5 Changes between 0.9.8h and 0.9.8i [xx XXX xxxx]
7 Changes between 0.9.8g and 0.9.8h [28 May 2008]
9 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
10 handshake which could lead to a cilent crash as found using the
11 Codenomicon TLS test suite (CVE-2008-1672)
12 [Steve Henson, Mark Cox]
14 *) Fix double free in TLS server name extensions which could lead to
15 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
18 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
20 Clear the error queue to ensure that error entries left from
21 older function calls do not interfere with the correct operation.
22 [Lutz Jaenicke, Erik de Castro Lopo]
24 *) Remove root CA certificates of commercial CAs:
26 The OpenSSL project does not recommend any specific CA and does not
27 have any policy with respect to including or excluding any CA.
28 Therefore it does not make any sense to ship an arbitrary selection
29 of root CA certificates with the OpenSSL software.
32 *) RSA OAEP patches to fix two separate invalid memory reads.
33 The first one involves inputs when 'lzero' is greater than
34 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
35 before the beginning of from). The second one involves inputs where
36 the 'db' section contains nothing but zeroes (there is a one-byte
37 invalid read after the end of 'db').
38 [Ivan Nestlerode <inestlerode@us.ibm.com>]
40 *) Partial backport from 0.9.9-dev:
42 Introduce bn_mul_mont (dedicated Montgomery multiplication
43 procedure) as a candidate for BIGNUM assembler implementation.
44 While 0.9.9-dev uses assembler for various architectures, only
45 x86_64 is available by default here in the 0.9.8 branch, and
46 32-bit x86 is available through a compile-time setting.
48 To try the 32-bit x86 assembler implementation, use Configure
49 option "enable-montasm" (which exists only for this backport).
51 As "enable-montasm" for 32-bit x86 disclaims code stability
52 anyway, in this constellation we activate additional code
53 backported from 0.9.9-dev for further performance improvements,
54 namely BN_from_montgomery_word. (To enable this otherwise,
55 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
57 [Andy Polyakov (backport partially by Bodo Moeller)]
59 *) Add TLS session ticket callback. This allows an application to set
60 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
61 values. This is useful for key rollover for example where several key
62 sets may exist with different names.
65 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
66 This was broken until now in 0.9.8 releases, such that the only way
67 a registered ENGINE could be used (assuming it initialises
68 successfully on the host) was to explicitly set it as the default
69 for the relevant algorithms. This is in contradiction with 0.9.7
70 behaviour and the documentation. With this fix, when an ENGINE is
71 registered into a given algorithm's table of implementations, the
72 'uptodate' flag is reset so that auto-discovery will be used next
73 time a new context for that algorithm attempts to select an
75 [Ian Lister (tweaked by Geoff Thorpe)]
77 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
78 implemention in the following ways:
80 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
83 Lack of BER streaming support means one pass streaming processing is
84 only supported if data is detached: setting the streaming flag is
85 ignored for embedded content.
87 CMS support is disabled by default and must be explicitly enabled
88 with the enable-cms configuration option.
91 *) Update the GMP engine glue to do direct copies between BIGNUM and
92 mpz_t when openssl and GMP use the same limb size. Otherwise the
93 existing "conversion via a text string export" trick is still used.
94 [Paul Sheer <paulsheer@gmail.com>]
96 *) Zlib compression BIO. This is a filter BIO which compressed and
97 uncompresses any data passed through it.
100 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
101 RFC3394 compatible AES key wrapping.
104 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
105 sets string data without copying. X509_ALGOR_set0() and
106 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
107 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
108 from an X509_ATTRIBUTE structure optionally checking it occurs only
109 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
113 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
114 to get the expected BN_FLG_CONSTTIME behavior.
115 [Bodo Moeller (Google)]
119 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
120 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
121 - added some more tests to do_tests.pl
122 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
123 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
124 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
125 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
126 - various changes to netware.pl to enable gcc-cross builds on Win32
128 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
129 - various changes to fix missing prototype warnings
130 - fixed x86nasm.pl to create correct asm files for NASM COFF output
131 - added AES, WHIRLPOOL and CPUID assembler code to build files
132 - added missing AES assembler make rules to mk1mf.pl
133 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
134 [Guenter Knauf <eflash@gmx.net>]
136 *) Implement certificate status request TLS extension defined in RFC3546.
137 A client can set the appropriate parameters and receive the encoded
138 OCSP response via a callback. A server can query the supplied parameters
139 and set the encoded OCSP response in the callback. Add simplified examples
140 to s_client and s_server.
143 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
146 + Binary incompatibility of ssl_ctx_st structure
147 + DTLS interoperation with non-compliant servers
148 + Don't call get_session_cb() without proposed session
149 + Fix ia64 assembler code
150 [Andy Polyakov, Steve Henson]
152 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
154 *) DTLS Handshake overhaul. There were longstanding issues with
155 OpenSSL DTLS implementation, which were making it impossible for
156 RFC 4347 compliant client to communicate with OpenSSL server.
157 Unfortunately just fixing these incompatibilities would "cut off"
158 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
159 server keeps tolerating non RFC compliant syntax. The opposite is
160 not true, 0.9.8f client can not communicate with earlier server.
161 This update even addresses CVE-2007-4995.
164 *) Changes to avoid need for function casts in OpenSSL: some compilers
165 (gcc 4.2 and later) reject their use.
166 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
169 *) Add RFC4507 support to OpenSSL. This includes the corrections in
170 RFC4507bis. The encrypted ticket format is an encrypted encoded
171 SSL_SESSION structure, that way new session features are automatically
174 If a client application caches session in an SSL_SESSION structure
175 support is transparent because tickets are now stored in the encoded
178 The SSL_CTX structure automatically generates keys for ticket
179 protection in servers so again support should be possible
180 with no application modification.
182 If a client or server wishes to disable RFC4507 support then the option
183 SSL_OP_NO_TICKET can be set.
185 Add a TLS extension debugging callback to allow the contents of any client
186 or server extensions to be examined.
188 This work was sponsored by Google.
191 *) Add initial support for TLS extensions, specifically for the server_name
192 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
193 have new members for a host name. The SSL data structure has an
194 additional member SSL_CTX *initial_ctx so that new sessions can be
195 stored in that context to allow for session resumption, even after the
196 SSL has been switched to a new SSL_CTX in reaction to a client's
197 server_name extension.
199 New functions (subject to change):
202 SSL_get_servername_type()
205 New CTRL codes and macros (subject to change):
207 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
208 - SSL_CTX_set_tlsext_servername_callback()
209 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
210 - SSL_CTX_set_tlsext_servername_arg()
211 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
213 openssl s_client has a new '-servername ...' option.
215 openssl s_server has new options '-servername_host ...', '-cert2 ...',
216 '-key2 ...', '-servername_fatal' (subject to change). This allows
217 testing the HostName extension for a specific single host name ('-cert'
218 and '-key' remain fallbacks for handshakes without HostName
219 negotiation). If the unrecogninzed_name alert has to be sent, this by
220 default is a warning; it becomes fatal with the '-servername_fatal'
223 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
225 *) Add AES and SSE2 assembly language support to VC++ build.
228 *) Mitigate attack on final subtraction in Montgomery reduction.
231 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
232 (which previously caused an internal error).
235 *) Squeeze another 10% out of IGE mode when in != out.
238 *) AES IGE mode speedup.
239 [Dean Gaudet (Google)]
241 *) Add the Korean symmetric 128-bit cipher SEED (see
242 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
243 add SEED ciphersuites from RFC 4162:
245 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
246 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
247 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
248 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
250 To minimize changes between patchlevels in the OpenSSL 0.9.8
251 series, SEED remains excluded from compilation unless OpenSSL
252 is configured with 'enable-seed'.
255 *) Mitigate branch prediction attacks, which can be practical if a
256 single processor is shared, allowing a spy process to extract
257 information. For detailed background information, see
258 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
259 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
260 and Necessary Software Countermeasures"). The core of the change
261 are new versions BN_div_no_branch() and
262 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
263 respectively, which are slower, but avoid the security-relevant
264 conditional branches. These are automatically called by BN_div()
265 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
266 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
267 remove a conditional branch.
269 BN_FLG_CONSTTIME is the new name for the previous
270 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
271 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
272 in the exponent causes BN_mod_exp_mont() to use the alternative
273 implementation in BN_mod_exp_mont_consttime().) The old name
274 remains as a deprecated alias.
276 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
277 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
278 constant-time implementations for more than just exponentiation.
279 Here too the old name is kept as a deprecated alias.
281 BN_BLINDING_new() will now use BN_dup() for the modulus so that
282 the BN_BLINDING structure gets an independent copy of the
283 modulus. This means that the previous "BIGNUM *m" argument to
284 BN_BLINDING_new() and to BN_BLINDING_create_param() now
285 essentially becomes "const BIGNUM *m", although we can't actually
286 change this in the header file before 0.9.9. It allows
287 RSA_setup_blinding() to use BN_with_flags() on the modulus to
288 enable BN_FLG_CONSTTIME.
290 [Matthew D Wood (Intel Corp)]
292 *) In the SSL/TLS server implementation, be strict about session ID
293 context matching (which matters if an application uses a single
294 external cache for different purposes). Previously,
295 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
296 set. This did ensure strict client verification, but meant that,
297 with applications using a single external cache for quite
298 different requirements, clients could circumvent ciphersuite
299 restrictions for a given session ID context by starting a session
300 in a different context.
303 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
304 a ciphersuite string such as "DEFAULT:RSA" cannot enable
305 authentication-only ciphersuites.
308 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
309 not complete and could lead to a possible single byte overflow
310 (CVE-2007-5135) [Ben Laurie]
312 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
314 *) Since AES128 and AES256 (and similarly Camellia128 and
315 Camellia256) share a single mask bit in the logic of
316 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
317 kludge to work properly if AES128 is available and AES256 isn't
318 (or if Camellia128 is available and Camellia256 isn't).
321 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
322 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
323 When a point or a seed is encoded in a BIT STRING, we need to
324 prevent the removal of trailing zero bits to get the proper DER
325 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
326 of a NamedBitList, for which trailing 0 bits need to be removed.)
329 *) Have SSL/TLS server implementation tolerate "mismatched" record
330 protocol version while receiving ClientHello even if the
331 ClientHello is fragmented. (The server can't insist on the
332 particular protocol version it has chosen before the ServerHello
333 message has informed the client about his choice.)
336 *) Add RFC 3779 support.
337 [Rob Austein for ARIN, Ben Laurie]
339 *) Load error codes if they are not already present instead of using a
340 static variable. This allows them to be cleanly unloaded and reloaded.
341 Improve header file function name parsing.
344 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
345 or CAPABILITY handshake as required by RFCs.
348 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
350 *) Introduce limits to prevent malicious keys being able to
351 cause a denial of service. (CVE-2006-2940)
352 [Steve Henson, Bodo Moeller]
354 *) Fix ASN.1 parsing of certain invalid structures that can result
355 in a denial of service. (CVE-2006-2937) [Steve Henson]
357 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
358 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
360 *) Fix SSL client code which could crash if connecting to a
361 malicious SSLv2 server. (CVE-2006-4343)
362 [Tavis Ormandy and Will Drewry, Google Security Team]
364 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
365 match only those. Before that, "AES256-SHA" would be interpreted
366 as a pattern and match "AES128-SHA" too (since AES128-SHA got
367 the same strength classification in 0.9.7h) as we currently only
368 have a single AES bit in the ciphersuite description bitmap.
369 That change, however, also applied to ciphersuite strings such as
370 "RC4-MD5" that intentionally matched multiple ciphersuites --
371 namely, SSL 2.0 ciphersuites in addition to the more common ones
372 from SSL 3.0/TLS 1.0.
374 So we change the selection algorithm again: Naming an explicit
375 ciphersuite selects this one ciphersuite, and any other similar
376 ciphersuite (same bitmap) from *other* protocol versions.
377 Thus, "RC4-MD5" again will properly select both the SSL 2.0
378 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
380 Since SSL 2.0 does not have any ciphersuites for which the
381 128/256 bit distinction would be relevant, this works for now.
382 The proper fix will be to use different bits for AES128 and
383 AES256, which would have avoided the problems from the beginning;
384 however, bits are scarce, so we can only do this in a new release
385 (not just a patchlevel) when we can change the SSL_CIPHER
386 definition to split the single 'unsigned long mask' bitmap into
387 multiple values to extend the available space.
391 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
393 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
394 (CVE-2006-4339) [Ben Laurie and Google Security Team]
396 *) Add AES IGE and biIGE modes.
399 *) Change the Unix randomness entropy gathering to use poll() when
400 possible instead of select(), since the latter has some
401 undesirable limitations.
402 [Darryl Miles via Richard Levitte and Bodo Moeller]
404 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
405 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
406 cannot be implicitly activated as part of, e.g., the "AES" alias.
407 However, please upgrade to OpenSSL 0.9.9[-dev] for
408 non-experimental use of the ECC ciphersuites to get TLS extension
409 support, which is required for curve and point format negotiation
410 to avoid potential handshake problems.
413 *) Disable rogue ciphersuites:
415 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
416 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
417 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
419 The latter two were purportedly from
420 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
423 Also deactivate the remaining ciphersuites from
424 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
425 unofficial, and the ID has long expired.
428 *) Fix RSA blinding Heisenbug (problems sometimes occured on
429 dual-core machines) and other potential thread-safety issues.
432 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
433 versions), which is now available for royalty-free use
434 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
435 Also, add Camellia TLS ciphersuites from RFC 4132.
437 To minimize changes between patchlevels in the OpenSSL 0.9.8
438 series, Camellia remains excluded from compilation unless OpenSSL
439 is configured with 'enable-camellia'.
442 *) Disable the padding bug check when compression is in use. The padding
443 bug check assumes the first packet is of even length, this is not
444 necessarily true if compresssion is enabled and can result in false
445 positives causing handshake failure. The actual bug test is ancient
446 code so it is hoped that implementations will either have fixed it by
447 now or any which still have the bug do not support compression.
450 Changes between 0.9.8a and 0.9.8b [04 May 2006]
452 *) When applying a cipher rule check to see if string match is an explicit
453 cipher suite and only match that one cipher suite if it is.
456 *) Link in manifests for VC++ if needed.
457 [Austin Ziegler <halostatue@gmail.com>]
459 *) Update support for ECC-based TLS ciphersuites according to
460 draft-ietf-tls-ecc-12.txt with proposed changes (but without
461 TLS extensions, which are supported starting with the 0.9.9
462 branch, not in the OpenSSL 0.9.8 branch).
465 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
466 opaque EVP_CIPHER_CTX handling.
469 *) Fixes and enhancements to zlib compression code. We now only use
470 "zlib1.dll" and use the default __cdecl calling convention on Win32
471 to conform with the standards mentioned here:
472 http://www.zlib.net/DLL_FAQ.txt
473 Static zlib linking now works on Windows and the new --with-zlib-include
474 --with-zlib-lib options to Configure can be used to supply the location
475 of the headers and library. Gracefully handle case where zlib library
479 *) Several fixes and enhancements to the OID generation code. The old code
480 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
481 handle numbers larger than ULONG_MAX, truncated printing and had a
482 non standard OBJ_obj2txt() behaviour.
485 *) Add support for building of engines under engine/ as shared libraries
486 under VC++ build system.
489 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
490 Hopefully, we will not see any false combination of paths any more.
493 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
495 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
496 (part of SSL_OP_ALL). This option used to disable the
497 countermeasure against man-in-the-middle protocol-version
498 rollback in the SSL 2.0 server implementation, which is a bad
499 idea. (CVE-2005-2969)
501 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
502 for Information Security, National Institute of Advanced Industrial
503 Science and Technology [AIST], Japan)]
505 *) Add two function to clear and return the verify parameter flags.
508 *) Keep cipherlists sorted in the source instead of sorting them at
509 runtime, thus removing the need for a lock.
512 *) Avoid some small subgroup attacks in Diffie-Hellman.
513 [Nick Mathewson and Ben Laurie]
515 *) Add functions for well-known primes.
518 *) Extended Windows CE support.
519 [Satoshi Nakamura and Andy Polyakov]
521 *) Initialize SSL_METHOD structures at compile time instead of during
522 runtime, thus removing the need for a lock.
525 *) Make PKCS7_decrypt() work even if no certificate is supplied by
526 attempting to decrypt each encrypted key in turn. Add support to
530 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
532 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
535 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
538 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
539 key into the same file any more.
542 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
545 *) Add -utf8 command line and config file option to 'ca'.
546 [Stefan <stf@udoma.org]
548 *) Removed the macro des_crypt(), as it seems to conflict with some
549 libraries. Use DES_crypt().
552 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
553 involves renaming the source and generated shared-libs for
554 both. The engines will accept the corrected or legacy ids
555 ('ncipher' and '4758_cca' respectively) when binding. NB,
556 this only applies when building 'shared'.
557 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
559 *) Add attribute functions to EVP_PKEY structure. Modify
560 PKCS12_create() to recognize a CSP name attribute and
561 use it. Make -CSP option work again in pkcs12 utility.
564 *) Add new functionality to the bn blinding code:
565 - automatic re-creation of the BN_BLINDING parameters after
566 a fixed number of uses (currently 32)
567 - add new function for parameter creation
568 - introduce flags to control the update behaviour of the
569 BN_BLINDING parameters
570 - hide BN_BLINDING structure
571 Add a second BN_BLINDING slot to the RSA structure to improve
572 performance when a single RSA object is shared among several
576 *) Add support for DTLS.
577 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
579 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
580 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
583 *) Remove buggy and incompletet DH cert support from
584 ssl/ssl_rsa.c and ssl/s3_both.c
587 *) Use SHA-1 instead of MD5 as the default digest algorithm for
588 the apps/openssl applications.
591 *) Compile clean with "-Wall -Wmissing-prototypes
592 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
593 DEBUG_SAFESTACK must also be set.
596 *) Change ./Configure so that certain algorithms can be disabled by default.
597 The new counterpiece to "no-xxx" is "enable-xxx".
599 The patented RC5 and MDC2 algorithms will now be disabled unless
600 "enable-rc5" and "enable-mdc2", respectively, are specified.
602 (IDEA remains enabled despite being patented. This is because IDEA
603 is frequently required for interoperability, and there is no license
604 fee for non-commercial use. As before, "no-idea" can be used to
605 avoid this algorithm.)
609 *) Add processing of proxy certificates (see RFC 3820). This work was
610 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
611 EGEE (Enabling Grids for E-science in Europe).
614 *) RC4 performance overhaul on modern architectures/implementations, such
615 as Intel P4, IA-64 and AMD64.
618 *) New utility extract-section.pl. This can be used specify an alternative
619 section number in a pod file instead of having to treat each file as
620 a separate case in Makefile. This can be done by adding two lines to the
623 =for comment openssl_section:XXX
625 The blank line is mandatory.
629 *) New arguments -certform, -keyform and -pass for s_client and s_server
630 to allow alternative format key and certificate files and passphrase
634 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
635 update associated structures and add various utility functions.
637 Add new policy related verify parameters, include policy checking in
638 standard verify code. Enhance 'smime' application with extra parameters
639 to support policy checking and print out.
642 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
643 Nehemiah processors. These extensions support AES encryption in hardware
644 as well as RNG (though RNG support is currently disabled).
645 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
647 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
650 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
651 [Andy Polyakov and a number of other people]
653 *) Improved PowerPC platform support. Most notably BIGNUM assembler
654 implementation contributed by IBM.
655 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
657 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
658 exponent rather than 'unsigned long'. There is a corresponding change to
659 the new 'rsa_keygen' element of the RSA_METHOD structure.
660 [Jelte Jansen, Geoff Thorpe]
662 *) Functionality for creating the initial serial number file is now
663 moved from CA.pl to the 'ca' utility with a new option -create_serial.
665 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
666 number file to 1, which is bound to cause problems. To avoid
667 the problems while respecting compatibility between different 0.9.7
668 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
669 CA.pl for serial number initialization. With the new release 0.9.8,
670 we can fix the problem directly in the 'ca' utility.)
673 *) Reduced header interdepencies by declaring more opaque objects in
674 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
675 give fewer recursive includes, which could break lazy source code - so
676 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
677 developers should define this symbol when building and using openssl to
678 ensure they track the recommended behaviour, interfaces, [etc], but
679 backwards-compatible behaviour prevails when this isn't defined.
682 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
685 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
686 This will generate a random key of the appropriate length based on the
687 cipher context. The EVP_CIPHER can provide its own random key generation
688 routine to support keys of a specific form. This is used in the des and
689 3des routines to generate a key of the correct parity. Update S/MIME
690 code to use new functions and hence generate correct parity DES keys.
691 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
692 valid (weak or incorrect parity).
695 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
696 as looking them up. This is useful when the verified structure may contain
697 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
698 present unless the new PKCS7_NO_CRL flag is asserted.
701 *) Extend ASN1 oid configuration module. It now additionally accepts the
704 shortName = some long name, 1.2.3.4
707 *) Reimplemented the BN_CTX implementation. There is now no more static
708 limitation on the number of variables it can handle nor the depth of the
709 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
710 information can now expand as required, and rather than having a single
711 static array of bignums, BN_CTX now uses a linked-list of such arrays
712 allowing it to expand on demand whilst maintaining the usefulness of
716 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
717 to allow all RSA operations to function using a single BN_CTX.
720 *) Preliminary support for certificate policy evaluation and checking. This
721 is initially intended to pass the tests outlined in "Conformance Testing
722 of Relying Party Client Certificate Path Processing Logic" v1.07.
725 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
726 remained unused and not that useful. A variety of other little bignum
727 tweaks and fixes have also been made continuing on from the audit (see
731 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
732 associated ASN1, EVP and SSL functions and old ASN1 macros.
735 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
736 and this should never fail. So the return value from the use of
737 BN_set_word() (which can fail due to needless expansion) is now deprecated;
738 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
741 *) BN_CTX_get() should return zero-valued bignums, providing the same
742 initialised value as BN_new().
743 [Geoff Thorpe, suggested by Ulf Möller]
745 *) Support for inhibitAnyPolicy certificate extension.
748 *) An audit of the BIGNUM code is underway, for which debugging code is
749 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
750 is considered valid when processing BIGNUMs, and causes execution to
751 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
752 further steps are taken to deliberately pollute unused data in BIGNUM
753 structures to try and expose faulty code further on. For now, openssl will
754 (in its default mode of operation) continue to tolerate the inconsistent
755 forms that it has tolerated in the past, but authors and packagers should
756 consider trying openssl and their own applications when compiled with
757 these debugging symbols defined. It will help highlight potential bugs in
758 their own code, and will improve the test coverage for OpenSSL itself. At
759 some point, these tighter rules will become openssl's default to improve
760 maintainability, though the assert()s and other overheads will remain only
761 in debugging configurations. See bn.h for more details.
762 [Geoff Thorpe, Nils Larsch, Ulf Möller]
764 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
765 that can only be obtained through BN_CTX_new() (which implicitly
766 initialises it). The presence of this function only made it possible
767 to overwrite an existing structure (and cause memory leaks).
770 *) Because of the callback-based approach for implementing LHASH as a
771 template type, lh_insert() adds opaque objects to hash-tables and
772 lh_doall() or lh_doall_arg() are typically used with a destructor callback
773 to clean up those corresponding objects before destroying the hash table
774 (and losing the object pointers). So some over-zealous constifications in
775 LHASH have been relaxed so that lh_insert() does not take (nor store) the
776 objects as "const" and the lh_doall[_arg] callback wrappers are not
777 prototyped to have "const" restrictions on the object pointers they are
778 given (and so aren't required to cast them away any more).
781 *) The tmdiff.h API was so ugly and minimal that our own timing utility
782 (speed) prefers to use its own implementation. The two implementations
783 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
784 its object type properly exposed (MS_TM) instead of casting to/from "char
785 *". This may still change yet if someone realises MS_TM and "ms_time_***"
786 aren't necessarily the greatest nomenclatures - but this is what was used
787 internally to the implementation so I've used that for now.
790 *) Ensure that deprecated functions do not get compiled when
791 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
792 the self-tests were still using deprecated key-generation functions so
793 these have been updated also.
796 *) Reorganise PKCS#7 code to separate the digest location functionality
797 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
798 New function PKCS7_set_digest() to set the digest type for PKCS#7
799 digestedData type. Add additional code to correctly generate the
800 digestedData type and add support for this type in PKCS7 initialization
804 *) New function PKCS7_set0_type_other() this initializes a PKCS7
805 structure of type "other".
808 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
809 sure the loop does correctly stop and breaking ("division by zero")
810 modulus operations are not performed. The (pre-generated) prime
811 table crypto/bn/bn_prime.h was already correct, but it could not be
812 re-generated on some platforms because of the "division by zero"
813 situation in the script.
814 [Ralf S. Engelschall]
816 *) Update support for ECC-based TLS ciphersuites according to
817 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
818 SHA-1 now is only used for "small" curves (where the
819 representation of a field element takes up to 24 bytes); for
820 larger curves, the field element resulting from ECDH is directly
821 used as premaster secret.
822 [Douglas Stebila (Sun Microsystems Laboratories)]
824 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
825 curve secp160r1 to the tests.
826 [Douglas Stebila (Sun Microsystems Laboratories)]
828 *) Add the possibility to load symbols globally with DSO.
829 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
831 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
832 control of the error stack.
835 *) Add support for STORE in ENGINE.
838 *) Add the STORE type. The intention is to provide a common interface
839 to certificate and key stores, be they simple file-based stores, or
840 HSM-type store, or LDAP stores, or...
841 NOTE: The code is currently UNTESTED and isn't really used anywhere.
844 *) Add a generic structure called OPENSSL_ITEM. This can be used to
845 pass a list of arguments to any function as well as provide a way
846 for a function to pass data back to the caller.
849 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
850 works like BUF_strdup() but can be used to duplicate a portion of
851 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
855 *) Add the function sk_find_ex() which works like sk_find(), but will
856 return an index to an element even if an exact match couldn't be
857 found. The index is guaranteed to point at the element where the
858 searched-for key would be inserted to preserve sorting order.
861 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
862 takes an extra flags argument for optional functionality. Currently,
863 the following flags are defined:
865 OBJ_BSEARCH_VALUE_ON_NOMATCH
866 This one gets OBJ_bsearch_ex() to return a pointer to the first
867 element where the comparing function returns a negative or zero
870 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
871 This one gets OBJ_bsearch_ex() to return a pointer to the first
872 element where the comparing function returns zero. This is useful
873 if there are more than one element where the comparing function
877 *) Make it possible to create self-signed certificates with 'openssl ca'
878 in such a way that the self-signed certificate becomes part of the
879 CA database and uses the same mechanisms for serial number generation
880 as all other certificate signing. The new flag '-selfsign' enables
881 this functionality. Adapt CA.sh and CA.pl.in.
884 *) Add functionality to check the public key of a certificate request
885 against a given private. This is useful to check that a certificate
886 request can be signed by that key (self-signing).
889 *) Make it possible to have multiple active certificates with the same
890 subject in the CA index file. This is done only if the keyword
891 'unique_subject' is set to 'no' in the main CA section (default
892 if 'CA_default') of the configuration file. The value is saved
893 with the database itself in a separate index attribute file,
894 named like the index file with '.attr' appended to the name.
897 *) Generate muti valued AVAs using '+' notation in config files for
901 *) Support for nameConstraints certificate extension.
904 *) Support for policyConstraints certificate extension.
907 *) Support for policyMappings certificate extension.
910 *) Make sure the default DSA_METHOD implementation only uses its
911 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
912 and change its own handlers to be NULL so as to remove unnecessary
913 indirection. This lets alternative implementations fallback to the
914 default implementation more easily.
917 *) Support for directoryName in GeneralName related extensions
921 *) Make it possible to link applications using Makefile.shared.
922 Make that possible even when linking against static libraries!
925 *) Support for single pass processing for S/MIME signing. This now
926 means that S/MIME signing can be done from a pipe, in addition
927 cleartext signing (multipart/signed type) is effectively streaming
928 and the signed data does not need to be all held in memory.
930 This is done with a new flag PKCS7_STREAM. When this flag is set
931 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
932 is done after the data is output (and digests calculated) in
936 *) Add full support for -rpath/-R, both in shared libraries and
937 applications, at least on the platforms where it's known how
941 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
942 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
943 will now compute a table of multiples of the generator that
944 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
945 faster (notably in the case of a single point multiplication,
947 [Nils Larsch, Bodo Moeller]
949 *) IPv6 support for certificate extensions. The various extensions
950 which use the IP:a.b.c.d can now take IPv6 addresses using the
951 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
955 *) Added an ENGINE that implements RSA by performing private key
956 exponentiations with the GMP library. The conversions to and from
957 GMP's mpz_t format aren't optimised nor are any montgomery forms
958 cached, and on x86 it appears OpenSSL's own performance has caught up.
959 However there are likely to be other architectures where GMP could
960 provide a boost. This ENGINE is not built in by default, but it can be
961 specified at Configure time and should be accompanied by the necessary
962 linker additions, eg;
963 ./config -DOPENSSL_USE_GMP -lgmp
966 *) "openssl engine" will not display ENGINE/DSO load failure errors when
967 testing availability of engines with "-t" - the old behaviour is
968 produced by increasing the feature's verbosity with "-tt".
