigate branch prediction attacks, which can be practical if a
451 single processor is shared, allowing a spy process to extract
452 information. For detailed background information, see
453 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
454 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
455 and Necessary Software Countermeasures"). The core of the change
456 are new versions BN_div_no_branch() and
457 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
458 respectively, which are slower, but avoid the security-relevant
459 conditional branches. These are automatically called by BN_div()
460 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
461 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
462 remove a conditional branch.
464 BN_FLG_CONSTTIME is the new name for the previous
465 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
466 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
467 in the exponent causes BN_mod_exp_mont() to use the alternative
468 implementation in BN_mod_exp_mont_consttime().) The old name
469 remains as a deprecated alias.
471 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
472 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
473 constant-time implementations for more than just exponentiation.
474 Here too the old name is kept as a deprecated alias.
476 BN_BLINDING_new() will now use BN_dup() for the modulus so that
477 the BN_BLINDING structure gets an independent copy of the
478 modulus. This means that the previous "BIGNUM *m" argument to
479 BN_BLINDING_new() and to BN_BLINDING_create_param() now
480 essentially becomes "const BIGNUM *m", although we can't actually
481 change this in the header file before 0.9.9. It allows
482 RSA_setup_blinding() to use BN_with_flags() on the modulus to
483 enable BN_FLG_CONSTTIME.
485 [Matthew D Wood (Intel Corp)]
487 *) In the SSL/TLS server implementation, be strict about session ID
488 context matching (which matters if an application uses a single
489 external cache for different purposes). Previously,
490 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
491 set. This did ensure strict client verification, but meant that,
492 with applications using a single external cache for quite
493 different requirements, clients could circumvent ciphersuite
494 restrictions for a given session ID context by starting a session
495 in a different context.
498 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
499 a ciphersuite string such as "DEFAULT:RSA" cannot enable
500 authentication-only ciphersuites.
503 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
504 not complete and could lead to a possible single byte overflow
505 (CVE-2007-5135) [Ben Laurie]
507 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
509 *) Since AES128 and AES256 (and similarly Camellia128 and
510 Camellia256) share a single mask bit in the logic of
511 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
512 kludge to work properly if AES128 is available and AES256 isn't
513 (or if Camellia128 is available and Camellia256 isn't).
516 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
517 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
518 When a point or a seed is encoded in a BIT STRING, we need to
519 prevent the removal of trailing zero bits to get the proper DER
520 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
521 of a NamedBitList, for which trailing 0 bits need to be removed.)
524 *) Have SSL/TLS server implementation tolerate "mismatched" record
525 protocol version while receiving ClientHello even if the
526 ClientHello is fragmented. (The server can't insist on the
527 particular protocol version it has chosen before the ServerHello
528 message has informed the client about his choice.)
531 *) Add RFC 3779 support.
532 [Rob Austein for ARIN, Ben Laurie]
534 *) Load error codes if they are not already present instead of using a
535 static variable. This allows them to be cleanly unloaded and reloaded.
536 Improve header file function name parsing.
539 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
540 or CAPABILITY handshake as required by RFCs.
543 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
545 *) Introduce limits to prevent malicious keys being able to
546 cause a denial of service. (CVE-2006-2940)
547 [Steve Henson, Bodo Moeller]
549 *) Fix ASN.1 parsing of certain invalid structures that can result
550 in a denial of service. (CVE-2006-2937) [Steve Henson]
552 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
553 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
555 *) Fix SSL client code which could crash if connecting to a
556 malicious SSLv2 server. (CVE-2006-4343)
557 [Tavis Ormandy and Will Drewry, Google Security Team]
559 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
560 match only those. Before that, "AES256-SHA" would be interpreted
561 as a pattern and match "AES128-SHA" too (since AES128-SHA got
562 the same strength classification in 0.9.7h) as we currently only
563 have a single AES bit in the ciphersuite description bitmap.
564 That change, however, also applied to ciphersuite strings such as
565 "RC4-MD5" that intentionally matched multiple ciphersuites --
566 namely, SSL 2.0 ciphersuites in addition to the more common ones
567 from SSL 3.0/TLS 1.0.
569 So we change the selection algorithm again: Naming an explicit
570 ciphersuite selects this one ciphersuite, and any other similar
571 ciphersuite (same bitmap) from *other* protocol versions.
572 Thus, "RC4-MD5" again will properly select both the SSL 2.0
573 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
575 Since SSL 2.0 does not have any ciphersuites for which the
576 128/256 bit distinction would be relevant, this works for now.
577 The proper fix will be to use different bits for AES128 and
578 AES256, which would have avoided the problems from the beginning;
579 however, bits are scarce, so we can only do this in a new release
580 (not just a patchlevel) when we can change the SSL_CIPHER
581 definition to split the single 'unsigned long mask' bitmap into
582 multiple values to extend the available space.
586 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
588 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
589 (CVE-2006-4339) [Ben Laurie and Google Security Team]
591 *) Add AES IGE and biIGE modes.
594 *) Change the Unix randomness entropy gathering to use poll() when
595 possible instead of select(), since the latter has some
596 undesirable limitations.
597 [Darryl Miles via Richard Levitte and Bodo Moeller]
599 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
600 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
601 cannot be implicitly activated as part of, e.g., the "AES" alias.
602 However, please upgrade to OpenSSL 0.9.9[-dev] for
603 non-experimental use of the ECC ciphersuites to get TLS extension
604 support, which is required for curve and point format negotiation
605 to avoid potential handshake problems.
608 *) Disable rogue ciphersuites:
610 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
611 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
612 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
614 The latter two were purportedly from
615 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
618 Also deactivate the remaining ciphersuites from
619 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
620 unofficial, and the ID has long expired.
623 *) Fix RSA blinding Heisenbug (problems sometimes occured on
624 dual-core machines) and other potential thread-safety issues.
627 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
628 versions), which is now available for royalty-free use
629 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
630 Also, add Camellia TLS ciphersuites from RFC 4132.
632 To minimize changes between patchlevels in the OpenSSL 0.9.8
633 series, Camellia remains excluded from compilation unless OpenSSL
634 is configured with 'enable-camellia'.
637 *) Disable the padding bug check when compression is in use. The padding
638 bug check assumes the first packet is of even length, this is not
639 necessarily true if compresssion is enabled and can result in false
640 positives causing handshake failure. The actual bug test is ancient
641 code so it is hoped that implementations will either have fixed it by
642 now or any which still have the bug do not support compression.
645 Changes between 0.9.8a and 0.9.8b [04 May 2006]
647 *) When applying a cipher rule check to see if string match is an explicit
648 cipher suite and only match that one cipher suite if it is.
651 *) Link in manifests for VC++ if needed.
652 [Austin Ziegler <halostatue@gmail.com>]
654 *) Update support for ECC-based TLS ciphersuites according to
655 draft-ietf-tls-ecc-12.txt with proposed changes (but without
656 TLS extensions, which are supported starting with the 0.9.9
657 branch, not in the OpenSSL 0.9.8 branch).
660 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
661 opaque EVP_CIPHER_CTX handling.
664 *) Fixes and enhancements to zlib compression code. We now only use
665 "zlib1.dll" and use the default __cdecl calling convention on Win32
666 to conform with the standards mentioned here:
667 http://www.zlib.net/DLL_FAQ.txt
668 Static zlib linking now works on Windows and the new --with-zlib-include
669 --with-zlib-lib options to Configure can be used to supply the location
670 of the headers and library. Gracefully handle case where zlib library
674 *) Several fixes and enhancements to the OID generation code. The old code
675 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
676 handle numbers larger than ULONG_MAX, truncated printing and had a
677 non standard OBJ_obj2txt() behaviour.
680 *) Add support for building of engines under engine/ as shared libraries
681 under VC++ build system.
684 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
685 Hopefully, we will not see any false combination of paths any more.
688 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
690 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
691 (part of SSL_OP_ALL). This option used to disable the
692 countermeasure against man-in-the-middle protocol-version
693 rollback in the SSL 2.0 server implementation, which is a bad
694 idea. (CVE-2005-2969)
696 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
697 for Information Security, National Institute of Advanced Industrial
698 Science and Technology [AIST], Japan)]
700 *) Add two function to clear and return the verify parameter flags.
703 *) Keep cipherlists sorted in the source instead of sorting them at
704 runtime, thus removing the need for a lock.
707 *) Avoid some small subgroup attacks in Diffie-Hellman.
708 [Nick Mathewson and Ben Laurie]
710 *) Add functions for well-known primes.
713 *) Extended Windows CE support.
714 [Satoshi Nakamura and Andy Polyakov]
716 *) Initialize SSL_METHOD structures at compile time instead of during
717 runtime, thus removing the need for a lock.
720 *) Make PKCS7_decrypt() work even if no certificate is supplied by
721 attempting to decrypt each encrypted key in turn. Add support to
725 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
727 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
730 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
733 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
734 key into the same file any more.
737 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
740 *) Add -utf8 command line and config file option to 'ca'.
741 [Stefan <stf@udoma.org]
743 *) Removed the macro des_crypt(), as it seems to conflict with some
744 libraries. Use DES_crypt().
747 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
748 involves renaming the source and generated shared-libs for
749 both. The engines will accept the corrected or legacy ids
750 ('ncipher' and '4758_cca' respectively) when binding. NB,
751 this only applies when building 'shared'.
752 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
754 *) Add attribute functions to EVP_PKEY structure. Modify
755 PKCS12_create() to recognize a CSP name attribute and
756 use it. Make -CSP option work again in pkcs12 utility.
759 *) Add new functionality to the bn blinding code:
760 - automatic re-creation of the BN_BLINDING parameters after
761 a fixed number of uses (currently 32)
762 - add new function for parameter creation
763 - introduce flags to control the update behaviour of the
764 BN_BLINDING parameters
765 - hide BN_BLINDING structure
766 Add a second BN_BLINDING slot to the RSA structure to improve
767 performance when a single RSA object is shared among several
771 *) Add support for DTLS.
772 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
774 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
775 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
778 *) Remove buggy and incompletet DH cert support from
779 ssl/ssl_rsa.c and ssl/s3_both.c
782 *) Use SHA-1 instead of MD5 as the default digest algorithm for
783 the apps/openssl applications.
786 *) Compile clean with "-Wall -Wmissing-prototypes
787 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
788 DEBUG_SAFESTACK must also be set.
791 *) Change ./Configure so that certain algorithms can be disabled by default.
792 The new counterpiece to "no-xxx" is "enable-xxx".
794 The patented RC5 and MDC2 algorithms will now be disabled unless
795 "enable-rc5" and "enable-mdc2", respectively, are specified.
797 (IDEA remains enabled despite being patented. This is because IDEA
798 is frequently required for interoperability, and there is no license
799 fee for non-commercial use. As before, "no-idea" can be used to
800 avoid this algorithm.)
804 *) Add processing of proxy certificates (see RFC 3820). This work was
805 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
806 EGEE (Enabling Grids for E-science in Europe).
809 *) RC4 performance overhaul on modern architectures/implementations, such
810 as Intel P4, IA-64 and AMD64.
813 *) New utility extract-section.pl. This can be used specify an alternative
814 section number in a pod file instead of having to treat each file as
815 a separate case in Makefile. This can be done by adding two lines to the
818 =for comment openssl_section:XXX
820 The blank line is mandatory.
824 *) New arguments -certform, -keyform and -pass for s_client and s_server
825 to allow alternative format key and certificate files and passphrase
829 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
830 update associated structures and add various utility functions.
832 Add new policy related verify parameters, include policy checking in
833 standard verify code. Enhance 'smime' application with extra parameters
834 to support policy checking and print out.
837 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
838 Nehemiah processors. These extensions support AES encryption in hardware
839 as well as RNG (though RNG support is currently disabled).
840 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
842 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
845 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
846 [Andy Polyakov and a number of other people]
848 *) Improved PowerPC platform support. Most notably BIGNUM assembler
849 implementation contributed by IBM.
850 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
852 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
853 exponent rather than 'unsigned long'. There is a corresponding change to
854 the new 'rsa_keygen' element of the RSA_METHOD structure.
855 [Jelte Jansen, Geoff Thorpe]
857 *) Functionality for creating the initial serial number file is now
858 moved from CA.pl to the 'ca' utility with a new option -create_serial.
860 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
861 number file to 1, which is bound to cause problems. To avoid
862 the problems while respecting compatibility between different 0.9.7
863 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
864 CA.pl for serial number initialization. With the new release 0.9.8,
865 we can fix the problem directly in the 'ca' utility.)
868 *) Reduced header interdepencies by declaring more opaque objects in
869 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
870 give fewer recursive includes, which could break lazy source code - so
871 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
872 developers should define this symbol when building and using openssl to
873 ensure they track the recommended behaviour, interfaces, [etc], but
874 backwards-compatible behaviour prevails when this isn't defined.
877 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
880 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
881 This will generate a random key of the appropriate length based on the
882 cipher context. The EVP_CIPHER can provide its own random key generation
883 routine to support keys of a specific form. This is used in the des and
884 3des routines to generate a key of the correct parity. Update S/MIME
885 code to use new functions and hence generate correct parity DES keys.
886 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
887 valid (weak or incorrect parity).
890 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
891 as looking them up. This is useful when the verified structure may contain
892 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
893 present unless the new PKCS7_NO_CRL flag is asserted.
896 *) Extend ASN1 oid configuration module. It now additionally accepts the
899 shortName = some long name, 1.2.3.4
902 *) Reimplemented the BN_CTX implementation. There is now no more static
903 limitation on the number of variables it can handle nor the depth of the
904 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
905 information can now expand as required, and rather than having a single
906 static array of bignums, BN_CTX now uses a linked-list of such arrays
907 allowing it to expand on demand whilst maintaining the usefulness of
911 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
912 to allow all RSA operations to function using a single BN_CTX.
915 *) Preliminary support for certificate policy evaluation and checking. This
916 is initially intended to pass the tests outlined in "Conformance Testing
917 of Relying Party Client Certificate Path Processing Logic" v1.07.
920 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
921 remained unused and not that useful. A variety of other little bignum
922 tweaks and fixes have also been made continuing on from the audit (see
926 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
927 associated ASN1, EVP and SSL functions and old ASN1 macros.
930 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
931 and this should never fail. So the return value from the use of
932 BN_set_word() (which can fail due to needless expansion) is now deprecated;
933 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
936 *) BN_CTX_get() should return zero-valued bignums, providing the same
937 initialised value as BN_new().
938 [Geoff Thorpe, suggested by Ulf Möller]
940 *) Support for inhibitAnyPolicy certificate extension.
943 *) An audit of the BIGNUM code is underway, for which debugging code is
944 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
945 is considered valid when processing BIGNUMs, and causes execution to
946 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
947 further steps are taken to deliberately pollute unused data in BIGNUM
948 structures to try and expose faulty code further on. For now, openssl will
949 (in its default mode of operation) continue to tolerate the inconsistent
950 forms that it has tolerated in the past, but authors and packagers should
951 consider trying openssl and their own applications when compiled with
952 these debugging symbols defined. It will help highlight potential bugs in
953 their own code, and will improve the test coverage for OpenSSL itself. At
954 some point, these tighter rules will become openssl's default to improve
955 maintainability, though the assert()s and other overheads will remain only
956 in debugging configurations. See bn.h for more details.
957 [Geoff Thorpe, Nils Larsch, Ulf Möller]
959 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
960 that can only be obtained through BN_CTX_new() (which implicitly
961 initialises it). The presence of this function only made it possible
962 to overwrite an existing structure (and cause memory leaks).
965 *) Because of the callback-based approach for implementing LHASH as a
966 template type, lh_insert() adds opaque objects to hash-tables and
967 lh_doall() or lh_doall_arg() are typically used with a destructor callback
968 to clean up those corresponding objects before destroying the hash table
969 (and losing the object pointers). So some over-zealous constifications in
970 LHASH have been relaxed so that lh_insert() does not take (nor store) the
971 objects as "const" and the lh_doall[_arg] callback wrappers are not
972 prototyped to have "const" restrictions on the object pointers they are
973 given (and so aren't required to cast them away any more).
976 *) The tmdiff.h API was so ugly and minimal that our own timing utility
977 (speed) prefers to use its own implementation. The two implementations
978 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
979 its object type properly exposed (MS_TM) instead of casting to/from "char
980 *". This may still change yet if someone realises MS_TM and "ms_time_***"
981 aren't necessarily the greatest nomenclatures - but this is what was used
982 internally to the implementation so I've used that for now.
985 *) Ensure that deprecated functions do not get compiled when
986 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
987 the self-tests were still using deprecated key-generation functions so
988 these have been updated also.
991 *) Reorganise PKCS#7 code to separate the digest location functionality
992 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
993 New function PKCS7_set_digest() to set the digest type for PKCS#7
994 digestedData type. Add additional code to correctly generate the
995 digestedData type and add support for this type in PKCS7 initialization
999 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1000 structure of type "other".
1003 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1004 sure the loop does correctly stop and breaking ("division by zero")
1005 modulus operations are not performed. The (pre-generated) prime
1006 table crypto/bn/bn_prime.h was already correct, but it could not be
1007 re-generated on some platforms because of the "division by zero"
1008 situation in the script.
1009 [Ralf S. Engelschall]
1011 *) Update support for ECC-based TLS ciphersuites according to
1012 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1013 SHA-1 now is only used for "small" curves (where the
1014 representation of a field element takes up to 24 bytes); for
1015 larger curves, the field element resulting from ECDH is directly
1016 used as premaster secret.
1017 [Douglas Stebila (Sun Microsystems Laboratories)]
1019 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1020 curve secp160r1 to the tests.
1021 [Douglas Stebila (Sun Microsystems Laboratories)]
1023 *) Add the possibility to load symbols globally with DSO.
1024 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1026 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1027 control of the error stack.
1030 *) Add support for STORE in ENGINE.
1033 *) Add the STORE type. The intention is to provide a common interface
1034 to certificate and key stores, be they simple file-based stores, or
1035 HSM-type store, or LDAP stores, or...
1036 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1039 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1040 pass a list of arguments to any function as well as provide a way
1041 for a function to pass data back to the caller.
1044 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1045 works like BUF_strdup() but can be used to duplicate a portion of
1046 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1050 *) Add the function sk_find_ex() which works like sk_find(), but will
1051 return an index to an element even if an exact match couldn't be
1052 found. The index is guaranteed to point at the element where the
1053 searched-for key would be inserted to preserve sorting order.
1056 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1057 takes an extra flags argument for optional functionality. Currently,
1058 the following flags are defined:
1060 OBJ_BSEARCH_VALUE_ON_NOMATCH
1061 This one gets OBJ_bsearch_ex() to return a pointer to the first
1062 element where the comparing function returns a negative or zero
1065 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1066 This one gets OBJ_bsearch_ex() to return a pointer to the first
1067 element where the comparing function returns zero. This is useful
1068 if there are more than one element where the comparing function
1072 *) Make it possible to create self-signed certificates with 'openssl ca'
1073 in such a way that the self-signed certificate becomes part of the
1074 CA database and uses the same mechanisms for serial number generation
1075 as all other certificate signing. The new flag '-selfsign' enables
1076 this functionality. Adapt CA.sh and CA.pl.in.
1079 *) Add functionality to check the public key of a certificate request
1080 against a given private. This is useful to check that a certificate
1081 request can be signed by that key (self-signing).
1084 *) Make it possible to have multiple active certificates with the same
1085 subject in the CA index file. This is done only if the keyword
1086 'unique_subject' is set to 'no' in the main CA section (default
1087 if 'CA_default') of the configuration file. The value is saved
1088 with the database itself in a separate index attribute file,
1089 named like the index file with '.attr' appended to the name.
1092 *) Generate muti valued AVAs using '+' notation in config files for
1096 *) Support for nameConstraints certificate extension.
1099 *) Support for policyConstraints certificate extension.
1102 *) Support for policyMappings certificate extension.
1105 *) Make sure the default DSA_METHOD implementation only uses its
1106 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1107 and change its own handlers to be NULL so as to remove unnecessary
1108 indirection. This lets alternative implementations fallback to the
1109 default implementation more easily.
1112 *) Support for directoryName in GeneralName related extensions
1116 *) Make it possible to link applications using Makefile.shared.
1117 Make that possible even when linking against static libraries!
1120 *) Support for single pass processing for S/MIME signing. This now
1121 means that S/MIME signing can be done from a pipe, in addition
1122 cleartext signing (multipart/signed type) is effectively streaming
1123 and the signed data does not need to be all held in memory.
1125 This is done with a new flag PKCS7_STREAM. When this flag is set
1126 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1127 is done after the data is output (and digests calculated) in
1128 SMIME_write_PKCS7().
1131 *) Add full support for -rpath/-R, both in shared libraries and
1132 applications, at least on the platforms where it's known how
1136 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1137 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1138 will now compute a table of multiples of the generator that
1139 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1140 faster (notably in the case of a single point multiplication,
1141 scalar * generator).
1142 [Nils Larsch, Bodo Moeller]
1144 *) IPv6 support for certificate extensions. The various extensions
1145 which use the IP:a.b.c.d can now take IPv6 addresses using the
1146 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1150 *) Added an ENGINE that implements RSA by performing private key
1151 exponentiations with the GMP library. The conversions to and from
1152 GMP's mpz_t format aren't optimised nor are any montgomery forms
1153 cached, and on x86 it appears OpenSSL's own performance has caught up.
1154 However there are likely to be other architectures where GMP could
1155 provide a boost. This ENGINE is not built in by default, but it can be
1156 specified at Configure time and should be accompanied by the necessary
1157 linker additions, eg;
1158 ./config -DOPENSSL_USE_GMP -lgmp
1161 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1162 testing availability of engines with "-t" - the old behaviour is
1163 produced by increasing the feature's verbosity with "-tt".
1166 *) ECDSA routines: under certain error conditions uninitialized BN objects
1167 could be freed. Solution: make sure initialization is performed early
1168 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1172 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1173 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1174 software implementations. For DSA and DH, parameter generation can
1175 also be overriden by providing the appropriate method callbacks.
1178 *) Change the "progress" mechanism used in key-generation and
1179 primality testing to functions that take a new BN_GENCB pointer in
1180 place of callback/argument pairs. The new API functions have "_ex"
1181 postfixes and the older functions are reimplemented as wrappers for
1182 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1183 declarations of the old functions to help (graceful) attempts to
1184 migrate to the new functions. Also, the new key-generation API
1185 functions operate on a caller-supplied key-structure and return
1186 success/failure rather than returning a key or NULL - this is to
1187 help make "keygen" another member function of RSA_METHOD etc.
1189 Example for using the new callback interface:
1191 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1195 BN_GENCB_set(&my_cb, my_callback, my_arg);
1197 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1198 /* For the meaning of a, b in calls to my_callback(), see the
1199 * documentation of the function that calls the callback.
1200 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1201 * my_callback should return 1 if it wants BN_is_prime_ex()
1202 * to continue, or 0 to stop.
1207 *) Change the ZLIB compression method to be stateful, and make it
1208 available to TLS with the number defined in
1209 draft-ietf-tls-compression-04.txt.
1212 *) Add the ASN.1 structures and functions for CertificatePair, which
1213 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1215 CertificatePair ::= SEQUENCE {
1216 forward [0] Certificate OPTIONAL,
1217 reverse [1] Certificate OPTIONAL,
1218 -- at least one of the pair shall be present -- }
1220 Also implement the PEM functions to read and write certificate
1221 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1223 This needed to be defined, mostly for the sake of the LDAP
1224 attribute crossCertificatePair, but may prove useful elsewhere as
1228 *) Make it possible to inhibit symlinking of shared libraries in
1229 Makefile.shared, for Cygwin's sake.
1232 *) Extend the BIGNUM API by creating a function
1233 void BN_set_negative(BIGNUM *a, int neg);
1234 and a macro that behave like
1235 int BN_is_negative(const BIGNUM *a);
1237 to avoid the need to access 'a->neg' directly in applications.
1240 *) Implement fast modular reduction for pseudo-Mersenne primes
1241 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1242 EC_GROUP_new_curve_GFp() will now automatically use this
1244 [Nils Larsch <nla@trustcenter.de>]
1246 *) Add new lock type (CRYPTO_LOCK_BN).
1249 *) Change the ENGINE framework to automatically load engines
1250 dynamically from specific directories unless they could be
1251 found to already be built in or loaded. Move all the
1252 current engines except for the cryptodev one to a new
1254 The engines in engines/ are built as shared libraries if
1255 the "shared" options was given to ./Configure or ./config.
1256 Otherwise, they are inserted in libcrypto.a.
1257 /usr/local/ssl/engines is the default directory for dynamic
1258 engines, but that can be overriden at configure time through
1259 the usual use of --prefix and/or --openssldir, and at run
1260 time with the environment variable OPENSSL_ENGINES.
1261 [Geoff Thorpe and Richard Levitte]
1263 *) Add Makefile.shared, a helper makefile to build shared
1264 libraries. Addapt Makefile.org.
1267 *) Add version info to Win32 DLLs.
1268 [Peter 'Luna' Runestig" <peter@runestig.com>]
1270 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1271 can be added using this API to created arbitrary PKCS#12
1272 files while avoiding the low level API.
1274 New options to PKCS12_create(), key or cert can be NULL and
1275 will then be omitted from the output file. The encryption
1276 algorithm NIDs can be set to -1 for no encryption, the mac
1277 iteration count can be set to 0 to omit the mac.
1279 Enhance pkcs12 utility by making the -nokeys and -nocerts
1280 options work when creating a PKCS#12 file. New option -nomac
1281 to omit the mac, NONE can be set for an encryption algorithm.
1282 New code is modified to use the enhanced PKCS12_create()
1283 instead of the low level API.
1286 *) Extend ASN1 encoder to support indefinite length constructed
1287 encoding. This can output sequences tags and octet strings in
1288 this form. Modify pk7_asn1.c to support indefinite length
1289 encoding. This is experimental and needs additional code to
1290 be useful, such as an ASN1 bio and some enhanced streaming
1293 Extend template encode functionality so that tagging is passed
1294 down to the template encoder.
1297 *) Let 'openssl req' fail if an argument to '-newkey' is not
1298 recognized instead of using RSA as a default.
1301 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1302 As these are not official, they are not included in "ALL";
1303 the "ECCdraft" ciphersuite group alias can be used to select them.
1304 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1306 *) Add ECDH engine support.
1307 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1309 *) Add ECDH in new directory crypto/ecdh/.
1310 [Douglas Stebila (Sun Microsystems Laboratories)]
1312 *) Let BN_rand_range() abort with an error after 100 iterations
1313 without success (which indicates a broken PRNG).
1316 *) Change BN_mod_sqrt() so that it verifies that the input value
1317 is really the square of the return value. (Previously,
1318 BN_mod_sqrt would show GIGO behaviour.)
1321 *) Add named elliptic curves over binary fields from X9.62, SECG,
1322 and WAP/WTLS; add OIDs that were still missing.
1324 [Sheueling Chang Shantz and Douglas Stebila
1325 (Sun Microsystems Laboratories)]
1327 *) Extend the EC library for elliptic curves over binary fields
1328 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1331 EC_GF2m_simple_method
1335 EC_GROUP_new_curve_GF2m
1336 EC_GROUP_set_curve_GF2m
1337 EC_GROUP_get_curve_GF2m
1338 EC_POINT_set_affine_coordinates_GF2m
1339 EC_POINT_get_affine_coordinates_GF2m
1340 EC_POINT_set_compressed_coordinates_GF2m
1342 Point compression for binary fields is disabled by default for
1343 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1346 As binary polynomials are represented as BIGNUMs, various members
1347 of the EC_GROUP and EC_POINT data structures can be shared
1348 between the implementations for prime fields and binary fields;
1349 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1350 are essentially identical to their ..._GFp counterparts.
1351 (For simplicity, the '..._GFp' prefix has been dropped from
1352 various internal method names.)
1354 An internal 'field_div' method (similar to 'field_mul' and
1355 'field_sqr') has been added; this is used only for binary fields.
1357 [Sheueling Chang Shantz and Douglas Stebila
1358 (Sun Microsystems Laboratories)]
1360 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1361 through methods ('mul', 'precompute_mult').
1363 The generic implementations (now internally called 'ec_wNAF_mul'
1364 and 'ec_wNAF_precomputed_mult') remain the default if these
1365 methods are undefined.
