5 Changes between 0.9.8h and 0.9.8i [xx XXX xxxx]
7 *) Expand ENGINE to support engine supplied SSL client certificate functions.
9 This work was sponsored by Logica.
12 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows keystores. Support for SSL/TLS client authentication too.
13 Not compiled unless enable-capieng specified to Configure.
15 This work was sponsored by Logica.
18 Changes between 0.9.8g and 0.9.8h [28 May 2008]
20 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
21 handshake which could lead to a cilent crash as found using the
22 Codenomicon TLS test suite (CVE-2008-1672)
23 [Steve Henson, Mark Cox]
25 *) Fix double free in TLS server name extensions which could lead to
26 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
29 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
31 Clear the error queue to ensure that error entries left from
32 older function calls do not interfere with the correct operation.
33 [Lutz Jaenicke, Erik de Castro Lopo]
35 *) Remove root CA certificates of commercial CAs:
37 The OpenSSL project does not recommend any specific CA and does not
38 have any policy with respect to including or excluding any CA.
39 Therefore it does not make any sense to ship an arbitrary selection
40 of root CA certificates with the OpenSSL software.
43 *) RSA OAEP patches to fix two separate invalid memory reads.
44 The first one involves inputs when 'lzero' is greater than
45 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
46 before the beginning of from). The second one involves inputs where
47 the 'db' section contains nothing but zeroes (there is a one-byte
48 invalid read after the end of 'db').
49 [Ivan Nestlerode <inestlerode@us.ibm.com>]
51 *) Partial backport from 0.9.9-dev:
53 Introduce bn_mul_mont (dedicated Montgomery multiplication
54 procedure) as a candidate for BIGNUM assembler implementation.
55 While 0.9.9-dev uses assembler for various architectures, only
56 x86_64 is available by default here in the 0.9.8 branch, and
57 32-bit x86 is available through a compile-time setting.
59 To try the 32-bit x86 assembler implementation, use Configure
60 option "enable-montasm" (which exists only for this backport).
62 As "enable-montasm" for 32-bit x86 disclaims code stability
63 anyway, in this constellation we activate additional code
64 backported from 0.9.9-dev for further performance improvements,
65 namely BN_from_montgomery_word. (To enable this otherwise,
66 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
68 [Andy Polyakov (backport partially by Bodo Moeller)]
70 *) Add TLS session ticket callback. This allows an application to set
71 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
72 values. This is useful for key rollover for example where several key
73 sets may exist with different names.
76 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
77 This was broken until now in 0.9.8 releases, such that the only way
78 a registered ENGINE could be used (assuming it initialises
79 successfully on the host) was to explicitly set it as the default
80 for the relevant algorithms. This is in contradiction with 0.9.7
81 behaviour and the documentation. With this fix, when an ENGINE is
82 registered into a given algorithm's table of implementations, the
83 'uptodate' flag is reset so that auto-discovery will be used next
84 time a new context for that algorithm attempts to select an
86 [Ian Lister (tweaked by Geoff Thorpe)]
88 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
89 implemention in the following ways:
91 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
94 Lack of BER streaming support means one pass streaming processing is
95 only supported if data is detached: setting the streaming flag is
96 ignored for embedded content.
98 CMS support is disabled by default and must be explicitly enabled
99 with the enable-cms configuration option.
102 *) Update the GMP engine glue to do direct copies between BIGNUM and
103 mpz_t when openssl and GMP use the same limb size. Otherwise the
104 existing "conversion via a text string export" trick is still used.
105 [Paul Sheer <paulsheer@gmail.com>]
107 *) Zlib compression BIO. This is a filter BIO which compressed and
108 uncompresses any data passed through it.
111 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
112 RFC3394 compatible AES key wrapping.
115 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
116 sets string data without copying. X509_ALGOR_set0() and
117 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
118 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
119 from an X509_ATTRIBUTE structure optionally checking it occurs only
120 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
124 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
125 to get the expected BN_FLG_CONSTTIME behavior.
