5 Changes between 0.9.8i and 0.9.8j [xx XXX xxxx]
7 *) Update Configure code and WIN32 build scripts to support experimental
8 code. This is surrounded by OPENSSL_EXPERIMENTAL_FOO and not compiled
9 in by default. Using the configuration option "enable-experimental-foo"
10 enables it. Use this option for JPAKE.
13 *) Use correct exit code if there is an error in dgst command.
14 [Steve Henson; problem pointed out by Roland Dirlewanger]
16 *) Add JPAKE support, including demo authentication in s_client and
20 *) Set the comparison function in v3_addr_canonize().
21 [Rob Austein <sra@hactrn.net>]
23 *) Add support for XMPP STARTTLS in s_client.
24 [Philip Paeps <philip@freebsd.org>]
26 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
27 to ensure that even with this option, only ciphersuites in the
28 server's preference list will be accepted. (Note that the option
29 applies only when resuming a session, so the earlier behavior was
30 just about the algorithm choice for symmetric cryptography.)
33 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
35 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
36 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
39 *) The fix in 0.9.8c that supposedly got rid of unsafe
40 double-checked locking was incomplete for RSA blinding,
41 addressing just one layer of what turns out to have been
42 doubly unsafe triple-checked locking.
44 So now fix this for real by retiring the MONT_HELPER macro
45 in crypto/rsa/rsa_eay.c.
47 [Bodo Moeller; problem pointed out by Marius Schilder]
49 *) Various precautionary measures:
51 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
53 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
54 (NB: This would require knowledge of the secret session ticket key
55 to exploit, in which case you'd be SOL either way.)
57 - Change bn_nist.c so that it will properly handle input BIGNUMs
58 outside the expected range.
60 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
63 [Neel Mehta, Bodo Moeller]
65 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
66 the load fails. Useful for distros.
67 [Ben Laurie and the FreeBSD team]
69 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
72 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
75 *) Expand ENGINE to support engine supplied SSL client certificate functions.
77 This work was sponsored by Logica.
80 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
81 keystores. Support for SSL/TLS client authentication too.
82 Not compiled unless enable-capieng specified to Configure.
84 This work was sponsored by Logica.
87 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
88 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
89 attribute creation routines such as certifcate requests and PKCS#12
93 Changes between 0.9.8g and 0.9.8h [28 May 2008]
95 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
96 handshake which could lead to a cilent crash as found using the
97 Codenomicon TLS test suite (CVE-2008-1672)
98 [Steve Henson, Mark Cox]
100 *) Fix double free in TLS server name extensions which could lead to
101 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
104 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
106 Clear the error queue to ensure that error entries left from
107 older function calls do not interfere with the correct operation.
108 [Lutz Jaenicke, Erik de Castro Lopo]
110 *) Remove root CA certificates of commercial CAs:
112 The OpenSSL project does not recommend any specific CA and does not
113 have any policy with respect to including or excluding any CA.
114 Therefore it does not make any sense to ship an arbitrary selection
115 of root CA certificates with the OpenSSL software.
118 *) RSA OAEP patches to fix two separate invalid memory reads.
119 The first one involves inputs when 'lzero' is greater than
120 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
121 before the beginning of from). The second one involves inputs where
122 the 'db' section contains nothing but zeroes (there is a one-byte
123 invalid read after the end of 'db').
124 [Ivan Nestlerode <inestlerode@us.ibm.com>]
126 *) Partial backport from 0.9.9-dev:
128 Introduce bn_mul_mont (dedicated Montgomery multiplication
129 procedure) as a candidate for BIGNUM assembler implementation.
130 While 0.9.9-dev uses assembler for various architectures, only
131 x86_64 is available by default here in the 0.9.8 branch, and
132 32-bit x86 is available through a compile-time setting.
134 To try the 32-bit x86 assembler implementation, use Configure
135 option "enable-montasm" (which exists only for this backport).
137 As "enable-montasm" for 32-bit x86 disclaims code stability
138 anyway, in this constellation we activate additional code
139 backported from 0.9.9-dev for further performance improvements,
140 namely BN_from_montgomery_word. (To enable this otherwise,
141 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
143 [Andy Polyakov (backport partially by Bodo Moeller)]
145 *) Add TLS session ticket callback. This allows an application to set
146 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
147 values. This is useful for key rollover for example where several key
148 sets may exist with different names.
151 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
152 This was broken until now in 0.9.8 releases, such that the only way
153 a registered ENGINE could be used (assuming it initialises
154 successfully on the host) was to explicitly set it as the default
155 for the relevant algorithms. This is in contradiction with 0.9.7
156 behaviour and the documentation. With this fix, when an ENGINE is
157 registered into a given algorithm's table of implementations, the
158 'uptodate' flag is reset so that auto-discovery will be used next
159 time a new context for that algorithm attempts to select an
161 [Ian Lister (tweaked by Geoff Thorpe)]
163 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
164 implemention in the following ways:
166 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
169 Lack of BER streaming support means one pass streaming processing is
170 only supported if data is detached: setting the streaming flag is
171 ignored for embedded content.
173 CMS support is disabled by default and must be explicitly enabled
174 with the enable-cms configuration option.
177 *) Update the GMP engine glue to do direct copies between BIGNUM and
178 mpz_t when openssl and GMP use the same limb size. Otherwise the
179 existing "conversion via a text string export" trick is still used.
180 [Paul Sheer <paulsheer@gmail.com>]
182 *) Zlib compression BIO. This is a filter BIO which compressed and
183 uncompresses any data passed through it.
186 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
187 RFC3394 compatible AES key wrapping.
190 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
191 sets string data without copying. X509_ALGOR_set0() and
192 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
193 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
194 from an X509_ATTRIBUTE structure optionally checking it occurs only
195 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
199 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
200 to get the expected BN_FLG_CONSTTIME behavior.
201 [Bodo Moeller (Google)]
205 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
206 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
207 - added some more tests to do_tests.pl
208 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
209 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
210 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
211 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
212 - various changes to netware.pl to enable gcc-cross builds on Win32
214 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
215 - various changes to fix missing prototype warnings
216 - fixed x86nasm.pl to create correct asm files for NASM COFF output
217 - added AES, WHIRLPOOL and CPUID assembler code to build files
218 - added missing AES assembler make rules to mk1mf.pl
219 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
220 [Guenter Knauf <eflash@gmx.net>]
222 *) Implement certificate status request TLS extension defined in RFC3546.
223 A client can set the appropriate parameters and receive the encoded
224 OCSP response via a callback. A server can query the supplied parameters
225 and set the encoded OCSP response in the callback. Add simplified examples
226 to s_client and s_server.
229 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
232 + Binary incompatibility of ssl_ctx_st structure
233 + DTLS interoperation with non-compliant servers
234 + Don't call get_session_cb() without proposed session
235 + Fix ia64 assembler code
236 [Andy Polyakov, Steve Henson]
238 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
240 *) DTLS Handshake overhaul. There were longstanding issues with
241 OpenSSL DTLS implementation, which were making it impossible for
242 RFC 4347 compliant client to communicate with OpenSSL server.
243 Unfortunately just fixing these incompatibilities would "cut off"
244 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
245 server keeps tolerating non RFC compliant syntax. The opposite is
246 not true, 0.9.8f client can not communicate with earlier server.
247 This update even addresses CVE-2007-4995.
250 *) Changes to avoid need for function casts in OpenSSL: some compilers
251 (gcc 4.2 and later) reject their use.
252 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
255 *) Add RFC4507 support to OpenSSL. This includes the corrections in
256 RFC4507bis. The encrypted ticket format is an encrypted encoded
257 SSL_SESSION structure, that way new session features are automatically
260 If a client application caches session in an SSL_SESSION structure
261 support is transparent because tickets are now stored in the encoded
264 The SSL_CTX structure automatically generates keys for ticket
265 protection in servers so again support should be possible
266 with no application modification.
268 If a client or server wishes to disable RFC4507 support then the option
269 SSL_OP_NO_TICKET can be set.
271 Add a TLS extension debugging callback to allow the contents of any client
272 or server extensions to be examined.
274 This work was sponsored by Google.
277 *) Add initial support for TLS extensions, specifically for the server_name
278 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
279 have new members for a host name. The SSL data structure has an
280 additional member SSL_CTX *initial_ctx so that new sessions can be
281 stored in that context to allow for session resumption, even after the
282 SSL has been switched to a new SSL_CTX in reaction to a client's
283 server_name extension.
285 New functions (subject to change):
288 SSL_get_servername_type()
291 New CTRL codes and macros (subject to change):
293 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
294 - SSL_CTX_set_tlsext_servername_callback()
295 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
296 - SSL_CTX_set_tlsext_servername_arg()
297 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
299 openssl s_client has a new '-servername ...' option.
301 openssl s_server has new options '-servername_host ...', '-cert2 ...',
302 '-key2 ...', '-servername_fatal' (subject to change). This allows
303 testing the HostName extension for a specific single host name ('-cert'
304 and '-key' remain fallbacks for handshakes without HostName
305 negotiation). If the unrecogninzed_name alert has to be sent, this by
306 default is a warning; it becomes fatal with the '-servername_fatal'
309 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
311 *) Add AES and SSE2 assembly language support to VC++ build.
314 *) Mitigate attack on final subtraction in Montgomery reduction.
317 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
318 (which previously caused an internal error).
321 *) Squeeze another 10% out of IGE mode when in != out.
324 *) AES IGE mode speedup.
325 [Dean Gaudet (Google)]
327 *) Add the Korean symmetric 128-bit cipher SEED (see
328 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
329 add SEED ciphersuites from RFC 4162:
331 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
332 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
333 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
334 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
336 To minimize changes between patchlevels in the OpenSSL 0.9.8
337 series, SEED remains excluded from compilation unless OpenSSL
338 is configured with 'enable-seed'.
341 *) Mitigate branch prediction attacks, which can be practical if a
342 single processor is shared, allowing a spy process to extract
343 information. For detailed background information, see
344 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
345 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
346 and Necessary Software Countermeasures"). The core of the change
347 are new versions BN_div_no_branch() and
348 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
349 respectively, which are slower, but avoid the security-relevant
350 conditional branches. These are automatically called by BN_div()
351 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
352 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
353 remove a conditional branch.
355 BN_FLG_CONSTTIME is the new name for the previous
356 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
357 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
358 in the exponent causes BN_mod_exp_mont() to use the alternative
359 implementation in BN_mod_exp_mont_consttime().) The old name
360 remains as a deprecated alias.
362 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
363 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
364 constant-time implementations for more than just exponentiation.
365 Here too the old name is kept as a deprecated alias.
367 BN_BLINDING_new() will now use BN_dup() for the modulus so that
368 the BN_BLINDING structure gets an independent copy of the
369 modulus. This means that the previous "BIGNUM *m" argument to
370 BN_BLINDING_new() and to BN_BLINDING_create_param() now
371 essentially becomes "const BIGNUM *m", although we can't actually
372 change this in the header file before 0.9.9. It allows
373 RSA_setup_blinding() to use BN_with_flags() on the modulus to
374 enable BN_FLG_CONSTTIME.
376 [Matthew D Wood (Intel Corp)]
378 *) In the SSL/TLS server implementation, be strict about session ID
379 context matching (which matters if an application uses a single
380 external cache for different purposes). Previously,
381 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
382 set. This did ensure strict client verification, but meant that,
383 with applications using a single external cache for quite
384 different requirements, clients could circumvent ciphersuite
385 restrictions for a given session ID context by starting a session
386 in a different context.
389 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
390 a ciphersuite string such as "DEFAULT:RSA" cannot enable
391 authentication-only ciphersuites.
394 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
395 not complete and could lead to a possible single byte overflow
396 (CVE-2007-5135) [Ben Laurie]
398 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
400 *) Since AES128 and AES256 (and similarly Camellia128 and
401 Camellia256) share a single mask bit in the logic of
402 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
403 kludge to work properly if AES128 is available and AES256 isn't
404 (or if Camellia128 is available and Camellia256 isn't).
407 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
408 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
409 When a point or a seed is encoded in a BIT STRING, we need to
410 prevent the removal of trailing zero bits to get the proper DER
411 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
412 of a NamedBitList, for which trailing 0 bits need to be removed.)
415 *) Have SSL/TLS server implementation tolerate "mismatched" record
416 protocol version while receiving ClientHello even if the
417 ClientHello is fragmented. (The server can't insist on the
418 particular protocol version it has chosen before the ServerHello
419 message has informed the client about his choice.)
422 *) Add RFC 3779 support.
423 [Rob Austein for ARIN, Ben Laurie]
425 *) Load error codes if they are not already present instead of using a
426 static variable. This allows them to be cleanly unloaded and reloaded.
427 Improve header file function name parsing.
430 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
431 or CAPABILITY handshake as required by RFCs.
434 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
436 *) Introduce limits to prevent malicious keys being able to
437 cause a denial of service. (CVE-2006-2940)
438 [Steve Henson, Bodo Moeller]
440 *) Fix ASN.1 parsing of certain invalid structures that can result
441 in a denial of service. (CVE-2006-2937) [Steve Henson]
443 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
444 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
446 *) Fix SSL client code which could crash if connecting to a
447 malicious SSLv2 server. (CVE-2006-4343)
448 [Tavis Ormandy and Will Drewry, Google Security Team]
450 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
451 match only those. Before that, "AES256-SHA" would be interpreted
452 as a pattern and match "AES128-SHA" too (since AES128-SHA got
453 the same strength classification in 0.9.7h) as we currently only
454 have a single AES bit in the ciphersuite description bitmap.
455 That change, however, also applied to ciphersuite strings such as
456 "RC4-MD5" that intentionally matched multiple ciphersuites --
457 namely, SSL 2.0 ciphersuites in addition to the more common ones
458 from SSL 3.0/TLS 1.0.
460 So we change the selection algorithm again: Naming an explicit
461 ciphersuite selects this one ciphersuite, and any other similar
462 ciphersuite (same bitmap) from *other* protocol versions.
463 Thus, "RC4-MD5" again will properly select both the SSL 2.0
464 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
466 Since SSL 2.0 does not have any ciphersuites for which the
467 128/256 bit distinction would be relevant, this works for now.
468 The proper fix will be to use different bits for AES128 and
469 AES256, which would have avoided the problems from the beginning;
470 however, bits are scarce, so we can only do this in a new release
471 (not just a patchlevel) when we can change the SSL_CIPHER
472 definition to split the single 'unsigned long mask' bitmap into
473 multiple values to extend the available space.
477 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
479 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
480 (CVE-2006-4339) [Ben Laurie and Google Security Team]
482 *) Add AES IGE and biIGE modes.
485 *) Change the Unix randomness entropy gathering to use poll() when
486 possible instead of select(), since the latter has some
487 undesirable limitations.
488 [Darryl Miles via Richard Levitte and Bodo Moeller]
490 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
491 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
492 cannot be implicitly activated as part of, e.g., the "AES" alias.
493 However, please upgrade to OpenSSL 0.9.9[-dev] for
494 non-experimental use of the ECC ciphersuites to get TLS extension
495 support, which is required for curve and point format negotiation
496 to avoid potential handshake problems.
499 *) Disable rogue ciphersuites:
501 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
502 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
503 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
505 The latter two were purportedly from
506 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
509 Also deactivate the remaining ciphersuites from
510 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
511 unofficial, and the ID has long expired.
514 *) Fix RSA blinding Heisenbug (problems sometimes occured on
515 dual-core machines) and other potential thread-safety issues.
518 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
519 versions), which is now available for royalty-free use
520 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
521 Also, add Camellia TLS ciphersuites from RFC 4132.
523 To minimize changes between patchlevels in the OpenSSL 0.9.8
524 series, Camellia remains excluded from compilation unless OpenSSL
525 is configured with 'enable-camellia'.
528 *) Disable the padding bug check when compression is in use. The padding
529 bug check assumes the first packet is of even length, this is not
530 necessarily true if compresssion is enabled and can result in false
531 positives causing handshake failure. The actual bug test is ancient
532 code so it is hoped that implementations will either have fixed it by
533 now or any which still have the bug do not support compression.
536 Changes between 0.9.8a and 0.9.8b [04 May 2006]
538 *) When applying a cipher rule check to see if string match is an explicit
539 cipher suite and only match that one cipher suite if it is.
542 *) Link in manifests for VC++ if needed.
543 [Austin Ziegler <halostatue@gmail.com>]
545 *) Update support for ECC-based TLS ciphersuites according to
546 draft-ietf-tls-ecc-12.txt with proposed changes (but without
547 TLS extensions, which are supported starting with the 0.9.9
548 branch, not in the OpenSSL 0.9.8 branch).
551 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
552 opaque EVP_CIPHER_CTX handling.
555 *) Fixes and enhancements to zlib compression code. We now only use
556 "zlib1.dll" and use the default __cdecl calling convention on Win32
557 to conform with the standards mentioned here:
558 http://www.zlib.net/DLL_FAQ.txt
559 Static zlib linking now works on Windows and the new --with-zlib-include
560 --with-zlib-lib options to Configure can be used to supply the location
561 of the headers and library. Gracefully handle case where zlib library
565 *) Several fixes and enhancements to the OID generation code. The old code
566 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
567 handle numbers larger than ULONG_MAX, truncated printing and had a
568 non standard OBJ_obj2txt() behaviour.
571 *) Add support for building of engines under engine/ as shared libraries
572 under VC++ build system.
575 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
576 Hopefully, we will not see any false combination of paths any more.
579 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
581 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
582 (part of SSL_OP_ALL). This option used to disable the
583 countermeasure against man-in-the-middle protocol-version
584 rollback in the SSL 2.0 server implementation, which is a bad
585 idea. (CVE-2005-2969)
587 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
588 for Information Security, National Institute of Advanced Industrial
589 Science and Technology [AIST], Japan)]
591 *) Add two function to clear and return the verify parameter flags.
594 *) Keep cipherlists sorted in the source instead of sorting them at
595 runtime, thus removing the need for a lock.
598 *) Avoid some small subgroup attacks in Diffie-Hellman.
599 [Nick Mathewson and Ben Laurie]
601 *) Add functions for well-known primes.
604 *) Extended Windows CE support.
605 [Satoshi Nakamura and Andy Polyakov]
607 *) Initialize SSL_METHOD structures at compile time instead of during
608 runtime, thus removing the need for a lock.
611 *) Make PKCS7_decrypt() work even if no certificate is supplied by
612 attempting to decrypt each encrypted key in turn. Add support to
616 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
618 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
621 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
624 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
625 key into the same file any more.
628 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
631 *) Add -utf8 command line and config file option to 'ca'.
632 [Stefan <stf@udoma.org]
634 *) Removed the macro des_crypt(), as it seems to conflict with some
635 libraries. Use DES_crypt().
638 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
639 involves renaming the source and generated shared-libs for
640 both. The engines will accept the corrected or legacy ids
641 ('ncipher' and '4758_cca' respectively) when binding. NB,
642 this only applies when building 'shared'.
643 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
645 *) Add attribute functions to EVP_PKEY structure. Modify
646 PKCS12_create() to recognize a CSP name attribute and
647 use it. Make -CSP option work again in pkcs12 utility.
650 *) Add new functionality to the bn blinding code:
651 - automatic re-creation of the BN_BLINDING parameters after
652 a fixed number of uses (currently 32)
653 - add new function for parameter creation
654 - introduce flags to control the update behaviour of the
655 BN_BLINDING parameters
656 - hide BN_BLINDING structure
657 Add a second BN_BLINDING slot to the RSA structure to improve
658 performance when a single RSA object is shared among several
662 *) Add support for DTLS.
663 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
665 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
666 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
669 *) Remove buggy and incompletet DH cert support from
670 ssl/ssl_rsa.c and ssl/s3_both.c
673 *) Use SHA-1 instead of MD5 as the default digest algorithm for
674 the apps/openssl applications.
677 *) Compile clean with "-Wall -Wmissing-prototypes
678 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
679 DEBUG_SAFESTACK must also be set.
682 *) Change ./Configure so that certain algorithms can be disabled by default.
683 The new counterpiece to "no-xxx" is "enable-xxx".
685 The patented RC5 and MDC2 algorithms will now be disabled unless
686 "enable-rc5" and "enable-mdc2", respectively, are specified.
688 (IDEA remains enabled despite being patented. This is because IDEA
689 is frequently required for interoperability, and there is no license
690 fee for non-commercial use. As before, "no-idea" can be used to
691 avoid this algorithm.)
695 *) Add processing of proxy certificates (see RFC 3820). This work was
696 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
697 EGEE (Enabling Grids for E-science in Europe).
700 *) RC4 performance overhaul on modern architectures/implementations, such
701 as Intel P4, IA-64 and AMD64.
704 *) New utility extract-section.pl. This can be used specify an alternative
705 section number in a pod file instead of having to treat each file as
706 a separate case in Makefile. This can be done by adding two lines to the
709 =for comment openssl_section:XXX
711 The blank line is mandatory.
715 *) New arguments -certform, -keyform and -pass for s_client and s_server
716 to allow alternative format key and certificate files and passphrase
720 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
721 update associated structures and add various utility functions.
723 Add new policy related verify parameters, include policy checking in
724 standard verify code. Enhance 'smime' application with extra parameters
725 to support policy checking and print out.
728 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
729 Nehemiah processors. These extensions support AES encryption in hardware
730 as well as RNG (though RNG support is currently disabled).
731 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
733 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
736 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
737 [Andy Polyakov and a number of other people]
739 *) Improved PowerPC platform support. Most notably BIGNUM assembler
740 implementation contributed by IBM.
741 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
743 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
744 exponent rather than 'unsigned long'. There is a corresponding change to
745 the new 'rsa_keygen' element of the RSA_METHOD structure.
746 [Jelte Jansen, Geoff Thorpe]
748 *) Functionality for creating the initial serial number file is now
749 moved from CA.pl to the 'ca' utility with a new option -create_serial.
751 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
752 number file to 1, which is bound to cause problems. To avoid
753 the problems while respecting compatibility between different 0.9.7
754 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
755 CA.pl for serial number initialization. With the new release 0.9.8,
756 we can fix the problem directly in the 'ca' utility.)
759 *) Reduced header interdepencies by declaring more opaque objects in
760 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
761 give fewer recursive includes, which could break lazy source code - so
762 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
763 developers should define this symbol when building and using openssl to
764 ensure they track the recommended behaviour, interfaces, [etc], but
765 backwards-compatible behaviour prevails when this isn't defined.
768 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
771 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
772 This will generate a random key of the appropriate length based on the
773 cipher context. The EVP_CIPHER can provide its own random key generation
774 routine to support keys of a specific form. This is used in the des and
775 3des routines to generate a key of the correct parity. Update S/MIME
776 code to use new functions and hence generate correct parity DES keys.
777 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
778 valid (weak or incorrect parity).
781 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
782 as looking them up. This is useful when the verified structure may contain
783 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
784 present unless the new PKCS7_NO_CRL flag is asserted.
787 *) Extend ASN1 oid configuration module. It now additionally accepts the
790 shortName = some long name, 1.2.3.4
793 *) Reimplemented the BN_CTX implementation. There is now no more static
794 limitation on the number of variables it can handle nor the depth of the
795 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
796 information can now expand as required, and rather than having a single
797 static array of bignums, BN_CTX now uses a linked-list of such arrays
798 allowing it to expand on demand whilst maintaining the usefulness of
802 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
803 to allow all RSA operations to function using a single BN_CTX.
806 *) Preliminary support for certificate policy evaluation and checking. This
807 is initially intended to pass the tests outlined in "Conformance Testing
808 of Relying Party Client Certificate Path Processing Logic" v1.07.
811 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
812 remained unused and not that useful. A variety of other little bignum
813 tweaks and fixes have also been made continuing on from the audit (see
817 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
818 associated ASN1, EVP and SSL functions and old ASN1 macros.
821 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
822 and this should never fail. So the return value from the use of
823 BN_set_word() (which can fail due to needless expansion) is now deprecated;
824 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
827 *) BN_CTX_get() should return zero-valued bignums, providing the same
828 initialised value as BN_new().
829 [Geoff Thorpe, suggested by Ulf Möller]
831 *) Support for inhibitAnyPolicy certificate extension.
834 *) An audit of the BIGNUM code is underway, for which debugging code is
835 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
836 is considered valid when processing BIGNUMs, and causes execution to
837 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
838 further steps are taken to deliberately pollute unused data in BIGNUM
839 structures to try and expose faulty code further on. For now, openssl will
840 (in its default mode of operation) continue to tolerate the inconsistent
841 forms that it has tolerated in the past, but authors and packagers should
842 consider trying openssl and their own applications when compiled with
843 these debugging symbols defined. It will help highlight potential bugs in
844 their own code, and will improve the test coverage for OpenSSL itself. At
845 some point, these tighter rules will become openssl's default to improve
846 maintainability, though the assert()s and other overheads will remain only
847 in debugging configurations. See bn.h for more details.
848 [Geoff Thorpe, Nils Larsch, Ulf Möller]
850 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
851 that can only be obtained through BN_CTX_new() (which implicitly
852 initialises it). The presence of this function only made it possible
853 to overwrite an existing structure (and cause memory leaks).
856 *) Because of the callback-based approach for implementing LHASH as a
857 template type, lh_insert() adds opaque objects to hash-tables and
858 lh_doall() or lh_doall_arg() are typically used with a destructor callback
859 to clean up those corresponding objects before destroying the hash table
860 (and losing the object pointers). So some over-zealous constifications in
861 LHASH have been relaxed so that lh_insert() does not take (nor store) the
862 objects as "const" and the lh_doall[_arg] callback wrappers are not
863 prototyped to have "const" restrictions on the object pointers they are
864 given (and so aren't required to cast them away any more).
867 *) The tmdiff.h API was so ugly and minimal that our own timing utility
868 (speed) prefers to use its own implementation. The two implementations
869 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
870 its object type properly exposed (MS_TM) instead of casting to/from "char
871 *". This may still change yet if someone realises MS_TM and "ms_time_***"
872 aren't necessarily the greatest nomenclatures - but this is what was used
873 internally to the implementation so I've used that for now.
876 *) Ensure that deprecated functions do not get compiled when
877 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
878 the self-tests were still using deprecated key-generation functions so
879 these have been updated also.
882 *) Reorganise PKCS#7 code to separate the digest location functionality
883 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
884 New function PKCS7_set_digest() to set the digest type for PKCS#7
885 digestedData type. Add additional code to correctly generate the
886 digestedData type and add support for this type in PKCS7 initialization
890 *) New function PKCS7_set0_type_other() this initializes a PKCS7
891 structure of type "other".
894 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
895 sure the loop does correctly stop and breaking ("division by zero")
896 modulus operations are not performed. The (pre-generated) prime
897 table crypto/bn/bn_prime.h was already correct, but it could not be
898 re-generated on some platforms because of the "division by zero"
899 situation in the script.
900 [Ralf S. Engelschall]
902 *) Update support for ECC-based TLS ciphersuites according to
903 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
904 SHA-1 now is only used for "small" curves (where the
905 representation of a field element takes up to 24 bytes); for
906 larger curves, the field element resulting from ECDH is directly
907 used as premaster secret.
908 [Douglas Stebila (Sun Microsystems Laboratories)]
910 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
911 curve secp160r1 to the tests.
912 [Douglas Stebila (Sun Microsystems Laboratories)]
914 *) Add the possibility to load symbols globally with DSO.
915 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
917 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
918 control of the error stack.
921 *) Add support for STORE in ENGINE.
924 *) Add the STORE type. The intention is to provide a common interface
925 to certificate and key stores, be they simple file-based stores, or
926 HSM-type store, or LDAP stores, or...
927 NOTE: The code is currently UNTESTED and isn't really used anywhere.
930 *) Add a generic structure called OPENSSL_ITEM. This can be used to
931 pass a list of arguments to any function as well as provide a way
932 for a function to pass data back to the caller.
935 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
936 works like BUF_strdup() but can be used to duplicate a portion of
937 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
941 *) Add the function sk_find_ex() which works like sk_find(), but will
942 return an index to an element even if an exact match couldn't be
943 found. The index is guaranteed to point at the element where the
944 searched-for key would be inserted to preserve sorting order.
947 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
948 takes an extra flags argument for optional functionality. Currently,
949 the following flags are defined:
951 OBJ_BSEARCH_VALUE_ON_NOMATCH
952 This one gets OBJ_bsearch_ex() to return a pointer to the first
953 element where the comparing function returns a negative or zero
956 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
957 This one gets OBJ_bsearch_ex() to return a pointer to the first
958 element where the comparing function returns zero. This is useful
959 if there are more than one element where the comparing function
963 *) Make it possible to create self-signed certificates with 'openssl ca'
964 in such a way that the self-signed certificate becomes part of the
965 CA database and uses the same mechanisms for serial number generation
966 as all other certificate signing. The new flag '-selfsign' enables
967 this functionality. Adapt CA.sh and CA.pl.in.
970 *) Add functionality to check the public key of a certificate request
971 against a given private. This is useful to check that a certificate
972 request can be signed by that key (self-signing).
975 *) Make it possible to have multiple active certificates with the same
976 subject in the CA index file. This is done only if the keyword
977 'unique_subject' is set to 'no' in the main CA section (default
978 if 'CA_default') of the configuration file. The value is saved
979 with the database itself in a separate index attribute file,
980 named like the index file with '.attr' appended to the name.
983 *) Generate muti valued AVAs using '+' notation in config files for
987 *) Support for nameConstraints certificate extension.
990 *) Support for policyConstraints certificate extension.
993 *) Support for policyMappings certificate extension.