1367 [Sheueling Chang Shantz and Douglas Stebila
1368 (Sun Microsystems Laboratories)]
1370 *) New function EC_GROUP_get_degree, which is defined through
1371 EC_METHOD. For curves over prime fields, this returns the bit
1372 length of the modulus.
1374 [Sheueling Chang Shantz and Douglas Stebila
1375 (Sun Microsystems Laboratories)]
1377 *) New functions EC_GROUP_dup, EC_POINT_dup.
1378 (These simply call ..._new and ..._copy).
1380 [Sheueling Chang Shantz and Douglas Stebila
1381 (Sun Microsystems Laboratories)]
1383 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1384 Polynomials are represented as BIGNUMs (where the sign bit is not
1385 used) in the following functions [macros]:
1388 BN_GF2m_sub [= BN_GF2m_add]
1389 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1390 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1391 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1393 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1394 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1395 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1396 BN_GF2m_cmp [= BN_ucmp]
1398 (Note that only the 'mod' functions are actually for fields GF(2^m).
1399 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1401 For some functions, an the irreducible polynomial defining a
1402 field can be given as an 'unsigned int[]' with strictly
1403 decreasing elements giving the indices of those bits that are set;
1404 i.e., p[] represents the polynomial
1405 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1407 p[0] > p[1] > ... > p[k] = 0.
1408 This applies to the following functions:
1413 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1414 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1416 BN_GF2m_mod_sqrt_arr
1417 BN_GF2m_mod_solve_quad_arr
1421 Conversion can be performed by the following functions:
1426 bntest.c has additional tests for binary polynomial arithmetic.
1428 Two implementations for BN_GF2m_mod_div() are available.
1429 The default algorithm simply uses BN_GF2m_mod_inv() and
1430 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1431 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1432 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1434 [Sheueling Chang Shantz and Douglas Stebila
1435 (Sun Microsystems Laboratories)]
1437 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1438 functionality is disabled at compile-time.
1439 [Douglas Stebila <douglas.stebila@sun.com>]
1441 *) Change default behaviour of 'openssl asn1parse' so that more
1442 information is visible when viewing, e.g., a certificate:
1444 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1445 mode the content of non-printable OCTET STRINGs is output in a
1446 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1447 avoid the appearance of a printable string.
1448 [Nils Larsch <nla@trustcenter.de>]
1450 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1452 EC_GROUP_set_asn1_flag()
1453 EC_GROUP_get_asn1_flag()
1454 EC_GROUP_set_point_conversion_form()
1455 EC_GROUP_get_point_conversion_form()
1456 These control ASN1 encoding details:
1457 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1458 has been set to OPENSSL_EC_NAMED_CURVE.
1459 - Points are encoded in uncompressed form by default; options for
1460 asn1_for are as for point2oct, namely
1461 POINT_CONVERSION_COMPRESSED
1462 POINT_CONVERSION_UNCOMPRESSED
1463 POINT_CONVERSION_HYBRID
1465 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1468 EC_GROUP_get0_seed()
1469 EC_GROUP_get_seed_len()
1470 This is used only for ASN1 purposes (so far).
1471 [Nils Larsch <nla@trustcenter.de>]
1473 *) Add 'field_type' member to EC_METHOD, which holds the NID
1474 of the appropriate field type OID. The new function
1475 EC_METHOD_get_field_type() returns this value.
1476 [Nils Larsch <nla@trustcenter.de>]
1481 EC_POINT_point2hex()
1482 EC_POINT_hex2point()
1483 providing useful interfaces to EC_POINT_point2oct() and
1484 EC_POINT_oct2point().
1485 [Nils Larsch <nla@trustcenter.de>]
1487 *) Change internals of the EC library so that the functions
1488 EC_GROUP_set_generator()
1489 EC_GROUP_get_generator()
1490 EC_GROUP_get_order()
1491 EC_GROUP_get_cofactor()
1492 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1493 to methods, which would lead to unnecessary code duplication when
1494 adding different types of curves.
1495 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1497 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1498 arithmetic, and such that modified wNAFs are generated
1499 (which avoid length expansion in many cases).
1502 *) Add a function EC_GROUP_check_discriminant() (defined via
1503 EC_METHOD) that verifies that the curve discriminant is non-zero.
1505 Add a function EC_GROUP_check() that makes some sanity tests
1506 on a EC_GROUP, its generator and order. This includes
1507 EC_GROUP_check_discriminant().
1508 [Nils Larsch <nla@trustcenter.de>]
1510 *) Add ECDSA in new directory crypto/ecdsa/.
1512 Add applications 'openssl ecparam' and 'openssl ecdsa'
1513 (these are based on 'openssl dsaparam' and 'openssl dsa').
1515 ECDSA support is also included in various other files across the
1516 library. Most notably,
1517 - 'openssl req' now has a '-newkey ecdsa:file' option;
1518 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1519 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1520 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1521 them suitable for ECDSA where domain parameters must be
1522 extracted before the specific public key;
1523 - ECDSA engine support has been added.
1524 [Nils Larsch <nla@trustcenter.de>]
1526 *) Include some named elliptic curves, and add OIDs from X9.62,
1527 SECG, and WAP/WTLS. Each curve can be obtained from the new
1529 EC_GROUP_new_by_curve_name(),
1530 and the list of available named curves can be obtained with
1531 EC_get_builtin_curves().
1532 Also add a 'curve_name' member to EC_GROUP objects, which can be
1534 EC_GROUP_set_curve_name()
1535 EC_GROUP_get_curve_name()
1536 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1538 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1539 was actually never needed) and in BN_mul(). The removal in BN_mul()
1540 required a small change in bn_mul_part_recursive() and the addition
1541 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1542 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1543 bn_sub_words() and bn_add_words() except they take arrays with
1547 Changes between 0.9.7m and 0.9.7n [xx XXX xxxx]
1549 *) In the SSL/TLS server implementation, be strict about session ID
1550 context matching (which matters if an application uses a single
1551 external cache for different purposes). Previously,
1552 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1553 set. This did ensure strict client verification, but meant that,
1554 with applications using a single external cache for quite
1555 different requirements, clients could circumvent ciphersuite
1556 restrictions for a given session ID context by starting a session
1557 in a different context.
1560 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1562 *) Cleanse PEM buffers before freeing them since they may contain
1564 [Benjamin Bennett <ben@psc.edu>]
1566 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1567 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1568 authentication-only ciphersuites.
1571 *) Since AES128 and AES256 share a single mask bit in the logic of
1572 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1573 kludge to work properly if AES128 is available and AES256 isn't.
1576 *) Expand security boundary to match 1.1.1 module.
1579 *) Remove redundant features: hash file source, editing of test vectors
1580 modify fipsld to use external fips_premain.c signature.
1583 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1584 run algorithm test programs.
1587 *) Make algorithm test programs more tolerant of whitespace.
1590 *) Have SSL/TLS server implementation tolerate "mismatched" record
1591 protocol version while receiving ClientHello even if the
1592 ClientHello is fragmented. (The server can't insist on the
1593 particular protocol version it has chosen before the ServerHello
1594 message has informed the client about his choice.)
1597 *) Load error codes if they are not already present instead of using a
1598 static variable. This allows them to be cleanly unloaded and reloaded.
1601 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1603 *) Introduce limits to prevent malicious keys being able to
1604 cause a denial of service. (CVE-2006-2940)
1605 [Steve Henson, Bodo Moeller]
1607 *) Fix ASN.1 parsing of certain invalid structures that can result
1608 in a denial of service. (CVE-2006-2937) [Steve Henson]
1610 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1611 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1613 *) Fix SSL client code which could crash if connecting to a
1614 malicious SSLv2 server. (CVE-2006-4343)
1615 [Tavis Ormandy and Will Drewry, Google Security Team]
1617 *) Change ciphersuite string processing so that an explicit
1618 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1619 will no longer include "AES128-SHA"), and any other similar
1620 ciphersuite (same bitmap) from *other* protocol versions (so that
1621 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1622 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1623 changes from 0.9.8b and 0.9.8d.
1626 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1628 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1629 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1631 *) Change the Unix randomness entropy gathering to use poll() when
1632 possible instead of select(), since the latter has some
1633 undesirable limitations.
1634 [Darryl Miles via Richard Levitte and Bodo Moeller]
1636 *) Disable rogue ciphersuites:
1638 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1639 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1640 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1642 The latter two were purportedly from
1643 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1646 Also deactive the remaining ciphersuites from
1647 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1648 unofficial, and the ID has long expired.
1651 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1652 dual-core machines) and other potential thread-safety issues.
1655 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1657 *) Adapt fipsld and the build system to link against the validated FIPS
1658 module in FIPS mode.
1661 *) Fixes for VC++ 2005 build under Windows.
1664 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1665 from a Windows bash shell such as MSYS. It is autodetected from the
1666 "config" script when run from a VC++ environment. Modify standard VC++
1667 build to use fipscanister.o from the GNU make build.
1670 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1672 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1673 The value now differs depending on if you build for FIPS or not.
1674 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1675 safely run with a non-FIPSed libcrypto, as it may crash because of
1676 the difference induced by this change.
1679 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1681 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1682 (part of SSL_OP_ALL). This option used to disable the
1683 countermeasure against man-in-the-middle protocol-version
1684 rollback in the SSL 2.0 server implementation, which is a bad
1685 idea. (CVE-2005-2969)
1687 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1688 for Information Security, National Institute of Advanced Industrial
1689 Science and Technology [AIST], Japan)]
1691 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1692 mainly for FIPS compliance and not fully integrated at this stage.
1695 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1696 the exponentiation using a fixed-length exponent. (Otherwise,
1697 the information leaked through timing could expose the secret key
1698 after many signatures; cf. Bleichenbacher's attack on DSA with
1702 *) Make a new fixed-window mod_exp implementation the default for
1703 RSA, DSA, and DH private-key operations so that the sequence of
1704 squares and multiplies and the memory access pattern are
1705 independent of the particular secret key. This will mitigate
1706 cache-timing and potential related attacks.
1708 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1709 and this is automatically used by BN_mod_exp_mont() if the new flag
1710 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1711 will use this BN flag for private exponents unless the flag
1712 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1713 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1715 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1717 *) Change the client implementation for SSLv23_method() and
1718 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1719 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1720 (Previously, the SSL 2.0 backwards compatible Client Hello
1721 message format would be used even with SSL_OP_NO_SSLv2.)
1724 *) Add support for smime-type MIME parameter in S/MIME messages which some
1728 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1729 a threadsafe manner. Modify rsa code to use new function and add calls
1730 to dsa and dh code (which had race conditions before).
1733 *) Include the fixed error library code in the C error file definitions
1734 instead of fixing them up at runtime. This keeps the error code
1735 structures constant.
1738 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1740 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1743 *) Fixes for newer kerberos headers. NB: the casts are needed because
1744 the 'length' field is signed on one version and unsigned on another
1745 with no (?) obvious way to tell the difference, without these VC++
1746 complains. Also the "definition" of FAR (blank) is no longer included
1747 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1748 some needed definitions.
1751 *) Undo Cygwin change.
1754 *) Added support for proxy certificates according to RFC 3820.
1755 Because they may be a security thread to unaware applications,
1756 they must be explicitely allowed in run-time. See
1757 docs/HOWTO/proxy_certificates.txt for further information.
1760 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1762 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1763 server and client random values. Previously
1764 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1765 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1767 This change has negligible security impact because:
1769 1. Server and client random values still have 24 bytes of pseudo random
1772 2. Server and client random values are sent in the clear in the initial
1775 3. The master secret is derived using the premaster secret (48 bytes in
1776 size for static RSA ciphersuites) as well as client server and random
1779 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1782 [Stephen Henson, reported by UK NISCC]
1784 *) Use Windows randomness collection on Cygwin.
1787 *) Fix hang in EGD/PRNGD query when communication socket is closed
1788 prematurely by EGD/PRNGD.
1789 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1791 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1794 *) Back-port of selected performance improvements from development
1795 branch, as well as improved support for PowerPC platforms.
1798 *) Add lots of checks for memory allocation failure, error codes to indicate
1799 failure and freeing up memory if a failure occurs.
1800 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1802 *) Add new -passin argument to dgst.
1805 *) Perform some character comparisons of different types in X509_NAME_cmp:
1806 this is needed for some certificates that reencode DNs into UTF8Strings
1807 (in violation of RFC3280) and can't or wont issue name rollover
1811 *) Make an explicit check during certificate validation to see that
1812 the CA setting in each certificate on the chain is correct. As a
1813 side effect always do the following basic checks on extensions,
1814 not just when there's an associated purpose to the check:
1816 - if there is an unhandled critical extension (unless the user
1817 has chosen to ignore this fault)
1818 - if the path length has been exceeded (if one is set at all)
1819 - that certain extensions fit the associated purpose (if one has
1823 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1825 *) Avoid a race condition when CRLs are checked in a multi threaded
1826 environment. This would happen due to the reordering of the revoked
1827 entries during signature checking and serial number lookup. Now the
1828 encoding is cached and the serial number sort performed under a lock.
1829 Add new STACK function sk_is_sorted().
1832 *) Add Delta CRL to the extension code.
1835 *) Various fixes to s3_pkt.c so alerts are sent properly.
1836 [David Holmes <d.holmes@f5.com>]
1838 *) Reduce the chances of duplicate issuer name and serial numbers (in
1839 violation of RFC3280) using the OpenSSL certificate creation utilities.
1840 This is done by creating a random 64 bit value for the initial serial
1841 number when a serial number file is created or when a self signed
1842 certificate is created using 'openssl req -x509'. The initial serial
1843 number file is created using 'openssl x509 -next_serial' in CA.pl
1844 rather than being initialized to 1.
1847 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1849 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1850 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1851 [Joe Orton, Steve Henson]
1853 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1855 [Joe Orton, Steve Henson]
1857 *) Make it possible to have multiple active certificates with the same
1858 subject in the CA index file. This is done only if the keyword
1859 'unique_subject' is set to 'no' in the main CA section (default
1860 if 'CA_default') of the configuration file. The value is saved
1861 with the database itself in a separate index attribute file,
1862 named like the index file with '.attr' appended to the name.
1865 *) X509 verify fixes. Disable broken certificate workarounds when
1866 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1867 keyUsage extension present. Don't accept CRLs with unhandled critical
1868 extensions: since verify currently doesn't process CRL extensions this
1869 rejects a CRL with *any* critical extensions. Add new verify error codes
1873 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1874 A clarification of RFC2560 will require the use of OCTET STRINGs and
1875 some implementations cannot handle the current raw format. Since OpenSSL
1876 copies and compares OCSP nonces as opaque blobs without any attempt at
1877 parsing them this should not create any compatibility issues.
1880 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1881 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1882 this HMAC (and other) operations are several times slower than OpenSSL
1886 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1887 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1889 *) Use the correct content when signing type "other".
1892 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1894 *) Fix various bugs revealed by running the NISCC test suite:
1896 Stop out of bounds reads in the ASN1 code when presented with
1897 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1899 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1901 If verify callback ignores invalid public key errors don't try to check
1902 certificate signature with the NULL public key.
1906 *) New -ignore_err option in ocsp application to stop the server
1907 exiting on the first error in a request.
1910 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1911 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1915 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1916 extra data after the compression methods not only for TLS 1.0
1917 but also for SSL 3.0 (as required by the specification).
1918 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1920 *) Change X509_certificate_type() to mark the key as exported/exportable
1921 when it's 512 *bits* long, not 512 bytes.
1924 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1925 blocks during encryption.
1928 *) Various fixes to base64 BIO and non blocking I/O. On write
1929 flushes were not handled properly if the BIO retried. On read
1930 data was not being buffered properly and had various logic bugs.
1931 This also affects blocking I/O when the data being decoded is a
1935 *) Various S/MIME bugfixes and compatibility changes:
1936 output correct application/pkcs7 MIME type if
1937 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1938 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1939 of files as .eml work). Correctly handle very long lines in MIME
1943 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1945 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1946 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1947 a protocol version number mismatch like a decryption error
1948 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1951 *) Turn on RSA blinding by default in the default implementation
1952 to avoid a timing attack. Applications that don't want it can call
1953 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1954 They would be ill-advised to do so in most cases.
1955 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1957 *) Change RSA blinding code so that it works when the PRNG is not
1958 seeded (in this case, the secret RSA exponent is abused as
1959 an unpredictable seed -- if it is not unpredictable, there
1960 is no point in blinding anyway). Make RSA blinding thread-safe
1961 by remembering the creator's thread ID in rsa->blinding and
1962 having all other threads use local one-time blinding factors
1963 (this requires more computation than sharing rsa->blinding, but
1964 avoids excessive locking; and if an RSA object is not shared
1965 between threads, blinding will still be very fast).
1968 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1969 ENGINE as defaults for all supported algorithms irrespective of
1970 the 'flags' parameter. 'flags' is now honoured, so applications
1971 should make sure they are passing it correctly.
1974 *) Target "mingw" now allows native Windows code to be generated in
1975 the Cygwin environment as well as with the MinGW compiler.
1978 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1980 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1981 via timing by performing a MAC computation even if incorrrect
1982 block cipher padding has been found. This is a countermeasure
1983 against active attacks where the attacker has to distinguish
1984 between bad padding and a MAC verification error. (CVE-2003-0078)
1986 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1987 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1988 Martin Vuagnoux (EPFL, Ilion)]
1990 *) Make the no-err option work as intended. The intention with no-err
1991 is not to have the whole error stack handling routines removed from
1992 libcrypto, it's only intended to remove all the function name and
1993 reason texts, thereby removing some of the footprint that may not
1994 be interesting if those errors aren't displayed anyway.
1996 NOTE: it's still possible for any application or module to have it's
1997 own set of error texts inserted. The routines are there, just not
1998 used by default when no-err is given.
2001 *) Add support for FreeBSD on IA64.
2002 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2004 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2005 Kerberos function mit_des_cbc_cksum(). Before this change,
2006 the value returned by DES_cbc_cksum() was like the one from
2007 mit_des_cbc_cksum(), except the bytes were swapped.
2008 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2010 *) Allow an application to disable the automatic SSL chain building.
2011 Before this a rather primitive chain build was always performed in
2012 ssl3_output_cert_chain(): an application had no way to send the
2013 correct chain if the automatic operation produced an incorrect result.
2015 Now the chain builder is disabled if either:
2017 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2019 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2021 The reasoning behind this is that an application would not want the
2022 auto chain building to take place if extra chain certificates are
2023 present and it might also want a means of sending no additional
2024 certificates (for example the chain has two certificates and the
2028 *) Add the possibility to build without the ENGINE framework.
2029 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2031 *) Under Win32 gmtime() can return NULL: check return value in
2032 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2035 *) DSA routines: under certain error conditions uninitialized BN objects
2036 could be freed. Solution: make sure initialization is performed early
2037 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2038 Nils Larsch <nla@trustcenter.de> via PR#459)
2041 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2042 checked on reconnect on the client side, therefore session resumption
2043 could still fail with a "ssl session id is different" error. This
2044 behaviour is masked when SSL_OP_ALL is used due to
2045 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2046 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2047 followup to PR #377.
2050 *) IA-32 assembler support enhancements: unified ELF targets, support
2051 for SCO/Caldera platforms, fix for Cygwin shared build.
2054 *) Add support for FreeBSD on sparc64. As a consequence, support for
2055 FreeBSD on non-x86 processors is separate from x86 processors on
2056 the config script, much like the NetBSD support.
2057 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2059 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2061 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2064 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2065 code (06) was taken as the first octet of the session ID and the last
2066 octet was ignored consequently. As a result SSLv2 client side session
2067 caching could not have worked due to the session ID mismatch between
2069 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2073 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2074 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2078 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2079 seems that in spite of existing for more than a year, many application
2080 author have done nothing to provide the necessary callbacks, which
2081 means that this particular engine will not work properly anywhere.
2082 This is a very unfortunate situation which forces us, in the name
2083 of usability, to give the hw_ncipher.c a static lock, which is part
2085 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2086 appear in 0.9.8 or later. We EXPECT application authors to have
2087 dealt properly with this when 0.9.8 is released (unless we actually
2088 make such changes in the libcrypto locking code that changes will
2089 have to be made anyway).
2092 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2093 octets have been read, EOF or an error occurs. Without this change
2094 some truncated ASN1 structures will not produce an error.
2097 *) Disable Heimdal support, since it hasn't been fully implemented.
2098 Still give the possibility to force the use of Heimdal, but with
2099 warnings and a request that patches get sent to openssl-dev.
2102 *) Add the VC-CE target, introduce the WINCE sysname, and add
2103 INSTALL.WCE and appropriate conditionals to make it build.
2104 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2106 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2107 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2108 edit numbers of the version.
2109 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2111 *) Introduce safe string copy and catenation functions
2112 (BUF_strlcpy() and BUF_strlcat()).
2113 [Ben Laurie (CHATS) and Richard Levitte]
2115 *) Avoid using fixed-size buffers for one-line DNs.
2116 [Ben Laurie (CHATS)]
2118 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2119 resizing buffers containing secrets, and use where appropriate.
2120 [Ben Laurie (CHATS)]
2122 *) Avoid using fixed size buffers for configuration file location.
2123 [Ben Laurie (CHATS)]
2125 *) Avoid filename truncation for various CA files.
2126 [Ben Laurie (CHATS)]
2128 *) Use sizeof in preference to magic numbers.
2129 [Ben Laurie (CHATS)]
2131 *) Avoid filename truncation in cert requests.
2132 [Ben Laurie (CHATS)]
2134 *) Add assertions to check for (supposedly impossible) buffer
2136 [Ben Laurie (CHATS)]
2138 *) Don't cache truncated DNS entries in the local cache (this could
2139 potentially lead to a spoofing attack).
2140 [Ben Laurie (CHATS)]
2142 *) Fix various buffers to be large enough for hex/decimal
2143 representations in a platform independent manner.
2144 [Ben Laurie (CHATS)]
2146 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2147 resizing buffers containing secrets, and use where appropriate.
2148 [Ben Laurie (CHATS)]
2150 *) Add BIO_indent() to avoid much slightly worrying code to do
2152 [Ben Laurie (CHATS)]
2154 *) Convert sprintf()/BIO_puts() to BIO_printf().
2155 [Ben Laurie (CHATS)]
2157 *) buffer_gets() could terminate with the buffer only half
2159 [Ben Laurie (CHATS)]
2161 *) Add assertions to prevent user-supplied crypto functions from
2162 overflowing internal buffers by having large block sizes, etc.
2163 [Ben Laurie (CHATS)]
2165 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2167 [Ben Laurie (CHATS)]
2169 *) Eliminate unused copy of key in RC4.
2170 [Ben Laurie (CHATS)]
2172 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2173 [Ben Laurie (CHATS)]
2175 *) Fix off-by-one error in EGD path.
2176 [Ben Laurie (CHATS)]
2178 *) If RANDFILE path is too long, ignore instead of truncating.
2179 [Ben Laurie (CHATS)]
2181 *) Eliminate unused and incorrectly sized X.509 structure
2183 [Ben Laurie (CHATS)]
2185 *) Eliminate unused and dangerous function knumber().
2186 [Ben Laurie (CHATS)]
2188 *) Eliminate unused and dangerous structure, KSSL_ERR.
2189 [Ben Laurie (CHATS)]
2191 *) Protect against overlong session ID context length in an encoded
2192 session object. Since these are local, this does not appear to be
2194 [Ben Laurie (CHATS)]
2196 *) Change from security patch (see 0.9.6e below) that did not affect
2197 the 0.9.6 release series:
2199 Remote buffer overflow in SSL3 protocol - an attacker could
2200 supply an oversized master key in Kerberos-enabled versions.
2202 [Ben Laurie (CHATS)]
2204 *) Change the SSL kerb5 codes to match RFC 2712.
2207 *) Make -nameopt work fully for req and add -reqopt switch.
2208 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2210 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2211 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2213 *) Make sure tests can be performed even if the corresponding algorithms
2214 have been removed entirely. This was also the last step to make
2215 OpenSSL compilable with DJGPP under all reasonable conditions.
2216 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2218 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2219 to allow version independent disabling of normally unselected ciphers,
2220 which may be activated as a side-effect of selecting a single cipher.
2222 (E.g., cipher list string "RSA" enables ciphersuites that are left
2223 out of "ALL" because they do not provide symmetric encryption.
2224 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2225 [Lutz Jaenicke, Bodo Moeller]
2227 *) Add appropriate support for separate platform-dependent build
2228 directories. The recommended way to make a platform-dependent
2229 build directory is the following (tested on Linux), maybe with
2232 # Place yourself outside of the OpenSSL source tree. In
2233 # this example, the environment variable OPENSSL_SOURCE
2234 # is assumed to contain the absolute OpenSSL source directory.
2235 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2236 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2237 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2238 mkdir -p `dirname $F`
2239 ln -s $OPENSSL_SOURCE/$F $F
2242 To be absolutely sure not to disturb the source tree, a "make clean"
2243 is a good thing. If it isn't successfull, don't worry about it,
2244 it probably means the source directory is very clean.
2247 *) Make sure any ENGINE control commands make local copies of string
2248 pointers passed to them whenever necessary. Otherwise it is possible
2249 the caller may have overwritten (or deallocated) the original string
2250 data when a later ENGINE operation tries to use the stored values.
2251 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2253 *) Improve diagnostics in file reading and command-line digests.
2254 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2256 *) Add AES modes CFB and OFB to the object database. Correct an
2257 error in AES-CFB decryption.
2260 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2261 allows existing EVP_CIPHER_CTX structures to be reused after
2262 calling EVP_*Final(). This behaviour is used by encryption
2263 BIOs and some applications. This has the side effect that
2264 applications must explicitly clean up cipher contexts with
2265 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2268 *) Check the values of dna and dnb in bn_mul_recursive before calling
2269 bn_mul_comba (a non zero value means the a or b arrays do not contain
2270 n2 elements) and fallback to bn_mul_normal if either is not zero.
2273 *) Fix escaping of non-ASCII characters when using the -subj option
2274 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2277 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2278 form for "surname", serialNumber has no short form.
2279 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2280 therefore remove "mail" short name for "internet 7".
2281 The OID for unique identifiers in X509 certificates is
2282 x500UniqueIdentifier, not uniqueIdentifier.
2283 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2286 *) Add an "init" command to the ENGINE config module and auto initialize
2287 ENGINEs. Without any "init" command the ENGINE will be initialized
2288 after all ctrl commands have been executed on it. If init=1 the
2289 ENGINE is initailized at that point (ctrls before that point are run
2290 on the uninitialized ENGINE and after on the initialized one). If
2291 init=0 then the ENGINE will not be iniatialized at all.
2294 *) Fix the 'app_verify_callback' interface so that the user-defined
2295 argument is actually passed to the callback: In the
2296 SSL_CTX_set_cert_verify_callback() prototype, the callback
2297 declaration has been changed from
2300 int (*cb)(X509_STORE_CTX *,void *);
2301 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2302 i=s->ctx->app_verify_callback(&ctx)
2303 has been changed into
2304 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2306 To update applications using SSL_CTX_set_cert_verify_callback(),
2307 a dummy argument can be added to their callback functions.
2308 [D. K. Smetters <smetters@parc.xerox.com>]
2310 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2311 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2313 *) Add and OPENSSL_LOAD_CONF define which will cause
2314 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2315 This allows older applications to transparently support certain
2316 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2317 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2318 load the config file and OPENSSL_add_all_algorithms_conf() which will
2319 always load it have also been added.
2322 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2323 Adjust NIDs and EVP layer.
2324 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2326 *) Config modules support in openssl utility.
2328 Most commands now load modules from the config file,
2329 though in a few (such as version) this isn't done
2330 because it couldn't be used for anything.
2332 In the case of ca and req the config file used is
2333 the same as the utility itself: that is the -config
2334 command line option can be used to specify an
2338 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2339 use "openssl_conf" if filename is NULL use default openssl config file.
2342 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2343 config section name. Add a new flag to tolerate a missing config file
2344 and move code to CONF_modules_load_file().
2347 *) Support for crypto accelerator cards from Accelerated Encryption
2348 Processing, www.aep.ie. (Use engine 'aep')
2349 The support was copied from 0.9.6c [engine] and adapted/corrected
2350 to work with the new engine framework.
2351 [AEP Inc. and Richard Levitte]
2353 *) Support for SureWare crypto accelerator cards from Baltimore
2354 Technologies. (Use engine 'sureware')
2355 The support was copied from 0.9.6c [engine] and adapted
2356 to work with the new engine framework.