971 *) ECDSA routines: under certain error conditions uninitialized BN objects
972 could be freed. Solution: make sure initialization is performed early
973 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
977 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
978 and DH_METHOD (eg. by ENGINE implementations) to override the normal
979 software implementations. For DSA and DH, parameter generation can
980 also be overriden by providing the appropriate method callbacks.
983 *) Change the "progress" mechanism used in key-generation and
984 primality testing to functions that take a new BN_GENCB pointer in
985 place of callback/argument pairs. The new API functions have "_ex"
986 postfixes and the older functions are reimplemented as wrappers for
987 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
988 declarations of the old functions to help (graceful) attempts to
989 migrate to the new functions. Also, the new key-generation API
990 functions operate on a caller-supplied key-structure and return
991 success/failure rather than returning a key or NULL - this is to
992 help make "keygen" another member function of RSA_METHOD etc.
994 Example for using the new callback interface:
996 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1000 BN_GENCB_set(&my_cb, my_callback, my_arg);
1002 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1003 /* For the meaning of a, b in calls to my_callback(), see the
1004 * documentation of the function that calls the callback.
1005 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1006 * my_callback should return 1 if it wants BN_is_prime_ex()
1007 * to continue, or 0 to stop.
1012 *) Change the ZLIB compression method to be stateful, and make it
1013 available to TLS with the number defined in
1014 draft-ietf-tls-compression-04.txt.
1017 *) Add the ASN.1 structures and functions for CertificatePair, which
1018 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1020 CertificatePair ::= SEQUENCE {
1021 forward [0] Certificate OPTIONAL,
1022 reverse [1] Certificate OPTIONAL,
1023 -- at least one of the pair shall be present -- }
1025 Also implement the PEM functions to read and write certificate
1026 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1028 This needed to be defined, mostly for the sake of the LDAP
1029 attribute crossCertificatePair, but may prove useful elsewhere as
1033 *) Make it possible to inhibit symlinking of shared libraries in
1034 Makefile.shared, for Cygwin's sake.
1037 *) Extend the BIGNUM API by creating a function
1038 void BN_set_negative(BIGNUM *a, int neg);
1039 and a macro that behave like
1040 int BN_is_negative(const BIGNUM *a);
1042 to avoid the need to access 'a->neg' directly in applications.
1045 *) Implement fast modular reduction for pseudo-Mersenne primes
1046 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1047 EC_GROUP_new_curve_GFp() will now automatically use this
1049 [Nils Larsch <nla@trustcenter.de>]
1051 *) Add new lock type (CRYPTO_LOCK_BN).
1054 *) Change the ENGINE framework to automatically load engines
1055 dynamically from specific directories unless they could be
1056 found to already be built in or loaded. Move all the
1057 current engines except for the cryptodev one to a new
1059 The engines in engines/ are built as shared libraries if
1060 the "shared" options was given to ./Configure or ./config.
1061 Otherwise, they are inserted in libcrypto.a.
1062 /usr/local/ssl/engines is the default directory for dynamic
1063 engines, but that can be overriden at configure time through
1064 the usual use of --prefix and/or --openssldir, and at run
1065 time with the environment variable OPENSSL_ENGINES.
1066 [Geoff Thorpe and Richard Levitte]
1068 *) Add Makefile.shared, a helper makefile to build shared
1069 libraries. Addapt Makefile.org.
1072 *) Add version info to Win32 DLLs.
1073 [Peter 'Luna' Runestig" <peter@runestig.com>]
1075 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1076 can be added using this API to created arbitrary PKCS#12
1077 files while avoiding the low level API.
1079 New options to PKCS12_create(), key or cert can be NULL and
1080 will then be omitted from the output file. The encryption
1081 algorithm NIDs can be set to -1 for no encryption, the mac
1082 iteration count can be set to 0 to omit the mac.
1084 Enhance pkcs12 utility by making the -nokeys and -nocerts
1085 options work when creating a PKCS#12 file. New option -nomac
1086 to omit the mac, NONE can be set for an encryption algorithm.
1087 New code is modified to use the enhanced PKCS12_create()
1088 instead of the low level API.
1091 *) Extend ASN1 encoder to support indefinite length constructed
1092 encoding. This can output sequences tags and octet strings in
1093 this form. Modify pk7_asn1.c to support indefinite length
1094 encoding. This is experimental and needs additional code to
1095 be useful, such as an ASN1 bio and some enhanced streaming
1098 Extend template encode functionality so that tagging is passed
1099 down to the template encoder.
1102 *) Let 'openssl req' fail if an argument to '-newkey' is not
1103 recognized instead of using RSA as a default.
1106 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1107 As these are not official, they are not included in "ALL";
1108 the "ECCdraft" ciphersuite group alias can be used to select them.
1109 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1111 *) Add ECDH engine support.
1112 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1114 *) Add ECDH in new directory crypto/ecdh/.
1115 [Douglas Stebila (Sun Microsystems Laboratories)]
1117 *) Let BN_rand_range() abort with an error after 100 iterations
1118 without success (which indicates a broken PRNG).
1121 *) Change BN_mod_sqrt() so that it verifies that the input value
1122 is really the square of the return value. (Previously,
1123 BN_mod_sqrt would show GIGO behaviour.)
1126 *) Add named elliptic curves over binary fields from X9.62, SECG,
1127 and WAP/WTLS; add OIDs that were still missing.
1129 [Sheueling Chang Shantz and Douglas Stebila
1130 (Sun Microsystems Laboratories)]
1132 *) Extend the EC library for elliptic curves over binary fields
1133 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1136 EC_GF2m_simple_method
1140 EC_GROUP_new_curve_GF2m
1141 EC_GROUP_set_curve_GF2m
1142 EC_GROUP_get_curve_GF2m
1143 EC_POINT_set_affine_coordinates_GF2m
1144 EC_POINT_get_affine_coordinates_GF2m
1145 EC_POINT_set_compressed_coordinates_GF2m
1147 Point compression for binary fields is disabled by default for
1148 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1151 As binary polynomials are represented as BIGNUMs, various members
1152 of the EC_GROUP and EC_POINT data structures can be shared
1153 between the implementations for prime fields and binary fields;
1154 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1155 are essentially identical to their ..._GFp counterparts.
1156 (For simplicity, the '..._GFp' prefix has been dropped from
1157 various internal method names.)
1159 An internal 'field_div' method (similar to 'field_mul' and
1160 'field_sqr') has been added; this is used only for binary fields.
1162 [Sheueling Chang Shantz and Douglas Stebila
1163 (Sun Microsystems Laboratories)]
1165 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1166 through methods ('mul', 'precompute_mult').
1168 The generic implementations (now internally called 'ec_wNAF_mul'
1169 and 'ec_wNAF_precomputed_mult') remain the default if these
1170 methods are undefined.
1172 [Sheueling Chang Shantz and Douglas Stebila
1173 (Sun Microsystems Laboratories)]
1175 *) New function EC_GROUP_get_degree, which is defined through
1176 EC_METHOD. For curves over prime fields, this returns the bit
1177 length of the modulus.
1179 [Sheueling Chang Shantz and Douglas Stebila
1180 (Sun Microsystems Laboratories)]
1182 *) New functions EC_GROUP_dup, EC_POINT_dup.
1183 (These simply call ..._new and ..._copy).
1185 [Sheueling Chang Shantz and Douglas Stebila
1186 (Sun Microsystems Laboratories)]
1188 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1189 Polynomials are represented as BIGNUMs (where the sign bit is not
1190 used) in the following functions [macros]:
1193 BN_GF2m_sub [= BN_GF2m_add]
1194 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1195 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1196 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1198 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1199 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1200 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1201 BN_GF2m_cmp [= BN_ucmp]
1203 (Note that only the 'mod' functions are actually for fields GF(2^m).
1204 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1206 For some functions, an the irreducible polynomial defining a
1207 field can be given as an 'unsigned int[]' with strictly
1208 decreasing elements giving the indices of those bits that are set;
1209 i.e., p[] represents the polynomial
1210 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1212 p[0] > p[1] > ... > p[k] = 0.
1213 This applies to the following functions:
1218 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1219 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1221 BN_GF2m_mod_sqrt_arr
1222 BN_GF2m_mod_solve_quad_arr
1226 Conversion can be performed by the following functions:
1231 bntest.c has additional tests for binary polynomial arithmetic.
1233 Two implementations for BN_GF2m_mod_div() are available.
1234 The default algorithm simply uses BN_GF2m_mod_inv() and
1235 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1236 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1237 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1239 [Sheueling Chang Shantz and Douglas Stebila
1240 (Sun Microsystems Laboratories)]
1242 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1243 functionality is disabled at compile-time.
1244 [Douglas Stebila <douglas.stebila@sun.com>]
1246 *) Change default behaviour of 'openssl asn1parse' so that more
1247 information is visible when viewing, e.g., a certificate:
1249 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1250 mode the content of non-printable OCTET STRINGs is output in a
1251 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1252 avoid the appearance of a printable string.
1253 [Nils Larsch <nla@trustcenter.de>]
1255 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1257 EC_GROUP_set_asn1_flag()
1258 EC_GROUP_get_asn1_flag()
1259 EC_GROUP_set_point_conversion_form()
1260 EC_GROUP_get_point_conversion_form()
1261 These control ASN1 encoding details:
1262 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1263 has been set to OPENSSL_EC_NAMED_CURVE.
1264 - Points are encoded in uncompressed form by default; options for
1265 asn1_for are as for point2oct, namely
1266 POINT_CONVERSION_COMPRESSED
1267 POINT_CONVERSION_UNCOMPRESSED
1268 POINT_CONVERSION_HYBRID
1270 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1273 EC_GROUP_get0_seed()
1274 EC_GROUP_get_seed_len()
1275 This is used only for ASN1 purposes (so far).
1276 [Nils Larsch <nla@trustcenter.de>]
1278 *) Add 'field_type' member to EC_METHOD, which holds the NID
1279 of the appropriate field type OID. The new function
1280 EC_METHOD_get_field_type() returns this value.
1281 [Nils Larsch <nla@trustcenter.de>]
1286 EC_POINT_point2hex()
1287 EC_POINT_hex2point()
1288 providing useful interfaces to EC_POINT_point2oct() and
1289 EC_POINT_oct2point().
1290 [Nils Larsch <nla@trustcenter.de>]
1292 *) Change internals of the EC library so that the functions
1293 EC_GROUP_set_generator()
1294 EC_GROUP_get_generator()
1295 EC_GROUP_get_order()
1296 EC_GROUP_get_cofactor()
1297 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1298 to methods, which would lead to unnecessary code duplication when
1299 adding different types of curves.
1300 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1302 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1303 arithmetic, and such that modified wNAFs are generated
1304 (which avoid length expansion in many cases).
1307 *) Add a function EC_GROUP_check_discriminant() (defined via
1308 EC_METHOD) that verifies that the curve discriminant is non-zero.
1310 Add a function EC_GROUP_check() that makes some sanity tests
1311 on a EC_GROUP, its generator and order. This includes
1312 EC_GROUP_check_discriminant().
1313 [Nils Larsch <nla@trustcenter.de>]
1315 *) Add ECDSA in new directory crypto/ecdsa/.
1317 Add applications 'openssl ecparam' and 'openssl ecdsa'
1318 (these are based on 'openssl dsaparam' and 'openssl dsa').
1320 ECDSA support is also included in various other files across the
1321 library. Most notably,
1322 - 'openssl req' now has a '-newkey ecdsa:file' option;
1323 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1324 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1325 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1326 them suitable for ECDSA where domain parameters must be
1327 extracted before the specific public key;
1328 - ECDSA engine support has been added.
1329 [Nils Larsch <nla@trustcenter.de>]
1331 *) Include some named elliptic curves, and add OIDs from X9.62,
1332 SECG, and WAP/WTLS. Each curve can be obtained from the new
1334 EC_GROUP_new_by_curve_name(),
1335 and the list of available named curves can be obtained with
1336 EC_get_builtin_curves().
1337 Also add a 'curve_name' member to EC_GROUP objects, which can be
1339 EC_GROUP_set_curve_name()
1340 EC_GROUP_get_curve_name()
1341 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1343 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1344 was actually never needed) and in BN_mul(). The removal in BN_mul()
1345 required a small change in bn_mul_part_recursive() and the addition
1346 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1347 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1348 bn_sub_words() and bn_add_words() except they take arrays with
1352 Changes between 0.9.7m and 0.9.7n [xx XXX xxxx]
1354 *) In the SSL/TLS server implementation, be strict about session ID
1355 context matching (which matters if an application uses a single
1356 external cache for different purposes). Previously,
1357 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1358 set. This did ensure strict client verification, but meant that,
1359 with applications using a single external cache for quite
1360 different requirements, clients could circumvent ciphersuite
1361 restrictions for a given session ID context by starting a session
1362 in a different context.
1365 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1367 *) Cleanse PEM buffers before freeing them since they may contain
1369 [Benjamin Bennett <ben@psc.edu>]
1371 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1372 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1373 authentication-only ciphersuites.
1376 *) Since AES128 and AES256 share a single mask bit in the logic of
1377 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1378 kludge to work properly if AES128 is available and AES256 isn't.
1381 *) Expand security boundary to match 1.1.1 module.
1384 *) Remove redundant features: hash file source, editing of test vectors
1385 modify fipsld to use external fips_premain.c signature.
1388 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1389 run algorithm test programs.
1392 *) Make algorithm test programs more tolerant of whitespace.
1395 *) Have SSL/TLS server implementation tolerate "mismatched" record
1396 protocol version while receiving ClientHello even if the
1397 ClientHello is fragmented. (The server can't insist on the
1398 particular protocol version it has chosen before the ServerHello
1399 message has informed the client about his choice.)
1402 *) Load error codes if they are not already present instead of using a
1403 static variable. This allows them to be cleanly unloaded and reloaded.
1406 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1408 *) Introduce limits to prevent malicious keys being able to
1409 cause a denial of service. (CVE-2006-2940)
1410 [Steve Henson, Bodo Moeller]
1412 *) Fix ASN.1 parsing of certain invalid structures that can result
1413 in a denial of service. (CVE-2006-2937) [Steve Henson]
1415 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1416 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1418 *) Fix SSL client code which could crash if connecting to a
1419 malicious SSLv2 server. (CVE-2006-4343)
1420 [Tavis Ormandy and Will Drewry, Google Security Team]
1422 *) Change ciphersuite string processing so that an explicit
1423 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1424 will no longer include "AES128-SHA"), and any other similar
1425 ciphersuite (same bitmap) from *other* protocol versions (so that
1426 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1427 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1428 changes from 0.9.8b and 0.9.8d.
1431 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1433 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1434 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1436 *) Change the Unix randomness entropy gathering to use poll() when
1437 possible instead of select(), since the latter has some
1438 undesirable limitations.
1439 [Darryl Miles via Richard Levitte and Bodo Moeller]
1441 *) Disable rogue ciphersuites:
1443 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1444 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1445 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1447 The latter two were purportedly from
1448 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1451 Also deactive the remaining ciphersuites from
1452 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1453 unofficial, and the ID has long expired.
1456 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1457 dual-core machines) and other potential thread-safety issues.
1460 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1462 *) Adapt fipsld and the build system to link against the validated FIPS
1463 module in FIPS mode.
1466 *) Fixes for VC++ 2005 build under Windows.
1469 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1470 from a Windows bash shell such as MSYS. It is autodetected from the
1471 "config" script when run from a VC++ environment. Modify standard VC++
1472 build to use fipscanister.o from the GNU make build.
1475 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1477 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1478 The value now differs depending on if you build for FIPS or not.
1479 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1480 safely run with a non-FIPSed libcrypto, as it may crash because of
1481 the difference induced by this change.
1484 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1486 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1487 (part of SSL_OP_ALL). This option used to disable the
1488 countermeasure against man-in-the-middle protocol-version
1489 rollback in the SSL 2.0 server implementation, which is a bad
1490 idea. (CVE-2005-2969)
1492 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1493 for Information Security, National Institute of Advanced Industrial
1494 Science and Technology [AIST], Japan)]
1496 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1497 mainly for FIPS compliance and not fully integrated at this stage.
1500 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1501 the exponentiation using a fixed-length exponent. (Otherwise,
1502 the information leaked through timing could expose the secret key
1503 after many signatures; cf. Bleichenbacher's attack on DSA with
1507 *) Make a new fixed-window mod_exp implementation the default for
1508 RSA, DSA, and DH private-key operations so that the sequence of
1509 squares and multiplies and the memory access pattern are
1510 independent of the particular secret key. This will mitigate
1511 cache-timing and potential related attacks.
1513 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1514 and this is automatically used by BN_mod_exp_mont() if the new flag
1515 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1516 will use this BN flag for private exponents unless the flag
1517 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1518 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1520 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1522 *) Change the client implementation for SSLv23_method() and
1523 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1524 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1525 (Previously, the SSL 2.0 backwards compatible Client Hello
1526 message format would be used even with SSL_OP_NO_SSLv2.)
1529 *) Add support for smime-type MIME parameter in S/MIME messages which some
1533 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1534 a threadsafe manner. Modify rsa code to use new function and add calls
1535 to dsa and dh code (which had race conditions before).
1538 *) Include the fixed error library code in the C error file definitions
1539 instead of fixing them up at runtime. This keeps the error code
1540 structures constant.
1543 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1545 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1548 *) Fixes for newer kerberos headers. NB: the casts are needed because
1549 the 'length' field is signed on one version and unsigned on another
1550 with no (?) obvious way to tell the difference, without these VC++
1551 complains. Also the "definition" of FAR (blank) is no longer included
1552 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1553 some needed definitions.
1556 *) Undo Cygwin change.
1559 *) Added support for proxy certificates according to RFC 3820.
1560 Because they may be a security thread to unaware applications,
1561 they must be explicitely allowed in run-time. See
1562 docs/HOWTO/proxy_certificates.txt for further information.
1565 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1567 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1568 server and client random values. Previously
1569 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1570 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1572 This change has negligible security impact because:
1574 1. Server and client random values still have 24 bytes of pseudo random
1577 2. Server and client random values are sent in the clear in the initial
1580 3. The master secret is derived using the premaster secret (48 bytes in
1581 size for static RSA ciphersuites) as well as client server and random
1584 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1587 [Stephen Henson, reported by UK NISCC]
1589 *) Use Windows randomness collection on Cygwin.
1592 *) Fix hang in EGD/PRNGD query when communication socket is closed
1593 prematurely by EGD/PRNGD.
1594 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1596 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1599 *) Back-port of selected performance improvements from development
1600 branch, as well as improved support for PowerPC platforms.
1603 *) Add lots of checks for memory allocation failure, error codes to indicate
1604 failure and freeing up memory if a failure occurs.
1605 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1607 *) Add new -passin argument to dgst.
1610 *) Perform some character comparisons of different types in X509_NAME_cmp:
1611 this is needed for some certificates that reencode DNs into UTF8Strings
1612 (in violation of RFC3280) and can't or wont issue name rollover
1616 *) Make an explicit check during certificate validation to see that
1617 the CA setting in each certificate on the chain is correct. As a
1618 side effect always do the following basic checks on extensions,
1619 not just when there's an associated purpose to the check:
1621 - if there is an unhandled critical extension (unless the user
1622 has chosen to ignore this fault)
1623 - if the path length has been exceeded (if one is set at all)
1624 - that certain extensions fit the associated purpose (if one has
1628 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1630 *) Avoid a race condition when CRLs are checked in a multi threaded
1631 environment. This would happen due to the reordering of the revoked
1632 entries during signature checking and serial number lookup. Now the
1633 encoding is cached and the serial number sort performed under a lock.
1634 Add new STACK function sk_is_sorted().
1637 *) Add Delta CRL to the extension code.
1640 *) Various fixes to s3_pkt.c so alerts are sent properly.
1641 [David Holmes <d.holmes@f5.com>]
1643 *) Reduce the chances of duplicate issuer name and serial numbers (in
1644 violation of RFC3280) using the OpenSSL certificate creation utilities.
1645 This is done by creating a random 64 bit value for the initial serial
1646 number when a serial number file is created or when a self signed
1647 certificate is created using 'openssl req -x509'. The initial serial
1648 number file is created using 'openssl x509 -next_serial' in CA.pl
1649 rather than being initialized to 1.
1652 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1654 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1655 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1656 [Joe Orton, Steve Henson]
1658 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1660 [Joe Orton, Steve Henson]
1662 *) Make it possible to have multiple active certificates with the same
1663 subject in the CA index file. This is done only if the keyword
1664 'unique_subject' is set to 'no' in the main CA section (default
1665 if 'CA_default') of the configuration file. The value is saved
1666 with the database itself in a separate index attribute file,
1667 named like the index file with '.attr' appended to the name.
1670 *) X509 verify fixes. Disable broken certificate workarounds when
1671 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1672 keyUsage extension present. Don't accept CRLs with unhandled critical
1673 extensions: since verify currently doesn't process CRL extensions this
1674 rejects a CRL with *any* critical extensions. Add new verify error codes
1678 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1679 A clarification of RFC2560 will require the use of OCTET STRINGs and
1680 some implementations cannot handle the current raw format. Since OpenSSL
1681 copies and compares OCSP nonces as opaque blobs without any attempt at
1682 parsing them this should not create any compatibility issues.
1685 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1686 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1687 this HMAC (and other) operations are several times slower than OpenSSL
1691 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1692 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1694 *) Use the correct content when signing type "other".
1697 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1699 *) Fix various bugs revealed by running the NISCC test suite:
1701 Stop out of bounds reads in the ASN1 code when presented with
1702 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1704 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1706 If verify callback ignores invalid public key errors don't try to check
1707 certificate signature with the NULL public key.
1711 *) New -ignore_err option in ocsp application to stop the server
1712 exiting on the first error in a request.
1715 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1716 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1720 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1721 extra data after the compression methods not only for TLS 1.0
1722 but also for SSL 3.0 (as required by the specification).
1723 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1725 *) Change X509_certificate_type() to mark the key as exported/exportable
1726 when it's 512 *bits* long, not 512 bytes.
1729 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1730 blocks during encryption.
1733 *) Various fixes to base64 BIO and non blocking I/O. On write
1734 flushes were not handled properly if the BIO retried. On read
1735 data was not being buffered properly and had various logic bugs.
1736 This also affects blocking I/O when the data being decoded is a
1740 *) Various S/MIME bugfixes and compatibility changes:
1741 output correct application/pkcs7 MIME type if
1742 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1743 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1744 of files as .eml work). Correctly handle very long lines in MIME
1748 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1750 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1751 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1752 a protocol version number mismatch like a decryption error
1753 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1756 *) Turn on RSA blinding by default in the default implementation
1757 to avoid a timing attack. Applications that don't want it can call
1758 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1759 They would be ill-advised to do so in most cases.
1760 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1762 *) Change RSA blinding code so that it works when the PRNG is not
1763 seeded (in this case, the secret RSA exponent is abused as
1764 an unpredictable seed -- if it is not unpredictable, there
1765 is no point in blinding anyway). Make RSA blinding thread-safe
1766 by remembering the creator's thread ID in rsa->blinding and
1767 having all other threads use local one-time blinding factors
1768 (this requires more computation than sharing rsa->blinding, but
1769 avoids excessive locking; and if an RSA object is not shared
1770 between threads, blinding will still be very fast).
1773 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1774 ENGINE as defaults for all supported algorithms irrespective of
1775 the 'flags' parameter. 'flags' is now honoured, so applications
1776 should make sure they are passing it correctly.
1779 *) Target "mingw" now allows native Windows code to be generated in
1780 the Cygwin environment as well as with the MinGW compiler.
1783 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1785 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1786 via timing by performing a MAC computation even if incorrrect
1787 block cipher padding has been found. This is a countermeasure
1788 against active attacks where the attacker has to distinguish
1789 between bad padding and a MAC verification error. (CVE-2003-0078)
1791 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1792 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1793 Martin Vuagnoux (EPFL, Ilion)]
1795 *) Make the no-err option work as intended. The intention with no-err
1796 is not to have the whole error stack handling routines removed from
1797 libcrypto, it's only intended to remove all the function name and
1798 reason texts, thereby removing some of the footprint that may not
1799 be interesting if those errors aren't displayed anyway.
1801 NOTE: it's still possible for any application or module to have it's
1802 own set of error texts inserted. The routines are there, just not
1803 used by default when no-err is given.
1806 *) Add support for FreeBSD on IA64.
1807 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1809 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1810 Kerberos function mit_des_cbc_cksum(). Before this change,
1811 the value returned by DES_cbc_cksum() was like the one from
1812 mit_des_cbc_cksum(), except the bytes were swapped.
1813 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1815 *) Allow an application to disable the automatic SSL chain building.
1816 Before this a rather primitive chain build was always performed in
1817 ssl3_output_cert_chain(): an application had no way to send the
1818 correct chain if the automatic operation produced an incorrect result.
1820 Now the chain builder is disabled if either:
1822 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1824 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1826 The reasoning behind this is that an application would not want the
1827 auto chain building to take place if extra chain certificates are
1828 present and it might also want a means of sending no additional
1829 certificates (for example the chain has two certificates and the
1833 *) Add the possibility to build without the ENGINE framework.
1834 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1836 *) Under Win32 gmtime() can return NULL: check return value in
1837 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1840 *) DSA routines: under certain error conditions uninitialized BN objects
1841 could be freed. Solution: make sure initialization is performed early
1842 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1843 Nils Larsch <nla@trustcenter.de> via PR#459)
1846 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1847 checked on reconnect on the client side, therefore session resumption
1848 could still fail with a "ssl session id is different" error. This
1849 behaviour is masked when SSL_OP_ALL is used due to
1850 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1851 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1852 followup to PR #377.
1855 *) IA-32 assembler support enhancements: unified ELF targets, support
1856 for SCO/Caldera platforms, fix for Cygwin shared build.
1859 *) Add support for FreeBSD on sparc64. As a consequence, support for
1860 FreeBSD on non-x86 processors is separate from x86 processors on
1861 the config script, much like the NetBSD support.
1862 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1864 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1866 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1869 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1870 code (06) was taken as the first octet of the session ID and the last
1871 octet was ignored consequently. As a result SSLv2 client side session
1872 caching could not have worked due to the session ID mismatch between
1874 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1878 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1879 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1883 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1884 seems that in spite of existing for more than a year, many application
1885 author have done nothing to provide the necessary callbacks, which
1886 means that this particular engine will not work properly anywhere.
1887 This is a very unfortunate situation which forces us, in the name
1888 of usability, to give the hw_ncipher.c a static lock, which is part
1890 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1891 appear in 0.9.8 or later. We EXPECT application authors to have
1892 dealt properly with this when 0.9.8 is released (unless we actually
1893 make such changes in the libcrypto locking code that changes will
1894 have to be made anyway).
1897 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1898 octets have been read, EOF or an error occurs. Without this change
1899 some truncated ASN1 structures will not produce an error.
1902 *) Disable Heimdal support, since it hasn't been fully implemented.
1903 Still give the possibility to force the use of Heimdal, but with
1904 warnings and a request that patches get sent to openssl-dev.
1907 *) Add the VC-CE target, introduce the WINCE sysname, and add
1908 INSTALL.WCE and appropriate conditionals to make it build.
1909 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1911 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1912 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1913 edit numbers of the version.
1914 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1916 *) Introduce safe string copy and catenation functions
1917 (BUF_strlcpy() and BUF_strlcat()).
1918 [Ben Laurie (CHATS) and Richard Levitte]
1920 *) Avoid using fixed-size buffers for one-line DNs.
1921 [Ben Laurie (CHATS)]
1923 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1924 resizing buffers containing secrets, and use where appropriate.
1925 [Ben Laurie (CHATS)]
1927 *) Avoid using fixed size buffers for configuration file location.
1928 [Ben Laurie (CHATS)]
1930 *) Avoid filename truncation for various CA files.
1931 [Ben Laurie (CHATS)]
1933 *) Use sizeof in preference to magic numbers.
1934 [Ben Laurie (CHATS)]
1936 *) Avoid filename truncation in cert requests.
1937 [Ben Laurie (CHATS)]
1939 *) Add assertions to check for (supposedly impossible) buffer
1941 [Ben Laurie (CHATS)]
1943 *) Don't cache truncated DNS entries in the local cache (this could
1944 potentially lead to a spoofing attack).
1945 [Ben Laurie (CHATS)]
1947 *) Fix various buffers to be large enough for hex/decimal
1948 representations in a platform independent manner.
1949 [Ben Laurie (CHATS)]
1951 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1952 resizing buffers containing secrets, and use where appropriate.
1953 [Ben Laurie (CHATS)]
1955 *) Add BIO_indent() to avoid much slightly worrying code to do
1957 [Ben Laurie (CHATS)]
1959 *) Convert sprintf()/BIO_puts() to BIO_printf().
1960 [Ben Laurie (CHATS)]
1962 *) buffer_gets() could terminate with the buffer only half
1964 [Ben Laurie (CHATS)]
1966 *) Add assertions to prevent user-supplied crypto functions from
1967 overflowing internal buffers by having large block sizes, etc.
1968 [Ben Laurie (CHATS)]
1970 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1972 [Ben Laurie (CHATS)]
1974 *) Eliminate unused copy of key in RC4.
1975 [Ben Laurie (CHATS)]
1977 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1978 [Ben Laurie (CHATS)]
1980 *) Fix off-by-one error in EGD path.