996 *) Make sure the default DSA_METHOD implementation only uses its
997 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
998 and change its own handlers to be NULL so as to remove unnecessary
999 indirection. This lets alternative implementations fallback to the
1000 default implementation more easily.
1003 *) Support for directoryName in GeneralName related extensions
1007 *) Make it possible to link applications using Makefile.shared.
1008 Make that possible even when linking against static libraries!
1011 *) Support for single pass processing for S/MIME signing. This now
1012 means that S/MIME signing can be done from a pipe, in addition
1013 cleartext signing (multipart/signed type) is effectively streaming
1014 and the signed data does not need to be all held in memory.
1016 This is done with a new flag PKCS7_STREAM. When this flag is set
1017 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1018 is done after the data is output (and digests calculated) in
1019 SMIME_write_PKCS7().
1022 *) Add full support for -rpath/-R, both in shared libraries and
1023 applications, at least on the platforms where it's known how
1027 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1028 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1029 will now compute a table of multiples of the generator that
1030 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1031 faster (notably in the case of a single point multiplication,
1032 scalar * generator).
1033 [Nils Larsch, Bodo Moeller]
1035 *) IPv6 support for certificate extensions. The various extensions
1036 which use the IP:a.b.c.d can now take IPv6 addresses using the
1037 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1041 *) Added an ENGINE that implements RSA by performing private key
1042 exponentiations with the GMP library. The conversions to and from
1043 GMP's mpz_t format aren't optimised nor are any montgomery forms
1044 cached, and on x86 it appears OpenSSL's own performance has caught up.
1045 However there are likely to be other architectures where GMP could
1046 provide a boost. This ENGINE is not built in by default, but it can be
1047 specified at Configure time and should be accompanied by the necessary
1048 linker additions, eg;
1049 ./config -DOPENSSL_USE_GMP -lgmp
1052 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1053 testing availability of engines with "-t" - the old behaviour is
1054 produced by increasing the feature's verbosity with "-tt".
1057 *) ECDSA routines: under certain error conditions uninitialized BN objects
1058 could be freed. Solution: make sure initialization is performed early
1059 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1063 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1064 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1065 software implementations. For DSA and DH, parameter generation can
1066 also be overriden by providing the appropriate method callbacks.
1069 *) Change the "progress" mechanism used in key-generation and
1070 primality testing to functions that take a new BN_GENCB pointer in
1071 place of callback/argument pairs. The new API functions have "_ex"
1072 postfixes and the older functions are reimplemented as wrappers for
1073 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1074 declarations of the old functions to help (graceful) attempts to
1075 migrate to the new functions. Also, the new key-generation API
1076 functions operate on a caller-supplied key-structure and return
1077 success/failure rather than returning a key or NULL - this is to
1078 help make "keygen" another member function of RSA_METHOD etc.
1080 Example for using the new callback interface:
1082 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1086 BN_GENCB_set(&my_cb, my_callback, my_arg);
1088 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1089 /* For the meaning of a, b in calls to my_callback(), see the
1090 * documentation of the function that calls the callback.
1091 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1092 * my_callback should return 1 if it wants BN_is_prime_ex()
1093 * to continue, or 0 to stop.
1098 *) Change the ZLIB compression method to be stateful, and make it
1099 available to TLS with the number defined in
1100 draft-ietf-tls-compression-04.txt.
1103 *) Add the ASN.1 structures and functions for CertificatePair, which
1104 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1106 CertificatePair ::= SEQUENCE {
1107 forward [0] Certificate OPTIONAL,
1108 reverse [1] Certificate OPTIONAL,
1109 -- at least one of the pair shall be present -- }
1111 Also implement the PEM functions to read and write certificate
1112 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1114 This needed to be defined, mostly for the sake of the LDAP
1115 attribute crossCertificatePair, but may prove useful elsewhere as
1119 *) Make it possible to inhibit symlinking of shared libraries in
1120 Makefile.shared, for Cygwin's sake.
1123 *) Extend the BIGNUM API by creating a function
1124 void BN_set_negative(BIGNUM *a, int neg);
1125 and a macro that behave like
1126 int BN_is_negative(const BIGNUM *a);
1128 to avoid the need to access 'a->neg' directly in applications.
1131 *) Implement fast modular reduction for pseudo-Mersenne primes
1132 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1133 EC_GROUP_new_curve_GFp() will now automatically use this
1135 [Nils Larsch <nla@trustcenter.de>]
1137 *) Add new lock type (CRYPTO_LOCK_BN).
1140 *) Change the ENGINE framework to automatically load engines
1141 dynamically from specific directories unless they could be
1142 found to already be built in or loaded. Move all the
1143 current engines except for the cryptodev one to a new
1145 The engines in engines/ are built as shared libraries if
1146 the "shared" options was given to ./Configure or ./config.
1147 Otherwise, they are inserted in libcrypto.a.
1148 /usr/local/ssl/engines is the default directory for dynamic
1149 engines, but that can be overriden at configure time through
1150 the usual use of --prefix and/or --openssldir, and at run
1151 time with the environment variable OPENSSL_ENGINES.
1152 [Geoff Thorpe and Richard Levitte]
1154 *) Add Makefile.shared, a helper makefile to build shared
1155 libraries. Addapt Makefile.org.
1158 *) Add version info to Win32 DLLs.
1159 [Peter 'Luna' Runestig" <peter@runestig.com>]
1161 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1162 can be added using this API to created arbitrary PKCS#12
1163 files while avoiding the low level API.
1165 New options to PKCS12_create(), key or cert can be NULL and
1166 will then be omitted from the output file. The encryption
1167 algorithm NIDs can be set to -1 for no encryption, the mac
1168 iteration count can be set to 0 to omit the mac.
1170 Enhance pkcs12 utility by making the -nokeys and -nocerts
1171 options work when creating a PKCS#12 file. New option -nomac
1172 to omit the mac, NONE can be set for an encryption algorithm.
1173 New code is modified to use the enhanced PKCS12_create()
1174 instead of the low level API.
1177 *) Extend ASN1 encoder to support indefinite length constructed
1178 encoding. This can output sequences tags and octet strings in
1179 this form. Modify pk7_asn1.c to support indefinite length
1180 encoding. This is experimental and needs additional code to
1181 be useful, such as an ASN1 bio and some enhanced streaming
1184 Extend template encode functionality so that tagging is passed
1185 down to the template encoder.
1188 *) Let 'openssl req' fail if an argument to '-newkey' is not
1189 recognized instead of using RSA as a default.
1192 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1193 As these are not official, they are not included in "ALL";
1194 the "ECCdraft" ciphersuite group alias can be used to select them.
1195 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1197 *) Add ECDH engine support.
1198 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1200 *) Add ECDH in new directory crypto/ecdh/.
1201 [Douglas Stebila (Sun Microsystems Laboratories)]
1203 *) Let BN_rand_range() abort with an error after 100 iterations
1204 without success (which indicates a broken PRNG).
1207 *) Change BN_mod_sqrt() so that it verifies that the input value
1208 is really the square of the return value. (Previously,
1209 BN_mod_sqrt would show GIGO behaviour.)
1212 *) Add named elliptic curves over binary fields from X9.62, SECG,
1213 and WAP/WTLS; add OIDs that were still missing.
1215 [Sheueling Chang Shantz and Douglas Stebila
1216 (Sun Microsystems Laboratories)]
1218 *) Extend the EC library for elliptic curves over binary fields
1219 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1222 EC_GF2m_simple_method
1226 EC_GROUP_new_curve_GF2m
1227 EC_GROUP_set_curve_GF2m
1228 EC_GROUP_get_curve_GF2m
1229 EC_POINT_set_affine_coordinates_GF2m
1230 EC_POINT_get_affine_coordinates_GF2m
1231 EC_POINT_set_compressed_coordinates_GF2m
1233 Point compression for binary fields is disabled by default for
1234 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1237 As binary polynomials are represented as BIGNUMs, various members
1238 of the EC_GROUP and EC_POINT data structures can be shared
1239 between the implementations for prime fields and binary fields;
1240 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1241 are essentially identical to their ..._GFp counterparts.
1242 (For simplicity, the '..._GFp' prefix has been dropped from
1243 various internal method names.)
1245 An internal 'field_div' method (similar to 'field_mul' and
1246 'field_sqr') has been added; this is used only for binary fields.
1248 [Sheueling Chang Shantz and Douglas Stebila
1249 (Sun Microsystems Laboratories)]
1251 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1252 through methods ('mul', 'precompute_mult').
1254 The generic implementations (now internally called 'ec_wNAF_mul'
1255 and 'ec_wNAF_precomputed_mult') remain the default if these
1256 methods are undefined.
1258 [Sheueling Chang Shantz and Douglas Stebila
1259 (Sun Microsystems Laboratories)]
1261 *) New function EC_GROUP_get_degree, which is defined through
1262 EC_METHOD. For curves over prime fields, this returns the bit
1263 length of the modulus.
1265 [Sheueling Chang Shantz and Douglas Stebila
1266 (Sun Microsystems Laboratories)]
1268 *) New functions EC_GROUP_dup, EC_POINT_dup.
1269 (These simply call ..._new and ..._copy).
1271 [Sheueling Chang Shantz and Douglas Stebila
1272 (Sun Microsystems Laboratories)]
1274 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1275 Polynomials are represented as BIGNUMs (where the sign bit is not
1276 used) in the following functions [macros]:
1279 BN_GF2m_sub [= BN_GF2m_add]
1280 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1281 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1282 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1284 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1285 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1286 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1287 BN_GF2m_cmp [= BN_ucmp]
1289 (Note that only the 'mod' functions are actually for fields GF(2^m).
1290 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1292 For some functions, an the irreducible polynomial defining a
1293 field can be given as an 'unsigned int[]' with strictly
1294 decreasing elements giving the indices of those bits that are set;
1295 i.e., p[] represents the polynomial
1296 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1298 p[0] > p[1] > ... > p[k] = 0.
1299 This applies to the following functions:
1304 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1305 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1307 BN_GF2m_mod_sqrt_arr
1308 BN_GF2m_mod_solve_quad_arr
1312 Conversion can be performed by the following functions:
1317 bntest.c has additional tests for binary polynomial arithmetic.
1319 Two implementations for BN_GF2m_mod_div() are available.
1320 The default algorithm simply uses BN_GF2m_mod_inv() and
1321 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1322 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1323 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1325 [Sheueling Chang Shantz and Douglas Stebila
1326 (Sun Microsystems Laboratories)]
1328 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1329 functionality is disabled at compile-time.
1330 [Douglas Stebila <douglas.stebila@sun.com>]
1332 *) Change default behaviour of 'openssl asn1parse' so that more
1333 information is visible when viewing, e.g., a certificate:
1335 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1336 mode the content of non-printable OCTET STRINGs is output in a
1337 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1338 avoid the appearance of a printable string.
1339 [Nils Larsch <nla@trustcenter.de>]
1341 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1343 EC_GROUP_set_asn1_flag()
1344 EC_GROUP_get_asn1_flag()
1345 EC_GROUP_set_point_conversion_form()
1346 EC_GROUP_get_point_conversion_form()
1347 These control ASN1 encoding details:
1348 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1349 has been set to OPENSSL_EC_NAMED_CURVE.
1350 - Points are encoded in uncompressed form by default; options for
1351 asn1_for are as for point2oct, namely
1352 POINT_CONVERSION_COMPRESSED
1353 POINT_CONVERSION_UNCOMPRESSED
1354 POINT_CONVERSION_HYBRID
1356 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1359 EC_GROUP_get0_seed()
1360 EC_GROUP_get_seed_len()
1361 This is used only for ASN1 purposes (so far).
1362 [Nils Larsch <nla@trustcenter.de>]
1364 *) Add 'field_type' member to EC_METHOD, which holds the NID
1365 of the appropriate field type OID. The new function
1366 EC_METHOD_get_field_type() returns this value.
1367 [Nils Larsch <nla@trustcenter.de>]
1372 EC_POINT_point2hex()
1373 EC_POINT_hex2point()
1374 providing useful interfaces to EC_POINT_point2oct() and
1375 EC_POINT_oct2point().
1376 [Nils Larsch <nla@trustcenter.de>]
1378 *) Change internals of the EC library so that the functions
1379 EC_GROUP_set_generator()
1380 EC_GROUP_get_generator()
1381 EC_GROUP_get_order()
1382 EC_GROUP_get_cofactor()
1383 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1384 to methods, which would lead to unnecessary code duplication when
1385 adding different types of curves.
1386 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1388 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1389 arithmetic, and such that modified wNAFs are generated
1390 (which avoid length expansion in many cases).
1393 *) Add a function EC_GROUP_check_discriminant() (defined via
1394 EC_METHOD) that verifies that the curve discriminant is non-zero.
1396 Add a function EC_GROUP_check() that makes some sanity tests
1397 on a EC_GROUP, its generator and order. This includes
1398 EC_GROUP_check_discriminant().
1399 [Nils Larsch <nla@trustcenter.de>]
1401 *) Add ECDSA in new directory crypto/ecdsa/.
1403 Add applications 'openssl ecparam' and 'openssl ecdsa'
1404 (these are based on 'openssl dsaparam' and 'openssl dsa').
1406 ECDSA support is also included in various other files across the
1407 library. Most notably,
1408 - 'openssl req' now has a '-newkey ecdsa:file' option;
1409 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1410 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1411 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1412 them suitable for ECDSA where domain parameters must be
1413 extracted before the specific public key;
1414 - ECDSA engine support has been added.
1415 [Nils Larsch <nla@trustcenter.de>]
1417 *) Include some named elliptic curves, and add OIDs from X9.62,
1418 SECG, and WAP/WTLS. Each curve can be obtained from the new
1420 EC_GROUP_new_by_curve_name(),
1421 and the list of available named curves can be obtained with
1422 EC_get_builtin_curves().
1423 Also add a 'curve_name' member to EC_GROUP objects, which can be
1425 EC_GROUP_set_curve_name()
1426 EC_GROUP_get_curve_name()
1427 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1429 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1430 was actually never needed) and in BN_mul(). The removal in BN_mul()
1431 required a small change in bn_mul_part_recursive() and the addition
1432 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1433 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1434 bn_sub_words() and bn_add_words() except they take arrays with
1438 Changes between 0.9.7m and 0.9.7n [xx XXX xxxx]
1440 *) In the SSL/TLS server implementation, be strict about session ID
1441 context matching (which matters if an application uses a single
1442 external cache for different purposes). Previously,
1443 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1444 set. This did ensure strict client verification, but meant that,
1445 with applications using a single external cache for quite
1446 different requirements, clients could circumvent ciphersuite
1447 restrictions for a given session ID context by starting a session
1448 in a different context.
1451 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1453 *) Cleanse PEM buffers before freeing them since they may contain
1455 [Benjamin Bennett <ben@psc.edu>]
1457 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1458 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1459 authentication-only ciphersuites.
1462 *) Since AES128 and AES256 share a single mask bit in the logic of
1463 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1464 kludge to work properly if AES128 is available and AES256 isn't.
1467 *) Expand security boundary to match 1.1.1 module.
1470 *) Remove redundant features: hash file source, editing of test vectors
1471 modify fipsld to use external fips_premain.c signature.
1474 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1475 run algorithm test programs.
1478 *) Make algorithm test programs more tolerant of whitespace.
1481 *) Have SSL/TLS server implementation tolerate "mismatched" record
1482 protocol version while receiving ClientHello even if the
1483 ClientHello is fragmented. (The server can't insist on the
1484 particular protocol version it has chosen before the ServerHello
1485 message has informed the client about his choice.)
1488 *) Load error codes if they are not already present instead of using a
1489 static variable. This allows them to be cleanly unloaded and reloaded.
1492 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1494 *) Introduce limits to prevent malicious keys being able to
1495 cause a denial of service. (CVE-2006-2940)
1496 [Steve Henson, Bodo Moeller]
1498 *) Fix ASN.1 parsing of certain invalid structures that can result
1499 in a denial of service. (CVE-2006-2937) [Steve Henson]
1501 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1502 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1504 *) Fix SSL client code which could crash if connecting to a
1505 malicious SSLv2 server. (CVE-2006-4343)
1506 [Tavis Ormandy and Will Drewry, Google Security Team]
1508 *) Change ciphersuite string processing so that an explicit
1509 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1510 will no longer include "AES128-SHA"), and any other similar
1511 ciphersuite (same bitmap) from *other* protocol versions (so that
1512 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1513 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1514 changes from 0.9.8b and 0.9.8d.
1517 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1519 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1520 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1522 *) Change the Unix randomness entropy gathering to use poll() when
1523 possible instead of select(), since the latter has some
1524 undesirable limitations.
1525 [Darryl Miles via Richard Levitte and Bodo Moeller]
1527 *) Disable rogue ciphersuites:
1529 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1530 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1531 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1533 The latter two were purportedly from
1534 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1537 Also deactive the remaining ciphersuites from
1538 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1539 unofficial, and the ID has long expired.
1542 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1543 dual-core machines) and other potential thread-safety issues.
1546 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1548 *) Adapt fipsld and the build system to link against the validated FIPS
1549 module in FIPS mode.
1552 *) Fixes for VC++ 2005 build under Windows.
1555 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1556 from a Windows bash shell such as MSYS. It is autodetected from the
1557 "config" script when run from a VC++ environment. Modify standard VC++
1558 build to use fipscanister.o from the GNU make build.
1561 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1563 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1564 The value now differs depending on if you build for FIPS or not.
1565 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1566 safely run with a non-FIPSed libcrypto, as it may crash because of
1567 the difference induced by this change.
1570 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1572 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1573 (part of SSL_OP_ALL). This option used to disable the
1574 countermeasure against man-in-the-middle protocol-version
1575 rollback in the SSL 2.0 server implementation, which is a bad
1576 idea. (CVE-2005-2969)
1578 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1579 for Information Security, National Institute of Advanced Industrial
1580 Science and Technology [AIST], Japan)]
1582 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1583 mainly for FIPS compliance and not fully integrated at this stage.
1586 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1587 the exponentiation using a fixed-length exponent. (Otherwise,
1588 the information leaked through timing could expose the secret key
1589 after many signatures; cf. Bleichenbacher's attack on DSA with
1593 *) Make a new fixed-window mod_exp implementation the default for
1594 RSA, DSA, and DH private-key operations so that the sequence of
1595 squares and multiplies and the memory access pattern are
1596 independent of the particular secret key. This will mitigate
1597 cache-timing and potential related attacks.
1599 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1600 and this is automatically used by BN_mod_exp_mont() if the new flag
1601 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1602 will use this BN flag for private exponents unless the flag
1603 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1604 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1606 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1608 *) Change the client implementation for SSLv23_method() and
1609 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1610 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1611 (Previously, the SSL 2.0 backwards compatible Client Hello
1612 message format would be used even with SSL_OP_NO_SSLv2.)
1615 *) Add support for smime-type MIME parameter in S/MIME messages which some
1619 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1620 a threadsafe manner. Modify rsa code to use new function and add calls
1621 to dsa and dh code (which had race conditions before).
1624 *) Include the fixed error library code in the C error file definitions
1625 instead of fixing them up at runtime. This keeps the error code
1626 structures constant.
1629 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1631 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1634 *) Fixes for newer kerberos headers. NB: the casts are needed because
1635 the 'length' field is signed on one version and unsigned on another
1636 with no (?) obvious way to tell the difference, without these VC++
1637 complains. Also the "definition" of FAR (blank) is no longer included
1638 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1639 some needed definitions.
1642 *) Undo Cygwin change.
1645 *) Added support for proxy certificates according to RFC 3820.
1646 Because they may be a security thread to unaware applications,
1647 they must be explicitely allowed in run-time. See
1648 docs/HOWTO/proxy_certificates.txt for further information.
1651 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1653 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1654 server and client random values. Previously
1655 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1656 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1658 This change has negligible security impact because:
1660 1. Server and client random values still have 24 bytes of pseudo random
1663 2. Server and client random values are sent in the clear in the initial
1666 3. The master secret is derived using the premaster secret (48 bytes in
1667 size for static RSA ciphersuites) as well as client server and random
1670 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1673 [Stephen Henson, reported by UK NISCC]
1675 *) Use Windows randomness collection on Cygwin.
1678 *) Fix hang in EGD/PRNGD query when communication socket is closed
1679 prematurely by EGD/PRNGD.
1680 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1682 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1685 *) Back-port of selected performance improvements from development
1686 branch, as well as improved support for PowerPC platforms.
1689 *) Add lots of checks for memory allocation failure, error codes to indicate
1690 failure and freeing up memory if a failure occurs.
1691 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1693 *) Add new -passin argument to dgst.
1696 *) Perform some character comparisons of different types in X509_NAME_cmp:
1697 this is needed for some certificates that reencode DNs into UTF8Strings
1698 (in violation of RFC3280) and can't or wont issue name rollover
1702 *) Make an explicit check during certificate validation to see that
1703 the CA setting in each certificate on the chain is correct. As a
1704 side effect always do the following basic checks on extensions,
1705 not just when there's an associated purpose to the check:
1707 - if there is an unhandled critical extension (unless the user
1708 has chosen to ignore this fault)
1709 - if the path length has been exceeded (if one is set at all)
1710 - that certain extensions fit the associated purpose (if one has
1714 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1716 *) Avoid a race condition when CRLs are checked in a multi threaded
1717 environment. This would happen due to the reordering of the revoked
1718 entries during signature checking and serial number lookup. Now the
1719 encoding is cached and the serial number sort performed under a lock.
1720 Add new STACK function sk_is_sorted().
1723 *) Add Delta CRL to the extension code.
1726 *) Various fixes to s3_pkt.c so alerts are sent properly.
1727 [David Holmes <d.holmes@f5.com>]
1729 *) Reduce the chances of duplicate issuer name and serial numbers (in
1730 violation of RFC3280) using the OpenSSL certificate creation utilities.
1731 This is done by creating a random 64 bit value for the initial serial
1732 number when a serial number file is created or when a self signed
1733 certificate is created using 'openssl req -x509'. The initial serial
1734 number file is created using 'openssl x509 -next_serial' in CA.pl
1735 rather than being initialized to 1.
1738 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1740 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1741 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1742 [Joe Orton, Steve Henson]
1744 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1746 [Joe Orton, Steve Henson]
1748 *) Make it possible to have multiple active certificates with the same
1749 subject in the CA index file. This is done only if the keyword
1750 'unique_subject' is set to 'no' in the main CA section (default
1751 if 'CA_default') of the configuration file. The value is saved
1752 with the database itself in a separate index attribute file,
1753 named like the index file with '.attr' appended to the name.
1756 *) X509 verify fixes. Disable broken certificate workarounds when
1757 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1758 keyUsage extension present. Don't accept CRLs with unhandled critical
1759 extensions: since verify currently doesn't process CRL extensions this
1760 rejects a CRL with *any* critical extensions. Add new verify error codes
1764 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1765 A clarification of RFC2560 will require the use of OCTET STRINGs and
1766 some implementations cannot handle the current raw format. Since OpenSSL
1767 copies and compares OCSP nonces as opaque blobs without any attempt at
1768 parsing them this should not create any compatibility issues.
1771 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1772 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1773 this HMAC (and other) operations are several times slower than OpenSSL
1777 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1778 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1780 *) Use the correct content when signing type "other".
1783 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1785 *) Fix various bugs revealed by running the NISCC test suite:
1787 Stop out of bounds reads in the ASN1 code when presented with
1788 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1790 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1792 If verify callback ignores invalid public key errors don't try to check
1793 certificate signature with the NULL public key.
1797 *) New -ignore_err option in ocsp application to stop the server
1798 exiting on the first error in a request.
1801 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1802 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1806 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1807 extra data after the compression methods not only for TLS 1.0
1808 but also for SSL 3.0 (as required by the specification).
1809 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1811 *) Change X509_certificate_type() to mark the key as exported/exportable
1812 when it's 512 *bits* long, not 512 bytes.
1815 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1816 blocks during encryption.
1819 *) Various fixes to base64 BIO and non blocking I/O. On write
1820 flushes were not handled properly if the BIO retried. On read
1821 data was not being buffered properly and had various logic bugs.
1822 This also affects blocking I/O when the data being decoded is a
1826 *) Various S/MIME bugfixes and compatibility changes:
1827 output correct application/pkcs7 MIME type if
1828 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1829 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1830 of files as .eml work). Correctly handle very long lines in MIME
1834 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1836 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1837 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1838 a protocol version number mismatch like a decryption error
1839 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1842 *) Turn on RSA blinding by default in the default implementation
1843 to avoid a timing attack. Applications that don't want it can call
1844 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1845 They would be ill-advised to do so in most cases.
1846 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1848 *) Change RSA blinding code so that it works when the PRNG is not
1849 seeded (in this case, the secret RSA exponent is abused as
1850 an unpredictable seed -- if it is not unpredictable, there
1851 is no point in blinding anyway). Make RSA blinding thread-safe
1852 by remembering the creator's thread ID in rsa->blinding and
1853 having all other threads use local one-time blinding factors
1854 (this requires more computation than sharing rsa->blinding, but
1855 avoids excessive locking; and if an RSA object is not shared
1856 between threads, blinding will still be very fast).
1859 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1860 ENGINE as defaults for all supported algorithms irrespective of
1861 the 'flags' parameter. 'flags' is now honoured, so applications
1862 should make sure they are passing it correctly.
1865 *) Target "mingw" now allows native Windows code to be generated in
1866 the Cygwin environment as well as with the MinGW compiler.
1869 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1871 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1872 via timing by performing a MAC computation even if incorrrect
1873 block cipher padding has been found. This is a countermeasure
1874 against active attacks where the attacker has to distinguish
1875 between bad padding and a MAC verification error. (CVE-2003-0078)
1877 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1878 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1879 Martin Vuagnoux (EPFL, Ilion)]
1881 *) Make the no-err option work as intended. The intention with no-err
1882 is not to have the whole error stack handling routines removed from
1883 libcrypto, it's only intended to remove all the function name and
1884 reason texts, thereby removing some of the footprint that may not
1885 be interesting if those errors aren't displayed anyway.
1887 NOTE: it's still possible for any application or module to have it's
1888 own set of error texts inserted. The routines are there, just not
1889 used by default when no-err is given.
1892 *) Add support for FreeBSD on IA64.
1893 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1895 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1896 Kerberos function mit_des_cbc_cksum(). Before this change,
1897 the value returned by DES_cbc_cksum() was like the one from
1898 mit_des_cbc_cksum(), except the bytes were swapped.
1899 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1901 *) Allow an application to disable the automatic SSL chain building.
1902 Before this a rather primitive chain build was always performed in
1903 ssl3_output_cert_chain(): an application had no way to send the
1904 correct chain if the automatic operation produced an incorrect result.
1906 Now the chain builder is disabled if either:
1908 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1910 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1912 The reasoning behind this is that an application would not want the
1913 auto chain building to take place if extra chain certificates are
1914 present and it might also want a means of sending no additional
1915 certificates (for example the chain has two certificates and the
1919 *) Add the possibility to build without the ENGINE framework.
1920 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1922 *) Under Win32 gmtime() can return NULL: check return value in
1923 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1926 *) DSA routines: under certain error conditions uninitialized BN objects
1927 could be freed. Solution: make sure initialization is performed early
1928 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1929 Nils Larsch <nla@trustcenter.de> via PR#459)
1932 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1933 checked on reconnect on the client side, therefore session resumption
1934 could still fail with a "ssl session id is different" error. This
1935 behaviour is masked when SSL_OP_ALL is used due to
1936 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1937 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1938 followup to PR #377.
1941 *) IA-32 assembler support enhancements: unified ELF targets, support
1942 for SCO/Caldera platforms, fix for Cygwin shared build.
1945 *) Add support for FreeBSD on sparc64. As a consequence, support for
1946 FreeBSD on non-x86 processors is separate from x86 processors on
1947 the config script, much like the NetBSD support.
1948 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1950 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1952 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1955 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1956 code (06) was taken as the first octet of the session ID and the last
1957 octet was ignored consequently. As a result SSLv2 client side session
1958 caching could not have worked due to the session ID mismatch between
1960 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1964 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1965 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1969 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1970 seems that in spite of existing for more than a year, many application
1971 author have done nothing to provide the necessary callbacks, which
1972 means that this particular engine will not work properly anywhere.
1973 This is a very unfortunate situation which forces us, in the name
1974 of usability, to give the hw_ncipher.c a static lock, which is part
1976 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1977 appear in 0.9.8 or later. We EXPECT application authors to have
1978 dealt properly with this when 0.9.8 is released (unless we actually
1979 make such changes in the libcrypto locking code that changes will
1980 have to be made anyway).
1983 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1984 octets have been read, EOF or an error occurs. Without this change
1985 some truncated ASN1 structures will not produce an error.
1988 *) Disable Heimdal support, since it hasn't been fully implemented.
1989 Still give the possibility to force the use of Heimdal, but with
1990 warnings and a request that patches get sent to openssl-dev.
1993 *) Add the VC-CE target, introduce the WINCE sysname, and add
1994 INSTALL.WCE and appropriate conditionals to make it build.
1995 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1997 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1998 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1999 edit numbers of the version.
2000 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2002 *) Introduce safe string copy and catenation functions
2003 (BUF_strlcpy() and BUF_strlcat()).
2004 [Ben Laurie (CHATS) and Richard Levitte]
2006 *) Avoid using fixed-size buffers for one-line DNs.
2007 [Ben Laurie (CHATS)]
2009 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2010 resizing buffers containing secrets, and use where appropriate.
2011 [Ben Laurie (CHATS)]
2013 *) Avoid using fixed size buffers for configuration file location.