2359 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2360 make the newer ENGINE framework commands for the CHIL engine work.
2361 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2363 *) Make it possible to produce shared libraries on ReliantUNIX.
2364 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2366 *) Add the configuration target debug-linux-ppro.
2367 Make 'openssl rsa' use the general key loading routines
2368 implemented in apps.c, and make those routines able to
2369 handle the key format FORMAT_NETSCAPE and the variant
2371 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2373 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2374 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2376 *) Add -keyform to rsautl, and document -engine.
2377 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2379 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2380 BIO_R_NO_SUCH_FILE error code rather than the generic
2381 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2384 *) Add new functions
2386 ERR_peek_last_error_line
2387 ERR_peek_last_error_line_data.
2388 These are similar to
2391 ERR_peek_error_line_data,
2392 but report on the latest error recorded rather than the first one
2393 still in the error queue.
2394 [Ben Laurie, Bodo Moeller]
2396 *) default_algorithms option in ENGINE config module. This allows things
2398 default_algorithms = ALL
2399 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2402 *) Prelminary ENGINE config module.
2405 *) New experimental application configuration code.
2408 *) Change the AES code to follow the same name structure as all other
2409 symmetric ciphers, and behave the same way. Move everything to
2410 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2411 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2413 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2414 [Ben Laurie and Theo de Raadt]
2416 *) Add option to output public keys in req command.
2417 [Massimiliano Pala madwolf@openca.org]
2419 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2420 (up to about 10% better than before for P-192 and P-224).
2423 *) New functions/macros
2425 SSL_CTX_set_msg_callback(ctx, cb)
2426 SSL_CTX_set_msg_callback_arg(ctx, arg)
2427 SSL_set_msg_callback(ssl, cb)
2428 SSL_set_msg_callback_arg(ssl, arg)
2430 to request calling a callback function
2432 void cb(int write_p, int version, int content_type,
2433 const void *buf, size_t len, SSL *ssl, void *arg)
2435 whenever a protocol message has been completely received
2436 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2437 protocol version according to which the SSL library interprets
2438 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2439 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2440 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2441 specification (change_cipher_spec(20), alert(21), handshake(22)).
2442 'buf' and 'len' point to the actual message, 'ssl' to the
2443 SSL object, and 'arg' is the application-defined value set by
2444 SSL[_CTX]_set_msg_callback_arg().
2446 'openssl s_client' and 'openssl s_server' have new '-msg' options
2447 to enable a callback that displays all protocol messages.
2450 *) Change the shared library support so shared libraries are built as
2451 soon as the corresponding static library is finished, and thereby get
2452 openssl and the test programs linked against the shared library.
2453 This still only happens when the keyword "shard" has been given to
2454 the configuration scripts.
2456 NOTE: shared library support is still an experimental thing, and
2457 backward binary compatibility is still not guaranteed.
2458 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2460 *) Add support for Subject Information Access extension.
2461 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2463 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2464 additional bytes when new memory had to be allocated, not just
2465 when reusing an existing buffer.
2468 *) New command line and configuration option 'utf8' for the req command.
2469 This allows field values to be specified as UTF8 strings.
2472 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2473 runs for the former and machine-readable output for the latter.
2476 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2477 of the e-mail address in the DN (i.e., it will go into a certificate
2478 extension only). The new configuration file option 'email_in_dn = no'
2479 has the same effect.
2480 [Massimiliano Pala madwolf@openca.org]
2482 *) Change all functions with names starting with des_ to be starting
2483 with DES_ instead. Add wrappers that are compatible with libdes,
2484 but are named _ossl_old_des_*. Finally, add macros that map the
2485 des_* symbols to the corresponding _ossl_old_des_* if libdes
2486 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2487 desired, the des_* symbols will be mapped to DES_*, with one
2490 Since we provide two compatibility mappings, the user needs to
2491 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2492 compatibility is desired. The default (i.e., when that macro
2493 isn't defined) is OpenSSL 0.9.6c compatibility.
2495 There are also macros that enable and disable the support of old
2496 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2497 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2498 are defined, the default will apply: to support the old des routines.
2500 In either case, one must include openssl/des.h to get the correct
2501 definitions. Do not try to just include openssl/des_old.h, that
2504 NOTE: This is a major break of an old API into a new one. Software
2505 authors are encouraged to switch to the DES_ style functions. Some
2506 time in the future, des_old.h and the libdes compatibility functions
2507 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2508 default), and then completely removed.
2511 *) Test for certificates which contain unsupported critical extensions.
2512 If such a certificate is found during a verify operation it is
2513 rejected by default: this behaviour can be overridden by either
2514 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2515 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2516 X509_supported_extension() has also been added which returns 1 if a
2517 particular extension is supported.
2520 *) Modify the behaviour of EVP cipher functions in similar way to digests
2521 to retain compatibility with existing code.
2524 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2525 compatibility with existing code. In particular the 'ctx' parameter does
2526 not have to be to be initialized before the call to EVP_DigestInit() and
2527 it is tidied up after a call to EVP_DigestFinal(). New function
2528 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2529 EVP_MD_CTX_copy() changed to not require the destination to be
2530 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2531 requires the destination to be valid.
2533 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2534 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2537 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2538 so that complete 'Handshake' protocol structures are kept in memory
2539 instead of overwriting 'msg_type' and 'length' with 'body' data.
2542 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2543 [Massimo Santin via Richard Levitte]
2545 *) Major restructuring to the underlying ENGINE code. This includes
2546 reduction of linker bloat, separation of pure "ENGINE" manipulation
2547 (initialisation, etc) from functionality dealing with implementations
2548 of specific crypto iterfaces. This change also introduces integrated
2549 support for symmetric ciphers and digest implementations - so ENGINEs
2550 can now accelerate these by providing EVP_CIPHER and EVP_MD
2551 implementations of their own. This is detailed in crypto/engine/README
2552 as it couldn't be adequately described here. However, there are a few
2553 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2554 were changed in the original introduction of ENGINE code have now
2555 reverted back - the hooking from this code to ENGINE is now a good
2556 deal more passive and at run-time, operations deal directly with
2557 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2558 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2559 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2560 they were not being used by the framework as there is no concept of a
2561 BIGNUM_METHOD and they could not be generalised to the new
2562 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2563 ENGINE_cpy() has been removed as it cannot be consistently defined in
2567 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2570 *) Change mkdef.pl to sort symbols that get the same entry number,
2571 and make sure the automatically generated functions ERR_load_*
2572 become part of libeay.num as well.
2575 *) New function SSL_renegotiate_pending(). This returns true once
2576 renegotiation has been requested (either SSL_renegotiate() call
2577 or HelloRequest/ClientHello receveived from the peer) and becomes
2578 false once a handshake has been completed.
2579 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2580 sends a HelloRequest, but does not ensure that a handshake takes
2581 place. SSL_renegotiate_pending() is useful for checking if the
2582 client has followed the request.)
2585 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2586 By default, clients may request session resumption even during
2587 renegotiation (if session ID contexts permit); with this option,
2588 session resumption is possible only in the first handshake.
2590 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2591 more bits available for options that should not be part of
2592 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2595 *) Add some demos for certificate and certificate request creation.
2598 *) Make maximum certificate chain size accepted from the peer application
2599 settable (SSL*_get/set_max_cert_list()), as proposed by
2600 "Douglas E. Engert" <deengert@anl.gov>.
2603 *) Add support for shared libraries for Unixware-7
2604 (Boyd Lynn Gerber <gerberb@zenez.com>).
2607 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2608 be done prior to destruction. Use this to unload error strings from
2609 ENGINEs that load their own error strings. NB: This adds two new API
2610 functions to "get" and "set" this destroy handler in an ENGINE.
2613 *) Alter all existing ENGINE implementations (except "openssl" and
2614 "openbsd") to dynamically instantiate their own error strings. This
2615 makes them more flexible to be built both as statically-linked ENGINEs
2616 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2617 Also, add stub code to each that makes building them as self-contained
2618 shared-libraries easier (see README.ENGINE).
2621 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2622 implementations into applications that are completely implemented in
2623 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2624 commands that can be used to configure what shared-library to load and
2625 to control aspects of the way it is handled. Also, made an update to
2626 the README.ENGINE file that brings its information up-to-date and
2627 provides some information and instructions on the "dynamic" ENGINE
2628 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2631 *) Make it possible to unload ranges of ERR strings with a new
2632 "ERR_unload_strings" function.
2635 *) Add a copy() function to EVP_MD.
2638 *) Make EVP_MD routines take a context pointer instead of just the
2639 md_data void pointer.
2642 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2643 that the digest can only process a single chunk of data
2644 (typically because it is provided by a piece of
2645 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2646 is only going to provide a single chunk of data, and hence the
2647 framework needn't accumulate the data for oneshot drivers.
2650 *) As with "ERR", make it possible to replace the underlying "ex_data"
2651 functions. This change also alters the storage and management of global
2652 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2653 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2654 index counters. The API functions that use this state have been changed
2655 to take a "class_index" rather than pointers to the class's local STACK
2656 and counter, and there is now an API function to dynamically create new
2657 classes. This centralisation allows us to (a) plug a lot of the
2658 thread-safety problems that existed, and (b) makes it possible to clean
2659 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2660 such data would previously have always leaked in application code and
2661 workarounds were in place to make the memory debugging turn a blind eye
2662 to it. Application code that doesn't use this new function will still
2663 leak as before, but their memory debugging output will announce it now
2664 rather than letting it slide.
2666 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2667 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2668 has a return value to indicate success or failure.
2671 *) Make it possible to replace the underlying "ERR" functions such that the
2672 global state (2 LHASH tables and 2 locks) is only used by the "default"
2673 implementation. This change also adds two functions to "get" and "set"
2674 the implementation prior to it being automatically set the first time
2675 any other ERR function takes place. Ie. an application can call "get",
2676 pass the return value to a module it has just loaded, and that module
2677 can call its own "set" function using that value. This means the
2678 module's "ERR" operations will use (and modify) the error state in the
2679 application and not in its own statically linked copy of OpenSSL code.
2682 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2683 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2684 the operation, and provides a more encapsulated way for external code
2685 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2686 to use these functions rather than manually incrementing the counts.
2688 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2691 *) Add EVP test program.
2694 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2697 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2698 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2699 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2700 These allow a CRL to be built without having to access X509_CRL fields
2701 directly. Modify 'ca' application to use new functions.
2704 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2705 bug workarounds. Rollback attack detection is a security feature.
2706 The problem will only arise on OpenSSL servers when TLSv1 is not
2707 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2708 Software authors not wanting to support TLSv1 will have special reasons
2709 for their choice and can explicitly enable this option.
2710 [Bodo Moeller, Lutz Jaenicke]
2712 *) Rationalise EVP so it can be extended: don't include a union of
2713 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2714 (similar to those existing for EVP_CIPHER_CTX).
2719 EVP_MD_CTX_init(&md); /* new function call */
2720 EVP_DigestInit(&md, EVP_sha1());
2721 EVP_DigestUpdate(&md, in, len);
2722 EVP_DigestFinal(&md, out, NULL);
2723 EVP_MD_CTX_cleanup(&md); /* new function call */
2727 *) Make DES key schedule conform to the usual scheme, as well as
2728 correcting its structure. This means that calls to DES functions
2729 now have to pass a pointer to a des_key_schedule instead of a
2730 plain des_key_schedule (which was actually always a pointer
2733 des_key_schedule ks;
2735 des_set_key_checked(..., &ks);
2736 des_ncbc_encrypt(..., &ks, ...);
2738 (Note that a later change renames 'des_...' into 'DES_...'.)
2741 *) Initial reduction of linker bloat: the use of some functions, such as
2742 PEM causes large amounts of unused functions to be linked in due to
2743 poor organisation. For example pem_all.c contains every PEM function
2744 which has a knock on effect of linking in large amounts of (unused)
2745 ASN1 code. Grouping together similar functions and splitting unrelated
2746 functions prevents this.
2749 *) Cleanup of EVP macros.
2752 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2753 correct _ecb suffix.
2756 *) Add initial OCSP responder support to ocsp application. The
2757 revocation information is handled using the text based index
2758 use by the ca application. The responder can either handle
2759 requests generated internally, supplied in files (for example
2760 via a CGI script) or using an internal minimal server.
2763 *) Add configuration choices to get zlib compression for TLS.
2766 *) Changes to Kerberos SSL for RFC 2712 compliance:
2767 1. Implemented real KerberosWrapper, instead of just using
2768 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2769 2. Implemented optional authenticator field of KerberosWrapper.
2771 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2772 and authenticator structs; see crypto/krb5/.
2774 Generalized Kerberos calls to support multiple Kerberos libraries.
2775 [Vern Staats <staatsvr@asc.hpc.mil>,
2776 Jeffrey Altman <jaltman@columbia.edu>
2777 via Richard Levitte]
2779 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2780 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2781 values for each of the key sizes rather than having just
2782 parameters (and 'speed' generating keys each time).
2785 *) Speed up EVP routines.
2788 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2789 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2790 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2791 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2793 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2794 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2795 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2798 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2800 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2803 *) Added the OS2-EMX target.
2804 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2806 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2807 to support NCONF routines in extension code. New function CONF_set_nconf()
2808 to allow functions which take an NCONF to also handle the old LHASH
2809 structure: this means that the old CONF compatible routines can be
2810 retained (in particular wrt extensions) without having to duplicate the
2811 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2814 *) Enhance the general user interface with mechanisms for inner control
2815 and with possibilities to have yes/no kind of prompts.
2818 *) Change all calls to low level digest routines in the library and
2819 applications to use EVP. Add missing calls to HMAC_cleanup() and
2820 don't assume HMAC_CTX can be copied using memcpy().
2821 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2823 *) Add the possibility to control engines through control names but with
2824 arbitrary arguments instead of just a string.
2825 Change the key loaders to take a UI_METHOD instead of a callback
2826 function pointer. NOTE: this breaks binary compatibility with earlier
2827 versions of OpenSSL [engine].
2828 Adapt the nCipher code for these new conditions and add a card insertion
2832 *) Enhance the general user interface with mechanisms to better support
2833 dialog box interfaces, application-defined prompts, the possibility
2834 to use defaults (for example default passwords from somewhere else)
2835 and interrupts/cancellations.
2838 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2839 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2842 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2843 tidy up some unnecessarily weird code in 'sk_new()').
2844 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2846 *) Change the key loading routines for ENGINEs to use the same kind
2847 callback (pem_password_cb) as all other routines that need this
2851 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2852 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2853 than this minimum value is recommended.
2856 *) New random seeder for OpenVMS, using the system process statistics
2857 that are easily reachable.
2860 *) Windows apparently can't transparently handle global
2861 variables defined in DLLs. Initialisations such as:
2863 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2865 wont compile. This is used by the any applications that need to
2866 declare their own ASN1 modules. This was fixed by adding the option
2867 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2868 needed for static libraries under Win32.
2871 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2872 setting of purpose and trust fields. New X509_STORE trust and
2873 purpose functions and tidy up setting in other SSL functions.
2876 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2877 structure. These are inherited by X509_STORE_CTX when it is
2878 initialised. This allows various defaults to be set in the
2879 X509_STORE structure (such as flags for CRL checking and custom
2880 purpose or trust settings) for functions which only use X509_STORE_CTX
2881 internally such as S/MIME.
2883 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2884 trust settings if they are not set in X509_STORE. This allows X509_STORE
2885 purposes and trust (in S/MIME for example) to override any set by default.
2887 Add command line options for CRL checking to smime, s_client and s_server
2891 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2892 are set then the CRL is looked up in the X509_STORE structure and
2893 its validity and signature checked, then if the certificate is found
2894 in the CRL the verify fails with a revoked error.
2896 Various new CRL related callbacks added to X509_STORE_CTX structure.
2898 Command line options added to 'verify' application to support this.
2900 This needs some additional work, such as being able to handle multiple
2901 CRLs with different times, extension based lookup (rather than just
2902 by subject name) and ultimately more complete V2 CRL extension
2906 *) Add a general user interface API (crypto/ui/). This is designed
2907 to replace things like des_read_password and friends (backward
2908 compatibility functions using this new API are provided).
2909 The purpose is to remove prompting functions from the DES code
2910 section as well as provide for prompting through dialog boxes in
2911 a window system and the like.
2914 *) Add "ex_data" support to ENGINE so implementations can add state at a
2915 per-structure level rather than having to store it globally.
2918 *) Make it possible for ENGINE structures to be copied when retrieved by
2919 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2920 This causes the "original" ENGINE structure to act like a template,
2921 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2922 operational state can be localised to each ENGINE structure, despite the
2923 fact they all share the same "methods". New ENGINE structures returned in
2924 this case have no functional references and the return value is the single
2925 structural reference. This matches the single structural reference returned
2926 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2930 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2931 needs to match any other type at all we need to manually clear the
2935 *) Changes to the "openssl engine" utility to include;
2936 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2937 about an ENGINE's available control commands.
2938 - executing control commands from command line arguments using the
2939 '-pre' and '-post' switches. '-post' is only used if '-t' is
2940 specified and the ENGINE is successfully initialised. The syntax for
2941 the individual commands are colon-separated, for example;
2942 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2945 *) New dynamic control command support for ENGINEs. ENGINEs can now
2946 declare their own commands (numbers), names (strings), descriptions,
2947 and input types for run-time discovery by calling applications. A
2948 subset of these commands are implicitly classed as "executable"
2949 depending on their input type, and only these can be invoked through
2950 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2951 can be based on user input, config files, etc). The distinction is
2952 that "executable" commands cannot return anything other than a boolean
2953 result and can only support numeric or string input, whereas some
2954 discoverable commands may only be for direct use through
2955 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2956 pointers, or other custom uses. The "executable" commands are to
2957 support parameterisations of ENGINE behaviour that can be
2958 unambiguously defined by ENGINEs and used consistently across any
2959 OpenSSL-based application. Commands have been added to all the
2960 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2961 control over shared-library paths without source code alterations.
2964 *) Changed all ENGINE implementations to dynamically allocate their
2965 ENGINEs rather than declaring them statically. Apart from this being
2966 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2967 this also allows the implementations to compile without using the
2968 internal engine_int.h header.
2971 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2972 'const' value. Any code that should be able to modify a RAND_METHOD
2973 should already have non-const pointers to it (ie. they should only
2974 modify their own ones).
2977 *) Made a variety of little tweaks to the ENGINE code.
2978 - "atalla" and "ubsec" string definitions were moved from header files
2979 to C code. "nuron" string definitions were placed in variables
2980 rather than hard-coded - allowing parameterisation of these values
2981 later on via ctrl() commands.
2982 - Removed unused "#if 0"'d code.
2983 - Fixed engine list iteration code so it uses ENGINE_free() to release
2984 structural references.
2985 - Constified the RAND_METHOD element of ENGINE structures.
2986 - Constified various get/set functions as appropriate and added
2987 missing functions (including a catch-all ENGINE_cpy that duplicates
2988 all ENGINE values onto a new ENGINE except reference counts/state).
2989 - Removed NULL parameter checks in get/set functions. Setting a method
2990 or function to NULL is a way of cancelling out a previously set
2991 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2992 and doesn't justify the extra error symbols and code.
2993 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2994 flags from engine_int.h to engine.h.
2995 - Changed prototypes for ENGINE handler functions (init(), finish(),
2996 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2999 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3000 to the algorithm using long division. The binary algorithm can be
3001 used only if the modulus is odd. On 32-bit systems, it is faster
3002 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3003 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3004 up to 450 bits. In 64-bit environments, the binary algorithm
3005 appears to be advantageous for much longer moduli; here we use it
3006 for moduli up to 2048 bits.
3009 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3010 could not support the combine flag in choice fields.
3013 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3014 extensions from a certificate request to the certificate.
3017 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3018 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3019 file: this allows the display of the certificate about to be
3020 signed to be customised, to allow certain fields to be included
3021 or excluded and extension details. The old system didn't display
3022 multicharacter strings properly, omitted fields not in the policy
3023 and couldn't display additional details such as extensions.
3026 *) Function EC_POINTs_mul for multiple scalar multiplication
3027 of an arbitrary number of elliptic curve points
3028 \sum scalars[i]*points[i],
3029 optionally including the generator defined for the EC_GROUP:
3030 scalar*generator + \sum scalars[i]*points[i].
3032 EC_POINT_mul is a simple wrapper function for the typical case
3033 that the point list has just one item (besides the optional
3037 *) First EC_METHODs for curves over GF(p):
3039 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3040 operations and provides various method functions that can also
3041 operate with faster implementations of modular arithmetic.
3043 EC_GFp_mont_method() reuses most functions that are part of
3044 EC_GFp_simple_method, but uses Montgomery arithmetic.
3046 [Bodo Moeller; point addition and point doubling
3047 implementation directly derived from source code provided by
3048 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3050 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3051 crypto/ec/ec_lib.c):
3053 Curves are EC_GROUP objects (with an optional group generator)
3054 based on EC_METHODs that are built into the library.
3056 Points are EC_POINT objects based on EC_GROUP objects.
3058 Most of the framework would be able to handle curves over arbitrary
3059 finite fields, but as there are no obvious types for fields other
3060 than GF(p), some functions are limited to that for now.
3063 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3064 that the file contains a complete HTTP response.
3067 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3068 change the def and num file printf format specifier from "%-40sXXX"
3069 to "%-39s XXX". The latter will always guarantee a space after the
3070 field while the former will cause them to run together if the field
3071 is 40 of more characters long.
3074 *) Constify the cipher and digest 'method' functions and structures
3075 and modify related functions to take constant EVP_MD and EVP_CIPHER
3079 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3080 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3083 *) Modify EVP_Digest*() routines so they now return values. Although the
3084 internal software routines can never fail additional hardware versions
3088 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3090 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3091 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3094 ERR_R_NESTED_ASN1_ERROR
3096 ERR_R_MISSING_ASN1_EOS
3097 were 4 .. 9, conflicting with
3098 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3100 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3101 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3103 Add new error code 'ERR_R_INTERNAL_ERROR'.
3106 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3110 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3111 sets the subject name for a new request or supersedes the
3112 subject name in a given request. Formats that can be parsed are
3113 'CN=Some Name, OU=myOU, C=IT'
3115 'CN=Some Name/OU=myOU/C=IT'.
3117 Add options '-batch' and '-verbose' to 'openssl req'.
3118 [Massimiliano Pala <madwolf@hackmasters.net>]
3120 *) Introduce the possibility to access global variables through
3121 functions on platform were that's the best way to handle exporting
3122 global variables in shared libraries. To enable this functionality,
3123 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3124 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3125 is normally done by Configure or something similar).
3127 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3128 in the source file (foo.c) like this:
3130 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3131 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3133 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3134 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3136 OPENSSL_DECLARE_GLOBAL(int,foo);
3137 #define foo OPENSSL_GLOBAL_REF(foo)
3138 OPENSSL_DECLARE_GLOBAL(double,bar);
3139 #define bar OPENSSL_GLOBAL_REF(bar)
3141 The #defines are very important, and therefore so is including the
3142 header file everywhere where the defined globals are used.
3144 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3145 of ASN.1 items, but that structure is a bit different.
3147 The largest change is in util/mkdef.pl which has been enhanced with
3148 better and easier to understand logic to choose which symbols should
3149 go into the Windows .def files as well as a number of fixes and code
3150 cleanup (among others, algorithm keywords are now sorted
3151 lexicographically to avoid constant rewrites).
3154 *) In BN_div() keep a copy of the sign of 'num' before writing the
3155 result to 'rm' because if rm==num the value will be overwritten
3156 and produce the wrong result if 'num' is negative: this caused
3157 problems with BN_mod() and BN_nnmod().
3160 *) Function OCSP_request_verify(). This checks the signature on an
3161 OCSP request and verifies the signer certificate. The signer
3162 certificate is just checked for a generic purpose and OCSP request
3166 *) Add OCSP_check_validity() function to check the validity of OCSP
3167 responses. OCSP responses are prepared in real time and may only
3168 be a few seconds old. Simply checking that the current time lies
3169 between thisUpdate and nextUpdate max reject otherwise valid responses
3170 caused by either OCSP responder or client clock inaccuracy. Instead
3171 we allow thisUpdate and nextUpdate to fall within a certain period of
3172 the current time. The age of the response can also optionally be
3173 checked. Two new options -validity_period and -status_age added to
3177 *) If signature or public key algorithm is unrecognized print out its
3178 OID rather that just UNKNOWN.
3181 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3182 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3183 ID to be generated from the issuer certificate alone which can then be
3184 passed to OCSP_id_issuer_cmp().
3187 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3188 ASN1 modules to export functions returning ASN1_ITEM pointers
3189 instead of the ASN1_ITEM structures themselves. This adds several
3190 new macros which allow the underlying ASN1 function/structure to
3191 be accessed transparently. As a result code should not use ASN1_ITEM
3192 references directly (such as &X509_it) but instead use the relevant
3193 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3194 use of the new ASN1 code on platforms where exporting structures
3195 is problematical (for example in shared libraries) but exporting
3196 functions returning pointers to structures is not.
3199 *) Add support for overriding the generation of SSL/TLS session IDs.
3200 These callbacks can be registered either in an SSL_CTX or per SSL.
3201 The purpose of this is to allow applications to control, if they wish,
3202 the arbitrary values chosen for use as session IDs, particularly as it
3203 can be useful for session caching in multiple-server environments. A
3204 command-line switch for testing this (and any client code that wishes
3205 to use such a feature) has been added to "s_server".
3206 [Geoff Thorpe, Lutz Jaenicke]
3208 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3209 of the form '#if defined(...) || defined(...) || ...' and
3210 '#if !defined(...) && !defined(...) && ...'. This also avoids
3211 the growing number of special cases it was previously handling.
3214 *) Make all configuration macros available for application by making
3215 sure they are available in opensslconf.h, by giving them names starting
3216 with "OPENSSL_" to avoid conflicts with other packages and by making
3217 sure e_os2.h will cover all platform-specific cases together with
3219 Additionally, it is now possible to define configuration/platform-
3220 specific names (called "system identities"). In the C code, these
3221 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3222 macro with the name beginning with "OPENSSL_SYS_", which is determined
3223 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3227 *) New option -set_serial to 'req' and 'x509' this allows the serial
3228 number to use to be specified on the command line. Previously self
3229 signed certificates were hard coded with serial number 0 and the
3230 CA options of 'x509' had to use a serial number in a file which was
3234 *) New options to 'ca' utility to support V2 CRL entry extensions.
3235 Currently CRL reason, invalidity date and hold instruction are
3236 supported. Add new CRL extensions to V3 code and some new objects.
3239 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3240 disable standard block padding (aka PKCS#5 padding) in the EVP
3241 API, which was previously mandatory. This means that the data is
3242 not padded in any way and so the total length much be a multiple
3243 of the block size, otherwise an error occurs.
3246 *) Initial (incomplete) OCSP SSL support.
3249 *) New function OCSP_parse_url(). This splits up a URL into its host,
3250 port and path components: primarily to parse OCSP URLs. New -url
3251 option to ocsp utility.
3254 *) New nonce behavior. The return value of OCSP_check_nonce() now
3255 reflects the various checks performed. Applications can decide
3256 whether to tolerate certain situations such as an absent nonce
3257 in a response when one was present in a request: the ocsp application
3258 just prints out a warning. New function OCSP_add1_basic_nonce()
3259 this is to allow responders to include a nonce in a response even if
3260 the request is nonce-less.
3263 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3264 skipped when using openssl x509 multiple times on a single input file,
3265 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3268 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3269 set string type: to handle setting ASN1_TIME structures. Fix ca
3270 utility to correctly initialize revocation date of CRLs.
3273 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3274 the clients preferred ciphersuites and rather use its own preferences.
3275 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3276 Internet Explorer by ensuring unchanged hash method during stepup.
3277 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3280 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3281 to aes and add a new 'exist' option to print out symbols that don't
3285 *) Additional options to ocsp utility to allow flags to be set and
3286 additional certificates supplied.
3289 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3290 OCSP client a number of certificate to only verify the response
3294 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3295 handle the new API. Currently only ECB, CBC modes supported. Add new
3298 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3299 Encryption Standard (AES) Ciphersuites for Transport Layer
3300 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3301 not enabled by default and were not part of the "ALL" ciphersuite
3302 alias because they were not yet official; they could be
3303 explicitly requested by specifying the "AESdraft" ciphersuite
3304 group alias. In the final release of OpenSSL 0.9.7, the group
3305 alias is called "AES" and is part of "ALL".)