1981 [Ben Laurie (CHATS)]
1983 *) If RANDFILE path is too long, ignore instead of truncating.
1984 [Ben Laurie (CHATS)]
1986 *) Eliminate unused and incorrectly sized X.509 structure
1988 [Ben Laurie (CHATS)]
1990 *) Eliminate unused and dangerous function knumber().
1991 [Ben Laurie (CHATS)]
1993 *) Eliminate unused and dangerous structure, KSSL_ERR.
1994 [Ben Laurie (CHATS)]
1996 *) Protect against overlong session ID context length in an encoded
1997 session object. Since these are local, this does not appear to be
1999 [Ben Laurie (CHATS)]
2001 *) Change from security patch (see 0.9.6e below) that did not affect
2002 the 0.9.6 release series:
2004 Remote buffer overflow in SSL3 protocol - an attacker could
2005 supply an oversized master key in Kerberos-enabled versions.
2007 [Ben Laurie (CHATS)]
2009 *) Change the SSL kerb5 codes to match RFC 2712.
2012 *) Make -nameopt work fully for req and add -reqopt switch.
2013 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2015 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2016 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2018 *) Make sure tests can be performed even if the corresponding algorithms
2019 have been removed entirely. This was also the last step to make
2020 OpenSSL compilable with DJGPP under all reasonable conditions.
2021 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2023 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2024 to allow version independent disabling of normally unselected ciphers,
2025 which may be activated as a side-effect of selecting a single cipher.
2027 (E.g., cipher list string "RSA" enables ciphersuites that are left
2028 out of "ALL" because they do not provide symmetric encryption.
2029 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2030 [Lutz Jaenicke, Bodo Moeller]
2032 *) Add appropriate support for separate platform-dependent build
2033 directories. The recommended way to make a platform-dependent
2034 build directory is the following (tested on Linux), maybe with
2037 # Place yourself outside of the OpenSSL source tree. In
2038 # this example, the environment variable OPENSSL_SOURCE
2039 # is assumed to contain the absolute OpenSSL source directory.
2040 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2041 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2042 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2043 mkdir -p `dirname $F`
2044 ln -s $OPENSSL_SOURCE/$F $F
2047 To be absolutely sure not to disturb the source tree, a "make clean"
2048 is a good thing. If it isn't successfull, don't worry about it,
2049 it probably means the source directory is very clean.
2052 *) Make sure any ENGINE control commands make local copies of string
2053 pointers passed to them whenever necessary. Otherwise it is possible
2054 the caller may have overwritten (or deallocated) the original string
2055 data when a later ENGINE operation tries to use the stored values.
2056 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2058 *) Improve diagnostics in file reading and command-line digests.
2059 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2061 *) Add AES modes CFB and OFB to the object database. Correct an
2062 error in AES-CFB decryption.
2065 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2066 allows existing EVP_CIPHER_CTX structures to be reused after
2067 calling EVP_*Final(). This behaviour is used by encryption
2068 BIOs and some applications. This has the side effect that
2069 applications must explicitly clean up cipher contexts with
2070 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2073 *) Check the values of dna and dnb in bn_mul_recursive before calling
2074 bn_mul_comba (a non zero value means the a or b arrays do not contain
2075 n2 elements) and fallback to bn_mul_normal if either is not zero.
2078 *) Fix escaping of non-ASCII characters when using the -subj option
2079 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2082 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2083 form for "surname", serialNumber has no short form.
2084 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2085 therefore remove "mail" short name for "internet 7".
2086 The OID for unique identifiers in X509 certificates is
2087 x500UniqueIdentifier, not uniqueIdentifier.
2088 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2091 *) Add an "init" command to the ENGINE config module and auto initialize
2092 ENGINEs. Without any "init" command the ENGINE will be initialized
2093 after all ctrl commands have been executed on it. If init=1 the
2094 ENGINE is initailized at that point (ctrls before that point are run
2095 on the uninitialized ENGINE and after on the initialized one). If
2096 init=0 then the ENGINE will not be iniatialized at all.
2099 *) Fix the 'app_verify_callback' interface so that the user-defined
2100 argument is actually passed to the callback: In the
2101 SSL_CTX_set_cert_verify_callback() prototype, the callback
2102 declaration has been changed from
2105 int (*cb)(X509_STORE_CTX *,void *);
2106 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2107 i=s->ctx->app_verify_callback(&ctx)
2108 has been changed into
2109 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2111 To update applications using SSL_CTX_set_cert_verify_callback(),
2112 a dummy argument can be added to their callback functions.
2113 [D. K. Smetters <smetters@parc.xerox.com>]
2115 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2116 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2118 *) Add and OPENSSL_LOAD_CONF define which will cause
2119 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2120 This allows older applications to transparently support certain
2121 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2122 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2123 load the config file and OPENSSL_add_all_algorithms_conf() which will
2124 always load it have also been added.
2127 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2128 Adjust NIDs and EVP layer.
2129 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2131 *) Config modules support in openssl utility.
2133 Most commands now load modules from the config file,
2134 though in a few (such as version) this isn't done
2135 because it couldn't be used for anything.
2137 In the case of ca and req the config file used is
2138 the same as the utility itself: that is the -config
2139 command line option can be used to specify an
2143 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2144 use "openssl_conf" if filename is NULL use default openssl config file.
2147 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2148 config section name. Add a new flag to tolerate a missing config file
2149 and move code to CONF_modules_load_file().
2152 *) Support for crypto accelerator cards from Accelerated Encryption
2153 Processing, www.aep.ie. (Use engine 'aep')
2154 The support was copied from 0.9.6c [engine] and adapted/corrected
2155 to work with the new engine framework.
2156 [AEP Inc. and Richard Levitte]
2158 *) Support for SureWare crypto accelerator cards from Baltimore
2159 Technologies. (Use engine 'sureware')
2160 The support was copied from 0.9.6c [engine] and adapted
2161 to work with the new engine framework.
2164 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2165 make the newer ENGINE framework commands for the CHIL engine work.
2166 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2168 *) Make it possible to produce shared libraries on ReliantUNIX.
2169 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2171 *) Add the configuration target debug-linux-ppro.
2172 Make 'openssl rsa' use the general key loading routines
2173 implemented in apps.c, and make those routines able to
2174 handle the key format FORMAT_NETSCAPE and the variant
2176 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2178 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2179 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2181 *) Add -keyform to rsautl, and document -engine.
2182 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2184 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2185 BIO_R_NO_SUCH_FILE error code rather than the generic
2186 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2189 *) Add new functions
2191 ERR_peek_last_error_line
2192 ERR_peek_last_error_line_data.
2193 These are similar to
2196 ERR_peek_error_line_data,
2197 but report on the latest error recorded rather than the first one
2198 still in the error queue.
2199 [Ben Laurie, Bodo Moeller]
2201 *) default_algorithms option in ENGINE config module. This allows things
2203 default_algorithms = ALL
2204 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2207 *) Prelminary ENGINE config module.
2210 *) New experimental application configuration code.
2213 *) Change the AES code to follow the same name structure as all other
2214 symmetric ciphers, and behave the same way. Move everything to
2215 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2216 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2218 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2219 [Ben Laurie and Theo de Raadt]
2221 *) Add option to output public keys in req command.
2222 [Massimiliano Pala madwolf@openca.org]
2224 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2225 (up to about 10% better than before for P-192 and P-224).
2228 *) New functions/macros
2230 SSL_CTX_set_msg_callback(ctx, cb)
2231 SSL_CTX_set_msg_callback_arg(ctx, arg)
2232 SSL_set_msg_callback(ssl, cb)
2233 SSL_set_msg_callback_arg(ssl, arg)
2235 to request calling a callback function
2237 void cb(int write_p, int version, int content_type,
2238 const void *buf, size_t len, SSL *ssl, void *arg)
2240 whenever a protocol message has been completely received
2241 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2242 protocol version according to which the SSL library interprets
2243 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2244 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2245 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2246 specification (change_cipher_spec(20), alert(21), handshake(22)).
2247 'buf' and 'len' point to the actual message, 'ssl' to the
2248 SSL object, and 'arg' is the application-defined value set by
2249 SSL[_CTX]_set_msg_callback_arg().
2251 'openssl s_client' and 'openssl s_server' have new '-msg' options
2252 to enable a callback that displays all protocol messages.
2255 *) Change the shared library support so shared libraries are built as
2256 soon as the corresponding static library is finished, and thereby get
2257 openssl and the test programs linked against the shared library.
2258 This still only happens when the keyword "shard" has been given to
2259 the configuration scripts.
2261 NOTE: shared library support is still an experimental thing, and
2262 backward binary compatibility is still not guaranteed.
2263 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2265 *) Add support for Subject Information Access extension.
2266 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2268 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2269 additional bytes when new memory had to be allocated, not just
2270 when reusing an existing buffer.
2273 *) New command line and configuration option 'utf8' for the req command.
2274 This allows field values to be specified as UTF8 strings.
2277 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2278 runs for the former and machine-readable output for the latter.
2281 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2282 of the e-mail address in the DN (i.e., it will go into a certificate
2283 extension only). The new configuration file option 'email_in_dn = no'
2284 has the same effect.
2285 [Massimiliano Pala madwolf@openca.org]
2287 *) Change all functions with names starting with des_ to be starting
2288 with DES_ instead. Add wrappers that are compatible with libdes,
2289 but are named _ossl_old_des_*. Finally, add macros that map the
2290 des_* symbols to the corresponding _ossl_old_des_* if libdes
2291 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2292 desired, the des_* symbols will be mapped to DES_*, with one
2295 Since we provide two compatibility mappings, the user needs to
2296 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2297 compatibility is desired. The default (i.e., when that macro
2298 isn't defined) is OpenSSL 0.9.6c compatibility.
2300 There are also macros that enable and disable the support of old
2301 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2302 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2303 are defined, the default will apply: to support the old des routines.
2305 In either case, one must include openssl/des.h to get the correct
2306 definitions. Do not try to just include openssl/des_old.h, that
2309 NOTE: This is a major break of an old API into a new one. Software
2310 authors are encouraged to switch to the DES_ style functions. Some
2311 time in the future, des_old.h and the libdes compatibility functions
2312 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2313 default), and then completely removed.
2316 *) Test for certificates which contain unsupported critical extensions.
2317 If such a certificate is found during a verify operation it is
2318 rejected by default: this behaviour can be overridden by either
2319 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2320 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2321 X509_supported_extension() has also been added which returns 1 if a
2322 particular extension is supported.
2325 *) Modify the behaviour of EVP cipher functions in similar way to digests
2326 to retain compatibility with existing code.
2329 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2330 compatibility with existing code. In particular the 'ctx' parameter does
2331 not have to be to be initialized before the call to EVP_DigestInit() and
2332 it is tidied up after a call to EVP_DigestFinal(). New function
2333 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2334 EVP_MD_CTX_copy() changed to not require the destination to be
2335 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2336 requires the destination to be valid.
2338 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2339 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2342 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2343 so that complete 'Handshake' protocol structures are kept in memory
2344 instead of overwriting 'msg_type' and 'length' with 'body' data.
2347 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2348 [Massimo Santin via Richard Levitte]
2350 *) Major restructuring to the underlying ENGINE code. This includes
2351 reduction of linker bloat, separation of pure "ENGINE" manipulation
2352 (initialisation, etc) from functionality dealing with implementations
2353 of specific crypto iterfaces. This change also introduces integrated
2354 support for symmetric ciphers and digest implementations - so ENGINEs
2355 can now accelerate these by providing EVP_CIPHER and EVP_MD
2356 implementations of their own. This is detailed in crypto/engine/README
2357 as it couldn't be adequately described here. However, there are a few
2358 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2359 were changed in the original introduction of ENGINE code have now
2360 reverted back - the hooking from this code to ENGINE is now a good
2361 deal more passive and at run-time, operations deal directly with
2362 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2363 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2364 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2365 they were not being used by the framework as there is no concept of a
2366 BIGNUM_METHOD and they could not be generalised to the new
2367 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2368 ENGINE_cpy() has been removed as it cannot be consistently defined in
2372 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2375 *) Change mkdef.pl to sort symbols that get the same entry number,
2376 and make sure the automatically generated functions ERR_load_*
2377 become part of libeay.num as well.
2380 *) New function SSL_renegotiate_pending(). This returns true once
2381 renegotiation has been requested (either SSL_renegotiate() call
2382 or HelloRequest/ClientHello receveived from the peer) and becomes
2383 false once a handshake has been completed.
2384 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2385 sends a HelloRequest, but does not ensure that a handshake takes
2386 place. SSL_renegotiate_pending() is useful for checking if the
2387 client has followed the request.)
2390 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2391 By default, clients may request session resumption even during
2392 renegotiation (if session ID contexts permit); with this option,
2393 session resumption is possible only in the first handshake.
2395 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2396 more bits available for options that should not be part of
2397 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2400 *) Add some demos for certificate and certificate request creation.
2403 *) Make maximum certificate chain size accepted from the peer application
2404 settable (SSL*_get/set_max_cert_list()), as proposed by
2405 "Douglas E. Engert" <deengert@anl.gov>.
2408 *) Add support for shared libraries for Unixware-7
2409 (Boyd Lynn Gerber <gerberb@zenez.com>).
2412 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2413 be done prior to destruction. Use this to unload error strings from
2414 ENGINEs that load their own error strings. NB: This adds two new API
2415 functions to "get" and "set" this destroy handler in an ENGINE.
2418 *) Alter all existing ENGINE implementations (except "openssl" and
2419 "openbsd") to dynamically instantiate their own error strings. This
2420 makes them more flexible to be built both as statically-linked ENGINEs
2421 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2422 Also, add stub code to each that makes building them as self-contained
2423 shared-libraries easier (see README.ENGINE).
2426 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2427 implementations into applications that are completely implemented in
2428 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2429 commands that can be used to configure what shared-library to load and
2430 to control aspects of the way it is handled. Also, made an update to
2431 the README.ENGINE file that brings its information up-to-date and
2432 provides some information and instructions on the "dynamic" ENGINE
2433 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2436 *) Make it possible to unload ranges of ERR strings with a new
2437 "ERR_unload_strings" function.
2440 *) Add a copy() function to EVP_MD.
2443 *) Make EVP_MD routines take a context pointer instead of just the
2444 md_data void pointer.
2447 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2448 that the digest can only process a single chunk of data
2449 (typically because it is provided by a piece of
2450 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2451 is only going to provide a single chunk of data, and hence the
2452 framework needn't accumulate the data for oneshot drivers.
2455 *) As with "ERR", make it possible to replace the underlying "ex_data"
2456 functions. This change also alters the storage and management of global
2457 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2458 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2459 index counters. The API functions that use this state have been changed
2460 to take a "class_index" rather than pointers to the class's local STACK
2461 and counter, and there is now an API function to dynamically create new
2462 classes. This centralisation allows us to (a) plug a lot of the
2463 thread-safety problems that existed, and (b) makes it possible to clean
2464 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2465 such data would previously have always leaked in application code and
2466 workarounds were in place to make the memory debugging turn a blind eye
2467 to it. Application code that doesn't use this new function will still
2468 leak as before, but their memory debugging output will announce it now
2469 rather than letting it slide.
2471 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2472 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2473 has a return value to indicate success or failure.
2476 *) Make it possible to replace the underlying "ERR" functions such that the
2477 global state (2 LHASH tables and 2 locks) is only used by the "default"
2478 implementation. This change also adds two functions to "get" and "set"
2479 the implementation prior to it being automatically set the first time
2480 any other ERR function takes place. Ie. an application can call "get",
2481 pass the return value to a module it has just loaded, and that module
2482 can call its own "set" function using that value. This means the
2483 module's "ERR" operations will use (and modify) the error state in the
2484 application and not in its own statically linked copy of OpenSSL code.
2487 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2488 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2489 the operation, and provides a more encapsulated way for external code
2490 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2491 to use these functions rather than manually incrementing the counts.
2493 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2496 *) Add EVP test program.
2499 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2502 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2503 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2504 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2505 These allow a CRL to be built without having to access X509_CRL fields
2506 directly. Modify 'ca' application to use new functions.
2509 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2510 bug workarounds. Rollback attack detection is a security feature.
2511 The problem will only arise on OpenSSL servers when TLSv1 is not
2512 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2513 Software authors not wanting to support TLSv1 will have special reasons
2514 for their choice and can explicitly enable this option.
2515 [Bodo Moeller, Lutz Jaenicke]
2517 *) Rationalise EVP so it can be extended: don't include a union of
2518 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2519 (similar to those existing for EVP_CIPHER_CTX).
2524 EVP_MD_CTX_init(&md); /* new function call */
2525 EVP_DigestInit(&md, EVP_sha1());
2526 EVP_DigestUpdate(&md, in, len);
2527 EVP_DigestFinal(&md, out, NULL);
2528 EVP_MD_CTX_cleanup(&md); /* new function call */
2532 *) Make DES key schedule conform to the usual scheme, as well as
2533 correcting its structure. This means that calls to DES functions
2534 now have to pass a pointer to a des_key_schedule instead of a
2535 plain des_key_schedule (which was actually always a pointer
2538 des_key_schedule ks;
2540 des_set_key_checked(..., &ks);
2541 des_ncbc_encrypt(..., &ks, ...);
2543 (Note that a later change renames 'des_...' into 'DES_...'.)
2546 *) Initial reduction of linker bloat: the use of some functions, such as
2547 PEM causes large amounts of unused functions to be linked in due to
2548 poor organisation. For example pem_all.c contains every PEM function
2549 which has a knock on effect of linking in large amounts of (unused)
2550 ASN1 code. Grouping together similar functions and splitting unrelated
2551 functions prevents this.
2554 *) Cleanup of EVP macros.
2557 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2558 correct _ecb suffix.
2561 *) Add initial OCSP responder support to ocsp application. The
2562 revocation information is handled using the text based index
2563 use by the ca application. The responder can either handle
2564 requests generated internally, supplied in files (for example
2565 via a CGI script) or using an internal minimal server.
2568 *) Add configuration choices to get zlib compression for TLS.
2571 *) Changes to Kerberos SSL for RFC 2712 compliance:
2572 1. Implemented real KerberosWrapper, instead of just using
2573 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2574 2. Implemented optional authenticator field of KerberosWrapper.
2576 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2577 and authenticator structs; see crypto/krb5/.
2579 Generalized Kerberos calls to support multiple Kerberos libraries.
2580 [Vern Staats <staatsvr@asc.hpc.mil>,
2581 Jeffrey Altman <jaltman@columbia.edu>
2582 via Richard Levitte]
2584 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2585 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2586 values for each of the key sizes rather than having just
2587 parameters (and 'speed' generating keys each time).
2590 *) Speed up EVP routines.
2593 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2594 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2595 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2596 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2598 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2599 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2600 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2603 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2605 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2608 *) Added the OS2-EMX target.
2609 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2611 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2612 to support NCONF routines in extension code. New function CONF_set_nconf()
2613 to allow functions which take an NCONF to also handle the old LHASH
2614 structure: this means that the old CONF compatible routines can be
2615 retained (in particular wrt extensions) without having to duplicate the
2616 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2619 *) Enhance the general user interface with mechanisms for inner control
2620 and with possibilities to have yes/no kind of prompts.
2623 *) Change all calls to low level digest routines in the library and
2624 applications to use EVP. Add missing calls to HMAC_cleanup() and
2625 don't assume HMAC_CTX can be copied using memcpy().
2626 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2628 *) Add the possibility to control engines through control names but with
2629 arbitrary arguments instead of just a string.
2630 Change the key loaders to take a UI_METHOD instead of a callback
2631 function pointer. NOTE: this breaks binary compatibility with earlier
2632 versions of OpenSSL [engine].
2633 Adapt the nCipher code for these new conditions and add a card insertion
2637 *) Enhance the general user interface with mechanisms to better support
2638 dialog box interfaces, application-defined prompts, the possibility
2639 to use defaults (for example default passwords from somewhere else)
2640 and interrupts/cancellations.
2643 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2644 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2647 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2648 tidy up some unnecessarily weird code in 'sk_new()').
2649 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2651 *) Change the key loading routines for ENGINEs to use the same kind
2652 callback (pem_password_cb) as all other routines that need this
2656 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2657 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2658 than this minimum value is recommended.
2661 *) New random seeder for OpenVMS, using the system process statistics
2662 that are easily reachable.
2665 *) Windows apparently can't transparently handle global
2666 variables defined in DLLs. Initialisations such as:
2668 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2670 wont compile. This is used by the any applications that need to
2671 declare their own ASN1 modules. This was fixed by adding the option
2672 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2673 needed for static libraries under Win32.
2676 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2677 setting of purpose and trust fields. New X509_STORE trust and
2678 purpose functions and tidy up setting in other SSL functions.
2681 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2682 structure. These are inherited by X509_STORE_CTX when it is
2683 initialised. This allows various defaults to be set in the
2684 X509_STORE structure (such as flags for CRL checking and custom
2685 purpose or trust settings) for functions which only use X509_STORE_CTX
2686 internally such as S/MIME.
2688 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2689 trust settings if they are not set in X509_STORE. This allows X509_STORE
2690 purposes and trust (in S/MIME for example) to override any set by default.
2692 Add command line options for CRL checking to smime, s_client and s_server
2696 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2697 are set then the CRL is looked up in the X509_STORE structure and
2698 its validity and signature checked, then if the certificate is found
2699 in the CRL the verify fails with a revoked error.
2701 Various new CRL related callbacks added to X509_STORE_CTX structure.
2703 Command line options added to 'verify' application to support this.
2705 This needs some additional work, such as being able to handle multiple
2706 CRLs with different times, extension based lookup (rather than just
2707 by subject name) and ultimately more complete V2 CRL extension
2711 *) Add a general user interface API (crypto/ui/). This is designed
2712 to replace things like des_read_password and friends (backward
2713 compatibility functions using this new API are provided).
2714 The purpose is to remove prompting functions from the DES code
2715 section as well as provide for prompting through dialog boxes in
2716 a window system and the like.
2719 *) Add "ex_data" support to ENGINE so implementations can add state at a
2720 per-structure level rather than having to store it globally.
2723 *) Make it possible for ENGINE structures to be copied when retrieved by
2724 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2725 This causes the "original" ENGINE structure to act like a template,
2726 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2727 operational state can be localised to each ENGINE structure, despite the
2728 fact they all share the same "methods". New ENGINE structures returned in
2729 this case have no functional references and the return value is the single
2730 structural reference. This matches the single structural reference returned
2731 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2735 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2736 needs to match any other type at all we need to manually clear the
2740 *) Changes to the "openssl engine" utility to include;
2741 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2742 about an ENGINE's available control commands.
2743 - executing control commands from command line arguments using the
2744 '-pre' and '-post' switches. '-post' is only used if '-t' is
2745 specified and the ENGINE is successfully initialised. The syntax for
2746 the individual commands are colon-separated, for example;
2747 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2750 *) New dynamic control command support for ENGINEs. ENGINEs can now
2751 declare their own commands (numbers), names (strings), descriptions,
2752 and input types for run-time discovery by calling applications. A
2753 subset of these commands are implicitly classed as "executable"
2754 depending on their input type, and only these can be invoked through
2755 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2756 can be based on user input, config files, etc). The distinction is
2757 that "executable" commands cannot return anything other than a boolean
2758 result and can only support numeric or string input, whereas some
2759 discoverable commands may only be for direct use through
2760 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2761 pointers, or other custom uses. The "executable" commands are to
2762 support parameterisations of ENGINE behaviour that can be
2763 unambiguously defined by ENGINEs and used consistently across any
2764 OpenSSL-based application. Commands have been added to all the
2765 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2766 control over shared-library paths without source code alterations.
2769 *) Changed all ENGINE implementations to dynamically allocate their
2770 ENGINEs rather than declaring them statically. Apart from this being
2771 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2772 this also allows the implementations to compile without using the
2773 internal engine_int.h header.
2776 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2777 'const' value. Any code that should be able to modify a RAND_METHOD
2778 should already have non-const pointers to it (ie. they should only
2779 modify their own ones).
2782 *) Made a variety of little tweaks to the ENGINE code.
2783 - "atalla" and "ubsec" string definitions were moved from header files
2784 to C code. "nuron" string definitions were placed in variables
2785 rather than hard-coded - allowing parameterisation of these values
2786 later on via ctrl() commands.
2787 - Removed unused "#if 0"'d code.
2788 - Fixed engine list iteration code so it uses ENGINE_free() to release
2789 structural references.
2790 - Constified the RAND_METHOD element of ENGINE structures.
2791 - Constified various get/set functions as appropriate and added
2792 missing functions (including a catch-all ENGINE_cpy that duplicates
2793 all ENGINE values onto a new ENGINE except reference counts/state).
2794 - Removed NULL parameter checks in get/set functions. Setting a method
2795 or function to NULL is a way of cancelling out a previously set
2796 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2797 and doesn't justify the extra error symbols and code.
2798 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2799 flags from engine_int.h to engine.h.
2800 - Changed prototypes for ENGINE handler functions (init(), finish(),
2801 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2804 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2805 to the algorithm using long division. The binary algorithm can be
2806 used only if the modulus is odd. On 32-bit systems, it is faster
2807 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2808 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2809 up to 450 bits. In 64-bit environments, the binary algorithm
2810 appears to be advantageous for much longer moduli; here we use it
2811 for moduli up to 2048 bits.
2814 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2815 could not support the combine flag in choice fields.
2818 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2819 extensions from a certificate request to the certificate.
2822 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2823 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2824 file: this allows the display of the certificate about to be
2825 signed to be customised, to allow certain fields to be included
2826 or excluded and extension details. The old system didn't display
2827 multicharacter strings properly, omitted fields not in the policy
2828 and couldn't display additional details such as extensions.
2831 *) Function EC_POINTs_mul for multiple scalar multiplication
2832 of an arbitrary number of elliptic curve points
2833 \sum scalars[i]*points[i],
2834 optionally including the generator defined for the EC_GROUP:
2835 scalar*generator + \sum scalars[i]*points[i].
2837 EC_POINT_mul is a simple wrapper function for the typical case
2838 that the point list has just one item (besides the optional
2842 *) First EC_METHODs for curves over GF(p):
2844 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2845 operations and provides various method functions that can also
2846 operate with faster implementations of modular arithmetic.
2848 EC_GFp_mont_method() reuses most functions that are part of
2849 EC_GFp_simple_method, but uses Montgomery arithmetic.
2851 [Bodo Moeller; point addition and point doubling
2852 implementation directly derived from source code provided by
2853 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2855 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2856 crypto/ec/ec_lib.c):
2858 Curves are EC_GROUP objects (with an optional group generator)
2859 based on EC_METHODs that are built into the library.
2861 Points are EC_POINT objects based on EC_GROUP objects.
2863 Most of the framework would be able to handle curves over arbitrary
2864 finite fields, but as there are no obvious types for fields other
2865 than GF(p), some functions are limited to that for now.
2868 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2869 that the file contains a complete HTTP response.
2872 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2873 change the def and num file printf format specifier from "%-40sXXX"
2874 to "%-39s XXX". The latter will always guarantee a space after the
2875 field while the former will cause them to run together if the field
2876 is 40 of more characters long.
2879 *) Constify the cipher and digest 'method' functions and structures
2880 and modify related functions to take constant EVP_MD and EVP_CIPHER
2884 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2885 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2888 *) Modify EVP_Digest*() routines so they now return values. Although the
2889 internal software routines can never fail additional hardware versions
2893 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2895 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2896 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2899 ERR_R_NESTED_ASN1_ERROR
2901 ERR_R_MISSING_ASN1_EOS
2902 were 4 .. 9, conflicting with
2903 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2905 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2906 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2908 Add new error code 'ERR_R_INTERNAL_ERROR'.
2911 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2915 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2916 sets the subject name for a new request or supersedes the
2917 subject name in a given request. Formats that can be parsed are
2918 'CN=Some Name, OU=myOU, C=IT'
2920 'CN=Some Name/OU=myOU/C=IT'.
2922 Add options '-batch' and '-verbose' to 'openssl req'.
2923 [Massimiliano Pala <madwolf@hackmasters.net>]
2925 *) Introduce the possibility to access global variables through
2926 functions on platform were that's the best way to handle exporting
2927 global variables in shared libraries. To enable this functionality,
2928 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2929 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2930 is normally done by Configure or something similar).
2932 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2933 in the source file (foo.c) like this:
2935 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2936 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2938 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2939 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2941 OPENSSL_DECLARE_GLOBAL(int,foo);
2942 #define foo OPENSSL_GLOBAL_REF(foo)
2943 OPENSSL_DECLARE_GLOBAL(double,bar);
2944 #define bar OPENSSL_GLOBAL_REF(bar)
2946 The #defines are very important, and therefore so is including the
2947 header file everywhere where the defined globals are used.
2949 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2950 of ASN.1 items, but that structure is a bit different.
2952 The largest change is in util/mkdef.pl which has been enhanced with
2953 better and easier to understand logic to choose which symbols should
2954 go into the Windows .def files as well as a number of fixes and code
2955 cleanup (among others, algorithm keywords are now sorted
2956 lexicographically to avoid constant rewrites).
