5 Changes between 0.9.8k and 0.9.8l [xx XXX xxxx]
7 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
8 X690 8.9.12 and can produce some misleading textual output of OIDs.
9 [Steve Henson, reported by Dan Kaminsky]
11 *) Delete MD2 from algorithm tables. This follows the recommendation in
12 several standards that it is not used in new applications due to
13 several cryptographic weaknesses. For binary compatibility reasons
14 the MD2 API is still compiled in by default.
17 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
21 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
22 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
24 [Guenter <lists@gknw.net>]
26 *) Fix the server certificate chain building code to use X509_verify_cert(),
27 it used to have an ad-hoc builder which was unable to cope with anything
28 other than a simple chain.
29 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
31 *) Don't check self signed certificate signatures in X509_verify_cert()
32 by default (a flag can override this): it just wastes time without
33 adding any security. As a useful side effect self signed root CAs
34 with non-FIPS digests are now usable in FIPS mode.
37 *) In dtls1_process_out_of_seq_message() the check if the current message
38 is already buffered was missing. For every new message was memory
39 allocated, allowing an attacker to perform an denial of service attack
40 with sending out of seq handshake messages until there is no memory
41 left. Additionally every future messege was buffered, even if the
42 sequence number made no sense and would be part of another handshake.
43 So only messages with sequence numbers less than 10 in advance will be
44 buffered. (CVE-2009-1378)
45 [Robin Seggelmann, discovered by Daniel Mentz]
47 *) Records are buffered if they arrive with a future epoch to be
48 processed after finishing the corresponding handshake. There is
49 currently no limitation to this buffer allowing an attacker to perform
50 a DOS attack with sending records with future epochs until there is no
51 memory left. This patch adds the pqueue_size() function to detemine
52 the size of a buffer and limits the record buffer to 100 entries.
54 [Robin Seggelmann, discovered by Daniel Mentz]
56 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
57 parent structure is freed. (CVE-2009-1379)
60 *) Handle non-blocking I/O properly in SSL_shutdown() call.
61 [Darryl Miles <darryl-mailinglists@netbauds.net>]
64 [Ilya O. <vrghost@gmail.com>]
66 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
68 *) Don't set val to NULL when freeing up structures, it is freed up by
69 underlying code. If sizeof(void *) > sizeof(long) this can result in
70 zeroing past the valid field. (CVE-2009-0789)
71 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
73 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
74 checked correctly. This would allow some invalid signed attributes to
75 appear to verify correctly. (CVE-2009-0591)
76 [Ivan Nestlerode <inestlerode@us.ibm.com>]
78 *) Reject UniversalString and BMPString types with invalid lengths. This
79 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
80 a legal length. (CVE-2009-0590)
83 *) Set S/MIME signing as the default purpose rather than setting it
84 unconditionally. This allows applications to override it at the store
88 *) Permit restricted recursion of ASN1 strings. This is needed in practice
89 to handle some structures.
92 *) Improve efficiency of mem_gets: don't search whole buffer each time
94 [Jeremy Shapiro <jnshapir@us.ibm.com>]
96 *) New -hex option for openssl rand.
99 *) Print out UTF8String and NumericString when parsing ASN1.
102 *) Support NumericString type for name components.
105 *) Allow CC in the environment to override the automatically chosen
106 compiler. Note that nothing is done to ensure flags work with the
110 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
112 *) Properly check EVP_VerifyFinal() and similar return values
114 [Ben Laurie, Bodo Moeller, Google Security Team]
116 *) Enable TLS extensions by default.
119 *) Allow the CHIL engine to be loaded, whether the application is
120 multithreaded or not. (This does not release the developer from the
121 obligation to set up the dynamic locking callbacks.)
122 [Sander Temme <sander@temme.net>]
124 *) Use correct exit code if there is an error in dgst command.
125 [Steve Henson; problem pointed out by Roland Dirlewanger]
127 *) Tweak Configure so that you need to say "experimental-jpake" to enable
128 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
131 *) Add experimental JPAKE support, including demo authentication in
132 s_client and s_server.
135 *) Set the comparison function in v3_addr_canonize().
136 [Rob Austein <sra@hactrn.net>]
138 *) Add support for XMPP STARTTLS in s_client.
139 [Philip Paeps <philip@freebsd.org>]
141 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
142 to ensure that even with this option, only ciphersuites in the
143 server's preference list will be accepted. (Note that the option
144 applies only when resuming a session, so the earlier behavior was
145 just about the algorithm choice for symmetric cryptography.)
148 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
150 *) Fix NULL pointer dereference if a DTLS server recieved
151 ChangeCipherSpec as first record (CVE-2009-1386)
152 [Bodo Moeller, discovered by Alex Lam]
154 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
155 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
158 *) The fix in 0.9.8c that supposedly got rid of unsafe
159 double-checked locking was incomplete for RSA blinding,
160 addressing just one layer of what turns out to have been
161 doubly unsafe triple-checked locking.
163 So now fix this for real by retiring the MONT_HELPER macro
164 in crypto/rsa/rsa_eay.c.
166 [Bodo Moeller; problem pointed out by Marius Schilder]
168 *) Various precautionary measures:
170 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
172 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
173 (NB: This would require knowledge of the secret session ticket key
174 to exploit, in which case you'd be SOL either way.)
176 - Change bn_nist.c so that it will properly handle input BIGNUMs
177 outside the expected range.
179 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
182 [Neel Mehta, Bodo Moeller]
184 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
185 the load fails. Useful for distros.
186 [Ben Laurie and the FreeBSD team]
188 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
191 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
194 *) Expand ENGINE to support engine supplied SSL client certificate functions.
196 This work was sponsored by Logica.
199 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
200 keystores. Support for SSL/TLS client authentication too.
201 Not compiled unless enable-capieng specified to Configure.
203 This work was sponsored by Logica.
206 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
207 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
208 attribute creation routines such as certifcate requests and PKCS#12
212 Changes between 0.9.8g and 0.9.8h [28 May 2008]
214 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
215 handshake which could lead to a cilent crash as found using the
216 Codenomicon TLS test suite (CVE-2008-1672)
217 [Steve Henson, Mark Cox]
219 *) Fix double free in TLS server name extensions which could lead to
220 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
223 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
225 Clear the error queue to ensure that error entries left from
226 older function calls do not interfere with the correct operation.
227 [Lutz Jaenicke, Erik de Castro Lopo]
229 *) Remove root CA certificates of commercial CAs:
231 The OpenSSL project does not recommend any specific CA and does not
232 have any policy with respect to including or excluding any CA.
233 Therefore it does not make any sense to ship an arbitrary selection
234 of root CA certificates with the OpenSSL software.
237 *) RSA OAEP patches to fix two separate invalid memory reads.
238 The first one involves inputs when 'lzero' is greater than
239 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
240 before the beginning of from). The second one involves inputs where
241 the 'db' section contains nothing but zeroes (there is a one-byte
242 invalid read after the end of 'db').
243 [Ivan Nestlerode <inestlerode@us.ibm.com>]
245 *) Partial backport from 0.9.9-dev:
247 Introduce bn_mul_mont (dedicated Montgomery multiplication
248 procedure) as a candidate for BIGNUM assembler implementation.
249 While 0.9.9-dev uses assembler for various architectures, only
250 x86_64 is available by default here in the 0.9.8 branch, and
251 32-bit x86 is available through a compile-time setting.
253 To try the 32-bit x86 assembler implementation, use Configure
254 option "enable-montasm" (which exists only for this backport).
256 As "enable-montasm" for 32-bit x86 disclaims code stability
257 anyway, in this constellation we activate additional code
258 backported from 0.9.9-dev for further performance improvements,
259 namely BN_from_montgomery_word. (To enable this otherwise,
260 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
262 [Andy Polyakov (backport partially by Bodo Moeller)]
264 *) Add TLS session ticket callback. This allows an application to set
265 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
266 values. This is useful for key rollover for example where several key
267 sets may exist with different names.
270 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
271 This was broken until now in 0.9.8 releases, such that the only way
272 a registered ENGINE could be used (assuming it initialises
273 successfully on the host) was to explicitly set it as the default
274 for the relevant algorithms. This is in contradiction with 0.9.7
275 behaviour and the documentation. With this fix, when an ENGINE is
276 registered into a given algorithm's table of implementations, the
277 'uptodate' flag is reset so that auto-discovery will be used next
278 time a new context for that algorithm attempts to select an
280 [Ian Lister (tweaked by Geoff Thorpe)]
282 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
283 implemention in the following ways:
285 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
288 Lack of BER streaming support means one pass streaming processing is
289 only supported if data is detached: setting the streaming flag is
290 ignored for embedded content.
292 CMS support is disabled by default and must be explicitly enabled
293 with the enable-cms configuration option.
296 *) Update the GMP engine glue to do direct copies between BIGNUM and
297 mpz_t when openssl and GMP use the same limb size. Otherwise the
298 existing "conversion via a text string export" trick is still used.
299 [Paul Sheer <paulsheer@gmail.com>]
301 *) Zlib compression BIO. This is a filter BIO which compressed and
302 uncompresses any data passed through it.
305 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
306 RFC3394 compatible AES key wrapping.
309 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
310 sets string data without copying. X509_ALGOR_set0() and
311 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
312 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
313 from an X509_ATTRIBUTE structure optionally checking it occurs only
314 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
318 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
319 to get the expected BN_FLG_CONSTTIME behavior.
320 [Bodo Moeller (Google)]
324 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
325 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
326 - added some more tests to do_tests.pl
327 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
328 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
329 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
330 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
331 - various changes to netware.pl to enable gcc-cross builds on Win32
333 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
334 - various changes to fix missing prototype warnings
335 - fixed x86nasm.pl to create correct asm files for NASM COFF output
336 - added AES, WHIRLPOOL and CPUID assembler code to build files
337 - added missing AES assembler make rules to mk1mf.pl
338 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
339 [Guenter Knauf <eflash@gmx.net>]
341 *) Implement certificate status request TLS extension defined in RFC3546.
342 A client can set the appropriate parameters and receive the encoded
343 OCSP response via a callback. A server can query the supplied parameters
344 and set the encoded OCSP response in the callback. Add simplified examples
345 to s_client and s_server.
348 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
351 + Binary incompatibility of ssl_ctx_st structure
352 + DTLS interoperation with non-compliant servers
353 + Don't call get_session_cb() without proposed session
354 + Fix ia64 assembler code
355 [Andy Polyakov, Steve Henson]
357 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
359 *) DTLS Handshake overhaul. There were longstanding issues with
360 OpenSSL DTLS implementation, which were making it impossible for
361 RFC 4347 compliant client to communicate with OpenSSL server.
362 Unfortunately just fixing these incompatibilities would "cut off"
363 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
364 server keeps tolerating non RFC compliant syntax. The opposite is
365 not true, 0.9.8f client can not communicate with earlier server.
366 This update even addresses CVE-2007-4995.
369 *) Changes to avoid need for function casts in OpenSSL: some compilers
370 (gcc 4.2 and later) reject their use.
371 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
374 *) Add RFC4507 support to OpenSSL. This includes the corrections in
375 RFC4507bis. The encrypted ticket format is an encrypted encoded
376 SSL_SESSION structure, that way new session features are automatically
379 If a client application caches session in an SSL_SESSION structure
380 support is transparent because tickets are now stored in the encoded
383 The SSL_CTX structure automatically generates keys for ticket
384 protection in servers so again support should be possible
385 with no application modification.
387 If a client or server wishes to disable RFC4507 support then the option
388 SSL_OP_NO_TICKET can be set.
390 Add a TLS extension debugging callback to allow the contents of any client
391 or server extensions to be examined.
393 This work was sponsored by Google.
396 *) Add initial support for TLS extensions, specifically for the server_name
397 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
398 have new members for a host name. The SSL data structure has an
399 additional member SSL_CTX *initial_ctx so that new sessions can be
400 stored in that context to allow for session resumption, even after the
401 SSL has been switched to a new SSL_CTX in reaction to a client's
402 server_name extension.
404 New functions (subject to change):
407 SSL_get_servername_type()
410 New CTRL codes and macros (subject to change):
412 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
413 - SSL_CTX_set_tlsext_servername_callback()
414 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
415 - SSL_CTX_set_tlsext_servername_arg()
416 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
418 openssl s_client has a new '-servername ...' option.
420 openssl s_server has new options '-servername_host ...', '-cert2 ...',
421 '-key2 ...', '-servername_fatal' (subject to change). This allows
422 testing the HostName extension for a specific single host name ('-cert'
423 and '-key' remain fallbacks for handshakes without HostName
424 negotiation). If the unrecogninzed_name alert has to be sent, this by
425 default is a warning; it becomes fatal with the '-servername_fatal'
428 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
430 *) Add AES and SSE2 assembly language support to VC++ build.
433 *) Mitigate attack on final subtraction in Montgomery reduction.
436 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
437 (which previously caused an internal error).
440 *) Squeeze another 10% out of IGE mode when in != out.
443 *) AES IGE mode speedup.
444 [Dean Gaudet (Google)]
446 *) Add the Korean symmetric 128-bit cipher SEED (see
447 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
448 add SEED ciphersuites from RFC 4162:
450 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
451 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
452 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
453 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
455 To minimize changes between patchlevels in the OpenSSL 0.9.8
456 series, SEED remains excluded from compilation unless OpenSSL
457 is configured with 'enable-seed'.
460 *) Mitigate branch prediction attacks, which can be practical if a
461 single processor is shared, allowing a spy process to extract
462 information. For detailed background information, see
463 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
464 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
465 and Necessary Software Countermeasures"). The core of the change
466 are new versions BN_div_no_branch() and
467 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
468 respectively, which are slower, but avoid the security-relevant
469 conditional branches. These are automatically called by BN_div()
470 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
471 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
472 remove a conditional branch.
474 BN_FLG_CONSTTIME is the new name for the previous
475 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
476 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
477 in the exponent causes BN_mod_exp_mont() to use the alternative
478 implementation in BN_mod_exp_mont_consttime().) The old name
479 remains as a deprecated alias.
481 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
482 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
483 constant-time implementations for more than just exponentiation.
484 Here too the old name is kept as a deprecated alias.
486 BN_BLINDING_new() will now use BN_dup() for the modulus so that
487 the BN_BLINDING structure gets an independent copy of the
488 modulus. This means that the previous "BIGNUM *m" argument to
489 BN_BLINDING_new() and to BN_BLINDING_create_param() now
490 essentially becomes "const BIGNUM *m", although we can't actually
491 change this in the header file before 0.9.9. It allows
492 RSA_setup_blinding() to use BN_with_flags() on the modulus to
493 enable BN_FLG_CONSTTIME.
495 [Matthew D Wood (Intel Corp)]
497 *) In the SSL/TLS server implementation, be strict about session ID
498 context matching (which matters if an application uses a single
499 external cache for different purposes). Previously,
500 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
501 set. This did ensure strict client verification, but meant that,
502 with applications using a single external cache for quite
503 different requirements, clients could circumvent ciphersuite
504 restrictions for a given session ID context by starting a session
505 in a different context.
508 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
509 a ciphersuite string such as "DEFAULT:RSA" cannot enable
510 authentication-only ciphersuites.
513 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
514 not complete and could lead to a possible single byte overflow
515 (CVE-2007-5135) [Ben Laurie]
517 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
519 *) Since AES128 and AES256 (and similarly Camellia128 and
520 Camellia256) share a single mask bit in the logic of
521 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
522 kludge to work properly if AES128 is available and AES256 isn't
523 (or if Camellia128 is available and Camellia256 isn't).
526 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
527 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
528 When a point or a seed is encoded in a BIT STRING, we need to
529 prevent the removal of trailing zero bits to get the proper DER
530 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
531 of a NamedBitList, for which trailing 0 bits need to be removed.)
534 *) Have SSL/TLS server implementation tolerate "mismatched" record
535 protocol version while receiving ClientHello even if the
536 ClientHello is fragmented. (The server can't insist on the
537 particular protocol version it has chosen before the ServerHello
538 message has informed the client about his choice.)
541 *) Add RFC 3779 support.
542 [Rob Austein for ARIN, Ben Laurie]
544 *) Load error codes if they are not already present instead of using a
545 static variable. This allows them to be cleanly unloaded and reloaded.
546 Improve header file function name parsing.
549 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
550 or CAPABILITY handshake as required by RFCs.
553 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
555 *) Introduce limits to prevent malicious keys being able to
556 cause a denial of service. (CVE-2006-2940)
557 [Steve Henson, Bodo Moeller]
559 *) Fix ASN.1 parsing of certain invalid structures that can result
560 in a denial of service. (CVE-2006-2937) [Steve Henson]
562 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
563 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
565 *) Fix SSL client code which could crash if connecting to a
566 malicious SSLv2 server. (CVE-2006-4343)
567 [Tavis Ormandy and Will Drewry, Google Security Team]
569 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
570 match only those. Before that, "AES256-SHA" would be interpreted
571 as a pattern and match "AES128-SHA" too (since AES128-SHA got
572 the same strength classification in 0.9.7h) as we currently only
573 have a single AES bit in the ciphersuite description bitmap.
574 That change, however, also applied to ciphersuite strings such as
575 "RC4-MD5" that intentionally matched multiple ciphersuites --
576 namely, SSL 2.0 ciphersuites in addition to the more common ones
577 from SSL 3.0/TLS 1.0.
579 So we change the selection algorithm again: Naming an explicit
580 ciphersuite selects this one ciphersuite, and any other similar
581 ciphersuite (same bitmap) from *other* protocol versions.
582 Thus, "RC4-MD5" again will properly select both the SSL 2.0
583 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
585 Since SSL 2.0 does not have any ciphersuites for which the
586 128/256 bit distinction would be relevant, this works for now.
587 The proper fix will be to use different bits for AES128 and
588 AES256, which would have avoided the problems from the beginning;
589 however, bits are scarce, so we can only do this in a new release
590 (not just a patchlevel) when we can change the SSL_CIPHER
591 definition to split the single 'unsigned long mask' bitmap into
592 multiple values to extend the available space.
596 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
598 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
599 (CVE-2006-4339) [Ben Laurie and Google Security Team]
601 *) Add AES IGE and biIGE modes.
604 *) Change the Unix randomness entropy gathering to use poll() when
605 possible instead of select(), since the latter has some
606 undesirable limitations.
607 [Darryl Miles via Richard Levitte and Bodo Moeller]
609 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
610 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
611 cannot be implicitly activated as part of, e.g., the "AES" alias.
612 However, please upgrade to OpenSSL 0.9.9[-dev] for
613 non-experimental use of the ECC ciphersuites to get TLS extension
614 support, which is required for curve and point format negotiation
615 to avoid potential handshake problems.
618 *) Disable rogue ciphersuites:
620 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
621 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
622 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
624 The latter two were purportedly from
625 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
628 Also deactivate the remaining ciphersuites from
629 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
630 unofficial, and the ID has long expired.
633 *) Fix RSA blinding Heisenbug (problems sometimes occured on
634 dual-core machines) and other potential thread-safety issues.
637 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
638 versions), which is now available for royalty-free use
639 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
640 Also, add Camellia TLS ciphersuites from RFC 4132.
642 To minimize changes between patchlevels in the OpenSSL 0.9.8
643 series, Camellia remains excluded from compilation unless OpenSSL
644 is configured with 'enable-camellia'.
647 *) Disable the padding bug check when compression is in use. The padding
648 bug check assumes the first packet is of even length, this is not
649 necessarily true if compresssion is enabled and can result in false
650 positives causing handshake failure. The actual bug test is ancient
651 code so it is hoped that implementations will either have fixed it by
652 now or any which still have the bug do not support compression.
655 Changes between 0.9.8a and 0.9.8b [04 May 2006]
657 *) When applying a cipher rule check to see if string match is an explicit
658 cipher suite and only match that one cipher suite if it is.
661 *) Link in manifests for VC++ if needed.
662 [Austin Ziegler <halostatue@gmail.com>]
664 *) Update support for ECC-based TLS ciphersuites according to
665 draft-ietf-tls-ecc-12.txt with proposed changes (but without
666 TLS extensions, which are supported starting with the 0.9.9
667 branch, not in the OpenSSL 0.9.8 branch).
670 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
671 opaque EVP_CIPHER_CTX handling.
674 *) Fixes and enhancements to zlib compression code. We now only use
675 "zlib1.dll" and use the default __cdecl calling convention on Win32
676 to conform with the standards mentioned here:
677 http://www.zlib.net/DLL_FAQ.txt
678 Static zlib linking now works on Windows and the new --with-zlib-include
679 --with-zlib-lib options to Configure can be used to supply the location
680 of the headers and library. Gracefully handle case where zlib library
684 *) Several fixes and enhancements to the OID generation code. The old code
685 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
686 handle numbers larger than ULONG_MAX, truncated printing and had a
687 non standard OBJ_obj2txt() behaviour.
690 *) Add support for building of engines under engine/ as shared libraries
691 under VC++ build system.
694 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
695 Hopefully, we will not see any false combination of paths any more.
698 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
700 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
701 (part of SSL_OP_ALL). This option used to disable the
702 countermeasure against man-in-the-middle protocol-version
703 rollback in the SSL 2.0 server implementation, which is a bad
704 idea. (CVE-2005-2969)
706 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
707 for Information Security, National Institute of Advanced Industrial
708 Science and Technology [AIST], Japan)]
710 *) Add two function to clear and return the verify parameter flags.
713 *) Keep cipherlists sorted in the source instead of sorting them at
714 runtime, thus removing the need for a lock.
717 *) Avoid some small subgroup attacks in Diffie-Hellman.
718 [Nick Mathewson and Ben Laurie]
720 *) Add functions for well-known primes.
723 *) Extended Windows CE support.
724 [Satoshi Nakamura and Andy Polyakov]
726 *) Initialize SSL_METHOD structures at compile time instead of during
727 runtime, thus removing the need for a lock.
730 *) Make PKCS7_decrypt() work even if no certificate is supplied by
731 attempting to decrypt each encrypted key in turn. Add support to
735 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
737 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
740 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
743 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
744 key into the same file any more.
747 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
750 *) Add -utf8 command line and config file option to 'ca'.
751 [Stefan <stf@udoma.org]
753 *) Removed the macro des_crypt(), as it seems to conflict with some
754 libraries. Use DES_crypt().
757 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
758 involves renaming the source and generated shared-libs for
759 both. The engines will accept the corrected or legacy ids
760 ('ncipher' and '4758_cca' respectively) when binding. NB,
761 this only applies when building 'shared'.
762 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
764 *) Add attribute functions to EVP_PKEY structure. Modify
765 PKCS12_create() to recognize a CSP name attribute and
766 use it. Make -CSP option work again in pkcs12 utility.
769 *) Add new functionality to the bn blinding code:
770 - automatic re-creation of the BN_BLINDING parameters after
771 a fixed number of uses (currently 32)
772 - add new function for parameter creation
773 - introduce flags to control the update behaviour of the
774 BN_BLINDING parameters
775 - hide BN_BLINDING structure
776 Add a second BN_BLINDING slot to the RSA structure to improve
777 performance when a single RSA object is shared among several
781 *) Add support for DTLS.
782 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
784 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
785 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
788 *) Remove buggy and incompletet DH cert support from
789 ssl/ssl_rsa.c and ssl/s3_both.c
792 *) Use SHA-1 instead of MD5 as the default digest algorithm for
793 the apps/openssl applications.
796 *) Compile clean with "-Wall -Wmissing-prototypes
797 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
798 DEBUG_SAFESTACK must also be set.
801 *) Change ./Configure so that certain algorithms can be disabled by default.
802 The new counterpiece to "no-xxx" is "enable-xxx".
804 The patented RC5 and MDC2 algorithms will now be disabled unless
805 "enable-rc5" and "enable-mdc2", respectively, are specified.
807 (IDEA remains enabled despite being patented. This is because IDEA
808 is frequently required for interoperability, and there is no license
809 fee for non-commercial use. As before, "no-idea" can be used to
810 avoid this algorithm.)
814 *) Add processing of proxy certificates (see RFC 3820). This work was
815 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
816 EGEE (Enabling Grids for E-science in Europe).
819 *) RC4 performance overhaul on modern architectures/implementations, such
820 as Intel P4, IA-64 and AMD64.
823 *) New utility extract-section.pl. This can be used specify an alternative
824 section number in a pod file instead of having to treat each file as
825 a separate case in Makefile. This can be done by adding two lines to the
828 =for comment openssl_section:XXX
830 The blank line is mandatory.
834 *) New arguments -certform, -keyform and -pass for s_client and s_server
835 to allow alternative format key and certificate files and passphrase
839 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
840 update associated structures and add various utility functions.
842 Add new policy related verify parameters, include policy checking in
843 standard verify code. Enhance 'smime' application with extra parameters
844 to support policy checking and print out.
847 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
848 Nehemiah processors. These extensions support AES encryption in hardware
849 as well as RNG (though RNG support is currently disabled).
850 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
852 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
855 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
856 [Andy Polyakov and a number of other people]
858 *) Improved PowerPC platform support. Most notably BIGNUM assembler
859 implementation contributed by IBM.
860 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
862 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
863 exponent rather than 'unsigned long'. There is a corresponding change to
864 the new 'rsa_keygen' element of the RSA_METHOD structure.
865 [Jelte Jansen, Geoff Thorpe]
867 *) Functionality for creating the initial serial number file is now
868 moved from CA.pl to the 'ca' utility with a new option -create_serial.
870 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
871 number file to 1, which is bound to cause problems. To avoid
872 the problems while respecting compatibility between different 0.9.7
873 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
874 CA.pl for serial number initialization. With the new release 0.9.8,
875 we can fix the problem directly in the 'ca' utility.)
878 *) Reduced header interdepencies by declaring more opaque objects in
879 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
880 give fewer recursive includes, which could break lazy source code - so
881 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
882 developers should define this symbol when building and using openssl to
883 ensure they track the recommended behaviour, interfaces, [etc], but
884 backwards-compatible behaviour prevails when this isn't defined.
887 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
890 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
891 This will generate a random key of the appropriate length based on the
892 cipher context. The EVP_CIPHER can provide its own random key generation
893 routine to support keys of a specific form. This is used in the des and
894 3des routines to generate a key of the correct parity. Update S/MIME
895 code to use new functions and hence generate correct parity DES keys.
896 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
897 valid (weak or incorrect parity).
900 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
901 as looking them up. This is useful when the verified structure may contain
902 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
903 present unless the new PKCS7_NO_CRL flag is asserted.
906 *) Extend ASN1 oid configuration module. It now additionally accepts the
909 shortName = some long name, 1.2.3.4
912 *) Reimplemented the BN_CTX implementation. There is now no more static
913 limitation on the number of variables it can handle nor the depth of the
914 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
915 information can now expand as required, and rather than having a single
916 static array of bignums, BN_CTX now uses a linked-list of such arrays
917 allowing it to expand on demand whilst maintaining the usefulness of
921 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
922 to allow all RSA operations to function using a single BN_CTX.
925 *) Preliminary support for certificate policy evaluation and checking. This
926 is initially intended to pass the tests outlined in "Conformance Testing
927 of Relying Party Client Certificate Path Processing Logic" v1.07.
930 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
931 remained unused and not that useful. A variety of other little bignum
932 tweaks and fixes have also been made continuing on from the audit (see
936 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
937 associated ASN1, EVP and SSL functions and old ASN1 macros.
940 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
941 and this should never fail. So the return value from the use of
942 BN_set_word() (which can fail due to needless expansion) is now deprecated;
943 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
946 *) BN_CTX_get() should return zero-valued bignums, providing the same
947 initialised value as BN_new().
948 [Geoff Thorpe, suggested by Ulf Möller]
950 *) Support for inhibitAnyPolicy certificate extension.
953 *) An audit of the BIGNUM code is underway, for which debugging code is
954 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
955 is considered valid when processing BIGNUMs, and causes execution to
956 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
957 further steps are taken to deliberately pollute unused data in BIGNUM
958 structures to try and expose faulty code further on. For now, openssl will
959 (in its default mode of operation) continue to tolerate the inconsistent
960 forms that it has tolerated in the past, but authors and packagers should
961 consider trying openssl and their own applications when compiled with
962 these debugging symbols defined. It will help highlight potential bugs in
963 their own code, and will improve the test coverage for OpenSSL itself. At
964 some point, these tighter rules will become openssl's default to improve
965 maintainability, though the assert()s and other overheads will remain only
966 in debugging configurations. See bn.h for more details.
967 [Geoff Thorpe, Nils Larsch, Ulf Möller]
969 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
970 that can only be obtained through BN_CTX_new() (which implicitly
971 initialises it). The presence of this function only made it possible
972 to overwrite an existing structure (and cause memory leaks).
975 *) Because of the callback-based approach for implementing LHASH as a
976 template type, lh_insert() adds opaque objects to hash-tables and
977 lh_doall() or lh_doall_arg() are typically used with a destructor callback
978 to clean up those corresponding objects before destroying the hash table
979 (and losing the object pointers). So some over-zealous constifications in
980 LHASH have been relaxed so that lh_insert() does not take (nor store) the
981 objects as "const" and the lh_doall[_arg] callback wrappers are not
982 prototyped to have "const" restrictions on the object pointers they are
983 given (and so aren't required to cast them away any more).
986 *) The tmdiff.h API was so ugly and minimal that our own timing utility
987 (speed) prefers to use its own implementation. The two implementations
988 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
989 its object type properly exposed (MS_TM) instead of casting to/from "char
990 *". This may still change yet if someone realises MS_TM and "ms_time_***"
991 aren't necessarily the greatest nomenclatures - but this is what was used
992 internally to the implementation so I've used that for now.
995 *) Ensure that deprecated functions do not get compiled when
996 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
997 the self-tests were still using deprecated key-generation functions so
998 these have been updated also.
1001 *) Reorganise PKCS#7 code to separate the digest location functionality
1002 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1003 New function PKCS7_set_digest() to set the digest type for PKCS#7
1004 digestedData type. Add additional code to correctly generate the
1005 digestedData type and add support for this type in PKCS7 initialization
1009 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1010 structure of type "other".