3306 [Ben Laurie, Steve Henson, Bodo Moeller]
3308 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3309 request to response.
3312 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3313 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3314 extract information from a certificate request. OCSP_response_create()
3315 creates a response and optionally adds a basic response structure.
3316 OCSP_basic_add1_status() adds a complete single response to a basic
3317 response and returns the OCSP_SINGLERESP structure just added (to allow
3318 extensions to be included for example). OCSP_basic_add1_cert() adds a
3319 certificate to a basic response and OCSP_basic_sign() signs a basic
3320 response with various flags. New helper functions ASN1_TIME_check()
3321 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3322 (converts ASN1_TIME to GeneralizedTime).
3325 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3326 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3327 structure from a certificate. X509_pubkey_digest() digests the public_key
3328 contents: this is used in various key identifiers.
3331 *) Make sk_sort() tolerate a NULL argument.
3332 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3334 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3335 passed by the function are trusted implicitly. If any of them signed the
3336 response then it is assumed to be valid and is not verified.
3339 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3340 to data. This was previously part of the PKCS7 ASN1 code. This
3341 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3342 [Steve Henson, reported by Kenneth R. Robinette
3343 <support@securenetterm.com>]
3345 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3346 routines: without these tracing memory leaks is very painful.
3347 Fix leaks in PKCS12 and PKCS7 routines.
3350 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3351 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3352 effectively meant GeneralizedTime would never be used. Now it
3353 is initialised to -1 but X509_time_adj() now has to check the value
3354 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3355 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3356 [Steve Henson, reported by Kenneth R. Robinette
3357 <support@securenetterm.com>]
3359 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3360 result in a zero length in the ASN1_INTEGER structure which was
3361 not consistent with the structure when d2i_ASN1_INTEGER() was used
3362 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3363 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3364 where it did not print out a minus for negative ASN1_INTEGER.
3367 *) Add summary printout to ocsp utility. The various functions which
3368 convert status values to strings have been renamed to:
3369 OCSP_response_status_str(), OCSP_cert_status_str() and
3370 OCSP_crl_reason_str() and are no longer static. New options
3371 to verify nonce values and to disable verification. OCSP response
3372 printout format cleaned up.
3375 *) Add additional OCSP certificate checks. These are those specified
3376 in RFC2560. This consists of two separate checks: the CA of the
3377 certificate being checked must either be the OCSP signer certificate
3378 or the issuer of the OCSP signer certificate. In the latter case the
3379 OCSP signer certificate must contain the OCSP signing extended key
3380 usage. This check is performed by attempting to match the OCSP
3381 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3382 in the OCSP_CERTID structures of the response.
3385 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3386 and related routines. This uses the standard OpenSSL certificate
3387 verify routines to perform initial checks (just CA validity) and
3388 to obtain the certificate chain. Then additional checks will be
3389 performed on the chain. Currently the root CA is checked to see
3390 if it is explicitly trusted for OCSP signing. This is used to set
3391 a root CA as a global signing root: that is any certificate that
3392 chains to that CA is an acceptable OCSP signing certificate.
3395 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3396 extensions from a separate configuration file.
3397 As when reading extensions from the main configuration file,
3398 the '-extensions ...' option may be used for specifying the
3400 [Massimiliano Pala <madwolf@comune.modena.it>]
3402 *) New OCSP utility. Allows OCSP requests to be generated or
3403 read. The request can be sent to a responder and the output
3404 parsed, outputed or printed in text form. Not complete yet:
3405 still needs to check the OCSP response validity.
3408 *) New subcommands for 'openssl ca':
3409 'openssl ca -status <serial>' prints the status of the cert with
3410 the given serial number (according to the index file).
3411 'openssl ca -updatedb' updates the expiry status of certificates
3413 [Massimiliano Pala <madwolf@comune.modena.it>]
3415 *) New '-newreq-nodes' command option to CA.pl. This is like
3416 '-newreq', but calls 'openssl req' with the '-nodes' option
3417 so that the resulting key is not encrypted.
3418 [Damien Miller <djm@mindrot.org>]
3420 *) New configuration for the GNU Hurd.
3421 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3423 *) Initial code to implement OCSP basic response verify. This
3424 is currently incomplete. Currently just finds the signer's
3425 certificate and verifies the signature on the response.
3428 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3429 value of OPENSSLDIR. This is available via the new '-d' option
3430 to 'openssl version', and is also included in 'openssl version -a'.
3433 *) Allowing defining memory allocation callbacks that will be given
3434 file name and line number information in additional arguments
3435 (a const char* and an int). The basic functionality remains, as
3436 well as the original possibility to just replace malloc(),
3437 realloc() and free() by functions that do not know about these
3438 additional arguments. To register and find out the current
3439 settings for extended allocation functions, the following
3440 functions are provided:
3442 CRYPTO_set_mem_ex_functions
3443 CRYPTO_set_locked_mem_ex_functions
3444 CRYPTO_get_mem_ex_functions
3445 CRYPTO_get_locked_mem_ex_functions
3447 These work the same way as CRYPTO_set_mem_functions and friends.
3448 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3449 extended allocation function is enabled.
3450 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3451 a conventional allocation function is enabled.
3452 [Richard Levitte, Bodo Moeller]
3454 *) Finish off removing the remaining LHASH function pointer casts.
3455 There should no longer be any prototype-casting required when using
3456 the LHASH abstraction, and any casts that remain are "bugs". See
3457 the callback types and macros at the head of lhash.h for details
3458 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3461 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3462 If /dev/[u]random devices are not available or do not return enough
3463 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3465 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3466 /etc/entropy will be queried once each in this sequence, quering stops
3467 when enough entropy was collected without querying more sockets.
3470 *) Change the Unix RAND_poll() variant to be able to poll several
3471 random devices, as specified by DEVRANDOM, until a sufficient amount
3472 of data has been collected. We spend at most 10 ms on each file
3473 (select timeout) and read in non-blocking mode. DEVRANDOM now
3474 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3475 (previously it was just the string "/dev/urandom"), so on typical
3476 platforms the 10 ms delay will never occur.
3477 Also separate out the Unix variant to its own file, rand_unix.c.
3478 For VMS, there's a currently-empty rand_vms.c.
3481 *) Move OCSP client related routines to ocsp_cl.c. These
3482 provide utility functions which an application needing
3483 to issue a request to an OCSP responder and analyse the
3484 response will typically need: as opposed to those which an
3485 OCSP responder itself would need which will be added later.
3487 OCSP_request_sign() signs an OCSP request with an API similar
3488 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3489 response. OCSP_response_get1_basic() extracts basic response
3490 from response. OCSP_resp_find_status(): finds and extracts status
3491 information from an OCSP_CERTID structure (which will be created
3492 when the request structure is built). These are built from lower
3493 level functions which work on OCSP_SINGLERESP structures but
3494 wont normally be used unless the application wishes to examine
3495 extensions in the OCSP response for example.
3497 Replace nonce routines with a pair of functions.
3498 OCSP_request_add1_nonce() adds a nonce value and optionally
3499 generates a random value. OCSP_check_nonce() checks the
3500 validity of the nonce in an OCSP response.
3503 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3504 This doesn't copy the supplied OCSP_CERTID and avoids the
3505 need to free up the newly created id. Change return type
3506 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3507 This can then be used to add extensions to the request.
3508 Deleted OCSP_request_new(), since most of its functionality
3509 is now in OCSP_REQUEST_new() (and the case insensitive name
3510 clash) apart from the ability to set the request name which
3511 will be added elsewhere.
3514 *) Update OCSP API. Remove obsolete extensions argument from
3515 various functions. Extensions are now handled using the new
3516 OCSP extension code. New simple OCSP HTTP function which
3517 can be used to send requests and parse the response.
3520 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3521 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3522 uses the special reorder version of SET OF to sort the attributes
3523 and reorder them to match the encoded order. This resolves a long
3524 standing problem: a verify on a PKCS7 structure just after signing
3525 it used to fail because the attribute order did not match the
3526 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3527 it uses the received order. This is necessary to tolerate some broken
3528 software that does not order SET OF. This is handled by encoding
3529 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3530 to produce the required SET OF.
3533 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3534 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3535 files to get correct declarations of the ASN.1 item variables.
3538 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3539 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3540 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3541 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3542 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3543 ASN1_ITEM and no wrapper functions.
3546 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3547 replace the old function pointer based I/O routines. Change most of
3548 the *_d2i_bio() and *_d2i_fp() functions to use these.
3551 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3552 lines, recognice more "algorithms" that can be deselected, and make
3553 it complain about algorithm deselection that isn't recognised.
3556 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3557 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3558 to use new functions. Add NO_ASN1_OLD which can be set to remove
3559 some old style ASN1 functions: this can be used to determine if old
3560 code will still work when these eventually go away.
3563 *) New extension functions for OCSP structures, these follow the
3564 same conventions as certificates and CRLs.
3567 *) New function X509V3_add1_i2d(). This automatically encodes and
3568 adds an extension. Its behaviour can be customised with various
3569 flags to append, replace or delete. Various wrappers added for
3570 certifcates and CRLs.
3573 *) Fix to avoid calling the underlying ASN1 print routine when
3574 an extension cannot be parsed. Correct a typo in the
3575 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3578 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3579 entries for variables.
3582 *) Add functionality to apps/openssl.c for detecting locking
3583 problems: As the program is single-threaded, all we have
3584 to do is register a locking callback using an array for
3585 storing which locks are currently held by the program.
3588 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3589 SSL_get_ex_data_X509_STORE_idx(), which is used in
3590 ssl_verify_cert_chain() and thus can be called at any time
3591 during TLS/SSL handshakes so that thread-safety is essential.
3592 Unfortunately, the ex_data design is not at all suited
3593 for multi-threaded use, so it probably should be abolished.
3596 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3597 [Broadcom, tweaked and integrated by Geoff Thorpe]
3599 *) Move common extension printing code to new function
3600 X509V3_print_extensions(). Reorganise OCSP print routines and
3601 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3604 *) New function X509_signature_print() to remove duplication in some
3608 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3609 set (this was treated exactly the same as SET OF previously). This
3610 is used to reorder the STACK representing the structure to match the
3611 encoding. This will be used to get round a problem where a PKCS7
3612 structure which was signed could not be verified because the STACK
3613 order did not reflect the encoded order.
3616 *) Reimplement the OCSP ASN1 module using the new code.
3619 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3620 for its ASN1 operations. The old style function pointers still exist
3621 for now but they will eventually go away.
3624 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3625 completely replaces the old ASN1 functionality with a table driven
3626 encoder and decoder which interprets an ASN1_ITEM structure describing
3627 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3628 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3629 has also been converted to the new form.
3632 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3633 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3634 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3635 for negative moduli.
3638 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3639 of not touching the result's sign bit.
3642 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3646 *) Changed the LHASH code to use prototypes for callbacks, and created
3647 macros to declare and implement thin (optionally static) functions
3648 that provide type-safety and avoid function pointer casting for the
3649 type-specific callbacks.
3652 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3654 [Veers Staats <staatsvr@asc.hpc.mil>,
3655 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3657 *) Reformat the FAQ so the different questions and answers can be divided
3658 in sections depending on the subject.
3661 *) Have the zlib compression code load ZLIB.DLL dynamically under
3665 *) New function BN_mod_sqrt for computing square roots modulo a prime
3666 (using the probabilistic Tonelli-Shanks algorithm unless
3667 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3668 be handled deterministically).
3669 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3671 *) Make BN_mod_inverse faster by explicitly handling small quotients
3672 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3673 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3676 *) New function BN_kronecker.
3679 *) Fix BN_gcd so that it works on negative inputs; the result is
3680 positive unless both parameters are zero.
3681 Previously something reasonably close to an infinite loop was
3682 possible because numbers could be growing instead of shrinking
3683 in the implementation of Euclid's algorithm.
3686 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3687 sign of the number in question.
3689 Fix BN_is_word(a,w) to work correctly for w == 0.
3691 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3692 because its test if the absolute value of 'a' equals 'w'.
3693 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3694 it exists mostly for use in the implementations of BN_is_zero(),
3695 BN_is_one(), and BN_is_word().
3698 *) New function BN_swap.
3701 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3702 the exponentiation functions are more likely to produce reasonable
3703 results on negative inputs.
3706 *) Change BN_mod_mul so that the result is always non-negative.
3707 Previously, it could be negative if one of the factors was negative;
3708 I don't think anyone really wanted that behaviour.
3711 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3712 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3713 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3714 and add new functions:
3723 BN_mod_lshift1_quick
3727 These functions always generate non-negative results.
3729 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3730 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3732 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3733 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3734 be reduced modulo m.
3735 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3738 The following entry accidentily appeared in the CHANGES file
3739 distributed with OpenSSL 0.9.7. The modifications described in
3740 it do *not* apply to OpenSSL 0.9.7.
3742 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3743 was actually never needed) and in BN_mul(). The removal in BN_mul()
3744 required a small change in bn_mul_part_recursive() and the addition
3745 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3746 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3747 bn_sub_words() and bn_add_words() except they take arrays with
3752 *) In 'openssl passwd', verify passwords read from the terminal
3753 unless the '-salt' option is used (which usually means that
3754 verification would just waste user's time since the resulting
3755 hash is going to be compared with some given password hash)
3756 or the new '-noverify' option is used.
3758 This is an incompatible change, but it does not affect
3759 non-interactive use of 'openssl passwd' (passwords on the command
3760 line, '-stdin' option, '-in ...' option) and thus should not
3764 *) Remove all references to RSAref, since there's no more need for it.
3767 *) Make DSO load along a path given through an environment variable
3768 (SHLIB_PATH) with shl_load().
3771 *) Constify the ENGINE code as a result of BIGNUM constification.
3772 Also constify the RSA code and most things related to it. In a
3773 few places, most notable in the depth of the ASN.1 code, ugly
3774 casts back to non-const were required (to be solved at a later
3778 *) Make it so the openssl application has all engines loaded by default.
3781 *) Constify the BIGNUM routines a little more.
3784 *) Add the following functions:
3786 ENGINE_load_cswift()
3788 ENGINE_load_atalla()
3790 ENGINE_load_builtin_engines()
3792 That way, an application can itself choose if external engines that
3793 are built-in in OpenSSL shall ever be used or not. The benefit is
3794 that applications won't have to be linked with libdl or other dso
3795 libraries unless it's really needed.
3797 Changed 'openssl engine' to load all engines on demand.
3798 Changed the engine header files to avoid the duplication of some
3799 declarations (they differed!).
3802 *) 'openssl engine' can now list capabilities.
3805 *) Better error reporting in 'openssl engine'.
3808 *) Never call load_dh_param(NULL) in s_server.
3811 *) Add engine application. It can currently list engines by name and
3812 identity, and test if they are actually available.
3815 *) Improve RPM specification file by forcing symbolic linking and making
3816 sure the installed documentation is also owned by root.root.
3817 [Damien Miller <djm@mindrot.org>]
3819 *) Give the OpenSSL applications more possibilities to make use of
3820 keys (public as well as private) handled by engines.
3823 *) Add OCSP code that comes from CertCo.
3826 *) Add VMS support for the Rijndael code.
3829 *) Added untested support for Nuron crypto accelerator.
3832 *) Add support for external cryptographic devices. This code was
3833 previously distributed separately as the "engine" branch.
3834 [Geoff Thorpe, Richard Levitte]
3836 *) Rework the filename-translation in the DSO code. It is now possible to
3837 have far greater control over how a "name" is turned into a filename
3838 depending on the operating environment and any oddities about the
3839 different shared library filenames on each system.
3842 *) Support threads on FreeBSD-elf in Configure.
3845 *) Fix for SHA1 assembly problem with MASM: it produces
3846 warnings about corrupt line number information when assembling
3847 with debugging information. This is caused by the overlapping
3849 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3852 NCONF_get_number() has no error checking at all. As a replacement,
3853 NCONF_get_number_e() is defined (_e for "error checking") and is
3854 promoted strongly. The old NCONF_get_number is kept around for
3855 binary backward compatibility.
3856 Make it possible for methods to load from something other than a BIO,
3857 by providing a function pointer that is given a name instead of a BIO.
3858 For example, this could be used to load configuration data from an
3862 *) Fix for non blocking accept BIOs. Added new I/O special reason
3863 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3864 with non blocking I/O was not possible because no retry code was
3865 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3869 *) Added the beginnings of Rijndael support.
3872 *) Fix for bug in DirectoryString mask setting. Add support for
3873 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3874 to allow certificate printing to more controllable, additional
3875 'certopt' option to 'x509' to allow new printing options to be
3879 *) Clean old EAY MD5 hack from e_os.h.
3882 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3884 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3885 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3886 [Joe Orton, Steve Henson]
3888 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3890 *) Fix additional bug revealed by the NISCC test suite:
3892 Stop bug triggering large recursion when presented with
3893 certain ASN.1 tags (CVE-2003-0851)
3896 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3898 *) Fix various bugs revealed by running the NISCC test suite:
3900 Stop out of bounds reads in the ASN1 code when presented with
3901 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3903 If verify callback ignores invalid public key errors don't try to check
3904 certificate signature with the NULL public key.
3908 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3909 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3913 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3914 extra data after the compression methods not only for TLS 1.0
3915 but also for SSL 3.0 (as required by the specification).
3916 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3918 *) Change X509_certificate_type() to mark the key as exported/exportable
3919 when it's 512 *bits* long, not 512 bytes.
3922 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3924 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3925 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3926 a protocol version number mismatch like a decryption error
3927 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3930 *) Turn on RSA blinding by default in the default implementation
3931 to avoid a timing attack. Applications that don't want it can call
3932 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3933 They would be ill-advised to do so in most cases.
3934 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3936 *) Change RSA blinding code so that it works when the PRNG is not
3937 seeded (in this case, the secret RSA exponent is abused as
3938 an unpredictable seed -- if it is not unpredictable, there
3939 is no point in blinding anyway). Make RSA blinding thread-safe
3940 by remembering the creator's thread ID in rsa->blinding and
3941 having all other threads use local one-time blinding factors
3942 (this requires more computation than sharing rsa->blinding, but
3943 avoids excessive locking; and if an RSA object is not shared
3944 between threads, blinding will still be very fast).
3947 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3949 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3950 via timing by performing a MAC computation even if incorrrect
3951 block cipher padding has been found. This is a countermeasure
3952 against active attacks where the attacker has to distinguish
3953 between bad padding and a MAC verification error. (CVE-2003-0078)
3955 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3956 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3957 Martin Vuagnoux (EPFL, Ilion)]
3959 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3961 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3962 memory from it's contents. This is done with a counter that will
3963 place alternating values in each byte. This can be used to solve
3964 two issues: 1) the removal of calls to memset() by highly optimizing
3965 compilers, and 2) cleansing with other values than 0, since those can
3966 be read through on certain media, for example a swap space on disk.
3969 *) Bugfix: client side session caching did not work with external caching,
3970 because the session->cipher setting was not restored when reloading
3971 from the external cache. This problem was masked, when
3972 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3973 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3976 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3977 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3978 [Zeev Lieber <zeev-l@yahoo.com>]
3980 *) Undo an undocumented change introduced in 0.9.6e which caused
3981 repeated calls to OpenSSL_add_all_ciphers() and
3982 OpenSSL_add_all_digests() to be ignored, even after calling
3986 *) Change the default configuration reader to deal with last line not
3987 being properly terminated.
3990 *) Change X509_NAME_cmp() so it applies the special rules on handling
3991 DN values that are of type PrintableString, as well as RDNs of type
3992 emailAddress where the value has the type ia5String.
3993 [stefank@valicert.com via Richard Levitte]
3995 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3996 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3997 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3998 the bitwise-OR of the two for use by the majority of applications
3999 wanting this behaviour, and update the docs. The documented
4000 behaviour and actual behaviour were inconsistent and had been
4001 changing anyway, so this is more a bug-fix than a behavioural
4003 [Geoff Thorpe, diagnosed by Nadav Har'El]
4005 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4006 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4009 *) Fix initialization code race conditions in
4010 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4011 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4012 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4013 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4014 ssl2_get_cipher_by_char(),
4015 ssl3_get_cipher_by_char().
4016 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4018 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4019 the cached sessions are flushed, as the remove_cb() might use ex_data
4020 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4021 (see [openssl.org #212]).
4022 [Geoff Thorpe, Lutz Jaenicke]
4024 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4025 length, instead of the encoding length to d2i_ASN1_OBJECT.
4028 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4030 *) [In 0.9.6g-engine release:]
4031 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4032 [Lynn Gazis <lgazis@rainbow.com>]
4034 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4036 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4037 and get fix the header length calculation.
4038 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4039 Alon Kantor <alonk@checkpoint.com> (and others),
4042 *) Use proper error handling instead of 'assertions' in buffer
4043 overflow checks added in 0.9.6e. This prevents DoS (the
4044 assertions could call abort()).
4045 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4047 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4049 *) Add various sanity checks to asn1_get_length() to reject
4050 the ASN1 length bytes if they exceed sizeof(long), will appear
4051 negative or the content length exceeds the length of the
4053 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4055 *) Fix cipher selection routines: ciphers without encryption had no flags
4056 for the cipher strength set and where therefore not handled correctly
4057 by the selection routines (PR #130).
4060 *) Fix EVP_dsa_sha macro.
4064 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4065 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4066 that was added in OpenSSL 0.9.6d.
4068 As the countermeasure turned out to be incompatible with some
4069 broken SSL implementations, the new option is part of SSL_OP_ALL.
4070 SSL_OP_ALL is usually employed when compatibility with weird SSL
4071 implementations is desired (e.g. '-bugs' option to 's_client' and
4072 's_server'), so the new option is automatically set in many
4076 *) Changes in security patch:
4078 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4079 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4080 Air Force Materiel Command, USAF, under agreement number
4083 *) Add various sanity checks to asn1_get_length() to reject
4084 the ASN1 length bytes if they exceed sizeof(long), will appear
4085 negative or the content length exceeds the length of the
4086 supplied buffer. (CVE-2002-0659)
4087 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4089 *) Assertions for various potential buffer overflows, not known to
4091 [Ben Laurie (CHATS)]
4093 *) Various temporary buffers to hold ASCII versions of integers were
4094 too small for 64 bit platforms. (CVE-2002-0655)
4095 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4097 *) Remote buffer overflow in SSL3 protocol - an attacker could
4098 supply an oversized session ID to a client. (CVE-2002-0656)
4099 [Ben Laurie (CHATS)]
4101 *) Remote buffer overflow in SSL2 protocol - an attacker could
4102 supply an oversized client master key. (CVE-2002-0656)
4103 [Ben Laurie (CHATS)]
4105 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4107 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4108 encoded as NULL) with id-dsa-with-sha1.
4109 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4111 *) Check various X509_...() return values in apps/req.c.
4112 [Nils Larsch <nla@trustcenter.de>]
4114 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4115 an end-of-file condition would erronously be flagged, when the CRLF
4116 was just at the end of a processed block. The bug was discovered when
4117 processing data through a buffering memory BIO handing the data to a
4118 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4119 <ptsekov@syntrex.com> and Nedelcho Stanev.
4122 *) Implement a countermeasure against a vulnerability recently found
4123 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4124 before application data chunks to avoid the use of known IVs
4125 with data potentially chosen by the attacker.
4128 *) Fix length checks in ssl3_get_client_hello().
4131 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4132 to prevent ssl3_read_internal() from incorrectly assuming that
4133 ssl3_read_bytes() found application data while handshake
4134 processing was enabled when in fact s->s3->in_read_app_data was
4135 merely automatically cleared during the initial handshake.
4136 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4138 *) Fix object definitions for Private and Enterprise: they were not
4139 recognized in their shortname (=lowercase) representation. Extend
4140 obj_dat.pl to issue an error when using undefined keywords instead
4141 of silently ignoring the problem (Svenning Sorensen
4142 <sss@sss.dnsalias.net>).
4145 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4146 generators, i.e. generators other than 2 and 5. (Previously, the
4147 code did not properly initialise the 'add' and 'rem' values to
4148 BN_generate_prime().)
4150 In the new general case, we do not insist that 'generator' is
4151 actually a primitive root: This requirement is rather pointless;
4152 a generator of the order-q subgroup is just as good, if not
4156 *) Map new X509 verification errors to alerts. Discovered and submitted by
4157 Tom Wu <tom@arcot.com>.
4160 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4161 returning non-zero before the data has been completely received
4162 when using non-blocking I/O.
4163 [Bodo Moeller; problem pointed out by John Hughes]
4165 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4166 [Ben Laurie, Lutz Jaenicke]
4168 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4169 Yoram Zahavi <YoramZ@gilian.com>).
4172 *) Add information about CygWin 1.3 and on, and preserve proper
4173 configuration for the versions before that.
4174 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4176 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4177 check whether we deal with a copy of a session and do not delete from
4178 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4179 <izhar@checkpoint.com>.
4182 *) Do not store session data into the internal session cache, if it
4183 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4184 flag is set). Proposed by Aslam <aslam@funk.com>.
4187 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4191 *) [In 0.9.6d-engine release:]
4192 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4193 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4195 *) Add the configuration target linux-s390x.
4196 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4198 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4199 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4200 variable as an indication that a ClientHello message has been
4201 received. As the flag value will be lost between multiple
4202 invocations of ssl3_accept when using non-blocking I/O, the
4203 function may not be aware that a handshake has actually taken
4204 place, thus preventing a new session from being added to the
4207 To avoid this problem, we now set s->new_session to 2 instead of
4208 using a local variable.
4209 [Lutz Jaenicke, Bodo Moeller]
4211 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4212 if the SSL_R_LENGTH_MISMATCH error is detected.
4213 [Geoff Thorpe, Bodo Moeller]
4215 *) New 'shared_ldflag' column in Configure platform table.
4218 *) Fix EVP_CIPHER_mode macro.
4219 ["Dan S. Camper" <dan@bti.net>]
4221 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4222 type, we must throw them away by setting rr->length to 0.
4223 [D P Chang <dpc@qualys.com>]
4225 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4227 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4228 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4229 worked incorrectly for those cases where range = 10..._2 and
4230 3*range is two bits longer than range.)
4233 *) Only add signing time to PKCS7 structures if it is not already
4237 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4238 OBJ_ld_ce should be OBJ_id_ce.
4239 Also some ip-pda OIDs in crypto/objects/objects.txt were
4240 incorrect (cf. RFC 3039).
4241 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4243 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4244 returns early because it has nothing to do.
4245 [Andy Schneider <andy.schneider@bjss.co.uk>]
4247 *) [In 0.9.6c-engine release:]
4248 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4249 [Andy Schneider <andy.schneider@bjss.co.uk>]
4251 *) [In 0.9.6c-engine release:]
4252 Add support for Cryptographic Appliance's keyserver technology.
4253 (Use engine 'keyclient')
4254 [Cryptographic Appliances and Geoff Thorpe]
4256 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4257 is called via tools/c89.sh because arguments have to be
4258 rearranged (all '-L' options must appear before the first object
4260 [Richard Shapiro <rshapiro@abinitio.com>]
4262 *) [In 0.9.6c-engine release:]
4263 Add support for Broadcom crypto accelerator cards, backported
4265 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4267 *) [In 0.9.6c-engine release:]
4268 Add support for SureWare crypto accelerator cards from
4269 Baltimore Technologies. (Use engine 'sureware')
4270 [Baltimore Technologies and Mark Cox]
4272 *) [In 0.9.6c-engine release:]
4273 Add support for crypto accelerator cards from Accelerated
4274 Encryption Processing, www.aep.ie. (Use engine 'aep')
4275 [AEP Inc. and Mark Cox]
4277 *) Add a configuration entry for gcc on UnixWare.
4278 [Gary Benson <gbenson@redhat.com>]
4280 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4281 messages are stored in a single piece (fixed-length part and
4282 variable-length part combined) and fix various bugs found on the way.
4285 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4286 instead. BIO_gethostbyname() does not know what timeouts are
4287 appropriate, so entries would stay in cache even when they have
4289 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4291 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4292 faced with a pathologically small ClientHello fragment that does
4293 not contain client_version: Instead of aborting with an error,
4294 simply choose the highest available protocol version (i.e.,
4295 TLS 1.0 unless it is disabled). In practice, ClientHello
4296 messages are never sent like this, but this change gives us
4297 strictly correct behaviour at least for TLS.