2959 *) In BN_div() keep a copy of the sign of 'num' before writing the
2960 result to 'rm' because if rm==num the value will be overwritten
2961 and produce the wrong result if 'num' is negative: this caused
2962 problems with BN_mod() and BN_nnmod().
2965 *) Function OCSP_request_verify(). This checks the signature on an
2966 OCSP request and verifies the signer certificate. The signer
2967 certificate is just checked for a generic purpose and OCSP request
2971 *) Add OCSP_check_validity() function to check the validity of OCSP
2972 responses. OCSP responses are prepared in real time and may only
2973 be a few seconds old. Simply checking that the current time lies
2974 between thisUpdate and nextUpdate max reject otherwise valid responses
2975 caused by either OCSP responder or client clock inaccuracy. Instead
2976 we allow thisUpdate and nextUpdate to fall within a certain period of
2977 the current time. The age of the response can also optionally be
2978 checked. Two new options -validity_period and -status_age added to
2982 *) If signature or public key algorithm is unrecognized print out its
2983 OID rather that just UNKNOWN.
2986 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2987 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2988 ID to be generated from the issuer certificate alone which can then be
2989 passed to OCSP_id_issuer_cmp().
2992 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2993 ASN1 modules to export functions returning ASN1_ITEM pointers
2994 instead of the ASN1_ITEM structures themselves. This adds several
2995 new macros which allow the underlying ASN1 function/structure to
2996 be accessed transparently. As a result code should not use ASN1_ITEM
2997 references directly (such as &X509_it) but instead use the relevant
2998 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2999 use of the new ASN1 code on platforms where exporting structures
3000 is problematical (for example in shared libraries) but exporting
3001 functions returning pointers to structures is not.
3004 *) Add support for overriding the generation of SSL/TLS session IDs.
3005 These callbacks can be registered either in an SSL_CTX or per SSL.
3006 The purpose of this is to allow applications to control, if they wish,
3007 the arbitrary values chosen for use as session IDs, particularly as it
3008 can be useful for session caching in multiple-server environments. A
3009 command-line switch for testing this (and any client code that wishes
3010 to use such a feature) has been added to "s_server".
3011 [Geoff Thorpe, Lutz Jaenicke]
3013 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3014 of the form '#if defined(...) || defined(...) || ...' and
3015 '#if !defined(...) && !defined(...) && ...'. This also avoids
3016 the growing number of special cases it was previously handling.
3019 *) Make all configuration macros available for application by making
3020 sure they are available in opensslconf.h, by giving them names starting
3021 with "OPENSSL_" to avoid conflicts with other packages and by making
3022 sure e_os2.h will cover all platform-specific cases together with
3024 Additionally, it is now possible to define configuration/platform-
3025 specific names (called "system identities"). In the C code, these
3026 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3027 macro with the name beginning with "OPENSSL_SYS_", which is determined
3028 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3032 *) New option -set_serial to 'req' and 'x509' this allows the serial
3033 number to use to be specified on the command line. Previously self
3034 signed certificates were hard coded with serial number 0 and the
3035 CA options of 'x509' had to use a serial number in a file which was
3039 *) New options to 'ca' utility to support V2 CRL entry extensions.
3040 Currently CRL reason, invalidity date and hold instruction are
3041 supported. Add new CRL extensions to V3 code and some new objects.
3044 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3045 disable standard block padding (aka PKCS#5 padding) in the EVP
3046 API, which was previously mandatory. This means that the data is
3047 not padded in any way and so the total length much be a multiple
3048 of the block size, otherwise an error occurs.
3051 *) Initial (incomplete) OCSP SSL support.
3054 *) New function OCSP_parse_url(). This splits up a URL into its host,
3055 port and path components: primarily to parse OCSP URLs. New -url
3056 option to ocsp utility.
3059 *) New nonce behavior. The return value of OCSP_check_nonce() now
3060 reflects the various checks performed. Applications can decide
3061 whether to tolerate certain situations such as an absent nonce
3062 in a response when one was present in a request: the ocsp application
3063 just prints out a warning. New function OCSP_add1_basic_nonce()
3064 this is to allow responders to include a nonce in a response even if
3065 the request is nonce-less.
3068 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3069 skipped when using openssl x509 multiple times on a single input file,
3070 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3073 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3074 set string type: to handle setting ASN1_TIME structures. Fix ca
3075 utility to correctly initialize revocation date of CRLs.
3078 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3079 the clients preferred ciphersuites and rather use its own preferences.
3080 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3081 Internet Explorer by ensuring unchanged hash method during stepup.
3082 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3085 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3086 to aes and add a new 'exist' option to print out symbols that don't
3090 *) Additional options to ocsp utility to allow flags to be set and
3091 additional certificates supplied.
3094 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3095 OCSP client a number of certificate to only verify the response
3099 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3100 handle the new API. Currently only ECB, CBC modes supported. Add new
3103 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3104 Encryption Standard (AES) Ciphersuites for Transport Layer
3105 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3106 not enabled by default and were not part of the "ALL" ciphersuite
3107 alias because they were not yet official; they could be
3108 explicitly requested by specifying the "AESdraft" ciphersuite
3109 group alias. In the final release of OpenSSL 0.9.7, the group
3110 alias is called "AES" and is part of "ALL".)
3111 [Ben Laurie, Steve Henson, Bodo Moeller]
3113 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3114 request to response.
3117 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3118 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3119 extract information from a certificate request. OCSP_response_create()
3120 creates a response and optionally adds a basic response structure.
3121 OCSP_basic_add1_status() adds a complete single response to a basic
3122 response and returns the OCSP_SINGLERESP structure just added (to allow
3123 extensions to be included for example). OCSP_basic_add1_cert() adds a
3124 certificate to a basic response and OCSP_basic_sign() signs a basic
3125 response with various flags. New helper functions ASN1_TIME_check()
3126 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3127 (converts ASN1_TIME to GeneralizedTime).
3130 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3131 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3132 structure from a certificate. X509_pubkey_digest() digests the public_key
3133 contents: this is used in various key identifiers.
3136 *) Make sk_sort() tolerate a NULL argument.
3137 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3139 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3140 passed by the function are trusted implicitly. If any of them signed the
3141 response then it is assumed to be valid and is not verified.
3144 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3145 to data. This was previously part of the PKCS7 ASN1 code. This
3146 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3147 [Steve Henson, reported by Kenneth R. Robinette
3148 <support@securenetterm.com>]
3150 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3151 routines: without these tracing memory leaks is very painful.
3152 Fix leaks in PKCS12 and PKCS7 routines.
3155 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3156 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3157 effectively meant GeneralizedTime would never be used. Now it
3158 is initialised to -1 but X509_time_adj() now has to check the value
3159 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3160 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3161 [Steve Henson, reported by Kenneth R. Robinette
3162 <support@securenetterm.com>]
3164 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3165 result in a zero length in the ASN1_INTEGER structure which was
3166 not consistent with the structure when d2i_ASN1_INTEGER() was used
3167 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3168 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3169 where it did not print out a minus for negative ASN1_INTEGER.
3172 *) Add summary printout to ocsp utility. The various functions which
3173 convert status values to strings have been renamed to:
3174 OCSP_response_status_str(), OCSP_cert_status_str() and
3175 OCSP_crl_reason_str() and are no longer static. New options
3176 to verify nonce values and to disable verification. OCSP response
3177 printout format cleaned up.
3180 *) Add additional OCSP certificate checks. These are those specified
3181 in RFC2560. This consists of two separate checks: the CA of the
3182 certificate being checked must either be the OCSP signer certificate
3183 or the issuer of the OCSP signer certificate. In the latter case the
3184 OCSP signer certificate must contain the OCSP signing extended key
3185 usage. This check is performed by attempting to match the OCSP
3186 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3187 in the OCSP_CERTID structures of the response.
3190 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3191 and related routines. This uses the standard OpenSSL certificate
3192 verify routines to perform initial checks (just CA validity) and
3193 to obtain the certificate chain. Then additional checks will be
3194 performed on the chain. Currently the root CA is checked to see
3195 if it is explicitly trusted for OCSP signing. This is used to set
3196 a root CA as a global signing root: that is any certificate that
3197 chains to that CA is an acceptable OCSP signing certificate.
3200 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3201 extensions from a separate configuration file.
3202 As when reading extensions from the main configuration file,
3203 the '-extensions ...' option may be used for specifying the
3205 [Massimiliano Pala <madwolf@comune.modena.it>]
3207 *) New OCSP utility. Allows OCSP requests to be generated or
3208 read. The request can be sent to a responder and the output
3209 parsed, outputed or printed in text form. Not complete yet:
3210 still needs to check the OCSP response validity.
3213 *) New subcommands for 'openssl ca':
3214 'openssl ca -status <serial>' prints the status of the cert with
3215 the given serial number (according to the index file).
3216 'openssl ca -updatedb' updates the expiry status of certificates
3218 [Massimiliano Pala <madwolf@comune.modena.it>]
3220 *) New '-newreq-nodes' command option to CA.pl. This is like
3221 '-newreq', but calls 'openssl req' with the '-nodes' option
3222 so that the resulting key is not encrypted.
3223 [Damien Miller <djm@mindrot.org>]
3225 *) New configuration for the GNU Hurd.
3226 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3228 *) Initial code to implement OCSP basic response verify. This
3229 is currently incomplete. Currently just finds the signer's
3230 certificate and verifies the signature on the response.
3233 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3234 value of OPENSSLDIR. This is available via the new '-d' option
3235 to 'openssl version', and is also included in 'openssl version -a'.
3238 *) Allowing defining memory allocation callbacks that will be given
3239 file name and line number information in additional arguments
3240 (a const char* and an int). The basic functionality remains, as
3241 well as the original possibility to just replace malloc(),
3242 realloc() and free() by functions that do not know about these
3243 additional arguments. To register and find out the current
3244 settings for extended allocation functions, the following
3245 functions are provided:
3247 CRYPTO_set_mem_ex_functions
3248 CRYPTO_set_locked_mem_ex_functions
3249 CRYPTO_get_mem_ex_functions
3250 CRYPTO_get_locked_mem_ex_functions
3252 These work the same way as CRYPTO_set_mem_functions and friends.
3253 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3254 extended allocation function is enabled.
3255 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3256 a conventional allocation function is enabled.
3257 [Richard Levitte, Bodo Moeller]
3259 *) Finish off removing the remaining LHASH function pointer casts.
3260 There should no longer be any prototype-casting required when using
3261 the LHASH abstraction, and any casts that remain are "bugs". See
3262 the callback types and macros at the head of lhash.h for details
3263 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3266 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3267 If /dev/[u]random devices are not available or do not return enough
3268 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3270 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3271 /etc/entropy will be queried once each in this sequence, quering stops
3272 when enough entropy was collected without querying more sockets.
3275 *) Change the Unix RAND_poll() variant to be able to poll several
3276 random devices, as specified by DEVRANDOM, until a sufficient amount
3277 of data has been collected. We spend at most 10 ms on each file
3278 (select timeout) and read in non-blocking mode. DEVRANDOM now
3279 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3280 (previously it was just the string "/dev/urandom"), so on typical
3281 platforms the 10 ms delay will never occur.
3282 Also separate out the Unix variant to its own file, rand_unix.c.
3283 For VMS, there's a currently-empty rand_vms.c.
3286 *) Move OCSP client related routines to ocsp_cl.c. These
3287 provide utility functions which an application needing
3288 to issue a request to an OCSP responder and analyse the
3289 response will typically need: as opposed to those which an
3290 OCSP responder itself would need which will be added later.
3292 OCSP_request_sign() signs an OCSP request with an API similar
3293 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3294 response. OCSP_response_get1_basic() extracts basic response
3295 from response. OCSP_resp_find_status(): finds and extracts status
3296 information from an OCSP_CERTID structure (which will be created
3297 when the request structure is built). These are built from lower
3298 level functions which work on OCSP_SINGLERESP structures but
3299 wont normally be used unless the application wishes to examine
3300 extensions in the OCSP response for example.
3302 Replace nonce routines with a pair of functions.
3303 OCSP_request_add1_nonce() adds a nonce value and optionally
3304 generates a random value. OCSP_check_nonce() checks the
3305 validity of the nonce in an OCSP response.
3308 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3309 This doesn't copy the supplied OCSP_CERTID and avoids the
3310 need to free up the newly created id. Change return type
3311 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3312 This can then be used to add extensions to the request.
3313 Deleted OCSP_request_new(), since most of its functionality
3314 is now in OCSP_REQUEST_new() (and the case insensitive name
3315 clash) apart from the ability to set the request name which
3316 will be added elsewhere.
3319 *) Update OCSP API. Remove obsolete extensions argument from
3320 various functions. Extensions are now handled using the new
3321 OCSP extension code. New simple OCSP HTTP function which
3322 can be used to send requests and parse the response.
3325 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3326 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3327 uses the special reorder version of SET OF to sort the attributes
3328 and reorder them to match the encoded order. This resolves a long
3329 standing problem: a verify on a PKCS7 structure just after signing
3330 it used to fail because the attribute order did not match the
3331 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3332 it uses the received order. This is necessary to tolerate some broken
3333 software that does not order SET OF. This is handled by encoding
3334 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3335 to produce the required SET OF.
3338 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3339 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3340 files to get correct declarations of the ASN.1 item variables.
3343 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3344 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3345 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3346 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3347 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3348 ASN1_ITEM and no wrapper functions.
3351 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3352 replace the old function pointer based I/O routines. Change most of
3353 the *_d2i_bio() and *_d2i_fp() functions to use these.
3356 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3357 lines, recognice more "algorithms" that can be deselected, and make
3358 it complain about algorithm deselection that isn't recognised.
3361 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3362 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3363 to use new functions. Add NO_ASN1_OLD which can be set to remove
3364 some old style ASN1 functions: this can be used to determine if old
3365 code will still work when these eventually go away.
3368 *) New extension functions for OCSP structures, these follow the
3369 same conventions as certificates and CRLs.
3372 *) New function X509V3_add1_i2d(). This automatically encodes and
3373 adds an extension. Its behaviour can be customised with various
3374 flags to append, replace or delete. Various wrappers added for
3375 certifcates and CRLs.
3378 *) Fix to avoid calling the underlying ASN1 print routine when
3379 an extension cannot be parsed. Correct a typo in the
3380 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3383 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3384 entries for variables.
3387 *) Add functionality to apps/openssl.c for detecting locking
3388 problems: As the program is single-threaded, all we have
3389 to do is register a locking callback using an array for
3390 storing which locks are currently held by the program.
3393 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3394 SSL_get_ex_data_X509_STORE_idx(), which is used in
3395 ssl_verify_cert_chain() and thus can be called at any time
3396 during TLS/SSL handshakes so that thread-safety is essential.
3397 Unfortunately, the ex_data design is not at all suited
3398 for multi-threaded use, so it probably should be abolished.
3401 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3402 [Broadcom, tweaked and integrated by Geoff Thorpe]
3404 *) Move common extension printing code to new function
3405 X509V3_print_extensions(). Reorganise OCSP print routines and
3406 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3409 *) New function X509_signature_print() to remove duplication in some
3413 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3414 set (this was treated exactly the same as SET OF previously). This
3415 is used to reorder the STACK representing the structure to match the
3416 encoding. This will be used to get round a problem where a PKCS7
3417 structure which was signed could not be verified because the STACK
3418 order did not reflect the encoded order.
3421 *) Reimplement the OCSP ASN1 module using the new code.
3424 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3425 for its ASN1 operations. The old style function pointers still exist
3426 for now but they will eventually go away.
3429 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3430 completely replaces the old ASN1 functionality with a table driven
3431 encoder and decoder which interprets an ASN1_ITEM structure describing
3432 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3433 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3434 has also been converted to the new form.
3437 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3438 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3439 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3440 for negative moduli.
3443 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3444 of not touching the result's sign bit.
3447 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3451 *) Changed the LHASH code to use prototypes for callbacks, and created
3452 macros to declare and implement thin (optionally static) functions
3453 that provide type-safety and avoid function pointer casting for the
3454 type-specific callbacks.
3457 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3459 [Veers Staats <staatsvr@asc.hpc.mil>,
3460 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3462 *) Reformat the FAQ so the different questions and answers can be divided
3463 in sections depending on the subject.
3466 *) Have the zlib compression code load ZLIB.DLL dynamically under
3470 *) New function BN_mod_sqrt for computing square roots modulo a prime
3471 (using the probabilistic Tonelli-Shanks algorithm unless
3472 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3473 be handled deterministically).
3474 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3476 *) Make BN_mod_inverse faster by explicitly handling small quotients
3477 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3478 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3481 *) New function BN_kronecker.
3484 *) Fix BN_gcd so that it works on negative inputs; the result is
3485 positive unless both parameters are zero.
3486 Previously something reasonably close to an infinite loop was
3487 possible because numbers could be growing instead of shrinking
3488 in the implementation of Euclid's algorithm.
3491 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3492 sign of the number in question.
3494 Fix BN_is_word(a,w) to work correctly for w == 0.
3496 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3497 because its test if the absolute value of 'a' equals 'w'.
3498 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3499 it exists mostly for use in the implementations of BN_is_zero(),
3500 BN_is_one(), and BN_is_word().
3503 *) New function BN_swap.
3506 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3507 the exponentiation functions are more likely to produce reasonable
3508 results on negative inputs.
3511 *) Change BN_mod_mul so that the result is always non-negative.
3512 Previously, it could be negative if one of the factors was negative;
3513 I don't think anyone really wanted that behaviour.
3516 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3517 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3518 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3519 and add new functions:
3528 BN_mod_lshift1_quick
3532 These functions always generate non-negative results.
3534 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3535 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3537 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3538 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3539 be reduced modulo m.
3540 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3543 The following entry accidentily appeared in the CHANGES file
3544 distributed with OpenSSL 0.9.7. The modifications described in
3545 it do *not* apply to OpenSSL 0.9.7.
3547 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3548 was actually never needed) and in BN_mul(). The removal in BN_mul()
3549 required a small change in bn_mul_part_recursive() and the addition
3550 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3551 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3552 bn_sub_words() and bn_add_words() except they take arrays with
3557 *) In 'openssl passwd', verify passwords read from the terminal
3558 unless the '-salt' option is used (which usually means that
3559 verification would just waste user's time since the resulting
3560 hash is going to be compared with some given password hash)
3561 or the new '-noverify' option is used.
3563 This is an incompatible change, but it does not affect
3564 non-interactive use of 'openssl passwd' (passwords on the command
3565 line, '-stdin' option, '-in ...' option) and thus should not
3569 *) Remove all references to RSAref, since there's no more need for it.
3572 *) Make DSO load along a path given through an environment variable
3573 (SHLIB_PATH) with shl_load().
3576 *) Constify the ENGINE code as a result of BIGNUM constification.
3577 Also constify the RSA code and most things related to it. In a
3578 few places, most notable in the depth of the ASN.1 code, ugly
3579 casts back to non-const were required (to be solved at a later
3583 *) Make it so the openssl application has all engines loaded by default.
3586 *) Constify the BIGNUM routines a little more.
3589 *) Add the following functions:
3591 ENGINE_load_cswift()
3593 ENGINE_load_atalla()
3595 ENGINE_load_builtin_engines()
3597 That way, an application can itself choose if external engines that
3598 are built-in in OpenSSL shall ever be used or not. The benefit is
3599 that applications won't have to be linked with libdl or other dso
3600 libraries unless it's really needed.
3602 Changed 'openssl engine' to load all engines on demand.
3603 Changed the engine header files to avoid the duplication of some
3604 declarations (they differed!).
3607 *) 'openssl engine' can now list capabilities.
3610 *) Better error reporting in 'openssl engine'.
3613 *) Never call load_dh_param(NULL) in s_server.
3616 *) Add engine application. It can currently list engines by name and
3617 identity, and test if they are actually available.
3620 *) Improve RPM specification file by forcing symbolic linking and making
3621 sure the installed documentation is also owned by root.root.
3622 [Damien Miller <djm@mindrot.org>]
3624 *) Give the OpenSSL applications more possibilities to make use of
3625 keys (public as well as private) handled by engines.
3628 *) Add OCSP code that comes from CertCo.
3631 *) Add VMS support for the Rijndael code.
3634 *) Added untested support for Nuron crypto accelerator.
3637 *) Add support for external cryptographic devices. This code was
3638 previously distributed separately as the "engine" branch.
3639 [Geoff Thorpe, Richard Levitte]
3641 *) Rework the filename-translation in the DSO code. It is now possible to
3642 have far greater control over how a "name" is turned into a filename
3643 depending on the operating environment and any oddities about the
3644 different shared library filenames on each system.
3647 *) Support threads on FreeBSD-elf in Configure.
3650 *) Fix for SHA1 assembly problem with MASM: it produces
3651 warnings about corrupt line number information when assembling
3652 with debugging information. This is caused by the overlapping
3654 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3657 NCONF_get_number() has no error checking at all. As a replacement,
3658 NCONF_get_number_e() is defined (_e for "error checking") and is
3659 promoted strongly. The old NCONF_get_number is kept around for
3660 binary backward compatibility.
3661 Make it possible for methods to load from something other than a BIO,
3662 by providing a function pointer that is given a name instead of a BIO.
3663 For example, this could be used to load configuration data from an
3667 *) Fix for non blocking accept BIOs. Added new I/O special reason
3668 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3669 with non blocking I/O was not possible because no retry code was
3670 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3674 *) Added the beginnings of Rijndael support.
3677 *) Fix for bug in DirectoryString mask setting. Add support for
3678 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3679 to allow certificate printing to more controllable, additional
3680 'certopt' option to 'x509' to allow new printing options to be
3684 *) Clean old EAY MD5 hack from e_os.h.
3687 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3689 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3690 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3691 [Joe Orton, Steve Henson]
3693 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3695 *) Fix additional bug revealed by the NISCC test suite:
3697 Stop bug triggering large recursion when presented with
3698 certain ASN.1 tags (CVE-2003-0851)
3701 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3703 *) Fix various bugs revealed by running the NISCC test suite:
3705 Stop out of bounds reads in the ASN1 code when presented with
3706 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3708 If verify callback ignores invalid public key errors don't try to check
3709 certificate signature with the NULL public key.
3713 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3714 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3718 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3719 extra data after the compression methods not only for TLS 1.0
3720 but also for SSL 3.0 (as required by the specification).
3721 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3723 *) Change X509_certificate_type() to mark the key as exported/exportable
3724 when it's 512 *bits* long, not 512 bytes.
3727 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3729 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3730 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3731 a protocol version number mismatch like a decryption error
3732 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3735 *) Turn on RSA blinding by default in the default implementation
3736 to avoid a timing attack. Applications that don't want it can call
3737 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3738 They would be ill-advised to do so in most cases.
3739 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3741 *) Change RSA blinding code so that it works when the PRNG is not
3742 seeded (in this case, the secret RSA exponent is abused as
3743 an unpredictable seed -- if it is not unpredictable, there
3744 is no point in blinding anyway). Make RSA blinding thread-safe
3745 by remembering the creator's thread ID in rsa->blinding and
3746 having all other threads use local one-time blinding factors
3747 (this requires more computation than sharing rsa->blinding, but
3748 avoids excessive locking; and if an RSA object is not shared
3749 between threads, blinding will still be very fast).
3752 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3754 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3755 via timing by performing a MAC computation even if incorrrect
3756 block cipher padding has been found. This is a countermeasure
3757 against active attacks where the attacker has to distinguish
3758 between bad padding and a MAC verification error. (CVE-2003-0078)
3760 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3761 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3762 Martin Vuagnoux (EPFL, Ilion)]
3764 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3766 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3767 memory from it's contents. This is done with a counter that will
3768 place alternating values in each byte. This can be used to solve
3769 two issues: 1) the removal of calls to memset() by highly optimizing
3770 compilers, and 2) cleansing with other values than 0, since those can
3771 be read through on certain media, for example a swap space on disk.
3774 *) Bugfix: client side session caching did not work with external caching,
3775 because the session->cipher setting was not restored when reloading
3776 from the external cache. This problem was masked, when
3777 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3778 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3781 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3782 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3783 [Zeev Lieber <zeev-l@yahoo.com>]
3785 *) Undo an undocumented change introduced in 0.9.6e which caused
3786 repeated calls to OpenSSL_add_all_ciphers() and
3787 OpenSSL_add_all_digests() to be ignored, even after calling
3791 *) Change the default configuration reader to deal with last line not
3792 being properly terminated.
3795 *) Change X509_NAME_cmp() so it applies the special rules on handling
3796 DN values that are of type PrintableString, as well as RDNs of type
3797 emailAddress where the value has the type ia5String.
3798 [stefank@valicert.com via Richard Levitte]
3800 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3801 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3802 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3803 the bitwise-OR of the two for use by the majority of applications
3804 wanting this behaviour, and update the docs. The documented
3805 behaviour and actual behaviour were inconsistent and had been
3806 changing anyway, so this is more a bug-fix than a behavioural
3808 [Geoff Thorpe, diagnosed by Nadav Har'El]
3810 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3811 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3814 *) Fix initialization code race conditions in
3815 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3816 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3817 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3818 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3819 ssl2_get_cipher_by_char(),
3820 ssl3_get_cipher_by_char().
3821 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3823 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3824 the cached sessions are flushed, as the remove_cb() might use ex_data
3825 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3826 (see [openssl.org #212]).
3827 [Geoff Thorpe, Lutz Jaenicke]
3829 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3830 length, instead of the encoding length to d2i_ASN1_OBJECT.
3833 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3835 *) [In 0.9.6g-engine release:]
3836 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3837 [Lynn Gazis <lgazis@rainbow.com>]
3839 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3841 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3842 and get fix the header length calculation.
3843 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3844 Alon Kantor <alonk@checkpoint.com> (and others),
3847 *) Use proper error handling instead of 'assertions' in buffer
3848 overflow checks added in 0.9.6e. This prevents DoS (the
3849 assertions could call abort()).
3850 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3852 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3854 *) Add various sanity checks to asn1_get_length() to reject
3855 the ASN1 length bytes if they exceed sizeof(long), will appear
3856 negative or the content length exceeds the length of the
3858 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3860 *) Fix cipher selection routines: ciphers without encryption had no flags
3861 for the cipher strength set and where therefore not handled correctly
3862 by the selection routines (PR #130).
3865 *) Fix EVP_dsa_sha macro.
3869 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3870 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3871 that was added in OpenSSL 0.9.6d.
3873 As the countermeasure turned out to be incompatible with some
3874 broken SSL implementations, the new option is part of SSL_OP_ALL.
3875 SSL_OP_ALL is usually employed when compatibility with weird SSL
3876 implementations is desired (e.g. '-bugs' option to 's_client' and
3877 's_server'), so the new option is automatically set in many
3881 *) Changes in security patch:
3883 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3884 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3885 Air Force Materiel Command, USAF, under agreement number
3888 *) Add various sanity checks to asn1_get_length() to reject
3889 the ASN1 length bytes if they exceed sizeof(long), will appear
3890 negative or the content length exceeds the length of the
3891 supplied buffer. (CVE-2002-0659)
3892 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3894 *) Assertions for various potential buffer overflows, not known to
3896 [Ben Laurie (CHATS)]
3898 *) Various temporary buffers to hold ASCII versions of integers were
3899 too small for 64 bit platforms. (CVE-2002-0655)
3900 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3902 *) Remote buffer overflow in SSL3 protocol - an attacker could
3903 supply an oversized session ID to a client. (CVE-2002-0656)
3904 [Ben Laurie (CHATS)]
3906 *) Remote buffer overflow in SSL2 protocol - an attacker could
3907 supply an oversized client master key. (CVE-2002-0656)
3908 [Ben Laurie (CHATS)]
3910 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3912 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3913 encoded as NULL) with id-dsa-with-sha1.
3914 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3916 *) Check various X509_...() return values in apps/req.c.
3917 [Nils Larsch <nla@trustcenter.de>]
3919 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3920 an end-of-file condition would erronously be flagged, when the CRLF
3921 was just at the end of a processed block. The bug was discovered when
3922 processing data through a buffering memory BIO handing the data to a
3923 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3924 <ptsekov@syntrex.com> and Nedelcho Stanev.
3927 *) Implement a countermeasure against a vulnerability recently found
3928 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3929 before application data chunks to avoid the use of known IVs
3930 with data potentially chosen by the attacker.
3933 *) Fix length checks in ssl3_get_client_hello().
3936 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3937 to prevent ssl3_read_internal() from incorrectly assuming that
3938 ssl3_read_bytes() found application data while handshake
3939 processing was enabled when in fact s->s3->in_read_app_data was
3940 merely automatically cleared during the initial handshake.
3941 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3943 *) Fix object definitions for Private and Enterprise: they were not
3944 recognized in their shortname (=lowercase) representation. Extend
3945 obj_dat.pl to issue an error when using undefined keywords instead
3946 of silently ignoring the problem (Svenning Sorensen
3947 <sss@sss.dnsalias.net>).
3950 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3951 generators, i.e. generators other than 2 and 5. (Previously, the
3952 code did not properly initialise the 'add' and 'rem' values to
3953 BN_generate_prime().)
3955 In the new general case, we do not insist that 'generator' is
3956 actually a primitive root: This requirement is rather pointless;
3957 a generator of the order-q subgroup is just as good, if not
3961 *) Map new X509 verification errors to alerts. Discovered and submitted by
3962 Tom Wu <tom@arcot.com>.
3965 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3966 returning non-zero before the data has been completely received
3967 when using non-blocking I/O.