1013 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1014 sure the loop does correctly stop and breaking ("division by zero")
1015 modulus operations are not performed. The (pre-generated) prime
1016 table crypto/bn/bn_prime.h was already correct, but it could not be
1017 re-generated on some platforms because of the "division by zero"
1018 situation in the script.
1019 [Ralf S. Engelschall]
1021 *) Update support for ECC-based TLS ciphersuites according to
1022 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1023 SHA-1 now is only used for "small" curves (where the
1024 representation of a field element takes up to 24 bytes); for
1025 larger curves, the field element resulting from ECDH is directly
1026 used as premaster secret.
1027 [Douglas Stebila (Sun Microsystems Laboratories)]
1029 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1030 curve secp160r1 to the tests.
1031 [Douglas Stebila (Sun Microsystems Laboratories)]
1033 *) Add the possibility to load symbols globally with DSO.
1034 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1036 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1037 control of the error stack.
1040 *) Add support for STORE in ENGINE.
1043 *) Add the STORE type. The intention is to provide a common interface
1044 to certificate and key stores, be they simple file-based stores, or
1045 HSM-type store, or LDAP stores, or...
1046 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1049 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1050 pass a list of arguments to any function as well as provide a way
1051 for a function to pass data back to the caller.
1054 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1055 works like BUF_strdup() but can be used to duplicate a portion of
1056 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1060 *) Add the function sk_find_ex() which works like sk_find(), but will
1061 return an index to an element even if an exact match couldn't be
1062 found. The index is guaranteed to point at the element where the
1063 searched-for key would be inserted to preserve sorting order.
1066 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1067 takes an extra flags argument for optional functionality. Currently,
1068 the following flags are defined:
1070 OBJ_BSEARCH_VALUE_ON_NOMATCH
1071 This one gets OBJ_bsearch_ex() to return a pointer to the first
1072 element where the comparing function returns a negative or zero
1075 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1076 This one gets OBJ_bsearch_ex() to return a pointer to the first
1077 element where the comparing function returns zero. This is useful
1078 if there are more than one element where the comparing function
1082 *) Make it possible to create self-signed certificates with 'openssl ca'
1083 in such a way that the self-signed certificate becomes part of the
1084 CA database and uses the same mechanisms for serial number generation
1085 as all other certificate signing. The new flag '-selfsign' enables
1086 this functionality. Adapt CA.sh and CA.pl.in.
1089 *) Add functionality to check the public key of a certificate request
1090 against a given private. This is useful to check that a certificate
1091 request can be signed by that key (self-signing).
1094 *) Make it possible to have multiple active certificates with the same
1095 subject in the CA index file. This is done only if the keyword
1096 'unique_subject' is set to 'no' in the main CA section (default
1097 if 'CA_default') of the configuration file. The value is saved
1098 with the database itself in a separate index attribute file,
1099 named like the index file with '.attr' appended to the name.
1102 *) Generate muti valued AVAs using '+' notation in config files for
1106 *) Support for nameConstraints certificate extension.
1109 *) Support for policyConstraints certificate extension.
1112 *) Support for policyMappings certificate extension.
1115 *) Make sure the default DSA_METHOD implementation only uses its
1116 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1117 and change its own handlers to be NULL so as to remove unnecessary
1118 indirection. This lets alternative implementations fallback to the
1119 default implementation more easily.
1122 *) Support for directoryName in GeneralName related extensions
1126 *) Make it possible to link applications using Makefile.shared.
1127 Make that possible even when linking against static libraries!
1130 *) Support for single pass processing for S/MIME signing. This now
1131 means that S/MIME signing can be done from a pipe, in addition
1132 cleartext signing (multipart/signed type) is effectively streaming
1133 and the signed data does not need to be all held in memory.
1135 This is done with a new flag PKCS7_STREAM. When this flag is set
1136 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1137 is done after the data is output (and digests calculated) in
1138 SMIME_write_PKCS7().
1141 *) Add full support for -rpath/-R, both in shared libraries and
1142 applications, at least on the platforms where it's known how
1146 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1147 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1148 will now compute a table of multiples of the generator that
1149 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1150 faster (notably in the case of a single point multiplication,
1151 scalar * generator).
1152 [Nils Larsch, Bodo Moeller]
1154 *) IPv6 support for certificate extensions. The various extensions
1155 which use the IP:a.b.c.d can now take IPv6 addresses using the
1156 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1160 *) Added an ENGINE that implements RSA by performing private key
1161 exponentiations with the GMP library. The conversions to and from
1162 GMP's mpz_t format aren't optimised nor are any montgomery forms
1163 cached, and on x86 it appears OpenSSL's own performance has caught up.
1164 However there are likely to be other architectures where GMP could
1165 provide a boost. This ENGINE is not built in by default, but it can be
1166 specified at Configure time and should be accompanied by the necessary
1167 linker additions, eg;
1168 ./config -DOPENSSL_USE_GMP -lgmp
1171 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1172 testing availability of engines with "-t" - the old behaviour is
1173 produced by increasing the feature's verbosity with "-tt".
1176 *) ECDSA routines: under certain error conditions uninitialized BN objects
1177 could be freed. Solution: make sure initialization is performed early
1178 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1182 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1183 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1184 software implementations. For DSA and DH, parameter generation can
1185 also be overriden by providing the appropriate method callbacks.
1188 *) Change the "progress" mechanism used in key-generation and
1189 primality testing to functions that take a new BN_GENCB pointer in
1190 place of callback/argument pairs. The new API functions have "_ex"
1191 postfixes and the older functions are reimplemented as wrappers for
1192 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1193 declarations of the old functions to help (graceful) attempts to
1194 migrate to the new functions. Also, the new key-generation API
1195 functions operate on a caller-supplied key-structure and return
1196 success/failure rather than returning a key or NULL - this is to
1197 help make "keygen" another member function of RSA_METHOD etc.
1199 Example for using the new callback interface:
1201 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1205 BN_GENCB_set(&my_cb, my_callback, my_arg);
1207 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1208 /* For the meaning of a, b in calls to my_callback(), see the
1209 * documentation of the function that calls the callback.
1210 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1211 * my_callback should return 1 if it wants BN_is_prime_ex()
1212 * to continue, or 0 to stop.
1217 *) Change the ZLIB compression method to be stateful, and make it
1218 available to TLS with the number defined in
1219 draft-ietf-tls-compression-04.txt.
1222 *) Add the ASN.1 structures and functions for CertificatePair, which
1223 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1225 CertificatePair ::= SEQUENCE {
1226 forward [0] Certificate OPTIONAL,
1227 reverse [1] Certificate OPTIONAL,
1228 -- at least one of the pair shall be present -- }
1230 Also implement the PEM functions to read and write certificate
1231 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1233 This needed to be defined, mostly for the sake of the LDAP
1234 attribute crossCertificatePair, but may prove useful elsewhere as
1238 *) Make it possible to inhibit symlinking of shared libraries in
1239 Makefile.shared, for Cygwin's sake.
1242 *) Extend the BIGNUM API by creating a function
1243 void BN_set_negative(BIGNUM *a, int neg);
1244 and a macro that behave like
1245 int BN_is_negative(const BIGNUM *a);
1247 to avoid the need to access 'a->neg' directly in applications.
1250 *) Implement fast modular reduction for pseudo-Mersenne primes
1251 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1252 EC_GROUP_new_curve_GFp() will now automatically use this
1254 [Nils Larsch <nla@trustcenter.de>]
1256 *) Add new lock type (CRYPTO_LOCK_BN).
1259 *) Change the ENGINE framework to automatically load engines
1260 dynamically from specific directories unless they could be
1261 found to already be built in or loaded. Move all the
1262 current engines except for the cryptodev one to a new
1264 The engines in engines/ are built as shared libraries if
1265 the "shared" options was given to ./Configure or ./config.
1266 Otherwise, they are inserted in libcrypto.a.
1267 /usr/local/ssl/engines is the default directory for dynamic
1268 engines, but that can be overriden at configure time through
1269 the usual use of --prefix and/or --openssldir, and at run
1270 time with the environment variable OPENSSL_ENGINES.
1271 [Geoff Thorpe and Richard Levitte]
1273 *) Add Makefile.shared, a helper makefile to build shared
1274 libraries. Addapt Makefile.org.
1277 *) Add version info to Win32 DLLs.
1278 [Peter 'Luna' Runestig" <peter@runestig.com>]
1280 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1281 can be added using this API to created arbitrary PKCS#12
1282 files while avoiding the low level API.
1284 New options to PKCS12_create(), key or cert can be NULL and
1285 will then be omitted from the output file. The encryption
1286 algorithm NIDs can be set to -1 for no encryption, the mac
1287 iteration count can be set to 0 to omit the mac.
1289 Enhance pkcs12 utility by making the -nokeys and -nocerts
1290 options work when creating a PKCS#12 file. New option -nomac
1291 to omit the mac, NONE can be set for an encryption algorithm.
1292 New code is modified to use the enhanced PKCS12_create()
1293 instead of the low level API.
1296 *) Extend ASN1 encoder to support indefinite length constructed
1297 encoding. This can output sequences tags and octet strings in
1298 this form. Modify pk7_asn1.c to support indefinite length
1299 encoding. This is experimental and needs additional code to
1300 be useful, such as an ASN1 bio and some enhanced streaming
1303 Extend template encode functionality so that tagging is passed
1304 down to the template encoder.
1307 *) Let 'openssl req' fail if an argument to '-newkey' is not
1308 recognized instead of using RSA as a default.
1311 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1312 As these are not official, they are not included in "ALL";
1313 the "ECCdraft" ciphersuite group alias can be used to select them.
1314 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1316 *) Add ECDH engine support.
1317 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1319 *) Add ECDH in new directory crypto/ecdh/.
1320 [Douglas Stebila (Sun Microsystems Laboratories)]
1322 *) Let BN_rand_range() abort with an error after 100 iterations
1323 without success (which indicates a broken PRNG).
1326 *) Change BN_mod_sqrt() so that it verifies that the input value
1327 is really the square of the return value. (Previously,
1328 BN_mod_sqrt would show GIGO behaviour.)
1331 *) Add named elliptic curves over binary fields from X9.62, SECG,
1332 and WAP/WTLS; add OIDs that were still missing.
1334 [Sheueling Chang Shantz and Douglas Stebila
1335 (Sun Microsystems Laboratories)]
1337 *) Extend the EC library for elliptic curves over binary fields
1338 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1341 EC_GF2m_simple_method
1345 EC_GROUP_new_curve_GF2m
1346 EC_GROUP_set_curve_GF2m
1347 EC_GROUP_get_curve_GF2m
1348 EC_POINT_set_affine_coordinates_GF2m
1349 EC_POINT_get_affine_coordinates_GF2m
1350 EC_POINT_set_compressed_coordinates_GF2m
1352 Point compression for binary fields is disabled by default for
1353 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1356 As binary polynomials are represented as BIGNUMs, various members
1357 of the EC_GROUP and EC_POINT data structures can be shared
1358 between the implementations for prime fields and binary fields;
1359 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1360 are essentially identical to their ..._GFp counterparts.
1361 (For simplicity, the '..._GFp' prefix has been dropped from
1362 various internal method names.)
1364 An internal 'field_div' method (similar to 'field_mul' and
1365 'field_sqr') has been added; this is used only for binary fields.
1367 [Sheueling Chang Shantz and Douglas Stebila
1368 (Sun Microsystems Laboratories)]
1370 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1371 through methods ('mul', 'precompute_mult').
1373 The generic implementations (now internally called 'ec_wNAF_mul'
1374 and 'ec_wNAF_precomputed_mult') remain the default if these
1375 methods are undefined.
1377 [Sheueling Chang Shantz and Douglas Stebila
1378 (Sun Microsystems Laboratories)]
1380 *) New function EC_GROUP_get_degree, which is defined through
1381 EC_METHOD. For curves over prime fields, this returns the bit
1382 length of the modulus.
1384 [Sheueling Chang Shantz and Douglas Stebila
1385 (Sun Microsystems Laboratories)]
1387 *) New functions EC_GROUP_dup, EC_POINT_dup.
1388 (These simply call ..._new and ..._copy).
1390 [Sheueling Chang Shantz and Douglas Stebila
1391 (Sun Microsystems Laboratories)]
1393 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1394 Polynomials are represented as BIGNUMs (where the sign bit is not
1395 used) in the following functions [macros]:
1398 BN_GF2m_sub [= BN_GF2m_add]
1399 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1400 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1401 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1403 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1404 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1405 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1406 BN_GF2m_cmp [= BN_ucmp]
1408 (Note that only the 'mod' functions are actually for fields GF(2^m).
1409 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1411 For some functions, an the irreducible polynomial defining a
1412 field can be given as an 'unsigned int[]' with strictly
1413 decreasing elements giving the indices of those bits that are set;
1414 i.e., p[] represents the polynomial
1415 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1417 p[0] > p[1] > ... > p[k] = 0.
1418 This applies to the following functions:
1423 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1424 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1426 BN_GF2m_mod_sqrt_arr
1427 BN_GF2m_mod_solve_quad_arr
1431 Conversion can be performed by the following functions:
1436 bntest.c has additional tests for binary polynomial arithmetic.
1438 Two implementations for BN_GF2m_mod_div() are available.
1439 The default algorithm simply uses BN_GF2m_mod_inv() and
1440 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1441 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1442 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1444 [Sheueling Chang Shantz and Douglas Stebila
1445 (Sun Microsystems Laboratories)]
1447 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1448 functionality is disabled at compile-time.
1449 [Douglas Stebila <douglas.stebila@sun.com>]
1451 *) Change default behaviour of 'openssl asn1parse' so that more
1452 information is visible when viewing, e.g., a certificate:
1454 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1455 mode the content of non-printable OCTET STRINGs is output in a
1456 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1457 avoid the appearance of a printable string.
1458 [Nils Larsch <nla@trustcenter.de>]
1460 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1462 EC_GROUP_set_asn1_flag()
1463 EC_GROUP_get_asn1_flag()
1464 EC_GROUP_set_point_conversion_form()
1465 EC_GROUP_get_point_conversion_form()
1466 These control ASN1 encoding details:
1467 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1468 has been set to OPENSSL_EC_NAMED_CURVE.
1469 - Points are encoded in uncompressed form by default; options for
1470 asn1_for are as for point2oct, namely
1471 POINT_CONVERSION_COMPRESSED
1472 POINT_CONVERSION_UNCOMPRESSED
1473 POINT_CONVERSION_HYBRID
1475 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1478 EC_GROUP_get0_seed()
1479 EC_GROUP_get_seed_len()
1480 This is used only for ASN1 purposes (so far).
1481 [Nils Larsch <nla@trustcenter.de>]
1483 *) Add 'field_type' member to EC_METHOD, which holds the NID
1484 of the appropriate field type OID. The new function
1485 EC_METHOD_get_field_type() returns this value.
1486 [Nils Larsch <nla@trustcenter.de>]
1491 EC_POINT_point2hex()
1492 EC_POINT_hex2point()
1493 providing useful interfaces to EC_POINT_point2oct() and
1494 EC_POINT_oct2point().
1495 [Nils Larsch <nla@trustcenter.de>]
1497 *) Change internals of the EC library so that the functions
1498 EC_GROUP_set_generator()
1499 EC_GROUP_get_generator()
1500 EC_GROUP_get_order()
1501 EC_GROUP_get_cofactor()
1502 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1503 to methods, which would lead to unnecessary code duplication when
1504 adding different types of curves.
1505 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1507 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1508 arithmetic, and such that modified wNAFs are generated
1509 (which avoid length expansion in many cases).
1512 *) Add a function EC_GROUP_check_discriminant() (defined via
1513 EC_METHOD) that verifies that the curve discriminant is non-zero.
1515 Add a function EC_GROUP_check() that makes some sanity tests
1516 on a EC_GROUP, its generator and order. This includes
1517 EC_GROUP_check_discriminant().
1518 [Nils Larsch <nla@trustcenter.de>]
1520 *) Add ECDSA in new directory crypto/ecdsa/.
1522 Add applications 'openssl ecparam' and 'openssl ecdsa'
1523 (these are based on 'openssl dsaparam' and 'openssl dsa').
1525 ECDSA support is also included in various other files across the
1526 library. Most notably,
1527 - 'openssl req' now has a '-newkey ecdsa:file' option;
1528 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1529 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1530 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1531 them suitable for ECDSA where domain parameters must be
1532 extracted before the specific public key;
1533 - ECDSA engine support has been added.
1534 [Nils Larsch <nla@trustcenter.de>]
1536 *) Include some named elliptic curves, and add OIDs from X9.62,
1537 SECG, and WAP/WTLS. Each curve can be obtained from the new
1539 EC_GROUP_new_by_curve_name(),
1540 and the list of available named curves can be obtained with
1541 EC_get_builtin_curves().
1542 Also add a 'curve_name' member to EC_GROUP objects, which can be
1544 EC_GROUP_set_curve_name()
1545 EC_GROUP_get_curve_name()
1546 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1548 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1549 was actually never needed) and in BN_mul(). The removal in BN_mul()
1550 required a small change in bn_mul_part_recursive() and the addition
1551 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1552 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1553 bn_sub_words() and bn_add_words() except they take arrays with
1557 Changes between 0.9.7m and 0.9.7n [xx XXX xxxx]
1559 *) In the SSL/TLS server implementation, be strict about session ID
1560 context matching (which matters if an application uses a single
1561 external cache for different purposes). Previously,
1562 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1563 set. This did ensure strict client verification, but meant that,
1564 with applications using a single external cache for quite
1565 different requirements, clients could circumvent ciphersuite
1566 restrictions for a given session ID context by starting a session
1567 in a different context.
1570 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1572 *) Cleanse PEM buffers before freeing them since they may contain
1574 [Benjamin Bennett <ben@psc.edu>]
1576 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1577 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1578 authentication-only ciphersuites.
1581 *) Since AES128 and AES256 share a single mask bit in the logic of
1582 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1583 kludge to work properly if AES128 is available and AES256 isn't.
1586 *) Expand security boundary to match 1.1.1 module.
1589 *) Remove redundant features: hash file source, editing of test vectors
1590 modify fipsld to use external fips_premain.c signature.
1593 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1594 run algorithm test programs.
1597 *) Make algorithm test programs more tolerant of whitespace.
1600 *) Have SSL/TLS server implementation tolerate "mismatched" record
1601 protocol version while receiving ClientHello even if the
1602 ClientHello is fragmented. (The server can't insist on the
1603 particular protocol version it has chosen before the ServerHello
1604 message has informed the client about his choice.)
1607 *) Load error codes if they are not already present instead of using a
1608 static variable. This allows them to be cleanly unloaded and reloaded.
1611 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1613 *) Introduce limits to prevent malicious keys being able to
1614 cause a denial of service. (CVE-2006-2940)
1615 [Steve Henson, Bodo Moeller]
1617 *) Fix ASN.1 parsing of certain invalid structures that can result
1618 in a denial of service. (CVE-2006-2937) [Steve Henson]
1620 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1621 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1623 *) Fix SSL client code which could crash if connecting to a
1624 malicious SSLv2 server. (CVE-2006-4343)
1625 [Tavis Ormandy and Will Drewry, Google Security Team]
1627 *) Change ciphersuite string processing so that an explicit
1628 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1629 will no longer include "AES128-SHA"), and any other similar
1630 ciphersuite (same bitmap) from *other* protocol versions (so that
1631 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1632 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1633 changes from 0.9.8b and 0.9.8d.
1636 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1638 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1639 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1641 *) Change the Unix randomness entropy gathering to use poll() when
1642 possible instead of select(), since the latter has some
1643 undesirable limitations.
1644 [Darryl Miles via Richard Levitte and Bodo Moeller]
1646 *) Disable rogue ciphersuites:
1648 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1649 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1650 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1652 The latter two were purportedly from
1653 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1656 Also deactive the remaining ciphersuites from
1657 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1658 unofficial, and the ID has long expired.
1661 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1662 dual-core machines) and other potential thread-safety issues.
1665 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1667 *) Adapt fipsld and the build system to link against the validated FIPS
1668 module in FIPS mode.
1671 *) Fixes for VC++ 2005 build under Windows.
1674 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1675 from a Windows bash shell such as MSYS. It is autodetected from the
1676 "config" script when run from a VC++ environment. Modify standard VC++
1677 build to use fipscanister.o from the GNU make build.
1680 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1682 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1683 The value now differs depending on if you build for FIPS or not.
1684 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1685 safely run with a non-FIPSed libcrypto, as it may crash because of
1686 the difference induced by this change.
1689 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1691 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1692 (part of SSL_OP_ALL). This option used to disable the
1693 countermeasure against man-in-the-middle protocol-version
1694 rollback in the SSL 2.0 server implementation, which is a bad
1695 idea. (CVE-2005-2969)
1697 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1698 for Information Security, National Institute of Advanced Industrial
1699 Science and Technology [AIST], Japan)]
1701 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1702 mainly for FIPS compliance and not fully integrated at this stage.
1705 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1706 the exponentiation using a fixed-length exponent. (Otherwise,
1707 the information leaked through timing could expose the secret key
1708 after many signatures; cf. Bleichenbacher's attack on DSA with
1712 *) Make a new fixed-window mod_exp implementation the default for
1713 RSA, DSA, and DH private-key operations so that the sequence of
1714 squares and multiplies and the memory access pattern are
1715 independent of the particular secret key. This will mitigate
1716 cache-timing and potential related attacks.
1718 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1719 and this is automatically used by BN_mod_exp_mont() if the new flag
1720 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1721 will use this BN flag for private exponents unless the flag
1722 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1723 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1725 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1727 *) Change the client implementation for SSLv23_method() and
1728 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1729 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1730 (Previously, the SSL 2.0 backwards compatible Client Hello
1731 message format would be used even with SSL_OP_NO_SSLv2.)
1734 *) Add support for smime-type MIME parameter in S/MIME messages which some
1738 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1739 a threadsafe manner. Modify rsa code to use new function and add calls
1740 to dsa and dh code (which had race conditions before).
1743 *) Include the fixed error library code in the C error file definitions
1744 instead of fixing them up at runtime. This keeps the error code
1745 structures constant.
1748 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1750 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1753 *) Fixes for newer kerberos headers. NB: the casts are needed because
1754 the 'length' field is signed on one version and unsigned on another
1755 with no (?) obvious way to tell the difference, without these VC++
1756 complains. Also the "definition" of FAR (blank) is no longer included
1757 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1758 some needed definitions.
1761 *) Undo Cygwin change.
1764 *) Added support for proxy certificates according to RFC 3820.
1765 Because they may be a security thread to unaware applications,
1766 they must be explicitely allowed in run-time. See
1767 docs/HOWTO/proxy_certificates.txt for further information.
1770 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1772 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1773 server and client random values. Previously
1774 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1775 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1777 This change has negligible security impact because:
1779 1. Server and client random values still have 24 bytes of pseudo random
1782 2. Server and client random values are sent in the clear in the initial
1785 3. The master secret is derived using the premaster secret (48 bytes in
1786 size for static RSA ciphersuites) as well as client server and random
1789 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1792 [Stephen Henson, reported by UK NISCC]
1794 *) Use Windows randomness collection on Cygwin.
1797 *) Fix hang in EGD/PRNGD query when communication socket is closed
1798 prematurely by EGD/PRNGD.
1799 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1801 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1804 *) Back-port of selected performance improvements from development
1805 branch, as well as improved support for PowerPC platforms.
1808 *) Add lots of checks for memory allocation failure, error codes to indicate
1809 failure and freeing up memory if a failure occurs.
1810 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1812 *) Add new -passin argument to dgst.
1815 *) Perform some character comparisons of different types in X509_NAME_cmp:
1816 this is needed for some certificates that reencode DNs into UTF8Strings
1817 (in violation of RFC3280) and can't or wont issue name rollover
1821 *) Make an explicit check during certificate validation to see that
1822 the CA setting in each certificate on the chain is correct. As a
1823 side effect always do the following basic checks on extensions,
1824 not just when there's an associated purpose to the check:
1826 - if there is an unhandled critical extension (unless the user
1827 has chosen to ignore this fault)
1828 - if the path length has been exceeded (if one is set at all)
1829 - that certain extensions fit the associated purpose (if one has
1833 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1835 *) Avoid a race condition when CRLs are checked in a multi threaded
1836 environment. This would happen due to the reordering of the revoked
1837 entries during signature checking and serial number lookup. Now the
1838 encoding is cached and the serial number sort performed under a lock.
1839 Add new STACK function sk_is_sorted().
1842 *) Add Delta CRL to the extension code.
1845 *) Various fixes to s3_pkt.c so alerts are sent properly.
1846 [David Holmes <d.holmes@f5.com>]
1848 *) Reduce the chances of duplicate issuer name and serial numbers (in
1849 violation of RFC3280) using the OpenSSL certificate creation utilities.
1850 This is done by creating a random 64 bit value for the initial serial
1851 number when a serial number file is created or when a self signed
1852 certificate is created using 'openssl req -x509'. The initial serial
1853 number file is created using 'openssl x509 -next_serial' in CA.pl
1854 rather than being initialized to 1.
1857 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1859 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1860 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1861 [Joe Orton, Steve Henson]
1863 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1865 [Joe Orton, Steve Henson]
1867 *) Make it possible to have multiple active certificates with the same
1868 subject in the CA index file. This is done only if the keyword
1869 'unique_subject' is set to 'no' in the main CA section (default
1870 if 'CA_default') of the configuration file. The value is saved
1871 with the database itself in a separate index attribute file,
1872 named like the index file with '.attr' appended to the name.
1875 *) X509 verify fixes. Disable broken certificate workarounds when
1876 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1877 keyUsage extension present. Don't accept CRLs with unhandled critical
1878 extensions: since verify currently doesn't process CRL extensions this
1879 rejects a CRL with *any* critical extensions. Add new verify error codes
1883 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1884 A clarification of RFC2560 will require the use of OCTET STRINGs and
1885 some implementations cannot handle the current raw format. Since OpenSSL
1886 copies and compares OCSP nonces as opaque blobs without any attempt at
1887 parsing them this should not create any compatibility issues.
1890 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1891 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1892 this HMAC (and other) operations are several times slower than OpenSSL
1896 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1897 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1899 *) Use the correct content when signing type "other".
1902 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1904 *) Fix various bugs revealed by running the NISCC test suite:
1906 Stop out of bounds reads in the ASN1 code when presented with
1907 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1909 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1911 If verify callback ignores invalid public key errors don't try to check
1912 certificate signature with the NULL public key.
1916 *) New -ignore_err option in ocsp application to stop the server
1917 exiting on the first error in a request.
1920 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1921 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1925 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1926 extra data after the compression methods not only for TLS 1.0
1927 but also for SSL 3.0 (as required by the specification).
1928 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1930 *) Change X509_certificate_type() to mark the key as exported/exportable
1931 when it's 512 *bits* long, not 512 bytes.
1934 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1935 blocks during encryption.
1938 *) Various fixes to base64 BIO and non blocking I/O. On write
1939 flushes were not handled properly if the BIO retried. On read
1940 data was not being buffered properly and had various logic bugs.
1941 This also affects blocking I/O when the data being decoded is a
1945 *) Various S/MIME bugfixes and compatibility changes:
1946 output correct application/pkcs7 MIME type if
1947 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1948 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1949 of files as .eml work). Correctly handle very long lines in MIME
1953 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1955 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1956 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1957 a protocol version number mismatch like a decryption error
1958 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1961 *) Turn on RSA blinding by default in the default implementation
1962 to avoid a timing attack. Applications that don't want it can call
1963 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1964 They would be ill-advised to do so in most cases.
1965 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1967 *) Change RSA blinding code so that it works when the PRNG is not
1968 seeded (in this case, the secret RSA exponent is abused as
1969 an unpredictable seed -- if it is not unpredictable, there
1970 is no point in blinding anyway). Make RSA blinding thread-safe
1971 by remembering the creator's thread ID in rsa->blinding and
1972 having all other threads use local one-time blinding factors
1973 (this requires more computation than sharing rsa->blinding, but
1974 avoids excessive locking; and if an RSA object is not shared
1975 between threads, blinding will still be very fast).
1978 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1979 ENGINE as defaults for all supported algorithms irrespective of
1980 the 'flags' parameter. 'flags' is now honoured, so applications
1981 should make sure they are passing it correctly.
1984 *) Target "mingw" now allows native Windows code to be generated in
1985 the Cygwin environment as well as with the MinGW compiler.
1988 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1990 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1991 via timing by performing a MAC computation even if incorrrect
1992 block cipher padding has been found. This is a countermeasure
1993 against active attacks where the attacker has to distinguish
1994 between bad padding and a MAC verification error. (CVE-2003-0078)
1996 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1997 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1998 Martin Vuagnoux (EPFL, Ilion)]
2000 *) Make the no-err option work as intended. The intention with no-err
2001 is not to have the whole error stack handling routines removed from
2002 libcrypto, it's only intended to remove all the function name and
2003 reason texts, thereby removing some of the footprint that may not
2004 be interesting if those errors aren't displayed anyway.
2006 NOTE: it's still possible for any application or module to have it's
2007 own set of error texts inserted. The routines are there, just not
2008 used by default when no-err is given.
2011 *) Add support for FreeBSD on IA64.
2012 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2014 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2015 Kerberos function mit_des_cbc_cksum(). Before this change,
2016 the value returned by DES_cbc_cksum() was like the one from
2017 mit_des_cbc_cksum(), except the bytes were swapped.
2018 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2020 *) Allow an application to disable the automatic SSL chain building.
2021 Before this a rather primitive chain build was always performed in
2022 ssl3_output_cert_chain(): an application had no way to send the
2023 correct chain if the automatic operation produced an incorrect result.
2025 Now the chain builder is disabled if either:
2027 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2029 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2031 The reasoning behind this is that an application would not want the
2032 auto chain building to take place if extra chain certificates are
2033 present and it might also want a means of sending no additional
2034 certificates (for example the chain has two certificates and the
2038 *) Add the possibility to build without the ENGINE framework.