2014 [Ben Laurie (CHATS)]
2016 *) Avoid filename truncation for various CA files.
2017 [Ben Laurie (CHATS)]
2019 *) Use sizeof in preference to magic numbers.
2020 [Ben Laurie (CHATS)]
2022 *) Avoid filename truncation in cert requests.
2023 [Ben Laurie (CHATS)]
2025 *) Add assertions to check for (supposedly impossible) buffer
2027 [Ben Laurie (CHATS)]
2029 *) Don't cache truncated DNS entries in the local cache (this could
2030 potentially lead to a spoofing attack).
2031 [Ben Laurie (CHATS)]
2033 *) Fix various buffers to be large enough for hex/decimal
2034 representations in a platform independent manner.
2035 [Ben Laurie (CHATS)]
2037 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2038 resizing buffers containing secrets, and use where appropriate.
2039 [Ben Laurie (CHATS)]
2041 *) Add BIO_indent() to avoid much slightly worrying code to do
2043 [Ben Laurie (CHATS)]
2045 *) Convert sprintf()/BIO_puts() to BIO_printf().
2046 [Ben Laurie (CHATS)]
2048 *) buffer_gets() could terminate with the buffer only half
2050 [Ben Laurie (CHATS)]
2052 *) Add assertions to prevent user-supplied crypto functions from
2053 overflowing internal buffers by having large block sizes, etc.
2054 [Ben Laurie (CHATS)]
2056 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2058 [Ben Laurie (CHATS)]
2060 *) Eliminate unused copy of key in RC4.
2061 [Ben Laurie (CHATS)]
2063 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2064 [Ben Laurie (CHATS)]
2066 *) Fix off-by-one error in EGD path.
2067 [Ben Laurie (CHATS)]
2069 *) If RANDFILE path is too long, ignore instead of truncating.
2070 [Ben Laurie (CHATS)]
2072 *) Eliminate unused and incorrectly sized X.509 structure
2074 [Ben Laurie (CHATS)]
2076 *) Eliminate unused and dangerous function knumber().
2077 [Ben Laurie (CHATS)]
2079 *) Eliminate unused and dangerous structure, KSSL_ERR.
2080 [Ben Laurie (CHATS)]
2082 *) Protect against overlong session ID context length in an encoded
2083 session object. Since these are local, this does not appear to be
2085 [Ben Laurie (CHATS)]
2087 *) Change from security patch (see 0.9.6e below) that did not affect
2088 the 0.9.6 release series:
2090 Remote buffer overflow in SSL3 protocol - an attacker could
2091 supply an oversized master key in Kerberos-enabled versions.
2093 [Ben Laurie (CHATS)]
2095 *) Change the SSL kerb5 codes to match RFC 2712.
2098 *) Make -nameopt work fully for req and add -reqopt switch.
2099 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2101 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2102 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2104 *) Make sure tests can be performed even if the corresponding algorithms
2105 have been removed entirely. This was also the last step to make
2106 OpenSSL compilable with DJGPP under all reasonable conditions.
2107 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2109 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2110 to allow version independent disabling of normally unselected ciphers,
2111 which may be activated as a side-effect of selecting a single cipher.
2113 (E.g., cipher list string "RSA" enables ciphersuites that are left
2114 out of "ALL" because they do not provide symmetric encryption.
2115 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2116 [Lutz Jaenicke, Bodo Moeller]
2118 *) Add appropriate support for separate platform-dependent build
2119 directories. The recommended way to make a platform-dependent
2120 build directory is the following (tested on Linux), maybe with
2123 # Place yourself outside of the OpenSSL source tree. In
2124 # this example, the environment variable OPENSSL_SOURCE
2125 # is assumed to contain the absolute OpenSSL source directory.
2126 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2127 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2128 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2129 mkdir -p `dirname $F`
2130 ln -s $OPENSSL_SOURCE/$F $F
2133 To be absolutely sure not to disturb the source tree, a "make clean"
2134 is a good thing. If it isn't successfull, don't worry about it,
2135 it probably means the source directory is very clean.
2138 *) Make sure any ENGINE control commands make local copies of string
2139 pointers passed to them whenever necessary. Otherwise it is possible
2140 the caller may have overwritten (or deallocated) the original string
2141 data when a later ENGINE operation tries to use the stored values.
2142 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2144 *) Improve diagnostics in file reading and command-line digests.
2145 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2147 *) Add AES modes CFB and OFB to the object database. Correct an
2148 error in AES-CFB decryption.
2151 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2152 allows existing EVP_CIPHER_CTX structures to be reused after
2153 calling EVP_*Final(). This behaviour is used by encryption
2154 BIOs and some applications. This has the side effect that
2155 applications must explicitly clean up cipher contexts with
2156 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2159 *) Check the values of dna and dnb in bn_mul_recursive before calling
2160 bn_mul_comba (a non zero value means the a or b arrays do not contain
2161 n2 elements) and fallback to bn_mul_normal if either is not zero.
2164 *) Fix escaping of non-ASCII characters when using the -subj option
2165 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2168 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2169 form for "surname", serialNumber has no short form.
2170 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2171 therefore remove "mail" short name for "internet 7".
2172 The OID for unique identifiers in X509 certificates is
2173 x500UniqueIdentifier, not uniqueIdentifier.
2174 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2177 *) Add an "init" command to the ENGINE config module and auto initialize
2178 ENGINEs. Without any "init" command the ENGINE will be initialized
2179 after all ctrl commands have been executed on it. If init=1 the
2180 ENGINE is initailized at that point (ctrls before that point are run
2181 on the uninitialized ENGINE and after on the initialized one). If
2182 init=0 then the ENGINE will not be iniatialized at all.
2185 *) Fix the 'app_verify_callback' interface so that the user-defined
2186 argument is actually passed to the callback: In the
2187 SSL_CTX_set_cert_verify_callback() prototype, the callback
2188 declaration has been changed from
2191 int (*cb)(X509_STORE_CTX *,void *);
2192 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2193 i=s->ctx->app_verify_callback(&ctx)
2194 has been changed into
2195 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2197 To update applications using SSL_CTX_set_cert_verify_callback(),
2198 a dummy argument can be added to their callback functions.
2199 [D. K. Smetters <smetters@parc.xerox.com>]
2201 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2202 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2204 *) Add and OPENSSL_LOAD_CONF define which will cause
2205 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2206 This allows older applications to transparently support certain
2207 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2208 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2209 load the config file and OPENSSL_add_all_algorithms_conf() which will
2210 always load it have also been added.
2213 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2214 Adjust NIDs and EVP layer.
2215 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2217 *) Config modules support in openssl utility.
2219 Most commands now load modules from the config file,
2220 though in a few (such as version) this isn't done
2221 because it couldn't be used for anything.
2223 In the case of ca and req the config file used is
2224 the same as the utility itself: that is the -config
2225 command line option can be used to specify an
2229 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2230 use "openssl_conf" if filename is NULL use default openssl config file.
2233 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2234 config section name. Add a new flag to tolerate a missing config file
2235 and move code to CONF_modules_load_file().
2238 *) Support for crypto accelerator cards from Accelerated Encryption
2239 Processing, www.aep.ie. (Use engine 'aep')
2240 The support was copied from 0.9.6c [engine] and adapted/corrected
2241 to work with the new engine framework.
2242 [AEP Inc. and Richard Levitte]
2244 *) Support for SureWare crypto accelerator cards from Baltimore
2245 Technologies. (Use engine 'sureware')
2246 The support was copied from 0.9.6c [engine] and adapted
2247 to work with the new engine framework.
2250 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2251 make the newer ENGINE framework commands for the CHIL engine work.
2252 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2254 *) Make it possible to produce shared libraries on ReliantUNIX.
2255 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2257 *) Add the configuration target debug-linux-ppro.
2258 Make 'openssl rsa' use the general key loading routines
2259 implemented in apps.c, and make those routines able to
2260 handle the key format FORMAT_NETSCAPE and the variant
2262 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2264 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2265 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2267 *) Add -keyform to rsautl, and document -engine.
2268 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2270 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2271 BIO_R_NO_SUCH_FILE error code rather than the generic
2272 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2275 *) Add new functions
2277 ERR_peek_last_error_line
2278 ERR_peek_last_error_line_data.
2279 These are similar to
2282 ERR_peek_error_line_data,
2283 but report on the latest error recorded rather than the first one
2284 still in the error queue.
2285 [Ben Laurie, Bodo Moeller]
2287 *) default_algorithms option in ENGINE config module. This allows things
2289 default_algorithms = ALL
2290 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2293 *) Prelminary ENGINE config module.
2296 *) New experimental application configuration code.
2299 *) Change the AES code to follow the same name structure as all other
2300 symmetric ciphers, and behave the same way. Move everything to
2301 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2302 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2304 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2305 [Ben Laurie and Theo de Raadt]
2307 *) Add option to output public keys in req command.
2308 [Massimiliano Pala madwolf@openca.org]
2310 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2311 (up to about 10% better than before for P-192 and P-224).
2314 *) New functions/macros
2316 SSL_CTX_set_msg_callback(ctx, cb)
2317 SSL_CTX_set_msg_callback_arg(ctx, arg)
2318 SSL_set_msg_callback(ssl, cb)
2319 SSL_set_msg_callback_arg(ssl, arg)
2321 to request calling a callback function
2323 void cb(int write_p, int version, int content_type,
2324 const void *buf, size_t len, SSL *ssl, void *arg)
2326 whenever a protocol message has been completely received
2327 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2328 protocol version according to which the SSL library interprets
2329 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2330 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2331 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2332 specification (change_cipher_spec(20), alert(21), handshake(22)).
2333 'buf' and 'len' point to the actual message, 'ssl' to the
2334 SSL object, and 'arg' is the application-defined value set by
2335 SSL[_CTX]_set_msg_callback_arg().
2337 'openssl s_client' and 'openssl s_server' have new '-msg' options
2338 to enable a callback that displays all protocol messages.
2341 *) Change the shared library support so shared libraries are built as
2342 soon as the corresponding static library is finished, and thereby get
2343 openssl and the test programs linked against the shared library.
2344 This still only happens when the keyword "shard" has been given to
2345 the configuration scripts.
2347 NOTE: shared library support is still an experimental thing, and
2348 backward binary compatibility is still not guaranteed.
2349 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2351 *) Add support for Subject Information Access extension.
2352 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2354 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2355 additional bytes when new memory had to be allocated, not just
2356 when reusing an existing buffer.
2359 *) New command line and configuration option 'utf8' for the req command.
2360 This allows field values to be specified as UTF8 strings.
2363 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2364 runs for the former and machine-readable output for the latter.
2367 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2368 of the e-mail address in the DN (i.e., it will go into a certificate
2369 extension only). The new configuration file option 'email_in_dn = no'
2370 has the same effect.
2371 [Massimiliano Pala madwolf@openca.org]
2373 *) Change all functions with names starting with des_ to be starting
2374 with DES_ instead. Add wrappers that are compatible with libdes,
2375 but are named _ossl_old_des_*. Finally, add macros that map the
2376 des_* symbols to the corresponding _ossl_old_des_* if libdes
2377 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2378 desired, the des_* symbols will be mapped to DES_*, with one
2381 Since we provide two compatibility mappings, the user needs to
2382 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2383 compatibility is desired. The default (i.e., when that macro
2384 isn't defined) is OpenSSL 0.9.6c compatibility.
2386 There are also macros that enable and disable the support of old
2387 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2388 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2389 are defined, the default will apply: to support the old des routines.
2391 In either case, one must include openssl/des.h to get the correct
2392 definitions. Do not try to just include openssl/des_old.h, that
2395 NOTE: This is a major break of an old API into a new one. Software
2396 authors are encouraged to switch to the DES_ style functions. Some
2397 time in the future, des_old.h and the libdes compatibility functions
2398 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2399 default), and then completely removed.
2402 *) Test for certificates which contain unsupported critical extensions.
2403 If such a certificate is found during a verify operation it is
2404 rejected by default: this behaviour can be overridden by either
2405 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2406 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2407 X509_supported_extension() has also been added which returns 1 if a
2408 particular extension is supported.
2411 *) Modify the behaviour of EVP cipher functions in similar way to digests
2412 to retain compatibility with existing code.
2415 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2416 compatibility with existing code. In particular the 'ctx' parameter does
2417 not have to be to be initialized before the call to EVP_DigestInit() and
2418 it is tidied up after a call to EVP_DigestFinal(). New function
2419 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2420 EVP_MD_CTX_copy() changed to not require the destination to be
2421 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2422 requires the destination to be valid.
2424 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2425 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2428 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2429 so that complete 'Handshake' protocol structures are kept in memory
2430 instead of overwriting 'msg_type' and 'length' with 'body' data.
2433 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2434 [Massimo Santin via Richard Levitte]
2436 *) Major restructuring to the underlying ENGINE code. This includes
2437 reduction of linker bloat, separation of pure "ENGINE" manipulation
2438 (initialisation, etc) from functionality dealing with implementations
2439 of specific crypto iterfaces. This change also introduces integrated
2440 support for symmetric ciphers and digest implementations - so ENGINEs
2441 can now accelerate these by providing EVP_CIPHER and EVP_MD
2442 implementations of their own. This is detailed in crypto/engine/README
2443 as it couldn't be adequately described here. However, there are a few
2444 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2445 were changed in the original introduction of ENGINE code have now
2446 reverted back - the hooking from this code to ENGINE is now a good
2447 deal more passive and at run-time, operations deal directly with
2448 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2449 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2450 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2451 they were not being used by the framework as there is no concept of a
2452 BIGNUM_METHOD and they could not be generalised to the new
2453 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2454 ENGINE_cpy() has been removed as it cannot be consistently defined in
2458 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2461 *) Change mkdef.pl to sort symbols that get the same entry number,
2462 and make sure the automatically generated functions ERR_load_*
2463 become part of libeay.num as well.
2466 *) New function SSL_renegotiate_pending(). This returns true once
2467 renegotiation has been requested (either SSL_renegotiate() call
2468 or HelloRequest/ClientHello receveived from the peer) and becomes
2469 false once a handshake has been completed.
2470 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2471 sends a HelloRequest, but does not ensure that a handshake takes
2472 place. SSL_renegotiate_pending() is useful for checking if the
2473 client has followed the request.)
2476 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2477 By default, clients may request session resumption even during
2478 renegotiation (if session ID contexts permit); with this option,
2479 session resumption is possible only in the first handshake.
2481 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2482 more bits available for options that should not be part of
2483 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2486 *) Add some demos for certificate and certificate request creation.
2489 *) Make maximum certificate chain size accepted from the peer application
2490 settable (SSL*_get/set_max_cert_list()), as proposed by
2491 "Douglas E. Engert" <deengert@anl.gov>.
2494 *) Add support for shared libraries for Unixware-7
2495 (Boyd Lynn Gerber <gerberb@zenez.com>).
2498 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2499 be done prior to destruction. Use this to unload error strings from
2500 ENGINEs that load their own error strings. NB: This adds two new API
2501 functions to "get" and "set" this destroy handler in an ENGINE.
2504 *) Alter all existing ENGINE implementations (except "openssl" and
2505 "openbsd") to dynamically instantiate their own error strings. This
2506 makes them more flexible to be built both as statically-linked ENGINEs
2507 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2508 Also, add stub code to each that makes building them as self-contained
2509 shared-libraries easier (see README.ENGINE).
2512 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2513 implementations into applications that are completely implemented in
2514 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2515 commands that can be used to configure what shared-library to load and
2516 to control aspects of the way it is handled. Also, made an update to
2517 the README.ENGINE file that brings its information up-to-date and
2518 provides some information and instructions on the "dynamic" ENGINE
2519 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2522 *) Make it possible to unload ranges of ERR strings with a new
2523 "ERR_unload_strings" function.
2526 *) Add a copy() function to EVP_MD.
2529 *) Make EVP_MD routines take a context pointer instead of just the
2530 md_data void pointer.
2533 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2534 that the digest can only process a single chunk of data
2535 (typically because it is provided by a piece of
2536 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2537 is only going to provide a single chunk of data, and hence the
2538 framework needn't accumulate the data for oneshot drivers.
2541 *) As with "ERR", make it possible to replace the underlying "ex_data"
2542 functions. This change also alters the storage and management of global
2543 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2544 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2545 index counters. The API functions that use this state have been changed
2546 to take a "class_index" rather than pointers to the class's local STACK
2547 and counter, and there is now an API function to dynamically create new
2548 classes. This centralisation allows us to (a) plug a lot of the
2549 thread-safety problems that existed, and (b) makes it possible to clean
2550 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2551 such data would previously have always leaked in application code and
2552 workarounds were in place to make the memory debugging turn a blind eye
2553 to it. Application code that doesn't use this new function will still
2554 leak as before, but their memory debugging output will announce it now
2555 rather than letting it slide.
2557 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2558 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2559 has a return value to indicate success or failure.
2562 *) Make it possible to replace the underlying "ERR" functions such that the
2563 global state (2 LHASH tables and 2 locks) is only used by the "default"
2564 implementation. This change also adds two functions to "get" and "set"
2565 the implementation prior to it being automatically set the first time
2566 any other ERR function takes place. Ie. an application can call "get",
2567 pass the return value to a module it has just loaded, and that module
2568 can call its own "set" function using that value. This means the
2569 module's "ERR" operations will use (and modify) the error state in the
2570 application and not in its own statically linked copy of OpenSSL code.
2573 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2574 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2575 the operation, and provides a more encapsulated way for external code
2576 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2577 to use these functions rather than manually incrementing the counts.
2579 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2582 *) Add EVP test program.
2585 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2588 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2589 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2590 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2591 These allow a CRL to be built without having to access X509_CRL fields
2592 directly. Modify 'ca' application to use new functions.
2595 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2596 bug workarounds. Rollback attack detection is a security feature.
2597 The problem will only arise on OpenSSL servers when TLSv1 is not
2598 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2599 Software authors not wanting to support TLSv1 will have special reasons
2600 for their choice and can explicitly enable this option.
2601 [Bodo Moeller, Lutz Jaenicke]
2603 *) Rationalise EVP so it can be extended: don't include a union of
2604 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2605 (similar to those existing for EVP_CIPHER_CTX).
2610 EVP_MD_CTX_init(&md); /* new function call */
2611 EVP_DigestInit(&md, EVP_sha1());
2612 EVP_DigestUpdate(&md, in, len);
2613 EVP_DigestFinal(&md, out, NULL);
2614 EVP_MD_CTX_cleanup(&md); /* new function call */
2618 *) Make DES key schedule conform to the usual scheme, as well as
2619 correcting its structure. This means that calls to DES functions
2620 now have to pass a pointer to a des_key_schedule instead of a
2621 plain des_key_schedule (which was actually always a pointer
2624 des_key_schedule ks;
2626 des_set_key_checked(..., &ks);
2627 des_ncbc_encrypt(..., &ks, ...);
2629 (Note that a later change renames 'des_...' into 'DES_...'.)
2632 *) Initial reduction of linker bloat: the use of some functions, such as
2633 PEM causes large amounts of unused functions to be linked in due to
2634 poor organisation. For example pem_all.c contains every PEM function
2635 which has a knock on effect of linking in large amounts of (unused)
2636 ASN1 code. Grouping together similar functions and splitting unrelated
2637 functions prevents this.
2640 *) Cleanup of EVP macros.
2643 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2644 correct _ecb suffix.
2647 *) Add initial OCSP responder support to ocsp application. The
2648 revocation information is handled using the text based index
2649 use by the ca application. The responder can either handle
2650 requests generated internally, supplied in files (for example
2651 via a CGI script) or using an internal minimal server.
2654 *) Add configuration choices to get zlib compression for TLS.
2657 *) Changes to Kerberos SSL for RFC 2712 compliance:
2658 1. Implemented real KerberosWrapper, instead of just using
2659 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2660 2. Implemented optional authenticator field of KerberosWrapper.
2662 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2663 and authenticator structs; see crypto/krb5/.
2665 Generalized Kerberos calls to support multiple Kerberos libraries.
2666 [Vern Staats <staatsvr@asc.hpc.mil>,
2667 Jeffrey Altman <jaltman@columbia.edu>
2668 via Richard Levitte]
2670 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2671 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2672 values for each of the key sizes rather than having just
2673 parameters (and 'speed' generating keys each time).
2676 *) Speed up EVP routines.
2679 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2680 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2681 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2682 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2684 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2685 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2686 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2689 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2691 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2694 *) Added the OS2-EMX target.
2695 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2697 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2698 to support NCONF routines in extension code. New function CONF_set_nconf()
2699 to allow functions which take an NCONF to also handle the old LHASH
2700 structure: this means that the old CONF compatible routines can be
2701 retained (in particular wrt extensions) without having to duplicate the
2702 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2705 *) Enhance the general user interface with mechanisms for inner control
2706 and with possibilities to have yes/no kind of prompts.
2709 *) Change all calls to low level digest routines in the library and
2710 applications to use EVP. Add missing calls to HMAC_cleanup() and
2711 don't assume HMAC_CTX can be copied using memcpy().
2712 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2714 *) Add the possibility to control engines through control names but with
2715 arbitrary arguments instead of just a string.
2716 Change the key loaders to take a UI_METHOD instead of a callback
2717 function pointer. NOTE: this breaks binary compatibility with earlier
2718 versions of OpenSSL [engine].
2719 Adapt the nCipher code for these new conditions and add a card insertion
2723 *) Enhance the general user interface with mechanisms to better support
2724 dialog box interfaces, application-defined prompts, the possibility
2725 to use defaults (for example default passwords from somewhere else)
2726 and interrupts/cancellations.
2729 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2730 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2733 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2734 tidy up some unnecessarily weird code in 'sk_new()').
2735 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2737 *) Change the key loading routines for ENGINEs to use the same kind
2738 callback (pem_password_cb) as all other routines that need this
2742 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2743 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2744 than this minimum value is recommended.
2747 *) New random seeder for OpenVMS, using the system process statistics
2748 that are easily reachable.
2751 *) Windows apparently can't transparently handle global
2752 variables defined in DLLs. Initialisations such as:
2754 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2756 wont compile. This is used by the any applications that need to
2757 declare their own ASN1 modules. This was fixed by adding the option
2758 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2759 needed for static libraries under Win32.
2762 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2763 setting of purpose and trust fields. New X509_STORE trust and
2764 purpose functions and tidy up setting in other SSL functions.
2767 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2768 structure. These are inherited by X509_STORE_CTX when it is
2769 initialised. This allows various defaults to be set in the
2770 X509_STORE structure (such as flags for CRL checking and custom
2771 purpose or trust settings) for functions which only use X509_STORE_CTX
2772 internally such as S/MIME.
2774 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2775 trust settings if they are not set in X509_STORE. This allows X509_STORE
2776 purposes and trust (in S/MIME for example) to override any set by default.
2778 Add command line options for CRL checking to smime, s_client and s_server
2782 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2783 are set then the CRL is looked up in the X509_STORE structure and
2784 its validity and signature checked, then if the certificate is found
2785 in the CRL the verify fails with a revoked error.
2787 Various new CRL related callbacks added to X509_STORE_CTX structure.
2789 Command line options added to 'verify' application to support this.
2791 This needs some additional work, such as being able to handle multiple
2792 CRLs with different times, extension based lookup (rather than just
2793 by subject name) and ultimately more complete V2 CRL extension
2797 *) Add a general user interface API (crypto/ui/). This is designed
2798 to replace things like des_read_password and friends (backward
2799 compatibility functions using this new API are provided).
2800 The purpose is to remove prompting functions from the DES code
2801 section as well as provide for prompting through dialog boxes in
2802 a window system and the like.
2805 *) Add "ex_data" support to ENGINE so implementations can add state at a
2806 per-structure level rather than having to store it globally.
2809 *) Make it possible for ENGINE structures to be copied when retrieved by
2810 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2811 This causes the "original" ENGINE structure to act like a template,
2812 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2813 operational state can be localised to each ENGINE structure, despite the
2814 fact they all share the same "methods". New ENGINE structures returned in
2815 this case have no functional references and the return value is the single
2816 structural reference. This matches the single structural reference returned
2817 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2821 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2822 needs to match any other type at all we need to manually clear the
2826 *) Changes to the "openssl engine" utility to include;
2827 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2828 about an ENGINE's available control commands.
2829 - executing control commands from command line arguments using the
2830 '-pre' and '-post' switches. '-post' is only used if '-t' is
2831 specified and the ENGINE is successfully initialised. The syntax for
2832 the individual commands are colon-separated, for example;
2833 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2836 *) New dynamic control command support for ENGINEs. ENGINEs can now
2837 declare their own commands (numbers), names (strings), descriptions,
2838 and input types for run-time discovery by calling applications. A
2839 subset of these commands are implicitly classed as "executable"
2840 depending on their input type, and only these can be invoked through
2841 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2842 can be based on user input, config files, etc). The distinction is
2843 that "executable" commands cannot return anything other than a boolean
2844 result and can only support numeric or string input, whereas some
2845 discoverable commands may only be for direct use through
2846 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2847 pointers, or other custom uses. The "executable" commands are to
2848 support parameterisations of ENGINE behaviour that can be
2849 unambiguously defined by ENGINEs and used consistently across any
2850 OpenSSL-based application. Commands have been added to all the
2851 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2852 control over shared-library paths without source code alterations.
2855 *) Changed all ENGINE implementations to dynamically allocate their
2856 ENGINEs rather than declaring them statically. Apart from this being
2857 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2858 this also allows the implementations to compile without using the
2859 internal engine_int.h header.
2862 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2863 'const' value. Any code that should be able to modify a RAND_METHOD
2864 should already have non-const pointers to it (ie. they should only
2865 modify their own ones).
2868 *) Made a variety of little tweaks to the ENGINE code.
2869 - "atalla" and "ubsec" string definitions were moved from header files
2870 to C code. "nuron" string definitions were placed in variables
2871 rather than hard-coded - allowing parameterisation of these values
2872 later on via ctrl() commands.
2873 - Removed unused "#if 0"'d code.
2874 - Fixed engine list iteration code so it uses ENGINE_free() to release
2875 structural references.
2876 - Constified the RAND_METHOD element of ENGINE structures.
2877 - Constified various get/set functions as appropriate and added
2878 missing functions (including a catch-all ENGINE_cpy that duplicates
2879 all ENGINE values onto a new ENGINE except reference counts/state).
2880 - Removed NULL parameter checks in get/set functions. Setting a method
2881 or function to NULL is a way of cancelling out a previously set
2882 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2883 and doesn't justify the extra error symbols and code.
2884 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2885 flags from engine_int.h to engine.h.
2886 - Changed prototypes for ENGINE handler functions (init(), finish(),
2887 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2890 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2891 to the algorithm using long division. The binary algorithm can be
2892 used only if the modulus is odd. On 32-bit systems, it is faster
2893 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2894 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2895 up to 450 bits. In 64-bit environments, the binary algorithm
2896 appears to be advantageous for much longer moduli; here we use it
2897 for moduli up to 2048 bits.
2900 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2901 could not support the combine flag in choice fields.
2904 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2905 extensions from a certificate request to the certificate.
2908 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2909 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2910 file: this allows the display of the certificate about to be
2911 signed to be customised, to allow certain fields to be included
2912 or excluded and extension details. The old system didn't display
2913 multicharacter strings properly, omitted fields not in the policy
2914 and couldn't display additional details such as extensions.
2917 *) Function EC_POINTs_mul for multiple scalar multiplication
2918 of an arbitrary number of elliptic curve points
2919 \sum scalars[i]*points[i],
2920 optionally including the generator defined for the EC_GROUP:
2921 scalar*generator + \sum scalars[i]*points[i].
2923 EC_POINT_mul is a simple wrapper function for the typical case
2924 that the point list has just one item (besides the optional
2928 *) First EC_METHODs for curves over GF(p):
2930 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2931 operations and provides various method functions that can also
2932 operate with faster implementations of modular arithmetic.
2934 EC_GFp_mont_method() reuses most functions that are part of
2935 EC_GFp_simple_method, but uses Montgomery arithmetic.
2937 [Bodo Moeller; point addition and point doubling
2938 implementation directly derived from source code provided by
2939 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2941 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2942 crypto/ec/ec_lib.c):
2944 Curves are EC_GROUP objects (with an optional group generator)
2945 based on EC_METHODs that are built into the library.
2947 Points are EC_POINT objects based on EC_GROUP objects.
2949 Most of the framework would be able to handle curves over arbitrary
2950 finite fields, but as there are no obvious types for fields other
2951 than GF(p), some functions are limited to that for now.
2954 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2955 that the file contains a complete HTTP response.
2958 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2959 change the def and num file printf format specifier from "%-40sXXX"
2960 to "%-39s XXX". The latter will always guarantee a space after the
2961 field while the former will cause them to run together if the field
2962 is 40 of more characters long.
2965 *) Constify the cipher and digest 'method' functions and structures
2966 and modify related functions to take constant EVP_MD and EVP_CIPHER
2970 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2971 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2974 *) Modify EVP_Digest*() routines so they now return values. Although the
2975 internal software routines can never fail additional hardware versions
2979 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2981 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2982 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2985 ERR_R_NESTED_ASN1_ERROR
2987 ERR_R_MISSING_ASN1_EOS
2988 were 4 .. 9, conflicting with
2989 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2991 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2992 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2994 Add new error code 'ERR_R_INTERNAL_ERROR'.
2997 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3001 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3002 sets the subject name for a new request or supersedes the
3003 subject name in a given request. Formats that can be parsed are
3004 'CN=Some Name, OU=myOU, C=IT'
3006 'CN=Some Name/OU=myOU/C=IT'.