3968 [Bodo Moeller; problem pointed out by John Hughes]
3970 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3971 [Ben Laurie, Lutz Jaenicke]
3973 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3974 Yoram Zahavi <YoramZ@gilian.com>).
3977 *) Add information about CygWin 1.3 and on, and preserve proper
3978 configuration for the versions before that.
3979 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3981 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3982 check whether we deal with a copy of a session and do not delete from
3983 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3984 <izhar@checkpoint.com>.
3987 *) Do not store session data into the internal session cache, if it
3988 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3989 flag is set). Proposed by Aslam <aslam@funk.com>.
3992 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3996 *) [In 0.9.6d-engine release:]
3997 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3998 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4000 *) Add the configuration target linux-s390x.
4001 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4003 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4004 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4005 variable as an indication that a ClientHello message has been
4006 received. As the flag value will be lost between multiple
4007 invocations of ssl3_accept when using non-blocking I/O, the
4008 function may not be aware that a handshake has actually taken
4009 place, thus preventing a new session from being added to the
4012 To avoid this problem, we now set s->new_session to 2 instead of
4013 using a local variable.
4014 [Lutz Jaenicke, Bodo Moeller]
4016 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4017 if the SSL_R_LENGTH_MISMATCH error is detected.
4018 [Geoff Thorpe, Bodo Moeller]
4020 *) New 'shared_ldflag' column in Configure platform table.
4023 *) Fix EVP_CIPHER_mode macro.
4024 ["Dan S. Camper" <dan@bti.net>]
4026 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4027 type, we must throw them away by setting rr->length to 0.
4028 [D P Chang <dpc@qualys.com>]
4030 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4032 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4033 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4034 worked incorrectly for those cases where range = 10..._2 and
4035 3*range is two bits longer than range.)
4038 *) Only add signing time to PKCS7 structures if it is not already
4042 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4043 OBJ_ld_ce should be OBJ_id_ce.
4044 Also some ip-pda OIDs in crypto/objects/objects.txt were
4045 incorrect (cf. RFC 3039).
4046 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4048 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4049 returns early because it has nothing to do.
4050 [Andy Schneider <andy.schneider@bjss.co.uk>]
4052 *) [In 0.9.6c-engine release:]
4053 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4054 [Andy Schneider <andy.schneider@bjss.co.uk>]
4056 *) [In 0.9.6c-engine release:]
4057 Add support for Cryptographic Appliance's keyserver technology.
4058 (Use engine 'keyclient')
4059 [Cryptographic Appliances and Geoff Thorpe]
4061 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4062 is called via tools/c89.sh because arguments have to be
4063 rearranged (all '-L' options must appear before the first object
4065 [Richard Shapiro <rshapiro@abinitio.com>]
4067 *) [In 0.9.6c-engine release:]
4068 Add support for Broadcom crypto accelerator cards, backported
4070 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4072 *) [In 0.9.6c-engine release:]
4073 Add support for SureWare crypto accelerator cards from
4074 Baltimore Technologies. (Use engine 'sureware')
4075 [Baltimore Technologies and Mark Cox]
4077 *) [In 0.9.6c-engine release:]
4078 Add support for crypto accelerator cards from Accelerated
4079 Encryption Processing, www.aep.ie. (Use engine 'aep')
4080 [AEP Inc. and Mark Cox]
4082 *) Add a configuration entry for gcc on UnixWare.
4083 [Gary Benson <gbenson@redhat.com>]
4085 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4086 messages are stored in a single piece (fixed-length part and
4087 variable-length part combined) and fix various bugs found on the way.
4090 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4091 instead. BIO_gethostbyname() does not know what timeouts are
4092 appropriate, so entries would stay in cache even when they have
4094 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4096 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4097 faced with a pathologically small ClientHello fragment that does
4098 not contain client_version: Instead of aborting with an error,
4099 simply choose the highest available protocol version (i.e.,
4100 TLS 1.0 unless it is disabled). In practice, ClientHello
4101 messages are never sent like this, but this change gives us
4102 strictly correct behaviour at least for TLS.
4105 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4106 never resets s->method to s->ctx->method when called from within
4107 one of the SSL handshake functions.
4108 [Bodo Moeller; problem pointed out by Niko Baric]
4110 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4111 (sent using the client's version number) if client_version is
4112 smaller than the protocol version in use. Also change
4113 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4114 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4115 the client will at least see that alert.
4118 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4122 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4123 client receives HelloRequest while in a handshake.
4124 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4126 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4127 should end in 'break', not 'goto end' which circuments various
4128 cleanups done in state SSL_ST_OK. But session related stuff
4129 must be disabled for SSL_ST_OK in the case that we just sent a
4132 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4133 before just sending a HelloRequest.
4134 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4136 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4137 reveal whether illegal block cipher padding was found or a MAC
4138 verification error occured. (Neither SSLerr() codes nor alerts
4139 are directly visible to potential attackers, but the information
4140 may leak via logfiles.)
4142 Similar changes are not required for the SSL 2.0 implementation
4143 because the number of padding bytes is sent in clear for SSL 2.0,
4144 and the extra bytes are just ignored. However ssl/s2_pkt.c
4145 failed to verify that the purported number of padding bytes is in
4149 *) Add OpenUNIX-8 support including shared libraries
4150 (Boyd Lynn Gerber <gerberb@zenez.com>).
4153 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4154 'wristwatch attack' using huge encoding parameters (cf.
4155 James H. Manger's CRYPTO 2001 paper). Note that the
4156 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4157 encoding parameters and hence was not vulnerable.
4160 *) BN_sqr() bug fix.
4161 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4163 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4164 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4165 followed by modular reduction.
4166 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4168 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4169 equivalent based on BN_pseudo_rand() instead of BN_rand().
4172 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4173 This function was broken, as the check for a new client hello message
4174 to handle SGC did not allow these large messages.
4175 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4178 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4181 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4182 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4185 *) Rework the configuration and shared library support for Tru64 Unix.
4186 The configuration part makes use of modern compiler features and
4187 still retains old compiler behavior for those that run older versions
4188 of the OS. The shared library support part includes a variant that
4189 uses the RPATH feature, and is available through the special
4190 configuration target "alpha-cc-rpath", which will never be selected
4192 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4194 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4195 with the same message size as in ssl3_get_certificate_request().
4196 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4197 messages might inadvertently be reject as too long.
4198 [Petr Lampa <lampa@fee.vutbr.cz>]
4200 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4203 *) Modified SSL library such that the verify_callback that has been set
4204 specificly for an SSL object with SSL_set_verify() is actually being
4205 used. Before the change, a verify_callback set with this function was
4206 ignored and the verify_callback() set in the SSL_CTX at the time of
4207 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4208 to allow the necessary settings.
4211 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4212 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4213 done automatically (in contradiction to the requirements of the C
4214 standard). This made problems when used from OpenSSH.
4217 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4218 dh->length and always used
4220 BN_rand_range(priv_key, dh->p).
4222 BN_rand_range() is not necessary for Diffie-Hellman, and this
4223 specific range makes Diffie-Hellman unnecessarily inefficient if
4224 dh->length (recommended exponent length) is much smaller than the
4225 length of dh->p. We could use BN_rand_range() if the order of
4226 the subgroup was stored in the DH structure, but we only have
4231 BN_rand(priv_key, l, ...)
4233 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4239 RSA_eay_public_encrypt
4240 RSA_eay_private_decrypt
4241 RSA_eay_private_encrypt (signing)
4242 RSA_eay_public_decrypt (signature verification)
4244 (default implementations for RSA_public_encrypt,
4245 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4246 always reject numbers >= n.
4249 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4250 to synchronize access to 'locking_thread'. This is necessary on
4251 systems where access to 'locking_thread' (an 'unsigned long'
4252 variable) is not atomic.
4255 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4256 *before* setting the 'crypto_lock_rand' flag. The previous code had
4257 a race condition if 0 is a valid thread ID.
4258 [Travis Vitek <vitek@roguewave.com>]
4260 *) Add support for shared libraries under Irix.
4261 [Albert Chin-A-Young <china@thewrittenword.com>]
4263 *) Add configuration option to build on Linux on both big-endian and
4265 [Ralf Baechle <ralf@uni-koblenz.de>]
4267 *) Add the possibility to create shared libraries on HP-UX.
4270 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4272 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4273 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4274 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4275 PRNG state recovery was possible based on the output of
4276 one PRNG request appropriately sized to gain knowledge on
4277 'md' followed by enough consecutive 1-byte PRNG requests
4278 to traverse all of 'state'.
4280 1. When updating 'md_local' (the current thread's copy of 'md')
4281 during PRNG output generation, hash all of the previous
4282 'md_local' value, not just the half used for PRNG output.
4284 2. Make the number of bytes from 'state' included into the hash
4285 independent from the number of PRNG bytes requested.
4287 The first measure alone would be sufficient to avoid
4288 Markku-Juhani's attack. (Actually it had never occurred
4289 to me that the half of 'md_local' used for chaining was the
4290 half from which PRNG output bytes were taken -- I had always
4291 assumed that the secret half would be used.) The second
4292 measure makes sure that additional data from 'state' is never
4293 mixed into 'md_local' in small portions; this heuristically
4294 further strengthens the PRNG.
4297 *) Fix crypto/bn/asm/mips3.s.
4300 *) When only the key is given to "enc", the IV is undefined. Print out
4301 an error message in this case.
4304 *) Handle special case when X509_NAME is empty in X509 printing routines.
4307 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4308 positive and less than q.
4311 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4312 used: it isn't thread safe and the add_lock_callback should handle
4314 [Paul Rose <Paul.Rose@bridge.com>]
4316 *) Verify that incoming data obeys the block size in
4317 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4321 [Ulf Möller, Bodo Möller]
4323 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4324 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4325 when fixing the server behaviour for backwards-compatible 'client
4326 hello' messages. (Note that the attack is impractical against
4327 SSL 3.0 and TLS 1.0 anyway because length and version checking
4328 means that the probability of guessing a valid ciphertext is
4329 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4332 Before 0.9.5, the countermeasure (hide the error by generating a
4333 random 'decryption result') did not work properly because
4334 ERR_clear_error() was missing, meaning that SSL_get_error() would
4335 detect the supposedly ignored error.
4337 Both problems are now fixed.
4340 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4341 (previously it was 1024).
4344 *) Fix for compatibility mode trust settings: ignore trust settings
4345 unless some valid trust or reject settings are present.
4348 *) Fix for blowfish EVP: its a variable length cipher.
4351 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4352 parameters in DSA public key structures and return an error in the
4353 DSA routines if parameters are absent.
4356 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4357 in the current directory if neither $RANDFILE nor $HOME was set.
4358 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4359 caused some confusion to Windows users who haven't defined $HOME.
4360 Thus RAND_file_name() is changed again: e_os.h can define a
4361 DEFAULT_HOME, which will be used if $HOME is not set.
4362 For Windows, we use "C:"; on other platforms, we still require
4363 environment variables.
4365 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4366 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4367 having multiple threads call RAND_poll() concurrently.
4370 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4371 combination of a flag and a thread ID variable.
4372 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4373 flag), *other* threads can enter ssleay_add_bytes without obeying
4374 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4375 that they do not hold after the first thread unsets add_do_not_lock).
4378 *) Change bctest again: '-x' expressions are not available in all
4382 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4384 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4385 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4387 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4388 the default extension for executables, if any. Also, make the perl
4389 scripts that use symlink() to test if it really exists and use "cp"
4390 if it doesn't. All this made OpenSSL compilable and installable in
4394 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4395 If SEQUENCE is length is indefinite just set c->slen to the total
4396 amount of data available.
4397 [Steve Henson, reported by shige@FreeBSD.org]
4398 [This change does not apply to 0.9.7.]
4400 *) Change bctest to avoid here-documents inside command substitution
4401 (workaround for FreeBSD /bin/sh bug).
4402 For compatibility with Ultrix, avoid shell functions (introduced
4403 in the bctest version that searches along $PATH).
4406 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4407 with des_encrypt() defined on some operating systems, like Solaris
4411 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4412 On the Importance of Eliminating Errors in Cryptographic
4413 Computations, J. Cryptology 14 (2001) 2, 101-119,
4414 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4417 *) MIPS assembler BIGNUM division bug fix.
4420 *) Disabled incorrect Alpha assembler code.
4423 *) Fix PKCS#7 decode routines so they correctly update the length
4424 after reading an EOC for the EXPLICIT tag.
4426 [This change does not apply to 0.9.7.]
4428 *) Fix bug in PKCS#12 key generation routines. This was triggered
4429 if a 3DES key was generated with a 0 initial byte. Include
4430 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4431 (but broken) behaviour.
4434 *) Enhance bctest to search for a working bc along $PATH and print
4436 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4438 *) Fix memory leaks in err.c: free err_data string if necessary;
4439 don't write to the wrong index in ERR_set_error_data.
4442 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4446 *) Replace rdtsc with _emit statements for VC++ version 5.
4447 [Jeremy Cooper <jeremy@baymoo.org>]
4449 *) Make it possible to reuse SSLv2 sessions.
4452 *) In copy_email() check for >= 0 as a return value for
4453 X509_NAME_get_index_by_NID() since 0 is a valid index.
4454 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4456 *) Avoid coredump with unsupported or invalid public keys by checking if
4457 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4458 PKCS7_verify() fails with non detached data.
4461 *) Don't use getenv in library functions when run as setuid/setgid.
4462 New function OPENSSL_issetugid().
4465 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4466 due to incorrect handling of multi-threading:
4468 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4470 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4472 3. Count how many times MemCheck_off() has been called so that
4473 nested use can be treated correctly. This also avoids
4474 inband-signalling in the previous code (which relied on the
4475 assumption that thread ID 0 is impossible).
4478 *) Add "-rand" option also to s_client and s_server.
4481 *) Fix CPU detection on Irix 6.x.
4482 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4483 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4485 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4488 [This change does not apply to 0.9.7.]
4490 *) Use the cached encoding of an X509_NAME structure rather than
4491 copying it. This is apparently the reason for the libsafe "errors"
4492 but the code is actually correct.
4495 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4496 Bleichenbacher's DSA attack.
4497 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4498 to be set and top=0 forces the highest bit to be set; top=-1 is new
4499 and leaves the highest bit random.
4500 [Ulf Moeller, Bodo Moeller]
4502 *) In the NCONF_...-based implementations for CONF_... queries
4503 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4504 a temporary CONF structure with the data component set to NULL
4505 (which gives segmentation faults in lh_retrieve).
4506 Instead, use NULL for the CONF pointer in CONF_get_string and
4507 CONF_get_number (which may use environment variables) and directly
4508 return NULL from CONF_get_section.
4511 *) Fix potential buffer overrun for EBCDIC.
4514 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4515 keyUsage if basicConstraints absent for a CA.
4518 *) Make SMIME_write_PKCS7() write mail header values with a format that
4519 is more generally accepted (no spaces before the semicolon), since
4520 some programs can't parse those values properly otherwise. Also make
4521 sure BIO's that break lines after each write do not create invalid
4525 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4526 macros previously used would not encode an empty SEQUENCE OF
4527 and break the signature.
4529 [This change does not apply to 0.9.7.]
4531 *) Zero the premaster secret after deriving the master secret in
4535 *) Add some EVP_add_digest_alias registrations (as found in
4536 OpenSSL_add_all_digests()) to SSL_library_init()
4537 aka OpenSSL_add_ssl_algorithms(). This provides improved
4538 compatibility with peers using X.509 certificates
4539 with unconventional AlgorithmIdentifier OIDs.
4542 *) Fix for Irix with NO_ASM.
4543 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4545 *) ./config script fixes.
4546 [Ulf Moeller, Richard Levitte]
4548 *) Fix 'openssl passwd -1'.
4551 *) Change PKCS12_key_gen_asc() so it can cope with non null
4552 terminated strings whose length is passed in the passlen
4553 parameter, for example from PEM callbacks. This was done
4554 by adding an extra length parameter to asc2uni().
4555 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4557 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4558 call failed, free the DSA structure.
4561 *) Fix to uni2asc() to cope with zero length Unicode strings.
4562 These are present in some PKCS#12 files.
4565 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4566 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4567 when writing a 32767 byte record.
4568 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4570 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4571 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4573 (RSA objects have a reference count access to which is protected
4574 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4575 so they are meant to be shared between threads.)
4576 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4577 "Reddie, Steven" <Steven.Reddie@ca.com>]
4579 *) Fix a deadlock in CRYPTO_mem_leaks().
4582 *) Use better test patterns in bntest.
4585 *) rand_win.c fix for Borland C.
4588 *) BN_rshift bugfix for n == 0.
4591 *) Add a 'bctest' script that checks for some known 'bc' bugs
4592 so that 'make test' does not abort just because 'bc' is broken.
4595 *) Store verify_result within SSL_SESSION also for client side to
4596 avoid potential security hole. (Re-used sessions on the client side
4597 always resulted in verify_result==X509_V_OK, not using the original
4598 result of the server certificate verification.)
4601 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4602 SSL3_RT_APPLICATION_DATA, return 0.
4603 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4607 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4608 releases, have been re-implemented by renaming the previous
4609 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4610 and ssl3_read_internal, respectively, and adding 'peek' parameters
4611 to them. The new ssl[23]_{read,peek} functions are calls to
4612 ssl[23]_read_internal with the 'peek' flag set appropriately.
4613 A 'peek' parameter has also been added to ssl3_read_bytes, which
4614 does the actual work for ssl3_read_internal.
4617 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4618 the method-specific "init()" handler. Also clean up ex_data after
4619 calling the method-specific "finish()" handler. Previously, this was
4620 happening the other way round.
4623 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4624 The previous value, 12, was not always sufficient for BN_mod_exp().
4627 *) Make sure that shared libraries get the internal name engine with
4628 the full version number and not just 0. This should mark the
4629 shared libraries as not backward compatible. Of course, this should
4630 be changed again when we can guarantee backward binary compatibility.
4633 *) Fix typo in get_cert_by_subject() in by_dir.c
4634 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4636 *) Rework the system to generate shared libraries:
4638 - Make note of the expected extension for the shared libraries and
4639 if there is a need for symbolic links from for example libcrypto.so.0
4640 to libcrypto.so.0.9.7. There is extended info in Configure for
4643 - Make as few rebuilds of the shared libraries as possible.
4645 - Still avoid linking the OpenSSL programs with the shared libraries.
4647 - When installing, install the shared libraries separately from the
4651 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4653 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4654 and not in SSL_clear because the latter is also used by the
4655 accept/connect functions; previously, the settings made by
4656 SSL_set_read_ahead would be lost during the handshake.
4657 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4659 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4660 Previously, it would create entries for disableed algorithms no
4664 *) Added several new manual pages for SSL_* function.
4667 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4669 *) In ssl23_get_client_hello, generate an error message when faced
4670 with an initial SSL 3.0/TLS record that is too small to contain the
4671 first two bytes of the ClientHello message, i.e. client_version.
4672 (Note that this is a pathologic case that probably has never happened
4673 in real life.) The previous approach was to use the version number
4674 from the record header as a substitute; but our protocol choice
4675 should not depend on that one because it is not authenticated
4676 by the Finished messages.
4679 *) More robust randomness gathering functions for Windows.
4680 [Jeffrey Altman <jaltman@columbia.edu>]
4682 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4683 not set then we don't setup the error code for issuer check errors
4684 to avoid possibly overwriting other errors which the callback does
4685 handle. If an application does set the flag then we assume it knows
4686 what it is doing and can handle the new informational codes
4690 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4691 a general "ANY" type, as such it should be able to decode anything
4692 including tagged types. However it didn't check the class so it would
4693 wrongly interpret tagged types in the same way as their universal
4694 counterpart and unknown types were just rejected. Changed so that the
4695 tagged and unknown types are handled in the same way as a SEQUENCE:
4696 that is the encoding is stored intact. There is also a new type
4697 "V_ASN1_OTHER" which is used when the class is not universal, in this
4698 case we have no idea what the actual type is so we just lump them all
4702 *) On VMS, stdout may very well lead to a file that is written to
4703 in a record-oriented fashion. That means that every write() will
4704 write a separate record, which will be read separately by the
4705 programs trying to read from it. This can be very confusing.
4707 The solution is to put a BIO filter in the way that will buffer
4708 text until a linefeed is reached, and then write everything a
4709 line at a time, so every record written will be an actual line,
4710 not chunks of lines and not (usually doesn't happen, but I've
4711 seen it once) several lines in one record. BIO_f_linebuffer() is
4714 Currently, it's a VMS-only method, because that's where it has
4715 been tested well enough.
4718 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4719 it can return incorrect results.
4720 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4721 but it was in 0.9.6-beta[12].)
4724 *) Disable the check for content being present when verifying detached
4725 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4726 include zero length content when signing messages.
4729 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4730 BIO_ctrl (for BIO pairs).
4733 *) Add DSO method for VMS.
4736 *) Bug fix: Montgomery multiplication could produce results with the
4740 *) Add RPM specification openssl.spec and modify it to build three
4741 packages. The default package contains applications, application
4742 documentation and run-time libraries. The devel package contains
4743 include files, static libraries and function documentation. The
4744 doc package contains the contents of the doc directory. The original
4745 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4748 *) Add a large number of documentation files for many SSL routines.
4749 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4751 *) Add a configuration entry for Sony News 4.
4752 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4754 *) Don't set the two most significant bits to one when generating a
4755 random number < q in the DSA library.
4758 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4759 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4760 the underlying transport is blocking) if a handshake took place.
4761 (The default behaviour is needed by applications such as s_client
4762 and s_server that use select() to determine when to use SSL_read;
4763 but for applications that know in advance when to expect data, it
4764 just makes things more complicated.)
4767 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4771 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4772 work better on such systems.
4773 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4775 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4776 Update PKCS12_parse() so it copies the friendlyName and the
4777 keyid to the certificates aux info.
4780 *) Fix bug in PKCS7_verify() which caused an infinite loop
4781 if there was more than one signature.
4782 [Sven Uszpelkat <su@celocom.de>]
4784 *) Major change in util/mkdef.pl to include extra information
4785 about each symbol, as well as presentig variables as well
4786 as functions. This change means that there's n more need
4787 to rebuild the .num files when some algorithms are excluded.
4790 *) Allow the verify time to be set by an application,
4791 rather than always using the current time.
4794 *) Phase 2 verify code reorganisation. The certificate
4795 verify code now looks up an issuer certificate by a
4796 number of criteria: subject name, authority key id
4797 and key usage. It also verifies self signed certificates
4798 by the same criteria. The main comparison function is
4799 X509_check_issued() which performs these checks.
4801 Lot of changes were necessary in order to support this
4802 without completely rewriting the lookup code.
4804 Authority and subject key identifier are now cached.
4806 The LHASH 'certs' is X509_STORE has now been replaced
4807 by a STACK_OF(X509_OBJECT). This is mainly because an
4808 LHASH can't store or retrieve multiple objects with
4809 the same hash value.
4811 As a result various functions (which were all internal
4812 use only) have changed to handle the new X509_STORE
4813 structure. This will break anything that messed round
4814 with X509_STORE internally.
4816 The functions X509_STORE_add_cert() now checks for an
4817 exact match, rather than just subject name.
4819 The X509_STORE API doesn't directly support the retrieval
4820 of multiple certificates matching a given criteria, however
4821 this can be worked round by performing a lookup first
4822 (which will fill the cache with candidate certificates)
4823 and then examining the cache for matches. This is probably
4824 the best we can do without throwing out X509_LOOKUP
4825 entirely (maybe later...).
4827 The X509_VERIFY_CTX structure has been enhanced considerably.
4829 All certificate lookup operations now go via a get_issuer()
4830 callback. Although this currently uses an X509_STORE it
4831 can be replaced by custom lookups. This is a simple way
4832 to bypass the X509_STORE hackery necessary to make this
4833 work and makes it possible to use more efficient techniques
4834 in future. A very simple version which uses a simple
4835 STACK for its trusted certificate store is also provided
4836 using X509_STORE_CTX_trusted_stack().
4838 The verify_cb() and verify() callbacks now have equivalents
4839 in the X509_STORE_CTX structure.
4841 X509_STORE_CTX also has a 'flags' field which can be used
4842 to customise the verify behaviour.
4845 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4846 excludes S/MIME capabilities.
4849 *) When a certificate request is read in keep a copy of the
4850 original encoding of the signed data and use it when outputing
4851 again. Signatures then use the original encoding rather than
4852 a decoded, encoded version which may cause problems if the
4853 request is improperly encoded.
4856 *) For consistency with other BIO_puts implementations, call
4857 buffer_write(b, ...) directly in buffer_puts instead of calling
4860 In BIO_puts, increment b->num_write as in BIO_write.
4861 [Peter.Sylvester@EdelWeb.fr]
4863 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4864 BN_zero, we may not return a BIGNUM with an array consisting of
4868 *) Avoid calling abort() from within the library when problems are
4869 detected, except if preprocessor symbols have been defined
4870 (such as REF_CHECK, BN_DEBUG etc.).
4873 *) New openssl application 'rsautl'. This utility can be
4874 used for low level RSA operations. DER public key
4875 BIO/fp routines also added.
4878 *) New Configure entry and patches for compiling on QNX 4.
4879 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4881 *) A demo state-machine implementation was sponsored by
4882 Nuron (http://www.nuron.com/) and is now available in
4883 demos/state_machine.
4886 *) New options added to the 'dgst' utility for signature
4887 generation and verification.
4890 *) Unrecognized PKCS#7 content types are now handled via a
4891 catch all ASN1_TYPE structure. This allows unsupported
4892 types to be stored as a "blob" and an application can
4893 encode and decode it manually.
4896 *) Fix various signed/unsigned issues to make a_strex.c
4898 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4900 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4901 length if passed a buffer. ASN1_INTEGER_to_BN failed
4902 if passed a NULL BN and its argument was negative.
4903 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4905 *) Modification to PKCS#7 encoding routines to output definite
4906 length encoding. Since currently the whole structures are in
4907 memory there's not real point in using indefinite length
4908 constructed encoding. However if OpenSSL is compiled with
4909 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4912 *) Added BIO_vprintf() and BIO_vsnprintf().
4915 *) Added more prefixes to parse for in the the strings written
4916 through a logging bio, to cover all the levels that are available
4917 through syslog. The prefixes are now:
4919 PANIC, EMERG, EMR => LOG_EMERG
4920 ALERT, ALR => LOG_ALERT
4921 CRIT, CRI => LOG_CRIT
4922 ERROR, ERR => LOG_ERR
4923 WARNING, WARN, WAR => LOG_WARNING
4924 NOTICE, NOTE, NOT => LOG_NOTICE
4925 INFO, INF => LOG_INFO
4926 DEBUG, DBG => LOG_DEBUG
4928 and as before, if none of those prefixes are present at the
4929 beginning of the string, LOG_ERR is chosen.
4931 On Win32, the LOG_* levels are mapped according to this:
4933 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4934 LOG_WARNING => EVENTLOG_WARNING_TYPE
4935 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4939 *) Made it possible to reconfigure with just the configuration
4940 argument "reconf" or "reconfigure". The command line arguments
4941 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4942 and are retrieved from there when reconfiguring.
4946 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4948 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4951 *) The obj_dat.pl script was messing up the sorting of object
4952 names. The reason was that it compared the quoted version
4953 of strings as a result "OCSP" > "OCSP Signing" because
4954 " > SPACE. Changed script to store unquoted versions of
4955 names and add quotes on output. It was also omitting some
4956 names from the lookup table if they were given a default
4957 value (that is if SN is missing it is given the same
4958 value as LN and vice versa), these are now added on the
4959 grounds that if an object has a name we should be able to
4960 look it up. Finally added warning output when duplicate
4961 short or long names are found.
4964 *) Changes needed for Tandem NSK.
4965 [Scott Uroff <scott@xypro.com>]
4967 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4968 RSA_padding_check_SSLv23(), special padding was never detected
4969 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4970 version rollback attacks was not effective.
4972 In s23_clnt.c, don't use special rollback-attack detection padding
4973 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4974 client; similarly, in s23_srvr.c, don't do the rollback check if
4975 SSL 2.0 is the only protocol enabled in the server.
4978 *) Make it possible to get hexdumps of unprintable data with 'openssl
4979 asn1parse'. By implication, the functions ASN1_parse_dump() and
4980 BIO_dump_indent() are added.
4983 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4984 these print out strings and name structures based on various
4985 flags including RFC2253 support and proper handling of
4986 multibyte characters. Added options to the 'x509' utility
4987 to allow the various flags to be set.
4990 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4991 Also change the functions X509_cmp_current_time() and
4992 X509_gmtime_adj() work with an ASN1_TIME structure,
4993 this will enable certificates using GeneralizedTime in validity
4994 dates to be checked.
4997 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4998 negative public key encodings) on by default,
4999 NO_NEG_PUBKEY_BUG can be set to disable it.
5002 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5003 content octets. An i2c_ASN1_OBJECT is unnecessary because
5004 the encoding can be trivially obtained from the structure.
5007 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5008 not read locks (CRYPTO_r_[un]lock).
5011 *) A first attempt at creating official support for shared
5012 libraries through configuration. I've kept it so the
5013 default is static libraries only, and the OpenSSL programs
5014 are always statically linked for now, but there are
5015 preparations for dynamic linking in place.