2039 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2041 *) Under Win32 gmtime() can return NULL: check return value in
2042 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2045 *) DSA routines: under certain error conditions uninitialized BN objects
2046 could be freed. Solution: make sure initialization is performed early
2047 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2048 Nils Larsch <nla@trustcenter.de> via PR#459)
2051 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2052 checked on reconnect on the client side, therefore session resumption
2053 could still fail with a "ssl session id is different" error. This
2054 behaviour is masked when SSL_OP_ALL is used due to
2055 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2056 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2057 followup to PR #377.
2060 *) IA-32 assembler support enhancements: unified ELF targets, support
2061 for SCO/Caldera platforms, fix for Cygwin shared build.
2064 *) Add support for FreeBSD on sparc64. As a consequence, support for
2065 FreeBSD on non-x86 processors is separate from x86 processors on
2066 the config script, much like the NetBSD support.
2067 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2069 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2071 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2074 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2075 code (06) was taken as the first octet of the session ID and the last
2076 octet was ignored consequently. As a result SSLv2 client side session
2077 caching could not have worked due to the session ID mismatch between
2079 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2083 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2084 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2088 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2089 seems that in spite of existing for more than a year, many application
2090 author have done nothing to provide the necessary callbacks, which
2091 means that this particular engine will not work properly anywhere.
2092 This is a very unfortunate situation which forces us, in the name
2093 of usability, to give the hw_ncipher.c a static lock, which is part
2095 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2096 appear in 0.9.8 or later. We EXPECT application authors to have
2097 dealt properly with this when 0.9.8 is released (unless we actually
2098 make such changes in the libcrypto locking code that changes will
2099 have to be made anyway).
2102 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2103 octets have been read, EOF or an error occurs. Without this change
2104 some truncated ASN1 structures will not produce an error.
2107 *) Disable Heimdal support, since it hasn't been fully implemented.
2108 Still give the possibility to force the use of Heimdal, but with
2109 warnings and a request that patches get sent to openssl-dev.
2112 *) Add the VC-CE target, introduce the WINCE sysname, and add
2113 INSTALL.WCE and appropriate conditionals to make it build.
2114 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2116 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2117 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2118 edit numbers of the version.
2119 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2121 *) Introduce safe string copy and catenation functions
2122 (BUF_strlcpy() and BUF_strlcat()).
2123 [Ben Laurie (CHATS) and Richard Levitte]
2125 *) Avoid using fixed-size buffers for one-line DNs.
2126 [Ben Laurie (CHATS)]
2128 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2129 resizing buffers containing secrets, and use where appropriate.
2130 [Ben Laurie (CHATS)]
2132 *) Avoid using fixed size buffers for configuration file location.
2133 [Ben Laurie (CHATS)]
2135 *) Avoid filename truncation for various CA files.
2136 [Ben Laurie (CHATS)]
2138 *) Use sizeof in preference to magic numbers.
2139 [Ben Laurie (CHATS)]
2141 *) Avoid filename truncation in cert requests.
2142 [Ben Laurie (CHATS)]
2144 *) Add assertions to check for (supposedly impossible) buffer
2146 [Ben Laurie (CHATS)]
2148 *) Don't cache truncated DNS entries in the local cache (this could
2149 potentially lead to a spoofing attack).
2150 [Ben Laurie (CHATS)]
2152 *) Fix various buffers to be large enough for hex/decimal
2153 representations in a platform independent manner.
2154 [Ben Laurie (CHATS)]
2156 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2157 resizing buffers containing secrets, and use where appropriate.
2158 [Ben Laurie (CHATS)]
2160 *) Add BIO_indent() to avoid much slightly worrying code to do
2162 [Ben Laurie (CHATS)]
2164 *) Convert sprintf()/BIO_puts() to BIO_printf().
2165 [Ben Laurie (CHATS)]
2167 *) buffer_gets() could terminate with the buffer only half
2169 [Ben Laurie (CHATS)]
2171 *) Add assertions to prevent user-supplied crypto functions from
2172 overflowing internal buffers by having large block sizes, etc.
2173 [Ben Laurie (CHATS)]
2175 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2177 [Ben Laurie (CHATS)]
2179 *) Eliminate unused copy of key in RC4.
2180 [Ben Laurie (CHATS)]
2182 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2183 [Ben Laurie (CHATS)]
2185 *) Fix off-by-one error in EGD path.
2186 [Ben Laurie (CHATS)]
2188 *) If RANDFILE path is too long, ignore instead of truncating.
2189 [Ben Laurie (CHATS)]
2191 *) Eliminate unused and incorrectly sized X.509 structure
2193 [Ben Laurie (CHATS)]
2195 *) Eliminate unused and dangerous function knumber().
2196 [Ben Laurie (CHATS)]
2198 *) Eliminate unused and dangerous structure, KSSL_ERR.
2199 [Ben Laurie (CHATS)]
2201 *) Protect against overlong session ID context length in an encoded
2202 session object. Since these are local, this does not appear to be
2204 [Ben Laurie (CHATS)]
2206 *) Change from security patch (see 0.9.6e below) that did not affect
2207 the 0.9.6 release series:
2209 Remote buffer overflow in SSL3 protocol - an attacker could
2210 supply an oversized master key in Kerberos-enabled versions.
2212 [Ben Laurie (CHATS)]
2214 *) Change the SSL kerb5 codes to match RFC 2712.
2217 *) Make -nameopt work fully for req and add -reqopt switch.
2218 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2220 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2221 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2223 *) Make sure tests can be performed even if the corresponding algorithms
2224 have been removed entirely. This was also the last step to make
2225 OpenSSL compilable with DJGPP under all reasonable conditions.
2226 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2228 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2229 to allow version independent disabling of normally unselected ciphers,
2230 which may be activated as a side-effect of selecting a single cipher.
2232 (E.g., cipher list string "RSA" enables ciphersuites that are left
2233 out of "ALL" because they do not provide symmetric encryption.
2234 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2235 [Lutz Jaenicke, Bodo Moeller]
2237 *) Add appropriate support for separate platform-dependent build
2238 directories. The recommended way to make a platform-dependent
2239 build directory is the following (tested on Linux), maybe with
2242 # Place yourself outside of the OpenSSL source tree. In
2243 # this example, the environment variable OPENSSL_SOURCE
2244 # is assumed to contain the absolute OpenSSL source directory.
2245 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2246 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2247 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2248 mkdir -p `dirname $F`
2249 ln -s $OPENSSL_SOURCE/$F $F
2252 To be absolutely sure not to disturb the source tree, a "make clean"
2253 is a good thing. If it isn't successfull, don't worry about it,
2254 it probably means the source directory is very clean.
2257 *) Make sure any ENGINE control commands make local copies of string
2258 pointers passed to them whenever necessary. Otherwise it is possible
2259 the caller may have overwritten (or deallocated) the original string
2260 data when a later ENGINE operation tries to use the stored values.
2261 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2263 *) Improve diagnostics in file reading and command-line digests.
2264 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2266 *) Add AES modes CFB and OFB to the object database. Correct an
2267 error in AES-CFB decryption.
2270 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2271 allows existing EVP_CIPHER_CTX structures to be reused after
2272 calling EVP_*Final(). This behaviour is used by encryption
2273 BIOs and some applications. This has the side effect that
2274 applications must explicitly clean up cipher contexts with
2275 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2278 *) Check the values of dna and dnb in bn_mul_recursive before calling
2279 bn_mul_comba (a non zero value means the a or b arrays do not contain
2280 n2 elements) and fallback to bn_mul_normal if either is not zero.
2283 *) Fix escaping of non-ASCII characters when using the -subj option
2284 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2287 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2288 form for "surname", serialNumber has no short form.
2289 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2290 therefore remove "mail" short name for "internet 7".
2291 The OID for unique identifiers in X509 certificates is
2292 x500UniqueIdentifier, not uniqueIdentifier.
2293 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2296 *) Add an "init" command to the ENGINE config module and auto initialize
2297 ENGINEs. Without any "init" command the ENGINE will be initialized
2298 after all ctrl commands have been executed on it. If init=1 the
2299 ENGINE is initailized at that point (ctrls before that point are run
2300 on the uninitialized ENGINE and after on the initialized one). If
2301 init=0 then the ENGINE will not be iniatialized at all.
2304 *) Fix the 'app_verify_callback' interface so that the user-defined
2305 argument is actually passed to the callback: In the
2306 SSL_CTX_set_cert_verify_callback() prototype, the callback
2307 declaration has been changed from
2310 int (*cb)(X509_STORE_CTX *,void *);
2311 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2312 i=s->ctx->app_verify_callback(&ctx)
2313 has been changed into
2314 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2316 To update applications using SSL_CTX_set_cert_verify_callback(),
2317 a dummy argument can be added to their callback functions.
2318 [D. K. Smetters <smetters@parc.xerox.com>]
2320 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2321 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2323 *) Add and OPENSSL_LOAD_CONF define which will cause
2324 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2325 This allows older applications to transparently support certain
2326 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2327 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2328 load the config file and OPENSSL_add_all_algorithms_conf() which will
2329 always load it have also been added.
2332 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2333 Adjust NIDs and EVP layer.
2334 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2336 *) Config modules support in openssl utility.
2338 Most commands now load modules from the config file,
2339 though in a few (such as version) this isn't done
2340 because it couldn't be used for anything.
2342 In the case of ca and req the config file used is
2343 the same as the utility itself: that is the -config
2344 command line option can be used to specify an
2348 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2349 use "openssl_conf" if filename is NULL use default openssl config file.
2352 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2353 config section name. Add a new flag to tolerate a missing config file
2354 and move code to CONF_modules_load_file().
2357 *) Support for crypto accelerator cards from Accelerated Encryption
2358 Processing, www.aep.ie. (Use engine 'aep')
2359 The support was copied from 0.9.6c [engine] and adapted/corrected
2360 to work with the new engine framework.
2361 [AEP Inc. and Richard Levitte]
2363 *) Support for SureWare crypto accelerator cards from Baltimore
2364 Technologies. (Use engine 'sureware')
2365 The support was copied from 0.9.6c [engine] and adapted
2366 to work with the new engine framework.
2369 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2370 make the newer ENGINE framework commands for the CHIL engine work.
2371 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2373 *) Make it possible to produce shared libraries on ReliantUNIX.
2374 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2376 *) Add the configuration target debug-linux-ppro.
2377 Make 'openssl rsa' use the general key loading routines
2378 implemented in apps.c, and make those routines able to
2379 handle the key format FORMAT_NETSCAPE and the variant
2381 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2383 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2384 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2386 *) Add -keyform to rsautl, and document -engine.
2387 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2389 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2390 BIO_R_NO_SUCH_FILE error code rather than the generic
2391 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2394 *) Add new functions
2396 ERR_peek_last_error_line
2397 ERR_peek_last_error_line_data.
2398 These are similar to
2401 ERR_peek_error_line_data,
2402 but report on the latest error recorded rather than the first one
2403 still in the error queue.
2404 [Ben Laurie, Bodo Moeller]
2406 *) default_algorithms option in ENGINE config module. This allows things
2408 default_algorithms = ALL
2409 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2412 *) Prelminary ENGINE config module.
2415 *) New experimental application configuration code.
2418 *) Change the AES code to follow the same name structure as all other
2419 symmetric ciphers, and behave the same way. Move everything to
2420 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2421 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2423 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2424 [Ben Laurie and Theo de Raadt]
2426 *) Add option to output public keys in req command.
2427 [Massimiliano Pala madwolf@openca.org]
2429 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2430 (up to about 10% better than before for P-192 and P-224).
2433 *) New functions/macros
2435 SSL_CTX_set_msg_callback(ctx, cb)
2436 SSL_CTX_set_msg_callback_arg(ctx, arg)
2437 SSL_set_msg_callback(ssl, cb)
2438 SSL_set_msg_callback_arg(ssl, arg)
2440 to request calling a callback function
2442 void cb(int write_p, int version, int content_type,
2443 const void *buf, size_t len, SSL *ssl, void *arg)
2445 whenever a protocol message has been completely received
2446 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2447 protocol version according to which the SSL library interprets
2448 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2449 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2450 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2451 specification (change_cipher_spec(20), alert(21), handshake(22)).
2452 'buf' and 'len' point to the actual message, 'ssl' to the
2453 SSL object, and 'arg' is the application-defined value set by
2454 SSL[_CTX]_set_msg_callback_arg().
2456 'openssl s_client' and 'openssl s_server' have new '-msg' options
2457 to enable a callback that displays all protocol messages.
2460 *) Change the shared library support so shared libraries are built as
2461 soon as the corresponding static library is finished, and thereby get
2462 openssl and the test programs linked against the shared library.
2463 This still only happens when the keyword "shard" has been given to
2464 the configuration scripts.
2466 NOTE: shared library support is still an experimental thing, and
2467 backward binary compatibility is still not guaranteed.
2468 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2470 *) Add support for Subject Information Access extension.
2471 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2473 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2474 additional bytes when new memory had to be allocated, not just
2475 when reusing an existing buffer.
2478 *) New command line and configuration option 'utf8' for the req command.
2479 This allows field values to be specified as UTF8 strings.
2482 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2483 runs for the former and machine-readable output for the latter.
2486 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2487 of the e-mail address in the DN (i.e., it will go into a certificate
2488 extension only). The new configuration file option 'email_in_dn = no'
2489 has the same effect.
2490 [Massimiliano Pala madwolf@openca.org]
2492 *) Change all functions with names starting with des_ to be starting
2493 with DES_ instead. Add wrappers that are compatible with libdes,
2494 but are named _ossl_old_des_*. Finally, add macros that map the
2495 des_* symbols to the corresponding _ossl_old_des_* if libdes
2496 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2497 desired, the des_* symbols will be mapped to DES_*, with one
2500 Since we provide two compatibility mappings, the user needs to
2501 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2502 compatibility is desired. The default (i.e., when that macro
2503 isn't defined) is OpenSSL 0.9.6c compatibility.
2505 There are also macros that enable and disable the support of old
2506 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2507 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2508 are defined, the default will apply: to support the old des routines.
2510 In either case, one must include openssl/des.h to get the correct
2511 definitions. Do not try to just include openssl/des_old.h, that
2514 NOTE: This is a major break of an old API into a new one. Software
2515 authors are encouraged to switch to the DES_ style functions. Some
2516 time in the future, des_old.h and the libdes compatibility functions
2517 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2518 default), and then completely removed.
2521 *) Test for certificates which contain unsupported critical extensions.
2522 If such a certificate is found during a verify operation it is
2523 rejected by default: this behaviour can be overridden by either
2524 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2525 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2526 X509_supported_extension() has also been added which returns 1 if a
2527 particular extension is supported.
2530 *) Modify the behaviour of EVP cipher functions in similar way to digests
2531 to retain compatibility with existing code.
2534 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2535 compatibility with existing code. In particular the 'ctx' parameter does
2536 not have to be to be initialized before the call to EVP_DigestInit() and
2537 it is tidied up after a call to EVP_DigestFinal(). New function
2538 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2539 EVP_MD_CTX_copy() changed to not require the destination to be
2540 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2541 requires the destination to be valid.
2543 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2544 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2547 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2548 so that complete 'Handshake' protocol structures are kept in memory
2549 instead of overwriting 'msg_type' and 'length' with 'body' data.
2552 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2553 [Massimo Santin via Richard Levitte]
2555 *) Major restructuring to the underlying ENGINE code. This includes
2556 reduction of linker bloat, separation of pure "ENGINE" manipulation
2557 (initialisation, etc) from functionality dealing with implementations
2558 of specific crypto iterfaces. This change also introduces integrated
2559 support for symmetric ciphers and digest implementations - so ENGINEs
2560 can now accelerate these by providing EVP_CIPHER and EVP_MD
2561 implementations of their own. This is detailed in crypto/engine/README
2562 as it couldn't be adequately described here. However, there are a few
2563 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2564 were changed in the original introduction of ENGINE code have now
2565 reverted back - the hooking from this code to ENGINE is now a good
2566 deal more passive and at run-time, operations deal directly with
2567 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2568 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2569 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2570 they were not being used by the framework as there is no concept of a
2571 BIGNUM_METHOD and they could not be generalised to the new
2572 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2573 ENGINE_cpy() has been removed as it cannot be consistently defined in
2577 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2580 *) Change mkdef.pl to sort symbols that get the same entry number,
2581 and make sure the automatically generated functions ERR_load_*
2582 become part of libeay.num as well.
2585 *) New function SSL_renegotiate_pending(). This returns true once
2586 renegotiation has been requested (either SSL_renegotiate() call
2587 or HelloRequest/ClientHello receveived from the peer) and becomes
2588 false once a handshake has been completed.
2589 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2590 sends a HelloRequest, but does not ensure that a handshake takes
2591 place. SSL_renegotiate_pending() is useful for checking if the
2592 client has followed the request.)
2595 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2596 By default, clients may request session resumption even during
2597 renegotiation (if session ID contexts permit); with this option,
2598 session resumption is possible only in the first handshake.
2600 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2601 more bits available for options that should not be part of
2602 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2605 *) Add some demos for certificate and certificate request creation.
2608 *) Make maximum certificate chain size accepted from the peer application
2609 settable (SSL*_get/set_max_cert_list()), as proposed by
2610 "Douglas E. Engert" <deengert@anl.gov>.
2613 *) Add support for shared libraries for Unixware-7
2614 (Boyd Lynn Gerber <gerberb@zenez.com>).
2617 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2618 be done prior to destruction. Use this to unload error strings from
2619 ENGINEs that load their own error strings. NB: This adds two new API
2620 functions to "get" and "set" this destroy handler in an ENGINE.
2623 *) Alter all existing ENGINE implementations (except "openssl" and
2624 "openbsd") to dynamically instantiate their own error strings. This
2625 makes them more flexible to be built both as statically-linked ENGINEs
2626 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2627 Also, add stub code to each that makes building them as self-contained
2628 shared-libraries easier (see README.ENGINE).
2631 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2632 implementations into applications that are completely implemented in
2633 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2634 commands that can be used to configure what shared-library to load and
2635 to control aspects of the way it is handled. Also, made an update to
2636 the README.ENGINE file that brings its information up-to-date and
2637 provides some information and instructions on the "dynamic" ENGINE
2638 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2641 *) Make it possible to unload ranges of ERR strings with a new
2642 "ERR_unload_strings" function.
2645 *) Add a copy() function to EVP_MD.
2648 *) Make EVP_MD routines take a context pointer instead of just the
2649 md_data void pointer.
2652 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2653 that the digest can only process a single chunk of data
2654 (typically because it is provided by a piece of
2655 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2656 is only going to provide a single chunk of data, and hence the
2657 framework needn't accumulate the data for oneshot drivers.
2660 *) As with "ERR", make it possible to replace the underlying "ex_data"
2661 functions. This change also alters the storage and management of global
2662 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2663 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2664 index counters. The API functions that use this state have been changed
2665 to take a "class_index" rather than pointers to the class's local STACK
2666 and counter, and there is now an API function to dynamically create new
2667 classes. This centralisation allows us to (a) plug a lot of the
2668 thread-safety problems that existed, and (b) makes it possible to clean
2669 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2670 such data would previously have always leaked in application code and
2671 workarounds were in place to make the memory debugging turn a blind eye
2672 to it. Application code that doesn't use this new function will still
2673 leak as before, but their memory debugging output will announce it now
2674 rather than letting it slide.
2676 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2677 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2678 has a return value to indicate success or failure.
2681 *) Make it possible to replace the underlying "ERR" functions such that the
2682 global state (2 LHASH tables and 2 locks) is only used by the "default"
2683 implementation. This change also adds two functions to "get" and "set"
2684 the implementation prior to it being automatically set the first time
2685 any other ERR function takes place. Ie. an application can call "get",
2686 pass the return value to a module it has just loaded, and that module
2687 can call its own "set" function using that value. This means the
2688 module's "ERR" operations will use (and modify) the error state in the
2689 application and not in its own statically linked copy of OpenSSL code.
2692 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2693 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2694 the operation, and provides a more encapsulated way for external code
2695 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2696 to use these functions rather than manually incrementing the counts.
2698 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2701 *) Add EVP test program.
2704 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2707 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2708 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2709 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2710 These allow a CRL to be built without having to access X509_CRL fields
2711 directly. Modify 'ca' application to use new functions.
2714 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2715 bug workarounds. Rollback attack detection is a security feature.
2716 The problem will only arise on OpenSSL servers when TLSv1 is not
2717 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2718 Software authors not wanting to support TLSv1 will have special reasons
2719 for their choice and can explicitly enable this option.
2720 [Bodo Moeller, Lutz Jaenicke]
2722 *) Rationalise EVP so it can be extended: don't include a union of
2723 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2724 (similar to those existing for EVP_CIPHER_CTX).
2729 EVP_MD_CTX_init(&md); /* new function call */
2730 EVP_DigestInit(&md, EVP_sha1());
2731 EVP_DigestUpdate(&md, in, len);
2732 EVP_DigestFinal(&md, out, NULL);
2733 EVP_MD_CTX_cleanup(&md); /* new function call */
2737 *) Make DES key schedule conform to the usual scheme, as well as
2738 correcting its structure. This means that calls to DES functions
2739 now have to pass a pointer to a des_key_schedule instead of a
2740 plain des_key_schedule (which was actually always a pointer
2743 des_key_schedule ks;
2745 des_set_key_checked(..., &ks);
2746 des_ncbc_encrypt(..., &ks, ...);
2748 (Note that a later change renames 'des_...' into 'DES_...'.)
2751 *) Initial reduction of linker bloat: the use of some functions, such as
2752 PEM causes large amounts of unused functions to be linked in due to
2753 poor organisation. For example pem_all.c contains every PEM function
2754 which has a knock on effect of linking in large amounts of (unused)
2755 ASN1 code. Grouping together similar functions and splitting unrelated
2756 functions prevents this.
2759 *) Cleanup of EVP macros.
2762 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2763 correct _ecb suffix.
2766 *) Add initial OCSP responder support to ocsp application. The
2767 revocation information is handled using the text based index
2768 use by the ca application. The responder can either handle
2769 requests generated internally, supplied in files (for example
2770 via a CGI script) or using an internal minimal server.
2773 *) Add configuration choices to get zlib compression for TLS.
2776 *) Changes to Kerberos SSL for RFC 2712 compliance:
2777 1. Implemented real KerberosWrapper, instead of just using
2778 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2779 2. Implemented optional authenticator field of KerberosWrapper.
2781 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2782 and authenticator structs; see crypto/krb5/.
2784 Generalized Kerberos calls to support multiple Kerberos libraries.
2785 [Vern Staats <staatsvr@asc.hpc.mil>,
2786 Jeffrey Altman <jaltman@columbia.edu>
2787 via Richard Levitte]
2789 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2790 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2791 values for each of the key sizes rather than having just
2792 parameters (and 'speed' generating keys each time).
2795 *) Speed up EVP routines.
2798 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2799 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2800 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2801 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2803 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2804 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2805 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2808 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2810 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2813 *) Added the OS2-EMX target.
2814 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2816 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2817 to support NCONF routines in extension code. New function CONF_set_nconf()
2818 to allow functions which take an NCONF to also handle the old LHASH
2819 structure: this means that the old CONF compatible routines can be
2820 retained (in particular wrt extensions) without having to duplicate the
2821 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2824 *) Enhance the general user interface with mechanisms for inner control
2825 and with possibilities to have yes/no kind of prompts.
2828 *) Change all calls to low level digest routines in the library and
2829 applications to use EVP. Add missing calls to HMAC_cleanup() and
2830 don't assume HMAC_CTX can be copied using memcpy().
2831 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2833 *) Add the possibility to control engines through control names but with
2834 arbitrary arguments instead of just a string.
2835 Change the key loaders to take a UI_METHOD instead of a callback
2836 function pointer. NOTE: this breaks binary compatibility with earlier
2837 versions of OpenSSL [engine].
2838 Adapt the nCipher code for these new conditions and add a card insertion
2842 *) Enhance the general user interface with mechanisms to better support
2843 dialog box interfaces, application-defined prompts, the possibility
2844 to use defaults (for example default passwords from somewhere else)
2845 and interrupts/cancellations.
2848 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2849 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2852 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2853 tidy up some unnecessarily weird code in 'sk_new()').
2854 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2856 *) Change the key loading routines for ENGINEs to use the same kind
2857 callback (pem_password_cb) as all other routines that need this
2861 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2862 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2863 than this minimum value is recommended.
2866 *) New random seeder for OpenVMS, using the system process statistics
2867 that are easily reachable.
2870 *) Windows apparently can't transparently handle global
2871 variables defined in DLLs. Initialisations such as:
2873 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2875 wont compile. This is used by the any applications that need to
2876 declare their own ASN1 modules. This was fixed by adding the option
2877 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2878 needed for static libraries under Win32.
2881 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2882 setting of purpose and trust fields. New X509_STORE trust and
2883 purpose functions and tidy up setting in other SSL functions.
2886 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2887 structure. These are inherited by X509_STORE_CTX when it is
2888 initialised. This allows various defaults to be set in the
2889 X509_STORE structure (such as flags for CRL checking and custom
2890 purpose or trust settings) for functions which only use X509_STORE_CTX
2891 internally such as S/MIME.
2893 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2894 trust settings if they are not set in X509_STORE. This allows X509_STORE
2895 purposes and trust (in S/MIME for example) to override any set by default.
2897 Add command line options for CRL checking to smime, s_client and s_server
2901 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2902 are set then the CRL is looked up in the X509_STORE structure and
2903 its validity and signature checked, then if the certificate is found
2904 in the CRL the verify fails with a revoked error.
2906 Various new CRL related callbacks added to X509_STORE_CTX structure.
2908 Command line options added to 'verify' application to support this.
2910 This needs some additional work, such as being able to handle multiple
2911 CRLs with different times, extension based lookup (rather than just
2912 by subject name) and ultimately more complete V2 CRL extension
2916 *) Add a general user interface API (crypto/ui/). This is designed
2917 to replace things like des_read_password and friends (backward
2918 compatibility functions using this new API are provided).
2919 The purpose is to remove prompting functions from the DES code
2920 section as well as provide for prompting through dialog boxes in
2921 a window system and the like.
2924 *) Add "ex_data" support to ENGINE so implementations can add state at a
2925 per-structure level rather than having to store it globally.
2928 *) Make it possible for ENGINE structures to be copied when retrieved by
2929 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2930 This causes the "original" ENGINE structure to act like a template,
2931 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2932 operational state can be localised to each ENGINE structure, despite the
2933 fact they all share the same "methods". New ENGINE structures returned in
2934 this case have no functional references and the return value is the single
2935 structural reference. This matches the single structural reference returned
2936 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2940 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2941 needs to match any other type at all we need to manually clear the
2945 *) Changes to the "openssl engine" utility to include;
2946 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2947 about an ENGINE's available control commands.
2948 - executing control commands from command line arguments using the
2949 '-pre' and '-post' switches. '-post' is only used if '-t' is
2950 specified and the ENGINE is successfully initialised. The syntax for
2951 the individual commands are colon-separated, for example;
2952 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2955 *) New dynamic control command support for ENGINEs. ENGINEs can now
2956 declare their own commands (numbers), names (strings), descriptions,
2957 and input types for run-time discovery by calling applications. A
2958 subset of these commands are implicitly classed as "executable"
2959 depending on their input type, and only these can be invoked through
2960 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2961 can be based on user input, config files, etc). The distinction is
2962 that "executable" commands cannot return anything other than a boolean
2963 result and can only support numeric or string input, whereas some
2964 discoverable commands may only be for direct use through
2965 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2966 pointers, or other custom uses. The "executable" commands are to
2967 support parameterisations of ENGINE behaviour that can be
2968 unambiguously defined by ENGINEs and used consistently across any
2969 OpenSSL-based application. Commands have been added to all the
2970 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2971 control over shared-library paths without source code alterations.
2974 *) Changed all ENGINE implementations to dynamically allocate their
2975 ENGINEs rather than declaring them statically. Apart from this being
2976 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2977 this also allows the implementations to compile without using the
2978 internal engine_int.h header.
2981 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2982 'const' value. Any code that should be able to modify a RAND_METHOD
2983 should already have non-const pointers to it (ie. they should only
2984 modify their own ones).
2987 *) Made a variety of little tweaks to the ENGINE code.
2988 - "atalla" and "ubsec" string definitions were moved from header files
2989 to C code. "nuron" string definitions were placed in variables
2990 rather than hard-coded - allowing parameterisation of these values
2991 later on via ctrl() commands.
2992 - Removed unused "#if 0"'d code.
2993 - Fixed engine list iteration code so it uses ENGINE_free() to release
2994 structural references.
2995 - Constified the RAND_METHOD element of ENGINE structures.
2996 - Constified various get/set functions as appropriate and added
2997 missing functions (including a catch-all ENGINE_cpy that duplicates
2998 all ENGINE values onto a new ENGINE except reference counts/state).
2999 - Removed NULL parameter checks in get/set functions. Setting a method
3000 or function to NULL is a way of cancelling out a previously set
3001 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3002 and doesn't justify the extra error symbols and code.
3003 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3004 flags from engine_int.h to engine.h.
3005 - Changed prototypes for ENGINE handler functions (init(), finish(),
3006 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3009 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3010 to the algorithm using long division. The binary algorithm can be
3011 used only if the modulus is odd. On 32-bit systems, it is faster
3012 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3013 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3014 up to 450 bits. In 64-bit environments, the binary algorithm
3015 appears to be advantageous for much longer moduli; here we use it
3016 for moduli up to 2048 bits.
3019 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3020 could not support the combine flag in choice fields.
3023 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3024 extensions from a certificate request to the certificate.
3027 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3028 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3029 file: this allows the display of the certificate about to be
3030 signed to be customised, to allow certain fields to be included
3031 or excluded and extension details. The old system didn't display
3032 multicharacter strings properly, omitted fields not in the policy
3033 and couldn't display additional details such as extensions.