3008 Add options '-batch' and '-verbose' to 'openssl req'.
3009 [Massimiliano Pala <madwolf@hackmasters.net>]
3011 *) Introduce the possibility to access global variables through
3012 functions on platform were that's the best way to handle exporting
3013 global variables in shared libraries. To enable this functionality,
3014 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3015 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3016 is normally done by Configure or something similar).
3018 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3019 in the source file (foo.c) like this:
3021 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3022 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3024 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3025 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3027 OPENSSL_DECLARE_GLOBAL(int,foo);
3028 #define foo OPENSSL_GLOBAL_REF(foo)
3029 OPENSSL_DECLARE_GLOBAL(double,bar);
3030 #define bar OPENSSL_GLOBAL_REF(bar)
3032 The #defines are very important, and therefore so is including the
3033 header file everywhere where the defined globals are used.
3035 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3036 of ASN.1 items, but that structure is a bit different.
3038 The largest change is in util/mkdef.pl which has been enhanced with
3039 better and easier to understand logic to choose which symbols should
3040 go into the Windows .def files as well as a number of fixes and code
3041 cleanup (among others, algorithm keywords are now sorted
3042 lexicographically to avoid constant rewrites).
3045 *) In BN_div() keep a copy of the sign of 'num' before writing the
3046 result to 'rm' because if rm==num the value will be overwritten
3047 and produce the wrong result if 'num' is negative: this caused
3048 problems with BN_mod() and BN_nnmod().
3051 *) Function OCSP_request_verify(). This checks the signature on an
3052 OCSP request and verifies the signer certificate. The signer
3053 certificate is just checked for a generic purpose and OCSP request
3057 *) Add OCSP_check_validity() function to check the validity of OCSP
3058 responses. OCSP responses are prepared in real time and may only
3059 be a few seconds old. Simply checking that the current time lies
3060 between thisUpdate and nextUpdate max reject otherwise valid responses
3061 caused by either OCSP responder or client clock inaccuracy. Instead
3062 we allow thisUpdate and nextUpdate to fall within a certain period of
3063 the current time. The age of the response can also optionally be
3064 checked. Two new options -validity_period and -status_age added to
3068 *) If signature or public key algorithm is unrecognized print out its
3069 OID rather that just UNKNOWN.
3072 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3073 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3074 ID to be generated from the issuer certificate alone which can then be
3075 passed to OCSP_id_issuer_cmp().
3078 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3079 ASN1 modules to export functions returning ASN1_ITEM pointers
3080 instead of the ASN1_ITEM structures themselves. This adds several
3081 new macros which allow the underlying ASN1 function/structure to
3082 be accessed transparently. As a result code should not use ASN1_ITEM
3083 references directly (such as &X509_it) but instead use the relevant
3084 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3085 use of the new ASN1 code on platforms where exporting structures
3086 is problematical (for example in shared libraries) but exporting
3087 functions returning pointers to structures is not.
3090 *) Add support for overriding the generation of SSL/TLS session IDs.
3091 These callbacks can be registered either in an SSL_CTX or per SSL.
3092 The purpose of this is to allow applications to control, if they wish,
3093 the arbitrary values chosen for use as session IDs, particularly as it
3094 can be useful for session caching in multiple-server environments. A
3095 command-line switch for testing this (and any client code that wishes
3096 to use such a feature) has been added to "s_server".
3097 [Geoff Thorpe, Lutz Jaenicke]
3099 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3100 of the form '#if defined(...) || defined(...) || ...' and
3101 '#if !defined(...) && !defined(...) && ...'. This also avoids
3102 the growing number of special cases it was previously handling.
3105 *) Make all configuration macros available for application by making
3106 sure they are available in opensslconf.h, by giving them names starting
3107 with "OPENSSL_" to avoid conflicts with other packages and by making
3108 sure e_os2.h will cover all platform-specific cases together with
3110 Additionally, it is now possible to define configuration/platform-
3111 specific names (called "system identities"). In the C code, these
3112 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3113 macro with the name beginning with "OPENSSL_SYS_", which is determined
3114 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3118 *) New option -set_serial to 'req' and 'x509' this allows the serial
3119 number to use to be specified on the command line. Previously self
3120 signed certificates were hard coded with serial number 0 and the
3121 CA options of 'x509' had to use a serial number in a file which was
3125 *) New options to 'ca' utility to support V2 CRL entry extensions.
3126 Currently CRL reason, invalidity date and hold instruction are
3127 supported. Add new CRL extensions to V3 code and some new objects.
3130 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3131 disable standard block padding (aka PKCS#5 padding) in the EVP
3132 API, which was previously mandatory. This means that the data is
3133 not padded in any way and so the total length much be a multiple
3134 of the block size, otherwise an error occurs.
3137 *) Initial (incomplete) OCSP SSL support.
3140 *) New function OCSP_parse_url(). This splits up a URL into its host,
3141 port and path components: primarily to parse OCSP URLs. New -url
3142 option to ocsp utility.
3145 *) New nonce behavior. The return value of OCSP_check_nonce() now
3146 reflects the various checks performed. Applications can decide
3147 whether to tolerate certain situations such as an absent nonce
3148 in a response when one was present in a request: the ocsp application
3149 just prints out a warning. New function OCSP_add1_basic_nonce()
3150 this is to allow responders to include a nonce in a response even if
3151 the request is nonce-less.
3154 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3155 skipped when using openssl x509 multiple times on a single input file,
3156 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3159 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3160 set string type: to handle setting ASN1_TIME structures. Fix ca
3161 utility to correctly initialize revocation date of CRLs.
3164 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3165 the clients preferred ciphersuites and rather use its own preferences.
3166 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3167 Internet Explorer by ensuring unchanged hash method during stepup.
3168 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3171 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3172 to aes and add a new 'exist' option to print out symbols that don't
3176 *) Additional options to ocsp utility to allow flags to be set and
3177 additional certificates supplied.
3180 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3181 OCSP client a number of certificate to only verify the response
3185 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3186 handle the new API. Currently only ECB, CBC modes supported. Add new
3189 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3190 Encryption Standard (AES) Ciphersuites for Transport Layer
3191 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3192 not enabled by default and were not part of the "ALL" ciphersuite
3193 alias because they were not yet official; they could be
3194 explicitly requested by specifying the "AESdraft" ciphersuite
3195 group alias. In the final release of OpenSSL 0.9.7, the group
3196 alias is called "AES" and is part of "ALL".)
3197 [Ben Laurie, Steve Henson, Bodo Moeller]
3199 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3200 request to response.
3203 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3204 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3205 extract information from a certificate request. OCSP_response_create()
3206 creates a response and optionally adds a basic response structure.
3207 OCSP_basic_add1_status() adds a complete single response to a basic
3208 response and returns the OCSP_SINGLERESP structure just added (to allow
3209 extensions to be included for example). OCSP_basic_add1_cert() adds a
3210 certificate to a basic response and OCSP_basic_sign() signs a basic
3211 response with various flags. New helper functions ASN1_TIME_check()
3212 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3213 (converts ASN1_TIME to GeneralizedTime).
3216 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3217 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3218 structure from a certificate. X509_pubkey_digest() digests the public_key
3219 contents: this is used in various key identifiers.
3222 *) Make sk_sort() tolerate a NULL argument.
3223 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3225 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3226 passed by the function are trusted implicitly. If any of them signed the
3227 response then it is assumed to be valid and is not verified.
3230 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3231 to data. This was previously part of the PKCS7 ASN1 code. This
3232 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3233 [Steve Henson, reported by Kenneth R. Robinette
3234 <support@securenetterm.com>]
3236 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3237 routines: without these tracing memory leaks is very painful.
3238 Fix leaks in PKCS12 and PKCS7 routines.
3241 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3242 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3243 effectively meant GeneralizedTime would never be used. Now it
3244 is initialised to -1 but X509_time_adj() now has to check the value
3245 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3246 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3247 [Steve Henson, reported by Kenneth R. Robinette
3248 <support@securenetterm.com>]
3250 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3251 result in a zero length in the ASN1_INTEGER structure which was
3252 not consistent with the structure when d2i_ASN1_INTEGER() was used
3253 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3254 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3255 where it did not print out a minus for negative ASN1_INTEGER.
3258 *) Add summary printout to ocsp utility. The various functions which
3259 convert status values to strings have been renamed to:
3260 OCSP_response_status_str(), OCSP_cert_status_str() and
3261 OCSP_crl_reason_str() and are no longer static. New options
3262 to verify nonce values and to disable verification. OCSP response
3263 printout format cleaned up.
3266 *) Add additional OCSP certificate checks. These are those specified
3267 in RFC2560. This consists of two separate checks: the CA of the
3268 certificate being checked must either be the OCSP signer certificate
3269 or the issuer of the OCSP signer certificate. In the latter case the
3270 OCSP signer certificate must contain the OCSP signing extended key
3271 usage. This check is performed by attempting to match the OCSP
3272 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3273 in the OCSP_CERTID structures of the response.
3276 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3277 and related routines. This uses the standard OpenSSL certificate
3278 verify routines to perform initial checks (just CA validity) and
3279 to obtain the certificate chain. Then additional checks will be
3280 performed on the chain. Currently the root CA is checked to see
3281 if it is explicitly trusted for OCSP signing. This is used to set
3282 a root CA as a global signing root: that is any certificate that
3283 chains to that CA is an acceptable OCSP signing certificate.
3286 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3287 extensions from a separate configuration file.
3288 As when reading extensions from the main configuration file,
3289 the '-extensions ...' option may be used for specifying the
3291 [Massimiliano Pala <madwolf@comune.modena.it>]
3293 *) New OCSP utility. Allows OCSP requests to be generated or
3294 read. The request can be sent to a responder and the output
3295 parsed, outputed or printed in text form. Not complete yet:
3296 still needs to check the OCSP response validity.
3299 *) New subcommands for 'openssl ca':
3300 'openssl ca -status <serial>' prints the status of the cert with
3301 the given serial number (according to the index file).
3302 'openssl ca -updatedb' updates the expiry status of certificates
3304 [Massimiliano Pala <madwolf@comune.modena.it>]
3306 *) New '-newreq-nodes' command option to CA.pl. This is like
3307 '-newreq', but calls 'openssl req' with the '-nodes' option
3308 so that the resulting key is not encrypted.
3309 [Damien Miller <djm@mindrot.org>]
3311 *) New configuration for the GNU Hurd.
3312 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3314 *) Initial code to implement OCSP basic response verify. This
3315 is currently incomplete. Currently just finds the signer's
3316 certificate and verifies the signature on the response.
3319 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3320 value of OPENSSLDIR. This is available via the new '-d' option
3321 to 'openssl version', and is also included in 'openssl version -a'.
3324 *) Allowing defining memory allocation callbacks that will be given
3325 file name and line number information in additional arguments
3326 (a const char* and an int). The basic functionality remains, as
3327 well as the original possibility to just replace malloc(),
3328 realloc() and free() by functions that do not know about these
3329 additional arguments. To register and find out the current
3330 settings for extended allocation functions, the following
3331 functions are provided:
3333 CRYPTO_set_mem_ex_functions
3334 CRYPTO_set_locked_mem_ex_functions
3335 CRYPTO_get_mem_ex_functions
3336 CRYPTO_get_locked_mem_ex_functions
3338 These work the same way as CRYPTO_set_mem_functions and friends.
3339 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3340 extended allocation function is enabled.
3341 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3342 a conventional allocation function is enabled.
3343 [Richard Levitte, Bodo Moeller]
3345 *) Finish off removing the remaining LHASH function pointer casts.
3346 There should no longer be any prototype-casting required when using
3347 the LHASH abstraction, and any casts that remain are "bugs". See
3348 the callback types and macros at the head of lhash.h for details
3349 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3352 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3353 If /dev/[u]random devices are not available or do not return enough
3354 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3356 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3357 /etc/entropy will be queried once each in this sequence, quering stops
3358 when enough entropy was collected without querying more sockets.
3361 *) Change the Unix RAND_poll() variant to be able to poll several
3362 random devices, as specified by DEVRANDOM, until a sufficient amount
3363 of data has been collected. We spend at most 10 ms on each file
3364 (select timeout) and read in non-blocking mode. DEVRANDOM now
3365 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3366 (previously it was just the string "/dev/urandom"), so on typical
3367 platforms the 10 ms delay will never occur.
3368 Also separate out the Unix variant to its own file, rand_unix.c.
3369 For VMS, there's a currently-empty rand_vms.c.
3372 *) Move OCSP client related routines to ocsp_cl.c. These
3373 provide utility functions which an application needing
3374 to issue a request to an OCSP responder and analyse the
3375 response will typically need: as opposed to those which an
3376 OCSP responder itself would need which will be added later.
3378 OCSP_request_sign() signs an OCSP request with an API similar
3379 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3380 response. OCSP_response_get1_basic() extracts basic response
3381 from response. OCSP_resp_find_status(): finds and extracts status
3382 information from an OCSP_CERTID structure (which will be created
3383 when the request structure is built). These are built from lower
3384 level functions which work on OCSP_SINGLERESP structures but
3385 wont normally be used unless the application wishes to examine
3386 extensions in the OCSP response for example.
3388 Replace nonce routines with a pair of functions.
3389 OCSP_request_add1_nonce() adds a nonce value and optionally
3390 generates a random value. OCSP_check_nonce() checks the
3391 validity of the nonce in an OCSP response.
3394 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3395 This doesn't copy the supplied OCSP_CERTID and avoids the
3396 need to free up the newly created id. Change return type
3397 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3398 This can then be used to add extensions to the request.
3399 Deleted OCSP_request_new(), since most of its functionality
3400 is now in OCSP_REQUEST_new() (and the case insensitive name
3401 clash) apart from the ability to set the request name which
3402 will be added elsewhere.
3405 *) Update OCSP API. Remove obsolete extensions argument from
3406 various functions. Extensions are now handled using the new
3407 OCSP extension code. New simple OCSP HTTP function which
3408 can be used to send requests and parse the response.
3411 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3412 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3413 uses the special reorder version of SET OF to sort the attributes
3414 and reorder them to match the encoded order. This resolves a long
3415 standing problem: a verify on a PKCS7 structure just after signing
3416 it used to fail because the attribute order did not match the
3417 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3418 it uses the received order. This is necessary to tolerate some broken
3419 software that does not order SET OF. This is handled by encoding
3420 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3421 to produce the required SET OF.
3424 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3425 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3426 files to get correct declarations of the ASN.1 item variables.
3429 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3430 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3431 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3432 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3433 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3434 ASN1_ITEM and no wrapper functions.
3437 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3438 replace the old function pointer based I/O routines. Change most of
3439 the *_d2i_bio() and *_d2i_fp() functions to use these.
3442 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3443 lines, recognice more "algorithms" that can be deselected, and make
3444 it complain about algorithm deselection that isn't recognised.
3447 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3448 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3449 to use new functions. Add NO_ASN1_OLD which can be set to remove
3450 some old style ASN1 functions: this can be used to determine if old
3451 code will still work when these eventually go away.
3454 *) New extension functions for OCSP structures, these follow the
3455 same conventions as certificates and CRLs.
3458 *) New function X509V3_add1_i2d(). This automatically encodes and
3459 adds an extension. Its behaviour can be customised with various
3460 flags to append, replace or delete. Various wrappers added for
3461 certifcates and CRLs.
3464 *) Fix to avoid calling the underlying ASN1 print routine when
3465 an extension cannot be parsed. Correct a typo in the
3466 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3469 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3470 entries for variables.
3473 *) Add functionality to apps/openssl.c for detecting locking
3474 problems: As the program is single-threaded, all we have
3475 to do is register a locking callback using an array for
3476 storing which locks are currently held by the program.
3479 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3480 SSL_get_ex_data_X509_STORE_idx(), which is used in
3481 ssl_verify_cert_chain() and thus can be called at any time
3482 during TLS/SSL handshakes so that thread-safety is essential.
3483 Unfortunately, the ex_data design is not at all suited
3484 for multi-threaded use, so it probably should be abolished.
3487 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3488 [Broadcom, tweaked and integrated by Geoff Thorpe]
3490 *) Move common extension printing code to new function
3491 X509V3_print_extensions(). Reorganise OCSP print routines and
3492 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3495 *) New function X509_signature_print() to remove duplication in some
3499 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3500 set (this was treated exactly the same as SET OF previously). This
3501 is used to reorder the STACK representing the structure to match the
3502 encoding. This will be used to get round a problem where a PKCS7
3503 structure which was signed could not be verified because the STACK
3504 order did not reflect the encoded order.
3507 *) Reimplement the OCSP ASN1 module using the new code.
3510 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3511 for its ASN1 operations. The old style function pointers still exist
3512 for now but they will eventually go away.
3515 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3516 completely replaces the old ASN1 functionality with a table driven
3517 encoder and decoder which interprets an ASN1_ITEM structure describing
3518 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3519 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3520 has also been converted to the new form.
3523 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3524 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3525 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3526 for negative moduli.
3529 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3530 of not touching the result's sign bit.
3533 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3537 *) Changed the LHASH code to use prototypes for callbacks, and created
3538 macros to declare and implement thin (optionally static) functions
3539 that provide type-safety and avoid function pointer casting for the
3540 type-specific callbacks.
3543 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3545 [Veers Staats <staatsvr@asc.hpc.mil>,
3546 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3548 *) Reformat the FAQ so the different questions and answers can be divided
3549 in sections depending on the subject.
3552 *) Have the zlib compression code load ZLIB.DLL dynamically under
3556 *) New function BN_mod_sqrt for computing square roots modulo a prime
3557 (using the probabilistic Tonelli-Shanks algorithm unless
3558 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3559 be handled deterministically).
3560 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3562 *) Make BN_mod_inverse faster by explicitly handling small quotients
3563 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3564 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3567 *) New function BN_kronecker.
3570 *) Fix BN_gcd so that it works on negative inputs; the result is
3571 positive unless both parameters are zero.
3572 Previously something reasonably close to an infinite loop was
3573 possible because numbers could be growing instead of shrinking
3574 in the implementation of Euclid's algorithm.
3577 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3578 sign of the number in question.
3580 Fix BN_is_word(a,w) to work correctly for w == 0.
3582 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3583 because its test if the absolute value of 'a' equals 'w'.
3584 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3585 it exists mostly for use in the implementations of BN_is_zero(),
3586 BN_is_one(), and BN_is_word().
3589 *) New function BN_swap.
3592 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3593 the exponentiation functions are more likely to produce reasonable
3594 results on negative inputs.
3597 *) Change BN_mod_mul so that the result is always non-negative.
3598 Previously, it could be negative if one of the factors was negative;
3599 I don't think anyone really wanted that behaviour.
3602 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3603 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3604 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3605 and add new functions:
3614 BN_mod_lshift1_quick
3618 These functions always generate non-negative results.
3620 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3621 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3623 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3624 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3625 be reduced modulo m.
3626 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3629 The following entry accidentily appeared in the CHANGES file
3630 distributed with OpenSSL 0.9.7. The modifications described in
3631 it do *not* apply to OpenSSL 0.9.7.
3633 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3634 was actually never needed) and in BN_mul(). The removal in BN_mul()
3635 required a small change in bn_mul_part_recursive() and the addition
3636 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3637 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3638 bn_sub_words() and bn_add_words() except they take arrays with
3643 *) In 'openssl passwd', verify passwords read from the terminal
3644 unless the '-salt' option is used (which usually means that
3645 verification would just waste user's time since the resulting
3646 hash is going to be compared with some given password hash)
3647 or the new '-noverify' option is used.
3649 This is an incompatible change, but it does not affect
3650 non-interactive use of 'openssl passwd' (passwords on the command
3651 line, '-stdin' option, '-in ...' option) and thus should not
3655 *) Remove all references to RSAref, since there's no more need for it.
3658 *) Make DSO load along a path given through an environment variable
3659 (SHLIB_PATH) with shl_load().
3662 *) Constify the ENGINE code as a result of BIGNUM constification.
3663 Also constify the RSA code and most things related to it. In a
3664 few places, most notable in the depth of the ASN.1 code, ugly
3665 casts back to non-const were required (to be solved at a later
3669 *) Make it so the openssl application has all engines loaded by default.
3672 *) Constify the BIGNUM routines a little more.
3675 *) Add the following functions:
3677 ENGINE_load_cswift()
3679 ENGINE_load_atalla()
3681 ENGINE_load_builtin_engines()
3683 That way, an application can itself choose if external engines that
3684 are built-in in OpenSSL shall ever be used or not. The benefit is
3685 that applications won't have to be linked with libdl or other dso
3686 libraries unless it's really needed.
3688 Changed 'openssl engine' to load all engines on demand.
3689 Changed the engine header files to avoid the duplication of some
3690 declarations (they differed!).
3693 *) 'openssl engine' can now list capabilities.
3696 *) Better error reporting in 'openssl engine'.
3699 *) Never call load_dh_param(NULL) in s_server.
3702 *) Add engine application. It can currently list engines by name and
3703 identity, and test if they are actually available.
3706 *) Improve RPM specification file by forcing symbolic linking and making
3707 sure the installed documentation is also owned by root.root.
3708 [Damien Miller <djm@mindrot.org>]
3710 *) Give the OpenSSL applications more possibilities to make use of
3711 keys (public as well as private) handled by engines.
3714 *) Add OCSP code that comes from CertCo.
3717 *) Add VMS support for the Rijndael code.
3720 *) Added untested support for Nuron crypto accelerator.
3723 *) Add support for external cryptographic devices. This code was
3724 previously distributed separately as the "engine" branch.
3725 [Geoff Thorpe, Richard Levitte]
3727 *) Rework the filename-translation in the DSO code. It is now possible to
3728 have far greater control over how a "name" is turned into a filename
3729 depending on the operating environment and any oddities about the
3730 different shared library filenames on each system.
3733 *) Support threads on FreeBSD-elf in Configure.
3736 *) Fix for SHA1 assembly problem with MASM: it produces
3737 warnings about corrupt line number information when assembling
3738 with debugging information. This is caused by the overlapping
3740 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3743 NCONF_get_number() has no error checking at all. As a replacement,
3744 NCONF_get_number_e() is defined (_e for "error checking") and is
3745 promoted strongly. The old NCONF_get_number is kept around for
3746 binary backward compatibility.
3747 Make it possible for methods to load from something other than a BIO,
3748 by providing a function pointer that is given a name instead of a BIO.
3749 For example, this could be used to load configuration data from an
3753 *) Fix for non blocking accept BIOs. Added new I/O special reason
3754 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3755 with non blocking I/O was not possible because no retry code was
3756 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3760 *) Added the beginnings of Rijndael support.
3763 *) Fix for bug in DirectoryString mask setting. Add support for
3764 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3765 to allow certificate printing to more controllable, additional
3766 'certopt' option to 'x509' to allow new printing options to be
3770 *) Clean old EAY MD5 hack from e_os.h.
3773 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3775 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3776 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3777 [Joe Orton, Steve Henson]
3779 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3781 *) Fix additional bug revealed by the NISCC test suite:
3783 Stop bug triggering large recursion when presented with
3784 certain ASN.1 tags (CVE-2003-0851)
3787 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3789 *) Fix various bugs revealed by running the NISCC test suite:
3791 Stop out of bounds reads in the ASN1 code when presented with
3792 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3794 If verify callback ignores invalid public key errors don't try to check
3795 certificate signature with the NULL public key.
3799 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3800 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3804 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3805 extra data after the compression methods not only for TLS 1.0
3806 but also for SSL 3.0 (as required by the specification).
3807 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3809 *) Change X509_certificate_type() to mark the key as exported/exportable
3810 when it's 512 *bits* long, not 512 bytes.
3813 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3815 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3816 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3817 a protocol version number mismatch like a decryption error
3818 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3821 *) Turn on RSA blinding by default in the default implementation
3822 to avoid a timing attack. Applications that don't want it can call
3823 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3824 They would be ill-advised to do so in most cases.
3825 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3827 *) Change RSA blinding code so that it works when the PRNG is not
3828 seeded (in this case, the secret RSA exponent is abused as
3829 an unpredictable seed -- if it is not unpredictable, there
3830 is no point in blinding anyway). Make RSA blinding thread-safe
3831 by remembering the creator's thread ID in rsa->blinding and
3832 having all other threads use local one-time blinding factors
3833 (this requires more computation than sharing rsa->blinding, but
3834 avoids excessive locking; and if an RSA object is not shared
3835 between threads, blinding will still be very fast).
3838 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3840 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3841 via timing by performing a MAC computation even if incorrrect
3842 block cipher padding has been found. This is a countermeasure
3843 against active attacks where the attacker has to distinguish
3844 between bad padding and a MAC verification error. (CVE-2003-0078)
3846 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3847 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3848 Martin Vuagnoux (EPFL, Ilion)]
3850 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3852 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3853 memory from it's contents. This is done with a counter that will
3854 place alternating values in each byte. This can be used to solve
3855 two issues: 1) the removal of calls to memset() by highly optimizing
3856 compilers, and 2) cleansing with other values than 0, since those can
3857 be read through on certain media, for example a swap space on disk.
3860 *) Bugfix: client side session caching did not work with external caching,
3861 because the session->cipher setting was not restored when reloading
3862 from the external cache. This problem was masked, when
3863 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3864 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3867 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3868 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3869 [Zeev Lieber <zeev-l@yahoo.com>]
3871 *) Undo an undocumented change introduced in 0.9.6e which caused
3872 repeated calls to OpenSSL_add_all_ciphers() and
3873 OpenSSL_add_all_digests() to be ignored, even after calling
3877 *) Change the default configuration reader to deal with last line not
3878 being properly terminated.
3881 *) Change X509_NAME_cmp() so it applies the special rules on handling
3882 DN values that are of type PrintableString, as well as RDNs of type
3883 emailAddress where the value has the type ia5String.
3884 [stefank@valicert.com via Richard Levitte]
3886 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3887 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3888 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3889 the bitwise-OR of the two for use by the majority of applications
3890 wanting this behaviour, and update the docs. The documented
3891 behaviour and actual behaviour were inconsistent and had been
3892 changing anyway, so this is more a bug-fix than a behavioural
3894 [Geoff Thorpe, diagnosed by Nadav Har'El]
3896 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3897 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3900 *) Fix initialization code race conditions in
3901 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3902 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3903 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3904 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3905 ssl2_get_cipher_by_char(),
3906 ssl3_get_cipher_by_char().
3907 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3909 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3910 the cached sessions are flushed, as the remove_cb() might use ex_data
3911 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3912 (see [openssl.org #212]).
3913 [Geoff Thorpe, Lutz Jaenicke]
3915 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3916 length, instead of the encoding length to d2i_ASN1_OBJECT.
3919 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3921 *) [In 0.9.6g-engine release:]
3922 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3923 [Lynn Gazis <lgazis@rainbow.com>]
3925 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3927 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3928 and get fix the header length calculation.
3929 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3930 Alon Kantor <alonk@checkpoint.com> (and others),
3933 *) Use proper error handling instead of 'assertions' in buffer
3934 overflow checks added in 0.9.6e. This prevents DoS (the
3935 assertions could call abort()).
3936 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3938 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3940 *) Add various sanity checks to asn1_get_length() to reject
3941 the ASN1 length bytes if they exceed sizeof(long), will appear
3942 negative or the content length exceeds the length of the
3944 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3946 *) Fix cipher selection routines: ciphers without encryption had no flags
3947 for the cipher strength set and where therefore not handled correctly
3948 by the selection routines (PR #130).
3951 *) Fix EVP_dsa_sha macro.
3955 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3956 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3957 that was added in OpenSSL 0.9.6d.
3959 As the countermeasure turned out to be incompatible with some
3960 broken SSL implementations, the new option is part of SSL_OP_ALL.
3961 SSL_OP_ALL is usually employed when compatibility with weird SSL
3962 implementations is desired (e.g. '-bugs' option to 's_client' and
3963 's_server'), so the new option is automatically set in many
3967 *) Changes in security patch:
3969 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3970 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3971 Air Force Materiel Command, USAF, under agreement number
3974 *) Add various sanity checks to asn1_get_length() to reject
3975 the ASN1 length bytes if they exceed sizeof(long), will appear
3976 negative or the content length exceeds the length of the
3977 supplied buffer. (CVE-2002-0659)
3978 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3980 *) Assertions for various potential buffer overflows, not known to
3982 [Ben Laurie (CHATS)]
3984 *) Various temporary buffers to hold ASCII versions of integers were
3985 too small for 64 bit platforms. (CVE-2002-0655)
3986 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3988 *) Remote buffer overflow in SSL3 protocol - an attacker could
3989 supply an oversized session ID to a client. (CVE-2002-0656)
3990 [Ben Laurie (CHATS)]
3992 *) Remote buffer overflow in SSL2 protocol - an attacker could
3993 supply an oversized client master key. (CVE-2002-0656)
3994 [Ben Laurie (CHATS)]
3996 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3998 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3999 encoded as NULL) with id-dsa-with-sha1.
4000 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4002 *) Check various X509_...() return values in apps/req.c.
4003 [Nils Larsch <nla@trustcenter.de>]
4005 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4006 an end-of-file condition would erronously be flagged, when the CRLF
4007 was just at the end of a processed block. The bug was discovered when
4008 processing data through a buffering memory BIO handing the data to a
4009 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4010 <ptsekov@syntrex.com> and Nedelcho Stanev.
4013 *) Implement a countermeasure against a vulnerability recently found
4014 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4015 before application data chunks to avoid the use of known IVs
4016 with data potentially chosen by the attacker.