5016 This has been tested on Linux and Tru64.
5019 *) Randomness polling function for Win9x, as described in:
5020 Peter Gutmann, Software Generation of Practically Strong
5024 *) Fix so PRNG is seeded in req if using an already existing
5028 *) New options to smime application. -inform and -outform
5029 allow alternative formats for the S/MIME message including
5030 PEM and DER. The -content option allows the content to be
5031 specified separately. This should allow things like Netscape
5032 form signing output easier to verify.
5035 *) Fix the ASN1 encoding of tags using the 'long form'.
5038 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5039 STRING types. These convert content octets to and from the
5040 underlying type. The actual tag and length octets are
5041 already assumed to have been read in and checked. These
5042 are needed because all other string types have virtually
5043 identical handling apart from the tag. By having versions
5044 of the ASN1 functions that just operate on content octets
5045 IMPLICIT tagging can be handled properly. It also allows
5046 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5047 and ASN1_INTEGER are identical apart from the tag.
5050 *) Change the handling of OID objects as follows:
5052 - New object identifiers are inserted in objects.txt, following
5053 the syntax given in objects.README.
5054 - objects.pl is used to process obj_mac.num and create a new
5056 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5059 This is currently kind of a hack, and the perl code in objects.pl
5060 isn't very elegant, but it works as I intended. The simplest way
5061 to check that it worked correctly is to look in obj_dat.h and
5062 check the array nid_objs and make sure the objects haven't moved
5063 around (this is important!). Additions are OK, as well as
5064 consistent name changes.
5067 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5070 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5071 The given file adds to whatever has already been seeded into the
5072 random pool through the RANDFILE configuration file option or
5073 environment variable, or the default random state file.
5076 *) mkstack.pl now sorts each macro group into lexical order.
5077 Previously the output order depended on the order the files
5078 appeared in the directory, resulting in needless rewriting
5082 *) Patches to make OpenSSL compile under Win32 again. Mostly
5083 work arounds for the VC++ problem that it treats func() as
5084 func(void). Also stripped out the parts of mkdef.pl that
5085 added extra typesafe functions: these no longer exist.
5088 *) Reorganisation of the stack code. The macros are now all
5089 collected in safestack.h . Each macro is defined in terms of
5090 a "stack macro" of the form SKM_<name>(type, a, b). The
5091 DEBUG_SAFESTACK is now handled in terms of function casts,
5092 this has the advantage of retaining type safety without the
5093 use of additional functions. If DEBUG_SAFESTACK is not defined
5094 then the non typesafe macros are used instead. Also modified the
5095 mkstack.pl script to handle the new form. Needs testing to see
5096 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5097 the default if no major problems. Similar behaviour for ASN1_SET_OF
5098 and PKCS12_STACK_OF.
5101 *) When some versions of IIS use the 'NET' form of private key the
5102 key derivation algorithm is different. Normally MD5(password) is
5103 used as a 128 bit RC4 key. In the modified case
5104 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5105 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5106 as the old Netscape_RSA functions except they have an additional
5107 'sgckey' parameter which uses the modified algorithm. Also added
5108 an -sgckey command line option to the rsa utility. Thanks to
5109 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5110 algorithm to openssl-dev.
5113 *) The evp_local.h macros were using 'c.##kname' which resulted in
5114 invalid expansion on some systems (SCO 5.0.5 for example).
5115 Corrected to 'c.kname'.
5116 [Phillip Porch <root@theporch.com>]
5118 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5119 a STACK of email addresses from a certificate or request, these look
5120 in the subject name and the subject alternative name extensions and
5121 omit any duplicate addresses.
5124 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5125 This makes DSA verification about 2 % faster.
5128 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5129 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5130 plus overhead for 1024 bit moduli).
5131 This makes exponentiations about 0.5 % faster for 1024 bit
5132 exponents (as measured by "openssl speed rsa2048").
5135 *) Rename memory handling macros to avoid conflicts with other
5137 Malloc => OPENSSL_malloc
5138 Malloc_locked => OPENSSL_malloc_locked
5139 Realloc => OPENSSL_realloc
5140 Free => OPENSSL_free
5143 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5144 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5147 *) CygWin32 support.
5148 [John Jarvie <jjarvie@newsguy.com>]
5150 *) The type-safe stack code has been rejigged. It is now only compiled
5151 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5152 by default all type-specific stack functions are "#define"d back to
5153 standard stack functions. This results in more streamlined output
5154 but retains the type-safety checking possibilities of the original
5158 *) The STACK code has been cleaned up, and certain type declarations
5159 that didn't make a lot of sense have been brought in line. This has
5160 also involved a cleanup of sorts in safestack.h to more correctly
5161 map type-safe stack functions onto their plain stack counterparts.
5162 This work has also resulted in a variety of "const"ifications of
5163 lots of the code, especially "_cmp" operations which should normally
5164 be prototyped with "const" parameters anyway.
5167 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5168 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5169 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5170 where all of 'md' is used each time the PRNG is used, but 'state'
5171 is used only indexed by a cyclic counter. As entropy may not be
5172 well distributed from the beginning, 'md' is important as a
5173 chaining variable. However, the output function chains only half
5174 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5175 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5176 in all of 'state' being rewritten, with the new values depending
5177 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5180 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5181 the handshake is continued after ssl_verify_cert_chain();
5182 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5183 can lead to 'unexplainable' connection aborts later.
5184 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5186 *) Major EVP API cipher revision.
5187 Add hooks for extra EVP features. This allows various cipher
5188 parameters to be set in the EVP interface. Support added for variable
5189 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5190 setting of RC2 and RC5 parameters.
5192 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5195 Remove lots of duplicated code from the EVP library. For example *every*
5196 cipher init() function handles the 'iv' in the same way according to the
5197 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5198 for CFB and OFB modes they zero ctx->num.
5200 New functionality allows removal of S/MIME code RC2 hack.
5202 Most of the routines have the same form and so can be declared in terms
5205 By shifting this to the top level EVP_CipherInit() it can be removed from
5206 all individual ciphers. If the cipher wants to handle IVs or keys
5207 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5210 Change lots of functions like EVP_EncryptUpdate() to now return a
5211 value: although software versions of the algorithms cannot fail
5212 any installed hardware versions can.
5215 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5216 this option is set, tolerate broken clients that send the negotiated
5217 protocol version number instead of the requested protocol version
5221 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5222 i.e. non-zero for export ciphersuites, zero otherwise.
5223 Previous versions had this flag inverted, inconsistent with
5224 rsa_tmp_cb (..._TMP_RSA_CB).
5225 [Bodo Moeller; problem reported by Amit Chopra]
5227 *) Add missing DSA library text string. Work around for some IIS
5228 key files with invalid SEQUENCE encoding.
5231 *) Add a document (doc/standards.txt) that list all kinds of standards
5232 and so on that are implemented in OpenSSL.
5235 *) Enhance c_rehash script. Old version would mishandle certificates
5236 with the same subject name hash and wouldn't handle CRLs at all.
5237 Added -fingerprint option to crl utility, to support new c_rehash
5241 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5244 *) Fix for SSL server purpose checking. Server checking was
5245 rejecting certificates which had extended key usage present
5246 but no ssl client purpose.
5247 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5249 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5250 is a little unclear about how a blank password is handled.
5251 Since the password in encoded as a BMPString with terminating
5252 double NULL a zero length password would end up as just the
5253 double NULL. However no password at all is different and is
5254 handled differently in the PKCS#12 key generation code. NS
5255 treats a blank password as zero length. MSIE treats it as no
5256 password on export: but it will try both on import. We now do
5257 the same: PKCS12_parse() tries zero length and no password if
5258 the password is set to "" or NULL (NULL is now a valid password:
5259 it wasn't before) as does the pkcs12 application.
5262 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5263 perror when PEM_read_bio_X509_REQ fails, the error message must
5264 be obtained from the error queue.
5267 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5268 it in ERR_remove_state if appropriate, and change ERR_get_state
5269 accordingly to avoid race conditions (this is necessary because
5270 thread_hash is no longer constant once set).
5273 *) Bugfix for linux-elf makefile.one.
5276 *) RSA_get_default_method() will now cause a default
5277 RSA_METHOD to be chosen if one doesn't exist already.
5278 Previously this was only set during a call to RSA_new()
5279 or RSA_new_method(NULL) meaning it was possible for
5280 RSA_get_default_method() to return NULL.
5283 *) Added native name translation to the existing DSO code
5284 that will convert (if the flag to do so is set) filenames
5285 that are sufficiently small and have no path information
5286 into a canonical native form. Eg. "blah" converted to
5287 "libblah.so" or "blah.dll" etc.
5290 *) New function ERR_error_string_n(e, buf, len) which is like
5291 ERR_error_string(e, buf), but writes at most 'len' bytes
5292 including the 0 terminator. For ERR_error_string_n, 'buf'
5294 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5296 *) CONF library reworked to become more general. A new CONF
5297 configuration file reader "class" is implemented as well as a
5298 new functions (NCONF_*, for "New CONF") to handle it. The now
5299 old CONF_* functions are still there, but are reimplemented to
5300 work in terms of the new functions. Also, a set of functions
5301 to handle the internal storage of the configuration data is
5302 provided to make it easier to write new configuration file
5303 reader "classes" (I can definitely see something reading a
5304 configuration file in XML format, for example), called _CONF_*,
5305 or "the configuration storage API"...
5307 The new configuration file reading functions are:
5309 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5310 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5312 NCONF_default, NCONF_WIN32
5314 NCONF_dump_fp, NCONF_dump_bio
5316 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5317 NCONF_new creates a new CONF object. This works in the same way
5318 as other interfaces in OpenSSL, like the BIO interface.
5319 NCONF_dump_* dump the internal storage of the configuration file,
5320 which is useful for debugging. All other functions take the same
5321 arguments as the old CONF_* functions wth the exception of the
5322 first that must be a `CONF *' instead of a `LHASH *'.
5324 To make it easer to use the new classes with the old CONF_* functions,
5325 the function CONF_set_default_method is provided.
5328 *) Add '-tls1' option to 'openssl ciphers', which was already
5329 mentioned in the documentation but had not been implemented.
5330 (This option is not yet really useful because even the additional
5331 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5334 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5335 OpenSSL-based applications) load shared libraries and bind to
5336 them in a portable way.
5337 [Geoff Thorpe, with contributions from Richard Levitte]
5339 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5341 *) Make sure _lrotl and _lrotr are only used with MSVC.
5343 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5344 (the default implementation of RAND_status).
5346 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5347 to '-clrext' (= clear extensions), as intended and documented.
5348 [Bodo Moeller; inconsistency pointed out by Michael Attili
5351 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5352 was larger than the MD block size.
5353 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5355 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5356 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5357 using the passed key: if the passed key was a private key the result
5358 of X509_print(), for example, would be to print out all the private key
5362 *) des_quad_cksum() byte order bug fix.
5363 [Ulf Möller, using the problem description in krb4-0.9.7, where
5364 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5366 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5368 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5370 *) For easily testing in shell scripts whether some command
5371 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5372 returns with exit code 0 iff no command of the given name is available.
5373 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5374 the output goes to stdout and nothing is printed to stderr.
5375 Additional arguments are always ignored.
5377 Since for each cipher there is a command of the same name,
5378 the 'no-cipher' compilation switches can be tested this way.
5380 ('openssl no-XXX' is not able to detect pseudo-commands such
5381 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5384 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5387 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5388 is set; it will be thrown away anyway because each handshake creates
5390 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5391 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5392 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5393 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5396 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5397 'Q' and 'R' lose their special meanings (quit/renegotiate).
5398 This is part of what -quiet does; unlike -quiet, -ign_eof
5399 does not suppress any output.
5402 *) Add compatibility options to the purpose and trust code. The
5403 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5404 accepts a certificate or CA, this was the previous behaviour,
5405 with all the associated security issues.
5407 X509_TRUST_COMPAT is the old trust behaviour: only and
5408 automatically trust self signed roots in certificate store. A
5409 new trust setting X509_TRUST_DEFAULT is used to specify that
5410 a purpose has no associated trust setting and it should instead
5411 use the value in the default purpose.
5414 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5415 and fix a memory leak.
5418 *) In util/mkerr.pl (which implements 'make errors'), preserve
5419 reason strings from the previous version of the .c file, as
5420 the default to have only downcase letters (and digits) in
5421 automatically generated reasons codes is not always appropriate.
5424 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5425 using strerror. Previously, ERR_reason_error_string() returned
5426 library names as reason strings for SYSerr; but SYSerr is a special
5427 case where small numbers are errno values, not library numbers.
5430 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5431 converts DSA parameters into DH parameters. (When creating parameters,
5432 DSA_generate_parameters is used.)
5435 *) Include 'length' (recommended exponent length) in C code generated
5436 by 'openssl dhparam -C'.
5439 *) The second argument to set_label in perlasm was already being used
5440 so couldn't be used as a "file scope" flag. Moved to third argument
5444 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5445 instead of RAND_bytes for encryption IVs and salts.
5448 *) Include RAND_status() into RAND_METHOD instead of implementing
5449 it only for md_rand.c Otherwise replacing the PRNG by calling
5450 RAND_set_rand_method would be impossible.
5453 *) Don't let DSA_generate_key() enter an infinite loop if the random
5454 number generation fails.
5457 *) New 'rand' application for creating pseudo-random output.
5460 *) Added configuration support for Linux/IA64
5461 [Rolf Haberrecker <rolf@suse.de>]
5463 *) Assembler module support for Mingw32.
5466 *) Shared library support for HPUX (in shlib/).
5467 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5469 *) Shared library support for Solaris gcc.
5470 [Lutz Behnke <behnke@trustcenter.de>]
5472 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5474 *) PKCS7_encrypt() was adding text MIME headers twice because they
5475 were added manually and by SMIME_crlf_copy().
5478 *) In bntest.c don't call BN_rand with zero bits argument.
5479 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5481 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5482 case was implemented. This caused BN_div_recp() to fail occasionally.
5485 *) Add an optional second argument to the set_label() in the perl
5486 assembly language builder. If this argument exists and is set
5487 to 1 it signals that the assembler should use a symbol whose
5488 scope is the entire file, not just the current function. This
5489 is needed with MASM which uses the format label:: for this scope.
5490 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5492 *) Change the ASN1 types so they are typedefs by default. Before
5493 almost all types were #define'd to ASN1_STRING which was causing
5494 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5498 *) Change names of new functions to the new get1/get0 naming
5499 convention: After 'get1', the caller owns a reference count
5500 and has to call ..._free; 'get0' returns a pointer to some
5501 data structure without incrementing reference counters.
5502 (Some of the existing 'get' functions increment a reference
5503 counter, some don't.)
5504 Similarly, 'set1' and 'add1' functions increase reference
5505 counters or duplicate objects.
5508 *) Allow for the possibility of temp RSA key generation failure:
5509 the code used to assume it always worked and crashed on failure.
5512 *) Fix potential buffer overrun problem in BIO_printf().
5513 [Ulf Möller, using public domain code by Patrick Powell; problem
5514 pointed out by David Sacerdote <das33@cornell.edu>]
5516 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5517 RAND_egd() and RAND_status(). In the command line application,
5518 the EGD socket can be specified like a seed file using RANDFILE
5522 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5523 Some CAs (e.g. Verisign) distribute certificates in this form.
5526 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5527 list to exclude them. This means that no special compilation option
5528 is needed to use anonymous DH: it just needs to be included in the
5532 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5533 EVP_MD_type. The old functionality is available in a new macro called
5534 EVP_MD_md(). Change code that uses it and update docs.
5537 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5538 where the 'void *' argument is replaced by a function pointer argument.
5539 Previously 'void *' was abused to point to functions, which works on
5540 many platforms, but is not correct. As these functions are usually
5541 called by macros defined in OpenSSL header files, most source code
5542 should work without changes.
5545 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5546 sections with information on -D... compiler switches used for
5547 compiling the library so that applications can see them. To enable
5548 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5549 must be defined. E.g.,
5550 #define OPENSSL_ALGORITHM_DEFINES
5551 #include <openssl/opensslconf.h>
5552 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5553 [Richard Levitte, Ulf and Bodo Möller]
5555 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5559 *) Change the 'other' type in certificate aux info to a STACK_OF
5560 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5561 the required ASN1 format: arbitrary types determined by an OID.
5564 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5565 argument to 'req'. This is not because the function is newer or
5566 better than others it just uses the work 'NEW' in the certificate
5567 request header lines. Some software needs this.
5570 *) Reorganise password command line arguments: now passwords can be
5571 obtained from various sources. Delete the PEM_cb function and make
5572 it the default behaviour: i.e. if the callback is NULL and the
5573 usrdata argument is not NULL interpret it as a null terminated pass
5574 phrase. If usrdata and the callback are NULL then the pass phrase
5575 is prompted for as usual.
5578 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5579 the support is automatically enabled. The resulting binaries will
5580 autodetect the card and use it if present.
5581 [Ben Laurie and Compaq Inc.]
5583 *) Work around for Netscape hang bug. This sends certificate request
5584 and server done in one record. Since this is perfectly legal in the
5585 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5586 the bugs/SSLv3 entry for more info.
5589 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5592 *) Add -rand argument to smime and pkcs12 applications and read/write
5596 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5599 *) Add command line password options to the remaining applications.
5602 *) Bug fix for BN_div_recp() for numerators with an even number of
5606 *) More tests in bntest.c, and changed test_bn output.
5609 *) ./config recognizes MacOS X now.
5612 *) Bug fix for BN_div() when the first words of num and divsor are
5613 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5616 *) Add support for various broken PKCS#8 formats, and command line
5617 options to produce them.
5620 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5621 get temporary BIGNUMs from a BN_CTX.
5624 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5628 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5629 include a #define from the old name to the new. The original intent
5630 was that statically linked binaries could for example just call
5631 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5632 link with digests. This never worked becayse SSLeay_add_all_digests()
5633 and SSLeay_add_all_ciphers() were in the same source file so calling
5634 one would link with the other. They are now in separate source files.
5637 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5640 *) Use a less unusual form of the Miller-Rabin primality test (it used
5641 a binary algorithm for exponentiation integrated into the Miller-Rabin
5642 loop, our standard modexp algorithms are faster).
5645 *) Support for the EBCDIC character set completed.
5646 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5648 *) Source code cleanups: use const where appropriate, eliminate casts,
5649 use void * instead of char * in lhash.
5652 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5653 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5654 this the server could overwrite ephemeral keys that the client
5658 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5659 using 50 iterations of the Rabin-Miller test.
5661 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5662 iterations of the Rabin-Miller test as required by the appendix
5663 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5664 As BN_is_prime_fasttest includes trial division, DSA parameter
5665 generation becomes much faster.
5667 This implies a change for the callback functions in DSA_is_prime
5668 and DSA_generate_parameters: The callback function is called once
5669 for each positive witness in the Rabin-Miller test, not just
5670 occasionally in the inner loop; and the parameters to the
5671 callback function now provide an iteration count for the outer
5672 loop rather than for the current invocation of the inner loop.
5673 DSA_generate_parameters additionally can call the callback
5674 function with an 'iteration count' of -1, meaning that a
5675 candidate has passed the trial division test (when q is generated
5676 from an application-provided seed, trial division is skipped).
5679 *) New function BN_is_prime_fasttest that optionally does trial
5680 division before starting the Rabin-Miller test and has
5681 an additional BN_CTX * argument (whereas BN_is_prime always
5682 has to allocate at least one BN_CTX).
5683 'callback(1, -1, cb_arg)' is called when a number has passed the
5684 trial division stage.
5687 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5691 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5694 *) New function BN_pseudo_rand().
5697 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5698 bignum version of BN_from_montgomery() with the working code from
5699 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5703 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5704 made it impossible to use the same SSL_SESSION data structure in
5705 SSL2 clients in multiple threads.
5708 *) The return value of RAND_load_file() no longer counts bytes obtained
5709 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5710 to seed the PRNG (previously an explicit byte count was required).
5711 [Ulf Möller, Bodo Möller]
5713 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5714 used (char *) instead of (void *) and had casts all over the place.
5717 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5720 *) Retain source code compatibility for BN_prime_checks macro:
5721 BN_is_prime(..., BN_prime_checks, ...) now uses
5722 BN_prime_checks_for_size to determine the appropriate number of
5723 Rabin-Miller iterations.
5726 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5727 DH_CHECK_P_NOT_SAFE_PRIME.
5728 (Check if this is true? OpenPGP calls them "strong".)
5731 *) Merge the functionality of "dh" and "gendh" programs into a new program
5732 "dhparam". The old programs are retained for now but will handle DH keys
5733 (instead of parameters) in future.
5736 *) Make the ciphers, s_server and s_client programs check the return values
5737 when a new cipher list is set.
5740 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5741 ciphers. Before when the 56bit ciphers were enabled the sorting was
5744 The syntax for the cipher sorting has been extended to support sorting by
5745 cipher-strength (using the strength_bits hard coded in the tables).
5746 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5748 Fix a bug in the cipher-command parser: when supplying a cipher command
5749 string with an "undefined" symbol (neither command nor alphanumeric
5750 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5751 an error is flagged.
5753 Due to the strength-sorting extension, the code of the
5754 ssl_create_cipher_list() function was completely rearranged. I hope that
5755 the readability was also increased :-)
5756 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5758 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5759 for the first serial number and places 2 in the serial number file. This
5760 avoids problems when the root CA is created with serial number zero and
5761 the first user certificate has the same issuer name and serial number
5765 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5766 the new code. Add documentation for this stuff.
5769 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5770 X509_*() to X509at_*() on the grounds that they don't handle X509
5771 structures and behave in an analagous way to the X509v3 functions:
5772 they shouldn't be called directly but wrapper functions should be used
5775 So we also now have some wrapper functions that call the X509at functions
5776 when passed certificate requests. (TO DO: similar things can be done with
5777 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5778 things. Some of these need some d2i or i2d and print functionality
5779 because they handle more complex structures.)
5782 *) Add missing #ifndefs that caused missing symbols when building libssl
5783 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5784 NO_RSA in ssl/s2*.c.
5785 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5787 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5788 has a return value which indicates the quality of the random data
5789 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5790 error queue. New function RAND_pseudo_bytes() generates output that is
5791 guaranteed to be unique but not unpredictable. RAND_add is like
5792 RAND_seed, but takes an extra argument for an entropy estimate
5793 (RAND_seed always assumes full entropy).
5796 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5797 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5798 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5799 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5800 false-positive rate of at most 2^-80 for random input.
5803 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5806 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5807 in the 0.9.5 release), this returns the chain
5808 from an X509_CTX structure with a dup of the stack and all
5809 the X509 reference counts upped: so the stack will exist
5810 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5813 Also make SSL_SESSION_print() print out the verify return
5817 *) Add manpage for the pkcs12 command. Also change the default
5818 behaviour so MAC iteration counts are used unless the new
5819 -nomaciter option is used. This improves file security and
5820 only older versions of MSIE (4.0 for example) need it.
5823 *) Honor the no-xxx Configure options when creating .DEF files.
5826 *) Add PKCS#10 attributes to field table: challengePassword,
5827 unstructuredName and unstructuredAddress. These are taken from
5828 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5829 international characters are used.
5831 More changes to X509_ATTRIBUTE code: allow the setting of types
5832 based on strings. Remove the 'loc' parameter when adding
5833 attributes because these will be a SET OF encoding which is sorted
5837 *) Initial changes to the 'req' utility to allow request generation
5838 automation. This will allow an application to just generate a template
5839 file containing all the field values and have req construct the
5842 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5843 used all over the place including certificate requests and PKCS#7
5844 structures. They are currently handled manually where necessary with
5845 some primitive wrappers for PKCS#7. The new functions behave in a
5846 manner analogous to the X509 extension functions: they allow
5847 attributes to be looked up by NID and added.
5849 Later something similar to the X509V3 code would be desirable to
5850 automatically handle the encoding, decoding and printing of the
5851 more complex types. The string types like challengePassword can
5852 be handled by the string table functions.
5854 Also modified the multi byte string table handling. Now there is
5855 a 'global mask' which masks out certain types. The table itself
5856 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5857 is useful when for example there is only one permissible type
5858 (as in countryName) and using the mask might result in no valid
5862 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5863 SSL_get_peer_finished to allow applications to obtain the latest
5864 Finished messages sent to the peer or expected from the peer,
5865 respectively. (SSL_get_peer_finished is usually the Finished message
5866 actually received from the peer, otherwise the protocol will be aborted.)
5868 As the Finished message are message digests of the complete handshake
5869 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5870 be used for external authentication procedures when the authentication
5871 provided by SSL/TLS is not desired or is not enough.
5874 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5875 the host supports BWX extension and if Compaq C is present on the
5876 $PATH. Just exploiting of the BWX extension results in 20-30%
5877 performance kick for some algorithms, e.g. DES and RC4 to mention
5878 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5882 *) Add support for MS "fast SGC". This is arguably a violation of the
5883 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5884 weak crypto and after checking the certificate is SGC a second one
5885 with strong crypto. MS SGC stops the first handshake after receiving
5886 the server certificate message and sends a second client hello. Since
5887 a server will typically do all the time consuming operations before
5888 expecting any further messages from the client (server key exchange
5889 is the most expensive) there is little difference between the two.
5891 To get OpenSSL to support MS SGC we have to permit a second client
5892 hello message after we have sent server done. In addition we have to
5893 reset the MAC if we do get this second client hello.
5896 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5897 if a DER encoded private key is RSA or DSA traditional format. Changed
5898 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5899 format DER encoded private key. Newer code should use PKCS#8 format which
5900 has the key type encoded in the ASN1 structure. Added DER private key
5901 support to pkcs8 application.
5904 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5905 ciphersuites has been selected (as required by the SSL 3/TLS 1
5906 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5907 is set, we interpret this as a request to violate the specification
5908 (the worst that can happen is a handshake failure, and 'correct'
5909 behaviour would result in a handshake failure anyway).
5912 *) In SSL_CTX_add_session, take into account that there might be multiple
5913 SSL_SESSION structures with the same session ID (e.g. when two threads
5914 concurrently obtain them from an external cache).
5915 The internal cache can handle only one SSL_SESSION with a given ID,
5916 so if there's a conflict, we now throw out the old one to achieve
5920 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5921 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5922 some routines that use cipher OIDs: some ciphers do not have OIDs
5923 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5927 *) Simplify the trust setting structure and code. Now we just have
5928 two sequences of OIDs for trusted and rejected settings. These will
5929 typically have values the same as the extended key usage extension
5930 and any application specific purposes.
5932 The trust checking code now has a default behaviour: it will just
5933 check for an object with the same NID as the passed id. Functions can
5934 be provided to override either the default behaviour or the behaviour
5935 for a given id. SSL client, server and email already have functions
5936 in place for compatibility: they check the NID and also return "trusted"
5937 if the certificate is self signed.
5940 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5941 traditional format into an EVP_PKEY structure.
5944 *) Add a password callback function PEM_cb() which either prompts for
5945 a password if usr_data is NULL or otherwise assumes it is a null
5946 terminated password. Allow passwords to be passed on command line
5947 environment or config files in a few more utilities.
5950 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5951 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5952 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5953 Update documentation.
5956 *) Support for ASN1 "NULL" type. This could be handled before by using
5957 ASN1_TYPE but there wasn't any function that would try to read a NULL
5958 and produce an error if it couldn't. For compatibility we also have
5959 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5960 don't allocate anything because they don't need to.
5963 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5965 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5967 *) Rebuild of the memory allocation routines used by OpenSSL code and
5968 possibly others as well. The purpose is to make an interface that
5969 provide hooks so anyone can build a separate set of allocation and
5970 deallocation routines to be used by OpenSSL, for example memory
5971 pool implementations, or something else, which was previously hard
5972 since Malloc(), Realloc() and Free() were defined as macros having
5973 the values malloc, realloc and free, respectively (except for Win32
5974 compilations). The same is provided for memory debugging code.
5975 OpenSSL already comes with functionality to find memory leaks, but
5976 this gives people a chance to debug other memory problems.
5978 With these changes, a new set of functions and macros have appeared:
5980 CRYPTO_set_mem_debug_functions() [F]
5981 CRYPTO_get_mem_debug_functions() [F]
5982 CRYPTO_dbg_set_options() [F]
5983 CRYPTO_dbg_get_options() [F]
5984 CRYPTO_malloc_debug_init() [M]
5986 The memory debug functions are NULL by default, unless the library
5987 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5988 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5989 gives the standard debugging functions that come with OpenSSL) or
5990 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5991 provided by the library user) must be used. When the standard
5992 debugging functions are used, CRYPTO_dbg_set_options can be used to
5993 request additional information:
5994 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5995 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5997 Also, things like CRYPTO_set_mem_functions will always give the
5998 expected result (the new set of functions is used for allocation
5999 and deallocation) at all times, regardless of platform and compiler
6002 To finish it up, some functions that were never use in any other
6003 way than through macros have a new API and new semantic:
6006 CRYPTO_dbg_realloc()
6009 All macros of value have retained their old syntax.
6010 [Richard Levitte and Bodo Moeller]
6012 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6013 ordering of SMIMECapabilities wasn't in "strength order" and there
6014 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6018 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6019 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6020 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6022 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6023 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6024 functionality to handle multipart/signed properly) and a utility
6025 called 'smime' to call all this stuff. This is based on code I
6026 originally wrote for Celo who have kindly allowed it to be
6027 included in OpenSSL.