4300 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4301 never resets s->method to s->ctx->method when called from within
4302 one of the SSL handshake functions.
4303 [Bodo Moeller; problem pointed out by Niko Baric]
4305 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4306 (sent using the client's version number) if client_version is
4307 smaller than the protocol version in use. Also change
4308 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4309 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4310 the client will at least see that alert.
4313 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4317 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4318 client receives HelloRequest while in a handshake.
4319 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4321 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4322 should end in 'break', not 'goto end' which circuments various
4323 cleanups done in state SSL_ST_OK. But session related stuff
4324 must be disabled for SSL_ST_OK in the case that we just sent a
4327 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4328 before just sending a HelloRequest.
4329 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4331 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4332 reveal whether illegal block cipher padding was found or a MAC
4333 verification error occured. (Neither SSLerr() codes nor alerts
4334 are directly visible to potential attackers, but the information
4335 may leak via logfiles.)
4337 Similar changes are not required for the SSL 2.0 implementation
4338 because the number of padding bytes is sent in clear for SSL 2.0,
4339 and the extra bytes are just ignored. However ssl/s2_pkt.c
4340 failed to verify that the purported number of padding bytes is in
4344 *) Add OpenUNIX-8 support including shared libraries
4345 (Boyd Lynn Gerber <gerberb@zenez.com>).
4348 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4349 'wristwatch attack' using huge encoding parameters (cf.
4350 James H. Manger's CRYPTO 2001 paper). Note that the
4351 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4352 encoding parameters and hence was not vulnerable.
4355 *) BN_sqr() bug fix.
4356 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4358 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4359 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4360 followed by modular reduction.
4361 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4363 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4364 equivalent based on BN_pseudo_rand() instead of BN_rand().
4367 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4368 This function was broken, as the check for a new client hello message
4369 to handle SGC did not allow these large messages.
4370 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4373 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4376 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4377 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4380 *) Rework the configuration and shared library support for Tru64 Unix.
4381 The configuration part makes use of modern compiler features and
4382 still retains old compiler behavior for those that run older versions
4383 of the OS. The shared library support part includes a variant that
4384 uses the RPATH feature, and is available through the special
4385 configuration target "alpha-cc-rpath", which will never be selected
4387 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4389 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4390 with the same message size as in ssl3_get_certificate_request().
4391 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4392 messages might inadvertently be reject as too long.
4393 [Petr Lampa <lampa@fee.vutbr.cz>]
4395 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4398 *) Modified SSL library such that the verify_callback that has been set
4399 specificly for an SSL object with SSL_set_verify() is actually being
4400 used. Before the change, a verify_callback set with this function was
4401 ignored and the verify_callback() set in the SSL_CTX at the time of
4402 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4403 to allow the necessary settings.
4406 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4407 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4408 done automatically (in contradiction to the requirements of the C
4409 standard). This made problems when used from OpenSSH.
4412 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4413 dh->length and always used
4415 BN_rand_range(priv_key, dh->p).
4417 BN_rand_range() is not necessary for Diffie-Hellman, and this
4418 specific range makes Diffie-Hellman unnecessarily inefficient if
4419 dh->length (recommended exponent length) is much smaller than the
4420 length of dh->p. We could use BN_rand_range() if the order of
4421 the subgroup was stored in the DH structure, but we only have
4426 BN_rand(priv_key, l, ...)
4428 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4434 RSA_eay_public_encrypt
4435 RSA_eay_private_decrypt
4436 RSA_eay_private_encrypt (signing)
4437 RSA_eay_public_decrypt (signature verification)
4439 (default implementations for RSA_public_encrypt,
4440 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4441 always reject numbers >= n.
4444 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4445 to synchronize access to 'locking_thread'. This is necessary on
4446 systems where access to 'locking_thread' (an 'unsigned long'
4447 variable) is not atomic.
4450 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4451 *before* setting the 'crypto_lock_rand' flag. The previous code had
4452 a race condition if 0 is a valid thread ID.
4453 [Travis Vitek <vitek@roguewave.com>]
4455 *) Add support for shared libraries under Irix.
4456 [Albert Chin-A-Young <china@thewrittenword.com>]
4458 *) Add configuration option to build on Linux on both big-endian and
4460 [Ralf Baechle <ralf@uni-koblenz.de>]
4462 *) Add the possibility to create shared libraries on HP-UX.
4465 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4467 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4468 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4469 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4470 PRNG state recovery was possible based on the output of
4471 one PRNG request appropriately sized to gain knowledge on
4472 'md' followed by enough consecutive 1-byte PRNG requests
4473 to traverse all of 'state'.
4475 1. When updating 'md_local' (the current thread's copy of 'md')
4476 during PRNG output generation, hash all of the previous
4477 'md_local' value, not just the half used for PRNG output.
4479 2. Make the number of bytes from 'state' included into the hash
4480 independent from the number of PRNG bytes requested.
4482 The first measure alone would be sufficient to avoid
4483 Markku-Juhani's attack. (Actually it had never occurred
4484 to me that the half of 'md_local' used for chaining was the
4485 half from which PRNG output bytes were taken -- I had always
4486 assumed that the secret half would be used.) The second
4487 measure makes sure that additional data from 'state' is never
4488 mixed into 'md_local' in small portions; this heuristically
4489 further strengthens the PRNG.
4492 *) Fix crypto/bn/asm/mips3.s.
4495 *) When only the key is given to "enc", the IV is undefined. Print out
4496 an error message in this case.
4499 *) Handle special case when X509_NAME is empty in X509 printing routines.
4502 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4503 positive and less than q.
4506 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4507 used: it isn't thread safe and the add_lock_callback should handle
4509 [Paul Rose <Paul.Rose@bridge.com>]
4511 *) Verify that incoming data obeys the block size in
4512 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4516 [Ulf Möller, Bodo Möller]
4518 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4519 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4520 when fixing the server behaviour for backwards-compatible 'client
4521 hello' messages. (Note that the attack is impractical against
4522 SSL 3.0 and TLS 1.0 anyway because length and version checking
4523 means that the probability of guessing a valid ciphertext is
4524 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4527 Before 0.9.5, the countermeasure (hide the error by generating a
4528 random 'decryption result') did not work properly because
4529 ERR_clear_error() was missing, meaning that SSL_get_error() would
4530 detect the supposedly ignored error.
4532 Both problems are now fixed.
4535 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4536 (previously it was 1024).
4539 *) Fix for compatibility mode trust settings: ignore trust settings
4540 unless some valid trust or reject settings are present.
4543 *) Fix for blowfish EVP: its a variable length cipher.
4546 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4547 parameters in DSA public key structures and return an error in the
4548 DSA routines if parameters are absent.
4551 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4552 in the current directory if neither $RANDFILE nor $HOME was set.
4553 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4554 caused some confusion to Windows users who haven't defined $HOME.
4555 Thus RAND_file_name() is changed again: e_os.h can define a
4556 DEFAULT_HOME, which will be used if $HOME is not set.
4557 For Windows, we use "C:"; on other platforms, we still require
4558 environment variables.
4560 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4561 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4562 having multiple threads call RAND_poll() concurrently.
4565 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4566 combination of a flag and a thread ID variable.
4567 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4568 flag), *other* threads can enter ssleay_add_bytes without obeying
4569 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4570 that they do not hold after the first thread unsets add_do_not_lock).
4573 *) Change bctest again: '-x' expressions are not available in all
4577 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4579 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4580 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4582 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4583 the default extension for executables, if any. Also, make the perl
4584 scripts that use symlink() to test if it really exists and use "cp"
4585 if it doesn't. All this made OpenSSL compilable and installable in
4589 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4590 If SEQUENCE is length is indefinite just set c->slen to the total
4591 amount of data available.
4592 [Steve Henson, reported by shige@FreeBSD.org]
4593 [This change does not apply to 0.9.7.]
4595 *) Change bctest to avoid here-documents inside command substitution
4596 (workaround for FreeBSD /bin/sh bug).
4597 For compatibility with Ultrix, avoid shell functions (introduced
4598 in the bctest version that searches along $PATH).
4601 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4602 with des_encrypt() defined on some operating systems, like Solaris
4606 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4607 On the Importance of Eliminating Errors in Cryptographic
4608 Computations, J. Cryptology 14 (2001) 2, 101-119,
4609 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4612 *) MIPS assembler BIGNUM division bug fix.
4615 *) Disabled incorrect Alpha assembler code.
4618 *) Fix PKCS#7 decode routines so they correctly update the length
4619 after reading an EOC for the EXPLICIT tag.
4621 [This change does not apply to 0.9.7.]
4623 *) Fix bug in PKCS#12 key generation routines. This was triggered
4624 if a 3DES key was generated with a 0 initial byte. Include
4625 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4626 (but broken) behaviour.
4629 *) Enhance bctest to search for a working bc along $PATH and print
4631 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4633 *) Fix memory leaks in err.c: free err_data string if necessary;
4634 don't write to the wrong index in ERR_set_error_data.
4637 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4641 *) Replace rdtsc with _emit statements for VC++ version 5.
4642 [Jeremy Cooper <jeremy@baymoo.org>]
4644 *) Make it possible to reuse SSLv2 sessions.
4647 *) In copy_email() check for >= 0 as a return value for
4648 X509_NAME_get_index_by_NID() since 0 is a valid index.
4649 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4651 *) Avoid coredump with unsupported or invalid public keys by checking if
4652 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4653 PKCS7_verify() fails with non detached data.
4656 *) Don't use getenv in library functions when run as setuid/setgid.
4657 New function OPENSSL_issetugid().
4660 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4661 due to incorrect handling of multi-threading:
4663 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4665 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4667 3. Count how many times MemCheck_off() has been called so that
4668 nested use can be treated correctly. This also avoids
4669 inband-signalling in the previous code (which relied on the
4670 assumption that thread ID 0 is impossible).
4673 *) Add "-rand" option also to s_client and s_server.
4676 *) Fix CPU detection on Irix 6.x.
4677 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4678 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4680 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4683 [This change does not apply to 0.9.7.]
4685 *) Use the cached encoding of an X509_NAME structure rather than
4686 copying it. This is apparently the reason for the libsafe "errors"
4687 but the code is actually correct.
4690 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4691 Bleichenbacher's DSA attack.
4692 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4693 to be set and top=0 forces the highest bit to be set; top=-1 is new
4694 and leaves the highest bit random.
4695 [Ulf Moeller, Bodo Moeller]
4697 *) In the NCONF_...-based implementations for CONF_... queries
4698 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4699 a temporary CONF structure with the data component set to NULL
4700 (which gives segmentation faults in lh_retrieve).
4701 Instead, use NULL for the CONF pointer in CONF_get_string and
4702 CONF_get_number (which may use environment variables) and directly
4703 return NULL from CONF_get_section.
4706 *) Fix potential buffer overrun for EBCDIC.
4709 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4710 keyUsage if basicConstraints absent for a CA.
4713 *) Make SMIME_write_PKCS7() write mail header values with a format that
4714 is more generally accepted (no spaces before the semicolon), since
4715 some programs can't parse those values properly otherwise. Also make
4716 sure BIO's that break lines after each write do not create invalid
4720 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4721 macros previously used would not encode an empty SEQUENCE OF
4722 and break the signature.
4724 [This change does not apply to 0.9.7.]
4726 *) Zero the premaster secret after deriving the master secret in
4730 *) Add some EVP_add_digest_alias registrations (as found in
4731 OpenSSL_add_all_digests()) to SSL_library_init()
4732 aka OpenSSL_add_ssl_algorithms(). This provides improved
4733 compatibility with peers using X.509 certificates
4734 with unconventional AlgorithmIdentifier OIDs.
4737 *) Fix for Irix with NO_ASM.
4738 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4740 *) ./config script fixes.
4741 [Ulf Moeller, Richard Levitte]
4743 *) Fix 'openssl passwd -1'.
4746 *) Change PKCS12_key_gen_asc() so it can cope with non null
4747 terminated strings whose length is passed in the passlen
4748 parameter, for example from PEM callbacks. This was done
4749 by adding an extra length parameter to asc2uni().
4750 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4752 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4753 call failed, free the DSA structure.
4756 *) Fix to uni2asc() to cope with zero length Unicode strings.
4757 These are present in some PKCS#12 files.
4760 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4761 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4762 when writing a 32767 byte record.
4763 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4765 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4766 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4768 (RSA objects have a reference count access to which is protected
4769 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4770 so they are meant to be shared between threads.)
4771 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4772 "Reddie, Steven" <Steven.Reddie@ca.com>]
4774 *) Fix a deadlock in CRYPTO_mem_leaks().
4777 *) Use better test patterns in bntest.
4780 *) rand_win.c fix for Borland C.
4783 *) BN_rshift bugfix for n == 0.
4786 *) Add a 'bctest' script that checks for some known 'bc' bugs
4787 so that 'make test' does not abort just because 'bc' is broken.
4790 *) Store verify_result within SSL_SESSION also for client side to
4791 avoid potential security hole. (Re-used sessions on the client side
4792 always resulted in verify_result==X509_V_OK, not using the original
4793 result of the server certificate verification.)
4796 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4797 SSL3_RT_APPLICATION_DATA, return 0.
4798 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4802 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4803 releases, have been re-implemented by renaming the previous
4804 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4805 and ssl3_read_internal, respectively, and adding 'peek' parameters
4806 to them. The new ssl[23]_{read,peek} functions are calls to
4807 ssl[23]_read_internal with the 'peek' flag set appropriately.
4808 A 'peek' parameter has also been added to ssl3_read_bytes, which
4809 does the actual work for ssl3_read_internal.
4812 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4813 the method-specific "init()" handler. Also clean up ex_data after
4814 calling the method-specific "finish()" handler. Previously, this was
4815 happening the other way round.
4818 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4819 The previous value, 12, was not always sufficient for BN_mod_exp().
4822 *) Make sure that shared libraries get the internal name engine with
4823 the full version number and not just 0. This should mark the
4824 shared libraries as not backward compatible. Of course, this should
4825 be changed again when we can guarantee backward binary compatibility.
4828 *) Fix typo in get_cert_by_subject() in by_dir.c
4829 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4831 *) Rework the system to generate shared libraries:
4833 - Make note of the expected extension for the shared libraries and
4834 if there is a need for symbolic links from for example libcrypto.so.0
4835 to libcrypto.so.0.9.7. There is extended info in Configure for
4838 - Make as few rebuilds of the shared libraries as possible.
4840 - Still avoid linking the OpenSSL programs with the shared libraries.
4842 - When installing, install the shared libraries separately from the
4846 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4848 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4849 and not in SSL_clear because the latter is also used by the
4850 accept/connect functions; previously, the settings made by
4851 SSL_set_read_ahead would be lost during the handshake.
4852 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4854 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4855 Previously, it would create entries for disableed algorithms no
4859 *) Added several new manual pages for SSL_* function.
4862 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4864 *) In ssl23_get_client_hello, generate an error message when faced
4865 with an initial SSL 3.0/TLS record that is too small to contain the
4866 first two bytes of the ClientHello message, i.e. client_version.
4867 (Note that this is a pathologic case that probably has never happened
4868 in real life.) The previous approach was to use the version number
4869 from the record header as a substitute; but our protocol choice
4870 should not depend on that one because it is not authenticated
4871 by the Finished messages.
4874 *) More robust randomness gathering functions for Windows.
4875 [Jeffrey Altman <jaltman@columbia.edu>]
4877 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4878 not set then we don't setup the error code for issuer check errors
4879 to avoid possibly overwriting other errors which the callback does
4880 handle. If an application does set the flag then we assume it knows
4881 what it is doing and can handle the new informational codes
4885 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4886 a general "ANY" type, as such it should be able to decode anything
4887 including tagged types. However it didn't check the class so it would
4888 wrongly interpret tagged types in the same way as their universal
4889 counterpart and unknown types were just rejected. Changed so that the
4890 tagged and unknown types are handled in the same way as a SEQUENCE:
4891 that is the encoding is stored intact. There is also a new type
4892 "V_ASN1_OTHER" which is used when the class is not universal, in this
4893 case we have no idea what the actual type is so we just lump them all
4897 *) On VMS, stdout may very well lead to a file that is written to
4898 in a record-oriented fashion. That means that every write() will
4899 write a separate record, which will be read separately by the
4900 programs trying to read from it. This can be very confusing.
4902 The solution is to put a BIO filter in the way that will buffer
4903 text until a linefeed is reached, and then write everything a
4904 line at a time, so every record written will be an actual line,
4905 not chunks of lines and not (usually doesn't happen, but I've
4906 seen it once) several lines in one record. BIO_f_linebuffer() is
4909 Currently, it's a VMS-only method, because that's where it has
4910 been tested well enough.
4913 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4914 it can return incorrect results.
4915 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4916 but it was in 0.9.6-beta[12].)
4919 *) Disable the check for content being present when verifying detached
4920 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4921 include zero length content when signing messages.
4924 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4925 BIO_ctrl (for BIO pairs).
4928 *) Add DSO method for VMS.
4931 *) Bug fix: Montgomery multiplication could produce results with the
4935 *) Add RPM specification openssl.spec and modify it to build three
4936 packages. The default package contains applications, application
4937 documentation and run-time libraries. The devel package contains
4938 include files, static libraries and function documentation. The
4939 doc package contains the contents of the doc directory. The original
4940 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4943 *) Add a large number of documentation files for many SSL routines.
4944 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4946 *) Add a configuration entry for Sony News 4.
4947 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4949 *) Don't set the two most significant bits to one when generating a
4950 random number < q in the DSA library.
4953 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4954 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4955 the underlying transport is blocking) if a handshake took place.
4956 (The default behaviour is needed by applications such as s_client
4957 and s_server that use select() to determine when to use SSL_read;
4958 but for applications that know in advance when to expect data, it
4959 just makes things more complicated.)
4962 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4966 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4967 work better on such systems.
4968 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4970 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4971 Update PKCS12_parse() so it copies the friendlyName and the
4972 keyid to the certificates aux info.
4975 *) Fix bug in PKCS7_verify() which caused an infinite loop
4976 if there was more than one signature.
4977 [Sven Uszpelkat <su@celocom.de>]
4979 *) Major change in util/mkdef.pl to include extra information
4980 about each symbol, as well as presentig variables as well
4981 as functions. This change means that there's n more need
4982 to rebuild the .num files when some algorithms are excluded.
4985 *) Allow the verify time to be set by an application,
4986 rather than always using the current time.
4989 *) Phase 2 verify code reorganisation. The certificate
4990 verify code now looks up an issuer certificate by a
4991 number of criteria: subject name, authority key id
4992 and key usage. It also verifies self signed certificates
4993 by the same criteria. The main comparison function is
4994 X509_check_issued() which performs these checks.
4996 Lot of changes were necessary in order to support this
4997 without completely rewriting the lookup code.
4999 Authority and subject key identifier are now cached.
5001 The LHASH 'certs' is X509_STORE has now been replaced
5002 by a STACK_OF(X509_OBJECT). This is mainly because an
5003 LHASH can't store or retrieve multiple objects with
5004 the same hash value.
5006 As a result various functions (which were all internal
5007 use only) have changed to handle the new X509_STORE
5008 structure. This will break anything that messed round
5009 with X509_STORE internally.
5011 The functions X509_STORE_add_cert() now checks for an
5012 exact match, rather than just subject name.
5014 The X509_STORE API doesn't directly support the retrieval
5015 of multiple certificates matching a given criteria, however
5016 this can be worked round by performing a lookup first
5017 (which will fill the cache with candidate certificates)
5018 and then examining the cache for matches. This is probably
5019 the best we can do without throwing out X509_LOOKUP
5020 entirely (maybe later...).
5022 The X509_VERIFY_CTX structure has been enhanced considerably.
5024 All certificate lookup operations now go via a get_issuer()
5025 callback. Although this currently uses an X509_STORE it
5026 can be replaced by custom lookups. This is a simple way
5027 to bypass the X509_STORE hackery necessary to make this
5028 work and makes it possible to use more efficient techniques
5029 in future. A very simple version which uses a simple
5030 STACK for its trusted certificate store is also provided
5031 using X509_STORE_CTX_trusted_stack().
5033 The verify_cb() and verify() callbacks now have equivalents
5034 in the X509_STORE_CTX structure.
5036 X509_STORE_CTX also has a 'flags' field which can be used
5037 to customise the verify behaviour.
5040 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5041 excludes S/MIME capabilities.
5044 *) When a certificate request is read in keep a copy of the
5045 original encoding of the signed data and use it when outputing
5046 again. Signatures then use the original encoding rather than
5047 a decoded, encoded version which may cause problems if the
5048 request is improperly encoded.
5051 *) For consistency with other BIO_puts implementations, call
5052 buffer_write(b, ...) directly in buffer_puts instead of calling
5055 In BIO_puts, increment b->num_write as in BIO_write.
5056 [Peter.Sylvester@EdelWeb.fr]
5058 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5059 BN_zero, we may not return a BIGNUM with an array consisting of
5063 *) Avoid calling abort() from within the library when problems are
5064 detected, except if preprocessor symbols have been defined
5065 (such as REF_CHECK, BN_DEBUG etc.).
5068 *) New openssl application 'rsautl'. This utility can be
5069 used for low level RSA operations. DER public key
5070 BIO/fp routines also added.
5073 *) New Configure entry and patches for compiling on QNX 4.
5074 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5076 *) A demo state-machine implementation was sponsored by
5077 Nuron (http://www.nuron.com/) and is now available in
5078 demos/state_machine.
5081 *) New options added to the 'dgst' utility for signature
5082 generation and verification.
5085 *) Unrecognized PKCS#7 content types are now handled via a
5086 catch all ASN1_TYPE structure. This allows unsupported
5087 types to be stored as a "blob" and an application can
5088 encode and decode it manually.
5091 *) Fix various signed/unsigned issues to make a_strex.c
5093 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5095 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5096 length if passed a buffer. ASN1_INTEGER_to_BN failed
5097 if passed a NULL BN and its argument was negative.
5098 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5100 *) Modification to PKCS#7 encoding routines to output definite
5101 length encoding. Since currently the whole structures are in
5102 memory there's not real point in using indefinite length
5103 constructed encoding. However if OpenSSL is compiled with
5104 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5107 *) Added BIO_vprintf() and BIO_vsnprintf().
5110 *) Added more prefixes to parse for in the the strings written
5111 through a logging bio, to cover all the levels that are available
5112 through syslog. The prefixes are now:
5114 PANIC, EMERG, EMR => LOG_EMERG
5115 ALERT, ALR => LOG_ALERT
5116 CRIT, CRI => LOG_CRIT
5117 ERROR, ERR => LOG_ERR
5118 WARNING, WARN, WAR => LOG_WARNING
5119 NOTICE, NOTE, NOT => LOG_NOTICE
5120 INFO, INF => LOG_INFO
5121 DEBUG, DBG => LOG_DEBUG
5123 and as before, if none of those prefixes are present at the
5124 beginning of the string, LOG_ERR is chosen.
5126 On Win32, the LOG_* levels are mapped according to this:
5128 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5129 LOG_WARNING => EVENTLOG_WARNING_TYPE
5130 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5134 *) Made it possible to reconfigure with just the configuration
5135 argument "reconf" or "reconfigure". The command line arguments
5136 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5137 and are retrieved from there when reconfiguring.
5141 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5143 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5146 *) The obj_dat.pl script was messing up the sorting of object
5147 names. The reason was that it compared the quoted version
5148 of strings as a result "OCSP" > "OCSP Signing" because
5149 " > SPACE. Changed script to store unquoted versions of
5150 names and add quotes on output. It was also omitting some
5151 names from the lookup table if they were given a default
5152 value (that is if SN is missing it is given the same
5153 value as LN and vice versa), these are now added on the
5154 grounds that if an object has a name we should be able to
5155 look it up. Finally added warning output when duplicate
5156 short or long names are found.
5159 *) Changes needed for Tandem NSK.
5160 [Scott Uroff <scott@xypro.com>]
5162 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5163 RSA_padding_check_SSLv23(), special padding was never detected
5164 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5165 version rollback attacks was not effective.
5167 In s23_clnt.c, don't use special rollback-attack detection padding
5168 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5169 client; similarly, in s23_srvr.c, don't do the rollback check if
5170 SSL 2.0 is the only protocol enabled in the server.
5173 *) Make it possible to get hexdumps of unprintable data with 'openssl
5174 asn1parse'. By implication, the functions ASN1_parse_dump() and
5175 BIO_dump_indent() are added.
5178 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5179 these print out strings and name structures based on various
5180 flags including RFC2253 support and proper handling of
5181 multibyte characters. Added options to the 'x509' utility
5182 to allow the various flags to be set.
5185 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5186 Also change the functions X509_cmp_current_time() and
5187 X509_gmtime_adj() work with an ASN1_TIME structure,
5188 this will enable certificates using GeneralizedTime in validity
5189 dates to be checked.
5192 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5193 negative public key encodings) on by default,
5194 NO_NEG_PUBKEY_BUG can be set to disable it.
5197 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5198 content octets. An i2c_ASN1_OBJECT is unnecessary because
5199 the encoding can be trivially obtained from the structure.
5202 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5203 not read locks (CRYPTO_r_[un]lock).
5206 *) A first attempt at creating official support for shared
5207 libraries through configuration. I've kept it so the
5208 default is static libraries only, and the OpenSSL programs
5209 are always statically linked for now, but there are
5210 preparations for dynamic linking in place.
5211 This has been tested on Linux and Tru64.
5214 *) Randomness polling function for Win9x, as described in:
5215 Peter Gutmann, Software Generation of Practically Strong
5219 *) Fix so PRNG is seeded in req if using an already existing
5223 *) New options to smime application. -inform and -outform
5224 allow alternative formats for the S/MIME message including
5225 PEM and DER. The -content option allows the content to be
5226 specified separately. This should allow things like Netscape
5227 form signing output easier to verify.
5230 *) Fix the ASN1 encoding of tags using the 'long form'.
5233 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5234 STRING types. These convert content octets to and from the
5235 underlying type. The actual tag and length octets are
5236 already assumed to have been read in and checked. These
5237 are needed because all other string types have virtually
5238 identical handling apart from the tag. By having versions
5239 of the ASN1 functions that just operate on content octets
5240 IMPLICIT tagging can be handled properly. It also allows
5241 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5242 and ASN1_INTEGER are identical apart from the tag.
5245 *) Change the handling of OID objects as follows:
5247 - New object identifiers are inserted in objects.txt, following
5248 the syntax given in objects.README.
5249 - objects.pl is used to process obj_mac.num and create a new
5251 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5254 This is currently kind of a hack, and the perl code in objects.pl
5255 isn't very elegant, but it works as I intended. The simplest way
5256 to check that it worked correctly is to look in obj_dat.h and
5257 check the array nid_objs and make sure the objects haven't moved
5258 around (this is important!). Additions are OK, as well as
5259 consistent name changes.
5262 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5265 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5266 The given file adds to whatever has already been seeded into the
5267 random pool through the RANDFILE configuration file option or
5268 environment variable, or the default random state file.
5271 *) mkstack.pl now sorts each macro group into lexical order.
5272 Previously the output order depended on the order the files
5273 appeared in the directory, resulting in needless rewriting
5277 *) Patches to make OpenSSL compile under Win32 again. Mostly
5278 work arounds for the VC++ problem that it treats func() as
5279 func(void). Also stripped out the parts of mkdef.pl that
5280 added extra typesafe functions: these no longer exist.
5283 *) Reorganisation of the stack code. The macros are now all
5284 collected in safestack.h . Each macro is defined in terms of
5285 a "stack macro" of the form SKM_<name>(type, a, b). The
5286 DEBUG_SAFESTACK is now handled in terms of function casts,
5287 this has the advantage of retaining type safety without the
5288 use of additional functions. If DEBUG_SAFESTACK is not defined
5289 then the non typesafe macros are used instead. Also modified the
5290 mkstack.pl script to handle the new form. Needs testing to see
5291 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5292 the default if no major problems. Similar behaviour for ASN1_SET_OF
5293 and PKCS12_STACK_OF.