3036 *) Function EC_POINTs_mul for multiple scalar multiplication
3037 of an arbitrary number of elliptic curve points
3038 \sum scalars[i]*points[i],
3039 optionally including the generator defined for the EC_GROUP:
3040 scalar*generator + \sum scalars[i]*points[i].
3042 EC_POINT_mul is a simple wrapper function for the typical case
3043 that the point list has just one item (besides the optional
3047 *) First EC_METHODs for curves over GF(p):
3049 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3050 operations and provides various method functions that can also
3051 operate with faster implementations of modular arithmetic.
3053 EC_GFp_mont_method() reuses most functions that are part of
3054 EC_GFp_simple_method, but uses Montgomery arithmetic.
3056 [Bodo Moeller; point addition and point doubling
3057 implementation directly derived from source code provided by
3058 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3060 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3061 crypto/ec/ec_lib.c):
3063 Curves are EC_GROUP objects (with an optional group generator)
3064 based on EC_METHODs that are built into the library.
3066 Points are EC_POINT objects based on EC_GROUP objects.
3068 Most of the framework would be able to handle curves over arbitrary
3069 finite fields, but as there are no obvious types for fields other
3070 than GF(p), some functions are limited to that for now.
3073 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3074 that the file contains a complete HTTP response.
3077 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3078 change the def and num file printf format specifier from "%-40sXXX"
3079 to "%-39s XXX". The latter will always guarantee a space after the
3080 field while the former will cause them to run together if the field
3081 is 40 of more characters long.
3084 *) Constify the cipher and digest 'method' functions and structures
3085 and modify related functions to take constant EVP_MD and EVP_CIPHER
3089 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3090 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3093 *) Modify EVP_Digest*() routines so they now return values. Although the
3094 internal software routines can never fail additional hardware versions
3098 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3100 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3101 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3104 ERR_R_NESTED_ASN1_ERROR
3106 ERR_R_MISSING_ASN1_EOS
3107 were 4 .. 9, conflicting with
3108 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3110 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3111 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3113 Add new error code 'ERR_R_INTERNAL_ERROR'.
3116 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3120 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3121 sets the subject name for a new request or supersedes the
3122 subject name in a given request. Formats that can be parsed are
3123 'CN=Some Name, OU=myOU, C=IT'
3125 'CN=Some Name/OU=myOU/C=IT'.
3127 Add options '-batch' and '-verbose' to 'openssl req'.
3128 [Massimiliano Pala <madwolf@hackmasters.net>]
3130 *) Introduce the possibility to access global variables through
3131 functions on platform were that's the best way to handle exporting
3132 global variables in shared libraries. To enable this functionality,
3133 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3134 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3135 is normally done by Configure or something similar).
3137 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3138 in the source file (foo.c) like this:
3140 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3141 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3143 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3144 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3146 OPENSSL_DECLARE_GLOBAL(int,foo);
3147 #define foo OPENSSL_GLOBAL_REF(foo)
3148 OPENSSL_DECLARE_GLOBAL(double,bar);
3149 #define bar OPENSSL_GLOBAL_REF(bar)
3151 The #defines are very important, and therefore so is including the
3152 header file everywhere where the defined globals are used.
3154 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3155 of ASN.1 items, but that structure is a bit different.
3157 The largest change is in util/mkdef.pl which has been enhanced with
3158 better and easier to understand logic to choose which symbols should
3159 go into the Windows .def files as well as a number of fixes and code
3160 cleanup (among others, algorithm keywords are now sorted
3161 lexicographically to avoid constant rewrites).
3164 *) In BN_div() keep a copy of the sign of 'num' before writing the
3165 result to 'rm' because if rm==num the value will be overwritten
3166 and produce the wrong result if 'num' is negative: this caused
3167 problems with BN_mod() and BN_nnmod().
3170 *) Function OCSP_request_verify(). This checks the signature on an
3171 OCSP request and verifies the signer certificate. The signer
3172 certificate is just checked for a generic purpose and OCSP request
3176 *) Add OCSP_check_validity() function to check the validity of OCSP
3177 responses. OCSP responses are prepared in real time and may only
3178 be a few seconds old. Simply checking that the current time lies
3179 between thisUpdate and nextUpdate max reject otherwise valid responses
3180 caused by either OCSP responder or client clock inaccuracy. Instead
3181 we allow thisUpdate and nextUpdate to fall within a certain period of
3182 the current time. The age of the response can also optionally be
3183 checked. Two new options -validity_period and -status_age added to
3187 *) If signature or public key algorithm is unrecognized print out its
3188 OID rather that just UNKNOWN.
3191 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3192 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3193 ID to be generated from the issuer certificate alone which can then be
3194 passed to OCSP_id_issuer_cmp().
3197 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3198 ASN1 modules to export functions returning ASN1_ITEM pointers
3199 instead of the ASN1_ITEM structures themselves. This adds several
3200 new macros which allow the underlying ASN1 function/structure to
3201 be accessed transparently. As a result code should not use ASN1_ITEM
3202 references directly (such as &X509_it) but instead use the relevant
3203 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3204 use of the new ASN1 code on platforms where exporting structures
3205 is problematical (for example in shared libraries) but exporting
3206 functions returning pointers to structures is not.
3209 *) Add support for overriding the generation of SSL/TLS session IDs.
3210 These callbacks can be registered either in an SSL_CTX or per SSL.
3211 The purpose of this is to allow applications to control, if they wish,
3212 the arbitrary values chosen for use as session IDs, particularly as it
3213 can be useful for session caching in multiple-server environments. A
3214 command-line switch for testing this (and any client code that wishes
3215 to use such a feature) has been added to "s_server".
3216 [Geoff Thorpe, Lutz Jaenicke]
3218 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3219 of the form '#if defined(...) || defined(...) || ...' and
3220 '#if !defined(...) && !defined(...) && ...'. This also avoids
3221 the growing number of special cases it was previously handling.
3224 *) Make all configuration macros available for application by making
3225 sure they are available in opensslconf.h, by giving them names starting
3226 with "OPENSSL_" to avoid conflicts with other packages and by making
3227 sure e_os2.h will cover all platform-specific cases together with
3229 Additionally, it is now possible to define configuration/platform-
3230 specific names (called "system identities"). In the C code, these
3231 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3232 macro with the name beginning with "OPENSSL_SYS_", which is determined
3233 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3237 *) New option -set_serial to 'req' and 'x509' this allows the serial
3238 number to use to be specified on the command line. Previously self
3239 signed certificates were hard coded with serial number 0 and the
3240 CA options of 'x509' had to use a serial number in a file which was
3244 *) New options to 'ca' utility to support V2 CRL entry extensions.
3245 Currently CRL reason, invalidity date and hold instruction are
3246 supported. Add new CRL extensions to V3 code and some new objects.
3249 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3250 disable standard block padding (aka PKCS#5 padding) in the EVP
3251 API, which was previously mandatory. This means that the data is
3252 not padded in any way and so the total length much be a multiple
3253 of the block size, otherwise an error occurs.
3256 *) Initial (incomplete) OCSP SSL support.
3259 *) New function OCSP_parse_url(). This splits up a URL into its host,
3260 port and path components: primarily to parse OCSP URLs. New -url
3261 option to ocsp utility.
3264 *) New nonce behavior. The return value of OCSP_check_nonce() now
3265 reflects the various checks performed. Applications can decide
3266 whether to tolerate certain situations such as an absent nonce
3267 in a response when one was present in a request: the ocsp application
3268 just prints out a warning. New function OCSP_add1_basic_nonce()
3269 this is to allow responders to include a nonce in a response even if
3270 the request is nonce-less.
3273 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3274 skipped when using openssl x509 multiple times on a single input file,
3275 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3278 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3279 set string type: to handle setting ASN1_TIME structures. Fix ca
3280 utility to correctly initialize revocation date of CRLs.
3283 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3284 the clients preferred ciphersuites and rather use its own preferences.
3285 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3286 Internet Explorer by ensuring unchanged hash method during stepup.
3287 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3290 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3291 to aes and add a new 'exist' option to print out symbols that don't
3295 *) Additional options to ocsp utility to allow flags to be set and
3296 additional certificates supplied.
3299 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3300 OCSP client a number of certificate to only verify the response
3304 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3305 handle the new API. Currently only ECB, CBC modes supported. Add new
3308 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3309 Encryption Standard (AES) Ciphersuites for Transport Layer
3310 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3311 not enabled by default and were not part of the "ALL" ciphersuite
3312 alias because they were not yet official; they could be
3313 explicitly requested by specifying the "AESdraft" ciphersuite
3314 group alias. In the final release of OpenSSL 0.9.7, the group
3315 alias is called "AES" and is part of "ALL".)
3316 [Ben Laurie, Steve Henson, Bodo Moeller]
3318 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3319 request to response.
3322 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3323 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3324 extract information from a certificate request. OCSP_response_create()
3325 creates a response and optionally adds a basic response structure.
3326 OCSP_basic_add1_status() adds a complete single response to a basic
3327 response and returns the OCSP_SINGLERESP structure just added (to allow
3328 extensions to be included for example). OCSP_basic_add1_cert() adds a
3329 certificate to a basic response and OCSP_basic_sign() signs a basic
3330 response with various flags. New helper functions ASN1_TIME_check()
3331 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3332 (converts ASN1_TIME to GeneralizedTime).
3335 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3336 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3337 structure from a certificate. X509_pubkey_digest() digests the public_key
3338 contents: this is used in various key identifiers.
3341 *) Make sk_sort() tolerate a NULL argument.
3342 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3344 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3345 passed by the function are trusted implicitly. If any of them signed the
3346 response then it is assumed to be valid and is not verified.
3349 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3350 to data. This was previously part of the PKCS7 ASN1 code. This
3351 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3352 [Steve Henson, reported by Kenneth R. Robinette
3353 <support@securenetterm.com>]
3355 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3356 routines: without these tracing memory leaks is very painful.
3357 Fix leaks in PKCS12 and PKCS7 routines.
3360 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3361 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3362 effectively meant GeneralizedTime would never be used. Now it
3363 is initialised to -1 but X509_time_adj() now has to check the value
3364 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3365 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3366 [Steve Henson, reported by Kenneth R. Robinette
3367 <support@securenetterm.com>]
3369 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3370 result in a zero length in the ASN1_INTEGER structure which was
3371 not consistent with the structure when d2i_ASN1_INTEGER() was used
3372 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3373 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3374 where it did not print out a minus for negative ASN1_INTEGER.
3377 *) Add summary printout to ocsp utility. The various functions which
3378 convert status values to strings have been renamed to:
3379 OCSP_response_status_str(), OCSP_cert_status_str() and
3380 OCSP_crl_reason_str() and are no longer static. New options
3381 to verify nonce values and to disable verification. OCSP response
3382 printout format cleaned up.
3385 *) Add additional OCSP certificate checks. These are those specified
3386 in RFC2560. This consists of two separate checks: the CA of the
3387 certificate being checked must either be the OCSP signer certificate
3388 or the issuer of the OCSP signer certificate. In the latter case the
3389 OCSP signer certificate must contain the OCSP signing extended key
3390 usage. This check is performed by attempting to match the OCSP
3391 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3392 in the OCSP_CERTID structures of the response.
3395 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3396 and related routines. This uses the standard OpenSSL certificate
3397 verify routines to perform initial checks (just CA validity) and
3398 to obtain the certificate chain. Then additional checks will be
3399 performed on the chain. Currently the root CA is checked to see
3400 if it is explicitly trusted for OCSP signing. This is used to set
3401 a root CA as a global signing root: that is any certificate that
3402 chains to that CA is an acceptable OCSP signing certificate.
3405 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3406 extensions from a separate configuration file.
3407 As when reading extensions from the main configuration file,
3408 the '-extensions ...' option may be used for specifying the
3410 [Massimiliano Pala <madwolf@comune.modena.it>]
3412 *) New OCSP utility. Allows OCSP requests to be generated or
3413 read. The request can be sent to a responder and the output
3414 parsed, outputed or printed in text form. Not complete yet:
3415 still needs to check the OCSP response validity.
3418 *) New subcommands for 'openssl ca':
3419 'openssl ca -status <serial>' prints the status of the cert with
3420 the given serial number (according to the index file).
3421 'openssl ca -updatedb' updates the expiry status of certificates
3423 [Massimiliano Pala <madwolf@comune.modena.it>]
3425 *) New '-newreq-nodes' command option to CA.pl. This is like
3426 '-newreq', but calls 'openssl req' with the '-nodes' option
3427 so that the resulting key is not encrypted.
3428 [Damien Miller <djm@mindrot.org>]
3430 *) New configuration for the GNU Hurd.
3431 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3433 *) Initial code to implement OCSP basic response verify. This
3434 is currently incomplete. Currently just finds the signer's
3435 certificate and verifies the signature on the response.
3438 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3439 value of OPENSSLDIR. This is available via the new '-d' option
3440 to 'openssl version', and is also included in 'openssl version -a'.
3443 *) Allowing defining memory allocation callbacks that will be given
3444 file name and line number information in additional arguments
3445 (a const char* and an int). The basic functionality remains, as
3446 well as the original possibility to just replace malloc(),
3447 realloc() and free() by functions that do not know about these
3448 additional arguments. To register and find out the current
3449 settings for extended allocation functions, the following
3450 functions are provided:
3452 CRYPTO_set_mem_ex_functions
3453 CRYPTO_set_locked_mem_ex_functions
3454 CRYPTO_get_mem_ex_functions
3455 CRYPTO_get_locked_mem_ex_functions
3457 These work the same way as CRYPTO_set_mem_functions and friends.
3458 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3459 extended allocation function is enabled.
3460 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3461 a conventional allocation function is enabled.
3462 [Richard Levitte, Bodo Moeller]
3464 *) Finish off removing the remaining LHASH function pointer casts.
3465 There should no longer be any prototype-casting required when using
3466 the LHASH abstraction, and any casts that remain are "bugs". See
3467 the callback types and macros at the head of lhash.h for details
3468 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3471 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3472 If /dev/[u]random devices are not available or do not return enough
3473 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3475 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3476 /etc/entropy will be queried once each in this sequence, quering stops
3477 when enough entropy was collected without querying more sockets.
3480 *) Change the Unix RAND_poll() variant to be able to poll several
3481 random devices, as specified by DEVRANDOM, until a sufficient amount
3482 of data has been collected. We spend at most 10 ms on each file
3483 (select timeout) and read in non-blocking mode. DEVRANDOM now
3484 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3485 (previously it was just the string "/dev/urandom"), so on typical
3486 platforms the 10 ms delay will never occur.
3487 Also separate out the Unix variant to its own file, rand_unix.c.
3488 For VMS, there's a currently-empty rand_vms.c.
3491 *) Move OCSP client related routines to ocsp_cl.c. These
3492 provide utility functions which an application needing
3493 to issue a request to an OCSP responder and analyse the
3494 response will typically need: as opposed to those which an
3495 OCSP responder itself would need which will be added later.
3497 OCSP_request_sign() signs an OCSP request with an API similar
3498 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3499 response. OCSP_response_get1_basic() extracts basic response
3500 from response. OCSP_resp_find_status(): finds and extracts status
3501 information from an OCSP_CERTID structure (which will be created
3502 when the request structure is built). These are built from lower
3503 level functions which work on OCSP_SINGLERESP structures but
3504 wont normally be used unless the application wishes to examine
3505 extensions in the OCSP response for example.
3507 Replace nonce routines with a pair of functions.
3508 OCSP_request_add1_nonce() adds a nonce value and optionally
3509 generates a random value. OCSP_check_nonce() checks the
3510 validity of the nonce in an OCSP response.
3513 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3514 This doesn't copy the supplied OCSP_CERTID and avoids the
3515 need to free up the newly created id. Change return type
3516 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3517 This can then be used to add extensions to the request.
3518 Deleted OCSP_request_new(), since most of its functionality
3519 is now in OCSP_REQUEST_new() (and the case insensitive name
3520 clash) apart from the ability to set the request name which
3521 will be added elsewhere.
3524 *) Update OCSP API. Remove obsolete extensions argument from
3525 various functions. Extensions are now handled using the new
3526 OCSP extension code. New simple OCSP HTTP function which
3527 can be used to send requests and parse the response.
3530 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3531 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3532 uses the special reorder version of SET OF to sort the attributes
3533 and reorder them to match the encoded order. This resolves a long
3534 standing problem: a verify on a PKCS7 structure just after signing
3535 it used to fail because the attribute order did not match the
3536 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3537 it uses the received order. This is necessary to tolerate some broken
3538 software that does not order SET OF. This is handled by encoding
3539 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3540 to produce the required SET OF.
3543 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3544 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3545 files to get correct declarations of the ASN.1 item variables.
3548 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3549 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3550 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3551 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3552 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3553 ASN1_ITEM and no wrapper functions.
3556 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3557 replace the old function pointer based I/O routines. Change most of
3558 the *_d2i_bio() and *_d2i_fp() functions to use these.
3561 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3562 lines, recognice more "algorithms" that can be deselected, and make
3563 it complain about algorithm deselection that isn't recognised.
3566 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3567 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3568 to use new functions. Add NO_ASN1_OLD which can be set to remove
3569 some old style ASN1 functions: this can be used to determine if old
3570 code will still work when these eventually go away.
3573 *) New extension functions for OCSP structures, these follow the
3574 same conventions as certificates and CRLs.
3577 *) New function X509V3_add1_i2d(). This automatically encodes and
3578 adds an extension. Its behaviour can be customised with various
3579 flags to append, replace or delete. Various wrappers added for
3580 certifcates and CRLs.
3583 *) Fix to avoid calling the underlying ASN1 print routine when
3584 an extension cannot be parsed. Correct a typo in the
3585 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3588 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3589 entries for variables.
3592 *) Add functionality to apps/openssl.c for detecting locking
3593 problems: As the program is single-threaded, all we have
3594 to do is register a locking callback using an array for
3595 storing which locks are currently held by the program.
3598 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3599 SSL_get_ex_data_X509_STORE_idx(), which is used in
3600 ssl_verify_cert_chain() and thus can be called at any time
3601 during TLS/SSL handshakes so that thread-safety is essential.
3602 Unfortunately, the ex_data design is not at all suited
3603 for multi-threaded use, so it probably should be abolished.
3606 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3607 [Broadcom, tweaked and integrated by Geoff Thorpe]
3609 *) Move common extension printing code to new function
3610 X509V3_print_extensions(). Reorganise OCSP print routines and
3611 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3614 *) New function X509_signature_print() to remove duplication in some
3618 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3619 set (this was treated exactly the same as SET OF previously). This
3620 is used to reorder the STACK representing the structure to match the
3621 encoding. This will be used to get round a problem where a PKCS7
3622 structure which was signed could not be verified because the STACK
3623 order did not reflect the encoded order.
3626 *) Reimplement the OCSP ASN1 module using the new code.
3629 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3630 for its ASN1 operations. The old style function pointers still exist
3631 for now but they will eventually go away.
3634 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3635 completely replaces the old ASN1 functionality with a table driven
3636 encoder and decoder which interprets an ASN1_ITEM structure describing
3637 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3638 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3639 has also been converted to the new form.
3642 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3643 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3644 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3645 for negative moduli.
3648 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3649 of not touching the result's sign bit.
3652 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3656 *) Changed the LHASH code to use prototypes for callbacks, and created
3657 macros to declare and implement thin (optionally static) functions
3658 that provide type-safety and avoid function pointer casting for the
3659 type-specific callbacks.
3662 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3664 [Veers Staats <staatsvr@asc.hpc.mil>,
3665 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3667 *) Reformat the FAQ so the different questions and answers can be divided
3668 in sections depending on the subject.
3671 *) Have the zlib compression code load ZLIB.DLL dynamically under
3675 *) New function BN_mod_sqrt for computing square roots modulo a prime
3676 (using the probabilistic Tonelli-Shanks algorithm unless
3677 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3678 be handled deterministically).
3679 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3681 *) Make BN_mod_inverse faster by explicitly handling small quotients
3682 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3683 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3686 *) New function BN_kronecker.
3689 *) Fix BN_gcd so that it works on negative inputs; the result is
3690 positive unless both parameters are zero.
3691 Previously something reasonably close to an infinite loop was
3692 possible because numbers could be growing instead of shrinking
3693 in the implementation of Euclid's algorithm.
3696 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3697 sign of the number in question.
3699 Fix BN_is_word(a,w) to work correctly for w == 0.
3701 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3702 because its test if the absolute value of 'a' equals 'w'.
3703 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3704 it exists mostly for use in the implementations of BN_is_zero(),
3705 BN_is_one(), and BN_is_word().
3708 *) New function BN_swap.
3711 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3712 the exponentiation functions are more likely to produce reasonable
3713 results on negative inputs.
3716 *) Change BN_mod_mul so that the result is always non-negative.
3717 Previously, it could be negative if one of the factors was negative;
3718 I don't think anyone really wanted that behaviour.
3721 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3722 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3723 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3724 and add new functions:
3733 BN_mod_lshift1_quick
3737 These functions always generate non-negative results.
3739 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3740 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3742 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3743 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3744 be reduced modulo m.
3745 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3748 The following entry accidentily appeared in the CHANGES file
3749 distributed with OpenSSL 0.9.7. The modifications described in
3750 it do *not* apply to OpenSSL 0.9.7.
3752 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3753 was actually never needed) and in BN_mul(). The removal in BN_mul()
3754 required a small change in bn_mul_part_recursive() and the addition
3755 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3756 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3757 bn_sub_words() and bn_add_words() except they take arrays with
3762 *) In 'openssl passwd', verify passwords read from the terminal
3763 unless the '-salt' option is used (which usually means that
3764 verification would just waste user's time since the resulting
3765 hash is going to be compared with some given password hash)
3766 or the new '-noverify' option is used.
3768 This is an incompatible change, but it does not affect
3769 non-interactive use of 'openssl passwd' (passwords on the command
3770 line, '-stdin' option, '-in ...' option) and thus should not
3774 *) Remove all references to RSAref, since there's no more need for it.
3777 *) Make DSO load along a path given through an environment variable
3778 (SHLIB_PATH) with shl_load().
3781 *) Constify the ENGINE code as a result of BIGNUM constification.
3782 Also constify the RSA code and most things related to it. In a
3783 few places, most notable in the depth of the ASN.1 code, ugly
3784 casts back to non-const were required (to be solved at a later
3788 *) Make it so the openssl application has all engines loaded by default.
3791 *) Constify the BIGNUM routines a little more.
3794 *) Add the following functions:
3796 ENGINE_load_cswift()
3798 ENGINE_load_atalla()
3800 ENGINE_load_builtin_engines()
3802 That way, an application can itself choose if external engines that
3803 are built-in in OpenSSL shall ever be used or not. The benefit is
3804 that applications won't have to be linked with libdl or other dso
3805 libraries unless it's really needed.
3807 Changed 'openssl engine' to load all engines on demand.
3808 Changed the engine header files to avoid the duplication of some
3809 declarations (they differed!).
3812 *) 'openssl engine' can now list capabilities.
3815 *) Better error reporting in 'openssl engine'.
3818 *) Never call load_dh_param(NULL) in s_server.
3821 *) Add engine application. It can currently list engines by name and
3822 identity, and test if they are actually available.
3825 *) Improve RPM specification file by forcing symbolic linking and making
3826 sure the installed documentation is also owned by root.root.
3827 [Damien Miller <djm@mindrot.org>]
3829 *) Give the OpenSSL applications more possibilities to make use of
3830 keys (public as well as private) handled by engines.
3833 *) Add OCSP code that comes from CertCo.
3836 *) Add VMS support for the Rijndael code.
3839 *) Added untested support for Nuron crypto accelerator.
3842 *) Add support for external cryptographic devices. This code was
3843 previously distributed separately as the "engine" branch.
3844 [Geoff Thorpe, Richard Levitte]
3846 *) Rework the filename-translation in the DSO code. It is now possible to
3847 have far greater control over how a "name" is turned into a filename
3848 depending on the operating environment and any oddities about the
3849 different shared library filenames on each system.
3852 *) Support threads on FreeBSD-elf in Configure.
3855 *) Fix for SHA1 assembly problem with MASM: it produces
3856 warnings about corrupt line number information when assembling
3857 with debugging information. This is caused by the overlapping
3859 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3862 NCONF_get_number() has no error checking at all. As a replacement,
3863 NCONF_get_number_e() is defined (_e for "error checking") and is
3864 promoted strongly. The old NCONF_get_number is kept around for
3865 binary backward compatibility.
3866 Make it possible for methods to load from something other than a BIO,
3867 by providing a function pointer that is given a name instead of a BIO.
3868 For example, this could be used to load configuration data from an
3872 *) Fix for non blocking accept BIOs. Added new I/O special reason
3873 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3874 with non blocking I/O was not possible because no retry code was
3875 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3879 *) Added the beginnings of Rijndael support.
3882 *) Fix for bug in DirectoryString mask setting. Add support for
3883 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3884 to allow certificate printing to more controllable, additional
3885 'certopt' option to 'x509' to allow new printing options to be
3889 *) Clean old EAY MD5 hack from e_os.h.
3892 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3894 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3895 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3896 [Joe Orton, Steve Henson]
3898 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3900 *) Fix additional bug revealed by the NISCC test suite:
3902 Stop bug triggering large recursion when presented with
3903 certain ASN.1 tags (CVE-2003-0851)
3906 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3908 *) Fix various bugs revealed by running the NISCC test suite:
3910 Stop out of bounds reads in the ASN1 code when presented with
3911 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3913 If verify callback ignores invalid public key errors don't try to check
3914 certificate signature with the NULL public key.
3918 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3919 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3923 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3924 extra data after the compression methods not only for TLS 1.0
3925 but also for SSL 3.0 (as required by the specification).
3926 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3928 *) Change X509_certificate_type() to mark the key as exported/exportable
3929 when it's 512 *bits* long, not 512 bytes.
3932 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3934 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3935 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3936 a protocol version number mismatch like a decryption error
3937 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3940 *) Turn on RSA blinding by default in the default implementation
3941 to avoid a timing attack. Applications that don't want it can call
3942 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3943 They would be ill-advised to do so in most cases.
3944 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3946 *) Change RSA blinding code so that it works when the PRNG is not
3947 seeded (in this case, the secret RSA exponent is abused as
3948 an unpredictable seed -- if it is not unpredictable, there
3949 is no point in blinding anyway). Make RSA blinding thread-safe
3950 by remembering the creator's thread ID in rsa->blinding and
3951 having all other threads use local one-time blinding factors
3952 (this requires more computation than sharing rsa->blinding, but
3953 avoids excessive locking; and if an RSA object is not shared
3954 between threads, blinding will still be very fast).
3957 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3959 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3960 via timing by performing a MAC computation even if incorrrect
3961 block cipher padding has been found. This is a countermeasure
3962 against active attacks where the attacker has to distinguish
3963 between bad padding and a MAC verification error. (CVE-2003-0078)
3965 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3966 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3967 Martin Vuagnoux (EPFL, Ilion)]
3969 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3971 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3972 memory from it's contents. This is done with a counter that will
3973 place alternating values in each byte. This can be used to solve
3974 two issues: 1) the removal of calls to memset() by highly optimizing
3975 compilers, and 2) cleansing with other values than 0, since those can
3976 be read through on certain media, for example a swap space on disk.
3979 *) Bugfix: client side session caching did not work with external caching,
3980 because the session->cipher setting was not restored when reloading
3981 from the external cache. This problem was masked, when
3982 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3983 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3986 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3987 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3988 [Zeev Lieber <zeev-l@yahoo.com>]
3990 *) Undo an undocumented change introduced in 0.9.6e which caused
3991 repeated calls to OpenSSL_add_all_ciphers() and
3992 OpenSSL_add_all_digests() to be ignored, even after calling
3996 *) Change the default configuration reader to deal with last line not
3997 being properly terminated.
4000 *) Change X509_NAME_cmp() so it applies the special rules on handling
4001 DN values that are of type PrintableString, as well as RDNs of type
4002 emailAddress where the value has the type ia5String.
4003 [stefank@valicert.com via Richard Levitte]
4005 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4006 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4007 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4008 the bitwise-OR of the two for use by the majority of applications
4009 wanting this behaviour, and update the docs. The documented
4010 behaviour and actual behaviour were inconsistent and had been
4011 changing anyway, so this is more a bug-fix than a behavioural
4013 [Geoff Thorpe, diagnosed by Nadav Har'El]
4015 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4016 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4019 *) Fix initialization code race conditions in
4020 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4021 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4022 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4023 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4024 ssl2_get_cipher_by_char(),
4025 ssl3_get_cipher_by_char().
4026 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4028 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4029 the cached sessions are flushed, as the remove_cb() might use ex_data
4030 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4031 (see [openssl.org #212]).
4032 [Geoff Thorpe, Lutz Jaenicke]
4034 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4035 length, instead of the encoding length to d2i_ASN1_OBJECT.
4038 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4040 *) [In 0.9.6g-engine release:]
4041 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4042 [Lynn Gazis <lgazis@rainbow.com>]
4044 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4046 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4047 and get fix the header length calculation.
4048 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4049 Alon Kantor <alonk@checkpoint.com> (and others),
4052 *) Use proper error handling instead of 'assertions' in buffer
4053 overflow checks added in 0.9.6e. This prevents DoS (the
4054 assertions could call abort()).
4055 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4057 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4059 *) Add various sanity checks to asn1_get_length() to reject
4060 the ASN1 length bytes if they exceed sizeof(long), will appear
4061 negative or the content length exceeds the length of the
4063 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4065 *) Fix cipher selection routines: ciphers without encryption had no flags
4066 for the cipher strength set and where therefore not handled correctly
4067 by the selection routines (PR #130).
4070 *) Fix EVP_dsa_sha macro.
4074 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4075 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4076 that was added in OpenSSL 0.9.6d.
4078 As the countermeasure turned out to be incompatible with some
4079 broken SSL implementations, the new option is part of SSL_OP_ALL.