6030 *) Add variants des_set_key_checked and des_set_key_unchecked of
6031 des_set_key (aka des_key_sched). Global variable des_check_key
6032 decides which of these is called by des_set_key; this way
6033 des_check_key behaves as it always did, but applications and
6034 the library itself, which was buggy for des_check_key == 1,
6035 have a cleaner way to pick the version they need.
6038 *) New function PKCS12_newpass() which changes the password of a
6042 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6043 dynamic mix. In both cases the ids can be used as an index into the
6044 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6045 functions so they accept a list of the field values and the
6046 application doesn't need to directly manipulate the X509_TRUST
6050 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6054 *) Modify the way the V3 extension code looks up extensions. This now
6055 works in a similar way to the object code: we have some "standard"
6056 extensions in a static table which is searched with OBJ_bsearch()
6057 and the application can add dynamic ones if needed. The file
6058 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6059 updated whenever a new extension is added to the core code and kept
6060 in ext_nid order. There is a simple program 'tabtest.c' which checks
6061 this. New extensions are not added too often so this file can readily
6062 be maintained manually.
6064 There are two big advantages in doing things this way. The extensions
6065 can be looked up immediately and no longer need to be "added" using
6066 X509V3_add_standard_extensions(): this function now does nothing.
6067 [Side note: I get *lots* of email saying the extension code doesn't
6068 work because people forget to call this function]
6069 Also no dynamic allocation is done unless new extensions are added:
6070 so if we don't add custom extensions there is no need to call
6071 X509V3_EXT_cleanup().
6074 *) Modify enc utility's salting as follows: make salting the default. Add a
6075 magic header, so unsalted files fail gracefully instead of just decrypting
6076 to garbage. This is because not salting is a big security hole, so people
6077 should be discouraged from doing it.
6080 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6081 digest to be passed on the command line but it only used this
6082 parameter when signing a certificate. Modified so all relevant
6083 operations are affected by the digest parameter including the
6084 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6085 DSA key was used because it didn't fix the digest.
6088 *) Initial certificate chain verify code. Currently tests the untrusted
6089 certificates for consistency with the verify purpose (which is set
6090 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6092 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6093 this is because it will reject chains with invalid extensions whereas
6094 every previous version of OpenSSL and SSLeay made no checks at all.
6096 Trust code: checks the root CA for the relevant trust settings. Trust
6097 settings have an initial value consistent with the verify purpose: e.g.
6098 if the verify purpose is for SSL client use it expects the CA to be
6099 trusted for SSL client use. However the default value can be changed to
6100 permit custom trust settings: one example of this would be to only trust
6101 certificates from a specific "secure" set of CAs.
6103 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6104 which should be used for version portability: especially since the
6105 verify structure is likely to change more often now.
6107 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6108 to set them. If not set then assume SSL clients will verify SSL servers
6111 Two new options to the verify program: -untrusted allows a set of
6112 untrusted certificates to be passed in and -purpose which sets the
6113 intended purpose of the certificate. If a purpose is set then the
6114 new chain verify code is used to check extension consistency.
6117 *) Support for the authority information access extension.
6120 *) Modify RSA and DSA PEM read routines to transparently handle
6121 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6122 public keys in a format compatible with certificate
6123 SubjectPublicKeyInfo structures. Unfortunately there were already
6124 functions called *_PublicKey_* which used various odd formats so
6125 these are retained for compatibility: however the DSA variants were
6126 never in a public release so they have been deleted. Changed dsa/rsa
6127 utilities to handle the new format: note no releases ever handled public
6128 keys so we should be OK.
6130 The primary motivation for this change is to avoid the same fiasco
6131 that dogs private keys: there are several incompatible private key
6132 formats some of which are standard and some OpenSSL specific and
6133 require various evil hacks to allow partial transparent handling and
6134 even then it doesn't work with DER formats. Given the option anything
6135 other than PKCS#8 should be dumped: but the other formats have to
6136 stay in the name of compatibility.
6138 With public keys and the benefit of hindsight one standard format
6139 is used which works with EVP_PKEY, RSA or DSA structures: though
6140 it clearly returns an error if you try to read the wrong kind of key.
6142 Added a -pubkey option to the 'x509' utility to output the public key.
6143 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6144 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6145 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6146 that do the same as the EVP_PKEY_assign_*() except they up the
6147 reference count of the added key (they don't "swallow" the
6151 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6152 CRLs would fail if the file contained no certificates or no CRLs:
6153 added a new function to read in both types and return the number
6154 read: this means that if none are read it will be an error. The
6155 DER versions of the certificate and CRL reader would always fail
6156 because it isn't possible to mix certificates and CRLs in DER format
6157 without choking one or the other routine. Changed this to just read
6158 a certificate: this is the best we can do. Also modified the code
6159 in apps/verify.c to take notice of return codes: it was previously
6160 attempting to read in certificates from NULL pointers and ignoring
6161 any errors: this is one reason why the cert and CRL reader seemed
6162 to work. It doesn't check return codes from the default certificate
6163 routines: these may well fail if the certificates aren't installed.
6166 *) Code to support otherName option in GeneralName.
6169 *) First update to verify code. Change the verify utility
6170 so it warns if it is passed a self signed certificate:
6171 for consistency with the normal behaviour. X509_verify
6172 has been modified to it will now verify a self signed
6173 certificate if *exactly* the same certificate appears
6174 in the store: it was previously impossible to trust a
6175 single self signed certificate. This means that:
6176 openssl verify ss.pem
6177 now gives a warning about a self signed certificate but
6178 openssl verify -CAfile ss.pem ss.pem
6182 *) For servers, store verify_result in SSL_SESSION data structure
6183 (and add it to external session representation).
6184 This is needed when client certificate verifications fails,
6185 but an application-provided verification callback (set by
6186 SSL_CTX_set_cert_verify_callback) allows accepting the session
6187 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6188 but returns 1): When the session is reused, we have to set
6189 ssl->verify_result to the appropriate error code to avoid
6191 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6193 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6194 case in PKCS7_dataInit() where the signed PKCS7 structure
6195 didn't contain any existing data because it was being created.
6196 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6198 *) Add a salt to the key derivation routines in enc.c. This
6199 forms the first 8 bytes of the encrypted file. Also add a
6200 -S option to allow a salt to be input on the command line.
6203 *) New function X509_cmp(). Oddly enough there wasn't a function
6204 to compare two certificates. We do this by working out the SHA1
6205 hash and comparing that. X509_cmp() will be needed by the trust
6209 *) SSL_get1_session() is like SSL_get_session(), but increments
6210 the reference count in the SSL_SESSION returned.
6211 [Geoff Thorpe <geoff@eu.c2.net>]
6213 *) Fix for 'req': it was adding a null to request attributes.
6214 Also change the X509_LOOKUP and X509_INFO code to handle
6215 certificate auxiliary information.
6218 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6222 *) Add the possibility to add extra information to the memory leak
6223 detecting output, to form tracebacks, showing from where each
6224 allocation was originated: CRYPTO_push_info("constant string") adds
6225 the string plus current file name and line number to a per-thread
6226 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6227 is like calling CYRPTO_pop_info() until the stack is empty.
6228 Also updated memory leak detection code to be multi-thread-safe.
6231 *) Add options -text and -noout to pkcs7 utility and delete the
6232 encryption options which never did anything. Update docs.
6235 *) Add options to some of the utilities to allow the pass phrase
6236 to be included on either the command line (not recommended on
6237 OSes like Unix) or read from the environment. Update the
6238 manpages and fix a few bugs.
6241 *) Add a few manpages for some of the openssl commands.
6244 *) Fix the -revoke option in ca. It was freeing up memory twice,
6245 leaking and not finding already revoked certificates.
6248 *) Extensive changes to support certificate auxiliary information.
6249 This involves the use of X509_CERT_AUX structure and X509_AUX
6250 functions. An X509_AUX function such as PEM_read_X509_AUX()
6251 can still read in a certificate file in the usual way but it
6252 will also read in any additional "auxiliary information". By
6253 doing things this way a fair degree of compatibility can be
6254 retained: existing certificates can have this information added
6255 using the new 'x509' options.
6257 Current auxiliary information includes an "alias" and some trust
6258 settings. The trust settings will ultimately be used in enhanced
6259 certificate chain verification routines: currently a certificate
6260 can only be trusted if it is self signed and then it is trusted
6264 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6265 The problem was that one of the replacement routines had not been working
6266 since SSLeay releases. For now the offending routine has been replaced
6267 with non-optimised assembler. Even so, this now gives around 95%
6268 performance improvement for 1024 bit RSA signs.
6271 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6272 handling. Most clients have the effective key size in bits equal to
6273 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6274 A few however don't do this and instead use the size of the decrypted key
6275 to determine the RC2 key length and the AlgorithmIdentifier to determine
6276 the effective key length. In this case the effective key length can still
6277 be 40 bits but the key length can be 168 bits for example. This is fixed
6278 by manually forcing an RC2 key into the EVP_PKEY structure because the
6279 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6280 the key length and effective key length are equal.
6283 *) Add a bunch of functions that should simplify the creation of
6284 X509_NAME structures. Now you should be able to do:
6285 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6286 and have it automatically work out the correct field type and fill in
6287 the structures. The more adventurous can try:
6288 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6289 and it will (hopefully) work out the correct multibyte encoding.
6292 *) Change the 'req' utility to use the new field handling and multibyte
6293 copy routines. Before the DN field creation was handled in an ad hoc
6294 way in req, ca, and x509 which was rather broken and didn't support
6295 BMPStrings or UTF8Strings. Since some software doesn't implement
6296 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6297 using the dirstring_type option. See the new comment in the default
6298 openssl.cnf for more info.
6301 *) Make crypto/rand/md_rand.c more robust:
6302 - Assure unique random numbers after fork().
6303 - Make sure that concurrent threads access the global counter and
6304 md serializably so that we never lose entropy in them
6305 or use exactly the same state in multiple threads.
6306 Access to the large state is not always serializable because
6307 the additional locking could be a performance killer, and
6308 md should be large enough anyway.
6311 *) New file apps/app_rand.c with commonly needed functionality
6312 for handling the random seed file.
6314 Use the random seed file in some applications that previously did not:
6316 dsaparam -genkey (which also ignored its '-rand' option),
6319 x509 (when signing).
6320 Except on systems with /dev/urandom, it is crucial to have a random
6321 seed file at least for key creation, DSA signing, and for DH exchanges;
6322 for RSA signatures we could do without one.
6324 gendh and gendsa (unlike genrsa) used to read only the first byte
6325 of each file listed in the '-rand' option. The function as previously
6326 found in genrsa is now in app_rand.c and is used by all programs
6327 that support '-rand'.
6330 *) In RAND_write_file, use mode 0600 for creating files;
6331 don't just chmod when it may be too late.
6334 *) Report an error from X509_STORE_load_locations
6335 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6338 *) New function ASN1_mbstring_copy() this copies a string in either
6339 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6340 into an ASN1_STRING type. A mask of permissible types is passed
6341 and it chooses the "minimal" type to use or an error if not type
6345 *) Add function equivalents to the various macros in asn1.h. The old
6346 macros are retained with an M_ prefix. Code inside the library can
6347 use the M_ macros. External code (including the openssl utility)
6348 should *NOT* in order to be "shared library friendly".
6351 *) Add various functions that can check a certificate's extensions
6352 to see if it usable for various purposes such as SSL client,
6353 server or S/MIME and CAs of these types. This is currently
6354 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6355 verification. Also added a -purpose flag to x509 utility to
6356 print out all the purposes.
6359 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6363 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6364 for, obtain and decode and extension and obtain its critical flag.
6365 This allows all the necessary extension code to be handled in a
6366 single function call.
6369 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6370 platforms. See crypto/rc4/rc4_enc.c for further details.
6373 *) New -noout option to asn1parse. This causes no output to be produced
6374 its main use is when combined with -strparse and -out to extract data
6375 from a file (which may not be in ASN.1 format).
6378 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6379 when producing the local key id.
6380 [Richard Levitte <levitte@stacken.kth.se>]
6382 *) New option -dhparam in s_server. This allows a DH parameter file to be
6383 stated explicitly. If it is not stated then it tries the first server
6384 certificate file. The previous behaviour hard coded the filename
6388 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6389 a public key to be input or output. For example:
6390 openssl rsa -in key.pem -pubout -out pubkey.pem
6391 Also added necessary DSA public key functions to handle this.
6394 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6395 in the message. This was handled by allowing
6396 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6397 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6399 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6400 to the end of the strings whereas this didn't. This would cause problems
6401 if strings read with d2i_ASN1_bytes() were later modified.
6402 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6404 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6405 data and it contains EOF it will end up returning an error. This is
6406 caused by input 46 bytes long. The cause is due to the way base64
6407 BIOs find the start of base64 encoded data. They do this by trying a
6408 trial decode on each line until they find one that works. When they
6409 do a flag is set and it starts again knowing it can pass all the
6410 data directly through the decoder. Unfortunately it doesn't reset
6411 the context it uses. This means that if EOF is reached an attempt
6412 is made to pass two EOFs through the context and this causes the
6413 resulting error. This can also cause other problems as well. As is
6414 usual with these problems it takes *ages* to find and the fix is
6415 trivial: move one line.
6416 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6418 *) Ugly workaround to get s_client and s_server working under Windows. The
6419 old code wouldn't work because it needed to select() on sockets and the
6420 tty (for keypresses and to see if data could be written). Win32 only
6421 supports select() on sockets so we select() with a 1s timeout on the
6422 sockets and then see if any characters are waiting to be read, if none
6423 are present then we retry, we also assume we can always write data to
6424 the tty. This isn't nice because the code then blocks until we've
6425 received a complete line of data and it is effectively polling the
6426 keyboard at 1s intervals: however it's quite a bit better than not
6427 working at all :-) A dedicated Windows application might handle this
6428 with an event loop for example.
6431 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6432 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6433 will be called when RSA_sign() and RSA_verify() are used. This is useful
6434 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6435 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6436 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6437 This necessitated the support of an extra signature type NID_md5_sha1
6438 for SSL signatures and modifications to the SSL library to use it instead
6439 of calling RSA_public_decrypt() and RSA_private_encrypt().
6442 *) Add new -verify -CAfile and -CApath options to the crl program, these
6443 will lookup a CRL issuers certificate and verify the signature in a
6444 similar way to the verify program. Tidy up the crl program so it
6445 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6446 less strict. It will now permit CRL extensions even if it is not
6447 a V2 CRL: this will allow it to tolerate some broken CRLs.
6450 *) Initialize all non-automatic variables each time one of the openssl
6451 sub-programs is started (this is necessary as they may be started
6452 multiple times from the "OpenSSL>" prompt).
6453 [Lennart Bang, Bodo Moeller]
6455 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6456 removing all other RSA functionality (this is what NO_RSA does). This
6457 is so (for example) those in the US can disable those operations covered
6458 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6462 *) Non-copying interface to BIO pairs.
6463 (still largely untested)
6466 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6467 ASCII string. This was handled independently in various places before.
6470 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6471 UTF8 strings a character at a time.
6474 *) Use client_version from client hello to select the protocol
6475 (s23_srvr.c) and for RSA client key exchange verification
6476 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6479 *) Add various utility functions to handle SPKACs, these were previously
6480 handled by poking round in the structure internals. Added new function
6481 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6482 print, verify and generate SPKACs. Based on an original idea from
6483 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6486 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6489 *) Allow the config file extension section to be overwritten on the
6490 command line. Based on an original idea from Massimiliano Pala
6491 <madwolf@comune.modena.it>. The new option is called -extensions
6492 and can be applied to ca, req and x509. Also -reqexts to override
6493 the request extensions in req and -crlexts to override the crl extensions
6497 *) Add new feature to the SPKAC handling in ca. Now you can include
6498 the same field multiple times by preceding it by "XXXX." for example:
6501 this is the same syntax as used in the req config file.
6504 *) Allow certificate extensions to be added to certificate requests. These
6505 are specified in a 'req_extensions' option of the req section of the
6506 config file. They can be printed out with the -text option to req but
6507 are otherwise ignored at present.
6510 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6511 data read consists of only the final block it would not decrypted because
6512 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6513 A misplaced 'break' also meant the decrypted final block might not be
6514 copied until the next read.
6517 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6518 a few extra parameters to the DH structure: these will be useful if
6519 for example we want the value of 'q' or implement X9.42 DH.
6522 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6523 provides hooks that allow the default DSA functions or functions on a
6524 "per key" basis to be replaced. This allows hardware acceleration and
6525 hardware key storage to be handled without major modification to the
6526 library. Also added low level modexp hooks and CRYPTO_EX structure and
6527 associated functions.
6530 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6531 as "read only": it can't be written to and the buffer it points to will
6532 not be freed. Reading from a read only BIO is much more efficient than
6533 a normal memory BIO. This was added because there are several times when
6534 an area of memory needs to be read from a BIO. The previous method was
6535 to create a memory BIO and write the data to it, this results in two
6536 copies of the data and an O(n^2) reading algorithm. There is a new
6537 function BIO_new_mem_buf() which creates a read only memory BIO from
6538 an area of memory. Also modified the PKCS#7 routines to use read only
6542 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6543 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6544 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6545 but a retry condition occured while trying to read the rest.
6548 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6549 NID_pkcs7_encrypted by default: this was wrong since this should almost
6550 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6551 the encrypted data type: this is a more sensible place to put it and it
6552 allows the PKCS#12 code to be tidied up that duplicated this
6556 *) Changed obj_dat.pl script so it takes its input and output files on
6557 the command line. This should avoid shell escape redirection problems
6561 *) Initial support for certificate extension requests, these are included
6562 in things like Xenroll certificate requests. Included functions to allow
6563 extensions to be obtained and added.
6566 *) -crlf option to s_client and s_server for sending newlines as
6567 CRLF (as required by many protocols).
6570 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6572 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6573 [Ralf S. Engelschall]
6575 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6576 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6578 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6582 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6583 DH parameters/keys (q is lost during that conversion, but the resulting
6584 DH parameters contain its length).
6586 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6587 much faster than DH_generate_parameters (which creates parameters
6588 where p = 2*q + 1), and also the smaller q makes DH computations
6589 much more efficient (160-bit exponentiation instead of 1024-bit
6590 exponentiation); so this provides a convenient way to support DHE
6591 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6592 utter importance to use
6593 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6595 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6596 when such DH parameters are used, because otherwise small subgroup
6597 attacks may become possible!
6600 *) Avoid memory leak in i2d_DHparams.
6603 *) Allow the -k option to be used more than once in the enc program:
6604 this allows the same encrypted message to be read by multiple recipients.
6607 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6608 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6609 it will always use the numerical form of the OID, even if it has a short
6613 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6614 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6615 otherwise bn_mod_exp was called. In the case of hardware keys for example
6616 no private key components need be present and it might store extra data
6617 in the RSA structure, which cannot be accessed from bn_mod_exp.
6618 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6619 private key operations.
6622 *) Added support for SPARC Linux.
6625 *) pem_password_cb function type incompatibly changed from
6626 typedef int pem_password_cb(char *buf, int size, int rwflag);
6628 ....(char *buf, int size, int rwflag, void *userdata);
6629 so that applications can pass data to their callbacks:
6630 The PEM[_ASN1]_{read,write}... functions and macros now take an
6631 additional void * argument, which is just handed through whenever
6632 the password callback is called.
6633 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6635 New function SSL_CTX_set_default_passwd_cb_userdata.
6637 Compatibility note: As many C implementations push function arguments
6638 onto the stack in reverse order, the new library version is likely to
6639 interoperate with programs that have been compiled with the old
6640 pem_password_cb definition (PEM_whatever takes some data that
6641 happens to be on the stack as its last argument, and the callback
6642 just ignores this garbage); but there is no guarantee whatsoever that
6645 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6646 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6647 problems not only on Windows, but also on some Unix platforms.
6648 To avoid problematic command lines, these definitions are now in an
6649 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6650 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6653 *) MIPS III/IV assembler module is reimplemented.
6656 *) More DES library cleanups: remove references to srand/rand and
6657 delete an unused file.
6660 *) Add support for the the free Netwide assembler (NASM) under Win32,
6661 since not many people have MASM (ml) and it can be hard to obtain.
6662 This is currently experimental but it seems to work OK and pass all
6663 the tests. Check out INSTALL.W32 for info.
6666 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6667 without temporary keys kept an extra copy of the server key,
6668 and connections with temporary keys did not free everything in case
6672 *) New function RSA_check_key and new openssl rsa option -check
6673 for verifying the consistency of RSA keys.
6674 [Ulf Moeller, Bodo Moeller]
6676 *) Various changes to make Win32 compile work:
6677 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6678 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6679 comparison" warnings.
6680 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6683 *) Add a debugging option to PKCS#5 v2 key generation function: when
6684 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6685 derived keys are printed to stderr.
6688 *) Copy the flags in ASN1_STRING_dup().
6689 [Roman E. Pavlov <pre@mo.msk.ru>]
6691 *) The x509 application mishandled signing requests containing DSA
6692 keys when the signing key was also DSA and the parameters didn't match.
6694 It was supposed to omit the parameters when they matched the signing key:
6695 the verifying software was then supposed to automatically use the CA's
6696 parameters if they were absent from the end user certificate.
6698 Omitting parameters is no longer recommended. The test was also
6699 the wrong way round! This was probably due to unusual behaviour in
6700 EVP_cmp_parameters() which returns 1 if the parameters match.
6701 This meant that parameters were omitted when they *didn't* match and
6702 the certificate was useless. Certificates signed with 'ca' didn't have
6704 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6706 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6707 The interface is as follows:
6708 Applications can use
6709 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6710 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6711 "off" is now the default.
6712 The library internally uses
6713 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6714 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6715 to disable memory-checking temporarily.
6717 Some inconsistent states that previously were possible (and were
6718 even the default) are now avoided.
6720 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6721 with each memory chunk allocated; this is occasionally more helpful
6722 than just having a counter.
6724 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6726 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6730 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6731 which largely parallels "options", but is for changing API behaviour,
6732 whereas "options" are about protocol behaviour.
6733 Initial "mode" flags are:
6735 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6736 a single record has been written.
6737 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6738 retries use the same buffer location.
6739 (But all of the contents must be
6743 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6746 *) Fix problems with no-hmac etc.
6747 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6749 *) New functions RSA_get_default_method(), RSA_set_method() and
6750 RSA_get_method(). These allows replacement of RSA_METHODs without having
6751 to mess around with the internals of an RSA structure.
6754 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6755 Also really enable memory leak checks in openssl.c and in some
6757 [Chad C. Mulligan, Bodo Moeller]
6759 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6760 up the length of negative integers. This has now been simplified to just
6761 store the length when it is first determined and use it later, rather
6762 than trying to keep track of where data is copied and updating it to
6764 [Steve Henson, reported by Brien Wheeler
6765 <bwheeler@authentica-security.com>]
6767 *) Add a new function PKCS7_signatureVerify. This allows the verification
6768 of a PKCS#7 signature but with the signing certificate passed to the
6769 function itself. This contrasts with PKCS7_dataVerify which assumes the
6770 certificate is present in the PKCS#7 structure. This isn't always the
6771 case: certificates can be omitted from a PKCS#7 structure and be
6772 distributed by "out of band" means (such as a certificate database).
6775 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6776 function prototypes in pem.h, also change util/mkdef.pl to add the
6777 necessary function names.
6780 *) mk1mf.pl (used by Windows builds) did not properly read the
6781 options set by Configure in the top level Makefile, and Configure
6782 was not even able to write more than one option correctly.
6783 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6786 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6787 file to be loaded from a BIO or FILE pointer. The BIO version will
6788 for example allow memory BIOs to contain config info.
6791 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6792 Whoever hopes to achieve shared-library compatibility across versions
6793 must use this, not the compile-time macro.
6794 (Exercise 0.9.4: Which is the minimum library version required by
6796 Note: All this applies only to multi-threaded programs, others don't
6800 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6801 through a BIO pair triggered the default case, i.e.
6802 SSLerr(...,SSL_R_UNKNOWN_STATE).
6805 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6806 can use the SSL library even if none of the specific BIOs is
6810 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6811 for the encoded length.
6812 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6814 *) Add initial documentation of the X509V3 functions.
6817 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6818 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6819 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6820 secure PKCS#8 private key format with a high iteration count.
6823 *) Fix determination of Perl interpreter: A perl or perl5
6824 _directory_ in $PATH was also accepted as the interpreter.
6825 [Ralf S. Engelschall]
6827 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6828 wrong with it but it was very old and did things like calling
6829 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6833 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6834 to use the new extension code.
6837 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6838 with macros. This should make it easier to change their form, add extra
6839 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6843 *) Add to configuration table a new entry that can specify an alternative
6844 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6845 according to Mark Crispin <MRC@Panda.COM>.
6849 *) DES CBC did not update the IV. Weird.
6852 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6853 Changing the behaviour of the former might break existing programs --
6854 where IV updating is needed, des_ncbc_encrypt can be used.
6857 *) When bntest is run from "make test" it drives bc to check its
6858 calculations, as well as internally checking them. If an internal check
6859 fails, it needs to cause bc to give a non-zero result or make test carries
6860 on without noticing the failure. Fixed.
6863 *) DES library cleanups.
6866 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6867 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6868 ciphers. NOTE: although the key derivation function has been verified
6869 against some published test vectors it has not been extensively tested
6870 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6874 *) Instead of "mkdir -p", which is not fully portable, use new
6875 Perl script "util/mkdir-p.pl".
6878 *) Rewrite the way password based encryption (PBE) is handled. It used to
6879 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6880 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6881 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6882 the 'parameter' field of the AlgorithmIdentifier is passed to the
6883 underlying key generation function so it must do its own ASN1 parsing.
6884 This has also changed the EVP_PBE_CipherInit() function which now has a
6885 'parameter' argument instead of literal salt and iteration count values
6886 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6889 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6890 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6891 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6892 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6893 value was just used as a "magic string" and not used directly its
6894 value doesn't matter.
6897 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6901 *) "linux-sparc64" configuration (ultrapenguin).
6902 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6903 "linux-sparc" configuration.
6904 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6906 *) config now generates no-xxx options for missing ciphers.
6909 *) Support the EBCDIC character set (work in progress).
6910 File ebcdic.c not yet included because it has a different license.
6911 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6913 *) Support BS2000/OSD-POSIX.
6914 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6916 *) Make callbacks for key generation use void * instead of char *.
6919 *) Make S/MIME samples compile (not yet tested).
6922 *) Additional typesafe stacks.
6925 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6929 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6931 *) New configuration variant "sco5-gcc".
6933 *) Updated some demos.
6934 [Sean O Riordain, Wade Scholine]
6936 *) Add missing BIO_free at exit of pkcs12 application.
6939 *) Fix memory leak in conf.c.
6942 *) Updates for Win32 to assembler version of MD5.
6945 *) Set #! path to perl in apps/der_chop to where we found it
6946 instead of using a fixed path.
6949 *) SHA library changes for irix64-mips4-cc.
6952 *) Improvements for VMS support.
6956 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6958 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6959 This also avoids the problems with SC4.2 and unpatched SC5.
6960 [Andy Polyakov <appro@fy.chalmers.se>]
6962 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6963 These are required because of the typesafe stack would otherwise break
6964 existing code. If old code used a structure member which used to be STACK
6965 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6966 sk_num or sk_value it would produce an error because the num, data members
6967 are not present in STACK_OF. Now it just produces a warning. sk_set
6968 replaces the old method of assigning a value to sk_value
6969 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6970 that does this will no longer work (and should use sk_set instead) but
6971 this could be regarded as a "questionable" behaviour anyway.
6974 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6975 correctly handle encrypted S/MIME data.
6978 *) Change type of various DES function arguments from des_cblock
6979 (which means, in function argument declarations, pointer to char)
6980 to des_cblock * (meaning pointer to array with 8 char elements),
6981 which allows the compiler to do more typechecking; it was like
6982 that back in SSLeay, but with lots of ugly casts.
6984 Introduce new type const_des_cblock.
6987 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6988 problems: find RecipientInfo structure that matches recipient certificate
6989 and initialise the ASN1 structures properly based on passed cipher.
6992 *) Belatedly make the BN tests actually check the results.
6995 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6996 to and from BNs: it was completely broken. New compilation option
6997 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6998 key elements as negative integers.
7001 *) Reorganize and speed up MD5.
7002 [Andy Polyakov <appro@fy.chalmers.se>]
7005 [Richard Levitte <richard@levitte.org>]
7007 *) New option -out to asn1parse to allow the parsed structure to be
7008 output to a file. This is most useful when combined with the -strparse
7009 option to examine the output of things like OCTET STRINGS.
7012 *) Make SSL library a little more fool-proof by not requiring any longer
7013 that SSL_set_{accept,connect}_state be called before
7014 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7015 in many applications because usually everything *appeared* to work as
7016 intended anyway -- now it really works as intended).
7019 *) Move openssl.cnf out of lib/.
7022 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7023 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7024 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7025 [Ralf S. Engelschall]
7027 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7028 handle PKCS#7 enveloped data properly.