4019 *) Fix length checks in ssl3_get_client_hello().
4022 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4023 to prevent ssl3_read_internal() from incorrectly assuming that
4024 ssl3_read_bytes() found application data while handshake
4025 processing was enabled when in fact s->s3->in_read_app_data was
4026 merely automatically cleared during the initial handshake.
4027 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4029 *) Fix object definitions for Private and Enterprise: they were not
4030 recognized in their shortname (=lowercase) representation. Extend
4031 obj_dat.pl to issue an error when using undefined keywords instead
4032 of silently ignoring the problem (Svenning Sorensen
4033 <sss@sss.dnsalias.net>).
4036 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4037 generators, i.e. generators other than 2 and 5. (Previously, the
4038 code did not properly initialise the 'add' and 'rem' values to
4039 BN_generate_prime().)
4041 In the new general case, we do not insist that 'generator' is
4042 actually a primitive root: This requirement is rather pointless;
4043 a generator of the order-q subgroup is just as good, if not
4047 *) Map new X509 verification errors to alerts. Discovered and submitted by
4048 Tom Wu <tom@arcot.com>.
4051 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4052 returning non-zero before the data has been completely received
4053 when using non-blocking I/O.
4054 [Bodo Moeller; problem pointed out by John Hughes]
4056 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4057 [Ben Laurie, Lutz Jaenicke]
4059 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4060 Yoram Zahavi <YoramZ@gilian.com>).
4063 *) Add information about CygWin 1.3 and on, and preserve proper
4064 configuration for the versions before that.
4065 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4067 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4068 check whether we deal with a copy of a session and do not delete from
4069 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4070 <izhar@checkpoint.com>.
4073 *) Do not store session data into the internal session cache, if it
4074 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4075 flag is set). Proposed by Aslam <aslam@funk.com>.
4078 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4082 *) [In 0.9.6d-engine release:]
4083 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4084 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4086 *) Add the configuration target linux-s390x.
4087 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4089 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4090 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4091 variable as an indication that a ClientHello message has been
4092 received. As the flag value will be lost between multiple
4093 invocations of ssl3_accept when using non-blocking I/O, the
4094 function may not be aware that a handshake has actually taken
4095 place, thus preventing a new session from being added to the
4098 To avoid this problem, we now set s->new_session to 2 instead of
4099 using a local variable.
4100 [Lutz Jaenicke, Bodo Moeller]
4102 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4103 if the SSL_R_LENGTH_MISMATCH error is detected.
4104 [Geoff Thorpe, Bodo Moeller]
4106 *) New 'shared_ldflag' column in Configure platform table.
4109 *) Fix EVP_CIPHER_mode macro.
4110 ["Dan S. Camper" <dan@bti.net>]
4112 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4113 type, we must throw them away by setting rr->length to 0.
4114 [D P Chang <dpc@qualys.com>]
4116 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4118 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4119 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4120 worked incorrectly for those cases where range = 10..._2 and
4121 3*range is two bits longer than range.)
4124 *) Only add signing time to PKCS7 structures if it is not already
4128 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4129 OBJ_ld_ce should be OBJ_id_ce.
4130 Also some ip-pda OIDs in crypto/objects/objects.txt were
4131 incorrect (cf. RFC 3039).
4132 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4134 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4135 returns early because it has nothing to do.
4136 [Andy Schneider <andy.schneider@bjss.co.uk>]
4138 *) [In 0.9.6c-engine release:]
4139 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4140 [Andy Schneider <andy.schneider@bjss.co.uk>]
4142 *) [In 0.9.6c-engine release:]
4143 Add support for Cryptographic Appliance's keyserver technology.
4144 (Use engine 'keyclient')
4145 [Cryptographic Appliances and Geoff Thorpe]
4147 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4148 is called via tools/c89.sh because arguments have to be
4149 rearranged (all '-L' options must appear before the first object
4151 [Richard Shapiro <rshapiro@abinitio.com>]
4153 *) [In 0.9.6c-engine release:]
4154 Add support for Broadcom crypto accelerator cards, backported
4156 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4158 *) [In 0.9.6c-engine release:]
4159 Add support for SureWare crypto accelerator cards from
4160 Baltimore Technologies. (Use engine 'sureware')
4161 [Baltimore Technologies and Mark Cox]
4163 *) [In 0.9.6c-engine release:]
4164 Add support for crypto accelerator cards from Accelerated
4165 Encryption Processing, www.aep.ie. (Use engine 'aep')
4166 [AEP Inc. and Mark Cox]
4168 *) Add a configuration entry for gcc on UnixWare.
4169 [Gary Benson <gbenson@redhat.com>]
4171 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4172 messages are stored in a single piece (fixed-length part and
4173 variable-length part combined) and fix various bugs found on the way.
4176 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4177 instead. BIO_gethostbyname() does not know what timeouts are
4178 appropriate, so entries would stay in cache even when they have
4180 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4182 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4183 faced with a pathologically small ClientHello fragment that does
4184 not contain client_version: Instead of aborting with an error,
4185 simply choose the highest available protocol version (i.e.,
4186 TLS 1.0 unless it is disabled). In practice, ClientHello
4187 messages are never sent like this, but this change gives us
4188 strictly correct behaviour at least for TLS.
4191 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4192 never resets s->method to s->ctx->method when called from within
4193 one of the SSL handshake functions.
4194 [Bodo Moeller; problem pointed out by Niko Baric]
4196 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4197 (sent using the client's version number) if client_version is
4198 smaller than the protocol version in use. Also change
4199 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4200 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4201 the client will at least see that alert.
4204 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4208 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4209 client receives HelloRequest while in a handshake.
4210 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4212 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4213 should end in 'break', not 'goto end' which circuments various
4214 cleanups done in state SSL_ST_OK. But session related stuff
4215 must be disabled for SSL_ST_OK in the case that we just sent a
4218 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4219 before just sending a HelloRequest.
4220 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4222 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4223 reveal whether illegal block cipher padding was found or a MAC
4224 verification error occured. (Neither SSLerr() codes nor alerts
4225 are directly visible to potential attackers, but the information
4226 may leak via logfiles.)
4228 Similar changes are not required for the SSL 2.0 implementation
4229 because the number of padding bytes is sent in clear for SSL 2.0,
4230 and the extra bytes are just ignored. However ssl/s2_pkt.c
4231 failed to verify that the purported number of padding bytes is in
4235 *) Add OpenUNIX-8 support including shared libraries
4236 (Boyd Lynn Gerber <gerberb@zenez.com>).
4239 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4240 'wristwatch attack' using huge encoding parameters (cf.
4241 James H. Manger's CRYPTO 2001 paper). Note that the
4242 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4243 encoding parameters and hence was not vulnerable.
4246 *) BN_sqr() bug fix.
4247 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4249 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4250 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4251 followed by modular reduction.
4252 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4254 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4255 equivalent based on BN_pseudo_rand() instead of BN_rand().
4258 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4259 This function was broken, as the check for a new client hello message
4260 to handle SGC did not allow these large messages.
4261 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4264 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4267 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4268 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4271 *) Rework the configuration and shared library support for Tru64 Unix.
4272 The configuration part makes use of modern compiler features and
4273 still retains old compiler behavior for those that run older versions
4274 of the OS. The shared library support part includes a variant that
4275 uses the RPATH feature, and is available through the special
4276 configuration target "alpha-cc-rpath", which will never be selected
4278 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4280 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4281 with the same message size as in ssl3_get_certificate_request().
4282 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4283 messages might inadvertently be reject as too long.
4284 [Petr Lampa <lampa@fee.vutbr.cz>]
4286 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4289 *) Modified SSL library such that the verify_callback that has been set
4290 specificly for an SSL object with SSL_set_verify() is actually being
4291 used. Before the change, a verify_callback set with this function was
4292 ignored and the verify_callback() set in the SSL_CTX at the time of
4293 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4294 to allow the necessary settings.
4297 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4298 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4299 done automatically (in contradiction to the requirements of the C
4300 standard). This made problems when used from OpenSSH.
4303 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4304 dh->length and always used
4306 BN_rand_range(priv_key, dh->p).
4308 BN_rand_range() is not necessary for Diffie-Hellman, and this
4309 specific range makes Diffie-Hellman unnecessarily inefficient if
4310 dh->length (recommended exponent length) is much smaller than the
4311 length of dh->p. We could use BN_rand_range() if the order of
4312 the subgroup was stored in the DH structure, but we only have
4317 BN_rand(priv_key, l, ...)
4319 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4325 RSA_eay_public_encrypt
4326 RSA_eay_private_decrypt
4327 RSA_eay_private_encrypt (signing)
4328 RSA_eay_public_decrypt (signature verification)
4330 (default implementations for RSA_public_encrypt,
4331 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4332 always reject numbers >= n.
4335 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4336 to synchronize access to 'locking_thread'. This is necessary on
4337 systems where access to 'locking_thread' (an 'unsigned long'
4338 variable) is not atomic.
4341 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4342 *before* setting the 'crypto_lock_rand' flag. The previous code had
4343 a race condition if 0 is a valid thread ID.
4344 [Travis Vitek <vitek@roguewave.com>]
4346 *) Add support for shared libraries under Irix.
4347 [Albert Chin-A-Young <china@thewrittenword.com>]
4349 *) Add configuration option to build on Linux on both big-endian and
4351 [Ralf Baechle <ralf@uni-koblenz.de>]
4353 *) Add the possibility to create shared libraries on HP-UX.
4356 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4358 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4359 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4360 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4361 PRNG state recovery was possible based on the output of
4362 one PRNG request appropriately sized to gain knowledge on
4363 'md' followed by enough consecutive 1-byte PRNG requests
4364 to traverse all of 'state'.
4366 1. When updating 'md_local' (the current thread's copy of 'md')
4367 during PRNG output generation, hash all of the previous
4368 'md_local' value, not just the half used for PRNG output.
4370 2. Make the number of bytes from 'state' included into the hash
4371 independent from the number of PRNG bytes requested.
4373 The first measure alone would be sufficient to avoid
4374 Markku-Juhani's attack. (Actually it had never occurred
4375 to me that the half of 'md_local' used for chaining was the
4376 half from which PRNG output bytes were taken -- I had always
4377 assumed that the secret half would be used.) The second
4378 measure makes sure that additional data from 'state' is never
4379 mixed into 'md_local' in small portions; this heuristically
4380 further strengthens the PRNG.
4383 *) Fix crypto/bn/asm/mips3.s.
4386 *) When only the key is given to "enc", the IV is undefined. Print out
4387 an error message in this case.
4390 *) Handle special case when X509_NAME is empty in X509 printing routines.
4393 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4394 positive and less than q.
4397 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4398 used: it isn't thread safe and the add_lock_callback should handle
4400 [Paul Rose <Paul.Rose@bridge.com>]
4402 *) Verify that incoming data obeys the block size in
4403 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4407 [Ulf Möller, Bodo Möller]
4409 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4410 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4411 when fixing the server behaviour for backwards-compatible 'client
4412 hello' messages. (Note that the attack is impractical against
4413 SSL 3.0 and TLS 1.0 anyway because length and version checking
4414 means that the probability of guessing a valid ciphertext is
4415 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4418 Before 0.9.5, the countermeasure (hide the error by generating a
4419 random 'decryption result') did not work properly because
4420 ERR_clear_error() was missing, meaning that SSL_get_error() would
4421 detect the supposedly ignored error.
4423 Both problems are now fixed.
4426 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4427 (previously it was 1024).
4430 *) Fix for compatibility mode trust settings: ignore trust settings
4431 unless some valid trust or reject settings are present.
4434 *) Fix for blowfish EVP: its a variable length cipher.
4437 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4438 parameters in DSA public key structures and return an error in the
4439 DSA routines if parameters are absent.
4442 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4443 in the current directory if neither $RANDFILE nor $HOME was set.
4444 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4445 caused some confusion to Windows users who haven't defined $HOME.
4446 Thus RAND_file_name() is changed again: e_os.h can define a
4447 DEFAULT_HOME, which will be used if $HOME is not set.
4448 For Windows, we use "C:"; on other platforms, we still require
4449 environment variables.
4451 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4452 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4453 having multiple threads call RAND_poll() concurrently.
4456 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4457 combination of a flag and a thread ID variable.
4458 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4459 flag), *other* threads can enter ssleay_add_bytes without obeying
4460 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4461 that they do not hold after the first thread unsets add_do_not_lock).
4464 *) Change bctest again: '-x' expressions are not available in all
4468 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4470 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4471 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4473 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4474 the default extension for executables, if any. Also, make the perl
4475 scripts that use symlink() to test if it really exists and use "cp"
4476 if it doesn't. All this made OpenSSL compilable and installable in
4480 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4481 If SEQUENCE is length is indefinite just set c->slen to the total
4482 amount of data available.
4483 [Steve Henson, reported by shige@FreeBSD.org]
4484 [This change does not apply to 0.9.7.]
4486 *) Change bctest to avoid here-documents inside command substitution
4487 (workaround for FreeBSD /bin/sh bug).
4488 For compatibility with Ultrix, avoid shell functions (introduced
4489 in the bctest version that searches along $PATH).
4492 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4493 with des_encrypt() defined on some operating systems, like Solaris
4497 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4498 On the Importance of Eliminating Errors in Cryptographic
4499 Computations, J. Cryptology 14 (2001) 2, 101-119,
4500 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4503 *) MIPS assembler BIGNUM division bug fix.
4506 *) Disabled incorrect Alpha assembler code.
4509 *) Fix PKCS#7 decode routines so they correctly update the length
4510 after reading an EOC for the EXPLICIT tag.
4512 [This change does not apply to 0.9.7.]
4514 *) Fix bug in PKCS#12 key generation routines. This was triggered
4515 if a 3DES key was generated with a 0 initial byte. Include
4516 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4517 (but broken) behaviour.
4520 *) Enhance bctest to search for a working bc along $PATH and print
4522 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4524 *) Fix memory leaks in err.c: free err_data string if necessary;
4525 don't write to the wrong index in ERR_set_error_data.
4528 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4532 *) Replace rdtsc with _emit statements for VC++ version 5.
4533 [Jeremy Cooper <jeremy@baymoo.org>]
4535 *) Make it possible to reuse SSLv2 sessions.
4538 *) In copy_email() check for >= 0 as a return value for
4539 X509_NAME_get_index_by_NID() since 0 is a valid index.
4540 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4542 *) Avoid coredump with unsupported or invalid public keys by checking if
4543 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4544 PKCS7_verify() fails with non detached data.
4547 *) Don't use getenv in library functions when run as setuid/setgid.
4548 New function OPENSSL_issetugid().
4551 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4552 due to incorrect handling of multi-threading:
4554 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4556 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4558 3. Count how many times MemCheck_off() has been called so that
4559 nested use can be treated correctly. This also avoids
4560 inband-signalling in the previous code (which relied on the
4561 assumption that thread ID 0 is impossible).
4564 *) Add "-rand" option also to s_client and s_server.
4567 *) Fix CPU detection on Irix 6.x.
4568 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4569 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4571 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4574 [This change does not apply to 0.9.7.]
4576 *) Use the cached encoding of an X509_NAME structure rather than
4577 copying it. This is apparently the reason for the libsafe "errors"
4578 but the code is actually correct.
4581 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4582 Bleichenbacher's DSA attack.
4583 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4584 to be set and top=0 forces the highest bit to be set; top=-1 is new
4585 and leaves the highest bit random.
4586 [Ulf Moeller, Bodo Moeller]
4588 *) In the NCONF_...-based implementations for CONF_... queries
4589 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4590 a temporary CONF structure with the data component set to NULL
4591 (which gives segmentation faults in lh_retrieve).
4592 Instead, use NULL for the CONF pointer in CONF_get_string and
4593 CONF_get_number (which may use environment variables) and directly
4594 return NULL from CONF_get_section.
4597 *) Fix potential buffer overrun for EBCDIC.
4600 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4601 keyUsage if basicConstraints absent for a CA.
4604 *) Make SMIME_write_PKCS7() write mail header values with a format that
4605 is more generally accepted (no spaces before the semicolon), since
4606 some programs can't parse those values properly otherwise. Also make
4607 sure BIO's that break lines after each write do not create invalid
4611 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4612 macros previously used would not encode an empty SEQUENCE OF
4613 and break the signature.
4615 [This change does not apply to 0.9.7.]
4617 *) Zero the premaster secret after deriving the master secret in
4621 *) Add some EVP_add_digest_alias registrations (as found in
4622 OpenSSL_add_all_digests()) to SSL_library_init()
4623 aka OpenSSL_add_ssl_algorithms(). This provides improved
4624 compatibility with peers using X.509 certificates
4625 with unconventional AlgorithmIdentifier OIDs.
4628 *) Fix for Irix with NO_ASM.
4629 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4631 *) ./config script fixes.
4632 [Ulf Moeller, Richard Levitte]
4634 *) Fix 'openssl passwd -1'.
4637 *) Change PKCS12_key_gen_asc() so it can cope with non null
4638 terminated strings whose length is passed in the passlen
4639 parameter, for example from PEM callbacks. This was done
4640 by adding an extra length parameter to asc2uni().
4641 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4643 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4644 call failed, free the DSA structure.
4647 *) Fix to uni2asc() to cope with zero length Unicode strings.
4648 These are present in some PKCS#12 files.
4651 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4652 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4653 when writing a 32767 byte record.
4654 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4656 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4657 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4659 (RSA objects have a reference count access to which is protected
4660 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4661 so they are meant to be shared between threads.)
4662 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4663 "Reddie, Steven" <Steven.Reddie@ca.com>]
4665 *) Fix a deadlock in CRYPTO_mem_leaks().
4668 *) Use better test patterns in bntest.
4671 *) rand_win.c fix for Borland C.
4674 *) BN_rshift bugfix for n == 0.
4677 *) Add a 'bctest' script that checks for some known 'bc' bugs
4678 so that 'make test' does not abort just because 'bc' is broken.
4681 *) Store verify_result within SSL_SESSION also for client side to
4682 avoid potential security hole. (Re-used sessions on the client side
4683 always resulted in verify_result==X509_V_OK, not using the original
4684 result of the server certificate verification.)
4687 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4688 SSL3_RT_APPLICATION_DATA, return 0.
4689 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4693 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4694 releases, have been re-implemented by renaming the previous
4695 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4696 and ssl3_read_internal, respectively, and adding 'peek' parameters
4697 to them. The new ssl[23]_{read,peek} functions are calls to
4698 ssl[23]_read_internal with the 'peek' flag set appropriately.
4699 A 'peek' parameter has also been added to ssl3_read_bytes, which
4700 does the actual work for ssl3_read_internal.
4703 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4704 the method-specific "init()" handler. Also clean up ex_data after
4705 calling the method-specific "finish()" handler. Previously, this was
4706 happening the other way round.
4709 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4710 The previous value, 12, was not always sufficient for BN_mod_exp().
4713 *) Make sure that shared libraries get the internal name engine with
4714 the full version number and not just 0. This should mark the
4715 shared libraries as not backward compatible. Of course, this should
4716 be changed again when we can guarantee backward binary compatibility.
4719 *) Fix typo in get_cert_by_subject() in by_dir.c
4720 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4722 *) Rework the system to generate shared libraries:
4724 - Make note of the expected extension for the shared libraries and
4725 if there is a need for symbolic links from for example libcrypto.so.0
4726 to libcrypto.so.0.9.7. There is extended info in Configure for
4729 - Make as few rebuilds of the shared libraries as possible.
4731 - Still avoid linking the OpenSSL programs with the shared libraries.
4733 - When installing, install the shared libraries separately from the
4737 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4739 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4740 and not in SSL_clear because the latter is also used by the
4741 accept/connect functions; previously, the settings made by
4742 SSL_set_read_ahead would be lost during the handshake.
4743 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4745 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4746 Previously, it would create entries for disableed algorithms no
4750 *) Added several new manual pages for SSL_* function.
4753 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4755 *) In ssl23_get_client_hello, generate an error message when faced
4756 with an initial SSL 3.0/TLS record that is too small to contain the
4757 first two bytes of the ClientHello message, i.e. client_version.
4758 (Note that this is a pathologic case that probably has never happened
4759 in real life.) The previous approach was to use the version number
4760 from the record header as a substitute; but our protocol choice
4761 should not depend on that one because it is not authenticated
4762 by the Finished messages.
4765 *) More robust randomness gathering functions for Windows.
4766 [Jeffrey Altman <jaltman@columbia.edu>]
4768 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4769 not set then we don't setup the error code for issuer check errors
4770 to avoid possibly overwriting other errors which the callback does
4771 handle. If an application does set the flag then we assume it knows
4772 what it is doing and can handle the new informational codes
4776 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4777 a general "ANY" type, as such it should be able to decode anything
4778 including tagged types. However it didn't check the class so it would
4779 wrongly interpret tagged types in the same way as their universal
4780 counterpart and unknown types were just rejected. Changed so that the
4781 tagged and unknown types are handled in the same way as a SEQUENCE:
4782 that is the encoding is stored intact. There is also a new type
4783 "V_ASN1_OTHER" which is used when the class is not universal, in this
4784 case we have no idea what the actual type is so we just lump them all
4788 *) On VMS, stdout may very well lead to a file that is written to
4789 in a record-oriented fashion. That means that every write() will
4790 write a separate record, which will be read separately by the
4791 programs trying to read from it. This can be very confusing.
4793 The solution is to put a BIO filter in the way that will buffer
4794 text until a linefeed is reached, and then write everything a
4795 line at a time, so every record written will be an actual line,
4796 not chunks of lines and not (usually doesn't happen, but I've
4797 seen it once) several lines in one record. BIO_f_linebuffer() is
4800 Currently, it's a VMS-only method, because that's where it has
4801 been tested well enough.
4804 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4805 it can return incorrect results.
4806 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4807 but it was in 0.9.6-beta[12].)
4810 *) Disable the check for content being present when verifying detached
4811 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4812 include zero length content when signing messages.
4815 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4816 BIO_ctrl (for BIO pairs).
4819 *) Add DSO method for VMS.
4822 *) Bug fix: Montgomery multiplication could produce results with the
4826 *) Add RPM specification openssl.spec and modify it to build three
4827 packages. The default package contains applications, application
4828 documentation and run-time libraries. The devel package contains
4829 include files, static libraries and function documentation. The
4830 doc package contains the contents of the doc directory. The original
4831 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4834 *) Add a large number of documentation files for many SSL routines.
4835 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4837 *) Add a configuration entry for Sony News 4.
4838 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4840 *) Don't set the two most significant bits to one when generating a
4841 random number < q in the DSA library.
4844 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4845 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4846 the underlying transport is blocking) if a handshake took place.
4847 (The default behaviour is needed by applications such as s_client
4848 and s_server that use select() to determine when to use SSL_read;
4849 but for applications that know in advance when to expect data, it
4850 just makes things more complicated.)
4853 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4857 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4858 work better on such systems.
4859 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4861 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4862 Update PKCS12_parse() so it copies the friendlyName and the
4863 keyid to the certificates aux info.
4866 *) Fix bug in PKCS7_verify() which caused an infinite loop
4867 if there was more than one signature.
4868 [Sven Uszpelkat <su@celocom.de>]
4870 *) Major change in util/mkdef.pl to include extra information
4871 about each symbol, as well as presentig variables as well
4872 as functions. This change means that there's n more need
4873 to rebuild the .num files when some algorithms are excluded.
4876 *) Allow the verify time to be set by an application,
4877 rather than always using the current time.
4880 *) Phase 2 verify code reorganisation. The certificate
4881 verify code now looks up an issuer certificate by a
4882 number of criteria: subject name, authority key id
4883 and key usage. It also verifies self signed certificates
4884 by the same criteria. The main comparison function is
4885 X509_check_issued() which performs these checks.
4887 Lot of changes were necessary in order to support this
4888 without completely rewriting the lookup code.
4890 Authority and subject key identifier are now cached.
4892 The LHASH 'certs' is X509_STORE has now been replaced
4893 by a STACK_OF(X509_OBJECT). This is mainly because an
4894 LHASH can't store or retrieve multiple objects with
4895 the same hash value.
4897 As a result various functions (which were all internal
4898 use only) have changed to handle the new X509_STORE
4899 structure. This will break anything that messed round
4900 with X509_STORE internally.
4902 The functions X509_STORE_add_cert() now checks for an
4903 exact match, rather than just subject name.
4905 The X509_STORE API doesn't directly support the retrieval
4906 of multiple certificates matching a given criteria, however
4907 this can be worked round by performing a lookup first
4908 (which will fill the cache with candidate certificates)
4909 and then examining the cache for matches. This is probably
4910 the best we can do without throwing out X509_LOOKUP
4911 entirely (maybe later...).
4913 The X509_VERIFY_CTX structure has been enhanced considerably.
4915 All certificate lookup operations now go via a get_issuer()
4916 callback. Although this currently uses an X509_STORE it
4917 can be replaced by custom lookups. This is a simple way
4918 to bypass the X509_STORE hackery necessary to make this
4919 work and makes it possible to use more efficient techniques
4920 in future. A very simple version which uses a simple
4921 STACK for its trusted certificate store is also provided
4922 using X509_STORE_CTX_trusted_stack().
4924 The verify_cb() and verify() callbacks now have equivalents
4925 in the X509_STORE_CTX structure.
4927 X509_STORE_CTX also has a 'flags' field which can be used
4928 to customise the verify behaviour.
4931 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4932 excludes S/MIME capabilities.
4935 *) When a certificate request is read in keep a copy of the
4936 original encoding of the signed data and use it when outputing
4937 again. Signatures then use the original encoding rather than
4938 a decoded, encoded version which may cause problems if the
4939 request is improperly encoded.
4942 *) For consistency with other BIO_puts implementations, call
4943 buffer_write(b, ...) directly in buffer_puts instead of calling
4946 In BIO_puts, increment b->num_write as in BIO_write.
4947 [Peter.Sylvester@EdelWeb.fr]
4949 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4950 BN_zero, we may not return a BIGNUM with an array consisting of
4954 *) Avoid calling abort() from within the library when problems are
4955 detected, except if preprocessor symbols have been defined
4956 (such as REF_CHECK, BN_DEBUG etc.).
4959 *) New openssl application 'rsautl'. This utility can be
4960 used for low level RSA operations. DER public key
4961 BIO/fp routines also added.
4964 *) New Configure entry and patches for compiling on QNX 4.
4965 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4967 *) A demo state-machine implementation was sponsored by
4968 Nuron (http://www.nuron.com/) and is now available in
4969 demos/state_machine.
4972 *) New options added to the 'dgst' utility for signature
4973 generation and verification.
4976 *) Unrecognized PKCS#7 content types are now handled via a
4977 catch all ASN1_TYPE structure. This allows unsupported
4978 types to be stored as a "blob" and an application can
4979 encode and decode it manually.
4982 *) Fix various signed/unsigned issues to make a_strex.c
4984 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4986 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4987 length if passed a buffer. ASN1_INTEGER_to_BN failed
4988 if passed a NULL BN and its argument was negative.
4989 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4991 *) Modification to PKCS#7 encoding routines to output definite
4992 length encoding. Since currently the whole structures are in
4993 memory there's not real point in using indefinite length
4994 constructed encoding. However if OpenSSL is compiled with
4995 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4998 *) Added BIO_vprintf() and BIO_vsnprintf().
5001 *) Added more prefixes to parse for in the the strings written
5002 through a logging bio, to cover all the levels that are available
5003 through syslog. The prefixes are now:
5005 PANIC, EMERG, EMR => LOG_EMERG
5006 ALERT, ALR => LOG_ALERT
5007 CRIT, CRI => LOG_CRIT
5008 ERROR, ERR => LOG_ERR
5009 WARNING, WARN, WAR => LOG_WARNING
5010 NOTICE, NOTE, NOT => LOG_NOTICE
5011 INFO, INF => LOG_INFO
5012 DEBUG, DBG => LOG_DEBUG
5014 and as before, if none of those prefixes are present at the
5015 beginning of the string, LOG_ERR is chosen.
5017 On Win32, the LOG_* levels are mapped according to this:
5019 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5020 LOG_WARNING => EVENTLOG_WARNING_TYPE
5021 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5025 *) Made it possible to reconfigure with just the configuration
5026 argument "reconf" or "reconfigure". The command line arguments
5027 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5028 and are retrieved from there when reconfiguring.
5032 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5034 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5037 *) The obj_dat.pl script was messing up the sorting of object
5038 names. The reason was that it compared the quoted version
5039 of strings as a result "OCSP" > "OCSP Signing" because
5040 " > SPACE. Changed script to store unquoted versions of
5041 names and add quotes on output. It was also omitting some
5042 names from the lookup table if they were given a default
5043 value (that is if SN is missing it is given the same
5044 value as LN and vice versa), these are now added on the
5045 grounds that if an object has a name we should be able to
5046 look it up. Finally added warning output when duplicate
5047 short or long names are found.
5050 *) Changes needed for Tandem NSK.
5051 [Scott Uroff <scott@xypro.com>]
5053 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5054 RSA_padding_check_SSLv23(), special padding was never detected
5055 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5056 version rollback attacks was not effective.
5058 In s23_clnt.c, don't use special rollback-attack detection padding
5059 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5060 client; similarly, in s23_srvr.c, don't do the rollback check if
5061 SSL 2.0 is the only protocol enabled in the server.
5064 *) Make it possible to get hexdumps of unprintable data with 'openssl
5065 asn1parse'. By implication, the functions ASN1_parse_dump() and
5066 BIO_dump_indent() are added.
5069 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5070 these print out strings and name structures based on various
5071 flags including RFC2253 support and proper handling of
5072 multibyte characters. Added options to the 'x509' utility
5073 to allow the various flags to be set.