5296 *) When some versions of IIS use the 'NET' form of private key the
5297 key derivation algorithm is different. Normally MD5(password) is
5298 used as a 128 bit RC4 key. In the modified case
5299 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5300 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5301 as the old Netscape_RSA functions except they have an additional
5302 'sgckey' parameter which uses the modified algorithm. Also added
5303 an -sgckey command line option to the rsa utility. Thanks to
5304 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5305 algorithm to openssl-dev.
5308 *) The evp_local.h macros were using 'c.##kname' which resulted in
5309 invalid expansion on some systems (SCO 5.0.5 for example).
5310 Corrected to 'c.kname'.
5311 [Phillip Porch <root@theporch.com>]
5313 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5314 a STACK of email addresses from a certificate or request, these look
5315 in the subject name and the subject alternative name extensions and
5316 omit any duplicate addresses.
5319 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5320 This makes DSA verification about 2 % faster.
5323 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5324 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5325 plus overhead for 1024 bit moduli).
5326 This makes exponentiations about 0.5 % faster for 1024 bit
5327 exponents (as measured by "openssl speed rsa2048").
5330 *) Rename memory handling macros to avoid conflicts with other
5332 Malloc => OPENSSL_malloc
5333 Malloc_locked => OPENSSL_malloc_locked
5334 Realloc => OPENSSL_realloc
5335 Free => OPENSSL_free
5338 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5339 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5342 *) CygWin32 support.
5343 [John Jarvie <jjarvie@newsguy.com>]
5345 *) The type-safe stack code has been rejigged. It is now only compiled
5346 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5347 by default all type-specific stack functions are "#define"d back to
5348 standard stack functions. This results in more streamlined output
5349 but retains the type-safety checking possibilities of the original
5353 *) The STACK code has been cleaned up, and certain type declarations
5354 that didn't make a lot of sense have been brought in line. This has
5355 also involved a cleanup of sorts in safestack.h to more correctly
5356 map type-safe stack functions onto their plain stack counterparts.
5357 This work has also resulted in a variety of "const"ifications of
5358 lots of the code, especially "_cmp" operations which should normally
5359 be prototyped with "const" parameters anyway.
5362 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5363 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5364 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5365 where all of 'md' is used each time the PRNG is used, but 'state'
5366 is used only indexed by a cyclic counter. As entropy may not be
5367 well distributed from the beginning, 'md' is important as a
5368 chaining variable. However, the output function chains only half
5369 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5370 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5371 in all of 'state' being rewritten, with the new values depending
5372 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5375 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5376 the handshake is continued after ssl_verify_cert_chain();
5377 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5378 can lead to 'unexplainable' connection aborts later.
5379 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5381 *) Major EVP API cipher revision.
5382 Add hooks for extra EVP features. This allows various cipher
5383 parameters to be set in the EVP interface. Support added for variable
5384 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5385 setting of RC2 and RC5 parameters.
5387 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5390 Remove lots of duplicated code from the EVP library. For example *every*
5391 cipher init() function handles the 'iv' in the same way according to the
5392 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5393 for CFB and OFB modes they zero ctx->num.
5395 New functionality allows removal of S/MIME code RC2 hack.
5397 Most of the routines have the same form and so can be declared in terms
5400 By shifting this to the top level EVP_CipherInit() it can be removed from
5401 all individual ciphers. If the cipher wants to handle IVs or keys
5402 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5405 Change lots of functions like EVP_EncryptUpdate() to now return a
5406 value: although software versions of the algorithms cannot fail
5407 any installed hardware versions can.
5410 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5411 this option is set, tolerate broken clients that send the negotiated
5412 protocol version number instead of the requested protocol version
5416 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5417 i.e. non-zero for export ciphersuites, zero otherwise.
5418 Previous versions had this flag inverted, inconsistent with
5419 rsa_tmp_cb (..._TMP_RSA_CB).
5420 [Bodo Moeller; problem reported by Amit Chopra]
5422 *) Add missing DSA library text string. Work around for some IIS
5423 key files with invalid SEQUENCE encoding.
5426 *) Add a document (doc/standards.txt) that list all kinds of standards
5427 and so on that are implemented in OpenSSL.
5430 *) Enhance c_rehash script. Old version would mishandle certificates
5431 with the same subject name hash and wouldn't handle CRLs at all.
5432 Added -fingerprint option to crl utility, to support new c_rehash
5436 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5439 *) Fix for SSL server purpose checking. Server checking was
5440 rejecting certificates which had extended key usage present
5441 but no ssl client purpose.
5442 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5444 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5445 is a little unclear about how a blank password is handled.
5446 Since the password in encoded as a BMPString with terminating
5447 double NULL a zero length password would end up as just the
5448 double NULL. However no password at all is different and is
5449 handled differently in the PKCS#12 key generation code. NS
5450 treats a blank password as zero length. MSIE treats it as no
5451 password on export: but it will try both on import. We now do
5452 the same: PKCS12_parse() tries zero length and no password if
5453 the password is set to "" or NULL (NULL is now a valid password:
5454 it wasn't before) as does the pkcs12 application.
5457 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5458 perror when PEM_read_bio_X509_REQ fails, the error message must
5459 be obtained from the error queue.
5462 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5463 it in ERR_remove_state if appropriate, and change ERR_get_state
5464 accordingly to avoid race conditions (this is necessary because
5465 thread_hash is no longer constant once set).
5468 *) Bugfix for linux-elf makefile.one.
5471 *) RSA_get_default_method() will now cause a default
5472 RSA_METHOD to be chosen if one doesn't exist already.
5473 Previously this was only set during a call to RSA_new()
5474 or RSA_new_method(NULL) meaning it was possible for
5475 RSA_get_default_method() to return NULL.
5478 *) Added native name translation to the existing DSO code
5479 that will convert (if the flag to do so is set) filenames
5480 that are sufficiently small and have no path information
5481 into a canonical native form. Eg. "blah" converted to
5482 "libblah.so" or "blah.dll" etc.
5485 *) New function ERR_error_string_n(e, buf, len) which is like
5486 ERR_error_string(e, buf), but writes at most 'len' bytes
5487 including the 0 terminator. For ERR_error_string_n, 'buf'
5489 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5491 *) CONF library reworked to become more general. A new CONF
5492 configuration file reader "class" is implemented as well as a
5493 new functions (NCONF_*, for "New CONF") to handle it. The now
5494 old CONF_* functions are still there, but are reimplemented to
5495 work in terms of the new functions. Also, a set of functions
5496 to handle the internal storage of the configuration data is
5497 provided to make it easier to write new configuration file
5498 reader "classes" (I can definitely see something reading a
5499 configuration file in XML format, for example), called _CONF_*,
5500 or "the configuration storage API"...
5502 The new configuration file reading functions are:
5504 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5505 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5507 NCONF_default, NCONF_WIN32
5509 NCONF_dump_fp, NCONF_dump_bio
5511 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5512 NCONF_new creates a new CONF object. This works in the same way
5513 as other interfaces in OpenSSL, like the BIO interface.
5514 NCONF_dump_* dump the internal storage of the configuration file,
5515 which is useful for debugging. All other functions take the same
5516 arguments as the old CONF_* functions wth the exception of the
5517 first that must be a `CONF *' instead of a `LHASH *'.
5519 To make it easer to use the new classes with the old CONF_* functions,
5520 the function CONF_set_default_method is provided.
5523 *) Add '-tls1' option to 'openssl ciphers', which was already
5524 mentioned in the documentation but had not been implemented.
5525 (This option is not yet really useful because even the additional
5526 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5529 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5530 OpenSSL-based applications) load shared libraries and bind to
5531 them in a portable way.
5532 [Geoff Thorpe, with contributions from Richard Levitte]
5534 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5536 *) Make sure _lrotl and _lrotr are only used with MSVC.
5538 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5539 (the default implementation of RAND_status).
5541 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5542 to '-clrext' (= clear extensions), as intended and documented.
5543 [Bodo Moeller; inconsistency pointed out by Michael Attili
5546 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5547 was larger than the MD block size.
5548 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5550 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5551 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5552 using the passed key: if the passed key was a private key the result
5553 of X509_print(), for example, would be to print out all the private key
5557 *) des_quad_cksum() byte order bug fix.
5558 [Ulf Möller, using the problem description in krb4-0.9.7, where
5559 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5561 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5563 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5565 *) For easily testing in shell scripts whether some command
5566 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5567 returns with exit code 0 iff no command of the given name is available.
5568 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5569 the output goes to stdout and nothing is printed to stderr.
5570 Additional arguments are always ignored.
5572 Since for each cipher there is a command of the same name,
5573 the 'no-cipher' compilation switches can be tested this way.
5575 ('openssl no-XXX' is not able to detect pseudo-commands such
5576 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5579 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5582 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5583 is set; it will be thrown away anyway because each handshake creates
5585 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5586 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5587 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5588 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5591 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5592 'Q' and 'R' lose their special meanings (quit/renegotiate).
5593 This is part of what -quiet does; unlike -quiet, -ign_eof
5594 does not suppress any output.
5597 *) Add compatibility options to the purpose and trust code. The
5598 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5599 accepts a certificate or CA, this was the previous behaviour,
5600 with all the associated security issues.
5602 X509_TRUST_COMPAT is the old trust behaviour: only and
5603 automatically trust self signed roots in certificate store. A
5604 new trust setting X509_TRUST_DEFAULT is used to specify that
5605 a purpose has no associated trust setting and it should instead
5606 use the value in the default purpose.
5609 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5610 and fix a memory leak.
5613 *) In util/mkerr.pl (which implements 'make errors'), preserve
5614 reason strings from the previous version of the .c file, as
5615 the default to have only downcase letters (and digits) in
5616 automatically generated reasons codes is not always appropriate.
5619 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5620 using strerror. Previously, ERR_reason_error_string() returned
5621 library names as reason strings for SYSerr; but SYSerr is a special
5622 case where small numbers are errno values, not library numbers.
5625 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5626 converts DSA parameters into DH parameters. (When creating parameters,
5627 DSA_generate_parameters is used.)
5630 *) Include 'length' (recommended exponent length) in C code generated
5631 by 'openssl dhparam -C'.
5634 *) The second argument to set_label in perlasm was already being used
5635 so couldn't be used as a "file scope" flag. Moved to third argument
5639 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5640 instead of RAND_bytes for encryption IVs and salts.
5643 *) Include RAND_status() into RAND_METHOD instead of implementing
5644 it only for md_rand.c Otherwise replacing the PRNG by calling
5645 RAND_set_rand_method would be impossible.
5648 *) Don't let DSA_generate_key() enter an infinite loop if the random
5649 number generation fails.
5652 *) New 'rand' application for creating pseudo-random output.
5655 *) Added configuration support for Linux/IA64
5656 [Rolf Haberrecker <rolf@suse.de>]
5658 *) Assembler module support for Mingw32.
5661 *) Shared library support for HPUX (in shlib/).
5662 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5664 *) Shared library support for Solaris gcc.
5665 [Lutz Behnke <behnke@trustcenter.de>]
5667 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5669 *) PKCS7_encrypt() was adding text MIME headers twice because they
5670 were added manually and by SMIME_crlf_copy().
5673 *) In bntest.c don't call BN_rand with zero bits argument.
5674 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5676 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5677 case was implemented. This caused BN_div_recp() to fail occasionally.
5680 *) Add an optional second argument to the set_label() in the perl
5681 assembly language builder. If this argument exists and is set
5682 to 1 it signals that the assembler should use a symbol whose
5683 scope is the entire file, not just the current function. This
5684 is needed with MASM which uses the format label:: for this scope.
5685 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5687 *) Change the ASN1 types so they are typedefs by default. Before
5688 almost all types were #define'd to ASN1_STRING which was causing
5689 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5693 *) Change names of new functions to the new get1/get0 naming
5694 convention: After 'get1', the caller owns a reference count
5695 and has to call ..._free; 'get0' returns a pointer to some
5696 data structure without incrementing reference counters.
5697 (Some of the existing 'get' functions increment a reference
5698 counter, some don't.)
5699 Similarly, 'set1' and 'add1' functions increase reference
5700 counters or duplicate objects.
5703 *) Allow for the possibility of temp RSA key generation failure:
5704 the code used to assume it always worked and crashed on failure.
5707 *) Fix potential buffer overrun problem in BIO_printf().
5708 [Ulf Möller, using public domain code by Patrick Powell; problem
5709 pointed out by David Sacerdote <das33@cornell.edu>]
5711 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5712 RAND_egd() and RAND_status(). In the command line application,
5713 the EGD socket can be specified like a seed file using RANDFILE
5717 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5718 Some CAs (e.g. Verisign) distribute certificates in this form.
5721 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5722 list to exclude them. This means that no special compilation option
5723 is needed to use anonymous DH: it just needs to be included in the
5727 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5728 EVP_MD_type. The old functionality is available in a new macro called
5729 EVP_MD_md(). Change code that uses it and update docs.
5732 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5733 where the 'void *' argument is replaced by a function pointer argument.
5734 Previously 'void *' was abused to point to functions, which works on
5735 many platforms, but is not correct. As these functions are usually
5736 called by macros defined in OpenSSL header files, most source code
5737 should work without changes.
5740 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5741 sections with information on -D... compiler switches used for
5742 compiling the library so that applications can see them. To enable
5743 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5744 must be defined. E.g.,
5745 #define OPENSSL_ALGORITHM_DEFINES
5746 #include <openssl/opensslconf.h>
5747 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5748 [Richard Levitte, Ulf and Bodo Möller]
5750 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5754 *) Change the 'other' type in certificate aux info to a STACK_OF
5755 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5756 the required ASN1 format: arbitrary types determined by an OID.
5759 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5760 argument to 'req'. This is not because the function is newer or
5761 better than others it just uses the work 'NEW' in the certificate
5762 request header lines. Some software needs this.
5765 *) Reorganise password command line arguments: now passwords can be
5766 obtained from various sources. Delete the PEM_cb function and make
5767 it the default behaviour: i.e. if the callback is NULL and the
5768 usrdata argument is not NULL interpret it as a null terminated pass
5769 phrase. If usrdata and the callback are NULL then the pass phrase
5770 is prompted for as usual.
5773 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5774 the support is automatically enabled. The resulting binaries will
5775 autodetect the card and use it if present.
5776 [Ben Laurie and Compaq Inc.]
5778 *) Work around for Netscape hang bug. This sends certificate request
5779 and server done in one record. Since this is perfectly legal in the
5780 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5781 the bugs/SSLv3 entry for more info.
5784 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5787 *) Add -rand argument to smime and pkcs12 applications and read/write
5791 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5794 *) Add command line password options to the remaining applications.
5797 *) Bug fix for BN_div_recp() for numerators with an even number of
5801 *) More tests in bntest.c, and changed test_bn output.
5804 *) ./config recognizes MacOS X now.
5807 *) Bug fix for BN_div() when the first words of num and divsor are
5808 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5811 *) Add support for various broken PKCS#8 formats, and command line
5812 options to produce them.
5815 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5816 get temporary BIGNUMs from a BN_CTX.
5819 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5823 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5824 include a #define from the old name to the new. The original intent
5825 was that statically linked binaries could for example just call
5826 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5827 link with digests. This never worked becayse SSLeay_add_all_digests()
5828 and SSLeay_add_all_ciphers() were in the same source file so calling
5829 one would link with the other. They are now in separate source files.
5832 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5835 *) Use a less unusual form of the Miller-Rabin primality test (it used
5836 a binary algorithm for exponentiation integrated into the Miller-Rabin
5837 loop, our standard modexp algorithms are faster).
5840 *) Support for the EBCDIC character set completed.
5841 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5843 *) Source code cleanups: use const where appropriate, eliminate casts,
5844 use void * instead of char * in lhash.
5847 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5848 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5849 this the server could overwrite ephemeral keys that the client
5853 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5854 using 50 iterations of the Rabin-Miller test.
5856 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5857 iterations of the Rabin-Miller test as required by the appendix
5858 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5859 As BN_is_prime_fasttest includes trial division, DSA parameter
5860 generation becomes much faster.
5862 This implies a change for the callback functions in DSA_is_prime
5863 and DSA_generate_parameters: The callback function is called once
5864 for each positive witness in the Rabin-Miller test, not just
5865 occasionally in the inner loop; and the parameters to the
5866 callback function now provide an iteration count for the outer
5867 loop rather than for the current invocation of the inner loop.
5868 DSA_generate_parameters additionally can call the callback
5869 function with an 'iteration count' of -1, meaning that a
5870 candidate has passed the trial division test (when q is generated
5871 from an application-provided seed, trial division is skipped).
5874 *) New function BN_is_prime_fasttest that optionally does trial
5875 division before starting the Rabin-Miller test and has
5876 an additional BN_CTX * argument (whereas BN_is_prime always
5877 has to allocate at least one BN_CTX).
5878 'callback(1, -1, cb_arg)' is called when a number has passed the
5879 trial division stage.
5882 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5886 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5889 *) New function BN_pseudo_rand().
5892 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5893 bignum version of BN_from_montgomery() with the working code from
5894 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5898 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5899 made it impossible to use the same SSL_SESSION data structure in
5900 SSL2 clients in multiple threads.
5903 *) The return value of RAND_load_file() no longer counts bytes obtained
5904 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5905 to seed the PRNG (previously an explicit byte count was required).
5906 [Ulf Möller, Bodo Möller]
5908 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5909 used (char *) instead of (void *) and had casts all over the place.
5912 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5915 *) Retain source code compatibility for BN_prime_checks macro:
5916 BN_is_prime(..., BN_prime_checks, ...) now uses
5917 BN_prime_checks_for_size to determine the appropriate number of
5918 Rabin-Miller iterations.
5921 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5922 DH_CHECK_P_NOT_SAFE_PRIME.
5923 (Check if this is true? OpenPGP calls them "strong".)
5926 *) Merge the functionality of "dh" and "gendh" programs into a new program
5927 "dhparam". The old programs are retained for now but will handle DH keys
5928 (instead of parameters) in future.
5931 *) Make the ciphers, s_server and s_client programs check the return values
5932 when a new cipher list is set.
5935 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5936 ciphers. Before when the 56bit ciphers were enabled the sorting was
5939 The syntax for the cipher sorting has been extended to support sorting by
5940 cipher-strength (using the strength_bits hard coded in the tables).
5941 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5943 Fix a bug in the cipher-command parser: when supplying a cipher command
5944 string with an "undefined" symbol (neither command nor alphanumeric
5945 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5946 an error is flagged.
5948 Due to the strength-sorting extension, the code of the
5949 ssl_create_cipher_list() function was completely rearranged. I hope that
5950 the readability was also increased :-)
5951 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5953 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5954 for the first serial number and places 2 in the serial number file. This
5955 avoids problems when the root CA is created with serial number zero and
5956 the first user certificate has the same issuer name and serial number
5960 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5961 the new code. Add documentation for this stuff.
5964 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5965 X509_*() to X509at_*() on the grounds that they don't handle X509
5966 structures and behave in an analagous way to the X509v3 functions:
5967 they shouldn't be called directly but wrapper functions should be used
5970 So we also now have some wrapper functions that call the X509at functions
5971 when passed certificate requests. (TO DO: similar things can be done with
5972 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5973 things. Some of these need some d2i or i2d and print functionality
5974 because they handle more complex structures.)
5977 *) Add missing #ifndefs that caused missing symbols when building libssl
5978 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5979 NO_RSA in ssl/s2*.c.
5980 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5982 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5983 has a return value which indicates the quality of the random data
5984 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5985 error queue. New function RAND_pseudo_bytes() generates output that is
5986 guaranteed to be unique but not unpredictable. RAND_add is like
5987 RAND_seed, but takes an extra argument for an entropy estimate
5988 (RAND_seed always assumes full entropy).
5991 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5992 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5993 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5994 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5995 false-positive rate of at most 2^-80 for random input.
5998 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6001 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6002 in the 0.9.5 release), this returns the chain
6003 from an X509_CTX structure with a dup of the stack and all
6004 the X509 reference counts upped: so the stack will exist
6005 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6008 Also make SSL_SESSION_print() print out the verify return
6012 *) Add manpage for the pkcs12 command. Also change the default
6013 behaviour so MAC iteration counts are used unless the new
6014 -nomaciter option is used. This improves file security and
6015 only older versions of MSIE (4.0 for example) need it.
6018 *) Honor the no-xxx Configure options when creating .DEF files.
6021 *) Add PKCS#10 attributes to field table: challengePassword,
6022 unstructuredName and unstructuredAddress. These are taken from
6023 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6024 international characters are used.
6026 More changes to X509_ATTRIBUTE code: allow the setting of types
6027 based on strings. Remove the 'loc' parameter when adding
6028 attributes because these will be a SET OF encoding which is sorted
6032 *) Initial changes to the 'req' utility to allow request generation
6033 automation. This will allow an application to just generate a template
6034 file containing all the field values and have req construct the
6037 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6038 used all over the place including certificate requests and PKCS#7
6039 structures. They are currently handled manually where necessary with
6040 some primitive wrappers for PKCS#7. The new functions behave in a
6041 manner analogous to the X509 extension functions: they allow
6042 attributes to be looked up by NID and added.
6044 Later something similar to the X509V3 code would be desirable to
6045 automatically handle the encoding, decoding and printing of the
6046 more complex types. The string types like challengePassword can
6047 be handled by the string table functions.
6049 Also modified the multi byte string table handling. Now there is
6050 a 'global mask' which masks out certain types. The table itself
6051 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6052 is useful when for example there is only one permissible type
6053 (as in countryName) and using the mask might result in no valid
6057 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6058 SSL_get_peer_finished to allow applications to obtain the latest
6059 Finished messages sent to the peer or expected from the peer,
6060 respectively. (SSL_get_peer_finished is usually the Finished message
6061 actually received from the peer, otherwise the protocol will be aborted.)
6063 As the Finished message are message digests of the complete handshake
6064 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6065 be used for external authentication procedures when the authentication
6066 provided by SSL/TLS is not desired or is not enough.
6069 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6070 the host supports BWX extension and if Compaq C is present on the
6071 $PATH. Just exploiting of the BWX extension results in 20-30%
6072 performance kick for some algorithms, e.g. DES and RC4 to mention
6073 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6077 *) Add support for MS "fast SGC". This is arguably a violation of the
6078 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6079 weak crypto and after checking the certificate is SGC a second one
6080 with strong crypto. MS SGC stops the first handshake after receiving
6081 the server certificate message and sends a second client hello. Since
6082 a server will typically do all the time consuming operations before
6083 expecting any further messages from the client (server key exchange
6084 is the most expensive) there is little difference between the two.
6086 To get OpenSSL to support MS SGC we have to permit a second client
6087 hello message after we have sent server done. In addition we have to
6088 reset the MAC if we do get this second client hello.
6091 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6092 if a DER encoded private key is RSA or DSA traditional format. Changed
6093 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6094 format DER encoded private key. Newer code should use PKCS#8 format which
6095 has the key type encoded in the ASN1 structure. Added DER private key
6096 support to pkcs8 application.
6099 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6100 ciphersuites has been selected (as required by the SSL 3/TLS 1
6101 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6102 is set, we interpret this as a request to violate the specification
6103 (the worst that can happen is a handshake failure, and 'correct'
6104 behaviour would result in a handshake failure anyway).
6107 *) In SSL_CTX_add_session, take into account that there might be multiple
6108 SSL_SESSION structures with the same session ID (e.g. when two threads
6109 concurrently obtain them from an external cache).
6110 The internal cache can handle only one SSL_SESSION with a given ID,
6111 so if there's a conflict, we now throw out the old one to achieve
6115 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6116 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6117 some routines that use cipher OIDs: some ciphers do not have OIDs
6118 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6122 *) Simplify the trust setting structure and code. Now we just have
6123 two sequences of OIDs for trusted and rejected settings. These will
6124 typically have values the same as the extended key usage extension
6125 and any application specific purposes.
6127 The trust checking code now has a default behaviour: it will just
6128 check for an object with the same NID as the passed id. Functions can
6129 be provided to override either the default behaviour or the behaviour
6130 for a given id. SSL client, server and email already have functions
6131 in place for compatibility: they check the NID and also return "trusted"
6132 if the certificate is self signed.
6135 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6136 traditional format into an EVP_PKEY structure.
6139 *) Add a password callback function PEM_cb() which either prompts for
6140 a password if usr_data is NULL or otherwise assumes it is a null
6141 terminated password. Allow passwords to be passed on command line
6142 environment or config files in a few more utilities.
6145 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6146 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6147 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6148 Update documentation.
6151 *) Support for ASN1 "NULL" type. This could be handled before by using
6152 ASN1_TYPE but there wasn't any function that would try to read a NULL
6153 and produce an error if it couldn't. For compatibility we also have
6154 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6155 don't allocate anything because they don't need to.
6158 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6160 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6162 *) Rebuild of the memory allocation routines used by OpenSSL code and
6163 possibly others as well. The purpose is to make an interface that
6164 provide hooks so anyone can build a separate set of allocation and
6165 deallocation routines to be used by OpenSSL, for example memory
6166 pool implementations, or something else, which was previously hard
6167 since Malloc(), Realloc() and Free() were defined as macros having
6168 the values malloc, realloc and free, respectively (except for Win32
6169 compilations). The same is provided for memory debugging code.
6170 OpenSSL already comes with functionality to find memory leaks, but
6171 this gives people a chance to debug other memory problems.
6173 With these changes, a new set of functions and macros have appeared:
6175 CRYPTO_set_mem_debug_functions() [F]
6176 CRYPTO_get_mem_debug_functions() [F]
6177 CRYPTO_dbg_set_options() [F]
6178 CRYPTO_dbg_get_options() [F]
6179 CRYPTO_malloc_debug_init() [M]
6181 The memory debug functions are NULL by default, unless the library
6182 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6183 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6184 gives the standard debugging functions that come with OpenSSL) or
6185 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6186 provided by the library user) must be used. When the standard
6187 debugging functions are used, CRYPTO_dbg_set_options can be used to
6188 request additional information:
6189 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6190 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6192 Also, things like CRYPTO_set_mem_functions will always give the
6193 expected result (the new set of functions is used for allocation
6194 and deallocation) at all times, regardless of platform and compiler
6197 To finish it up, some functions that were never use in any other
6198 way than through macros have a new API and new semantic:
6201 CRYPTO_dbg_realloc()
6204 All macros of value have retained their old syntax.
6205 [Richard Levitte and Bodo Moeller]
6207 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6208 ordering of SMIMECapabilities wasn't in "strength order" and there
6209 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6213 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6214 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6215 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6217 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6218 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6219 functionality to handle multipart/signed properly) and a utility
6220 called 'smime' to call all this stuff. This is based on code I
6221 originally wrote for Celo who have kindly allowed it to be
6222 included in OpenSSL.
6225 *) Add variants des_set_key_checked and des_set_key_unchecked of
6226 des_set_key (aka des_key_sched). Global variable des_check_key
6227 decides which of these is called by des_set_key; this way
6228 des_check_key behaves as it always did, but applications and
6229 the library itself, which was buggy for des_check_key == 1,
6230 have a cleaner way to pick the version they need.
6233 *) New function PKCS12_newpass() which changes the password of a
6237 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6238 dynamic mix. In both cases the ids can be used as an index into the
6239 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6240 functions so they accept a list of the field values and the
6241 application doesn't need to directly manipulate the X509_TRUST
6245 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6249 *) Modify the way the V3 extension code looks up extensions. This now
6250 works in a similar way to the object code: we have some "standard"
6251 extensions in a static table which is searched with OBJ_bsearch()
6252 and the application can add dynamic ones if needed. The file
6253 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6254 updated whenever a new extension is added to the core code and kept
6255 in ext_nid order. There is a simple program 'tabtest.c' which checks
6256 this. New extensions are not added too often so this file can readily
6257 be maintained manually.
6259 There are two big advantages in doing things this way. The extensions
6260 can be looked up immediately and no longer need to be "added" using
6261 X509V3_add_standard_extensions(): this function now does nothing.
6262 [Side note: I get *lots* of email saying the extension code doesn't
6263 work because people forget to call this function]
6264 Also no dynamic allocation is done unless new extensions are added:
6265 so if we don't add custom extensions there is no need to call
6266 X509V3_EXT_cleanup().
6269 *) Modify enc utility's salting as follows: make salting the default. Add a
6270 magic header, so unsalted files fail gracefully instead of just decrypting
6271 to garbage. This is because not salting is a big security hole, so people
6272 should be discouraged from doing it.