126 [Bodo Moeller (Google)]
130 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
131 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
132 - added some more tests to do_tests.pl
133 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
134 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
135 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
136 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
137 - various changes to netware.pl to enable gcc-cross builds on Win32
139 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
140 - various changes to fix missing prototype warnings
141 - fixed x86nasm.pl to create correct asm files for NASM COFF output
142 - added AES, WHIRLPOOL and CPUID assembler code to build files
143 - added missing AES assembler make rules to mk1mf.pl
144 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
145 [Guenter Knauf <eflash@gmx.net>]
147 *) Implement certificate status request TLS extension defined in RFC3546.
148 A client can set the appropriate parameters and receive the encoded
149 OCSP response via a callback. A server can query the supplied parameters
150 and set the encoded OCSP response in the callback. Add simplified examples
151 to s_client and s_server.
154 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
157 + Binary incompatibility of ssl_ctx_st structure
158 + DTLS interoperation with non-compliant servers
159 + Don't call get_session_cb() without proposed session
160 + Fix ia64 assembler code
161 [Andy Polyakov, Steve Henson]
163 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
165 *) DTLS Handshake overhaul. There were longstanding issues with
166 OpenSSL DTLS implementation, which were making it impossible for
167 RFC 4347 compliant client to communicate with OpenSSL server.
168 Unfortunately just fixing these incompatibilities would "cut off"
169 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
170 server keeps tolerating non RFC compliant syntax. The opposite is
171 not true, 0.9.8f client can not communicate with earlier server.
172 This update even addresses CVE-2007-4995.
175 *) Changes to avoid need for function casts in OpenSSL: some compilers
176 (gcc 4.2 and later) reject their use.
177 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
180 *) Add RFC4507 support to OpenSSL. This includes the corrections in
181 RFC4507bis. The encrypted ticket format is an encrypted encoded
182 SSL_SESSION structure, that way new session features are automatically
185 If a client application caches session in an SSL_SESSION structure
186 support is transparent because tickets are now stored in the encoded
189 The SSL_CTX structure automatically generates keys for ticket
190 protection in servers so again support should be possible
191 with no application modification.
193 If a client or server wishes to disable RFC4507 support then the option
194 SSL_OP_NO_TICKET can be set.
196 Add a TLS extension debugging callback to allow the contents of any client
197 or server extensions to be examined.
199 This work was sponsored by Google.
202 *) Add initial support for TLS extensions, specifically for the server_name
203 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
204 have new members for a host name. The SSL data structure has an
205 additional member SSL_CTX *initial_ctx so that new sessions can be
206 stored in that context to allow for session resumption, even after the
207 SSL has been switched to a new SSL_CTX in reaction to a client's
208 server_name extension.
210 New functions (subject to change):
213 SSL_get_servername_type()
216 New CTRL codes and macros (subject to change):
218 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
219 - SSL_CTX_set_tlsext_servername_callback()
220 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
221 - SSL_CTX_set_tlsext_servername_arg()
222 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
224 openssl s_client has a new '-servername ...' option.
226 openssl s_server has new options '-servername_host ...', '-cert2 ...',
227 '-key2 ...', '-servername_fatal' (subject to change). This allows
228 testing the HostName extension for a specific single host name ('-cert'
229 and '-key' remain fallbacks for handshakes without HostName
230 negotiation). If the unrecogninzed_name alert has to be sent, this by
231 default is a warning; it becomes fatal with the '-servername_fatal'
234 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
236 *) Add AES and SSE2 assembly language support to VC++ build.
239 *) Mitigate attack on final subtraction in Montgomery reduction.
242 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
243 (which previously caused an internal error).
246 *) Squeeze another 10% out of IGE mode when in != out.
249 *) AES IGE mode speedup.