4080 SSL_OP_ALL is usually employed when compatibility with weird SSL
4081 implementations is desired (e.g. '-bugs' option to 's_client' and
4082 's_server'), so the new option is automatically set in many
4086 *) Changes in security patch:
4088 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4089 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4090 Air Force Materiel Command, USAF, under agreement number
4093 *) Add various sanity checks to asn1_get_length() to reject
4094 the ASN1 length bytes if they exceed sizeof(long), will appear
4095 negative or the content length exceeds the length of the
4096 supplied buffer. (CVE-2002-0659)
4097 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4099 *) Assertions for various potential buffer overflows, not known to
4101 [Ben Laurie (CHATS)]
4103 *) Various temporary buffers to hold ASCII versions of integers were
4104 too small for 64 bit platforms. (CVE-2002-0655)
4105 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4107 *) Remote buffer overflow in SSL3 protocol - an attacker could
4108 supply an oversized session ID to a client. (CVE-2002-0656)
4109 [Ben Laurie (CHATS)]
4111 *) Remote buffer overflow in SSL2 protocol - an attacker could
4112 supply an oversized client master key. (CVE-2002-0656)
4113 [Ben Laurie (CHATS)]
4115 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4117 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4118 encoded as NULL) with id-dsa-with-sha1.
4119 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4121 *) Check various X509_...() return values in apps/req.c.
4122 [Nils Larsch <nla@trustcenter.de>]
4124 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4125 an end-of-file condition would erronously be flagged, when the CRLF
4126 was just at the end of a processed block. The bug was discovered when
4127 processing data through a buffering memory BIO handing the data to a
4128 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4129 <ptsekov@syntrex.com> and Nedelcho Stanev.
4132 *) Implement a countermeasure against a vulnerability recently found
4133 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4134 before application data chunks to avoid the use of known IVs
4135 with data potentially chosen by the attacker.
4138 *) Fix length checks in ssl3_get_client_hello().
4141 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4142 to prevent ssl3_read_internal() from incorrectly assuming that
4143 ssl3_read_bytes() found application data while handshake
4144 processing was enabled when in fact s->s3->in_read_app_data was
4145 merely automatically cleared during the initial handshake.
4146 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4148 *) Fix object definitions for Private and Enterprise: they were not
4149 recognized in their shortname (=lowercase) representation. Extend
4150 obj_dat.pl to issue an error when using undefined keywords instead
4151 of silently ignoring the problem (Svenning Sorensen
4152 <sss@sss.dnsalias.net>).
4155 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4156 generators, i.e. generators other than 2 and 5. (Previously, the
4157 code did not properly initialise the 'add' and 'rem' values to
4158 BN_generate_prime().)
4160 In the new general case, we do not insist that 'generator' is
4161 actually a primitive root: This requirement is rather pointless;
4162 a generator of the order-q subgroup is just as good, if not
4166 *) Map new X509 verification errors to alerts. Discovered and submitted by
4167 Tom Wu <tom@arcot.com>.
4170 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4171 returning non-zero before the data has been completely received
4172 when using non-blocking I/O.
4173 [Bodo Moeller; problem pointed out by John Hughes]
4175 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4176 [Ben Laurie, Lutz Jaenicke]
4178 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4179 Yoram Zahavi <YoramZ@gilian.com>).
4182 *) Add information about CygWin 1.3 and on, and preserve proper
4183 configuration for the versions before that.
4184 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4186 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4187 check whether we deal with a copy of a session and do not delete from
4188 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4189 <izhar@checkpoint.com>.
4192 *) Do not store session data into the internal session cache, if it
4193 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4194 flag is set). Proposed by Aslam <aslam@funk.com>.
4197 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4201 *) [In 0.9.6d-engine release:]
4202 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4203 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4205 *) Add the configuration target linux-s390x.
4206 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4208 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4209 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4210 variable as an indication that a ClientHello message has been
4211 received. As the flag value will be lost between multiple
4212 invocations of ssl3_accept when using non-blocking I/O, the
4213 function may not be aware that a handshake has actually taken
4214 place, thus preventing a new session from being added to the
4217 To avoid this problem, we now set s->new_session to 2 instead of
4218 using a local variable.
4219 [Lutz Jaenicke, Bodo Moeller]
4221 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4222 if the SSL_R_LENGTH_MISMATCH error is detected.
4223 [Geoff Thorpe, Bodo Moeller]
4225 *) New 'shared_ldflag' column in Configure platform table.
4228 *) Fix EVP_CIPHER_mode macro.
4229 ["Dan S. Camper" <dan@bti.net>]
4231 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4232 type, we must throw them away by setting rr->length to 0.
4233 [D P Chang <dpc@qualys.com>]
4235 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4237 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4238 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4239 worked incorrectly for those cases where range = 10..._2 and
4240 3*range is two bits longer than range.)
4243 *) Only add signing time to PKCS7 structures if it is not already
4247 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4248 OBJ_ld_ce should be OBJ_id_ce.
4249 Also some ip-pda OIDs in crypto/objects/objects.txt were
4250 incorrect (cf. RFC 3039).
4251 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4253 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4254 returns early because it has nothing to do.
4255 [Andy Schneider <andy.schneider@bjss.co.uk>]
4257 *) [In 0.9.6c-engine release:]
4258 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4259 [Andy Schneider <andy.schneider@bjss.co.uk>]
4261 *) [In 0.9.6c-engine release:]
4262 Add support for Cryptographic Appliance's keyserver technology.
4263 (Use engine 'keyclient')
4264 [Cryptographic Appliances and Geoff Thorpe]
4266 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4267 is called via tools/c89.sh because arguments have to be
4268 rearranged (all '-L' options must appear before the first object
4270 [Richard Shapiro <rshapiro@abinitio.com>]
4272 *) [In 0.9.6c-engine release:]
4273 Add support for Broadcom crypto accelerator cards, backported
4275 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4277 *) [In 0.9.6c-engine release:]
4278 Add support for SureWare crypto accelerator cards from
4279 Baltimore Technologies. (Use engine 'sureware')
4280 [Baltimore Technologies and Mark Cox]
4282 *) [In 0.9.6c-engine release:]
4283 Add support for crypto accelerator cards from Accelerated
4284 Encryption Processing, www.aep.ie. (Use engine 'aep')
4285 [AEP Inc. and Mark Cox]
4287 *) Add a configuration entry for gcc on UnixWare.
4288 [Gary Benson <gbenson@redhat.com>]
4290 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4291 messages are stored in a single piece (fixed-length part and
4292 variable-length part combined) and fix various bugs found on the way.
4295 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4296 instead. BIO_gethostbyname() does not know what timeouts are
4297 appropriate, so entries would stay in cache even when they have
4299 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4301 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4302 faced with a pathologically small ClientHello fragment that does
4303 not contain client_version: Instead of aborting with an error,
4304 simply choose the highest available protocol version (i.e.,
4305 TLS 1.0 unless it is disabled). In practice, ClientHello
4306 messages are never sent like this, but this change gives us
4307 strictly correct behaviour at least for TLS.
4310 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4311 never resets s->method to s->ctx->method when called from within
4312 one of the SSL handshake functions.
4313 [Bodo Moeller; problem pointed out by Niko Baric]
4315 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4316 (sent using the client's version number) if client_version is
4317 smaller than the protocol version in use. Also change
4318 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4319 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4320 the client will at least see that alert.
4323 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4327 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4328 client receives HelloRequest while in a handshake.
4329 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4331 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4332 should end in 'break', not 'goto end' which circuments various
4333 cleanups done in state SSL_ST_OK. But session related stuff
4334 must be disabled for SSL_ST_OK in the case that we just sent a
4337 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4338 before just sending a HelloRequest.
4339 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4341 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4342 reveal whether illegal block cipher padding was found or a MAC
4343 verification error occured. (Neither SSLerr() codes nor alerts
4344 are directly visible to potential attackers, but the information
4345 may leak via logfiles.)
4347 Similar changes are not required for the SSL 2.0 implementation
4348 because the number of padding bytes is sent in clear for SSL 2.0,
4349 and the extra bytes are just ignored. However ssl/s2_pkt.c
4350 failed to verify that the purported number of padding bytes is in
4354 *) Add OpenUNIX-8 support including shared libraries
4355 (Boyd Lynn Gerber <gerberb@zenez.com>).
4358 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4359 'wristwatch attack' using huge encoding parameters (cf.
4360 James H. Manger's CRYPTO 2001 paper). Note that the
4361 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4362 encoding parameters and hence was not vulnerable.
4365 *) BN_sqr() bug fix.
4366 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4368 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4369 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4370 followed by modular reduction.
4371 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4373 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4374 equivalent based on BN_pseudo_rand() instead of BN_rand().
4377 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4378 This function was broken, as the check for a new client hello message
4379 to handle SGC did not allow these large messages.
4380 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4383 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4386 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4387 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4390 *) Rework the configuration and shared library support for Tru64 Unix.
4391 The configuration part makes use of modern compiler features and
4392 still retains old compiler behavior for those that run older versions
4393 of the OS. The shared library support part includes a variant that
4394 uses the RPATH feature, and is available through the special
4395 configuration target "alpha-cc-rpath", which will never be selected
4397 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4399 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4400 with the same message size as in ssl3_get_certificate_request().
4401 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4402 messages might inadvertently be reject as too long.
4403 [Petr Lampa <lampa@fee.vutbr.cz>]
4405 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4408 *) Modified SSL library such that the verify_callback that has been set
4409 specificly for an SSL object with SSL_set_verify() is actually being
4410 used. Before the change, a verify_callback set with this function was
4411 ignored and the verify_callback() set in the SSL_CTX at the time of
4412 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4413 to allow the necessary settings.
4416 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4417 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4418 done automatically (in contradiction to the requirements of the C
4419 standard). This made problems when used from OpenSSH.
4422 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4423 dh->length and always used
4425 BN_rand_range(priv_key, dh->p).
4427 BN_rand_range() is not necessary for Diffie-Hellman, and this
4428 specific range makes Diffie-Hellman unnecessarily inefficient if
4429 dh->length (recommended exponent length) is much smaller than the
4430 length of dh->p. We could use BN_rand_range() if the order of
4431 the subgroup was stored in the DH structure, but we only have
4436 BN_rand(priv_key, l, ...)
4438 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4444 RSA_eay_public_encrypt
4445 RSA_eay_private_decrypt
4446 RSA_eay_private_encrypt (signing)
4447 RSA_eay_public_decrypt (signature verification)
4449 (default implementations for RSA_public_encrypt,
4450 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4451 always reject numbers >= n.
4454 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4455 to synchronize access to 'locking_thread'. This is necessary on
4456 systems where access to 'locking_thread' (an 'unsigned long'
4457 variable) is not atomic.
4460 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4461 *before* setting the 'crypto_lock_rand' flag. The previous code had
4462 a race condition if 0 is a valid thread ID.
4463 [Travis Vitek <vitek@roguewave.com>]
4465 *) Add support for shared libraries under Irix.
4466 [Albert Chin-A-Young <china@thewrittenword.com>]
4468 *) Add configuration option to build on Linux on both big-endian and
4470 [Ralf Baechle <ralf@uni-koblenz.de>]
4472 *) Add the possibility to create shared libraries on HP-UX.
4475 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4477 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4478 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4479 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4480 PRNG state recovery was possible based on the output of
4481 one PRNG request appropriately sized to gain knowledge on
4482 'md' followed by enough consecutive 1-byte PRNG requests
4483 to traverse all of 'state'.
4485 1. When updating 'md_local' (the current thread's copy of 'md')
4486 during PRNG output generation, hash all of the previous
4487 'md_local' value, not just the half used for PRNG output.
4489 2. Make the number of bytes from 'state' included into the hash
4490 independent from the number of PRNG bytes requested.
4492 The first measure alone would be sufficient to avoid
4493 Markku-Juhani's attack. (Actually it had never occurred
4494 to me that the half of 'md_local' used for chaining was the
4495 half from which PRNG output bytes were taken -- I had always
4496 assumed that the secret half would be used.) The second
4497 measure makes sure that additional data from 'state' is never
4498 mixed into 'md_local' in small portions; this heuristically
4499 further strengthens the PRNG.
4502 *) Fix crypto/bn/asm/mips3.s.
4505 *) When only the key is given to "enc", the IV is undefined. Print out
4506 an error message in this case.
4509 *) Handle special case when X509_NAME is empty in X509 printing routines.
4512 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4513 positive and less than q.
4516 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4517 used: it isn't thread safe and the add_lock_callback should handle
4519 [Paul Rose <Paul.Rose@bridge.com>]
4521 *) Verify that incoming data obeys the block size in
4522 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4526 [Ulf Möller, Bodo Möller]
4528 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4529 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4530 when fixing the server behaviour for backwards-compatible 'client
4531 hello' messages. (Note that the attack is impractical against
4532 SSL 3.0 and TLS 1.0 anyway because length and version checking
4533 means that the probability of guessing a valid ciphertext is
4534 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4537 Before 0.9.5, the countermeasure (hide the error by generating a
4538 random 'decryption result') did not work properly because
4539 ERR_clear_error() was missing, meaning that SSL_get_error() would
4540 detect the supposedly ignored error.
4542 Both problems are now fixed.
4545 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4546 (previously it was 1024).
4549 *) Fix for compatibility mode trust settings: ignore trust settings
4550 unless some valid trust or reject settings are present.
4553 *) Fix for blowfish EVP: its a variable length cipher.
4556 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4557 parameters in DSA public key structures and return an error in the
4558 DSA routines if parameters are absent.
4561 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4562 in the current directory if neither $RANDFILE nor $HOME was set.
4563 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4564 caused some confusion to Windows users who haven't defined $HOME.
4565 Thus RAND_file_name() is changed again: e_os.h can define a
4566 DEFAULT_HOME, which will be used if $HOME is not set.
4567 For Windows, we use "C:"; on other platforms, we still require
4568 environment variables.
4570 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4571 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4572 having multiple threads call RAND_poll() concurrently.
4575 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4576 combination of a flag and a thread ID variable.
4577 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4578 flag), *other* threads can enter ssleay_add_bytes without obeying
4579 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4580 that they do not hold after the first thread unsets add_do_not_lock).
4583 *) Change bctest again: '-x' expressions are not available in all
4587 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4589 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4590 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4592 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4593 the default extension for executables, if any. Also, make the perl
4594 scripts that use symlink() to test if it really exists and use "cp"
4595 if it doesn't. All this made OpenSSL compilable and installable in
4599 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4600 If SEQUENCE is length is indefinite just set c->slen to the total
4601 amount of data available.
4602 [Steve Henson, reported by shige@FreeBSD.org]
4603 [This change does not apply to 0.9.7.]
4605 *) Change bctest to avoid here-documents inside command substitution
4606 (workaround for FreeBSD /bin/sh bug).
4607 For compatibility with Ultrix, avoid shell functions (introduced
4608 in the bctest version that searches along $PATH).
4611 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4612 with des_encrypt() defined on some operating systems, like Solaris
4616 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4617 On the Importance of Eliminating Errors in Cryptographic
4618 Computations, J. Cryptology 14 (2001) 2, 101-119,
4619 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4622 *) MIPS assembler BIGNUM division bug fix.
4625 *) Disabled incorrect Alpha assembler code.
4628 *) Fix PKCS#7 decode routines so they correctly update the length
4629 after reading an EOC for the EXPLICIT tag.
4631 [This change does not apply to 0.9.7.]
4633 *) Fix bug in PKCS#12 key generation routines. This was triggered
4634 if a 3DES key was generated with a 0 initial byte. Include
4635 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4636 (but broken) behaviour.
4639 *) Enhance bctest to search for a working bc along $PATH and print
4641 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4643 *) Fix memory leaks in err.c: free err_data string if necessary;
4644 don't write to the wrong index in ERR_set_error_data.
4647 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4651 *) Replace rdtsc with _emit statements for VC++ version 5.
4652 [Jeremy Cooper <jeremy@baymoo.org>]
4654 *) Make it possible to reuse SSLv2 sessions.
4657 *) In copy_email() check for >= 0 as a return value for
4658 X509_NAME_get_index_by_NID() since 0 is a valid index.
4659 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4661 *) Avoid coredump with unsupported or invalid public keys by checking if
4662 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4663 PKCS7_verify() fails with non detached data.
4666 *) Don't use getenv in library functions when run as setuid/setgid.
4667 New function OPENSSL_issetugid().
4670 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4671 due to incorrect handling of multi-threading:
4673 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4675 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4677 3. Count how many times MemCheck_off() has been called so that
4678 nested use can be treated correctly. This also avoids
4679 inband-signalling in the previous code (which relied on the
4680 assumption that thread ID 0 is impossible).
4683 *) Add "-rand" option also to s_client and s_server.
4686 *) Fix CPU detection on Irix 6.x.
4687 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4688 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4690 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4693 [This change does not apply to 0.9.7.]
4695 *) Use the cached encoding of an X509_NAME structure rather than
4696 copying it. This is apparently the reason for the libsafe "errors"
4697 but the code is actually correct.
4700 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4701 Bleichenbacher's DSA attack.
4702 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4703 to be set and top=0 forces the highest bit to be set; top=-1 is new
4704 and leaves the highest bit random.
4705 [Ulf Moeller, Bodo Moeller]
4707 *) In the NCONF_...-based implementations for CONF_... queries
4708 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4709 a temporary CONF structure with the data component set to NULL
4710 (which gives segmentation faults in lh_retrieve).
4711 Instead, use NULL for the CONF pointer in CONF_get_string and
4712 CONF_get_number (which may use environment variables) and directly
4713 return NULL from CONF_get_section.
4716 *) Fix potential buffer overrun for EBCDIC.
4719 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4720 keyUsage if basicConstraints absent for a CA.
4723 *) Make SMIME_write_PKCS7() write mail header values with a format that
4724 is more generally accepted (no spaces before the semicolon), since
4725 some programs can't parse those values properly otherwise. Also make
4726 sure BIO's that break lines after each write do not create invalid
4730 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4731 macros previously used would not encode an empty SEQUENCE OF
4732 and break the signature.
4734 [This change does not apply to 0.9.7.]
4736 *) Zero the premaster secret after deriving the master secret in
4740 *) Add some EVP_add_digest_alias registrations (as found in
4741 OpenSSL_add_all_digests()) to SSL_library_init()
4742 aka OpenSSL_add_ssl_algorithms(). This provides improved
4743 compatibility with peers using X.509 certificates
4744 with unconventional AlgorithmIdentifier OIDs.
4747 *) Fix for Irix with NO_ASM.
4748 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4750 *) ./config script fixes.
4751 [Ulf Moeller, Richard Levitte]
4753 *) Fix 'openssl passwd -1'.
4756 *) Change PKCS12_key_gen_asc() so it can cope with non null
4757 terminated strings whose length is passed in the passlen
4758 parameter, for example from PEM callbacks. This was done
4759 by adding an extra length parameter to asc2uni().
4760 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4762 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4763 call failed, free the DSA structure.
4766 *) Fix to uni2asc() to cope with zero length Unicode strings.
4767 These are present in some PKCS#12 files.
4770 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4771 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4772 when writing a 32767 byte record.
4773 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4775 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4776 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4778 (RSA objects have a reference count access to which is protected
4779 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4780 so they are meant to be shared between threads.)
4781 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4782 "Reddie, Steven" <Steven.Reddie@ca.com>]
4784 *) Fix a deadlock in CRYPTO_mem_leaks().
4787 *) Use better test patterns in bntest.
4790 *) rand_win.c fix for Borland C.
4793 *) BN_rshift bugfix for n == 0.
4796 *) Add a 'bctest' script that checks for some known 'bc' bugs
4797 so that 'make test' does not abort just because 'bc' is broken.
4800 *) Store verify_result within SSL_SESSION also for client side to
4801 avoid potential security hole. (Re-used sessions on the client side
4802 always resulted in verify_result==X509_V_OK, not using the original
4803 result of the server certificate verification.)
4806 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4807 SSL3_RT_APPLICATION_DATA, return 0.
4808 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4812 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4813 releases, have been re-implemented by renaming the previous
4814 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4815 and ssl3_read_internal, respectively, and adding 'peek' parameters
4816 to them. The new ssl[23]_{read,peek} functions are calls to
4817 ssl[23]_read_internal with the 'peek' flag set appropriately.
4818 A 'peek' parameter has also been added to ssl3_read_bytes, which
4819 does the actual work for ssl3_read_internal.
4822 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4823 the method-specific "init()" handler. Also clean up ex_data after
4824 calling the method-specific "finish()" handler. Previously, this was
4825 happening the other way round.
4828 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4829 The previous value, 12, was not always sufficient for BN_mod_exp().
4832 *) Make sure that shared libraries get the internal name engine with
4833 the full version number and not just 0. This should mark the
4834 shared libraries as not backward compatible. Of course, this should
4835 be changed again when we can guarantee backward binary compatibility.
4838 *) Fix typo in get_cert_by_subject() in by_dir.c
4839 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4841 *) Rework the system to generate shared libraries:
4843 - Make note of the expected extension for the shared libraries and
4844 if there is a need for symbolic links from for example libcrypto.so.0
4845 to libcrypto.so.0.9.7. There is extended info in Configure for
4848 - Make as few rebuilds of the shared libraries as possible.
4850 - Still avoid linking the OpenSSL programs with the shared libraries.
4852 - When installing, install the shared libraries separately from the
4856 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4858 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4859 and not in SSL_clear because the latter is also used by the
4860 accept/connect functions; previously, the settings made by
4861 SSL_set_read_ahead would be lost during the handshake.
4862 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4864 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4865 Previously, it would create entries for disableed algorithms no
4869 *) Added several new manual pages for SSL_* function.
4872 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4874 *) In ssl23_get_client_hello, generate an error message when faced
4875 with an initial SSL 3.0/TLS record that is too small to contain the
4876 first two bytes of the ClientHello message, i.e. client_version.
4877 (Note that this is a pathologic case that probably has never happened
4878 in real life.) The previous approach was to use the version number
4879 from the record header as a substitute; but our protocol choice
4880 should not depend on that one because it is not authenticated
4881 by the Finished messages.
4884 *) More robust randomness gathering functions for Windows.
4885 [Jeffrey Altman <jaltman@columbia.edu>]
4887 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4888 not set then we don't setup the error code for issuer check errors
4889 to avoid possibly overwriting other errors which the callback does
4890 handle. If an application does set the flag then we assume it knows
4891 what it is doing and can handle the new informational codes
4895 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4896 a general "ANY" type, as such it should be able to decode anything
4897 including tagged types. However it didn't check the class so it would
4898 wrongly interpret tagged types in the same way as their universal
4899 counterpart and unknown types were just rejected. Changed so that the
4900 tagged and unknown types are handled in the same way as a SEQUENCE:
4901 that is the encoding is stored intact. There is also a new type
4902 "V_ASN1_OTHER" which is used when the class is not universal, in this
4903 case we have no idea what the actual type is so we just lump them all
4907 *) On VMS, stdout may very well lead to a file that is written to
4908 in a record-oriented fashion. That means that every write() will
4909 write a separate record, which will be read separately by the
4910 programs trying to read from it. This can be very confusing.
4912 The solution is to put a BIO filter in the way that will buffer
4913 text until a linefeed is reached, and then write everything a
4914 line at a time, so every record written will be an actual line,
4915 not chunks of lines and not (usually doesn't happen, but I've
4916 seen it once) several lines in one record. BIO_f_linebuffer() is
4919 Currently, it's a VMS-only method, because that's where it has
4920 been tested well enough.
4923 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4924 it can return incorrect results.
4925 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4926 but it was in 0.9.6-beta[12].)
4929 *) Disable the check for content being present when verifying detached
4930 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4931 include zero length content when signing messages.
4934 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4935 BIO_ctrl (for BIO pairs).
4938 *) Add DSO method for VMS.
4941 *) Bug fix: Montgomery multiplication could produce results with the
4945 *) Add RPM specification openssl.spec and modify it to build three
4946 packages. The default package contains applications, application
4947 documentation and run-time libraries. The devel package contains
4948 include files, static libraries and function documentation. The
4949 doc package contains the contents of the doc directory. The original
4950 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4953 *) Add a large number of documentation files for many SSL routines.
4954 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4956 *) Add a configuration entry for Sony News 4.
4957 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4959 *) Don't set the two most significant bits to one when generating a
4960 random number < q in the DSA library.
4963 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4964 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4965 the underlying transport is blocking) if a handshake took place.
4966 (The default behaviour is needed by applications such as s_client
4967 and s_server that use select() to determine when to use SSL_read;
4968 but for applications that know in advance when to expect data, it
4969 just makes things more complicated.)
4972 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4976 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4977 work better on such systems.
4978 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4980 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4981 Update PKCS12_parse() so it copies the friendlyName and the
4982 keyid to the certificates aux info.
4985 *) Fix bug in PKCS7_verify() which caused an infinite loop
4986 if there was more than one signature.
4987 [Sven Uszpelkat <su@celocom.de>]
4989 *) Major change in util/mkdef.pl to include extra information
4990 about each symbol, as well as presentig variables as well
4991 as functions. This change means that there's n more need
4992 to rebuild the .num files when some algorithms are excluded.
4995 *) Allow the verify time to be set by an application,
4996 rather than always using the current time.
4999 *) Phase 2 verify code reorganisation. The certificate
5000 verify code now looks up an issuer certificate by a
5001 number of criteria: subject name, authority key id
5002 and key usage. It also verifies self signed certificates
5003 by the same criteria. The main comparison function is
5004 X509_check_issued() which performs these checks.
5006 Lot of changes were necessary in order to support this
5007 without completely rewriting the lookup code.
5009 Authority and subject key identifier are now cached.
5011 The LHASH 'certs' is X509_STORE has now been replaced
5012 by a STACK_OF(X509_OBJECT). This is mainly because an
5013 LHASH can't store or retrieve multiple objects with
5014 the same hash value.
5016 As a result various functions (which were all internal
5017 use only) have changed to handle the new X509_STORE
5018 structure. This will break anything that messed round
5019 with X509_STORE internally.
5021 The functions X509_STORE_add_cert() now checks for an
5022 exact match, rather than just subject name.
5024 The X509_STORE API doesn't directly support the retrieval
5025 of multiple certificates matching a given criteria, however
5026 this can be worked round by performing a lookup first
5027 (which will fill the cache with candidate certificates)
5028 and then examining the cache for matches. This is probably
5029 the best we can do without throwing out X509_LOOKUP
5030 entirely (maybe later...).
5032 The X509_VERIFY_CTX structure has been enhanced considerably.
5034 All certificate lookup operations now go via a get_issuer()
5035 callback. Although this currently uses an X509_STORE it
5036 can be replaced by custom lookups. This is a simple way
5037 to bypass the X509_STORE hackery necessary to make this
5038 work and makes it possible to use more efficient techniques
5039 in future. A very simple version which uses a simple
5040 STACK for its trusted certificate store is also provided
5041 using X509_STORE_CTX_trusted_stack().
5043 The verify_cb() and verify() callbacks now have equivalents
5044 in the X509_STORE_CTX structure.
5046 X509_STORE_CTX also has a 'flags' field which can be used
5047 to customise the verify behaviour.
5050 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5051 excludes S/MIME capabilities.
5054 *) When a certificate request is read in keep a copy of the
5055 original encoding of the signed data and use it when outputing
5056 again. Signatures then use the original encoding rather than
5057 a decoded, encoded version which may cause problems if the
5058 request is improperly encoded.
5061 *) For consistency with other BIO_puts implementations, call
5062 buffer_write(b, ...) directly in buffer_puts instead of calling
5065 In BIO_puts, increment b->num_write as in BIO_write.
5066 [Peter.Sylvester@EdelWeb.fr]
5068 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5069 BN_zero, we may not return a BIGNUM with an array consisting of
5073 *) Avoid calling abort() from within the library when problems are
5074 detected, except if preprocessor symbols have been defined
5075 (such as REF_CHECK, BN_DEBUG etc.).
5078 *) New openssl application 'rsautl'. This utility can be
5079 used for low level RSA operations. DER public key
5080 BIO/fp routines also added.
5083 *) New Configure entry and patches for compiling on QNX 4.
5084 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5086 *) A demo state-machine implementation was sponsored by
5087 Nuron (http://www.nuron.com/) and is now available in
5088 demos/state_machine.
5091 *) New options added to the 'dgst' utility for signature
5092 generation and verification.
5095 *) Unrecognized PKCS#7 content types are now handled via a
5096 catch all ASN1_TYPE structure. This allows unsupported
5097 types to be stored as a "blob" and an application can
5098 encode and decode it manually.
5101 *) Fix various signed/unsigned issues to make a_strex.c
5103 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5105 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5106 length if passed a buffer. ASN1_INTEGER_to_BN failed
5107 if passed a NULL BN and its argument was negative.
5108 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5110 *) Modification to PKCS#7 encoding routines to output definite
5111 length encoding. Since currently the whole structures are in
5112 memory there's not real point in using indefinite length
5113 constructed encoding. However if OpenSSL is compiled with
5114 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5117 *) Added BIO_vprintf() and BIO_vsnprintf().
5120 *) Added more prefixes to parse for in the the strings written
5121 through a logging bio, to cover all the levels that are available
5122 through syslog. The prefixes are now:
5124 PANIC, EMERG, EMR => LOG_EMERG
5125 ALERT, ALR => LOG_ALERT
5126 CRIT, CRI => LOG_CRIT
5127 ERROR, ERR => LOG_ERR
5128 WARNING, WARN, WAR => LOG_WARNING
5129 NOTICE, NOTE, NOT => LOG_NOTICE
5130 INFO, INF => LOG_INFO
5131 DEBUG, DBG => LOG_DEBUG
5133 and as before, if none of those prefixes are present at the
5134 beginning of the string, LOG_ERR is chosen.
5136 On Win32, the LOG_* levels are mapped according to this:
5138 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5139 LOG_WARNING => EVENTLOG_WARNING_TYPE
5140 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5144 *) Made it possible to reconfigure with just the configuration
5145 argument "reconf" or "reconfigure". The command line arguments
5146 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5147 and are retrieved from there when reconfiguring.