6275 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6276 digest to be passed on the command line but it only used this
6277 parameter when signing a certificate. Modified so all relevant
6278 operations are affected by the digest parameter including the
6279 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6280 DSA key was used because it didn't fix the digest.
6283 *) Initial certificate chain verify code. Currently tests the untrusted
6284 certificates for consistency with the verify purpose (which is set
6285 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6287 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6288 this is because it will reject chains with invalid extensions whereas
6289 every previous version of OpenSSL and SSLeay made no checks at all.
6291 Trust code: checks the root CA for the relevant trust settings. Trust
6292 settings have an initial value consistent with the verify purpose: e.g.
6293 if the verify purpose is for SSL client use it expects the CA to be
6294 trusted for SSL client use. However the default value can be changed to
6295 permit custom trust settings: one example of this would be to only trust
6296 certificates from a specific "secure" set of CAs.
6298 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6299 which should be used for version portability: especially since the
6300 verify structure is likely to change more often now.
6302 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6303 to set them. If not set then assume SSL clients will verify SSL servers
6306 Two new options to the verify program: -untrusted allows a set of
6307 untrusted certificates to be passed in and -purpose which sets the
6308 intended purpose of the certificate. If a purpose is set then the
6309 new chain verify code is used to check extension consistency.
6312 *) Support for the authority information access extension.
6315 *) Modify RSA and DSA PEM read routines to transparently handle
6316 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6317 public keys in a format compatible with certificate
6318 SubjectPublicKeyInfo structures. Unfortunately there were already
6319 functions called *_PublicKey_* which used various odd formats so
6320 these are retained for compatibility: however the DSA variants were
6321 never in a public release so they have been deleted. Changed dsa/rsa
6322 utilities to handle the new format: note no releases ever handled public
6323 keys so we should be OK.
6325 The primary motivation for this change is to avoid the same fiasco
6326 that dogs private keys: there are several incompatible private key
6327 formats some of which are standard and some OpenSSL specific and
6328 require various evil hacks to allow partial transparent handling and
6329 even then it doesn't work with DER formats. Given the option anything
6330 other than PKCS#8 should be dumped: but the other formats have to
6331 stay in the name of compatibility.
6333 With public keys and the benefit of hindsight one standard format
6334 is used which works with EVP_PKEY, RSA or DSA structures: though
6335 it clearly returns an error if you try to read the wrong kind of key.
6337 Added a -pubkey option to the 'x509' utility to output the public key.
6338 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6339 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6340 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6341 that do the same as the EVP_PKEY_assign_*() except they up the
6342 reference count of the added key (they don't "swallow" the
6346 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6347 CRLs would fail if the file contained no certificates or no CRLs:
6348 added a new function to read in both types and return the number
6349 read: this means that if none are read it will be an error. The
6350 DER versions of the certificate and CRL reader would always fail
6351 because it isn't possible to mix certificates and CRLs in DER format
6352 without choking one or the other routine. Changed this to just read
6353 a certificate: this is the best we can do. Also modified the code
6354 in apps/verify.c to take notice of return codes: it was previously
6355 attempting to read in certificates from NULL pointers and ignoring
6356 any errors: this is one reason why the cert and CRL reader seemed
6357 to work. It doesn't check return codes from the default certificate
6358 routines: these may well fail if the certificates aren't installed.
6361 *) Code to support otherName option in GeneralName.
6364 *) First update to verify code. Change the verify utility
6365 so it warns if it is passed a self signed certificate:
6366 for consistency with the normal behaviour. X509_verify
6367 has been modified to it will now verify a self signed
6368 certificate if *exactly* the same certificate appears
6369 in the store: it was previously impossible to trust a
6370 single self signed certificate. This means that:
6371 openssl verify ss.pem
6372 now gives a warning about a self signed certificate but
6373 openssl verify -CAfile ss.pem ss.pem
6377 *) For servers, store verify_result in SSL_SESSION data structure
6378 (and add it to external session representation).
6379 This is needed when client certificate verifications fails,
6380 but an application-provided verification callback (set by
6381 SSL_CTX_set_cert_verify_callback) allows accepting the session
6382 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6383 but returns 1): When the session is reused, we have to set
6384 ssl->verify_result to the appropriate error code to avoid
6386 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6388 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6389 case in PKCS7_dataInit() where the signed PKCS7 structure
6390 didn't contain any existing data because it was being created.
6391 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6393 *) Add a salt to the key derivation routines in enc.c. This
6394 forms the first 8 bytes of the encrypted file. Also add a
6395 -S option to allow a salt to be input on the command line.
6398 *) New function X509_cmp(). Oddly enough there wasn't a function
6399 to compare two certificates. We do this by working out the SHA1
6400 hash and comparing that. X509_cmp() will be needed by the trust
6404 *) SSL_get1_session() is like SSL_get_session(), but increments
6405 the reference count in the SSL_SESSION returned.
6406 [Geoff Thorpe <geoff@eu.c2.net>]
6408 *) Fix for 'req': it was adding a null to request attributes.
6409 Also change the X509_LOOKUP and X509_INFO code to handle
6410 certificate auxiliary information.
6413 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6417 *) Add the possibility to add extra information to the memory leak
6418 detecting output, to form tracebacks, showing from where each
6419 allocation was originated: CRYPTO_push_info("constant string") adds
6420 the string plus current file name and line number to a per-thread
6421 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6422 is like calling CYRPTO_pop_info() until the stack is empty.
6423 Also updated memory leak detection code to be multi-thread-safe.
6426 *) Add options -text and -noout to pkcs7 utility and delete the
6427 encryption options which never did anything. Update docs.
6430 *) Add options to some of the utilities to allow the pass phrase
6431 to be included on either the command line (not recommended on
6432 OSes like Unix) or read from the environment. Update the
6433 manpages and fix a few bugs.
6436 *) Add a few manpages for some of the openssl commands.
6439 *) Fix the -revoke option in ca. It was freeing up memory twice,
6440 leaking and not finding already revoked certificates.
6443 *) Extensive changes to support certificate auxiliary information.
6444 This involves the use of X509_CERT_AUX structure and X509_AUX
6445 functions. An X509_AUX function such as PEM_read_X509_AUX()
6446 can still read in a certificate file in the usual way but it
6447 will also read in any additional "auxiliary information". By
6448 doing things this way a fair degree of compatibility can be
6449 retained: existing certificates can have this information added
6450 using the new 'x509' options.
6452 Current auxiliary information includes an "alias" and some trust
6453 settings. The trust settings will ultimately be used in enhanced
6454 certificate chain verification routines: currently a certificate
6455 can only be trusted if it is self signed and then it is trusted
6459 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6460 The problem was that one of the replacement routines had not been working
6461 since SSLeay releases. For now the offending routine has been replaced
6462 with non-optimised assembler. Even so, this now gives around 95%
6463 performance improvement for 1024 bit RSA signs.
6466 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6467 handling. Most clients have the effective key size in bits equal to
6468 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6469 A few however don't do this and instead use the size of the decrypted key
6470 to determine the RC2 key length and the AlgorithmIdentifier to determine
6471 the effective key length. In this case the effective key length can still
6472 be 40 bits but the key length can be 168 bits for example. This is fixed
6473 by manually forcing an RC2 key into the EVP_PKEY structure because the
6474 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6475 the key length and effective key length are equal.
6478 *) Add a bunch of functions that should simplify the creation of
6479 X509_NAME structures. Now you should be able to do:
6480 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6481 and have it automatically work out the correct field type and fill in
6482 the structures. The more adventurous can try:
6483 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6484 and it will (hopefully) work out the correct multibyte encoding.
6487 *) Change the 'req' utility to use the new field handling and multibyte
6488 copy routines. Before the DN field creation was handled in an ad hoc
6489 way in req, ca, and x509 which was rather broken and didn't support
6490 BMPStrings or UTF8Strings. Since some software doesn't implement
6491 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6492 using the dirstring_type option. See the new comment in the default
6493 openssl.cnf for more info.
6496 *) Make crypto/rand/md_rand.c more robust:
6497 - Assure unique random numbers after fork().
6498 - Make sure that concurrent threads access the global counter and
6499 md serializably so that we never lose entropy in them
6500 or use exactly the same state in multiple threads.
6501 Access to the large state is not always serializable because
6502 the additional locking could be a performance killer, and
6503 md should be large enough anyway.
6506 *) New file apps/app_rand.c with commonly needed functionality
6507 for handling the random seed file.
6509 Use the random seed file in some applications that previously did not:
6511 dsaparam -genkey (which also ignored its '-rand' option),
6514 x509 (when signing).
6515 Except on systems with /dev/urandom, it is crucial to have a random
6516 seed file at least for key creation, DSA signing, and for DH exchanges;
6517 for RSA signatures we could do without one.
6519 gendh and gendsa (unlike genrsa) used to read only the first byte
6520 of each file listed in the '-rand' option. The function as previously
6521 found in genrsa is now in app_rand.c and is used by all programs
6522 that support '-rand'.
6525 *) In RAND_write_file, use mode 0600 for creating files;
6526 don't just chmod when it may be too late.
6529 *) Report an error from X509_STORE_load_locations
6530 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6533 *) New function ASN1_mbstring_copy() this copies a string in either
6534 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6535 into an ASN1_STRING type. A mask of permissible types is passed
6536 and it chooses the "minimal" type to use or an error if not type
6540 *) Add function equivalents to the various macros in asn1.h. The old
6541 macros are retained with an M_ prefix. Code inside the library can
6542 use the M_ macros. External code (including the openssl utility)
6543 should *NOT* in order to be "shared library friendly".
6546 *) Add various functions that can check a certificate's extensions
6547 to see if it usable for various purposes such as SSL client,
6548 server or S/MIME and CAs of these types. This is currently
6549 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6550 verification. Also added a -purpose flag to x509 utility to
6551 print out all the purposes.
6554 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6558 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6559 for, obtain and decode and extension and obtain its critical flag.
6560 This allows all the necessary extension code to be handled in a
6561 single function call.
6564 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6565 platforms. See crypto/rc4/rc4_enc.c for further details.
6568 *) New -noout option to asn1parse. This causes no output to be produced
6569 its main use is when combined with -strparse and -out to extract data
6570 from a file (which may not be in ASN.1 format).
6573 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6574 when producing the local key id.
6575 [Richard Levitte <levitte@stacken.kth.se>]
6577 *) New option -dhparam in s_server. This allows a DH parameter file to be
6578 stated explicitly. If it is not stated then it tries the first server
6579 certificate file. The previous behaviour hard coded the filename
6583 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6584 a public key to be input or output. For example:
6585 openssl rsa -in key.pem -pubout -out pubkey.pem
6586 Also added necessary DSA public key functions to handle this.
6589 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6590 in the message. This was handled by allowing
6591 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6592 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6594 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6595 to the end of the strings whereas this didn't. This would cause problems
6596 if strings read with d2i_ASN1_bytes() were later modified.
6597 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6599 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6600 data and it contains EOF it will end up returning an error. This is
6601 caused by input 46 bytes long. The cause is due to the way base64
6602 BIOs find the start of base64 encoded data. They do this by trying a
6603 trial decode on each line until they find one that works. When they
6604 do a flag is set and it starts again knowing it can pass all the
6605 data directly through the decoder. Unfortunately it doesn't reset
6606 the context it uses. This means that if EOF is reached an attempt
6607 is made to pass two EOFs through the context and this causes the
6608 resulting error. This can also cause other problems as well. As is
6609 usual with these problems it takes *ages* to find and the fix is
6610 trivial: move one line.
6611 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6613 *) Ugly workaround to get s_client and s_server working under Windows. The
6614 old code wouldn't work because it needed to select() on sockets and the
6615 tty (for keypresses and to see if data could be written). Win32 only
6616 supports select() on sockets so we select() with a 1s timeout on the
6617 sockets and then see if any characters are waiting to be read, if none
6618 are present then we retry, we also assume we can always write data to
6619 the tty. This isn't nice because the code then blocks until we've
6620 received a complete line of data and it is effectively polling the
6621 keyboard at 1s intervals: however it's quite a bit better than not
6622 working at all :-) A dedicated Windows application might handle this
6623 with an event loop for example.
6626 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6627 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6628 will be called when RSA_sign() and RSA_verify() are used. This is useful
6629 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6630 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6631 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6632 This necessitated the support of an extra signature type NID_md5_sha1
6633 for SSL signatures and modifications to the SSL library to use it instead
6634 of calling RSA_public_decrypt() and RSA_private_encrypt().
6637 *) Add new -verify -CAfile and -CApath options to the crl program, these
6638 will lookup a CRL issuers certificate and verify the signature in a
6639 similar way to the verify program. Tidy up the crl program so it
6640 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6641 less strict. It will now permit CRL extensions even if it is not
6642 a V2 CRL: this will allow it to tolerate some broken CRLs.
6645 *) Initialize all non-automatic variables each time one of the openssl
6646 sub-programs is started (this is necessary as they may be started
6647 multiple times from the "OpenSSL>" prompt).
6648 [Lennart Bang, Bodo Moeller]
6650 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6651 removing all other RSA functionality (this is what NO_RSA does). This
6652 is so (for example) those in the US can disable those operations covered
6653 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6657 *) Non-copying interface to BIO pairs.
6658 (still largely untested)
6661 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6662 ASCII string. This was handled independently in various places before.
6665 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6666 UTF8 strings a character at a time.
6669 *) Use client_version from client hello to select the protocol
6670 (s23_srvr.c) and for RSA client key exchange verification
6671 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6674 *) Add various utility functions to handle SPKACs, these were previously
6675 handled by poking round in the structure internals. Added new function
6676 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6677 print, verify and generate SPKACs. Based on an original idea from
6678 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6681 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6684 *) Allow the config file extension section to be overwritten on the
6685 command line. Based on an original idea from Massimiliano Pala
6686 <madwolf@comune.modena.it>. The new option is called -extensions
6687 and can be applied to ca, req and x509. Also -reqexts to override
6688 the request extensions in req and -crlexts to override the crl extensions
6692 *) Add new feature to the SPKAC handling in ca. Now you can include
6693 the same field multiple times by preceding it by "XXXX." for example:
6696 this is the same syntax as used in the req config file.
6699 *) Allow certificate extensions to be added to certificate requests. These
6700 are specified in a 'req_extensions' option of the req section of the
6701 config file. They can be printed out with the -text option to req but
6702 are otherwise ignored at present.
6705 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6706 data read consists of only the final block it would not decrypted because
6707 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6708 A misplaced 'break' also meant the decrypted final block might not be
6709 copied until the next read.
6712 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6713 a few extra parameters to the DH structure: these will be useful if
6714 for example we want the value of 'q' or implement X9.42 DH.
6717 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6718 provides hooks that allow the default DSA functions or functions on a
6719 "per key" basis to be replaced. This allows hardware acceleration and
6720 hardware key storage to be handled without major modification to the
6721 library. Also added low level modexp hooks and CRYPTO_EX structure and
6722 associated functions.
6725 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6726 as "read only": it can't be written to and the buffer it points to will
6727 not be freed. Reading from a read only BIO is much more efficient than
6728 a normal memory BIO. This was added because there are several times when
6729 an area of memory needs to be read from a BIO. The previous method was
6730 to create a memory BIO and write the data to it, this results in two
6731 copies of the data and an O(n^2) reading algorithm. There is a new
6732 function BIO_new_mem_buf() which creates a read only memory BIO from
6733 an area of memory. Also modified the PKCS#7 routines to use read only
6737 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6738 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6739 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6740 but a retry condition occured while trying to read the rest.
6743 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6744 NID_pkcs7_encrypted by default: this was wrong since this should almost
6745 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6746 the encrypted data type: this is a more sensible place to put it and it
6747 allows the PKCS#12 code to be tidied up that duplicated this
6751 *) Changed obj_dat.pl script so it takes its input and output files on
6752 the command line. This should avoid shell escape redirection problems
6756 *) Initial support for certificate extension requests, these are included
6757 in things like Xenroll certificate requests. Included functions to allow
6758 extensions to be obtained and added.
6761 *) -crlf option to s_client and s_server for sending newlines as
6762 CRLF (as required by many protocols).
6765 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6767 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6768 [Ralf S. Engelschall]
6770 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6771 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6773 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6777 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6778 DH parameters/keys (q is lost during that conversion, but the resulting
6779 DH parameters contain its length).
6781 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6782 much faster than DH_generate_parameters (which creates parameters
6783 where p = 2*q + 1), and also the smaller q makes DH computations
6784 much more efficient (160-bit exponentiation instead of 1024-bit
6785 exponentiation); so this provides a convenient way to support DHE
6786 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6787 utter importance to use
6788 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6790 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6791 when such DH parameters are used, because otherwise small subgroup
6792 attacks may become possible!
6795 *) Avoid memory leak in i2d_DHparams.
6798 *) Allow the -k option to be used more than once in the enc program:
6799 this allows the same encrypted message to be read by multiple recipients.
6802 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6803 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6804 it will always use the numerical form of the OID, even if it has a short
6808 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6809 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6810 otherwise bn_mod_exp was called. In the case of hardware keys for example
6811 no private key components need be present and it might store extra data
6812 in the RSA structure, which cannot be accessed from bn_mod_exp.
6813 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6814 private key operations.
6817 *) Added support for SPARC Linux.
6820 *) pem_password_cb function type incompatibly changed from
6821 typedef int pem_password_cb(char *buf, int size, int rwflag);
6823 ....(char *buf, int size, int rwflag, void *userdata);
6824 so that applications can pass data to their callbacks:
6825 The PEM[_ASN1]_{read,write}... functions and macros now take an
6826 additional void * argument, which is just handed through whenever
6827 the password callback is called.
6828 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6830 New function SSL_CTX_set_default_passwd_cb_userdata.
6832 Compatibility note: As many C implementations push function arguments
6833 onto the stack in reverse order, the new library version is likely to
6834 interoperate with programs that have been compiled with the old
6835 pem_password_cb definition (PEM_whatever takes some data that
6836 happens to be on the stack as its last argument, and the callback
6837 just ignores this garbage); but there is no guarantee whatsoever that
6840 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6841 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6842 problems not only on Windows, but also on some Unix platforms.
6843 To avoid problematic command lines, these definitions are now in an
6844 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6845 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6848 *) MIPS III/IV assembler module is reimplemented.
6851 *) More DES library cleanups: remove references to srand/rand and
6852 delete an unused file.
6855 *) Add support for the the free Netwide assembler (NASM) under Win32,
6856 since not many people have MASM (ml) and it can be hard to obtain.
6857 This is currently experimental but it seems to work OK and pass all
6858 the tests. Check out INSTALL.W32 for info.
6861 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6862 without temporary keys kept an extra copy of the server key,
6863 and connections with temporary keys did not free everything in case
6867 *) New function RSA_check_key and new openssl rsa option -check
6868 for verifying the consistency of RSA keys.
6869 [Ulf Moeller, Bodo Moeller]
6871 *) Various changes to make Win32 compile work:
6872 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6873 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6874 comparison" warnings.
6875 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6878 *) Add a debugging option to PKCS#5 v2 key generation function: when
6879 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6880 derived keys are printed to stderr.
6883 *) Copy the flags in ASN1_STRING_dup().
6884 [Roman E. Pavlov <pre@mo.msk.ru>]
6886 *) The x509 application mishandled signing requests containing DSA
6887 keys when the signing key was also DSA and the parameters didn't match.
6889 It was supposed to omit the parameters when they matched the signing key:
6890 the verifying software was then supposed to automatically use the CA's
6891 parameters if they were absent from the end user certificate.
6893 Omitting parameters is no longer recommended. The test was also
6894 the wrong way round! This was probably due to unusual behaviour in
6895 EVP_cmp_parameters() which returns 1 if the parameters match.
6896 This meant that parameters were omitted when they *didn't* match and
6897 the certificate was useless. Certificates signed with 'ca' didn't have
6899 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6901 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6902 The interface is as follows:
6903 Applications can use
6904 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6905 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6906 "off" is now the default.
6907 The library internally uses
6908 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6909 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6910 to disable memory-checking temporarily.
6912 Some inconsistent states that previously were possible (and were
6913 even the default) are now avoided.
6915 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6916 with each memory chunk allocated; this is occasionally more helpful
6917 than just having a counter.
6919 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6921 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6925 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6926 which largely parallels "options", but is for changing API behaviour,
6927 whereas "options" are about protocol behaviour.
6928 Initial "mode" flags are:
6930 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6931 a single record has been written.
6932 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6933 retries use the same buffer location.
6934 (But all of the contents must be
6938 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6941 *) Fix problems with no-hmac etc.
6942 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6944 *) New functions RSA_get_default_method(), RSA_set_method() and
6945 RSA_get_method(). These allows replacement of RSA_METHODs without having
6946 to mess around with the internals of an RSA structure.
6949 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6950 Also really enable memory leak checks in openssl.c and in some
6952 [Chad C. Mulligan, Bodo Moeller]
6954 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6955 up the length of negative integers. This has now been simplified to just
6956 store the length when it is first determined and use it later, rather
6957 than trying to keep track of where data is copied and updating it to
6959 [Steve Henson, reported by Brien Wheeler
6960 <bwheeler@authentica-security.com>]
6962 *) Add a new function PKCS7_signatureVerify. This allows the verification
6963 of a PKCS#7 signature but with the signing certificate passed to the
6964 function itself. This contrasts with PKCS7_dataVerify which assumes the
6965 certificate is present in the PKCS#7 structure. This isn't always the
6966 case: certificates can be omitted from a PKCS#7 structure and be
6967 distributed by "out of band" means (such as a certificate database).
6970 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6971 function prototypes in pem.h, also change util/mkdef.pl to add the
6972 necessary function names.
6975 *) mk1mf.pl (used by Windows builds) did not properly read the
6976 options set by Configure in the top level Makefile, and Configure
6977 was not even able to write more than one option correctly.
6978 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6981 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6982 file to be loaded from a BIO or FILE pointer. The BIO version will
6983 for example allow memory BIOs to contain config info.
6986 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6987 Whoever hopes to achieve shared-library compatibility across versions
6988 must use this, not the compile-time macro.
6989 (Exercise 0.9.4: Which is the minimum library version required by
6991 Note: All this applies only to multi-threaded programs, others don't
6995 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6996 through a BIO pair triggered the default case, i.e.
6997 SSLerr(...,SSL_R_UNKNOWN_STATE).
7000 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7001 can use the SSL library even if none of the specific BIOs is
7005 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7006 for the encoded length.
7007 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7009 *) Add initial documentation of the X509V3 functions.
7012 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7013 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7014 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7015 secure PKCS#8 private key format with a high iteration count.
7018 *) Fix determination of Perl interpreter: A perl or perl5
7019 _directory_ in $PATH was also accepted as the interpreter.
7020 [Ralf S. Engelschall]
7022 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7023 wrong with it but it was very old and did things like calling
7024 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7028 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7029 to use the new extension code.
7032 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7033 with macros. This should make it easier to change their form, add extra
7034 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7038 *) Add to configuration table a new entry that can specify an alternative
7039 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7040 according to Mark Crispin <MRC@Panda.COM>.
7044 *) DES CBC did not update the IV. Weird.
7047 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7048 Changing the behaviour of the former might break existing programs --
7049 where IV updating is needed, des_ncbc_encrypt can be used.
7052 *) When bntest is run from "make test" it drives bc to check its
7053 calculations, as well as internally checking them. If an internal check
7054 fails, it needs to cause bc to give a non-zero result or make test carries
7055 on without noticing the failure. Fixed.
7058 *) DES library cleanups.
7061 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7062 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7063 ciphers. NOTE: although the key derivation function has been verified
7064 against some published test vectors it has not been extensively tested
7065 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7069 *) Instead of "mkdir -p", which is not fully portable, use new
7070 Perl script "util/mkdir-p.pl".
7073 *) Rewrite the way password based encryption (PBE) is handled. It used to
7074 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7075 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7076 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7077 the 'parameter' field of the AlgorithmIdentifier is passed to the
7078 underlying key generation function so it must do its own ASN1 parsing.
7079 This has also changed the EVP_PBE_CipherInit() function which now has a
7080 'parameter' argument instead of literal salt and iteration count values
7081 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7084 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7085 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7086 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7087 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7088 value was just used as a "magic string" and not used directly its
7089 value doesn't matter.
7092 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7096 *) "linux-sparc64" configuration (ultrapenguin).
7097 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7098 "linux-sparc" configuration.
7099 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7101 *) config now generates no-xxx options for missing ciphers.
7104 *) Support the EBCDIC character set (work in progress).
7105 File ebcdic.c not yet included because it has a different license.
7106 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7108 *) Support BS2000/OSD-POSIX.
7109 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7111 *) Make callbacks for key generation use void * instead of char *.
7114 *) Make S/MIME samples compile (not yet tested).
7117 *) Additional typesafe stacks.
7120 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7124 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7126 *) New configuration variant "sco5-gcc".
7128 *) Updated some demos.
7129 [Sean O Riordain, Wade Scholine]
7131 *) Add missing BIO_free at exit of pkcs12 application.
7134 *) Fix memory leak in conf.c.
7137 *) Updates for Win32 to assembler version of MD5.
7140 *) Set #! path to perl in apps/der_chop to where we found it
7141 instead of using a fixed path.
7144 *) SHA library changes for irix64-mips4-cc.
7147 *) Improvements for VMS support.
7151 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7153 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7154 This also avoids the problems with SC4.2 and unpatched SC5.
7155 [Andy Polyakov <appro@fy.chalmers.se>]
7157 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7158 These are required because of the typesafe stack would otherwise break
7159 existing code. If old code used a structure member which used to be STACK
7160 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7161 sk_num or sk_value it would produce an error because the num, data members
7162 are not present in STACK_OF. Now it just produces a warning. sk_set
7163 replaces the old method of assigning a value to sk_value
7164 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7165 that does this will no longer work (and should use sk_set instead) but
7166 this could be regarded as a "questionable" behaviour anyway.
7169 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7170 correctly handle encrypted S/MIME data.
7173 *) Change type of various DES function arguments from des_cblock
7174 (which means, in function argument declarations, pointer to char)
7175 to des_cblock * (meaning pointer to array with 8 char elements),
7176 which allows the compiler to do more typechecking; it was like
7177 that back in SSLeay, but with lots of ugly casts.
7179 Introduce new type const_des_cblock.
7182 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7183 problems: find RecipientInfo structure that matches recipient certificate
7184 and initialise the ASN1 structures properly based on passed cipher.
7187 *) Belatedly make the BN tests actually check the results.
7190 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7191 to and from BNs: it was completely broken. New compilation option
7192 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7193 key elements as negative integers.
7196 *) Reorganize and speed up MD5.
7197 [Andy Polyakov <appro@fy.chalmers.se>]
7200 [Richard Levitte <richard@levitte.org>]
7202 *) New option -out to asn1parse to allow the parsed structure to be
7203 output to a file. This is most useful when combined with the -strparse
7204 option to examine the output of things like OCTET STRINGS.
7207 *) Make SSL library a little more fool-proof by not requiring any longer
7208 that SSL_set_{accept,connect}_state be called before
7209 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7210 in many applications because usually everything *appeared* to work as
7211 intended anyway -- now it really works as intended).
7214 *) Move openssl.cnf out of lib/.
7217 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7218 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7219 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7220 [Ralf S. Engelschall]
7222 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7223 handle PKCS#7 enveloped data properly.
7224 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7226 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7227 copying pointers. The cert_st handling is changed by this in
7228 various ways (and thus what used to be known as ctx->default_cert
7229 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7230 any longer when s->cert does not give us what we need).
7231 ssl_cert_instantiate becomes obsolete by this change.
7232 As soon as we've got the new code right (possibly it already is?),
7233 we have solved a couple of bugs of the earlier code where s->cert
7234 was used as if it could not have been shared with other SSL structures.
7236 Note that using the SSL API in certain dirty ways now will result
7237 in different behaviour than observed with earlier library versions:
7238 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7239 does not influence s as it used to.
7241 In order to clean up things more thoroughly, inside SSL_SESSION
7242 we don't use CERT any longer, but a new structure SESS_CERT
7243 that holds per-session data (if available); currently, this is
7244 the peer's certificate chain and, for clients, the server's certificate
7245 and temporary key. CERT holds only those values that can have
7246 meaningful defaults in an SSL_CTX.