250 [Dean Gaudet (Google)]
252 *) Add the Korean symmetric 128-bit cipher SEED (see
253 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
254 add SEED ciphersuites from RFC 4162:
256 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
257 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
258 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
259 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
261 To minimize changes between patchlevels in the OpenSSL 0.9.8
262 series, SEED remains excluded from compilation unless OpenSSL
263 is configured with 'enable-seed'.
266 *) Mitigate branch prediction attacks, which can be practical if a
267 single processor is shared, allowing a spy process to extract
268 information. For detailed background information, see
269 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
270 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
271 and Necessary Software Countermeasures"). The core of the change
272 are new versions BN_div_no_branch() and
273 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
274 respectively, which are slower, but avoid the security-relevant
275 conditional branches. These are automatically called by BN_div()
276 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
277 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
278 remove a conditional branch.
280 BN_FLG_CONSTTIME is the new name for the previous
281 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
282 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
283 in the exponent causes BN_mod_exp_mont() to use the alternative
284 implementation in BN_mod_exp_mont_consttime().) The old name
285 remains as a deprecated alias.
287 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
288 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
289 constant-time implementations for more than just exponentiation.
290 Here too the old name is kept as a deprecated alias.
292 BN_BLINDING_new() will now use BN_dup() for the modulus so that
293 the BN_BLINDING structure gets an independent copy of the
294 modulus. This means that the previous "BIGNUM *m" argument to
295 BN_BLINDING_new() and to BN_BLINDING_create_param() now
296 essentially becomes "const BIGNUM *m", although we can't actually
297 change this in the header file before 0.9.9. It allows
298 RSA_setup_blinding() to use BN_with_flags() on the modulus to
299 enable BN_FLG_CONSTTIME.
301 [Matthew D Wood (Intel Corp)]
303 *) In the SSL/TLS server implementation, be strict about session ID
304 context matching (which matters if an application uses a single
305 external cache for different purposes). Previously,
306 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
307 set. This did ensure strict client verification, but meant that,
308 with applications using a single external cache for quite
309 different requirements, clients could circumvent ciphersuite
310 restrictions for a given session ID context by starting a session
311 in a different context.
314 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
315 a ciphersuite string such as "DEFAULT:RSA" cannot enable
316 authentication-only ciphersuites.
319 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
320 not complete and could lead to a possible single byte overflow
321 (CVE-2007-5135) [Ben Laurie]
323 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
325 *) Since AES128 and AES256 (and similarly Camellia128 and
326 Camellia256) share a single mask bit in the logic of
327 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
328 kludge to work properly if AES128 is available and AES256 isn't
329 (or if Camellia128 is available and Camellia256 isn't).
332 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
333 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
334 When a point or a seed is encoded in a BIT STRING, we need to
335 prevent the removal of trailing zero bits to get the proper DER
336 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
337 of a NamedBitList, for which trailing 0 bits need to be removed.)
340 *) Have SSL/TLS server implementation tolerate "mismatched" record
341 protocol version while receiving ClientHello even if the
342 ClientHello is fragmented. (The server can't insist on the
343 particular protocol version it has chosen before the ServerHello
344 message has informed the client about his choice.)
347 *) Add RFC 3779 support.
348 [Rob Austein for ARIN, Ben Laurie]
350 *) Load error codes if they are not already present instead of using a
351 static variable. This allows them to be cleanly unloaded and reloaded.
352 Improve header file function name parsing.
355 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
356 or CAPABILITY handshake as required by RFCs.
359 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
361 *) Introduce limits to prevent malicious keys being able to
362 cause a denial of service. (CVE-2006-2940)
363 [Steve Henson, Bodo Moeller]
365 *) Fix ASN.1 parsing of certain invalid structures that can result
366 in a denial of service. (CVE-2006-2937) [Steve Henson]
368 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
369 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
371 *) Fix SSL client code which could crash if connecting to a
372 malicious SSLv2 server. (CVE-2006-4343)
373 [Tavis Ormandy and Will Drewry, Google Security Team]
375 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
376 match only those. Before that, "AES256-SHA" would be interpreted
377 as a pattern and match "AES128-SHA" too (since AES128-SHA got
378 the same strength classification in 0.9.7h) as we currently only
379 have a single AES bit in the ciphersuite description bitmap.