5076 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5077 Also change the functions X509_cmp_current_time() and
5078 X509_gmtime_adj() work with an ASN1_TIME structure,
5079 this will enable certificates using GeneralizedTime in validity
5080 dates to be checked.
5083 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5084 negative public key encodings) on by default,
5085 NO_NEG_PUBKEY_BUG can be set to disable it.
5088 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5089 content octets. An i2c_ASN1_OBJECT is unnecessary because
5090 the encoding can be trivially obtained from the structure.
5093 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5094 not read locks (CRYPTO_r_[un]lock).
5097 *) A first attempt at creating official support for shared
5098 libraries through configuration. I've kept it so the
5099 default is static libraries only, and the OpenSSL programs
5100 are always statically linked for now, but there are
5101 preparations for dynamic linking in place.
5102 This has been tested on Linux and Tru64.
5105 *) Randomness polling function for Win9x, as described in:
5106 Peter Gutmann, Software Generation of Practically Strong
5110 *) Fix so PRNG is seeded in req if using an already existing
5114 *) New options to smime application. -inform and -outform
5115 allow alternative formats for the S/MIME message including
5116 PEM and DER. The -content option allows the content to be
5117 specified separately. This should allow things like Netscape
5118 form signing output easier to verify.
5121 *) Fix the ASN1 encoding of tags using the 'long form'.
5124 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5125 STRING types. These convert content octets to and from the
5126 underlying type. The actual tag and length octets are
5127 already assumed to have been read in and checked. These
5128 are needed because all other string types have virtually
5129 identical handling apart from the tag. By having versions
5130 of the ASN1 functions that just operate on content octets
5131 IMPLICIT tagging can be handled properly. It also allows
5132 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5133 and ASN1_INTEGER are identical apart from the tag.
5136 *) Change the handling of OID objects as follows:
5138 - New object identifiers are inserted in objects.txt, following
5139 the syntax given in objects.README.
5140 - objects.pl is used to process obj_mac.num and create a new
5142 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5145 This is currently kind of a hack, and the perl code in objects.pl
5146 isn't very elegant, but it works as I intended. The simplest way
5147 to check that it worked correctly is to look in obj_dat.h and
5148 check the array nid_objs and make sure the objects haven't moved
5149 around (this is important!). Additions are OK, as well as
5150 consistent name changes.
5153 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5156 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5157 The given file adds to whatever has already been seeded into the
5158 random pool through the RANDFILE configuration file option or
5159 environment variable, or the default random state file.
5162 *) mkstack.pl now sorts each macro group into lexical order.
5163 Previously the output order depended on the order the files
5164 appeared in the directory, resulting in needless rewriting
5168 *) Patches to make OpenSSL compile under Win32 again. Mostly
5169 work arounds for the VC++ problem that it treats func() as
5170 func(void). Also stripped out the parts of mkdef.pl that
5171 added extra typesafe functions: these no longer exist.
5174 *) Reorganisation of the stack code. The macros are now all
5175 collected in safestack.h . Each macro is defined in terms of
5176 a "stack macro" of the form SKM_<name>(type, a, b). The
5177 DEBUG_SAFESTACK is now handled in terms of function casts,
5178 this has the advantage of retaining type safety without the
5179 use of additional functions. If DEBUG_SAFESTACK is not defined
5180 then the non typesafe macros are used instead. Also modified the
5181 mkstack.pl script to handle the new form. Needs testing to see
5182 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5183 the default if no major problems. Similar behaviour for ASN1_SET_OF
5184 and PKCS12_STACK_OF.
5187 *) When some versions of IIS use the 'NET' form of private key the
5188 key derivation algorithm is different. Normally MD5(password) is
5189 used as a 128 bit RC4 key. In the modified case
5190 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5191 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5192 as the old Netscape_RSA functions except they have an additional
5193 'sgckey' parameter which uses the modified algorithm. Also added
5194 an -sgckey command line option to the rsa utility. Thanks to
5195 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5196 algorithm to openssl-dev.
5199 *) The evp_local.h macros were using 'c.##kname' which resulted in
5200 invalid expansion on some systems (SCO 5.0.5 for example).
5201 Corrected to 'c.kname'.
5202 [Phillip Porch <root@theporch.com>]
5204 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5205 a STACK of email addresses from a certificate or request, these look
5206 in the subject name and the subject alternative name extensions and
5207 omit any duplicate addresses.
5210 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5211 This makes DSA verification about 2 % faster.
5214 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5215 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5216 plus overhead for 1024 bit moduli).
5217 This makes exponentiations about 0.5 % faster for 1024 bit
5218 exponents (as measured by "openssl speed rsa2048").
5221 *) Rename memory handling macros to avoid conflicts with other
5223 Malloc => OPENSSL_malloc
5224 Malloc_locked => OPENSSL_malloc_locked
5225 Realloc => OPENSSL_realloc
5226 Free => OPENSSL_free
5229 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5230 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5233 *) CygWin32 support.
5234 [John Jarvie <jjarvie@newsguy.com>]
5236 *) The type-safe stack code has been rejigged. It is now only compiled
5237 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5238 by default all type-specific stack functions are "#define"d back to
5239 standard stack functions. This results in more streamlined output
5240 but retains the type-safety checking possibilities of the original
5244 *) The STACK code has been cleaned up, and certain type declarations
5245 that didn't make a lot of sense have been brought in line. This has
5246 also involved a cleanup of sorts in safestack.h to more correctly
5247 map type-safe stack functions onto their plain stack counterparts.
5248 This work has also resulted in a variety of "const"ifications of
5249 lots of the code, especially "_cmp" operations which should normally
5250 be prototyped with "const" parameters anyway.
5253 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5254 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5255 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5256 where all of 'md' is used each time the PRNG is used, but 'state'
5257 is used only indexed by a cyclic counter. As entropy may not be
5258 well distributed from the beginning, 'md' is important as a
5259 chaining variable. However, the output function chains only half
5260 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5261 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5262 in all of 'state' being rewritten, with the new values depending
5263 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5266 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5267 the handshake is continued after ssl_verify_cert_chain();
5268 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5269 can lead to 'unexplainable' connection aborts later.
5270 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5272 *) Major EVP API cipher revision.
5273 Add hooks for extra EVP features. This allows various cipher
5274 parameters to be set in the EVP interface. Support added for variable
5275 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5276 setting of RC2 and RC5 parameters.
5278 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5281 Remove lots of duplicated code from the EVP library. For example *every*
5282 cipher init() function handles the 'iv' in the same way according to the
5283 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5284 for CFB and OFB modes they zero ctx->num.
5286 New functionality allows removal of S/MIME code RC2 hack.
5288 Most of the routines have the same form and so can be declared in terms
5291 By shifting this to the top level EVP_CipherInit() it can be removed from
5292 all individual ciphers. If the cipher wants to handle IVs or keys
5293 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5296 Change lots of functions like EVP_EncryptUpdate() to now return a
5297 value: although software versions of the algorithms cannot fail
5298 any installed hardware versions can.
5301 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5302 this option is set, tolerate broken clients that send the negotiated
5303 protocol version number instead of the requested protocol version
5307 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5308 i.e. non-zero for export ciphersuites, zero otherwise.
5309 Previous versions had this flag inverted, inconsistent with
5310 rsa_tmp_cb (..._TMP_RSA_CB).
5311 [Bodo Moeller; problem reported by Amit Chopra]
5313 *) Add missing DSA library text string. Work around for some IIS
5314 key files with invalid SEQUENCE encoding.
5317 *) Add a document (doc/standards.txt) that list all kinds of standards
5318 and so on that are implemented in OpenSSL.
5321 *) Enhance c_rehash script. Old version would mishandle certificates
5322 with the same subject name hash and wouldn't handle CRLs at all.
5323 Added -fingerprint option to crl utility, to support new c_rehash
5327 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5330 *) Fix for SSL server purpose checking. Server checking was
5331 rejecting certificates which had extended key usage present
5332 but no ssl client purpose.
5333 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5335 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5336 is a little unclear about how a blank password is handled.
5337 Since the password in encoded as a BMPString with terminating
5338 double NULL a zero length password would end up as just the
5339 double NULL. However no password at all is different and is
5340 handled differently in the PKCS#12 key generation code. NS
5341 treats a blank password as zero length. MSIE treats it as no
5342 password on export: but it will try both on import. We now do
5343 the same: PKCS12_parse() tries zero length and no password if
5344 the password is set to "" or NULL (NULL is now a valid password:
5345 it wasn't before) as does the pkcs12 application.
5348 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5349 perror when PEM_read_bio_X509_REQ fails, the error message must
5350 be obtained from the error queue.
5353 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5354 it in ERR_remove_state if appropriate, and change ERR_get_state
5355 accordingly to avoid race conditions (this is necessary because
5356 thread_hash is no longer constant once set).
5359 *) Bugfix for linux-elf makefile.one.
5362 *) RSA_get_default_method() will now cause a default
5363 RSA_METHOD to be chosen if one doesn't exist already.
5364 Previously this was only set during a call to RSA_new()
5365 or RSA_new_method(NULL) meaning it was possible for
5366 RSA_get_default_method() to return NULL.
5369 *) Added native name translation to the existing DSO code
5370 that will convert (if the flag to do so is set) filenames
5371 that are sufficiently small and have no path information
5372 into a canonical native form. Eg. "blah" converted to
5373 "libblah.so" or "blah.dll" etc.
5376 *) New function ERR_error_string_n(e, buf, len) which is like
5377 ERR_error_string(e, buf), but writes at most 'len' bytes
5378 including the 0 terminator. For ERR_error_string_n, 'buf'
5380 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5382 *) CONF library reworked to become more general. A new CONF
5383 configuration file reader "class" is implemented as well as a
5384 new functions (NCONF_*, for "New CONF") to handle it. The now
5385 old CONF_* functions are still there, but are reimplemented to
5386 work in terms of the new functions. Also, a set of functions
5387 to handle the internal storage of the configuration data is
5388 provided to make it easier to write new configuration file
5389 reader "classes" (I can definitely see something reading a
5390 configuration file in XML format, for example), called _CONF_*,
5391 or "the configuration storage API"...
5393 The new configuration file reading functions are:
5395 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5396 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5398 NCONF_default, NCONF_WIN32
5400 NCONF_dump_fp, NCONF_dump_bio
5402 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5403 NCONF_new creates a new CONF object. This works in the same way
5404 as other interfaces in OpenSSL, like the BIO interface.
5405 NCONF_dump_* dump the internal storage of the configuration file,
5406 which is useful for debugging. All other functions take the same
5407 arguments as the old CONF_* functions wth the exception of the
5408 first that must be a `CONF *' instead of a `LHASH *'.
5410 To make it easer to use the new classes with the old CONF_* functions,
5411 the function CONF_set_default_method is provided.
5414 *) Add '-tls1' option to 'openssl ciphers', which was already
5415 mentioned in the documentation but had not been implemented.
5416 (This option is not yet really useful because even the additional
5417 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5420 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5421 OpenSSL-based applications) load shared libraries and bind to
5422 them in a portable way.
5423 [Geoff Thorpe, with contributions from Richard Levitte]
5425 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5427 *) Make sure _lrotl and _lrotr are only used with MSVC.
5429 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5430 (the default implementation of RAND_status).
5432 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5433 to '-clrext' (= clear extensions), as intended and documented.
5434 [Bodo Moeller; inconsistency pointed out by Michael Attili
5437 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5438 was larger than the MD block size.
5439 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5441 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5442 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5443 using the passed key: if the passed key was a private key the result
5444 of X509_print(), for example, would be to print out all the private key
5448 *) des_quad_cksum() byte order bug fix.
5449 [Ulf Möller, using the problem description in krb4-0.9.7, where
5450 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5452 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5454 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5456 *) For easily testing in shell scripts whether some command
5457 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5458 returns with exit code 0 iff no command of the given name is available.
5459 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5460 the output goes to stdout and nothing is printed to stderr.
5461 Additional arguments are always ignored.
5463 Since for each cipher there is a command of the same name,
5464 the 'no-cipher' compilation switches can be tested this way.
5466 ('openssl no-XXX' is not able to detect pseudo-commands such
5467 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5470 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5473 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5474 is set; it will be thrown away anyway because each handshake creates
5476 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5477 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5478 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5479 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5482 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5483 'Q' and 'R' lose their special meanings (quit/renegotiate).
5484 This is part of what -quiet does; unlike -quiet, -ign_eof
5485 does not suppress any output.
5488 *) Add compatibility options to the purpose and trust code. The
5489 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5490 accepts a certificate or CA, this was the previous behaviour,
5491 with all the associated security issues.
5493 X509_TRUST_COMPAT is the old trust behaviour: only and
5494 automatically trust self signed roots in certificate store. A
5495 new trust setting X509_TRUST_DEFAULT is used to specify that
5496 a purpose has no associated trust setting and it should instead
5497 use the value in the default purpose.
5500 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5501 and fix a memory leak.
5504 *) In util/mkerr.pl (which implements 'make errors'), preserve
5505 reason strings from the previous version of the .c file, as
5506 the default to have only downcase letters (and digits) in
5507 automatically generated reasons codes is not always appropriate.
5510 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5511 using strerror. Previously, ERR_reason_error_string() returned
5512 library names as reason strings for SYSerr; but SYSerr is a special
5513 case where small numbers are errno values, not library numbers.
5516 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5517 converts DSA parameters into DH parameters. (When creating parameters,
5518 DSA_generate_parameters is used.)
5521 *) Include 'length' (recommended exponent length) in C code generated
5522 by 'openssl dhparam -C'.
5525 *) The second argument to set_label in perlasm was already being used
5526 so couldn't be used as a "file scope" flag. Moved to third argument
5530 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5531 instead of RAND_bytes for encryption IVs and salts.
5534 *) Include RAND_status() into RAND_METHOD instead of implementing
5535 it only for md_rand.c Otherwise replacing the PRNG by calling
5536 RAND_set_rand_method would be impossible.
5539 *) Don't let DSA_generate_key() enter an infinite loop if the random
5540 number generation fails.
5543 *) New 'rand' application for creating pseudo-random output.
5546 *) Added configuration support for Linux/IA64
5547 [Rolf Haberrecker <rolf@suse.de>]
5549 *) Assembler module support for Mingw32.
5552 *) Shared library support for HPUX (in shlib/).
5553 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5555 *) Shared library support for Solaris gcc.
5556 [Lutz Behnke <behnke@trustcenter.de>]
5558 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5560 *) PKCS7_encrypt() was adding text MIME headers twice because they
5561 were added manually and by SMIME_crlf_copy().
5564 *) In bntest.c don't call BN_rand with zero bits argument.
5565 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5567 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5568 case was implemented. This caused BN_div_recp() to fail occasionally.
5571 *) Add an optional second argument to the set_label() in the perl
5572 assembly language builder. If this argument exists and is set
5573 to 1 it signals that the assembler should use a symbol whose
5574 scope is the entire file, not just the current function. This
5575 is needed with MASM which uses the format label:: for this scope.
5576 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5578 *) Change the ASN1 types so they are typedefs by default. Before
5579 almost all types were #define'd to ASN1_STRING which was causing
5580 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5584 *) Change names of new functions to the new get1/get0 naming
5585 convention: After 'get1', the caller owns a reference count
5586 and has to call ..._free; 'get0' returns a pointer to some
5587 data structure without incrementing reference counters.
5588 (Some of the existing 'get' functions increment a reference
5589 counter, some don't.)
5590 Similarly, 'set1' and 'add1' functions increase reference
5591 counters or duplicate objects.
5594 *) Allow for the possibility of temp RSA key generation failure:
5595 the code used to assume it always worked and crashed on failure.
5598 *) Fix potential buffer overrun problem in BIO_printf().
5599 [Ulf Möller, using public domain code by Patrick Powell; problem
5600 pointed out by David Sacerdote <das33@cornell.edu>]
5602 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5603 RAND_egd() and RAND_status(). In the command line application,
5604 the EGD socket can be specified like a seed file using RANDFILE
5608 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5609 Some CAs (e.g. Verisign) distribute certificates in this form.
5612 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5613 list to exclude them. This means that no special compilation option
5614 is needed to use anonymous DH: it just needs to be included in the
5618 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5619 EVP_MD_type. The old functionality is available in a new macro called
5620 EVP_MD_md(). Change code that uses it and update docs.
5623 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5624 where the 'void *' argument is replaced by a function pointer argument.
5625 Previously 'void *' was abused to point to functions, which works on
5626 many platforms, but is not correct. As these functions are usually
5627 called by macros defined in OpenSSL header files, most source code
5628 should work without changes.
5631 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5632 sections with information on -D... compiler switches used for
5633 compiling the library so that applications can see them. To enable
5634 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5635 must be defined. E.g.,
5636 #define OPENSSL_ALGORITHM_DEFINES
5637 #include <openssl/opensslconf.h>
5638 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5639 [Richard Levitte, Ulf and Bodo Möller]
5641 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5645 *) Change the 'other' type in certificate aux info to a STACK_OF
5646 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5647 the required ASN1 format: arbitrary types determined by an OID.
5650 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5651 argument to 'req'. This is not because the function is newer or
5652 better than others it just uses the work 'NEW' in the certificate
5653 request header lines. Some software needs this.
5656 *) Reorganise password command line arguments: now passwords can be
5657 obtained from various sources. Delete the PEM_cb function and make
5658 it the default behaviour: i.e. if the callback is NULL and the
5659 usrdata argument is not NULL interpret it as a null terminated pass
5660 phrase. If usrdata and the callback are NULL then the pass phrase
5661 is prompted for as usual.
5664 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5665 the support is automatically enabled. The resulting binaries will
5666 autodetect the card and use it if present.
5667 [Ben Laurie and Compaq Inc.]
5669 *) Work around for Netscape hang bug. This sends certificate request
5670 and server done in one record. Since this is perfectly legal in the
5671 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5672 the bugs/SSLv3 entry for more info.
5675 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5678 *) Add -rand argument to smime and pkcs12 applications and read/write
5682 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5685 *) Add command line password options to the remaining applications.
5688 *) Bug fix for BN_div_recp() for numerators with an even number of
5692 *) More tests in bntest.c, and changed test_bn output.
5695 *) ./config recognizes MacOS X now.
5698 *) Bug fix for BN_div() when the first words of num and divsor are
5699 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5702 *) Add support for various broken PKCS#8 formats, and command line
5703 options to produce them.
5706 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5707 get temporary BIGNUMs from a BN_CTX.
5710 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5714 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5715 include a #define from the old name to the new. The original intent
5716 was that statically linked binaries could for example just call
5717 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5718 link with digests. This never worked becayse SSLeay_add_all_digests()
5719 and SSLeay_add_all_ciphers() were in the same source file so calling
5720 one would link with the other. They are now in separate source files.
5723 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5726 *) Use a less unusual form of the Miller-Rabin primality test (it used
5727 a binary algorithm for exponentiation integrated into the Miller-Rabin
5728 loop, our standard modexp algorithms are faster).
5731 *) Support for the EBCDIC character set completed.
5732 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5734 *) Source code cleanups: use const where appropriate, eliminate casts,
5735 use void * instead of char * in lhash.
5738 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5739 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5740 this the server could overwrite ephemeral keys that the client
5744 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5745 using 50 iterations of the Rabin-Miller test.
5747 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5748 iterations of the Rabin-Miller test as required by the appendix
5749 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5750 As BN_is_prime_fasttest includes trial division, DSA parameter
5751 generation becomes much faster.
5753 This implies a change for the callback functions in DSA_is_prime
5754 and DSA_generate_parameters: The callback function is called once
5755 for each positive witness in the Rabin-Miller test, not just
5756 occasionally in the inner loop; and the parameters to the
5757 callback function now provide an iteration count for the outer
5758 loop rather than for the current invocation of the inner loop.
5759 DSA_generate_parameters additionally can call the callback
5760 function with an 'iteration count' of -1, meaning that a
5761 candidate has passed the trial division test (when q is generated
5762 from an application-provided seed, trial division is skipped).
5765 *) New function BN_is_prime_fasttest that optionally does trial
5766 division before starting the Rabin-Miller test and has
5767 an additional BN_CTX * argument (whereas BN_is_prime always
5768 has to allocate at least one BN_CTX).
5769 'callback(1, -1, cb_arg)' is called when a number has passed the
5770 trial division stage.
5773 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5777 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5780 *) New function BN_pseudo_rand().
5783 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5784 bignum version of BN_from_montgomery() with the working code from
5785 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5789 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5790 made it impossible to use the same SSL_SESSION data structure in
5791 SSL2 clients in multiple threads.
5794 *) The return value of RAND_load_file() no longer counts bytes obtained
5795 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5796 to seed the PRNG (previously an explicit byte count was required).
5797 [Ulf Möller, Bodo Möller]
5799 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5800 used (char *) instead of (void *) and had casts all over the place.
5803 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5806 *) Retain source code compatibility for BN_prime_checks macro:
5807 BN_is_prime(..., BN_prime_checks, ...) now uses
5808 BN_prime_checks_for_size to determine the appropriate number of
5809 Rabin-Miller iterations.
5812 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5813 DH_CHECK_P_NOT_SAFE_PRIME.
5814 (Check if this is true? OpenPGP calls them "strong".)
5817 *) Merge the functionality of "dh" and "gendh" programs into a new program
5818 "dhparam". The old programs are retained for now but will handle DH keys
5819 (instead of parameters) in future.
5822 *) Make the ciphers, s_server and s_client programs check the return values
5823 when a new cipher list is set.
5826 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5827 ciphers. Before when the 56bit ciphers were enabled the sorting was
5830 The syntax for the cipher sorting has been extended to support sorting by
5831 cipher-strength (using the strength_bits hard coded in the tables).
5832 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5834 Fix a bug in the cipher-command parser: when supplying a cipher command
5835 string with an "undefined" symbol (neither command nor alphanumeric
5836 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5837 an error is flagged.
5839 Due to the strength-sorting extension, the code of the
5840 ssl_create_cipher_list() function was completely rearranged. I hope that
5841 the readability was also increased :-)
5842 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5844 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5845 for the first serial number and places 2 in the serial number file. This
5846 avoids problems when the root CA is created with serial number zero and
5847 the first user certificate has the same issuer name and serial number
5851 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5852 the new code. Add documentation for this stuff.
5855 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5856 X509_*() to X509at_*() on the grounds that they don't handle X509
5857 structures and behave in an analagous way to the X509v3 functions:
5858 they shouldn't be called directly but wrapper functions should be used
5861 So we also now have some wrapper functions that call the X509at functions
5862 when passed certificate requests. (TO DO: similar things can be done with
5863 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5864 things. Some of these need some d2i or i2d and print functionality
5865 because they handle more complex structures.)
5868 *) Add missing #ifndefs that caused missing symbols when building libssl
5869 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5870 NO_RSA in ssl/s2*.c.
5871 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5873 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5874 has a return value which indicates the quality of the random data
5875 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5876 error queue. New function RAND_pseudo_bytes() generates output that is
5877 guaranteed to be unique but not unpredictable. RAND_add is like
5878 RAND_seed, but takes an extra argument for an entropy estimate
5879 (RAND_seed always assumes full entropy).
5882 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5883 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5884 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5885 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5886 false-positive rate of at most 2^-80 for random input.
5889 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5892 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5893 in the 0.9.5 release), this returns the chain
5894 from an X509_CTX structure with a dup of the stack and all
5895 the X509 reference counts upped: so the stack will exist
5896 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5899 Also make SSL_SESSION_print() print out the verify return
5903 *) Add manpage for the pkcs12 command. Also change the default
5904 behaviour so MAC iteration counts are used unless the new
5905 -nomaciter option is used. This improves file security and
5906 only older versions of MSIE (4.0 for example) need it.
5909 *) Honor the no-xxx Configure options when creating .DEF files.
5912 *) Add PKCS#10 attributes to field table: challengePassword,
5913 unstructuredName and unstructuredAddress. These are taken from
5914 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5915 international characters are used.
5917 More changes to X509_ATTRIBUTE code: allow the setting of types
5918 based on strings. Remove the 'loc' parameter when adding
5919 attributes because these will be a SET OF encoding which is sorted
5923 *) Initial changes to the 'req' utility to allow request generation
5924 automation. This will allow an application to just generate a template
5925 file containing all the field values and have req construct the
5928 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5929 used all over the place including certificate requests and PKCS#7
5930 structures. They are currently handled manually where necessary with
5931 some primitive wrappers for PKCS#7. The new functions behave in a
5932 manner analogous to the X509 extension functions: they allow
5933 attributes to be looked up by NID and added.
5935 Later something similar to the X509V3 code would be desirable to
5936 automatically handle the encoding, decoding and printing of the
5937 more complex types. The string types like challengePassword can
5938 be handled by the string table functions.
5940 Also modified the multi byte string table handling. Now there is
5941 a 'global mask' which masks out certain types. The table itself
5942 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5943 is useful when for example there is only one permissible type
5944 (as in countryName) and using the mask might result in no valid
5948 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5949 SSL_get_peer_finished to allow applications to obtain the latest
5950 Finished messages sent to the peer or expected from the peer,
5951 respectively. (SSL_get_peer_finished is usually the Finished message
5952 actually received from the peer, otherwise the protocol will be aborted.)
5954 As the Finished message are message digests of the complete handshake
5955 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5956 be used for external authentication procedures when the authentication
5957 provided by SSL/TLS is not desired or is not enough.
5960 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5961 the host supports BWX extension and if Compaq C is present on the
5962 $PATH. Just exploiting of the BWX extension results in 20-30%
5963 performance kick for some algorithms, e.g. DES and RC4 to mention
5964 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5968 *) Add support for MS "fast SGC". This is arguably a violation of the
5969 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5970 weak crypto and after checking the certificate is SGC a second one
5971 with strong crypto. MS SGC stops the first handshake after receiving
5972 the server certificate message and sends a second client hello. Since
5973 a server will typically do all the time consuming operations before
5974 expecting any further messages from the client (server key exchange
5975 is the most expensive) there is little difference between the two.
5977 To get OpenSSL to support MS SGC we have to permit a second client
5978 hello message after we have sent server done. In addition we have to
5979 reset the MAC if we do get this second client hello.
5982 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5983 if a DER encoded private key is RSA or DSA traditional format. Changed
5984 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5985 format DER encoded private key. Newer code should use PKCS#8 format which
5986 has the key type encoded in the ASN1 structure. Added DER private key
5987 support to pkcs8 application.
5990 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5991 ciphersuites has been selected (as required by the SSL 3/TLS 1
5992 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5993 is set, we interpret this as a request to violate the specification
5994 (the worst that can happen is a handshake failure, and 'correct'
5995 behaviour would result in a handshake failure anyway).
5998 *) In SSL_CTX_add_session, take into account that there might be multiple
5999 SSL_SESSION structures with the same session ID (e.g. when two threads
6000 concurrently obtain them from an external cache).
6001 The internal cache can handle only one SSL_SESSION with a given ID,
6002 so if there's a conflict, we now throw out the old one to achieve
6006 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6007 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6008 some routines that use cipher OIDs: some ciphers do not have OIDs
6009 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6013 *) Simplify the trust setting structure and code. Now we just have
6014 two sequences of OIDs for trusted and rejected settings. These will
6015 typically have values the same as the extended key usage extension
6016 and any application specific purposes.
6018 The trust checking code now has a default behaviour: it will just
6019 check for an object with the same NID as the passed id. Functions can
6020 be provided to override either the default behaviour or the behaviour
6021 for a given id. SSL client, server and email already have functions
6022 in place for compatibility: they check the NID and also return "trusted"
6023 if the certificate is self signed.
6026 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6027 traditional format into an EVP_PKEY structure.
6030 *) Add a password callback function PEM_cb() which either prompts for
6031 a password if usr_data is NULL or otherwise assumes it is a null
6032 terminated password. Allow passwords to be passed on command line
6033 environment or config files in a few more utilities.
6036 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6037 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6038 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6039 Update documentation.
6042 *) Support for ASN1 "NULL" type. This could be handled before by using
6043 ASN1_TYPE but there wasn't any function that would try to read a NULL
6044 and produce an error if it couldn't. For compatibility we also have
6045 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6046 don't allocate anything because they don't need to.
6049 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6051 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6053 *) Rebuild of the memory allocation routines used by OpenSSL code and
6054 possibly others as well. The purpose is to make an interface that
6055 provide hooks so anyone can build a separate set of allocation and
6056 deallocation routines to be used by OpenSSL, for example memory
6057 pool implementations, or something else, which was previously hard
6058 since Malloc(), Realloc() and Free() were defined as macros having
6059 the values malloc, realloc and free, respectively (except for Win32
6060 compilations). The same is provided for memory debugging code.
6061 OpenSSL already comes with functionality to find memory leaks, but
6062 this gives people a chance to debug other memory problems.