5151 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5153 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5156 *) The obj_dat.pl script was messing up the sorting of object
5157 names. The reason was that it compared the quoted version
5158 of strings as a result "OCSP" > "OCSP Signing" because
5159 " > SPACE. Changed script to store unquoted versions of
5160 names and add quotes on output. It was also omitting some
5161 names from the lookup table if they were given a default
5162 value (that is if SN is missing it is given the same
5163 value as LN and vice versa), these are now added on the
5164 grounds that if an object has a name we should be able to
5165 look it up. Finally added warning output when duplicate
5166 short or long names are found.
5169 *) Changes needed for Tandem NSK.
5170 [Scott Uroff <scott@xypro.com>]
5172 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5173 RSA_padding_check_SSLv23(), special padding was never detected
5174 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5175 version rollback attacks was not effective.
5177 In s23_clnt.c, don't use special rollback-attack detection padding
5178 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5179 client; similarly, in s23_srvr.c, don't do the rollback check if
5180 SSL 2.0 is the only protocol enabled in the server.
5183 *) Make it possible to get hexdumps of unprintable data with 'openssl
5184 asn1parse'. By implication, the functions ASN1_parse_dump() and
5185 BIO_dump_indent() are added.
5188 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5189 these print out strings and name structures based on various
5190 flags including RFC2253 support and proper handling of
5191 multibyte characters. Added options to the 'x509' utility
5192 to allow the various flags to be set.
5195 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5196 Also change the functions X509_cmp_current_time() and
5197 X509_gmtime_adj() work with an ASN1_TIME structure,
5198 this will enable certificates using GeneralizedTime in validity
5199 dates to be checked.
5202 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5203 negative public key encodings) on by default,
5204 NO_NEG_PUBKEY_BUG can be set to disable it.
5207 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5208 content octets. An i2c_ASN1_OBJECT is unnecessary because
5209 the encoding can be trivially obtained from the structure.
5212 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5213 not read locks (CRYPTO_r_[un]lock).
5216 *) A first attempt at creating official support for shared
5217 libraries through configuration. I've kept it so the
5218 default is static libraries only, and the OpenSSL programs
5219 are always statically linked for now, but there are
5220 preparations for dynamic linking in place.
5221 This has been tested on Linux and Tru64.
5224 *) Randomness polling function for Win9x, as described in:
5225 Peter Gutmann, Software Generation of Practically Strong
5229 *) Fix so PRNG is seeded in req if using an already existing
5233 *) New options to smime application. -inform and -outform
5234 allow alternative formats for the S/MIME message including
5235 PEM and DER. The -content option allows the content to be
5236 specified separately. This should allow things like Netscape
5237 form signing output easier to verify.
5240 *) Fix the ASN1 encoding of tags using the 'long form'.
5243 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5244 STRING types. These convert content octets to and from the
5245 underlying type. The actual tag and length octets are
5246 already assumed to have been read in and checked. These
5247 are needed because all other string types have virtually
5248 identical handling apart from the tag. By having versions
5249 of the ASN1 functions that just operate on content octets
5250 IMPLICIT tagging can be handled properly. It also allows
5251 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5252 and ASN1_INTEGER are identical apart from the tag.
5255 *) Change the handling of OID objects as follows:
5257 - New object identifiers are inserted in objects.txt, following
5258 the syntax given in objects.README.
5259 - objects.pl is used to process obj_mac.num and create a new
5261 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5264 This is currently kind of a hack, and the perl code in objects.pl
5265 isn't very elegant, but it works as I intended. The simplest way
5266 to check that it worked correctly is to look in obj_dat.h and
5267 check the array nid_objs and make sure the objects haven't moved
5268 around (this is important!). Additions are OK, as well as
5269 consistent name changes.
5272 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5275 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5276 The given file adds to whatever has already been seeded into the
5277 random pool through the RANDFILE configuration file option or
5278 environment variable, or the default random state file.
5281 *) mkstack.pl now sorts each macro group into lexical order.
5282 Previously the output order depended on the order the files
5283 appeared in the directory, resulting in needless rewriting
5287 *) Patches to make OpenSSL compile under Win32 again. Mostly
5288 work arounds for the VC++ problem that it treats func() as
5289 func(void). Also stripped out the parts of mkdef.pl that
5290 added extra typesafe functions: these no longer exist.
5293 *) Reorganisation of the stack code. The macros are now all
5294 collected in safestack.h . Each macro is defined in terms of
5295 a "stack macro" of the form SKM_<name>(type, a, b). The
5296 DEBUG_SAFESTACK is now handled in terms of function casts,
5297 this has the advantage of retaining type safety without the
5298 use of additional functions. If DEBUG_SAFESTACK is not defined
5299 then the non typesafe macros are used instead. Also modified the
5300 mkstack.pl script to handle the new form. Needs testing to see
5301 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5302 the default if no major problems. Similar behaviour for ASN1_SET_OF
5303 and PKCS12_STACK_OF.
5306 *) When some versions of IIS use the 'NET' form of private key the
5307 key derivation algorithm is different. Normally MD5(password) is
5308 used as a 128 bit RC4 key. In the modified case
5309 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5310 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5311 as the old Netscape_RSA functions except they have an additional
5312 'sgckey' parameter which uses the modified algorithm. Also added
5313 an -sgckey command line option to the rsa utility. Thanks to
5314 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5315 algorithm to openssl-dev.
5318 *) The evp_local.h macros were using 'c.##kname' which resulted in
5319 invalid expansion on some systems (SCO 5.0.5 for example).
5320 Corrected to 'c.kname'.
5321 [Phillip Porch <root@theporch.com>]
5323 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5324 a STACK of email addresses from a certificate or request, these look
5325 in the subject name and the subject alternative name extensions and
5326 omit any duplicate addresses.
5329 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5330 This makes DSA verification about 2 % faster.
5333 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5334 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5335 plus overhead for 1024 bit moduli).
5336 This makes exponentiations about 0.5 % faster for 1024 bit
5337 exponents (as measured by "openssl speed rsa2048").
5340 *) Rename memory handling macros to avoid conflicts with other
5342 Malloc => OPENSSL_malloc
5343 Malloc_locked => OPENSSL_malloc_locked
5344 Realloc => OPENSSL_realloc
5345 Free => OPENSSL_free
5348 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5349 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5352 *) CygWin32 support.
5353 [John Jarvie <jjarvie@newsguy.com>]
5355 *) The type-safe stack code has been rejigged. It is now only compiled
5356 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5357 by default all type-specific stack functions are "#define"d back to
5358 standard stack functions. This results in more streamlined output
5359 but retains the type-safety checking possibilities of the original
5363 *) The STACK code has been cleaned up, and certain type declarations
5364 that didn't make a lot of sense have been brought in line. This has
5365 also involved a cleanup of sorts in safestack.h to more correctly
5366 map type-safe stack functions onto their plain stack counterparts.
5367 This work has also resulted in a variety of "const"ifications of
5368 lots of the code, especially "_cmp" operations which should normally
5369 be prototyped with "const" parameters anyway.
5372 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5373 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5374 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5375 where all of 'md' is used each time the PRNG is used, but 'state'
5376 is used only indexed by a cyclic counter. As entropy may not be
5377 well distributed from the beginning, 'md' is important as a
5378 chaining variable. However, the output function chains only half
5379 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5380 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5381 in all of 'state' being rewritten, with the new values depending
5382 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5385 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5386 the handshake is continued after ssl_verify_cert_chain();
5387 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5388 can lead to 'unexplainable' connection aborts later.
5389 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5391 *) Major EVP API cipher revision.
5392 Add hooks for extra EVP features. This allows various cipher
5393 parameters to be set in the EVP interface. Support added for variable
5394 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5395 setting of RC2 and RC5 parameters.
5397 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5400 Remove lots of duplicated code from the EVP library. For example *every*
5401 cipher init() function handles the 'iv' in the same way according to the
5402 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5403 for CFB and OFB modes they zero ctx->num.
5405 New functionality allows removal of S/MIME code RC2 hack.
5407 Most of the routines have the same form and so can be declared in terms
5410 By shifting this to the top level EVP_CipherInit() it can be removed from
5411 all individual ciphers. If the cipher wants to handle IVs or keys
5412 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5415 Change lots of functions like EVP_EncryptUpdate() to now return a
5416 value: although software versions of the algorithms cannot fail
5417 any installed hardware versions can.
5420 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5421 this option is set, tolerate broken clients that send the negotiated
5422 protocol version number instead of the requested protocol version
5426 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5427 i.e. non-zero for export ciphersuites, zero otherwise.
5428 Previous versions had this flag inverted, inconsistent with
5429 rsa_tmp_cb (..._TMP_RSA_CB).
5430 [Bodo Moeller; problem reported by Amit Chopra]
5432 *) Add missing DSA library text string. Work around for some IIS
5433 key files with invalid SEQUENCE encoding.
5436 *) Add a document (doc/standards.txt) that list all kinds of standards
5437 and so on that are implemented in OpenSSL.
5440 *) Enhance c_rehash script. Old version would mishandle certificates
5441 with the same subject name hash and wouldn't handle CRLs at all.
5442 Added -fingerprint option to crl utility, to support new c_rehash
5446 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5449 *) Fix for SSL server purpose checking. Server checking was
5450 rejecting certificates which had extended key usage present
5451 but no ssl client purpose.
5452 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5454 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5455 is a little unclear about how a blank password is handled.
5456 Since the password in encoded as a BMPString with terminating
5457 double NULL a zero length password would end up as just the
5458 double NULL. However no password at all is different and is
5459 handled differently in the PKCS#12 key generation code. NS
5460 treats a blank password as zero length. MSIE treats it as no
5461 password on export: but it will try both on import. We now do
5462 the same: PKCS12_parse() tries zero length and no password if
5463 the password is set to "" or NULL (NULL is now a valid password:
5464 it wasn't before) as does the pkcs12 application.
5467 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5468 perror when PEM_read_bio_X509_REQ fails, the error message must
5469 be obtained from the error queue.
5472 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5473 it in ERR_remove_state if appropriate, and change ERR_get_state
5474 accordingly to avoid race conditions (this is necessary because
5475 thread_hash is no longer constant once set).
5478 *) Bugfix for linux-elf makefile.one.
5481 *) RSA_get_default_method() will now cause a default
5482 RSA_METHOD to be chosen if one doesn't exist already.
5483 Previously this was only set during a call to RSA_new()
5484 or RSA_new_method(NULL) meaning it was possible for
5485 RSA_get_default_method() to return NULL.
5488 *) Added native name translation to the existing DSO code
5489 that will convert (if the flag to do so is set) filenames
5490 that are sufficiently small and have no path information
5491 into a canonical native form. Eg. "blah" converted to
5492 "libblah.so" or "blah.dll" etc.
5495 *) New function ERR_error_string_n(e, buf, len) which is like
5496 ERR_error_string(e, buf), but writes at most 'len' bytes
5497 including the 0 terminator. For ERR_error_string_n, 'buf'
5499 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5501 *) CONF library reworked to become more general. A new CONF
5502 configuration file reader "class" is implemented as well as a
5503 new functions (NCONF_*, for "New CONF") to handle it. The now
5504 old CONF_* functions are still there, but are reimplemented to
5505 work in terms of the new functions. Also, a set of functions
5506 to handle the internal storage of the configuration data is
5507 provided to make it easier to write new configuration file
5508 reader "classes" (I can definitely see something reading a
5509 configuration file in XML format, for example), called _CONF_*,
5510 or "the configuration storage API"...
5512 The new configuration file reading functions are:
5514 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5515 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5517 NCONF_default, NCONF_WIN32
5519 NCONF_dump_fp, NCONF_dump_bio
5521 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5522 NCONF_new creates a new CONF object. This works in the same way
5523 as other interfaces in OpenSSL, like the BIO interface.
5524 NCONF_dump_* dump the internal storage of the configuration file,
5525 which is useful for debugging. All other functions take the same
5526 arguments as the old CONF_* functions wth the exception of the
5527 first that must be a `CONF *' instead of a `LHASH *'.
5529 To make it easer to use the new classes with the old CONF_* functions,
5530 the function CONF_set_default_method is provided.
5533 *) Add '-tls1' option to 'openssl ciphers', which was already
5534 mentioned in the documentation but had not been implemented.
5535 (This option is not yet really useful because even the additional
5536 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5539 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5540 OpenSSL-based applications) load shared libraries and bind to
5541 them in a portable way.
5542 [Geoff Thorpe, with contributions from Richard Levitte]
5544 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5546 *) Make sure _lrotl and _lrotr are only used with MSVC.
5548 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5549 (the default implementation of RAND_status).
5551 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5552 to '-clrext' (= clear extensions), as intended and documented.
5553 [Bodo Moeller; inconsistency pointed out by Michael Attili
5556 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5557 was larger than the MD block size.
5558 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5560 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5561 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5562 using the passed key: if the passed key was a private key the result
5563 of X509_print(), for example, would be to print out all the private key
5567 *) des_quad_cksum() byte order bug fix.
5568 [Ulf Möller, using the problem description in krb4-0.9.7, where
5569 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5571 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5573 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5575 *) For easily testing in shell scripts whether some command
5576 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5577 returns with exit code 0 iff no command of the given name is available.
5578 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5579 the output goes to stdout and nothing is printed to stderr.
5580 Additional arguments are always ignored.
5582 Since for each cipher there is a command of the same name,
5583 the 'no-cipher' compilation switches can be tested this way.
5585 ('openssl no-XXX' is not able to detect pseudo-commands such
5586 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5589 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5592 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5593 is set; it will be thrown away anyway because each handshake creates
5595 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5596 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5597 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5598 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5601 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5602 'Q' and 'R' lose their special meanings (quit/renegotiate).
5603 This is part of what -quiet does; unlike -quiet, -ign_eof
5604 does not suppress any output.
5607 *) Add compatibility options to the purpose and trust code. The
5608 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5609 accepts a certificate or CA, this was the previous behaviour,
5610 with all the associated security issues.
5612 X509_TRUST_COMPAT is the old trust behaviour: only and
5613 automatically trust self signed roots in certificate store. A
5614 new trust setting X509_TRUST_DEFAULT is used to specify that
5615 a purpose has no associated trust setting and it should instead
5616 use the value in the default purpose.
5619 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5620 and fix a memory leak.
5623 *) In util/mkerr.pl (which implements 'make errors'), preserve
5624 reason strings from the previous version of the .c file, as
5625 the default to have only downcase letters (and digits) in
5626 automatically generated reasons codes is not always appropriate.
5629 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5630 using strerror. Previously, ERR_reason_error_string() returned
5631 library names as reason strings for SYSerr; but SYSerr is a special
5632 case where small numbers are errno values, not library numbers.
5635 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5636 converts DSA parameters into DH parameters. (When creating parameters,
5637 DSA_generate_parameters is used.)
5640 *) Include 'length' (recommended exponent length) in C code generated
5641 by 'openssl dhparam -C'.
5644 *) The second argument to set_label in perlasm was already being used
5645 so couldn't be used as a "file scope" flag. Moved to third argument
5649 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5650 instead of RAND_bytes for encryption IVs and salts.
5653 *) Include RAND_status() into RAND_METHOD instead of implementing
5654 it only for md_rand.c Otherwise replacing the PRNG by calling
5655 RAND_set_rand_method would be impossible.
5658 *) Don't let DSA_generate_key() enter an infinite loop if the random
5659 number generation fails.
5662 *) New 'rand' application for creating pseudo-random output.
5665 *) Added configuration support for Linux/IA64
5666 [Rolf Haberrecker <rolf@suse.de>]
5668 *) Assembler module support for Mingw32.
5671 *) Shared library support for HPUX (in shlib/).
5672 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5674 *) Shared library support for Solaris gcc.
5675 [Lutz Behnke <behnke@trustcenter.de>]
5677 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5679 *) PKCS7_encrypt() was adding text MIME headers twice because they
5680 were added manually and by SMIME_crlf_copy().
5683 *) In bntest.c don't call BN_rand with zero bits argument.
5684 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5686 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5687 case was implemented. This caused BN_div_recp() to fail occasionally.
5690 *) Add an optional second argument to the set_label() in the perl
5691 assembly language builder. If this argument exists and is set
5692 to 1 it signals that the assembler should use a symbol whose
5693 scope is the entire file, not just the current function. This
5694 is needed with MASM which uses the format label:: for this scope.
5695 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5697 *) Change the ASN1 types so they are typedefs by default. Before
5698 almost all types were #define'd to ASN1_STRING which was causing
5699 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5703 *) Change names of new functions to the new get1/get0 naming
5704 convention: After 'get1', the caller owns a reference count
5705 and has to call ..._free; 'get0' returns a pointer to some
5706 data structure without incrementing reference counters.
5707 (Some of the existing 'get' functions increment a reference
5708 counter, some don't.)
5709 Similarly, 'set1' and 'add1' functions increase reference
5710 counters or duplicate objects.
5713 *) Allow for the possibility of temp RSA key generation failure:
5714 the code used to assume it always worked and crashed on failure.
5717 *) Fix potential buffer overrun problem in BIO_printf().
5718 [Ulf Möller, using public domain code by Patrick Powell; problem
5719 pointed out by David Sacerdote <das33@cornell.edu>]
5721 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5722 RAND_egd() and RAND_status(). In the command line application,
5723 the EGD socket can be specified like a seed file using RANDFILE
5727 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5728 Some CAs (e.g. Verisign) distribute certificates in this form.
5731 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5732 list to exclude them. This means that no special compilation option
5733 is needed to use anonymous DH: it just needs to be included in the
5737 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5738 EVP_MD_type. The old functionality is available in a new macro called
5739 EVP_MD_md(). Change code that uses it and update docs.
5742 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5743 where the 'void *' argument is replaced by a function pointer argument.
5744 Previously 'void *' was abused to point to functions, which works on
5745 many platforms, but is not correct. As these functions are usually
5746 called by macros defined in OpenSSL header files, most source code
5747 should work without changes.
5750 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5751 sections with information on -D... compiler switches used for
5752 compiling the library so that applications can see them. To enable
5753 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5754 must be defined. E.g.,
5755 #define OPENSSL_ALGORITHM_DEFINES
5756 #include <openssl/opensslconf.h>
5757 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5758 [Richard Levitte, Ulf and Bodo Möller]
5760 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5764 *) Change the 'other' type in certificate aux info to a STACK_OF
5765 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5766 the required ASN1 format: arbitrary types determined by an OID.
5769 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5770 argument to 'req'. This is not because the function is newer or
5771 better than others it just uses the work 'NEW' in the certificate
5772 request header lines. Some software needs this.
5775 *) Reorganise password command line arguments: now passwords can be
5776 obtained from various sources. Delete the PEM_cb function and make
5777 it the default behaviour: i.e. if the callback is NULL and the
5778 usrdata argument is not NULL interpret it as a null terminated pass
5779 phrase. If usrdata and the callback are NULL then the pass phrase
5780 is prompted for as usual.
5783 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5784 the support is automatically enabled. The resulting binaries will
5785 autodetect the card and use it if present.
5786 [Ben Laurie and Compaq Inc.]
5788 *) Work around for Netscape hang bug. This sends certificate request
5789 and server done in one record. Since this is perfectly legal in the
5790 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5791 the bugs/SSLv3 entry for more info.
5794 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5797 *) Add -rand argument to smime and pkcs12 applications and read/write
5801 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5804 *) Add command line password options to the remaining applications.
5807 *) Bug fix for BN_div_recp() for numerators with an even number of
5811 *) More tests in bntest.c, and changed test_bn output.
5814 *) ./config recognizes MacOS X now.
5817 *) Bug fix for BN_div() when the first words of num and divsor are
5818 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5821 *) Add support for various broken PKCS#8 formats, and command line
5822 options to produce them.
5825 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5826 get temporary BIGNUMs from a BN_CTX.
5829 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5833 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5834 include a #define from the old name to the new. The original intent
5835 was that statically linked binaries could for example just call
5836 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5837 link with digests. This never worked becayse SSLeay_add_all_digests()
5838 and SSLeay_add_all_ciphers() were in the same source file so calling
5839 one would link with the other. They are now in separate source files.
5842 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5845 *) Use a less unusual form of the Miller-Rabin primality test (it used
5846 a binary algorithm for exponentiation integrated into the Miller-Rabin
5847 loop, our standard modexp algorithms are faster).
5850 *) Support for the EBCDIC character set completed.
5851 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5853 *) Source code cleanups: use const where appropriate, eliminate casts,
5854 use void * instead of char * in lhash.
5857 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5858 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5859 this the server could overwrite ephemeral keys that the client
5863 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5864 using 50 iterations of the Rabin-Miller test.
5866 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5867 iterations of the Rabin-Miller test as required by the appendix
5868 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5869 As BN_is_prime_fasttest includes trial division, DSA parameter
5870 generation becomes much faster.
5872 This implies a change for the callback functions in DSA_is_prime
5873 and DSA_generate_parameters: The callback function is called once
5874 for each positive witness in the Rabin-Miller test, not just
5875 occasionally in the inner loop; and the parameters to the
5876 callback function now provide an iteration count for the outer
5877 loop rather than for the current invocation of the inner loop.
5878 DSA_generate_parameters additionally can call the callback
5879 function with an 'iteration count' of -1, meaning that a
5880 candidate has passed the trial division test (when q is generated
5881 from an application-provided seed, trial division is skipped).
5884 *) New function BN_is_prime_fasttest that optionally does trial
5885 division before starting the Rabin-Miller test and has
5886 an additional BN_CTX * argument (whereas BN_is_prime always
5887 has to allocate at least one BN_CTX).
5888 'callback(1, -1, cb_arg)' is called when a number has passed the
5889 trial division stage.
5892 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5896 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5899 *) New function BN_pseudo_rand().
5902 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5903 bignum version of BN_from_montgomery() with the working code from
5904 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5908 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5909 made it impossible to use the same SSL_SESSION data structure in
5910 SSL2 clients in multiple threads.
5913 *) The return value of RAND_load_file() no longer counts bytes obtained
5914 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5915 to seed the PRNG (previously an explicit byte count was required).
5916 [Ulf Möller, Bodo Möller]
5918 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5919 used (char *) instead of (void *) and had casts all over the place.
5922 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5925 *) Retain source code compatibility for BN_prime_checks macro:
5926 BN_is_prime(..., BN_prime_checks, ...) now uses
5927 BN_prime_checks_for_size to determine the appropriate number of
5928 Rabin-Miller iterations.
5931 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5932 DH_CHECK_P_NOT_SAFE_PRIME.
5933 (Check if this is true? OpenPGP calls them "strong".)
5936 *) Merge the functionality of "dh" and "gendh" programs into a new program
5937 "dhparam". The old programs are retained for now but will handle DH keys
5938 (instead of parameters) in future.
5941 *) Make the ciphers, s_server and s_client programs check the return values
5942 when a new cipher list is set.
5945 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5946 ciphers. Before when the 56bit ciphers were enabled the sorting was
5949 The syntax for the cipher sorting has been extended to support sorting by
5950 cipher-strength (using the strength_bits hard coded in the tables).
5951 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5953 Fix a bug in the cipher-command parser: when supplying a cipher command
5954 string with an "undefined" symbol (neither command nor alphanumeric
5955 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5956 an error is flagged.
5958 Due to the strength-sorting extension, the code of the
5959 ssl_create_cipher_list() function was completely rearranged. I hope that
5960 the readability was also increased :-)
5961 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5963 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5964 for the first serial number and places 2 in the serial number file. This
5965 avoids problems when the root CA is created with serial number zero and
5966 the first user certificate has the same issuer name and serial number
5970 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5971 the new code. Add documentation for this stuff.
5974 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5975 X509_*() to X509at_*() on the grounds that they don't handle X509
5976 structures and behave in an analagous way to the X509v3 functions:
5977 they shouldn't be called directly but wrapper functions should be used
5980 So we also now have some wrapper functions that call the X509at functions
5981 when passed certificate requests. (TO DO: similar things can be done with
5982 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5983 things. Some of these need some d2i or i2d and print functionality
5984 because they handle more complex structures.)
5987 *) Add missing #ifndefs that caused missing symbols when building libssl
5988 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5989 NO_RSA in ssl/s2*.c.
5990 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5992 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5993 has a return value which indicates the quality of the random data
5994 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5995 error queue. New function RAND_pseudo_bytes() generates output that is
5996 guaranteed to be unique but not unpredictable. RAND_add is like
5997 RAND_seed, but takes an extra argument for an entropy estimate
5998 (RAND_seed always assumes full entropy).
6001 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6002 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6003 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6004 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6005 false-positive rate of at most 2^-80 for random input.
6008 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6011 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6012 in the 0.9.5 release), this returns the chain
6013 from an X509_CTX structure with a dup of the stack and all
6014 the X509 reference counts upped: so the stack will exist
6015 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6018 Also make SSL_SESSION_print() print out the verify return
6022 *) Add manpage for the pkcs12 command. Also change the default
6023 behaviour so MAC iteration counts are used unless the new
6024 -nomaciter option is used. This improves file security and
6025 only older versions of MSIE (4.0 for example) need it.
6028 *) Honor the no-xxx Configure options when creating .DEF files.
6031 *) Add PKCS#10 attributes to field table: challengePassword,
6032 unstructuredName and unstructuredAddress. These are taken from
6033 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6034 international characters are used.
6036 More changes to X509_ATTRIBUTE code: allow the setting of types
6037 based on strings. Remove the 'loc' parameter when adding
6038 attributes because these will be a SET OF encoding which is sorted
6042 *) Initial changes to the 'req' utility to allow request generation
6043 automation. This will allow an application to just generate a template
6044 file containing all the field values and have req construct the
6047 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6048 used all over the place including certificate requests and PKCS#7
6049 structures. They are currently handled manually where necessary with
6050 some primitive wrappers for PKCS#7. The new functions behave in a
6051 manner analogous to the X509 extension functions: they allow
6052 attributes to be looked up by NID and added.
6054 Later something similar to the X509V3 code would be desirable to
6055 automatically handle the encoding, decoding and printing of the
6056 more complex types. The string types like challengePassword can
6057 be handled by the string table functions.
6059 Also modified the multi byte string table handling. Now there is
6060 a 'global mask' which masks out certain types. The table itself
6061 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6062 is useful when for example there is only one permissible type
6063 (as in countryName) and using the mask might result in no valid
6067 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6068 SSL_get_peer_finished to allow applications to obtain the latest
6069 Finished messages sent to the peer or expected from the peer,
6070 respectively. (SSL_get_peer_finished is usually the Finished message
6071 actually received from the peer, otherwise the protocol will be aborted.)
6073 As the Finished message are message digests of the complete handshake
6074 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6075 be used for external authentication procedures when the authentication
6076 provided by SSL/TLS is not desired or is not enough.
6079 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6080 the host supports BWX extension and if Compaq C is present on the
6081 $PATH. Just exploiting of the BWX extension results in 20-30%
6082 performance kick for some algorithms, e.g. DES and RC4 to mention
6083 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6087 *) Add support for MS "fast SGC". This is arguably a violation of the
6088 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6089 weak crypto and after checking the certificate is SGC a second one
6090 with strong crypto. MS SGC stops the first handshake after receiving
6091 the server certificate message and sends a second client hello. Since
6092 a server will typically do all the time consuming operations before
6093 expecting any further messages from the client (server key exchange
6094 is the most expensive) there is little difference between the two.
6096 To get OpenSSL to support MS SGC we have to permit a second client
6097 hello message after we have sent server done. In addition we have to
6098 reset the MAC if we do get this second client hello.
6101 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6102 if a DER encoded private key is RSA or DSA traditional format. Changed
6103 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6104 format DER encoded private key. Newer code should use PKCS#8 format which
6105 has the key type encoded in the ASN1 structure. Added DER private key
6106 support to pkcs8 application.
6109 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6110 ciphersuites has been selected (as required by the SSL 3/TLS 1
6111 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6112 is set, we interpret this as a request to violate the specification
6113 (the worst that can happen is a handshake failure, and 'correct'
6114 behaviour would result in a handshake failure anyway).
6117 *) In SSL_CTX_add_session, take into account that there might be multiple
6118 SSL_SESSION structures with the same session ID (e.g. when two threads
6119 concurrently obtain them from an external cache).
6120 The internal cache can handle only one SSL_SESSION with a given ID,
6121 so if there's a conflict, we now throw out the old one to achieve
6125 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6126 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6127 some routines that use cipher OIDs: some ciphers do not have OIDs
6128 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6132 *) Simplify the trust setting structure and code. Now we just have
6133 two sequences of OIDs for trusted and rejected settings. These will
6134 typically have values the same as the extended key usage extension
6135 and any application specific purposes.
6137 The trust checking code now has a default behaviour: it will just
6138 check for an object with the same NID as the passed id. Functions can
6139 be provided to override either the default behaviour or the behaviour
6140 for a given id. SSL client, server and email already have functions
6141 in place for compatibility: they check the NID and also return "trusted"
6142 if the certificate is self signed.
6145 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6146 traditional format into an EVP_PKEY structure.
6149 *) Add a password callback function PEM_cb() which either prompts for
6150 a password if usr_data is NULL or otherwise assumes it is a null
6151 terminated password. Allow passwords to be passed on command line
6152 environment or config files in a few more utilities.
6155 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6156 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6157 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6158 Update documentation.
6161 *) Support for ASN1 "NULL" type. This could be handled before by using
6162 ASN1_TYPE but there wasn't any function that would try to read a NULL
6163 and produce an error if it couldn't. For compatibility we also have
6164 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6165 don't allocate anything because they don't need to.
6168 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6170 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6172 *) Rebuild of the memory allocation routines used by OpenSSL code and
6173 possibly others as well. The purpose is to make an interface that
6174 provide hooks so anyone can build a separate set of allocation and
6175 deallocation routines to be used by OpenSSL, for example memory
6176 pool implementations, or something else, which was previously hard
6177 since Malloc(), Realloc() and Free() were defined as macros having
6178 the values malloc, realloc and free, respectively (except for Win32
6179 compilations). The same is provided for memory debugging code.