380 That change, however, also applied to ciphersuite strings such as
381 "RC4-MD5" that intentionally matched multiple ciphersuites --
382 namely, SSL 2.0 ciphersuites in addition to the more common ones
383 from SSL 3.0/TLS 1.0.
385 So we change the selection algorithm again: Naming an explicit
386 ciphersuite selects this one ciphersuite, and any other similar
387 ciphersuite (same bitmap) from *other* protocol versions.
388 Thus, "RC4-MD5" again will properly select both the SSL 2.0
389 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
391 Since SSL 2.0 does not have any ciphersuites for which the
392 128/256 bit distinction would be relevant, this works for now.
393 The proper fix will be to use different bits for AES128 and
394 AES256, which would have avoided the problems from the beginning;
395 however, bits are scarce, so we can only do this in a new release
396 (not just a patchlevel) when we can change the SSL_CIPHER
397 definition to split the single 'unsigned long mask' bitmap into
398 multiple values to extend the available space.
402 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
404 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
405 (CVE-2006-4339) [Ben Laurie and Google Security Team]
407 *) Add AES IGE and biIGE modes.
410 *) Change the Unix randomness entropy gathering to use poll() when
411 possible instead of select(), since the latter has some
412 undesirable limitations.
413 [Darryl Miles via Richard Levitte and Bodo Moeller]
415 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
416 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
417 cannot be implicitly activated as part of, e.g., the "AES" alias.
418 However, please upgrade to OpenSSL 0.9.9[-dev] for
419 non-experimental use of the ECC ciphersuites to get TLS extension
420 support, which is required for curve and point format negotiation
421 to avoid potential handshake problems.
424 *) Disable rogue ciphersuites:
426 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
427 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
428 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
430 The latter two were purportedly from
431 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
434 Also deactivate the remaining ciphersuites from
435 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
436 unofficial, and the ID has long expired.
439 *) Fix RSA blinding Heisenbug (problems sometimes occured on
440 dual-core machines) and other potential thread-safety issues.
443 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
444 versions), which is now available for royalty-free use
445 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
446 Also, add Camellia TLS ciphersuites from RFC 4132.
448 To minimize changes between patchlevels in the OpenSSL 0.9.8
449 series, Camellia remains excluded from compilation unless OpenSSL
450 is configured with 'enable-camellia'.
453 *) Disable the padding bug check when compression is in use. The padding
454 bug check assumes the first packet is of even length, this is not
455 necessarily true if compresssion is enabled and can result in false
456 positives causing handshake failure. The actual bug test is ancient
457 code so it is hoped that implementations will either have fixed it by
458 now or any which still have the bug do not support compression.
461 Changes between 0.9.8a and 0.9.8b [04 May 2006]
463 *) When applying a cipher rule check to see if string match is an explicit
464 cipher suite and only match that one cipher suite if it is.
467 *) Link in manifests for VC++ if needed.
468 [Austin Ziegler <halostatue@gmail.com>]
470 *) Update support for ECC-based TLS ciphersuites according to
471 draft-ietf-tls-ecc-12.txt with proposed changes (but without
472 TLS extensions, which are supported starting with the 0.9.9
473 branch, not in the OpenSSL 0.9.8 branch).
476 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
477 opaque EVP_CIPHER_CTX handling.