7249 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7250 from the internal representation. Various PKCS#7 fixes: remove some
7251 evil casts and set the enc_dig_alg field properly based on the signing
7255 *) Allow PKCS#12 password to be set from the command line or the
7256 environment. Let 'ca' get its config file name from the environment
7257 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7261 *) Allow certificate policies extension to use an IA5STRING for the
7262 organization field. This is contrary to the PKIX definition but
7263 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7267 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7268 without disallowing inline assembler and the like for non-pedantic builds.
7271 *) Support Borland C++ builder.
7272 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7277 *) SHA-1 cleanups and performance enhancements.
7278 [Andy Polyakov <appro@fy.chalmers.se>]
7280 *) Sparc v8plus assembler for the bignum library.
7281 [Andy Polyakov <appro@fy.chalmers.se>]
7283 *) Accept any -xxx and +xxx compiler options in Configure.
7286 *) Update HPUX configuration.
7289 *) Add missing sk_<type>_unshift() function to safestack.h
7290 [Ralf S. Engelschall]
7292 *) New function SSL_CTX_use_certificate_chain_file that sets the
7293 "extra_cert"s in addition to the certificate. (This makes sense
7294 only for "PEM" format files, as chains as a whole are not
7298 *) Support verify_depth from the SSL API.
7299 x509_vfy.c had what can be considered an off-by-one-error:
7300 Its depth (which was not part of the external interface)
7301 was actually counting the number of certificates in a chain;
7302 now it really counts the depth.
7305 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7306 instead of X509err, which often resulted in confusing error
7307 messages since the error codes are not globally unique
7308 (e.g. an alleged error in ssl3_accept when a certificate
7309 didn't match the private key).
7311 *) New function SSL_CTX_set_session_id_context that allows to set a default
7312 value (so that you don't need SSL_set_session_id_context for each
7313 connection using the SSL_CTX).
7316 *) OAEP decoding bug fix.
7319 *) Support INSTALL_PREFIX for package builders, as proposed by
7323 *) New Configure options "threads" and "no-threads". For systems
7324 where the proper compiler options are known (currently Solaris
7325 and Linux), "threads" is the default.
7328 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7331 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7332 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7333 such as /usr/local/bin.
7336 *) "make linux-shared" to build shared libraries.
7337 [Niels Poppe <niels@netbox.org>]
7339 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7342 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7343 extension adding in x509 utility.
7346 *) Remove NOPROTO sections and error code comments.
7349 *) Partial rewrite of the DEF file generator to now parse the ANSI
7353 *) New Configure options --prefix=DIR and --openssldir=DIR.
7356 *) Complete rewrite of the error code script(s). It is all now handled
7357 by one script at the top level which handles error code gathering,
7358 header rewriting and C source file generation. It should be much better
7359 than the old method: it now uses a modified version of Ulf's parser to
7360 read the ANSI prototypes in all header files (thus the old K&R definitions
7361 aren't needed for error creation any more) and do a better job of
7362 translating function codes into names. The old 'ASN1 error code imbedded
7363 in a comment' is no longer necessary and it doesn't use .err files which
7364 have now been deleted. Also the error code call doesn't have to appear all
7365 on one line (which resulted in some large lines...).
7368 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7371 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7372 0 (which usually indicates a closed connection), but continue reading.
7375 *) Fix some race conditions.
7378 *) Add support for CRL distribution points extension. Add Certificate
7379 Policies and CRL distribution points documentation.
7382 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7385 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7386 8 of keying material. Merlin has also confirmed interop with this fix
7387 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7388 [Merlin Hughes <merlin@baltimore.ie>]
7390 *) Fix lots of warnings.
7391 [Richard Levitte <levitte@stacken.kth.se>]
7393 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7394 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7395 [Richard Levitte <levitte@stacken.kth.se>]
7397 *) Fix problems with sizeof(long) == 8.
7398 [Andy Polyakov <appro@fy.chalmers.se>]
7400 *) Change functions to ANSI C.
7403 *) Fix typos in error codes.
7404 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7406 *) Remove defunct assembler files from Configure.
7409 *) SPARC v8 assembler BIGNUM implementation.
7410 [Andy Polyakov <appro@fy.chalmers.se>]
7412 *) Support for Certificate Policies extension: both print and set.
7413 Various additions to support the r2i method this uses.
7416 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7417 return a const string when you are expecting an allocated buffer.
7420 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7421 types DirectoryString and DisplayText.
7424 *) Add code to allow r2i extensions to access the configuration database,
7425 add an LHASH database driver and add several ctx helper functions.
7428 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7429 fail when they extended the size of a BIGNUM.
7432 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7433 support typesafe stack.
7436 *) Fix typo in SSL_[gs]et_options().
7437 [Nils Frostberg <nils@medcom.se>]
7439 *) Delete various functions and files that belonged to the (now obsolete)
7440 old X509V3 handling code.
7443 *) New Configure option "rsaref".
7446 *) Don't auto-generate pem.h.
7449 *) Introduce type-safe ASN.1 SETs.
7452 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7453 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7455 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7456 that links with OpenSSL (well at least cause lots of warnings), but fear
7457 not: the conversion is trivial, and it eliminates loads of evil casts. A
7458 few STACKed things have been converted already. Feel free to convert more.
7459 In the fullness of time, I'll do away with the STACK type altogether.
7462 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7463 specified in <certfile> by updating the entry in the index.txt file.
7464 This way one no longer has to edit the index.txt file manually for
7465 revoking a certificate. The -revoke option does the gory details now.
7466 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7468 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7469 `-text' option at all and this way the `-noout -text' combination was
7470 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7471 [Ralf S. Engelschall]
7473 *) Make sure a corresponding plain text error message exists for the
7474 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7475 verify callback function determined that a certificate was revoked.
7476 [Ralf S. Engelschall]
7478 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7479 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7480 all available cipers including rc5, which was forgotten until now.
7481 In order to let the testing shell script know which algorithms
7482 are available, a new (up to now undocumented) command
7483 "openssl list-cipher-commands" is used.
7486 *) Bugfix: s_client occasionally would sleep in select() when
7487 it should have checked SSL_pending() first.
7490 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7491 the raw DSA values prior to ASN.1 encoding.
7494 *) Tweaks to Configure
7495 [Niels Poppe <niels@netbox.org>]
7497 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7501 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7504 *) New config option to avoid instructions that are illegal on the 80386.
7505 The default code is faster, but requires at least a 486.
7508 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7509 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7510 same as SSL2_VERSION anyway.
7513 *) New "-showcerts" option for s_client.
7516 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7517 application. Various cleanups and fixes.
7520 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7521 modify error routines to work internally. Add error codes and PBE init
7522 to library startup routines.
7525 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7526 packing functions to asn1 and evp. Changed function names and error
7527 codes along the way.
7530 *) PKCS12 integration: and so it begins... First of several patches to
7531 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7532 objects to objects.h
7535 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7536 and display support for Thawte strong extranet extension.
7539 *) Add LinuxPPC support.
7540 [Jeff Dubrule <igor@pobox.org>]
7542 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7543 bn_div_words in alpha.s.
7544 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7546 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7547 OAEP isn't supported when OpenSSL is built with RSAref.
7548 [Ulf Moeller <ulf@fitug.de>]
7550 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7551 so they no longer are missing under -DNOPROTO.
7552 [Soren S. Jorvang <soren@t.dk>]
7555 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7557 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7558 doesn't work when the session is reused. Coming soon!
7561 *) Fix a security hole, that allows sessions to be reused in the wrong
7562 context thus bypassing client cert protection! All software that uses
7563 client certs and session caches in multiple contexts NEEDS PATCHING to
7564 allow session reuse! A fuller solution is in the works.
7565 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7567 *) Some more source tree cleanups (removed obsolete files
7568 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7569 permission on "config" script to be executable) and a fix for the INSTALL
7571 [Ulf Moeller <ulf@fitug.de>]
7573 *) Remove some legacy and erroneous uses of malloc, free instead of
7575 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7577 *) Make rsa_oaep_test return non-zero on error.
7578 [Ulf Moeller <ulf@fitug.de>]
7580 *) Add support for native Solaris shared libraries. Configure
7581 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7582 if someone would make that last step automatic.
7583 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7585 *) ctx_size was not built with the right compiler during "make links". Fixed.
7588 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7589 except NULL ciphers". This means the default cipher list will no longer
7590 enable NULL ciphers. They need to be specifically enabled e.g. with
7591 the string "DEFAULT:eNULL".
7594 *) Fix to RSA private encryption routines: if p < q then it would
7595 occasionally produce an invalid result. This will only happen with
7596 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7599 *) Be less restrictive and allow also `perl util/perlpath.pl
7600 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7601 because this way one can also use an interpreter named `perl5' (which is
7602 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7603 installed as `perl').
7604 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7606 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7607 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7609 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7610 advapi32.lib to Win32 build and change the pem test comparision
7611 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7612 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7613 and crypto/des/ede_cbcm_enc.c.
7616 *) DES quad checksum was broken on big-endian architectures. Fixed.
7619 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7620 Win32 test batch file so it (might) work again. The Win32 test batch file
7621 is horrible: I feel ill....
7624 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7625 in e_os.h. Audit of header files to check ANSI and non ANSI
7626 sections: 10 functions were absent from non ANSI section and not exported
7627 from Windows DLLs. Fixed up libeay.num for new functions.
7630 *) Make `openssl version' output lines consistent.
7631 [Ralf S. Engelschall]
7633 *) Fix Win32 symbol export lists for BIO functions: Added
7634 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7635 to ms/libeay{16,32}.def.
7636 [Ralf S. Engelschall]
7638 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7639 fine under Unix and passes some trivial tests I've now added. But the
7640 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7641 added to make sure no one expects that this stuff really works in the
7642 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7643 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7645 [Ralf S. Engelschall]
7647 *) Fix the generation of two part addresses in perl.
7648 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7650 *) Add config entry for Linux on MIPS.
7651 [John Tobey <jtobey@channel1.com>]
7653 *) Make links whenever Configure is run, unless we are on Windoze.
7656 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7657 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7661 *) Add a useful kludge to allow package maintainers to specify compiler and
7662 other platforms details on the command line without having to patch the
7663 Configure script everytime: One now can use ``perl Configure
7664 <id>:<details>'', i.e. platform ids are allowed to have details appended
7665 to them (seperated by colons). This is treated as there would be a static
7666 pre-configured entry in Configure's %table under key <id> with value
7667 <details> and ``perl Configure <id>'' is called. So, when you want to
7668 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7669 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7670 now, which overrides the FreeBSD-elf entry on-the-fly.
7671 [Ralf S. Engelschall]
7673 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7676 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7677 on the `perl Configure ...' command line. This way one can compile
7678 OpenSSL libraries with Position Independent Code (PIC) which is needed
7679 for linking it into DSOs.
7680 [Ralf S. Engelschall]
7682 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7686 *) Cleaned up the LICENSE document: The official contact for any license
7687 questions now is the OpenSSL core team under openssl-core@openssl.org.
7688 And add a paragraph about the dual-license situation to make sure people
7689 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7690 to the OpenSSL toolkit.
7691 [Ralf S. Engelschall]
7693 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7694 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7695 Additonally cleaned up the `make links' target: Remove unnecessary
7696 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7697 to speed processing and no longer clutter the display with confusing
7698 stuff. Instead only the actually done links are displayed.
7699 [Ralf S. Engelschall]
7701 *) Permit null encryption ciphersuites, used for authentication only. It used
7702 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7703 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7707 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7708 signed attributes when verifying signatures (this would break them),
7709 the detached data encoding was wrong and public keys obtained using
7710 X509_get_pubkey() weren't freed.
7713 *) Add text documentation for the BUFFER functions. Also added a work around
7714 to a Win95 console bug. This was triggered by the password read stuff: the
7715 last character typed gets carried over to the next fread(). If you were
7716 generating a new cert request using 'req' for example then the last
7717 character of the passphrase would be CR which would then enter the first
7721 *) Added the new `Includes OpenSSL Cryptography Software' button as
7722 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7723 button and can be used by applications based on OpenSSL to show the
7724 relationship to the OpenSSL project.
7725 [Ralf S. Engelschall]
7727 *) Remove confusing variables in function signatures in files
7728 ssl/ssl_lib.c and ssl/ssl.h.
7729 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7731 *) Don't install bss_file.c under PREFIX/include/
7732 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7734 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7735 functions that return function pointers and has support for NT specific
7736 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7737 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7738 unsigned to signed types: this was killing the Win32 compile.
7741 *) Add new certificate file to stack functions,
7742 SSL_add_dir_cert_subjects_to_stack() and
7743 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7744 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7745 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7746 This means that Apache-SSL and similar packages don't have to mess around
7747 to add as many CAs as they want to the preferred list.
7750 *) Experiment with doxygen documentation. Currently only partially applied to
7752 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7753 openssl.doxy as the configuration file.
7756 *) Get rid of remaining C++-style comments which strict C compilers hate.
7757 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7759 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7760 compiled in by default: it has problems with large keys.
7763 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7764 DH private keys and/or callback functions which directly correspond to
7765 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7766 is needed for applications which have to configure certificates on a
7767 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7769 For the RSA certificate situation is makes no difference, but
7770 for the DSA certificate situation this fixes the "no shared cipher"
7771 problem where the OpenSSL cipher selection procedure failed because the
7772 temporary keys were not overtaken from the context and the API provided
7773 no way to reconfigure them.
7774 The new functions now let applications reconfigure the stuff and they
7775 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7776 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7777 non-public-API function ssl_cert_instantiate() is used as a helper
7778 function and also to reduce code redundancy inside ssl_rsa.c.
7779 [Ralf S. Engelschall]
7781 *) Move s_server -dcert and -dkey options out of the undocumented feature
7782 area because they are useful for the DSA situation and should be
7783 recognized by the users.
7784 [Ralf S. Engelschall]
7786 *) Fix the cipher decision scheme for export ciphers: the export bits are
7787 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7788 SSL_EXP_MASK. So, the original variable has to be used instead of the
7789 already masked variable.
7790 [Richard Levitte <levitte@stacken.kth.se>]
7792 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7793 [Richard Levitte <levitte@stacken.kth.se>]
7795 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7796 from `int' to `unsigned int' because it's a length and initialized by
7797 EVP_DigestFinal() which expects an `unsigned int *'.
7798 [Richard Levitte <levitte@stacken.kth.se>]
7800 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7801 script. Instead use the usual Shell->Perl transition trick.
7802 [Ralf S. Engelschall]
7804 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7805 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7806 -noout -modulus' as it's already the case for `openssl rsa -noout
7807 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7808 currently the public key is printed (a decision which was already done by
7809 `openssl dsa -modulus' in the past) which serves a similar purpose.
7810 Additionally the NO_RSA no longer completely removes the whole -modulus
7811 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7813 [Ralf S. Engelschall]
7815 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7816 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7817 [Arne Ansper <arne@ats.cyber.ee>]
7819 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7820 to be added. Now both 'req' and 'ca' can use new objects defined in the
7824 *) Add cool BIO that does syslog (or event log on NT).
7825 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7827 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7828 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7829 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7830 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7833 *) Add preliminary config info for new extension code.
7836 *) Make RSA_NO_PADDING really use no padding.
7837 [Ulf Moeller <ulf@fitug.de>]
7839 *) Generate errors when private/public key check is done.
7842 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7843 for some CRL extensions and new objects added.
7846 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7847 key usage extension and fuller support for authority key id.
7850 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7851 padding method for RSA, which is recommended for new applications in PKCS
7852 #1 v2.0 (RFC 2437, October 1998).
7853 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7854 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7855 against Bleichbacher's attack on RSA.
7856 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7859 *) Updates to the new SSL compression code
7860 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7862 *) Fix so that the version number in the master secret, when passed
7863 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7864 (because the server will not accept higher), that the version number
7865 is 0x03,0x01, not 0x03,0x00
7866 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7868 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7869 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7870 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7873 *) Support for RAW extensions where an arbitrary extension can be
7874 created by including its DER encoding. See apps/openssl.cnf for
7878 *) Make sure latest Perl versions don't interpret some generated C array
7879 code as Perl array code in the crypto/err/err_genc.pl script.
7880 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7882 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7883 not many people have the assembler. Various Win32 compilation fixes and
7884 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7888 *) Modify configure script 'Configure' to automatically create crypto/date.h
7889 file under Win32 and also build pem.h from pem.org. New script
7890 util/mkfiles.pl to create the MINFO file on environments that can't do a
7891 'make files': perl util/mkfiles.pl >MINFO should work.
7894 *) Major rework of DES function declarations, in the pursuit of correctness
7895 and purity. As a result, many evil casts evaporated, and some weirdness,
7896 too. You may find this causes warnings in your code. Zapping your evil
7897 casts will probably fix them. Mostly.
7900 *) Fix for a typo in asn1.h. Bug fix to object creation script
7901 obj_dat.pl. It considered a zero in an object definition to mean
7902 "end of object": none of the objects in objects.h have any zeros
7903 so it wasn't spotted.
7904 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7906 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7907 Masking (CBCM). In the absence of test vectors, the best I have been able
7908 to do is check that the decrypt undoes the encrypt, so far. Send me test
7909 vectors if you have them.
7912 *) Correct calculation of key length for export ciphers (too much space was
7913 allocated for null ciphers). This has not been tested!
7916 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7917 message is now correct (it understands "crypto" and "ssl" on its
7918 command line). There is also now an "update" option. This will update
7919 the util/ssleay.num and util/libeay.num files with any new functions.
7921 perl util/mkdef.pl crypto ssl update
7922 it will update them.
7925 *) Overhauled the Perl interface (perl/*):
7926 - ported BN stuff to OpenSSL's different BN library
7927 - made the perl/ source tree CVS-aware
7928 - renamed the package from SSLeay to OpenSSL (the files still contain
7929 their history because I've copied them in the repository)
7930 - removed obsolete files (the test scripts will be replaced
7931 by better Test::Harness variants in the future)
7932 [Ralf S. Engelschall]
7934 *) First cut for a very conservative source tree cleanup:
7935 1. merge various obsolete readme texts into doc/ssleay.txt
7936 where we collect the old documents and readme texts.
7937 2. remove the first part of files where I'm already sure that we no
7938 longer need them because of three reasons: either they are just temporary
7939 files which were left by Eric or they are preserved original files where
7940 I've verified that the diff is also available in the CVS via "cvs diff
7941 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7942 the crypto/md/ stuff).
7943 [Ralf S. Engelschall]
7945 *) More extension code. Incomplete support for subject and issuer alt
7946 name, issuer and authority key id. Change the i2v function parameters
7947 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7948 what that's for :-) Fix to ASN1 macro which messed up
7949 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7952 *) Preliminary support for ENUMERATED type. This is largely copied from the
7956 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7957 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7959 *) Make sure `make rehash' target really finds the `openssl' program.
7960 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7962 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7963 like to hear about it if this slows down other processors.
7966 *) Add CygWin32 platform information to Configure script.
7967 [Alan Batie <batie@aahz.jf.intel.com>]
7969 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7970 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7972 *) New program nseq to manipulate netscape certificate sequences
7975 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7979 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7980 but the BN code had some problems that would cause failures when
7981 doing certificate verification and some other functions.
7982 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7984 *) Add ASN1 and PEM code to support netscape certificate sequences.
7987 *) Add ASN1 and PEM code to support netscape certificate sequences.
7990 *) Add several PKIX and private extended key usage OIDs.
7993 *) Modify the 'ca' program to handle the new extension code. Modify
7994 openssl.cnf for new extension format, add comments.
7997 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7998 and add a sample to openssl.cnf so req -x509 now adds appropriate
8002 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8003 error code, add initial support to X509_print() and x509 application.
8006 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8007 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8008 stuff is currently isolated and isn't even compiled yet.
8011 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8012 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8013 Removed the versions check from X509 routines when loading extensions:
8014 this allows certain broken certificates that don't set the version
8015 properly to be processed.
8018 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8019 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8020 can still be regenerated with "make depend".
8023 *) Spelling mistake in C version of CAST-128.
8024 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8026 *) Changes to the error generation code. The perl script err-code.pl
8027 now reads in the old error codes and retains the old numbers, only
8028 adding new ones if necessary. It also only changes the .err files if new
8029 codes are added. The makefiles have been modified to only insert errors
8030 when needed (to avoid needlessly modifying header files). This is done
8031 by only inserting errors if the .err file is newer than the auto generated
8032 C file. To rebuild all the error codes from scratch (the old behaviour)
8033 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8034 or delete all the .err files.
8037 *) CAST-128 was incorrectly implemented for short keys. The C version has
8038 been fixed, but is untested. The assembler versions are also fixed, but
8039 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8040 to regenerate it if needed.
8041 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8042 Hagino <itojun@kame.net>]
8044 *) File was opened incorrectly in randfile.c.
8045 [Ulf Möller <ulf@fitug.de>]
8047 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8048 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8049 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8050 al: it's just almost always a UTCTime. Note this patch adds new error
8051 codes so do a "make errors" if there are problems.
8054 *) Correct Linux 1 recognition in config.
8055 [Ulf Möller <ulf@fitug.de>]
8057 *) Remove pointless MD5 hash when using DSA keys in ca.
8058 [Anonymous <nobody@replay.com>]
8060 *) Generate an error if given an empty string as a cert directory. Also
8061 generate an error if handed NULL (previously returned 0 to indicate an
8062 error, but didn't set one).
8063 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8065 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8068 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8069 parameters. This was causing a warning which killed off the Win32 compile.
8072 *) Remove C++ style comments from crypto/bn/bn_local.h.
8073 [Neil Costigan <neil.costigan@celocom.com>]
8075 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8076 based on a text string, looking up short and long names and finally
8077 "dot" format. The "dot" format stuff didn't work. Added new function
8078 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8079 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8080 OID is not part of the table.
8083 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8084 X509_LOOKUP_by_alias().
8087 *) Sort openssl functions by name.
8090 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8091 encryption from sample DSA keys (in case anyone is interested the password
8095 *) Make _all_ *_free functions accept a NULL pointer.
8096 [Frans Heymans <fheymans@isaserver.be>]
8098 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8100 [Anonymous <nobody@replay.com>]
8102 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8103 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8105 *) Don't blow it for numeric -newkey arguments to apps/req.
8106 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8108 *) Temp key "for export" tests were wrong in s3_srvr.c.
8109 [Anonymous <nobody@replay.com>]
8111 *) Add prototype for temp key callback functions
8112 SSL_CTX_set_tmp_{rsa,dh}_callback().
8115 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8116 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8119 *) X509_name_add_entry() freed the wrong thing after an error.
8120 [Arne Ansper <arne@ats.cyber.ee>]
8122 *) rsa_eay.c would attempt to free a NULL context.
8123 [Arne Ansper <arne@ats.cyber.ee>]
8125 *) BIO_s_socket() had a broken should_retry() on Windoze.
8126 [Arne Ansper <arne@ats.cyber.ee>]
8128 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8129 [Arne Ansper <arne@ats.cyber.ee>]
8131 *) Make sure the already existing X509_STORE->depth variable is initialized
8132 in X509_STORE_new(), but document the fact that this variable is still
8133 unused in the certificate verification process.
8134 [Ralf S. Engelschall]
8136 *) Fix the various library and apps files to free up pkeys obtained from
8137 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8140 *) Fix reference counting in X509_PUBKEY_get(). This makes
8141 demos/maurice/example2.c work, amongst others, probably.
8142 [Steve Henson and Ben Laurie]
8144 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8145 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8146 are no longer created. This way we have a single and consistent command
8147 line interface `openssl <command>', similar to `cvs <command>'.
8148 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8150 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8151 BIT STRING wrapper always have zero unused bits.
8154 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8157 *) Make the top-level INSTALL documentation easier to understand.
8160 *) Makefiles updated to exit if an error occurs in a sub-directory
8161 make (including if user presses ^C) [Paul Sutton]
8163 *) Make Montgomery context stuff explicit in RSA data structure.
8166 *) Fix build order of pem and err to allow for generated pem.h.
8169 *) Fix renumbering bug in X509_NAME_delete_entry().
8172 *) Enhanced the err-ins.pl script so it makes the error library number
8173 global and can add a library name. This is needed for external ASN1 and
8174 other error libraries.
8177 *) Fixed sk_insert which never worked properly.
8180 *) Fix ASN1 macros so they can handle indefinite length construted
8181 EXPLICIT tags. Some non standard certificates use these: they can now
8185 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8186 into a single doc/ssleay.txt bundle. This way the information is still
8187 preserved but no longer messes up this directory. Now it's new room for
8188 the new set of documenation files.
8189 [Ralf S. Engelschall]
8191 *) SETs were incorrectly DER encoded. This was a major pain, because they
8192 shared code with SEQUENCEs, which aren't coded the same. This means that
8193 almost everything to do with SETs or SEQUENCEs has either changed name or
8194 number of arguments.
8195 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8197 *) Fix test data to work with the above.
8200 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8201 was already fixed by Eric for 0.9.1 it seems.
8202 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8204 *) Autodetect FreeBSD3.
8207 *) Fix various bugs in Configure. This affects the following platforms:
8211 unixware-2.0-pentium
8215 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8216 before they are needed.
8219 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8223 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8225 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8226 changed SSLeay to OpenSSL in version strings.
8227 [Ralf S. Engelschall]
8229 *) Some fixups to the top-level documents.
8232 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8233 because the symlink to include/ was missing.
8234 [Ralf S. Engelschall]
8236 *) Incorporated the popular no-RSA/DSA-only patches
8237 which allow to compile a RSA-free SSLeay.
8238 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8240 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8241 when "ssleay" is still not found.
8242 [Ralf S. Engelschall]
8244 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8245 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8247 *) Updated the README file.
8248 [Ralf S. Engelschall]
8250 *) Added various .cvsignore files in the CVS repository subdirs
8251 to make a "cvs update" really silent.
8252 [Ralf S. Engelschall]
8254 *) Recompiled the error-definition header files and added
8255 missing symbols to the Win32 linker tables.
8256 [Ralf S. Engelschall]
8258 *) Cleaned up the top-level documents;
8259 o new files: CHANGES and LICENSE
8260 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8261 o merged COPYRIGHT into LICENSE
8262 o removed obsolete TODO file
8263 o renamed MICROSOFT to INSTALL.W32
8264 [Ralf S. Engelschall]
8266 *) Removed dummy files from the 0.9.1b source tree:
8267 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8268 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8269 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8270 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8271 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8272 [Ralf S. Engelschall]
8274 *) Added various platform portability fixes.
8277 *) The Genesis of the OpenSSL rpject:
8278 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8279 Young and Tim J. Hudson created while they were working for C2Net until
8281 [The OpenSSL Project]
8284 Changes between 0.9.0b and 0.9.1b [not released]
8286 *) Updated a few CA certificates under certs/
8289 *) Changed some BIGNUM api stuff.
8292 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8293 DGUX x86, Linux Alpha, etc.
8296 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8297 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8301 *) Add -strparse option to asn1pars program which parses nested
8303 [Dr Stephen Henson <shenson@bigfoot.com>]
8305 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8308 *) DSA fix for "ca" program.
8311 *) Added "-genkey" option to "dsaparam" program.
8314 *) Added RIPE MD160 (rmd160) message digest.
8317 *) Added -a (all) option to "ssleay version" command.
8320 *) Added PLATFORM define which is the id given to Configure.
8323 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8326 *) Extended the ASN.1 parser routines.
8329 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8332 *) Added a BN_CTX to the BN library.
8335 *) Fixed the weak key values in DES library
8338 *) Changed API in EVP library for cipher aliases.
8341 *) Added support for RC2/64bit cipher.
8344 *) Converted the lhash library to the crypto/mem.c functions.
8347 *) Added more recognized ASN.1 object ids.
8350 *) Added more RSA padding checks for SSL/TLS.
8353 *) Added BIO proxy/filter functionality.
8356 *) Added extra_certs to SSL_CTX which can be used
8357 send extra CA certificates to the client in the CA cert chain sending
8358 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8361 *) Now Fortezza is denied in the authentication phase because
8362 this is key exchange mechanism is not supported by SSLeay at all.
8365 *) Additional PKCS1 checks.
8368 *) Support the string "TLSv1" for all TLS v1 ciphers.
8371 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8372 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8375 *) Fixed a few memory leaks.
8378 *) Fixed various code and comment typos.
8381 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8382 bytes sent in the client random.
8383 [Edward Bishop <ebishop@spyglass.com>]