6180 OpenSSL already comes with functionality to find memory leaks, but
6181 this gives people a chance to debug other memory problems.
6183 With these changes, a new set of functions and macros have appeared:
6185 CRYPTO_set_mem_debug_functions() [F]
6186 CRYPTO_get_mem_debug_functions() [F]
6187 CRYPTO_dbg_set_options() [F]
6188 CRYPTO_dbg_get_options() [F]
6189 CRYPTO_malloc_debug_init() [M]
6191 The memory debug functions are NULL by default, unless the library
6192 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6193 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6194 gives the standard debugging functions that come with OpenSSL) or
6195 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6196 provided by the library user) must be used. When the standard
6197 debugging functions are used, CRYPTO_dbg_set_options can be used to
6198 request additional information:
6199 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6200 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6202 Also, things like CRYPTO_set_mem_functions will always give the
6203 expected result (the new set of functions is used for allocation
6204 and deallocation) at all times, regardless of platform and compiler
6207 To finish it up, some functions that were never use in any other
6208 way than through macros have a new API and new semantic:
6211 CRYPTO_dbg_realloc()
6214 All macros of value have retained their old syntax.
6215 [Richard Levitte and Bodo Moeller]
6217 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6218 ordering of SMIMECapabilities wasn't in "strength order" and there
6219 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6223 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6224 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6225 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6227 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6228 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6229 functionality to handle multipart/signed properly) and a utility
6230 called 'smime' to call all this stuff. This is based on code I
6231 originally wrote for Celo who have kindly allowed it to be
6232 included in OpenSSL.
6235 *) Add variants des_set_key_checked and des_set_key_unchecked of
6236 des_set_key (aka des_key_sched). Global variable des_check_key
6237 decides which of these is called by des_set_key; this way
6238 des_check_key behaves as it always did, but applications and
6239 the library itself, which was buggy for des_check_key == 1,
6240 have a cleaner way to pick the version they need.
6243 *) New function PKCS12_newpass() which changes the password of a
6247 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6248 dynamic mix. In both cases the ids can be used as an index into the
6249 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6250 functions so they accept a list of the field values and the
6251 application doesn't need to directly manipulate the X509_TRUST
6255 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6259 *) Modify the way the V3 extension code looks up extensions. This now
6260 works in a similar way to the object code: we have some "standard"
6261 extensions in a static table which is searched with OBJ_bsearch()
6262 and the application can add dynamic ones if needed. The file
6263 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6264 updated whenever a new extension is added to the core code and kept
6265 in ext_nid order. There is a simple program 'tabtest.c' which checks
6266 this. New extensions are not added too often so this file can readily
6267 be maintained manually.
6269 There are two big advantages in doing things this way. The extensions
6270 can be looked up immediately and no longer need to be "added" using
6271 X509V3_add_standard_extensions(): this function now does nothing.
6272 [Side note: I get *lots* of email saying the extension code doesn't
6273 work because people forget to call this function]
6274 Also no dynamic allocation is done unless new extensions are added:
6275 so if we don't add custom extensions there is no need to call
6276 X509V3_EXT_cleanup().
6279 *) Modify enc utility's salting as follows: make salting the default. Add a
6280 magic header, so unsalted files fail gracefully instead of just decrypting
6281 to garbage. This is because not salting is a big security hole, so people
6282 should be discouraged from doing it.
6285 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6286 digest to be passed on the command line but it only used this
6287 parameter when signing a certificate. Modified so all relevant
6288 operations are affected by the digest parameter including the
6289 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6290 DSA key was used because it didn't fix the digest.
6293 *) Initial certificate chain verify code. Currently tests the untrusted
6294 certificates for consistency with the verify purpose (which is set
6295 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6297 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6298 this is because it will reject chains with invalid extensions whereas
6299 every previous version of OpenSSL and SSLeay made no checks at all.
6301 Trust code: checks the root CA for the relevant trust settings. Trust
6302 settings have an initial value consistent with the verify purpose: e.g.
6303 if the verify purpose is for SSL client use it expects the CA to be
6304 trusted for SSL client use. However the default value can be changed to
6305 permit custom trust settings: one example of this would be to only trust
6306 certificates from a specific "secure" set of CAs.
6308 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6309 which should be used for version portability: especially since the
6310 verify structure is likely to change more often now.
6312 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6313 to set them. If not set then assume SSL clients will verify SSL servers
6316 Two new options to the verify program: -untrusted allows a set of
6317 untrusted certificates to be passed in and -purpose which sets the
6318 intended purpose of the certificate. If a purpose is set then the
6319 new chain verify code is used to check extension consistency.
6322 *) Support for the authority information access extension.
6325 *) Modify RSA and DSA PEM read routines to transparently handle
6326 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6327 public keys in a format compatible with certificate
6328 SubjectPublicKeyInfo structures. Unfortunately there were already
6329 functions called *_PublicKey_* which used various odd formats so
6330 these are retained for compatibility: however the DSA variants were
6331 never in a public release so they have been deleted. Changed dsa/rsa
6332 utilities to handle the new format: note no releases ever handled public
6333 keys so we should be OK.
6335 The primary motivation for this change is to avoid the same fiasco
6336 that dogs private keys: there are several incompatible private key
6337 formats some of which are standard and some OpenSSL specific and
6338 require various evil hacks to allow partial transparent handling and
6339 even then it doesn't work with DER formats. Given the option anything
6340 other than PKCS#8 should be dumped: but the other formats have to
6341 stay in the name of compatibility.
6343 With public keys and the benefit of hindsight one standard format
6344 is used which works with EVP_PKEY, RSA or DSA structures: though
6345 it clearly returns an error if you try to read the wrong kind of key.
6347 Added a -pubkey option to the 'x509' utility to output the public key.
6348 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6349 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6350 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6351 that do the same as the EVP_PKEY_assign_*() except they up the
6352 reference count of the added key (they don't "swallow" the
6356 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6357 CRLs would fail if the file contained no certificates or no CRLs:
6358 added a new function to read in both types and return the number
6359 read: this means that if none are read it will be an error. The
6360 DER versions of the certificate and CRL reader would always fail
6361 because it isn't possible to mix certificates and CRLs in DER format
6362 without choking one or the other routine. Changed this to just read
6363 a certificate: this is the best we can do. Also modified the code
6364 in apps/verify.c to take notice of return codes: it was previously
6365 attempting to read in certificates from NULL pointers and ignoring
6366 any errors: this is one reason why the cert and CRL reader seemed
6367 to work. It doesn't check return codes from the default certificate
6368 routines: these may well fail if the certificates aren't installed.
6371 *) Code to support otherName option in GeneralName.
6374 *) First update to verify code. Change the verify utility
6375 so it warns if it is passed a self signed certificate:
6376 for consistency with the normal behaviour. X509_verify
6377 has been modified to it will now verify a self signed
6378 certificate if *exactly* the same certificate appears
6379 in the store: it was previously impossible to trust a
6380 single self signed certificate. This means that:
6381 openssl verify ss.pem
6382 now gives a warning about a self signed certificate but
6383 openssl verify -CAfile ss.pem ss.pem
6387 *) For servers, store verify_result in SSL_SESSION data structure
6388 (and add it to external session representation).
6389 This is needed when client certificate verifications fails,
6390 but an application-provided verification callback (set by
6391 SSL_CTX_set_cert_verify_callback) allows accepting the session
6392 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6393 but returns 1): When the session is reused, we have to set
6394 ssl->verify_result to the appropriate error code to avoid
6396 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6398 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6399 case in PKCS7_dataInit() where the signed PKCS7 structure
6400 didn't contain any existing data because it was being created.
6401 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6403 *) Add a salt to the key derivation routines in enc.c. This
6404 forms the first 8 bytes of the encrypted file. Also add a
6405 -S option to allow a salt to be input on the command line.
6408 *) New function X509_cmp(). Oddly enough there wasn't a function
6409 to compare two certificates. We do this by working out the SHA1
6410 hash and comparing that. X509_cmp() will be needed by the trust
6414 *) SSL_get1_session() is like SSL_get_session(), but increments
6415 the reference count in the SSL_SESSION returned.
6416 [Geoff Thorpe <geoff@eu.c2.net>]
6418 *) Fix for 'req': it was adding a null to request attributes.
6419 Also change the X509_LOOKUP and X509_INFO code to handle
6420 certificate auxiliary information.
6423 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6427 *) Add the possibility to add extra information to the memory leak
6428 detecting output, to form tracebacks, showing from where each
6429 allocation was originated: CRYPTO_push_info("constant string") adds
6430 the string plus current file name and line number to a per-thread
6431 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6432 is like calling CYRPTO_pop_info() until the stack is empty.
6433 Also updated memory leak detection code to be multi-thread-safe.
6436 *) Add options -text and -noout to pkcs7 utility and delete the
6437 encryption options which never did anything. Update docs.
6440 *) Add options to some of the utilities to allow the pass phrase
6441 to be included on either the command line (not recommended on
6442 OSes like Unix) or read from the environment. Update the
6443 manpages and fix a few bugs.
6446 *) Add a few manpages for some of the openssl commands.
6449 *) Fix the -revoke option in ca. It was freeing up memory twice,
6450 leaking and not finding already revoked certificates.
6453 *) Extensive changes to support certificate auxiliary information.
6454 This involves the use of X509_CERT_AUX structure and X509_AUX
6455 functions. An X509_AUX function such as PEM_read_X509_AUX()
6456 can still read in a certificate file in the usual way but it
6457 will also read in any additional "auxiliary information". By
6458 doing things this way a fair degree of compatibility can be
6459 retained: existing certificates can have this information added
6460 using the new 'x509' options.
6462 Current auxiliary information includes an "alias" and some trust
6463 settings. The trust settings will ultimately be used in enhanced
6464 certificate chain verification routines: currently a certificate
6465 can only be trusted if it is self signed and then it is trusted
6469 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6470 The problem was that one of the replacement routines had not been working
6471 since SSLeay releases. For now the offending routine has been replaced
6472 with non-optimised assembler. Even so, this now gives around 95%
6473 performance improvement for 1024 bit RSA signs.
6476 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6477 handling. Most clients have the effective key size in bits equal to
6478 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6479 A few however don't do this and instead use the size of the decrypted key
6480 to determine the RC2 key length and the AlgorithmIdentifier to determine
6481 the effective key length. In this case the effective key length can still
6482 be 40 bits but the key length can be 168 bits for example. This is fixed
6483 by manually forcing an RC2 key into the EVP_PKEY structure because the
6484 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6485 the key length and effective key length are equal.
6488 *) Add a bunch of functions that should simplify the creation of
6489 X509_NAME structures. Now you should be able to do:
6490 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6491 and have it automatically work out the correct field type and fill in
6492 the structures. The more adventurous can try:
6493 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6494 and it will (hopefully) work out the correct multibyte encoding.
6497 *) Change the 'req' utility to use the new field handling and multibyte
6498 copy routines. Before the DN field creation was handled in an ad hoc
6499 way in req, ca, and x509 which was rather broken and didn't support
6500 BMPStrings or UTF8Strings. Since some software doesn't implement
6501 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6502 using the dirstring_type option. See the new comment in the default
6503 openssl.cnf for more info.
6506 *) Make crypto/rand/md_rand.c more robust:
6507 - Assure unique random numbers after fork().
6508 - Make sure that concurrent threads access the global counter and
6509 md serializably so that we never lose entropy in them
6510 or use exactly the same state in multiple threads.
6511 Access to the large state is not always serializable because
6512 the additional locking could be a performance killer, and
6513 md should be large enough anyway.
6516 *) New file apps/app_rand.c with commonly needed functionality
6517 for handling the random seed file.
6519 Use the random seed file in some applications that previously did not:
6521 dsaparam -genkey (which also ignored its '-rand' option),
6524 x509 (when signing).
6525 Except on systems with /dev/urandom, it is crucial to have a random
6526 seed file at least for key creation, DSA signing, and for DH exchanges;
6527 for RSA signatures we could do without one.
6529 gendh and gendsa (unlike genrsa) used to read only the first byte
6530 of each file listed in the '-rand' option. The function as previously
6531 found in genrsa is now in app_rand.c and is used by all programs
6532 that support '-rand'.
6535 *) In RAND_write_file, use mode 0600 for creating files;
6536 don't just chmod when it may be too late.
6539 *) Report an error from X509_STORE_load_locations
6540 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6543 *) New function ASN1_mbstring_copy() this copies a string in either
6544 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6545 into an ASN1_STRING type. A mask of permissible types is passed
6546 and it chooses the "minimal" type to use or an error if not type
6550 *) Add function equivalents to the various macros in asn1.h. The old
6551 macros are retained with an M_ prefix. Code inside the library can
6552 use the M_ macros. External code (including the openssl utility)
6553 should *NOT* in order to be "shared library friendly".
6556 *) Add various functions that can check a certificate's extensions
6557 to see if it usable for various purposes such as SSL client,
6558 server or S/MIME and CAs of these types. This is currently
6559 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6560 verification. Also added a -purpose flag to x509 utility to
6561 print out all the purposes.
6564 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6568 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6569 for, obtain and decode and extension and obtain its critical flag.
6570 This allows all the necessary extension code to be handled in a
6571 single function call.
6574 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6575 platforms. See crypto/rc4/rc4_enc.c for further details.
6578 *) New -noout option to asn1parse. This causes no output to be produced
6579 its main use is when combined with -strparse and -out to extract data
6580 from a file (which may not be in ASN.1 format).
6583 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6584 when producing the local key id.
6585 [Richard Levitte <levitte@stacken.kth.se>]
6587 *) New option -dhparam in s_server. This allows a DH parameter file to be
6588 stated explicitly. If it is not stated then it tries the first server
6589 certificate file. The previous behaviour hard coded the filename
6593 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6594 a public key to be input or output. For example:
6595 openssl rsa -in key.pem -pubout -out pubkey.pem
6596 Also added necessary DSA public key functions to handle this.
6599 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6600 in the message. This was handled by allowing
6601 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6602 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6604 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6605 to the end of the strings whereas this didn't. This would cause problems
6606 if strings read with d2i_ASN1_bytes() were later modified.
6607 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6609 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6610 data and it contains EOF it will end up returning an error. This is
6611 caused by input 46 bytes long. The cause is due to the way base64
6612 BIOs find the start of base64 encoded data. They do this by trying a
6613 trial decode on each line until they find one that works. When they
6614 do a flag is set and it starts again knowing it can pass all the
6615 data directly through the decoder. Unfortunately it doesn't reset
6616 the context it uses. This means that if EOF is reached an attempt
6617 is made to pass two EOFs through the context and this causes the
6618 resulting error. This can also cause other problems as well. As is
6619 usual with these problems it takes *ages* to find and the fix is
6620 trivial: move one line.
6621 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6623 *) Ugly workaround to get s_client and s_server working under Windows. The
6624 old code wouldn't work because it needed to select() on sockets and the
6625 tty (for keypresses and to see if data could be written). Win32 only
6626 supports select() on sockets so we select() with a 1s timeout on the
6627 sockets and then see if any characters are waiting to be read, if none
6628 are present then we retry, we also assume we can always write data to
6629 the tty. This isn't nice because the code then blocks until we've
6630 received a complete line of data and it is effectively polling the
6631 keyboard at 1s intervals: however it's quite a bit better than not
6632 working at all :-) A dedicated Windows application might handle this
6633 with an event loop for example.
6636 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6637 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6638 will be called when RSA_sign() and RSA_verify() are used. This is useful
6639 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6640 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6641 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6642 This necessitated the support of an extra signature type NID_md5_sha1
6643 for SSL signatures and modifications to the SSL library to use it instead
6644 of calling RSA_public_decrypt() and RSA_private_encrypt().
6647 *) Add new -verify -CAfile and -CApath options to the crl program, these
6648 will lookup a CRL issuers certificate and verify the signature in a
6649 similar way to the verify program. Tidy up the crl program so it
6650 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6651 less strict. It will now permit CRL extensions even if it is not
6652 a V2 CRL: this will allow it to tolerate some broken CRLs.
6655 *) Initialize all non-automatic variables each time one of the openssl
6656 sub-programs is started (this is necessary as they may be started
6657 multiple times from the "OpenSSL>" prompt).
6658 [Lennart Bang, Bodo Moeller]
6660 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6661 removing all other RSA functionality (this is what NO_RSA does). This
6662 is so (for example) those in the US can disable those operations covered
6663 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6667 *) Non-copying interface to BIO pairs.
6668 (still largely untested)
6671 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6672 ASCII string. This was handled independently in various places before.
6675 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6676 UTF8 strings a character at a time.
6679 *) Use client_version from client hello to select the protocol
6680 (s23_srvr.c) and for RSA client key exchange verification
6681 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6684 *) Add various utility functions to handle SPKACs, these were previously
6685 handled by poking round in the structure internals. Added new function
6686 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6687 print, verify and generate SPKACs. Based on an original idea from
6688 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6691 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6694 *) Allow the config file extension section to be overwritten on the
6695 command line. Based on an original idea from Massimiliano Pala
6696 <madwolf@comune.modena.it>. The new option is called -extensions
6697 and can be applied to ca, req and x509. Also -reqexts to override
6698 the request extensions in req and -crlexts to override the crl extensions
6702 *) Add new feature to the SPKAC handling in ca. Now you can include
6703 the same field multiple times by preceding it by "XXXX." for example:
6706 this is the same syntax as used in the req config file.
6709 *) Allow certificate extensions to be added to certificate requests. These
6710 are specified in a 'req_extensions' option of the req section of the
6711 config file. They can be printed out with the -text option to req but
6712 are otherwise ignored at present.
6715 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6716 data read consists of only the final block it would not decrypted because
6717 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6718 A misplaced 'break' also meant the decrypted final block might not be
6719 copied until the next read.
6722 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6723 a few extra parameters to the DH structure: these will be useful if
6724 for example we want the value of 'q' or implement X9.42 DH.
6727 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6728 provides hooks that allow the default DSA functions or functions on a
6729 "per key" basis to be replaced. This allows hardware acceleration and
6730 hardware key storage to be handled without major modification to the
6731 library. Also added low level modexp hooks and CRYPTO_EX structure and
6732 associated functions.
6735 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6736 as "read only": it can't be written to and the buffer it points to will
6737 not be freed. Reading from a read only BIO is much more efficient than
6738 a normal memory BIO. This was added because there are several times when
6739 an area of memory needs to be read from a BIO. The previous method was
6740 to create a memory BIO and write the data to it, this results in two
6741 copies of the data and an O(n^2) reading algorithm. There is a new
6742 function BIO_new_mem_buf() which creates a read only memory BIO from
6743 an area of memory. Also modified the PKCS#7 routines to use read only
6747 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6748 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6749 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6750 but a retry condition occured while trying to read the rest.
6753 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6754 NID_pkcs7_encrypted by default: this was wrong since this should almost
6755 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6756 the encrypted data type: this is a more sensible place to put it and it
6757 allows the PKCS#12 code to be tidied up that duplicated this
6761 *) Changed obj_dat.pl script so it takes its input and output files on
6762 the command line. This should avoid shell escape redirection problems
6766 *) Initial support for certificate extension requests, these are included
6767 in things like Xenroll certificate requests. Included functions to allow
6768 extensions to be obtained and added.
6771 *) -crlf option to s_client and s_server for sending newlines as
6772 CRLF (as required by many protocols).
6775 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6777 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6778 [Ralf S. Engelschall]
6780 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6781 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6783 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6787 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6788 DH parameters/keys (q is lost during that conversion, but the resulting
6789 DH parameters contain its length).
6791 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6792 much faster than DH_generate_parameters (which creates parameters
6793 where p = 2*q + 1), and also the smaller q makes DH computations
6794 much more efficient (160-bit exponentiation instead of 1024-bit
6795 exponentiation); so this provides a convenient way to support DHE
6796 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6797 utter importance to use
6798 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6800 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6801 when such DH parameters are used, because otherwise small subgroup
6802 attacks may become possible!
6805 *) Avoid memory leak in i2d_DHparams.
6808 *) Allow the -k option to be used more than once in the enc program:
6809 this allows the same encrypted message to be read by multiple recipients.
6812 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6813 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6814 it will always use the numerical form of the OID, even if it has a short
6818 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6819 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6820 otherwise bn_mod_exp was called. In the case of hardware keys for example
6821 no private key components need be present and it might store extra data
6822 in the RSA structure, which cannot be accessed from bn_mod_exp.
6823 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6824 private key operations.
6827 *) Added support for SPARC Linux.
6830 *) pem_password_cb function type incompatibly changed from
6831 typedef int pem_password_cb(char *buf, int size, int rwflag);
6833 ....(char *buf, int size, int rwflag, void *userdata);
6834 so that applications can pass data to their callbacks:
6835 The PEM[_ASN1]_{read,write}... functions and macros now take an
6836 additional void * argument, which is just handed through whenever
6837 the password callback is called.
6838 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6840 New function SSL_CTX_set_default_passwd_cb_userdata.
6842 Compatibility note: As many C implementations push function arguments
6843 onto the stack in reverse order, the new library version is likely to
6844 interoperate with programs that have been compiled with the old
6845 pem_password_cb definition (PEM_whatever takes some data that
6846 happens to be on the stack as its last argument, and the callback
6847 just ignores this garbage); but there is no guarantee whatsoever that
6850 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6851 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6852 problems not only on Windows, but also on some Unix platforms.
6853 To avoid problematic command lines, these definitions are now in an
6854 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6855 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6858 *) MIPS III/IV assembler module is reimplemented.
6861 *) More DES library cleanups: remove references to srand/rand and
6862 delete an unused file.
6865 *) Add support for the the free Netwide assembler (NASM) under Win32,
6866 since not many people have MASM (ml) and it can be hard to obtain.
6867 This is currently experimental but it seems to work OK and pass all
6868 the tests. Check out INSTALL.W32 for info.
6871 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6872 without temporary keys kept an extra copy of the server key,
6873 and connections with temporary keys did not free everything in case
6877 *) New function RSA_check_key and new openssl rsa option -check
6878 for verifying the consistency of RSA keys.
6879 [Ulf Moeller, Bodo Moeller]
6881 *) Various changes to make Win32 compile work:
6882 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6883 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6884 comparison" warnings.
6885 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6888 *) Add a debugging option to PKCS#5 v2 key generation function: when
6889 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6890 derived keys are printed to stderr.
6893 *) Copy the flags in ASN1_STRING_dup().
6894 [Roman E. Pavlov <pre@mo.msk.ru>]
6896 *) The x509 application mishandled signing requests containing DSA
6897 keys when the signing key was also DSA and the parameters didn't match.
6899 It was supposed to omit the parameters when they matched the signing key:
6900 the verifying software was then supposed to automatically use the CA's
6901 parameters if they were absent from the end user certificate.
6903 Omitting parameters is no longer recommended. The test was also
6904 the wrong way round! This was probably due to unusual behaviour in
6905 EVP_cmp_parameters() which returns 1 if the parameters match.
6906 This meant that parameters were omitted when they *didn't* match and
6907 the certificate was useless. Certificates signed with 'ca' didn't have
6909 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6911 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6912 The interface is as follows:
6913 Applications can use
6914 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6915 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6916 "off" is now the default.
6917 The library internally uses
6918 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6919 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6920 to disable memory-checking temporarily.
6922 Some inconsistent states that previously were possible (and were
6923 even the default) are now avoided.
6925 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6926 with each memory chunk allocated; this is occasionally more helpful
6927 than just having a counter.
6929 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6931 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6935 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6936 which largely parallels "options", but is for changing API behaviour,
6937 whereas "options" are about protocol behaviour.
6938 Initial "mode" flags are:
6940 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6941 a single record has been written.
6942 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6943 retries use the same buffer location.
6944 (But all of the contents must be
6948 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6951 *) Fix problems with no-hmac etc.
6952 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6954 *) New functions RSA_get_default_method(), RSA_set_method() and
6955 RSA_get_method(). These allows replacement of RSA_METHODs without having
6956 to mess around with the internals of an RSA structure.
6959 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6960 Also really enable memory leak checks in openssl.c and in some
6962 [Chad C. Mulligan, Bodo Moeller]
6964 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6965 up the length of negative integers. This has now been simplified to just
6966 store the length when it is first determined and use it later, rather
6967 than trying to keep track of where data is copied and updating it to
6969 [Steve Henson, reported by Brien Wheeler
6970 <bwheeler@authentica-security.com>]
6972 *) Add a new function PKCS7_signatureVerify. This allows the verification
6973 of a PKCS#7 signature but with the signing certificate passed to the
6974 function itself. This contrasts with PKCS7_dataVerify which assumes the
6975 certificate is present in the PKCS#7 structure. This isn't always the
6976 case: certificates can be omitted from a PKCS#7 structure and be
6977 distributed by "out of band" means (such as a certificate database).
6980 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6981 function prototypes in pem.h, also change util/mkdef.pl to add the
6982 necessary function names.
6985 *) mk1mf.pl (used by Windows builds) did not properly read the
6986 options set by Configure in the top level Makefile, and Configure
6987 was not even able to write more than one option correctly.
6988 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6991 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6992 file to be loaded from a BIO or FILE pointer. The BIO version will
6993 for example allow memory BIOs to contain config info.
6996 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6997 Whoever hopes to achieve shared-library compatibility across versions
6998 must use this, not the compile-time macro.
6999 (Exercise 0.9.4: Which is the minimum library version required by
7001 Note: All this applies only to multi-threaded programs, others don't
7005 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7006 through a BIO pair triggered the default case, i.e.
7007 SSLerr(...,SSL_R_UNKNOWN_STATE).
7010 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7011 can use the SSL library even if none of the specific BIOs is
7015 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7016 for the encoded length.
7017 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7019 *) Add initial documentation of the X509V3 functions.
7022 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7023 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7024 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7025 secure PKCS#8 private key format with a high iteration count.
7028 *) Fix determination of Perl interpreter: A perl or perl5
7029 _directory_ in $PATH was also accepted as the interpreter.
7030 [Ralf S. Engelschall]
7032 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7033 wrong with it but it was very old and did things like calling
7034 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7038 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7039 to use the new extension code.
7042 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7043 with macros. This should make it easier to change their form, add extra
7044 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7048 *) Add to configuration table a new entry that can specify an alternative
7049 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7050 according to Mark Crispin <MRC@Panda.COM>.
7054 *) DES CBC did not update the IV. Weird.
7057 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7058 Changing the behaviour of the former might break existing programs --
7059 where IV updating is needed, des_ncbc_encrypt can be used.
7062 *) When bntest is run from "make test" it drives bc to check its
7063 calculations, as well as internally checking them. If an internal check
7064 fails, it needs to cause bc to give a non-zero result or make test carries
7065 on without noticing the failure. Fixed.
7068 *) DES library cleanups.
7071 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7072 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7073 ciphers. NOTE: although the key derivation function has been verified
7074 against some published test vectors it has not been extensively tested
7075 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7079 *) Instead of "mkdir -p", which is not fully portable, use new
7080 Perl script "util/mkdir-p.pl".
7083 *) Rewrite the way password based encryption (PBE) is handled. It used to
7084 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7085 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7086 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7087 the 'parameter' field of the AlgorithmIdentifier is passed to the
7088 underlying key generation function so it must do its own ASN1 parsing.
7089 This has also changed the EVP_PBE_CipherInit() function which now has a
7090 'parameter' argument instead of literal salt and iteration count values
7091 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7094 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7095 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7096 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7097 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7098 value was just used as a "magic string" and not used directly its
7099 value doesn't matter.
7102 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7106 *) "linux-sparc64" configuration (ultrapenguin).
7107 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7108 "linux-sparc" configuration.
7109 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7111 *) config now generates no-xxx options for missing ciphers.
7114 *) Support the EBCDIC character set (work in progress).
7115 File ebcdic.c not yet included because it has a different license.
7116 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7118 *) Support BS2000/OSD-POSIX.
7119 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7121 *) Make callbacks for key generation use void * instead of char *.
7124 *) Make S/MIME samples compile (not yet tested).
7127 *) Additional typesafe stacks.
7130 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7134 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7136 *) New configuration variant "sco5-gcc".
7138 *) Updated some demos.
7139 [Sean O Riordain, Wade Scholine]
7141 *) Add missing BIO_free at exit of pkcs12 application.
7144 *) Fix memory leak in conf.c.
7147 *) Updates for Win32 to assembler version of MD5.
7150 *) Set #! path to perl in apps/der_chop to where we found it
7151 instead of using a fixed path.
7154 *) SHA library changes for irix64-mips4-cc.
7157 *) Improvements for VMS support.
7161 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7163 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7164 This also avoids the problems with SC4.2 and unpatched SC5.
7165 [Andy Polyakov <appro@fy.chalmers.se>]
7167 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7168 These are required because of the typesafe stack would otherwise break
7169 existing code. If old code used a structure member which used to be STACK
7170 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7171 sk_num or sk_value it would produce an error because the num, data members
7172 are not present in STACK_OF. Now it just produces a warning. sk_set
7173 replaces the old method of assigning a value to sk_value
7174 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7175 that does this will no longer work (and should use sk_set instead) but
7176 this could be regarded as a "questionable" behaviour anyway.
7179 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7180 correctly handle encrypted S/MIME data.
7183 *) Change type of various DES function arguments from des_cblock
7184 (which means, in function argument declarations, pointer to char)
7185 to des_cblock * (meaning pointer to array with 8 char elements),
7186 which allows the compiler to do more typechecking; it was like
7187 that back in SSLeay, but with lots of ugly casts.
7189 Introduce new type const_des_cblock.
7192 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7193 problems: find RecipientInfo structure that matches recipient certificate
7194 and initialise the ASN1 structures properly based on passed cipher.
7197 *) Belatedly make the BN tests actually check the results.
7200 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7201 to and from BNs: it was completely broken. New compilation option
7202 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7203 key elements as negative integers.
7206 *) Reorganize and speed up MD5.