480 *) Fixes and enhancements to zlib compression code. We now only use
481 "zlib1.dll" and use the default __cdecl calling convention on Win32
482 to conform with the standards mentioned here:
483 http://www.zlib.net/DLL_FAQ.txt
484 Static zlib linking now works on Windows and the new --with-zlib-include
485 --with-zlib-lib options to Configure can be used to supply the location
486 of the headers and library. Gracefully handle case where zlib library
490 *) Several fixes and enhancements to the OID generation code. The old code
491 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
492 handle numbers larger than ULONG_MAX, truncated printing and had a
493 non standard OBJ_obj2txt() behaviour.
496 *) Add support for building of engines under engine/ as shared libraries
497 under VC++ build system.
500 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
501 Hopefully, we will not see any false combination of paths any more.
504 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
506 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
507 (part of SSL_OP_ALL). This option used to disable the
508 countermeasure against man-in-the-middle protocol-version
509 rollback in the SSL 2.0 server implementation, which is a bad
510 idea. (CVE-2005-2969)
512 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
513 for Information Security, National Institute of Advanced Industrial
514 Science and Technology [AIST], Japan)]
516 *) Add two function to clear and return the verify parameter flags.
519 *) Keep cipherlists sorted in the source instead of sorting them at
520 runtime, thus removing the need for a lock.
523 *) Avoid some small subgroup attacks in Diffie-Hellman.
524 [Nick Mathewson and Ben Laurie]
526 *) Add functions for well-known primes.
529 *) Extended Windows CE support.
530 [Satoshi Nakamura and Andy Polyakov]
532 *) Initialize SSL_METHOD structures at compile time instead of during
533 runtime, thus removing the need for a lock.
536 *) Make PKCS7_decrypt() work even if no certificate is supplied by
537 attempting to decrypt each encrypted key in turn. Add support to
541 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
543 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
546 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
549 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
550 key into the same file any more.
553 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
556 *) Add -utf8 command line and config file option to 'ca'.
557 [Stefan <stf@udoma.org]
559 *) Removed the macro des_crypt(), as it seems to conflict with some
560 libraries. Use DES_crypt().
563 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
564 involves renaming the source and generated shared-libs for
565 both. The engines will accept the corrected or legacy ids
566 ('ncipher' and '4758_cca' respectively) when binding. NB,
567 this only applies when building 'shared'.
568 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
570 *) Add attribute functions to EVP_PKEY structure. Modify
571 PKCS12_create() to recognize a CSP name attribute and
572 use it. Make -CSP option work again in pkcs12 utility.
575 *) Add new functionality to the bn blinding code:
576 - automatic re-creation of the BN_BLINDING parameters after
577 a fixed number of uses (currently 32)
578 - add new function for parameter creation
579 - introduce flags to control the update behaviour of the
580 BN_BLINDING parameters
581 - hide BN_BLINDING structure
582 Add a second BN_BLINDING slot to the RSA structure to improve
583 performance when a single RSA object is shared among several
587 *) Add support for DTLS.
588 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
590 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
591 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
594 *) Remove buggy and incompletet DH cert support from
595 ssl/ssl_rsa.c and ssl/s3_both.c
598 *) Use SHA-1 instead of MD5 as the default digest algorithm for
599 the apps/openssl applications.
602 *) Compile clean with "-Wall -Wmissing-prototypes
603 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
604 DEBUG_SAFESTACK must also be set.
607 *) Change ./Configure so that certain algorithms can be disabled by default.
608 The new counterpiece to "no-xxx" is "enable-xxx".
610 The patented RC5 and MDC2 algorithms will now be disabled unless
611 "enable-rc5" and "enable-mdc2", respectively, are specified.
613 (IDEA remains enabled despite being patented. This is because IDEA
614 is frequently required for interoperability, and there is no license
615 fee for non-commercial use. As before, "no-idea" can be used to
616 avoid this algorithm.)
620 *) Add processing of proxy certificates (see RFC 3820). This work was
621 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
622 EGEE (Enabling Grids for E-science in Europe).
625 *) RC4 performance overhaul on modern architectures/implementations, such
626 as Intel P4, IA-64 and AMD64.