6064 With these changes, a new set of functions and macros have appeared:
6066 CRYPTO_set_mem_debug_functions() [F]
6067 CRYPTO_get_mem_debug_functions() [F]
6068 CRYPTO_dbg_set_options() [F]
6069 CRYPTO_dbg_get_options() [F]
6070 CRYPTO_malloc_debug_init() [M]
6072 The memory debug functions are NULL by default, unless the library
6073 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6074 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6075 gives the standard debugging functions that come with OpenSSL) or
6076 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6077 provided by the library user) must be used. When the standard
6078 debugging functions are used, CRYPTO_dbg_set_options can be used to
6079 request additional information:
6080 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6081 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6083 Also, things like CRYPTO_set_mem_functions will always give the
6084 expected result (the new set of functions is used for allocation
6085 and deallocation) at all times, regardless of platform and compiler
6088 To finish it up, some functions that were never use in any other
6089 way than through macros have a new API and new semantic:
6092 CRYPTO_dbg_realloc()
6095 All macros of value have retained their old syntax.
6096 [Richard Levitte and Bodo Moeller]
6098 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6099 ordering of SMIMECapabilities wasn't in "strength order" and there
6100 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6104 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6105 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6106 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6108 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6109 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6110 functionality to handle multipart/signed properly) and a utility
6111 called 'smime' to call all this stuff. This is based on code I
6112 originally wrote for Celo who have kindly allowed it to be
6113 included in OpenSSL.
6116 *) Add variants des_set_key_checked and des_set_key_unchecked of
6117 des_set_key (aka des_key_sched). Global variable des_check_key
6118 decides which of these is called by des_set_key; this way
6119 des_check_key behaves as it always did, but applications and
6120 the library itself, which was buggy for des_check_key == 1,
6121 have a cleaner way to pick the version they need.
6124 *) New function PKCS12_newpass() which changes the password of a
6128 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6129 dynamic mix. In both cases the ids can be used as an index into the
6130 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6131 functions so they accept a list of the field values and the
6132 application doesn't need to directly manipulate the X509_TRUST
6136 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6140 *) Modify the way the V3 extension code looks up extensions. This now
6141 works in a similar way to the object code: we have some "standard"
6142 extensions in a static table which is searched with OBJ_bsearch()
6143 and the application can add dynamic ones if needed. The file
6144 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6145 updated whenever a new extension is added to the core code and kept
6146 in ext_nid order. There is a simple program 'tabtest.c' which checks
6147 this. New extensions are not added too often so this file can readily
6148 be maintained manually.
6150 There are two big advantages in doing things this way. The extensions
6151 can be looked up immediately and no longer need to be "added" using
6152 X509V3_add_standard_extensions(): this function now does nothing.
6153 [Side note: I get *lots* of email saying the extension code doesn't
6154 work because people forget to call this function]
6155 Also no dynamic allocation is done unless new extensions are added:
6156 so if we don't add custom extensions there is no need to call
6157 X509V3_EXT_cleanup().
6160 *) Modify enc utility's salting as follows: make salting the default. Add a
6161 magic header, so unsalted files fail gracefully instead of just decrypting
6162 to garbage. This is because not salting is a big security hole, so people
6163 should be discouraged from doing it.
6166 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6167 digest to be passed on the command line but it only used this
6168 parameter when signing a certificate. Modified so all relevant
6169 operations are affected by the digest parameter including the
6170 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6171 DSA key was used because it didn't fix the digest.
6174 *) Initial certificate chain verify code. Currently tests the untrusted
6175 certificates for consistency with the verify purpose (which is set
6176 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6178 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6179 this is because it will reject chains with invalid extensions whereas
6180 every previous version of OpenSSL and SSLeay made no checks at all.
6182 Trust code: checks the root CA for the relevant trust settings. Trust
6183 settings have an initial value consistent with the verify purpose: e.g.
6184 if the verify purpose is for SSL client use it expects the CA to be
6185 trusted for SSL client use. However the default value can be changed to
6186 permit custom trust settings: one example of this would be to only trust
6187 certificates from a specific "secure" set of CAs.
6189 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6190 which should be used for version portability: especially since the
6191 verify structure is likely to change more often now.
6193 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6194 to set them. If not set then assume SSL clients will verify SSL servers
6197 Two new options to the verify program: -untrusted allows a set of
6198 untrusted certificates to be passed in and -purpose which sets the
6199 intended purpose of the certificate. If a purpose is set then the
6200 new chain verify code is used to check extension consistency.
6203 *) Support for the authority information access extension.
6206 *) Modify RSA and DSA PEM read routines to transparently handle
6207 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6208 public keys in a format compatible with certificate
6209 SubjectPublicKeyInfo structures. Unfortunately there were already
6210 functions called *_PublicKey_* which used various odd formats so
6211 these are retained for compatibility: however the DSA variants were
6212 never in a public release so they have been deleted. Changed dsa/rsa
6213 utilities to handle the new format: note no releases ever handled public
6214 keys so we should be OK.
6216 The primary motivation for this change is to avoid the same fiasco
6217 that dogs private keys: there are several incompatible private key
6218 formats some of which are standard and some OpenSSL specific and
6219 require various evil hacks to allow partial transparent handling and
6220 even then it doesn't work with DER formats. Given the option anything
6221 other than PKCS#8 should be dumped: but the other formats have to
6222 stay in the name of compatibility.
6224 With public keys and the benefit of hindsight one standard format
6225 is used which works with EVP_PKEY, RSA or DSA structures: though
6226 it clearly returns an error if you try to read the wrong kind of key.
6228 Added a -pubkey option to the 'x509' utility to output the public key.
6229 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6230 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6231 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6232 that do the same as the EVP_PKEY_assign_*() except they up the
6233 reference count of the added key (they don't "swallow" the
6237 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6238 CRLs would fail if the file contained no certificates or no CRLs:
6239 added a new function to read in both types and return the number
6240 read: this means that if none are read it will be an error. The
6241 DER versions of the certificate and CRL reader would always fail
6242 because it isn't possible to mix certificates and CRLs in DER format
6243 without choking one or the other routine. Changed this to just read
6244 a certificate: this is the best we can do. Also modified the code
6245 in apps/verify.c to take notice of return codes: it was previously
6246 attempting to read in certificates from NULL pointers and ignoring
6247 any errors: this is one reason why the cert and CRL reader seemed
6248 to work. It doesn't check return codes from the default certificate
6249 routines: these may well fail if the certificates aren't installed.
6252 *) Code to support otherName option in GeneralName.
6255 *) First update to verify code. Change the verify utility
6256 so it warns if it is passed a self signed certificate:
6257 for consistency with the normal behaviour. X509_verify
6258 has been modified to it will now verify a self signed
6259 certificate if *exactly* the same certificate appears
6260 in the store: it was previously impossible to trust a
6261 single self signed certificate. This means that:
6262 openssl verify ss.pem
6263 now gives a warning about a self signed certificate but
6264 openssl verify -CAfile ss.pem ss.pem
6268 *) For servers, store verify_result in SSL_SESSION data structure
6269 (and add it to external session representation).
6270 This is needed when client certificate verifications fails,
6271 but an application-provided verification callback (set by
6272 SSL_CTX_set_cert_verify_callback) allows accepting the session
6273 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6274 but returns 1): When the session is reused, we have to set
6275 ssl->verify_result to the appropriate error code to avoid
6277 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6279 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6280 case in PKCS7_dataInit() where the signed PKCS7 structure
6281 didn't contain any existing data because it was being created.
6282 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6284 *) Add a salt to the key derivation routines in enc.c. This
6285 forms the first 8 bytes of the encrypted file. Also add a
6286 -S option to allow a salt to be input on the command line.
6289 *) New function X509_cmp(). Oddly enough there wasn't a function
6290 to compare two certificates. We do this by working out the SHA1
6291 hash and comparing that. X509_cmp() will be needed by the trust
6295 *) SSL_get1_session() is like SSL_get_session(), but increments
6296 the reference count in the SSL_SESSION returned.
6297 [Geoff Thorpe <geoff@eu.c2.net>]
6299 *) Fix for 'req': it was adding a null to request attributes.
6300 Also change the X509_LOOKUP and X509_INFO code to handle
6301 certificate auxiliary information.
6304 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6308 *) Add the possibility to add extra information to the memory leak
6309 detecting output, to form tracebacks, showing from where each
6310 allocation was originated: CRYPTO_push_info("constant string") adds
6311 the string plus current file name and line number to a per-thread
6312 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6313 is like calling CYRPTO_pop_info() until the stack is empty.
6314 Also updated memory leak detection code to be multi-thread-safe.
6317 *) Add options -text and -noout to pkcs7 utility and delete the
6318 encryption options which never did anything. Update docs.
6321 *) Add options to some of the utilities to allow the pass phrase
6322 to be included on either the command line (not recommended on
6323 OSes like Unix) or read from the environment. Update the
6324 manpages and fix a few bugs.
6327 *) Add a few manpages for some of the openssl commands.
6330 *) Fix the -revoke option in ca. It was freeing up memory twice,
6331 leaking and not finding already revoked certificates.
6334 *) Extensive changes to support certificate auxiliary information.
6335 This involves the use of X509_CERT_AUX structure and X509_AUX
6336 functions. An X509_AUX function such as PEM_read_X509_AUX()
6337 can still read in a certificate file in the usual way but it
6338 will also read in any additional "auxiliary information". By
6339 doing things this way a fair degree of compatibility can be
6340 retained: existing certificates can have this information added
6341 using the new 'x509' options.
6343 Current auxiliary information includes an "alias" and some trust
6344 settings. The trust settings will ultimately be used in enhanced
6345 certificate chain verification routines: currently a certificate
6346 can only be trusted if it is self signed and then it is trusted
6350 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6351 The problem was that one of the replacement routines had not been working
6352 since SSLeay releases. For now the offending routine has been replaced
6353 with non-optimised assembler. Even so, this now gives around 95%
6354 performance improvement for 1024 bit RSA signs.
6357 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6358 handling. Most clients have the effective key size in bits equal to
6359 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6360 A few however don't do this and instead use the size of the decrypted key
6361 to determine the RC2 key length and the AlgorithmIdentifier to determine
6362 the effective key length. In this case the effective key length can still
6363 be 40 bits but the key length can be 168 bits for example. This is fixed
6364 by manually forcing an RC2 key into the EVP_PKEY structure because the
6365 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6366 the key length and effective key length are equal.
6369 *) Add a bunch of functions that should simplify the creation of
6370 X509_NAME structures. Now you should be able to do:
6371 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6372 and have it automatically work out the correct field type and fill in
6373 the structures. The more adventurous can try:
6374 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6375 and it will (hopefully) work out the correct multibyte encoding.
6378 *) Change the 'req' utility to use the new field handling and multibyte
6379 copy routines. Before the DN field creation was handled in an ad hoc
6380 way in req, ca, and x509 which was rather broken and didn't support
6381 BMPStrings or UTF8Strings. Since some software doesn't implement
6382 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6383 using the dirstring_type option. See the new comment in the default
6384 openssl.cnf for more info.
6387 *) Make crypto/rand/md_rand.c more robust:
6388 - Assure unique random numbers after fork().
6389 - Make sure that concurrent threads access the global counter and
6390 md serializably so that we never lose entropy in them
6391 or use exactly the same state in multiple threads.
6392 Access to the large state is not always serializable because
6393 the additional locking could be a performance killer, and
6394 md should be large enough anyway.
6397 *) New file apps/app_rand.c with commonly needed functionality
6398 for handling the random seed file.
6400 Use the random seed file in some applications that previously did not:
6402 dsaparam -genkey (which also ignored its '-rand' option),
6405 x509 (when signing).
6406 Except on systems with /dev/urandom, it is crucial to have a random
6407 seed file at least for key creation, DSA signing, and for DH exchanges;
6408 for RSA signatures we could do without one.
6410 gendh and gendsa (unlike genrsa) used to read only the first byte
6411 of each file listed in the '-rand' option. The function as previously
6412 found in genrsa is now in app_rand.c and is used by all programs
6413 that support '-rand'.
6416 *) In RAND_write_file, use mode 0600 for creating files;
6417 don't just chmod when it may be too late.
6420 *) Report an error from X509_STORE_load_locations
6421 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6424 *) New function ASN1_mbstring_copy() this copies a string in either
6425 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6426 into an ASN1_STRING type. A mask of permissible types is passed
6427 and it chooses the "minimal" type to use or an error if not type
6431 *) Add function equivalents to the various macros in asn1.h. The old
6432 macros are retained with an M_ prefix. Code inside the library can
6433 use the M_ macros. External code (including the openssl utility)
6434 should *NOT* in order to be "shared library friendly".
6437 *) Add various functions that can check a certificate's extensions
6438 to see if it usable for various purposes such as SSL client,
6439 server or S/MIME and CAs of these types. This is currently
6440 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6441 verification. Also added a -purpose flag to x509 utility to
6442 print out all the purposes.
6445 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6449 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6450 for, obtain and decode and extension and obtain its critical flag.
6451 This allows all the necessary extension code to be handled in a
6452 single function call.
6455 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6456 platforms. See crypto/rc4/rc4_enc.c for further details.
6459 *) New -noout option to asn1parse. This causes no output to be produced
6460 its main use is when combined with -strparse and -out to extract data
6461 from a file (which may not be in ASN.1 format).
6464 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6465 when producing the local key id.
6466 [Richard Levitte <levitte@stacken.kth.se>]
6468 *) New option -dhparam in s_server. This allows a DH parameter file to be
6469 stated explicitly. If it is not stated then it tries the first server
6470 certificate file. The previous behaviour hard coded the filename
6474 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6475 a public key to be input or output. For example:
6476 openssl rsa -in key.pem -pubout -out pubkey.pem
6477 Also added necessary DSA public key functions to handle this.
6480 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6481 in the message. This was handled by allowing
6482 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6483 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6485 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6486 to the end of the strings whereas this didn't. This would cause problems
6487 if strings read with d2i_ASN1_bytes() were later modified.
6488 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6490 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6491 data and it contains EOF it will end up returning an error. This is
6492 caused by input 46 bytes long. The cause is due to the way base64
6493 BIOs find the start of base64 encoded data. They do this by trying a
6494 trial decode on each line until they find one that works. When they
6495 do a flag is set and it starts again knowing it can pass all the
6496 data directly through the decoder. Unfortunately it doesn't reset
6497 the context it uses. This means that if EOF is reached an attempt
6498 is made to pass two EOFs through the context and this causes the
6499 resulting error. This can also cause other problems as well. As is
6500 usual with these problems it takes *ages* to find and the fix is
6501 trivial: move one line.
6502 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6504 *) Ugly workaround to get s_client and s_server working under Windows. The
6505 old code wouldn't work because it needed to select() on sockets and the
6506 tty (for keypresses and to see if data could be written). Win32 only
6507 supports select() on sockets so we select() with a 1s timeout on the
6508 sockets and then see if any characters are waiting to be read, if none
6509 are present then we retry, we also assume we can always write data to
6510 the tty. This isn't nice because the code then blocks until we've
6511 received a complete line of data and it is effectively polling the
6512 keyboard at 1s intervals: however it's quite a bit better than not
6513 working at all :-) A dedicated Windows application might handle this
6514 with an event loop for example.
6517 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6518 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6519 will be called when RSA_sign() and RSA_verify() are used. This is useful
6520 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6521 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6522 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6523 This necessitated the support of an extra signature type NID_md5_sha1
6524 for SSL signatures and modifications to the SSL library to use it instead
6525 of calling RSA_public_decrypt() and RSA_private_encrypt().
6528 *) Add new -verify -CAfile and -CApath options to the crl program, these
6529 will lookup a CRL issuers certificate and verify the signature in a
6530 similar way to the verify program. Tidy up the crl program so it
6531 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6532 less strict. It will now permit CRL extensions even if it is not
6533 a V2 CRL: this will allow it to tolerate some broken CRLs.
6536 *) Initialize all non-automatic variables each time one of the openssl
6537 sub-programs is started (this is necessary as they may be started
6538 multiple times from the "OpenSSL>" prompt).
6539 [Lennart Bang, Bodo Moeller]
6541 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6542 removing all other RSA functionality (this is what NO_RSA does). This
6543 is so (for example) those in the US can disable those operations covered
6544 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6548 *) Non-copying interface to BIO pairs.
6549 (still largely untested)
6552 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6553 ASCII string. This was handled independently in various places before.
6556 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6557 UTF8 strings a character at a time.
6560 *) Use client_version from client hello to select the protocol
6561 (s23_srvr.c) and for RSA client key exchange verification
6562 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6565 *) Add various utility functions to handle SPKACs, these were previously
6566 handled by poking round in the structure internals. Added new function
6567 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6568 print, verify and generate SPKACs. Based on an original idea from
6569 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6572 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6575 *) Allow the config file extension section to be overwritten on the
6576 command line. Based on an original idea from Massimiliano Pala
6577 <madwolf@comune.modena.it>. The new option is called -extensions
6578 and can be applied to ca, req and x509. Also -reqexts to override
6579 the request extensions in req and -crlexts to override the crl extensions
6583 *) Add new feature to the SPKAC handling in ca. Now you can include
6584 the same field multiple times by preceding it by "XXXX." for example:
6587 this is the same syntax as used in the req config file.
6590 *) Allow certificate extensions to be added to certificate requests. These
6591 are specified in a 'req_extensions' option of the req section of the
6592 config file. They can be printed out with the -text option to req but
6593 are otherwise ignored at present.
6596 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6597 data read consists of only the final block it would not decrypted because
6598 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6599 A misplaced 'break' also meant the decrypted final block might not be
6600 copied until the next read.
6603 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6604 a few extra parameters to the DH structure: these will be useful if
6605 for example we want the value of 'q' or implement X9.42 DH.
6608 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6609 provides hooks that allow the default DSA functions or functions on a
6610 "per key" basis to be replaced. This allows hardware acceleration and
6611 hardware key storage to be handled without major modification to the
6612 library. Also added low level modexp hooks and CRYPTO_EX structure and
6613 associated functions.
6616 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6617 as "read only": it can't be written to and the buffer it points to will
6618 not be freed. Reading from a read only BIO is much more efficient than
6619 a normal memory BIO. This was added because there are several times when
6620 an area of memory needs to be read from a BIO. The previous method was
6621 to create a memory BIO and write the data to it, this results in two
6622 copies of the data and an O(n^2) reading algorithm. There is a new
6623 function BIO_new_mem_buf() which creates a read only memory BIO from
6624 an area of memory. Also modified the PKCS#7 routines to use read only
6628 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6629 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6630 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6631 but a retry condition occured while trying to read the rest.
6634 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6635 NID_pkcs7_encrypted by default: this was wrong since this should almost
6636 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6637 the encrypted data type: this is a more sensible place to put it and it
6638 allows the PKCS#12 code to be tidied up that duplicated this
6642 *) Changed obj_dat.pl script so it takes its input and output files on
6643 the command line. This should avoid shell escape redirection problems
6647 *) Initial support for certificate extension requests, these are included
6648 in things like Xenroll certificate requests. Included functions to allow
6649 extensions to be obtained and added.
6652 *) -crlf option to s_client and s_server for sending newlines as
6653 CRLF (as required by many protocols).
6656 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6658 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6659 [Ralf S. Engelschall]
6661 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6662 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6664 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6668 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6669 DH parameters/keys (q is lost during that conversion, but the resulting
6670 DH parameters contain its length).
6672 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6673 much faster than DH_generate_parameters (which creates parameters
6674 where p = 2*q + 1), and also the smaller q makes DH computations
6675 much more efficient (160-bit exponentiation instead of 1024-bit
6676 exponentiation); so this provides a convenient way to support DHE
6677 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6678 utter importance to use
6679 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6681 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6682 when such DH parameters are used, because otherwise small subgroup
6683 attacks may become possible!
6686 *) Avoid memory leak in i2d_DHparams.
6689 *) Allow the -k option to be used more than once in the enc program:
6690 this allows the same encrypted message to be read by multiple recipients.
6693 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6694 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6695 it will always use the numerical form of the OID, even if it has a short
6699 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6700 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6701 otherwise bn_mod_exp was called. In the case of hardware keys for example
6702 no private key components need be present and it might store extra data
6703 in the RSA structure, which cannot be accessed from bn_mod_exp.
6704 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6705 private key operations.
6708 *) Added support for SPARC Linux.
6711 *) pem_password_cb function type incompatibly changed from
6712 typedef int pem_password_cb(char *buf, int size, int rwflag);
6714 ....(char *buf, int size, int rwflag, void *userdata);
6715 so that applications can pass data to their callbacks:
6716 The PEM[_ASN1]_{read,write}... functions and macros now take an
6717 additional void * argument, which is just handed through whenever
6718 the password callback is called.
6719 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6721 New function SSL_CTX_set_default_passwd_cb_userdata.
6723 Compatibility note: As many C implementations push function arguments
6724 onto the stack in reverse order, the new library version is likely to
6725 interoperate with programs that have been compiled with the old
6726 pem_password_cb definition (PEM_whatever takes some data that
6727 happens to be on the stack as its last argument, and the callback
6728 just ignores this garbage); but there is no guarantee whatsoever that
6731 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6732 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6733 problems not only on Windows, but also on some Unix platforms.
6734 To avoid problematic command lines, these definitions are now in an
6735 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6736 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6739 *) MIPS III/IV assembler module is reimplemented.
6742 *) More DES library cleanups: remove references to srand/rand and
6743 delete an unused file.
6746 *) Add support for the the free Netwide assembler (NASM) under Win32,
6747 since not many people have MASM (ml) and it can be hard to obtain.
6748 This is currently experimental but it seems to work OK and pass all
6749 the tests. Check out INSTALL.W32 for info.
6752 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6753 without temporary keys kept an extra copy of the server key,
6754 and connections with temporary keys did not free everything in case
6758 *) New function RSA_check_key and new openssl rsa option -check
6759 for verifying the consistency of RSA keys.
6760 [Ulf Moeller, Bodo Moeller]
6762 *) Various changes to make Win32 compile work:
6763 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6764 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6765 comparison" warnings.
6766 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6769 *) Add a debugging option to PKCS#5 v2 key generation function: when
6770 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6771 derived keys are printed to stderr.
6774 *) Copy the flags in ASN1_STRING_dup().
6775 [Roman E. Pavlov <pre@mo.msk.ru>]
6777 *) The x509 application mishandled signing requests containing DSA
6778 keys when the signing key was also DSA and the parameters didn't match.
6780 It was supposed to omit the parameters when they matched the signing key:
6781 the verifying software was then supposed to automatically use the CA's
6782 parameters if they were absent from the end user certificate.
6784 Omitting parameters is no longer recommended. The test was also
6785 the wrong way round! This was probably due to unusual behaviour in
6786 EVP_cmp_parameters() which returns 1 if the parameters match.
6787 This meant that parameters were omitted when they *didn't* match and
6788 the certificate was useless. Certificates signed with 'ca' didn't have
6790 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6792 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6793 The interface is as follows:
6794 Applications can use
6795 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6796 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6797 "off" is now the default.
6798 The library internally uses
6799 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6800 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6801 to disable memory-checking temporarily.
6803 Some inconsistent states that previously were possible (and were
6804 even the default) are now avoided.
6806 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6807 with each memory chunk allocated; this is occasionally more helpful
6808 than just having a counter.
6810 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6812 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6816 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6817 which largely parallels "options", but is for changing API behaviour,
6818 whereas "options" are about protocol behaviour.
6819 Initial "mode" flags are:
6821 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6822 a single record has been written.
6823 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6824 retries use the same buffer location.
6825 (But all of the contents must be
6829 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6832 *) Fix problems with no-hmac etc.
6833 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6835 *) New functions RSA_get_default_method(), RSA_set_method() and
6836 RSA_get_method(). These allows replacement of RSA_METHODs without having
6837 to mess around with the internals of an RSA structure.
6840 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6841 Also really enable memory leak checks in openssl.c and in some
6843 [Chad C. Mulligan, Bodo Moeller]
6845 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6846 up the length of negative integers. This has now been simplified to just
6847 store the length when it is first determined and use it later, rather
6848 than trying to keep track of where data is copied and updating it to
6850 [Steve Henson, reported by Brien Wheeler
6851 <bwheeler@authentica-security.com>]
6853 *) Add a new function PKCS7_signatureVerify. This allows the verification
6854 of a PKCS#7 signature but with the signing certificate passed to the
6855 function itself. This contrasts with PKCS7_dataVerify which assumes the
6856 certificate is present in the PKCS#7 structure. This isn't always the
6857 case: certificates can be omitted from a PKCS#7 structure and be
6858 distributed by "out of band" means (such as a certificate database).
6861 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6862 function prototypes in pem.h, also change util/mkdef.pl to add the
6863 necessary function names.
6866 *) mk1mf.pl (used by Windows builds) did not properly read the
6867 options set by Configure in the top level Makefile, and Configure
6868 was not even able to write more than one option correctly.
6869 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6872 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6873 file to be loaded from a BIO or FILE pointer. The BIO version will
6874 for example allow memory BIOs to contain config info.
6877 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6878 Whoever hopes to achieve shared-library compatibility across versions
6879 must use this, not the compile-time macro.
6880 (Exercise 0.9.4: Which is the minimum library version required by
6882 Note: All this applies only to multi-threaded programs, others don't
6886 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6887 through a BIO pair triggered the default case, i.e.
6888 SSLerr(...,SSL_R_UNKNOWN_STATE).
6891 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6892 can use the SSL library even if none of the specific BIOs is
6896 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6897 for the encoded length.
6898 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6900 *) Add initial documentation of the X509V3 functions.
6903 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6904 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6905 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6906 secure PKCS#8 private key format with a high iteration count.
6909 *) Fix determination of Perl interpreter: A perl or perl5
6910 _directory_ in $PATH was also accepted as the interpreter.
6911 [Ralf S. Engelschall]
6913 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6914 wrong with it but it was very old and did things like calling
6915 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6919 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6920 to use the new extension code.
6923 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6924 with macros. This should make it easier to change their form, add extra
6925 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6929 *) Add to configuration table a new entry that can specify an alternative
6930 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6931 according to Mark Crispin <MRC@Panda.COM>.
6935 *) DES CBC did not update the IV. Weird.
6938 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6939 Changing the behaviour of the former might break existing programs --
6940 where IV updating is needed, des_ncbc_encrypt can be used.
6943 *) When bntest is run from "make test" it drives bc to check its
6944 calculations, as well as internally checking them. If an internal check
6945 fails, it needs to cause bc to give a non-zero result or make test carries
6946 on without noticing the failure. Fixed.
6949 *) DES library cleanups.
6952 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6953 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6954 ciphers. NOTE: although the key derivation function has been verified
6955 against some published test vectors it has not been extensively tested
6956 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6960 *) Instead of "mkdir -p", which is not fully portable, use new
6961 Perl script "util/mkdir-p.pl".
6964 *) Rewrite the way password based encryption (PBE) is handled. It used to
6965 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6966 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6967 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6968 the 'parameter' field of the AlgorithmIdentifier is passed to the
6969 underlying key generation function so it must do its own ASN1 parsing.
6970 This has also changed the EVP_PBE_CipherInit() function which now has a
6971 'parameter' argument instead of literal salt and iteration count values
6972 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6975 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6976 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6977 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6978 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6979 value was just used as a "magic string" and not used directly its
6980 value doesn't matter.
6983 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6987 *) "linux-sparc64" configuration (ultrapenguin).
6988 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6989 "linux-sparc" configuration.
6990 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6992 *) config now generates no-xxx options for missing ciphers.
6995 *) Support the EBCDIC character set (work in progress).
6996 File ebcdic.c not yet included because it has a different license.
6997 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6999 *) Support BS2000/OSD-POSIX.
7000 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7002 *) Make callbacks for key generation use void * instead of char *.
7005 *) Make S/MIME samples compile (not yet tested).
7008 *) Additional typesafe stacks.
7011 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7015 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7017 *) New configuration variant "sco5-gcc".
7019 *) Updated some demos.
7020 [Sean O Riordain, Wade Scholine]
7022 *) Add missing BIO_free at exit of pkcs12 application.
7025 *) Fix memory leak in conf.c.
7028 *) Updates for Win32 to assembler version of MD5.
7031 *) Set #! path to perl in apps/der_chop to where we found it
7032 instead of using a fixed path.
7035 *) SHA library changes for irix64-mips4-cc.
7038 *) Improvements for VMS support.
7042 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7044 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7045 This also avoids the problems with SC4.2 and unpatched SC5.
7046 [Andy Polyakov <appro@fy.chalmers.se>]
7048 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7049 These are required because of the typesafe stack would otherwise break
7050 existing code. If old code used a structure member which used to be STACK
7051 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7052 sk_num or sk_value it would produce an error because the num, data members
7053 are not present in STACK_OF. Now it just produces a warning. sk_set
7054 replaces the old method of assigning a value to sk_value
7055 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7056 that does this will no longer work (and should use sk_set instead) but
7057 this could be regarded as a "questionable" behaviour anyway.
7060 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7061 correctly handle encrypted S/MIME data.
7064 *) Change type of various DES function arguments from des_cblock
7065 (which means, in function argument declarations, pointer to char)
7066 to des_cblock * (meaning pointer to array with 8 char elements),
7067 which allows the compiler to do more typechecking; it was like
7068 that back in SSLeay, but with lots of ugly casts.
7070 Introduce new type const_des_cblock.
7073 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7074 problems: find RecipientInfo structure that matches recipient certificate
7075 and initialise the ASN1 structures properly based on passed cipher.
7078 *) Belatedly make the BN tests actually check the results.
7081 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7082 to and from BNs: it was completely broken. New compilation option
7083 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7084 key elements as negative integers.
7087 *) Reorganize and speed up MD5.
7088 [Andy Polyakov <appro@fy.chalmers.se>]
7091 [Richard Levitte <richard@levitte.org>]
7093 *) New option -out to asn1parse to allow the parsed structure to be
7094 output to a file. This is most useful when combined with the -strparse
7095 option to examine the output of things like OCTET STRINGS.
7098 *) Make SSL library a little more fool-proof by not requiring any longer
7099 that SSL_set_{accept,connect}_state be called before
7100 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7101 in many applications because usually everything *appeared* to work as
7102 intended anyway -- now it really works as intended).