7207 [Andy Polyakov <appro@fy.chalmers.se>]
7210 [Richard Levitte <richard@levitte.org>]
7212 *) New option -out to asn1parse to allow the parsed structure to be
7213 output to a file. This is most useful when combined with the -strparse
7214 option to examine the output of things like OCTET STRINGS.
7217 *) Make SSL library a little more fool-proof by not requiring any longer
7218 that SSL_set_{accept,connect}_state be called before
7219 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7220 in many applications because usually everything *appeared* to work as
7221 intended anyway -- now it really works as intended).
7224 *) Move openssl.cnf out of lib/.
7227 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7228 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7229 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7230 [Ralf S. Engelschall]
7232 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7233 handle PKCS#7 enveloped data properly.
7234 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7236 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7237 copying pointers. The cert_st handling is changed by this in
7238 various ways (and thus what used to be known as ctx->default_cert
7239 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7240 any longer when s->cert does not give us what we need).
7241 ssl_cert_instantiate becomes obsolete by this change.
7242 As soon as we've got the new code right (possibly it already is?),
7243 we have solved a couple of bugs of the earlier code where s->cert
7244 was used as if it could not have been shared with other SSL structures.
7246 Note that using the SSL API in certain dirty ways now will result
7247 in different behaviour than observed with earlier library versions:
7248 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7249 does not influence s as it used to.
7251 In order to clean up things more thoroughly, inside SSL_SESSION
7252 we don't use CERT any longer, but a new structure SESS_CERT
7253 that holds per-session data (if available); currently, this is
7254 the peer's certificate chain and, for clients, the server's certificate
7255 and temporary key. CERT holds only those values that can have
7256 meaningful defaults in an SSL_CTX.
7259 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7260 from the internal representation. Various PKCS#7 fixes: remove some
7261 evil casts and set the enc_dig_alg field properly based on the signing
7265 *) Allow PKCS#12 password to be set from the command line or the
7266 environment. Let 'ca' get its config file name from the environment
7267 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7271 *) Allow certificate policies extension to use an IA5STRING for the
7272 organization field. This is contrary to the PKIX definition but
7273 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7277 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7278 without disallowing inline assembler and the like for non-pedantic builds.
7281 *) Support Borland C++ builder.
7282 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7287 *) SHA-1 cleanups and performance enhancements.
7288 [Andy Polyakov <appro@fy.chalmers.se>]
7290 *) Sparc v8plus assembler for the bignum library.
7291 [Andy Polyakov <appro@fy.chalmers.se>]
7293 *) Accept any -xxx and +xxx compiler options in Configure.
7296 *) Update HPUX configuration.
7299 *) Add missing sk_<type>_unshift() function to safestack.h
7300 [Ralf S. Engelschall]
7302 *) New function SSL_CTX_use_certificate_chain_file that sets the
7303 "extra_cert"s in addition to the certificate. (This makes sense
7304 only for "PEM" format files, as chains as a whole are not
7308 *) Support verify_depth from the SSL API.
7309 x509_vfy.c had what can be considered an off-by-one-error:
7310 Its depth (which was not part of the external interface)
7311 was actually counting the number of certificates in a chain;
7312 now it really counts the depth.
7315 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7316 instead of X509err, which often resulted in confusing error
7317 messages since the error codes are not globally unique
7318 (e.g. an alleged error in ssl3_accept when a certificate
7319 didn't match the private key).
7321 *) New function SSL_CTX_set_session_id_context that allows to set a default
7322 value (so that you don't need SSL_set_session_id_context for each
7323 connection using the SSL_CTX).
7326 *) OAEP decoding bug fix.
7329 *) Support INSTALL_PREFIX for package builders, as proposed by
7333 *) New Configure options "threads" and "no-threads". For systems
7334 where the proper compiler options are known (currently Solaris
7335 and Linux), "threads" is the default.
7338 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7341 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7342 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7343 such as /usr/local/bin.
7346 *) "make linux-shared" to build shared libraries.
7347 [Niels Poppe <niels@netbox.org>]
7349 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7352 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7353 extension adding in x509 utility.
7356 *) Remove NOPROTO sections and error code comments.
7359 *) Partial rewrite of the DEF file generator to now parse the ANSI
7363 *) New Configure options --prefix=DIR and --openssldir=DIR.
7366 *) Complete rewrite of the error code script(s). It is all now handled
7367 by one script at the top level which handles error code gathering,
7368 header rewriting and C source file generation. It should be much better
7369 than the old method: it now uses a modified version of Ulf's parser to
7370 read the ANSI prototypes in all header files (thus the old K&R definitions
7371 aren't needed for error creation any more) and do a better job of
7372 translating function codes into names. The old 'ASN1 error code imbedded
7373 in a comment' is no longer necessary and it doesn't use .err files which
7374 have now been deleted. Also the error code call doesn't have to appear all
7375 on one line (which resulted in some large lines...).
7378 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7381 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7382 0 (which usually indicates a closed connection), but continue reading.
7385 *) Fix some race conditions.
7388 *) Add support for CRL distribution points extension. Add Certificate
7389 Policies and CRL distribution points documentation.
7392 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7395 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7396 8 of keying material. Merlin has also confirmed interop with this fix
7397 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7398 [Merlin Hughes <merlin@baltimore.ie>]
7400 *) Fix lots of warnings.
7401 [Richard Levitte <levitte@stacken.kth.se>]
7403 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7404 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7405 [Richard Levitte <levitte@stacken.kth.se>]
7407 *) Fix problems with sizeof(long) == 8.
7408 [Andy Polyakov <appro@fy.chalmers.se>]
7410 *) Change functions to ANSI C.
7413 *) Fix typos in error codes.
7414 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7416 *) Remove defunct assembler files from Configure.
7419 *) SPARC v8 assembler BIGNUM implementation.
7420 [Andy Polyakov <appro@fy.chalmers.se>]
7422 *) Support for Certificate Policies extension: both print and set.
7423 Various additions to support the r2i method this uses.
7426 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7427 return a const string when you are expecting an allocated buffer.
7430 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7431 types DirectoryString and DisplayText.
7434 *) Add code to allow r2i extensions to access the configuration database,
7435 add an LHASH database driver and add several ctx helper functions.
7438 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7439 fail when they extended the size of a BIGNUM.
7442 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7443 support typesafe stack.
7446 *) Fix typo in SSL_[gs]et_options().
7447 [Nils Frostberg <nils@medcom.se>]
7449 *) Delete various functions and files that belonged to the (now obsolete)
7450 old X509V3 handling code.
7453 *) New Configure option "rsaref".
7456 *) Don't auto-generate pem.h.
7459 *) Introduce type-safe ASN.1 SETs.
7462 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7463 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7465 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7466 that links with OpenSSL (well at least cause lots of warnings), but fear
7467 not: the conversion is trivial, and it eliminates loads of evil casts. A
7468 few STACKed things have been converted already. Feel free to convert more.
7469 In the fullness of time, I'll do away with the STACK type altogether.
7472 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7473 specified in <certfile> by updating the entry in the index.txt file.
7474 This way one no longer has to edit the index.txt file manually for
7475 revoking a certificate. The -revoke option does the gory details now.
7476 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7478 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7479 `-text' option at all and this way the `-noout -text' combination was
7480 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7481 [Ralf S. Engelschall]
7483 *) Make sure a corresponding plain text error message exists for the
7484 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7485 verify callback function determined that a certificate was revoked.
7486 [Ralf S. Engelschall]
7488 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7489 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7490 all available cipers including rc5, which was forgotten until now.
7491 In order to let the testing shell script know which algorithms
7492 are available, a new (up to now undocumented) command
7493 "openssl list-cipher-commands" is used.
7496 *) Bugfix: s_client occasionally would sleep in select() when
7497 it should have checked SSL_pending() first.
7500 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7501 the raw DSA values prior to ASN.1 encoding.
7504 *) Tweaks to Configure
7505 [Niels Poppe <niels@netbox.org>]
7507 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7511 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7514 *) New config option to avoid instructions that are illegal on the 80386.
7515 The default code is faster, but requires at least a 486.
7518 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7519 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7520 same as SSL2_VERSION anyway.
7523 *) New "-showcerts" option for s_client.
7526 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7527 application. Various cleanups and fixes.
7530 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7531 modify error routines to work internally. Add error codes and PBE init
7532 to library startup routines.
7535 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7536 packing functions to asn1 and evp. Changed function names and error
7537 codes along the way.
7540 *) PKCS12 integration: and so it begins... First of several patches to
7541 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7542 objects to objects.h
7545 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7546 and display support for Thawte strong extranet extension.
7549 *) Add LinuxPPC support.
7550 [Jeff Dubrule <igor@pobox.org>]
7552 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7553 bn_div_words in alpha.s.
7554 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7556 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7557 OAEP isn't supported when OpenSSL is built with RSAref.
7558 [Ulf Moeller <ulf@fitug.de>]
7560 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7561 so they no longer are missing under -DNOPROTO.
7562 [Soren S. Jorvang <soren@t.dk>]
7565 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7567 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7568 doesn't work when the session is reused. Coming soon!
7571 *) Fix a security hole, that allows sessions to be reused in the wrong
7572 context thus bypassing client cert protection! All software that uses
7573 client certs and session caches in multiple contexts NEEDS PATCHING to
7574 allow session reuse! A fuller solution is in the works.
7575 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7577 *) Some more source tree cleanups (removed obsolete files
7578 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7579 permission on "config" script to be executable) and a fix for the INSTALL
7581 [Ulf Moeller <ulf@fitug.de>]
7583 *) Remove some legacy and erroneous uses of malloc, free instead of
7585 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7587 *) Make rsa_oaep_test return non-zero on error.
7588 [Ulf Moeller <ulf@fitug.de>]
7590 *) Add support for native Solaris shared libraries. Configure
7591 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7592 if someone would make that last step automatic.
7593 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7595 *) ctx_size was not built with the right compiler during "make links". Fixed.
7598 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7599 except NULL ciphers". This means the default cipher list will no longer
7600 enable NULL ciphers. They need to be specifically enabled e.g. with
7601 the string "DEFAULT:eNULL".
7604 *) Fix to RSA private encryption routines: if p < q then it would
7605 occasionally produce an invalid result. This will only happen with
7606 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7609 *) Be less restrictive and allow also `perl util/perlpath.pl
7610 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7611 because this way one can also use an interpreter named `perl5' (which is
7612 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7613 installed as `perl').
7614 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7616 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7617 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7619 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7620 advapi32.lib to Win32 build and change the pem test comparision
7621 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7622 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7623 and crypto/des/ede_cbcm_enc.c.
7626 *) DES quad checksum was broken on big-endian architectures. Fixed.
7629 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7630 Win32 test batch file so it (might) work again. The Win32 test batch file
7631 is horrible: I feel ill....
7634 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7635 in e_os.h. Audit of header files to check ANSI and non ANSI
7636 sections: 10 functions were absent from non ANSI section and not exported
7637 from Windows DLLs. Fixed up libeay.num for new functions.
7640 *) Make `openssl version' output lines consistent.
7641 [Ralf S. Engelschall]
7643 *) Fix Win32 symbol export lists for BIO functions: Added
7644 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7645 to ms/libeay{16,32}.def.
7646 [Ralf S. Engelschall]
7648 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7649 fine under Unix and passes some trivial tests I've now added. But the
7650 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7651 added to make sure no one expects that this stuff really works in the
7652 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7653 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7655 [Ralf S. Engelschall]
7657 *) Fix the generation of two part addresses in perl.
7658 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7660 *) Add config entry for Linux on MIPS.
7661 [John Tobey <jtobey@channel1.com>]
7663 *) Make links whenever Configure is run, unless we are on Windoze.
7666 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7667 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7671 *) Add a useful kludge to allow package maintainers to specify compiler and
7672 other platforms details on the command line without having to patch the
7673 Configure script everytime: One now can use ``perl Configure
7674 <id>:<details>'', i.e. platform ids are allowed to have details appended
7675 to them (seperated by colons). This is treated as there would be a static
7676 pre-configured entry in Configure's %table under key <id> with value
7677 <details> and ``perl Configure <id>'' is called. So, when you want to
7678 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7679 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7680 now, which overrides the FreeBSD-elf entry on-the-fly.
7681 [Ralf S. Engelschall]
7683 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7686 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7687 on the `perl Configure ...' command line. This way one can compile
7688 OpenSSL libraries with Position Independent Code (PIC) which is needed
7689 for linking it into DSOs.
7690 [Ralf S. Engelschall]
7692 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7696 *) Cleaned up the LICENSE document: The official contact for any license
7697 questions now is the OpenSSL core team under openssl-core@openssl.org.
7698 And add a paragraph about the dual-license situation to make sure people
7699 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7700 to the OpenSSL toolkit.
7701 [Ralf S. Engelschall]
7703 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7704 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7705 Additonally cleaned up the `make links' target: Remove unnecessary
7706 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7707 to speed processing and no longer clutter the display with confusing
7708 stuff. Instead only the actually done links are displayed.
7709 [Ralf S. Engelschall]
7711 *) Permit null encryption ciphersuites, used for authentication only. It used
7712 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7713 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7717 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7718 signed attributes when verifying signatures (this would break them),
7719 the detached data encoding was wrong and public keys obtained using
7720 X509_get_pubkey() weren't freed.
7723 *) Add text documentation for the BUFFER functions. Also added a work around
7724 to a Win95 console bug. This was triggered by the password read stuff: the
7725 last character typed gets carried over to the next fread(). If you were
7726 generating a new cert request using 'req' for example then the last
7727 character of the passphrase would be CR which would then enter the first
7731 *) Added the new `Includes OpenSSL Cryptography Software' button as
7732 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7733 button and can be used by applications based on OpenSSL to show the
7734 relationship to the OpenSSL project.
7735 [Ralf S. Engelschall]
7737 *) Remove confusing variables in function signatures in files
7738 ssl/ssl_lib.c and ssl/ssl.h.
7739 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7741 *) Don't install bss_file.c under PREFIX/include/
7742 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7744 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7745 functions that return function pointers and has support for NT specific
7746 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7747 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7748 unsigned to signed types: this was killing the Win32 compile.
7751 *) Add new certificate file to stack functions,
7752 SSL_add_dir_cert_subjects_to_stack() and
7753 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7754 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7755 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7756 This means that Apache-SSL and similar packages don't have to mess around
7757 to add as many CAs as they want to the preferred list.
7760 *) Experiment with doxygen documentation. Currently only partially applied to
7762 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7763 openssl.doxy as the configuration file.
7766 *) Get rid of remaining C++-style comments which strict C compilers hate.
7767 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7769 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7770 compiled in by default: it has problems with large keys.
7773 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7774 DH private keys and/or callback functions which directly correspond to
7775 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7776 is needed for applications which have to configure certificates on a
7777 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7779 For the RSA certificate situation is makes no difference, but
7780 for the DSA certificate situation this fixes the "no shared cipher"
7781 problem where the OpenSSL cipher selection procedure failed because the
7782 temporary keys were not overtaken from the context and the API provided
7783 no way to reconfigure them.
7784 The new functions now let applications reconfigure the stuff and they
7785 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7786 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7787 non-public-API function ssl_cert_instantiate() is used as a helper
7788 function and also to reduce code redundancy inside ssl_rsa.c.
7789 [Ralf S. Engelschall]
7791 *) Move s_server -dcert and -dkey options out of the undocumented feature
7792 area because they are useful for the DSA situation and should be
7793 recognized by the users.
7794 [Ralf S. Engelschall]
7796 *) Fix the cipher decision scheme for export ciphers: the export bits are
7797 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7798 SSL_EXP_MASK. So, the original variable has to be used instead of the
7799 already masked variable.
7800 [Richard Levitte <levitte@stacken.kth.se>]
7802 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7803 [Richard Levitte <levitte@stacken.kth.se>]
7805 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7806 from `int' to `unsigned int' because it's a length and initialized by
7807 EVP_DigestFinal() which expects an `unsigned int *'.
7808 [Richard Levitte <levitte@stacken.kth.se>]
7810 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7811 script. Instead use the usual Shell->Perl transition trick.
7812 [Ralf S. Engelschall]
7814 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7815 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7816 -noout -modulus' as it's already the case for `openssl rsa -noout
7817 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7818 currently the public key is printed (a decision which was already done by
7819 `openssl dsa -modulus' in the past) which serves a similar purpose.
7820 Additionally the NO_RSA no longer completely removes the whole -modulus
7821 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7823 [Ralf S. Engelschall]
7825 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7826 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7827 [Arne Ansper <arne@ats.cyber.ee>]
7829 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7830 to be added. Now both 'req' and 'ca' can use new objects defined in the
7834 *) Add cool BIO that does syslog (or event log on NT).
7835 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7837 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7838 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7839 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7840 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7843 *) Add preliminary config info for new extension code.
7846 *) Make RSA_NO_PADDING really use no padding.
7847 [Ulf Moeller <ulf@fitug.de>]
7849 *) Generate errors when private/public key check is done.
7852 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7853 for some CRL extensions and new objects added.
7856 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7857 key usage extension and fuller support for authority key id.
7860 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7861 padding method for RSA, which is recommended for new applications in PKCS
7862 #1 v2.0 (RFC 2437, October 1998).
7863 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7864 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7865 against Bleichbacher's attack on RSA.
7866 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7869 *) Updates to the new SSL compression code
7870 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7872 *) Fix so that the version number in the master secret, when passed
7873 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7874 (because the server will not accept higher), that the version number
7875 is 0x03,0x01, not 0x03,0x00
7876 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7878 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7879 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7880 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7883 *) Support for RAW extensions where an arbitrary extension can be
7884 created by including its DER encoding. See apps/openssl.cnf for
7888 *) Make sure latest Perl versions don't interpret some generated C array
7889 code as Perl array code in the crypto/err/err_genc.pl script.
7890 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7892 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7893 not many people have the assembler. Various Win32 compilation fixes and
7894 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7898 *) Modify configure script 'Configure' to automatically create crypto/date.h
7899 file under Win32 and also build pem.h from pem.org. New script
7900 util/mkfiles.pl to create the MINFO file on environments that can't do a
7901 'make files': perl util/mkfiles.pl >MINFO should work.
7904 *) Major rework of DES function declarations, in the pursuit of correctness
7905 and purity. As a result, many evil casts evaporated, and some weirdness,
7906 too. You may find this causes warnings in your code. Zapping your evil
7907 casts will probably fix them. Mostly.
7910 *) Fix for a typo in asn1.h. Bug fix to object creation script
7911 obj_dat.pl. It considered a zero in an object definition to mean
7912 "end of object": none of the objects in objects.h have any zeros
7913 so it wasn't spotted.
7914 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7916 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7917 Masking (CBCM). In the absence of test vectors, the best I have been able
7918 to do is check that the decrypt undoes the encrypt, so far. Send me test
7919 vectors if you have them.
7922 *) Correct calculation of key length for export ciphers (too much space was
7923 allocated for null ciphers). This has not been tested!
7926 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7927 message is now correct (it understands "crypto" and "ssl" on its
7928 command line). There is also now an "update" option. This will update
7929 the util/ssleay.num and util/libeay.num files with any new functions.
7931 perl util/mkdef.pl crypto ssl update
7932 it will update them.
7935 *) Overhauled the Perl interface (perl/*):
7936 - ported BN stuff to OpenSSL's different BN library
7937 - made the perl/ source tree CVS-aware
7938 - renamed the package from SSLeay to OpenSSL (the files still contain
7939 their history because I've copied them in the repository)
7940 - removed obsolete files (the test scripts will be replaced
7941 by better Test::Harness variants in the future)
7942 [Ralf S. Engelschall]
7944 *) First cut for a very conservative source tree cleanup:
7945 1. merge various obsolete readme texts into doc/ssleay.txt
7946 where we collect the old documents and readme texts.
7947 2. remove the first part of files where I'm already sure that we no
7948 longer need them because of three reasons: either they are just temporary
7949 files which were left by Eric or they are preserved original files where
7950 I've verified that the diff is also available in the CVS via "cvs diff
7951 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7952 the crypto/md/ stuff).
7953 [Ralf S. Engelschall]
7955 *) More extension code. Incomplete support for subject and issuer alt
7956 name, issuer and authority key id. Change the i2v function parameters
7957 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7958 what that's for :-) Fix to ASN1 macro which messed up
7959 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7962 *) Preliminary support for ENUMERATED type. This is largely copied from the
7966 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7967 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7969 *) Make sure `make rehash' target really finds the `openssl' program.
7970 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7972 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7973 like to hear about it if this slows down other processors.
7976 *) Add CygWin32 platform information to Configure script.
7977 [Alan Batie <batie@aahz.jf.intel.com>]
7979 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7980 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7982 *) New program nseq to manipulate netscape certificate sequences
7985 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7989 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7990 but the BN code had some problems that would cause failures when
7991 doing certificate verification and some other functions.
7992 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7994 *) Add ASN1 and PEM code to support netscape certificate sequences.
7997 *) Add ASN1 and PEM code to support netscape certificate sequences.
8000 *) Add several PKIX and private extended key usage OIDs.
8003 *) Modify the 'ca' program to handle the new extension code. Modify
8004 openssl.cnf for new extension format, add comments.
8007 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8008 and add a sample to openssl.cnf so req -x509 now adds appropriate
8012 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8013 error code, add initial support to X509_print() and x509 application.
8016 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8017 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8018 stuff is currently isolated and isn't even compiled yet.
8021 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8022 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8023 Removed the versions check from X509 routines when loading extensions:
8024 this allows certain broken certificates that don't set the version
8025 properly to be processed.
8028 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8029 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8030 can still be regenerated with "make depend".
8033 *) Spelling mistake in C version of CAST-128.
8034 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8036 *) Changes to the error generation code. The perl script err-code.pl
8037 now reads in the old error codes and retains the old numbers, only
8038 adding new ones if necessary. It also only changes the .err files if new
8039 codes are added. The makefiles have been modified to only insert errors
8040 when needed (to avoid needlessly modifying header files). This is done
8041 by only inserting errors if the .err file is newer than the auto generated
8042 C file. To rebuild all the error codes from scratch (the old behaviour)
8043 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8044 or delete all the .err files.
8047 *) CAST-128 was incorrectly implemented for short keys. The C version has
8048 been fixed, but is untested. The assembler versions are also fixed, but
8049 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8050 to regenerate it if needed.
8051 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8052 Hagino <itojun@kame.net>]
8054 *) File was opened incorrectly in randfile.c.
8055 [Ulf Möller <ulf@fitug.de>]
8057 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8058 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8059 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8060 al: it's just almost always a UTCTime. Note this patch adds new error
8061 codes so do a "make errors" if there are problems.
8064 *) Correct Linux 1 recognition in config.
8065 [Ulf Möller <ulf@fitug.de>]
8067 *) Remove pointless MD5 hash when using DSA keys in ca.
8068 [Anonymous <nobody@replay.com>]
8070 *) Generate an error if given an empty string as a cert directory. Also
8071 generate an error if handed NULL (previously returned 0 to indicate an
8072 error, but didn't set one).
8073 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8075 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8078 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8079 parameters. This was causing a warning which killed off the Win32 compile.
8082 *) Remove C++ style comments from crypto/bn/bn_local.h.
8083 [Neil Costigan <neil.costigan@celocom.com>]
8085 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8086 based on a text string, looking up short and long names and finally
8087 "dot" format. The "dot" format stuff didn't work. Added new function
8088 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8089 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8090 OID is not part of the table.
8093 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8094 X509_LOOKUP_by_alias().
8097 *) Sort openssl functions by name.
8100 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8101 encryption from sample DSA keys (in case anyone is interested the password
8105 *) Make _all_ *_free functions accept a NULL pointer.
8106 [Frans Heymans <fheymans@isaserver.be>]
8108 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8110 [Anonymous <nobody@replay.com>]
8112 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8113 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8115 *) Don't blow it for numeric -newkey arguments to apps/req.
8116 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8118 *) Temp key "for export" tests were wrong in s3_srvr.c.
8119 [Anonymous <nobody@replay.com>]
8121 *) Add prototype for temp key callback functions
8122 SSL_CTX_set_tmp_{rsa,dh}_callback().
8125 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8126 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8129 *) X509_name_add_entry() freed the wrong thing after an error.
8130 [Arne Ansper <arne@ats.cyber.ee>]
8132 *) rsa_eay.c would attempt to free a NULL context.
8133 [Arne Ansper <arne@ats.cyber.ee>]
8135 *) BIO_s_socket() had a broken should_retry() on Windoze.
8136 [Arne Ansper <arne@ats.cyber.ee>]
8138 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8139 [Arne Ansper <arne@ats.cyber.ee>]
8141 *) Make sure the already existing X509_STORE->depth variable is initialized
8142 in X509_STORE_new(), but document the fact that this variable is still
8143 unused in the certificate verification process.
8144 [Ralf S. Engelschall]
8146 *) Fix the various library and apps files to free up pkeys obtained from
8147 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8150 *) Fix reference counting in X509_PUBKEY_get(). This makes
8151 demos/maurice/example2.c work, amongst others, probably.
8152 [Steve Henson and Ben Laurie]
8154 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8155 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8156 are no longer created. This way we have a single and consistent command
8157 line interface `openssl <command>', similar to `cvs <command>'.
8158 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8160 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8161 BIT STRING wrapper always have zero unused bits.
8164 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8167 *) Make the top-level INSTALL documentation easier to understand.
8170 *) Makefiles updated to exit if an error occurs in a sub-directory
8171 make (including if user presses ^C) [Paul Sutton]
8173 *) Make Montgomery context stuff explicit in RSA data structure.
8176 *) Fix build order of pem and err to allow for generated pem.h.
8179 *) Fix renumbering bug in X509_NAME_delete_entry().
8182 *) Enhanced the err-ins.pl script so it makes the error library number
8183 global and can add a library name. This is needed for external ASN1 and
8184 other error libraries.
8187 *) Fixed sk_insert which never worked properly.
8190 *) Fix ASN1 macros so they can handle indefinite length construted
8191 EXPLICIT tags. Some non standard certificates use these: they can now
8195 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8196 into a single doc/ssleay.txt bundle. This way the information is still
8197 preserved but no longer messes up this directory. Now it's new room for
8198 the new set of documenation files.
8199 [Ralf S. Engelschall]
8201 *) SETs were incorrectly DER encoded. This was a major pain, because they
8202 shared code with SEQUENCEs, which aren't coded the same. This means that
8203 almost everything to do with SETs or SEQUENCEs has either changed name or
8204 number of arguments.
8205 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8207 *) Fix test data to work with the above.
8210 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8211 was already fixed by Eric for 0.9.1 it seems.
8212 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8214 *) Autodetect FreeBSD3.
8217 *) Fix various bugs in Configure. This affects the following platforms:
8221 unixware-2.0-pentium
8225 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8226 before they are needed.
8229 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8233 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8235 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8236 changed SSLeay to OpenSSL in version strings.
8237 [Ralf S. Engelschall]
8239 *) Some fixups to the top-level documents.
8242 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8243 because the symlink to include/ was missing.
8244 [Ralf S. Engelschall]
8246 *) Incorporated the popular no-RSA/DSA-only patches
8247 which allow to compile a RSA-free SSLeay.
8248 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8250 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8251 when "ssleay" is still not found.
8252 [Ralf S. Engelschall]
8254 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8255 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8257 *) Updated the README file.
8258 [Ralf S. Engelschall]
8260 *) Added various .cvsignore files in the CVS repository subdirs
8261 to make a "cvs update" really silent.
8262 [Ralf S. Engelschall]
8264 *) Recompiled the error-definition header files and added
8265 missing symbols to the Win32 linker tables.
8266 [Ralf S. Engelschall]
8268 *) Cleaned up the top-level documents;
8269 o new files: CHANGES and LICENSE
8270 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8271 o merged COPYRIGHT into LICENSE
8272 o removed obsolete TODO file
8273 o renamed MICROSOFT to INSTALL.W32
8274 [Ralf S. Engelschall]
8276 *) Removed dummy files from the 0.9.1b source tree:
8277 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8278 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8279 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8280 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8281 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8282 [Ralf S. Engelschall]
8284 *) Added various platform portability fixes.
8287 *) The Genesis of the OpenSSL rpject:
8288 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8289 Young and Tim J. Hudson created while they were working for C2Net until
8291 [The OpenSSL Project]
8294 Changes between 0.9.0b and 0.9.1b [not released]
8296 *) Updated a few CA certificates under certs/
8299 *) Changed some BIGNUM api stuff.
8302 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8303 DGUX x86, Linux Alpha, etc.
8306 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8307 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8311 *) Add -strparse option to asn1pars program which parses nested
8313 [Dr Stephen Henson <shenson@bigfoot.com>]
8315 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8318 *) DSA fix for "ca" program.
8321 *) Added "-genkey" option to "dsaparam" program.
8324 *) Added RIPE MD160 (rmd160) message digest.
8327 *) Added -a (all) option to "ssleay version" command.
8330 *) Added PLATFORM define which is the id given to Configure.
8333 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8336 *) Extended the ASN.1 parser routines.
8339 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8342 *) Added a BN_CTX to the BN library.
8345 *) Fixed the weak key values in DES library
8348 *) Changed API in EVP library for cipher aliases.
8351 *) Added support for RC2/64bit cipher.
8354 *) Converted the lhash library to the crypto/mem.c functions.
8357 *) Added more recognized ASN.1 object ids.
8360 *) Added more RSA padding checks for SSL/TLS.
8363 *) Added BIO proxy/filter functionality.
8366 *) Added extra_certs to SSL_CTX which can be used
8367 send extra CA certificates to the client in the CA cert chain sending
8368 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8371 *) Now Fortezza is denied in the authentication phase because
8372 this is key exchange mechanism is not supported by SSLeay at all.
8375 *) Additional PKCS1 checks.
8378 *) Support the string "TLSv1" for all TLS v1 ciphers.
8381 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8382 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8385 *) Fixed a few memory leaks.
8388 *) Fixed various code and comment typos.
8391 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8392 bytes sent in the client random.
8393 [Edward Bishop <ebishop@spyglass.com>]