7029 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7031 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7032 copying pointers. The cert_st handling is changed by this in
7033 various ways (and thus what used to be known as ctx->default_cert
7034 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7035 any longer when s->cert does not give us what we need).
7036 ssl_cert_instantiate becomes obsolete by this change.
7037 As soon as we've got the new code right (possibly it already is?),
7038 we have solved a couple of bugs of the earlier code where s->cert
7039 was used as if it could not have been shared with other SSL structures.
7041 Note that using the SSL API in certain dirty ways now will result
7042 in different behaviour than observed with earlier library versions:
7043 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7044 does not influence s as it used to.
7046 In order to clean up things more thoroughly, inside SSL_SESSION
7047 we don't use CERT any longer, but a new structure SESS_CERT
7048 that holds per-session data (if available); currently, this is
7049 the peer's certificate chain and, for clients, the server's certificate
7050 and temporary key. CERT holds only those values that can have
7051 meaningful defaults in an SSL_CTX.
7054 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7055 from the internal representation. Various PKCS#7 fixes: remove some
7056 evil casts and set the enc_dig_alg field properly based on the signing
7060 *) Allow PKCS#12 password to be set from the command line or the
7061 environment. Let 'ca' get its config file name from the environment
7062 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7066 *) Allow certificate policies extension to use an IA5STRING for the
7067 organization field. This is contrary to the PKIX definition but
7068 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7072 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7073 without disallowing inline assembler and the like for non-pedantic builds.
7076 *) Support Borland C++ builder.
7077 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7082 *) SHA-1 cleanups and performance enhancements.
7083 [Andy Polyakov <appro@fy.chalmers.se>]
7085 *) Sparc v8plus assembler for the bignum library.
7086 [Andy Polyakov <appro@fy.chalmers.se>]
7088 *) Accept any -xxx and +xxx compiler options in Configure.
7091 *) Update HPUX configuration.
7094 *) Add missing sk_<type>_unshift() function to safestack.h
7095 [Ralf S. Engelschall]
7097 *) New function SSL_CTX_use_certificate_chain_file that sets the
7098 "extra_cert"s in addition to the certificate. (This makes sense
7099 only for "PEM" format files, as chains as a whole are not
7103 *) Support verify_depth from the SSL API.
7104 x509_vfy.c had what can be considered an off-by-one-error:
7105 Its depth (which was not part of the external interface)
7106 was actually counting the number of certificates in a chain;
7107 now it really counts the depth.
7110 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7111 instead of X509err, which often resulted in confusing error
7112 messages since the error codes are not globally unique
7113 (e.g. an alleged error in ssl3_accept when a certificate
7114 didn't match the private key).
7116 *) New function SSL_CTX_set_session_id_context that allows to set a default
7117 value (so that you don't need SSL_set_session_id_context for each
7118 connection using the SSL_CTX).
7121 *) OAEP decoding bug fix.
7124 *) Support INSTALL_PREFIX for package builders, as proposed by
7128 *) New Configure options "threads" and "no-threads". For systems
7129 where the proper compiler options are known (currently Solaris
7130 and Linux), "threads" is the default.
7133 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7136 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7137 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7138 such as /usr/local/bin.
7141 *) "make linux-shared" to build shared libraries.
7142 [Niels Poppe <niels@netbox.org>]
7144 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7147 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7148 extension adding in x509 utility.
7151 *) Remove NOPROTO sections and error code comments.
7154 *) Partial rewrite of the DEF file generator to now parse the ANSI
7158 *) New Configure options --prefix=DIR and --openssldir=DIR.
7161 *) Complete rewrite of the error code script(s). It is all now handled
7162 by one script at the top level which handles error code gathering,
7163 header rewriting and C source file generation. It should be much better
7164 than the old method: it now uses a modified version of Ulf's parser to
7165 read the ANSI prototypes in all header files (thus the old K&R definitions
7166 aren't needed for error creation any more) and do a better job of
7167 translating function codes into names. The old 'ASN1 error code imbedded
7168 in a comment' is no longer necessary and it doesn't use .err files which
7169 have now been deleted. Also the error code call doesn't have to appear all
7170 on one line (which resulted in some large lines...).
7173 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7176 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7177 0 (which usually indicates a closed connection), but continue reading.
7180 *) Fix some race conditions.
7183 *) Add support for CRL distribution points extension. Add Certificate
7184 Policies and CRL distribution points documentation.
7187 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7190 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7191 8 of keying material. Merlin has also confirmed interop with this fix
7192 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7193 [Merlin Hughes <merlin@baltimore.ie>]
7195 *) Fix lots of warnings.
7196 [Richard Levitte <levitte@stacken.kth.se>]
7198 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7199 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7200 [Richard Levitte <levitte@stacken.kth.se>]
7202 *) Fix problems with sizeof(long) == 8.
7203 [Andy Polyakov <appro@fy.chalmers.se>]
7205 *) Change functions to ANSI C.
7208 *) Fix typos in error codes.
7209 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7211 *) Remove defunct assembler files from Configure.
7214 *) SPARC v8 assembler BIGNUM implementation.
7215 [Andy Polyakov <appro@fy.chalmers.se>]
7217 *) Support for Certificate Policies extension: both print and set.
7218 Various additions to support the r2i method this uses.
7221 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7222 return a const string when you are expecting an allocated buffer.
7225 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7226 types DirectoryString and DisplayText.
7229 *) Add code to allow r2i extensions to access the configuration database,
7230 add an LHASH database driver and add several ctx helper functions.
7233 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7234 fail when they extended the size of a BIGNUM.
7237 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7238 support typesafe stack.
7241 *) Fix typo in SSL_[gs]et_options().
7242 [Nils Frostberg <nils@medcom.se>]
7244 *) Delete various functions and files that belonged to the (now obsolete)
7245 old X509V3 handling code.
7248 *) New Configure option "rsaref".
7251 *) Don't auto-generate pem.h.
7254 *) Introduce type-safe ASN.1 SETs.
7257 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7258 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7260 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7261 that links with OpenSSL (well at least cause lots of warnings), but fear
7262 not: the conversion is trivial, and it eliminates loads of evil casts. A
7263 few STACKed things have been converted already. Feel free to convert more.
7264 In the fullness of time, I'll do away with the STACK type altogether.
7267 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7268 specified in <certfile> by updating the entry in the index.txt file.
7269 This way one no longer has to edit the index.txt file manually for
7270 revoking a certificate. The -revoke option does the gory details now.
7271 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7273 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7274 `-text' option at all and this way the `-noout -text' combination was
7275 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7276 [Ralf S. Engelschall]
7278 *) Make sure a corresponding plain text error message exists for the
7279 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7280 verify callback function determined that a certificate was revoked.
7281 [Ralf S. Engelschall]
7283 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7284 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7285 all available cipers including rc5, which was forgotten until now.
7286 In order to let the testing shell script know which algorithms
7287 are available, a new (up to now undocumented) command
7288 "openssl list-cipher-commands" is used.
7291 *) Bugfix: s_client occasionally would sleep in select() when
7292 it should have checked SSL_pending() first.
7295 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7296 the raw DSA values prior to ASN.1 encoding.
7299 *) Tweaks to Configure
7300 [Niels Poppe <niels@netbox.org>]
7302 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7306 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7309 *) New config option to avoid instructions that are illegal on the 80386.
7310 The default code is faster, but requires at least a 486.
7313 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7314 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7315 same as SSL2_VERSION anyway.
7318 *) New "-showcerts" option for s_client.
7321 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7322 application. Various cleanups and fixes.
7325 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7326 modify error routines to work internally. Add error codes and PBE init
7327 to library startup routines.
7330 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7331 packing functions to asn1 and evp. Changed function names and error
7332 codes along the way.
7335 *) PKCS12 integration: and so it begins... First of several patches to
7336 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7337 objects to objects.h
7340 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7341 and display support for Thawte strong extranet extension.
7344 *) Add LinuxPPC support.
7345 [Jeff Dubrule <igor@pobox.org>]
7347 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7348 bn_div_words in alpha.s.
7349 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7351 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7352 OAEP isn't supported when OpenSSL is built with RSAref.
7353 [Ulf Moeller <ulf@fitug.de>]
7355 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7356 so they no longer are missing under -DNOPROTO.
7357 [Soren S. Jorvang <soren@t.dk>]
7360 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7362 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7363 doesn't work when the session is reused. Coming soon!
7366 *) Fix a security hole, that allows sessions to be reused in the wrong
7367 context thus bypassing client cert protection! All software that uses
7368 client certs and session caches in multiple contexts NEEDS PATCHING to
7369 allow session reuse! A fuller solution is in the works.
7370 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7372 *) Some more source tree cleanups (removed obsolete files
7373 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7374 permission on "config" script to be executable) and a fix for the INSTALL
7376 [Ulf Moeller <ulf@fitug.de>]
7378 *) Remove some legacy and erroneous uses of malloc, free instead of
7380 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7382 *) Make rsa_oaep_test return non-zero on error.
7383 [Ulf Moeller <ulf@fitug.de>]
7385 *) Add support for native Solaris shared libraries. Configure
7386 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7387 if someone would make that last step automatic.
7388 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7390 *) ctx_size was not built with the right compiler during "make links". Fixed.
7393 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7394 except NULL ciphers". This means the default cipher list will no longer
7395 enable NULL ciphers. They need to be specifically enabled e.g. with
7396 the string "DEFAULT:eNULL".
7399 *) Fix to RSA private encryption routines: if p < q then it would
7400 occasionally produce an invalid result. This will only happen with
7401 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7404 *) Be less restrictive and allow also `perl util/perlpath.pl
7405 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7406 because this way one can also use an interpreter named `perl5' (which is
7407 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7408 installed as `perl').
7409 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7411 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7412 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7414 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7415 advapi32.lib to Win32 build and change the pem test comparision
7416 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7417 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7418 and crypto/des/ede_cbcm_enc.c.
7421 *) DES quad checksum was broken on big-endian architectures. Fixed.
7424 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7425 Win32 test batch file so it (might) work again. The Win32 test batch file
7426 is horrible: I feel ill....
7429 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7430 in e_os.h. Audit of header files to check ANSI and non ANSI
7431 sections: 10 functions were absent from non ANSI section and not exported
7432 from Windows DLLs. Fixed up libeay.num for new functions.
7435 *) Make `openssl version' output lines consistent.
7436 [Ralf S. Engelschall]
7438 *) Fix Win32 symbol export lists for BIO functions: Added
7439 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7440 to ms/libeay{16,32}.def.
7441 [Ralf S. Engelschall]
7443 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7444 fine under Unix and passes some trivial tests I've now added. But the
7445 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7446 added to make sure no one expects that this stuff really works in the
7447 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7448 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7450 [Ralf S. Engelschall]
7452 *) Fix the generation of two part addresses in perl.
7453 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7455 *) Add config entry for Linux on MIPS.
7456 [John Tobey <jtobey@channel1.com>]
7458 *) Make links whenever Configure is run, unless we are on Windoze.
7461 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7462 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7466 *) Add a useful kludge to allow package maintainers to specify compiler and
7467 other platforms details on the command line without having to patch the
7468 Configure script everytime: One now can use ``perl Configure
7469 <id>:<details>'', i.e. platform ids are allowed to have details appended
7470 to them (seperated by colons). This is treated as there would be a static
7471 pre-configured entry in Configure's %table under key <id> with value
7472 <details> and ``perl Configure <id>'' is called. So, when you want to
7473 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7474 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7475 now, which overrides the FreeBSD-elf entry on-the-fly.
7476 [Ralf S. Engelschall]
7478 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7481 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7482 on the `perl Configure ...' command line. This way one can compile
7483 OpenSSL libraries with Position Independent Code (PIC) which is needed
7484 for linking it into DSOs.
7485 [Ralf S. Engelschall]
7487 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7491 *) Cleaned up the LICENSE document: The official contact for any license
7492 questions now is the OpenSSL core team under openssl-core@openssl.org.
7493 And add a paragraph about the dual-license situation to make sure people
7494 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7495 to the OpenSSL toolkit.
7496 [Ralf S. Engelschall]
7498 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7499 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7500 Additonally cleaned up the `make links' target: Remove unnecessary
7501 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7502 to speed processing and no longer clutter the display with confusing
7503 stuff. Instead only the actually done links are displayed.
7504 [Ralf S. Engelschall]
7506 *) Permit null encryption ciphersuites, used for authentication only. It used
7507 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7508 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7512 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7513 signed attributes when verifying signatures (this would break them),
7514 the detached data encoding was wrong and public keys obtained using
7515 X509_get_pubkey() weren't freed.
7518 *) Add text documentation for the BUFFER functions. Also added a work around
7519 to a Win95 console bug. This was triggered by the password read stuff: the
7520 last character typed gets carried over to the next fread(). If you were
7521 generating a new cert request using 'req' for example then the last
7522 character of the passphrase would be CR which would then enter the first
7526 *) Added the new `Includes OpenSSL Cryptography Software' button as
7527 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7528 button and can be used by applications based on OpenSSL to show the
7529 relationship to the OpenSSL project.
7530 [Ralf S. Engelschall]
7532 *) Remove confusing variables in function signatures in files
7533 ssl/ssl_lib.c and ssl/ssl.h.
7534 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7536 *) Don't install bss_file.c under PREFIX/include/
7537 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7539 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7540 functions that return function pointers and has support for NT specific
7541 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7542 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7543 unsigned to signed types: this was killing the Win32 compile.
7546 *) Add new certificate file to stack functions,
7547 SSL_add_dir_cert_subjects_to_stack() and
7548 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7549 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7550 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7551 This means that Apache-SSL and similar packages don't have to mess around
7552 to add as many CAs as they want to the preferred list.
7555 *) Experiment with doxygen documentation. Currently only partially applied to
7557 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7558 openssl.doxy as the configuration file.
7561 *) Get rid of remaining C++-style comments which strict C compilers hate.
7562 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7564 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7565 compiled in by default: it has problems with large keys.
7568 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7569 DH private keys and/or callback functions which directly correspond to
7570 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7571 is needed for applications which have to configure certificates on a
7572 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7574 For the RSA certificate situation is makes no difference, but
7575 for the DSA certificate situation this fixes the "no shared cipher"
7576 problem where the OpenSSL cipher selection procedure failed because the
7577 temporary keys were not overtaken from the context and the API provided
7578 no way to reconfigure them.
7579 The new functions now let applications reconfigure the stuff and they
7580 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7581 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7582 non-public-API function ssl_cert_instantiate() is used as a helper
7583 function and also to reduce code redundancy inside ssl_rsa.c.
7584 [Ralf S. Engelschall]
7586 *) Move s_server -dcert and -dkey options out of the undocumented feature
7587 area because they are useful for the DSA situation and should be
7588 recognized by the users.
7589 [Ralf S. Engelschall]
7591 *) Fix the cipher decision scheme for export ciphers: the export bits are
7592 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7593 SSL_EXP_MASK. So, the original variable has to be used instead of the
7594 already masked variable.
7595 [Richard Levitte <levitte@stacken.kth.se>]
7597 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7598 [Richard Levitte <levitte@stacken.kth.se>]
7600 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7601 from `int' to `unsigned int' because it's a length and initialized by
7602 EVP_DigestFinal() which expects an `unsigned int *'.
7603 [Richard Levitte <levitte@stacken.kth.se>]
7605 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7606 script. Instead use the usual Shell->Perl transition trick.
7607 [Ralf S. Engelschall]
7609 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7610 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7611 -noout -modulus' as it's already the case for `openssl rsa -noout
7612 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7613 currently the public key is printed (a decision which was already done by
7614 `openssl dsa -modulus' in the past) which serves a similar purpose.
7615 Additionally the NO_RSA no longer completely removes the whole -modulus
7616 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7618 [Ralf S. Engelschall]
7620 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7621 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7622 [Arne Ansper <arne@ats.cyber.ee>]
7624 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7625 to be added. Now both 'req' and 'ca' can use new objects defined in the
7629 *) Add cool BIO that does syslog (or event log on NT).
7630 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7632 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7633 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7634 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7635 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7638 *) Add preliminary config info for new extension code.
7641 *) Make RSA_NO_PADDING really use no padding.
7642 [Ulf Moeller <ulf@fitug.de>]
7644 *) Generate errors when private/public key check is done.
7647 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7648 for some CRL extensions and new objects added.
7651 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7652 key usage extension and fuller support for authority key id.
7655 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7656 padding method for RSA, which is recommended for new applications in PKCS
7657 #1 v2.0 (RFC 2437, October 1998).
7658 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7659 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7660 against Bleichbacher's attack on RSA.
7661 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7664 *) Updates to the new SSL compression code
7665 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7667 *) Fix so that the version number in the master secret, when passed
7668 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7669 (because the server will not accept higher), that the version number
7670 is 0x03,0x01, not 0x03,0x00
7671 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7673 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7674 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7675 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7678 *) Support for RAW extensions where an arbitrary extension can be
7679 created by including its DER encoding. See apps/openssl.cnf for
7683 *) Make sure latest Perl versions don't interpret some generated C array
7684 code as Perl array code in the crypto/err/err_genc.pl script.
7685 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7687 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7688 not many people have the assembler. Various Win32 compilation fixes and
7689 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7693 *) Modify configure script 'Configure' to automatically create crypto/date.h
7694 file under Win32 and also build pem.h from pem.org. New script
7695 util/mkfiles.pl to create the MINFO file on environments that can't do a
7696 'make files': perl util/mkfiles.pl >MINFO should work.
7699 *) Major rework of DES function declarations, in the pursuit of correctness
7700 and purity. As a result, many evil casts evaporated, and some weirdness,
7701 too. You may find this causes warnings in your code. Zapping your evil
7702 casts will probably fix them. Mostly.
7705 *) Fix for a typo in asn1.h. Bug fix to object creation script
7706 obj_dat.pl. It considered a zero in an object definition to mean
7707 "end of object": none of the objects in objects.h have any zeros
7708 so it wasn't spotted.
7709 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7711 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7712 Masking (CBCM). In the absence of test vectors, the best I have been able
7713 to do is check that the decrypt undoes the encrypt, so far. Send me test
7714 vectors if you have them.
7717 *) Correct calculation of key length for export ciphers (too much space was
7718 allocated for null ciphers). This has not been tested!
7721 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7722 message is now correct (it understands "crypto" and "ssl" on its
7723 command line). There is also now an "update" option. This will update
7724 the util/ssleay.num and util/libeay.num files with any new functions.
7726 perl util/mkdef.pl crypto ssl update
7727 it will update them.
7730 *) Overhauled the Perl interface (perl/*):
7731 - ported BN stuff to OpenSSL's different BN library
7732 - made the perl/ source tree CVS-aware
7733 - renamed the package from SSLeay to OpenSSL (the files still contain
7734 their history because I've copied them in the repository)
7735 - removed obsolete files (the test scripts will be replaced
7736 by better Test::Harness variants in the future)
7737 [Ralf S. Engelschall]
7739 *) First cut for a very conservative source tree cleanup:
7740 1. merge various obsolete readme texts into doc/ssleay.txt
7741 where we collect the old documents and readme texts.
7742 2. remove the first part of files where I'm already sure that we no
7743 longer need them because of three reasons: either they are just temporary
7744 files which were left by Eric or they are preserved original files where
7745 I've verified that the diff is also available in the CVS via "cvs diff
7746 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7747 the crypto/md/ stuff).
7748 [Ralf S. Engelschall]
7750 *) More extension code. Incomplete support for subject and issuer alt
7751 name, issuer and authority key id. Change the i2v function parameters
7752 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7753 what that's for :-) Fix to ASN1 macro which messed up
7754 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7757 *) Preliminary support for ENUMERATED type. This is largely copied from the
7761 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7762 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7764 *) Make sure `make rehash' target really finds the `openssl' program.
7765 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7767 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7768 like to hear about it if this slows down other processors.
7771 *) Add CygWin32 platform information to Configure script.
7772 [Alan Batie <batie@aahz.jf.intel.com>]
7774 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7775 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7777 *) New program nseq to manipulate netscape certificate sequences
7780 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7784 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7785 but the BN code had some problems that would cause failures when
7786 doing certificate verification and some other functions.
7787 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7789 *) Add ASN1 and PEM code to support netscape certificate sequences.
7792 *) Add ASN1 and PEM code to support netscape certificate sequences.
7795 *) Add several PKIX and private extended key usage OIDs.
7798 *) Modify the 'ca' program to handle the new extension code. Modify
7799 openssl.cnf for new extension format, add comments.
7802 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7803 and add a sample to openssl.cnf so req -x509 now adds appropriate
7807 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7808 error code, add initial support to X509_print() and x509 application.
7811 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7812 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7813 stuff is currently isolated and isn't even compiled yet.
7816 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7817 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7818 Removed the versions check from X509 routines when loading extensions:
7819 this allows certain broken certificates that don't set the version
7820 properly to be processed.
7823 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7824 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7825 can still be regenerated with "make depend".
7828 *) Spelling mistake in C version of CAST-128.
7829 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7831 *) Changes to the error generation code. The perl script err-code.pl
7832 now reads in the old error codes and retains the old numbers, only
7833 adding new ones if necessary. It also only changes the .err files if new
7834 codes are added. The makefiles have been modified to only insert errors
7835 when needed (to avoid needlessly modifying header files). This is done
7836 by only inserting errors if the .err file is newer than the auto generated
7837 C file. To rebuild all the error codes from scratch (the old behaviour)
7838 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7839 or delete all the .err files.
7842 *) CAST-128 was incorrectly implemented for short keys. The C version has
7843 been fixed, but is untested. The assembler versions are also fixed, but
7844 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7845 to regenerate it if needed.
7846 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7847 Hagino <itojun@kame.net>]
7849 *) File was opened incorrectly in randfile.c.
7850 [Ulf Möller <ulf@fitug.de>]
7852 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7853 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7854 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7855 al: it's just almost always a UTCTime. Note this patch adds new error
7856 codes so do a "make errors" if there are problems.
7859 *) Correct Linux 1 recognition in config.
7860 [Ulf Möller <ulf@fitug.de>]
7862 *) Remove pointless MD5 hash when using DSA keys in ca.
7863 [Anonymous <nobody@replay.com>]
7865 *) Generate an error if given an empty string as a cert directory. Also
7866 generate an error if handed NULL (previously returned 0 to indicate an
7867 error, but didn't set one).
7868 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7870 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7873 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7874 parameters. This was causing a warning which killed off the Win32 compile.
7877 *) Remove C++ style comments from crypto/bn/bn_local.h.
7878 [Neil Costigan <neil.costigan@celocom.com>]
7880 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7881 based on a text string, looking up short and long names and finally
7882 "dot" format. The "dot" format stuff didn't work. Added new function
7883 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7884 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7885 OID is not part of the table.
7888 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7889 X509_LOOKUP_by_alias().
7892 *) Sort openssl functions by name.
7895 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7896 encryption from sample DSA keys (in case anyone is interested the password
7900 *) Make _all_ *_free functions accept a NULL pointer.
7901 [Frans Heymans <fheymans@isaserver.be>]
7903 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7905 [Anonymous <nobody@replay.com>]
7907 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7908 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7910 *) Don't blow it for numeric -newkey arguments to apps/req.
7911 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7913 *) Temp key "for export" tests were wrong in s3_srvr.c.
7914 [Anonymous <nobody@replay.com>]
7916 *) Add prototype for temp key callback functions
7917 SSL_CTX_set_tmp_{rsa,dh}_callback().
7920 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7921 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7924 *) X509_name_add_entry() freed the wrong thing after an error.
7925 [Arne Ansper <arne@ats.cyber.ee>]
7927 *) rsa_eay.c would attempt to free a NULL context.
7928 [Arne Ansper <arne@ats.cyber.ee>]
7930 *) BIO_s_socket() had a broken should_retry() on Windoze.
7931 [Arne Ansper <arne@ats.cyber.ee>]
7933 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7934 [Arne Ansper <arne@ats.cyber.ee>]
7936 *) Make sure the already existing X509_STORE->depth variable is initialized
7937 in X509_STORE_new(), but document the fact that this variable is still
7938 unused in the certificate verification process.
7939 [Ralf S. Engelschall]
7941 *) Fix the various library and apps files to free up pkeys obtained from
7942 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7945 *) Fix reference counting in X509_PUBKEY_get(). This makes
7946 demos/maurice/example2.c work, amongst others, probably.
7947 [Steve Henson and Ben Laurie]
7949 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7950 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7951 are no longer created. This way we have a single and consistent command
7952 line interface `openssl <command>', similar to `cvs <command>'.
7953 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7955 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7956 BIT STRING wrapper always have zero unused bits.
7959 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7962 *) Make the top-level INSTALL documentation easier to understand.
7965 *) Makefiles updated to exit if an error occurs in a sub-directory
7966 make (including if user presses ^C) [Paul Sutton]
7968 *) Make Montgomery context stuff explicit in RSA data structure.
7971 *) Fix build order of pem and err to allow for generated pem.h.
7974 *) Fix renumbering bug in X509_NAME_delete_entry().
7977 *) Enhanced the err-ins.pl script so it makes the error library number
7978 global and can add a library name. This is needed for external ASN1 and
7979 other error libraries.
7982 *) Fixed sk_insert which never worked properly.
7985 *) Fix ASN1 macros so they can handle indefinite length construted
7986 EXPLICIT tags. Some non standard certificates use these: they can now
7990 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7991 into a single doc/ssleay.txt bundle. This way the information is still
7992 preserved but no longer messes up this directory. Now it's new room for
7993 the new set of documenation files.
7994 [Ralf S. Engelschall]
7996 *) SETs were incorrectly DER encoded. This was a major pain, because they
7997 shared code with SEQUENCEs, which aren't coded the same. This means that
7998 almost everything to do with SETs or SEQUENCEs has either changed name or
7999 number of arguments.
8000 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8002 *) Fix test data to work with the above.
8005 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8006 was already fixed by Eric for 0.9.1 it seems.
8007 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8009 *) Autodetect FreeBSD3.
8012 *) Fix various bugs in Configure. This affects the following platforms:
8016 unixware-2.0-pentium
8020 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8021 before they are needed.
8024 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8028 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8030 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8031 changed SSLeay to OpenSSL in version strings.
8032 [Ralf S. Engelschall]
8034 *) Some fixups to the top-level documents.
8037 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8038 because the symlink to include/ was missing.
8039 [Ralf S. Engelschall]
8041 *) Incorporated the popular no-RSA/DSA-only patches
8042 which allow to compile a RSA-free SSLeay.
8043 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8045 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8046 when "ssleay" is still not found.
8047 [Ralf S. Engelschall]
8049 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8050 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8052 *) Updated the README file.
8053 [Ralf S. Engelschall]
8055 *) Added various .cvsignore files in the CVS repository subdirs
8056 to make a "cvs update" really silent.
8057 [Ralf S. Engelschall]
8059 *) Recompiled the error-definition header files and added
8060 missing symbols to the Win32 linker tables.
8061 [Ralf S. Engelschall]
8063 *) Cleaned up the top-level documents;
8064 o new files: CHANGES and LICENSE
8065 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8066 o merged COPYRIGHT into LICENSE
8067 o removed obsolete TODO file
8068 o renamed MICROSOFT to INSTALL.W32
8069 [Ralf S. Engelschall]
8071 *) Removed dummy files from the 0.9.1b source tree:
8072 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8073 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8074 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8075 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8076 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8077 [Ralf S. Engelschall]
8079 *) Added various platform portability fixes.
8082 *) The Genesis of the OpenSSL rpject:
8083 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8084 Young and Tim J. Hudson created while they were working for C2Net until
8086 [The OpenSSL Project]
8089 Changes between 0.9.0b and 0.9.1b [not released]
8091 *) Updated a few CA certificates under certs/
8094 *) Changed some BIGNUM api stuff.
8097 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8098 DGUX x86, Linux Alpha, etc.
8101 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8102 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8106 *) Add -strparse option to asn1pars program which parses nested
8108 [Dr Stephen Henson <shenson@bigfoot.com>]
8110 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8113 *) DSA fix for "ca" program.
8116 *) Added "-genkey" option to "dsaparam" program.
8119 *) Added RIPE MD160 (rmd160) message digest.
8122 *) Added -a (all) option to "ssleay version" command.
8125 *) Added PLATFORM define which is the id given to Configure.
8128 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8131 *) Extended the ASN.1 parser routines.
8134 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8137 *) Added a BN_CTX to the BN library.
8140 *) Fixed the weak key values in DES library
8143 *) Changed API in EVP library for cipher aliases.
8146 *) Added support for RC2/64bit cipher.
8149 *) Converted the lhash library to the crypto/mem.c functions.
8152 *) Added more recognized ASN.1 object ids.
8155 *) Added more RSA padding checks for SSL/TLS.
8158 *) Added BIO proxy/filter functionality.
8161 *) Added extra_certs to SSL_CTX which can be used
8162 send extra CA certificates to the client in the CA cert chain sending
8163 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8166 *) Now Fortezza is denied in the authentication phase because
8167 this is key exchange mechanism is not supported by SSLeay at all.
8170 *) Additional PKCS1 checks.
8173 *) Support the string "TLSv1" for all TLS v1 ciphers.
8176 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8177 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8180 *) Fixed a few memory leaks.
8183 *) Fixed various code and comment typos.
8186 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8187 bytes sent in the client random.
8188 [Edward Bishop <ebishop@spyglass.com>]