7105 *) Move openssl.cnf out of lib/.
7108 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7109 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7110 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7111 [Ralf S. Engelschall]
7113 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7114 handle PKCS#7 enveloped data properly.
7115 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7117 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7118 copying pointers. The cert_st handling is changed by this in
7119 various ways (and thus what used to be known as ctx->default_cert
7120 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7121 any longer when s->cert does not give us what we need).
7122 ssl_cert_instantiate becomes obsolete by this change.
7123 As soon as we've got the new code right (possibly it already is?),
7124 we have solved a couple of bugs of the earlier code where s->cert
7125 was used as if it could not have been shared with other SSL structures.
7127 Note that using the SSL API in certain dirty ways now will result
7128 in different behaviour than observed with earlier library versions:
7129 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7130 does not influence s as it used to.
7132 In order to clean up things more thoroughly, inside SSL_SESSION
7133 we don't use CERT any longer, but a new structure SESS_CERT
7134 that holds per-session data (if available); currently, this is
7135 the peer's certificate chain and, for clients, the server's certificate
7136 and temporary key. CERT holds only those values that can have
7137 meaningful defaults in an SSL_CTX.
7140 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7141 from the internal representation. Various PKCS#7 fixes: remove some
7142 evil casts and set the enc_dig_alg field properly based on the signing
7146 *) Allow PKCS#12 password to be set from the command line or the
7147 environment. Let 'ca' get its config file name from the environment
7148 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7152 *) Allow certificate policies extension to use an IA5STRING for the
7153 organization field. This is contrary to the PKIX definition but
7154 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7158 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7159 without disallowing inline assembler and the like for non-pedantic builds.
7162 *) Support Borland C++ builder.
7163 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7168 *) SHA-1 cleanups and performance enhancements.
7169 [Andy Polyakov <appro@fy.chalmers.se>]
7171 *) Sparc v8plus assembler for the bignum library.
7172 [Andy Polyakov <appro@fy.chalmers.se>]
7174 *) Accept any -xxx and +xxx compiler options in Configure.
7177 *) Update HPUX configuration.
7180 *) Add missing sk_<type>_unshift() function to safestack.h
7181 [Ralf S. Engelschall]
7183 *) New function SSL_CTX_use_certificate_chain_file that sets the
7184 "extra_cert"s in addition to the certificate. (This makes sense
7185 only for "PEM" format files, as chains as a whole are not
7189 *) Support verify_depth from the SSL API.
7190 x509_vfy.c had what can be considered an off-by-one-error:
7191 Its depth (which was not part of the external interface)
7192 was actually counting the number of certificates in a chain;
7193 now it really counts the depth.
7196 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7197 instead of X509err, which often resulted in confusing error
7198 messages since the error codes are not globally unique
7199 (e.g. an alleged error in ssl3_accept when a certificate
7200 didn't match the private key).
7202 *) New function SSL_CTX_set_session_id_context that allows to set a default
7203 value (so that you don't need SSL_set_session_id_context for each
7204 connection using the SSL_CTX).
7207 *) OAEP decoding bug fix.
7210 *) Support INSTALL_PREFIX for package builders, as proposed by
7214 *) New Configure options "threads" and "no-threads". For systems
7215 where the proper compiler options are known (currently Solaris
7216 and Linux), "threads" is the default.
7219 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7222 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7223 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7224 such as /usr/local/bin.
7227 *) "make linux-shared" to build shared libraries.
7228 [Niels Poppe <niels@netbox.org>]
7230 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7233 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7234 extension adding in x509 utility.
7237 *) Remove NOPROTO sections and error code comments.
7240 *) Partial rewrite of the DEF file generator to now parse the ANSI
7244 *) New Configure options --prefix=DIR and --openssldir=DIR.
7247 *) Complete rewrite of the error code script(s). It is all now handled
7248 by one script at the top level which handles error code gathering,
7249 header rewriting and C source file generation. It should be much better
7250 than the old method: it now uses a modified version of Ulf's parser to
7251 read the ANSI prototypes in all header files (thus the old K&R definitions
7252 aren't needed for error creation any more) and do a better job of
7253 translating function codes into names. The old 'ASN1 error code imbedded
7254 in a comment' is no longer necessary and it doesn't use .err files which
7255 have now been deleted. Also the error code call doesn't have to appear all
7256 on one line (which resulted in some large lines...).
7259 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7262 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7263 0 (which usually indicates a closed connection), but continue reading.
7266 *) Fix some race conditions.
7269 *) Add support for CRL distribution points extension. Add Certificate
7270 Policies and CRL distribution points documentation.
7273 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7276 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7277 8 of keying material. Merlin has also confirmed interop with this fix
7278 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7279 [Merlin Hughes <merlin@baltimore.ie>]
7281 *) Fix lots of warnings.
7282 [Richard Levitte <levitte@stacken.kth.se>]
7284 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7285 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7286 [Richard Levitte <levitte@stacken.kth.se>]
7288 *) Fix problems with sizeof(long) == 8.
7289 [Andy Polyakov <appro@fy.chalmers.se>]
7291 *) Change functions to ANSI C.
7294 *) Fix typos in error codes.
7295 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7297 *) Remove defunct assembler files from Configure.
7300 *) SPARC v8 assembler BIGNUM implementation.
7301 [Andy Polyakov <appro@fy.chalmers.se>]
7303 *) Support for Certificate Policies extension: both print and set.
7304 Various additions to support the r2i method this uses.
7307 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7308 return a const string when you are expecting an allocated buffer.
7311 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7312 types DirectoryString and DisplayText.
7315 *) Add code to allow r2i extensions to access the configuration database,
7316 add an LHASH database driver and add several ctx helper functions.
7319 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7320 fail when they extended the size of a BIGNUM.
7323 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7324 support typesafe stack.
7327 *) Fix typo in SSL_[gs]et_options().
7328 [Nils Frostberg <nils@medcom.se>]
7330 *) Delete various functions and files that belonged to the (now obsolete)
7331 old X509V3 handling code.
7334 *) New Configure option "rsaref".
7337 *) Don't auto-generate pem.h.
7340 *) Introduce type-safe ASN.1 SETs.
7343 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7344 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7346 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7347 that links with OpenSSL (well at least cause lots of warnings), but fear
7348 not: the conversion is trivial, and it eliminates loads of evil casts. A
7349 few STACKed things have been converted already. Feel free to convert more.
7350 In the fullness of time, I'll do away with the STACK type altogether.
7353 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7354 specified in <certfile> by updating the entry in the index.txt file.
7355 This way one no longer has to edit the index.txt file manually for
7356 revoking a certificate. The -revoke option does the gory details now.
7357 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7359 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7360 `-text' option at all and this way the `-noout -text' combination was
7361 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7362 [Ralf S. Engelschall]
7364 *) Make sure a corresponding plain text error message exists for the
7365 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7366 verify callback function determined that a certificate was revoked.
7367 [Ralf S. Engelschall]
7369 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7370 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7371 all available cipers including rc5, which was forgotten until now.
7372 In order to let the testing shell script know which algorithms
7373 are available, a new (up to now undocumented) command
7374 "openssl list-cipher-commands" is used.
7377 *) Bugfix: s_client occasionally would sleep in select() when
7378 it should have checked SSL_pending() first.
7381 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7382 the raw DSA values prior to ASN.1 encoding.
7385 *) Tweaks to Configure
7386 [Niels Poppe <niels@netbox.org>]
7388 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7392 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7395 *) New config option to avoid instructions that are illegal on the 80386.
7396 The default code is faster, but requires at least a 486.
7399 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7400 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7401 same as SSL2_VERSION anyway.
7404 *) New "-showcerts" option for s_client.
7407 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7408 application. Various cleanups and fixes.
7411 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7412 modify error routines to work internally. Add error codes and PBE init
7413 to library startup routines.
7416 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7417 packing functions to asn1 and evp. Changed function names and error
7418 codes along the way.
7421 *) PKCS12 integration: and so it begins... First of several patches to
7422 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7423 objects to objects.h
7426 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7427 and display support for Thawte strong extranet extension.
7430 *) Add LinuxPPC support.
7431 [Jeff Dubrule <igor@pobox.org>]
7433 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7434 bn_div_words in alpha.s.
7435 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7437 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7438 OAEP isn't supported when OpenSSL is built with RSAref.
7439 [Ulf Moeller <ulf@fitug.de>]
7441 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7442 so they no longer are missing under -DNOPROTO.
7443 [Soren S. Jorvang <soren@t.dk>]
7446 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7448 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7449 doesn't work when the session is reused. Coming soon!
7452 *) Fix a security hole, that allows sessions to be reused in the wrong
7453 context thus bypassing client cert protection! All software that uses
7454 client certs and session caches in multiple contexts NEEDS PATCHING to
7455 allow session reuse! A fuller solution is in the works.
7456 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7458 *) Some more source tree cleanups (removed obsolete files
7459 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7460 permission on "config" script to be executable) and a fix for the INSTALL
7462 [Ulf Moeller <ulf@fitug.de>]
7464 *) Remove some legacy and erroneous uses of malloc, free instead of
7466 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7468 *) Make rsa_oaep_test return non-zero on error.
7469 [Ulf Moeller <ulf@fitug.de>]
7471 *) Add support for native Solaris shared libraries. Configure
7472 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7473 if someone would make that last step automatic.
7474 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7476 *) ctx_size was not built with the right compiler during "make links". Fixed.
7479 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7480 except NULL ciphers". This means the default cipher list will no longer
7481 enable NULL ciphers. They need to be specifically enabled e.g. with
7482 the string "DEFAULT:eNULL".
7485 *) Fix to RSA private encryption routines: if p < q then it would
7486 occasionally produce an invalid result. This will only happen with
7487 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7490 *) Be less restrictive and allow also `perl util/perlpath.pl
7491 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7492 because this way one can also use an interpreter named `perl5' (which is
7493 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7494 installed as `perl').
7495 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7497 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7498 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7500 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7501 advapi32.lib to Win32 build and change the pem test comparision
7502 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7503 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7504 and crypto/des/ede_cbcm_enc.c.
7507 *) DES quad checksum was broken on big-endian architectures. Fixed.
7510 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7511 Win32 test batch file so it (might) work again. The Win32 test batch file
7512 is horrible: I feel ill....
7515 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7516 in e_os.h. Audit of header files to check ANSI and non ANSI
7517 sections: 10 functions were absent from non ANSI section and not exported
7518 from Windows DLLs. Fixed up libeay.num for new functions.
7521 *) Make `openssl version' output lines consistent.
7522 [Ralf S. Engelschall]
7524 *) Fix Win32 symbol export lists for BIO functions: Added
7525 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7526 to ms/libeay{16,32}.def.
7527 [Ralf S. Engelschall]
7529 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7530 fine under Unix and passes some trivial tests I've now added. But the
7531 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7532 added to make sure no one expects that this stuff really works in the
7533 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7534 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7536 [Ralf S. Engelschall]
7538 *) Fix the generation of two part addresses in perl.
7539 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7541 *) Add config entry for Linux on MIPS.
7542 [John Tobey <jtobey@channel1.com>]
7544 *) Make links whenever Configure is run, unless we are on Windoze.
7547 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7548 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7552 *) Add a useful kludge to allow package maintainers to specify compiler and
7553 other platforms details on the command line without having to patch the
7554 Configure script everytime: One now can use ``perl Configure
7555 <id>:<details>'', i.e. platform ids are allowed to have details appended
7556 to them (seperated by colons). This is treated as there would be a static
7557 pre-configured entry in Configure's %table under key <id> with value
7558 <details> and ``perl Configure <id>'' is called. So, when you want to
7559 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7560 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7561 now, which overrides the FreeBSD-elf entry on-the-fly.
7562 [Ralf S. Engelschall]
7564 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7567 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7568 on the `perl Configure ...' command line. This way one can compile
7569 OpenSSL libraries with Position Independent Code (PIC) which is needed
7570 for linking it into DSOs.
7571 [Ralf S. Engelschall]
7573 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7577 *) Cleaned up the LICENSE document: The official contact for any license
7578 questions now is the OpenSSL core team under openssl-core@openssl.org.
7579 And add a paragraph about the dual-license situation to make sure people
7580 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7581 to the OpenSSL toolkit.
7582 [Ralf S. Engelschall]
7584 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7585 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7586 Additonally cleaned up the `make links' target: Remove unnecessary
7587 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7588 to speed processing and no longer clutter the display with confusing
7589 stuff. Instead only the actually done links are displayed.
7590 [Ralf S. Engelschall]
7592 *) Permit null encryption ciphersuites, used for authentication only. It used
7593 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7594 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7598 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7599 signed attributes when verifying signatures (this would break them),
7600 the detached data encoding was wrong and public keys obtained using
7601 X509_get_pubkey() weren't freed.
7604 *) Add text documentation for the BUFFER functions. Also added a work around
7605 to a Win95 console bug. This was triggered by the password read stuff: the
7606 last character typed gets carried over to the next fread(). If you were
7607 generating a new cert request using 'req' for example then the last
7608 character of the passphrase would be CR which would then enter the first
7612 *) Added the new `Includes OpenSSL Cryptography Software' button as
7613 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7614 button and can be used by applications based on OpenSSL to show the
7615 relationship to the OpenSSL project.
7616 [Ralf S. Engelschall]
7618 *) Remove confusing variables in function signatures in files
7619 ssl/ssl_lib.c and ssl/ssl.h.
7620 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7622 *) Don't install bss_file.c under PREFIX/include/
7623 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7625 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7626 functions that return function pointers and has support for NT specific
7627 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7628 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7629 unsigned to signed types: this was killing the Win32 compile.
7632 *) Add new certificate file to stack functions,
7633 SSL_add_dir_cert_subjects_to_stack() and
7634 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7635 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7636 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7637 This means that Apache-SSL and similar packages don't have to mess around
7638 to add as many CAs as they want to the preferred list.
7641 *) Experiment with doxygen documentation. Currently only partially applied to
7643 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7644 openssl.doxy as the configuration file.
7647 *) Get rid of remaining C++-style comments which strict C compilers hate.
7648 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7650 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7651 compiled in by default: it has problems with large keys.
7654 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7655 DH private keys and/or callback functions which directly correspond to
7656 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7657 is needed for applications which have to configure certificates on a
7658 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7660 For the RSA certificate situation is makes no difference, but
7661 for the DSA certificate situation this fixes the "no shared cipher"
7662 problem where the OpenSSL cipher selection procedure failed because the
7663 temporary keys were not overtaken from the context and the API provided
7664 no way to reconfigure them.
7665 The new functions now let applications reconfigure the stuff and they
7666 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7667 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7668 non-public-API function ssl_cert_instantiate() is used as a helper
7669 function and also to reduce code redundancy inside ssl_rsa.c.
7670 [Ralf S. Engelschall]
7672 *) Move s_server -dcert and -dkey options out of the undocumented feature
7673 area because they are useful for the DSA situation and should be
7674 recognized by the users.
7675 [Ralf S. Engelschall]
7677 *) Fix the cipher decision scheme for export ciphers: the export bits are
7678 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7679 SSL_EXP_MASK. So, the original variable has to be used instead of the
7680 already masked variable.
7681 [Richard Levitte <levitte@stacken.kth.se>]
7683 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7684 [Richard Levitte <levitte@stacken.kth.se>]
7686 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7687 from `int' to `unsigned int' because it's a length and initialized by
7688 EVP_DigestFinal() which expects an `unsigned int *'.
7689 [Richard Levitte <levitte@stacken.kth.se>]
7691 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7692 script. Instead use the usual Shell->Perl transition trick.
7693 [Ralf S. Engelschall]
7695 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7696 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7697 -noout -modulus' as it's already the case for `openssl rsa -noout
7698 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7699 currently the public key is printed (a decision which was already done by
7700 `openssl dsa -modulus' in the past) which serves a similar purpose.
7701 Additionally the NO_RSA no longer completely removes the whole -modulus
7702 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7704 [Ralf S. Engelschall]
7706 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7707 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7708 [Arne Ansper <arne@ats.cyber.ee>]
7710 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7711 to be added. Now both 'req' and 'ca' can use new objects defined in the
7715 *) Add cool BIO that does syslog (or event log on NT).
7716 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7718 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7719 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7720 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7721 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7724 *) Add preliminary config info for new extension code.
7727 *) Make RSA_NO_PADDING really use no padding.
7728 [Ulf Moeller <ulf@fitug.de>]
7730 *) Generate errors when private/public key check is done.
7733 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7734 for some CRL extensions and new objects added.
7737 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7738 key usage extension and fuller support for authority key id.
7741 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7742 padding method for RSA, which is recommended for new applications in PKCS
7743 #1 v2.0 (RFC 2437, October 1998).
7744 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7745 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7746 against Bleichbacher's attack on RSA.
7747 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7750 *) Updates to the new SSL compression code
7751 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7753 *) Fix so that the version number in the master secret, when passed
7754 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7755 (because the server will not accept higher), that the version number
7756 is 0x03,0x01, not 0x03,0x00
7757 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7759 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7760 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7761 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7764 *) Support for RAW extensions where an arbitrary extension can be
7765 created by including its DER encoding. See apps/openssl.cnf for
7769 *) Make sure latest Perl versions don't interpret some generated C array
7770 code as Perl array code in the crypto/err/err_genc.pl script.
7771 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7773 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7774 not many people have the assembler. Various Win32 compilation fixes and
7775 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7779 *) Modify configure script 'Configure' to automatically create crypto/date.h
7780 file under Win32 and also build pem.h from pem.org. New script
7781 util/mkfiles.pl to create the MINFO file on environments that can't do a
7782 'make files': perl util/mkfiles.pl >MINFO should work.
7785 *) Major rework of DES function declarations, in the pursuit of correctness
7786 and purity. As a result, many evil casts evaporated, and some weirdness,
7787 too. You may find this causes warnings in your code. Zapping your evil
7788 casts will probably fix them. Mostly.
7791 *) Fix for a typo in asn1.h. Bug fix to object creation script
7792 obj_dat.pl. It considered a zero in an object definition to mean
7793 "end of object": none of the objects in objects.h have any zeros
7794 so it wasn't spotted.
7795 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7797 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7798 Masking (CBCM). In the absence of test vectors, the best I have been able
7799 to do is check that the decrypt undoes the encrypt, so far. Send me test
7800 vectors if you have them.
7803 *) Correct calculation of key length for export ciphers (too much space was
7804 allocated for null ciphers). This has not been tested!
7807 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7808 message is now correct (it understands "crypto" and "ssl" on its
7809 command line). There is also now an "update" option. This will update
7810 the util/ssleay.num and util/libeay.num files with any new functions.
7812 perl util/mkdef.pl crypto ssl update
7813 it will update them.
7816 *) Overhauled the Perl interface (perl/*):
7817 - ported BN stuff to OpenSSL's different BN library
7818 - made the perl/ source tree CVS-aware
7819 - renamed the package from SSLeay to OpenSSL (the files still contain
7820 their history because I've copied them in the repository)
7821 - removed obsolete files (the test scripts will be replaced
7822 by better Test::Harness variants in the future)
7823 [Ralf S. Engelschall]
7825 *) First cut for a very conservative source tree cleanup:
7826 1. merge various obsolete readme texts into doc/ssleay.txt
7827 where we collect the old documents and readme texts.
7828 2. remove the first part of files where I'm already sure that we no
7829 longer need them because of three reasons: either they are just temporary
7830 files which were left by Eric or they are preserved original files where
7831 I've verified that the diff is also available in the CVS via "cvs diff
7832 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7833 the crypto/md/ stuff).
7834 [Ralf S. Engelschall]
7836 *) More extension code. Incomplete support for subject and issuer alt
7837 name, issuer and authority key id. Change the i2v function parameters
7838 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7839 what that's for :-) Fix to ASN1 macro which messed up
7840 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7843 *) Preliminary support for ENUMERATED type. This is largely copied from the
7847 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7848 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7850 *) Make sure `make rehash' target really finds the `openssl' program.
7851 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7853 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7854 like to hear about it if this slows down other processors.
7857 *) Add CygWin32 platform information to Configure script.
7858 [Alan Batie <batie@aahz.jf.intel.com>]
7860 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7861 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7863 *) New program nseq to manipulate netscape certificate sequences
7866 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7870 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7871 but the BN code had some problems that would cause failures when
7872 doing certificate verification and some other functions.
7873 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7875 *) Add ASN1 and PEM code to support netscape certificate sequences.
7878 *) Add ASN1 and PEM code to support netscape certificate sequences.
7881 *) Add several PKIX and private extended key usage OIDs.
7884 *) Modify the 'ca' program to handle the new extension code. Modify
7885 openssl.cnf for new extension format, add comments.
7888 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7889 and add a sample to openssl.cnf so req -x509 now adds appropriate
7893 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7894 error code, add initial support to X509_print() and x509 application.
7897 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7898 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7899 stuff is currently isolated and isn't even compiled yet.
7902 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7903 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7904 Removed the versions check from X509 routines when loading extensions:
7905 this allows certain broken certificates that don't set the version
7906 properly to be processed.
7909 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7910 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7911 can still be regenerated with "make depend".
7914 *) Spelling mistake in C version of CAST-128.
7915 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7917 *) Changes to the error generation code. The perl script err-code.pl
7918 now reads in the old error codes and retains the old numbers, only
7919 adding new ones if necessary. It also only changes the .err files if new
7920 codes are added. The makefiles have been modified to only insert errors
7921 when needed (to avoid needlessly modifying header files). This is done
7922 by only inserting errors if the .err file is newer than the auto generated
7923 C file. To rebuild all the error codes from scratch (the old behaviour)
7924 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7925 or delete all the .err files.
7928 *) CAST-128 was incorrectly implemented for short keys. The C version has
7929 been fixed, but is untested. The assembler versions are also fixed, but
7930 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7931 to regenerate it if needed.
7932 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7933 Hagino <itojun@kame.net>]
7935 *) File was opened incorrectly in randfile.c.
7936 [Ulf Möller <ulf@fitug.de>]
7938 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7939 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7940 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7941 al: it's just almost always a UTCTime. Note this patch adds new error
7942 codes so do a "make errors" if there are problems.
7945 *) Correct Linux 1 recognition in config.
7946 [Ulf Möller <ulf@fitug.de>]
7948 *) Remove pointless MD5 hash when using DSA keys in ca.
7949 [Anonymous <nobody@replay.com>]
7951 *) Generate an error if given an empty string as a cert directory. Also
7952 generate an error if handed NULL (previously returned 0 to indicate an
7953 error, but didn't set one).
7954 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7956 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7959 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7960 parameters. This was causing a warning which killed off the Win32 compile.
7963 *) Remove C++ style comments from crypto/bn/bn_local.h.
7964 [Neil Costigan <neil.costigan@celocom.com>]
7966 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7967 based on a text string, looking up short and long names and finally
7968 "dot" format. The "dot" format stuff didn't work. Added new function
7969 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7970 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7971 OID is not part of the table.
7974 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7975 X509_LOOKUP_by_alias().
7978 *) Sort openssl functions by name.
7981 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7982 encryption from sample DSA keys (in case anyone is interested the password
7986 *) Make _all_ *_free functions accept a NULL pointer.
7987 [Frans Heymans <fheymans@isaserver.be>]
7989 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7991 [Anonymous <nobody@replay.com>]
7993 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7994 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7996 *) Don't blow it for numeric -newkey arguments to apps/req.
7997 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7999 *) Temp key "for export" tests were wrong in s3_srvr.c.
8000 [Anonymous <nobody@replay.com>]
8002 *) Add prototype for temp key callback functions
8003 SSL_CTX_set_tmp_{rsa,dh}_callback().
8006 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8007 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8010 *) X509_name_add_entry() freed the wrong thing after an error.
8011 [Arne Ansper <arne@ats.cyber.ee>]
8013 *) rsa_eay.c would attempt to free a NULL context.
8014 [Arne Ansper <arne@ats.cyber.ee>]
8016 *) BIO_s_socket() had a broken should_retry() on Windoze.
8017 [Arne Ansper <arne@ats.cyber.ee>]
8019 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8020 [Arne Ansper <arne@ats.cyber.ee>]
8022 *) Make sure the already existing X509_STORE->depth variable is initialized
8023 in X509_STORE_new(), but document the fact that this variable is still
8024 unused in the certificate verification process.
8025 [Ralf S. Engelschall]
8027 *) Fix the various library and apps files to free up pkeys obtained from
8028 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8031 *) Fix reference counting in X509_PUBKEY_get(). This makes
8032 demos/maurice/example2.c work, amongst others, probably.
8033 [Steve Henson and Ben Laurie]
8035 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8036 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8037 are no longer created. This way we have a single and consistent command
8038 line interface `openssl <command>', similar to `cvs <command>'.
8039 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8041 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8042 BIT STRING wrapper always have zero unused bits.
8045 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8048 *) Make the top-level INSTALL documentation easier to understand.
8051 *) Makefiles updated to exit if an error occurs in a sub-directory
8052 make (including if user presses ^C) [Paul Sutton]
8054 *) Make Montgomery context stuff explicit in RSA data structure.
8057 *) Fix build order of pem and err to allow for generated pem.h.
8060 *) Fix renumbering bug in X509_NAME_delete_entry().
8063 *) Enhanced the err-ins.pl script so it makes the error library number
8064 global and can add a library name. This is needed for external ASN1 and
8065 other error libraries.
8068 *) Fixed sk_insert which never worked properly.
8071 *) Fix ASN1 macros so they can handle indefinite length construted
8072 EXPLICIT tags. Some non standard certificates use these: they can now
8076 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8077 into a single doc/ssleay.txt bundle. This way the information is still
8078 preserved but no longer messes up this directory. Now it's new room for
8079 the new set of documenation files.
8080 [Ralf S. Engelschall]
8082 *) SETs were incorrectly DER encoded. This was a major pain, because they
8083 shared code with SEQUENCEs, which aren't coded the same. This means that
8084 almost everything to do with SETs or SEQUENCEs has either changed name or
8085 number of arguments.
8086 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8088 *) Fix test data to work with the above.
8091 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8092 was already fixed by Eric for 0.9.1 it seems.
8093 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8095 *) Autodetect FreeBSD3.
8098 *) Fix various bugs in Configure. This affects the following platforms:
8102 unixware-2.0-pentium
8106 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8107 before they are needed.
8110 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8114 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8116 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8117 changed SSLeay to OpenSSL in version strings.
8118 [Ralf S. Engelschall]
8120 *) Some fixups to the top-level documents.
8123 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8124 because the symlink to include/ was missing.
8125 [Ralf S. Engelschall]
8127 *) Incorporated the popular no-RSA/DSA-only patches
8128 which allow to compile a RSA-free SSLeay.
8129 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8131 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8132 when "ssleay" is still not found.
8133 [Ralf S. Engelschall]
8135 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8136 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8138 *) Updated the README file.
8139 [Ralf S. Engelschall]
8141 *) Added various .cvsignore files in the CVS repository subdirs
8142 to make a "cvs update" really silent.
8143 [Ralf S. Engelschall]
8145 *) Recompiled the error-definition header files and added
8146 missing symbols to the Win32 linker tables.
8147 [Ralf S. Engelschall]
8149 *) Cleaned up the top-level documents;
8150 o new files: CHANGES and LICENSE
8151 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8152 o merged COPYRIGHT into LICENSE
8153 o removed obsolete TODO file
8154 o renamed MICROSOFT to INSTALL.W32
8155 [Ralf S. Engelschall]
8157 *) Removed dummy files from the 0.9.1b source tree:
8158 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8159 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8160 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8161 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8162 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8163 [Ralf S. Engelschall]
8165 *) Added various platform portability fixes.
8168 *) The Genesis of the OpenSSL rpject:
8169 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8170 Young and Tim J. Hudson created while they were working for C2Net until
8172 [The OpenSSL Project]
8175 Changes between 0.9.0b and 0.9.1b [not released]
8177 *) Updated a few CA certificates under certs/
8180 *) Changed some BIGNUM api stuff.
8183 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8184 DGUX x86, Linux Alpha, etc.
8187 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8188 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8192 *) Add -strparse option to asn1pars program which parses nested
8194 [Dr Stephen Henson <shenson@bigfoot.com>]
8196 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8199 *) DSA fix for "ca" program.
8202 *) Added "-genkey" option to "dsaparam" program.
8205 *) Added RIPE MD160 (rmd160) message digest.
8208 *) Added -a (all) option to "ssleay version" command.
8211 *) Added PLATFORM define which is the id given to Configure.
8214 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8217 *) Extended the ASN.1 parser routines.
8220 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8223 *) Added a BN_CTX to the BN library.
8226 *) Fixed the weak key values in DES library
8229 *) Changed API in EVP library for cipher aliases.
8232 *) Added support for RC2/64bit cipher.
8235 *) Converted the lhash library to the crypto/mem.c functions.
8238 *) Added more recognized ASN.1 object ids.
8241 *) Added more RSA padding checks for SSL/TLS.
8244 *) Added BIO proxy/filter functionality.
8247 *) Added extra_certs to SSL_CTX which can be used
8248 send extra CA certificates to the client in the CA cert chain sending
8249 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8252 *) Now Fortezza is denied in the authentication phase because
8253 this is key exchange mechanism is not supported by SSLeay at all.
8256 *) Additional PKCS1 checks.
8259 *) Support the string "TLSv1" for all TLS v1 ciphers.
8262 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8263 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8266 *) Fixed a few memory leaks.
8269 *) Fixed various code and comment typos.
8272 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8273 bytes sent in the client random.
8274 [Edward Bishop <ebishop@spyglass.com>]