629 *) New utility extract-section.pl. This can be used specify an alternative
630 section number in a pod file instead of having to treat each file as
631 a separate case in Makefile. This can be done by adding two lines to the
634 =for comment openssl_section:XXX
636 The blank line is mandatory.
640 *) New arguments -certform, -keyform and -pass for s_client and s_server
641 to allow alternative format key and certificate files and passphrase
645 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
646 update associated structures and add various utility functions.
648 Add new policy related verify parameters, include policy checking in
649 standard verify code. Enhance 'smime' application with extra parameters
650 to support policy checking and print out.
653 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
654 Nehemiah processors. These extensions support AES encryption in hardware
655 as well as RNG (though RNG support is currently disabled).
656 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
658 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
661 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
662 [Andy Polyakov and a number of other people]
664 *) Improved PowerPC platform support. Most notably BIGNUM assembler
665 implementation contributed by IBM.
666 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
668 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
669 exponent rather than 'unsigned long'. There is a corresponding change to
670 the new 'rsa_keygen' element of the RSA_METHOD structure.
671 [Jelte Jansen, Geoff Thorpe]
673 *) Functionality for creating the initial serial number file is now
674 moved from CA.pl to the 'ca' utility with a new option -create_serial.
676 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
677 number file to 1, which is bound to cause problems. To avoid
678 the problems while respecting compatibility between different 0.9.7
679 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
680 CA.pl for serial number initialization. With the new release 0.9.8,
681 we can fix the problem directly in the 'ca' utility.)
684 *) Reduced header interdepencies by declaring more opaque objects in
685 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
686 give fewer recursive includes, which could break lazy source code - so
687 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
688 developers should define this symbol when building and using openssl to
689 ensure they track the recommended behaviour, interfaces, [etc], but
690 backwards-compatible behaviour prevails when this isn't defined.
693 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
696 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
697 This will generate a random key of the appropriate length based on the
698 cipher context. The EVP_CIPHER can provide its own random key generation
699 routine to support keys of a specific form. This is used in the des and
700 3des routines to generate a key of the correct parity. Update S/MIME
701 code to use new functions and hence generate correct parity DES keys.
702 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
703 valid (weak or incorrect parity).
706 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
707 as looking them up. This is useful when the verified structure may contain
708 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
709 present unless the new PKCS7_NO_CRL flag is asserted.
712 *) Extend ASN1 oid configuration module. It now additionally accepts the
715 shortName = some long name, 1.2.3.4
718 *) Reimplemented the BN_CTX implementation. There is now no more static
719 limitation on the number of variables it can handle nor the depth of the
720 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
721 information can now expand as required, and rather than having a single
722 static array of bignums, BN_CTX now uses a linked-list of such arrays
723 allowing it to expand on demand whilst maintaining the usefulness of
727 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
728 to allow all RSA operations to function using a single BN_CTX.
731 *) Preliminary support for certificate policy evaluation and checking. This
732 is initially intended to pass the tests outlined in "Conformance Testing
733 of Relying Party Client Certificate Path Processing Logic" v1.07.
736 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
737 remained unused and not that useful. A variety of other little bignum
738 tweaks and fixes have also been made continuing on from the audit (see
742 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
743 associated ASN1, EVP and SSL functions and old ASN1 macros.
746 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
747 and this should never fail. So the return value from the use of
748 BN_set_word() (which can fail due to needless expansion) is now deprecated;
749 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
752 *) BN_CTX_get() should return zero-valued bignums, providing the same
753 initialised value as BN_new().
754 [Geoff Thorpe, suggested by Ulf Möller]
756 *) Support for inhibitAnyPolicy certificate extension.
759 *) An audit of the BIGNUM code is underway, for which debugging code is
760 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
761 is considered valid when processing BIGNUMs, and causes execution to
762 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
763 further steps are taken to deliberately pollute unused data in BIGNUM
764 structures to try and expose faulty code further on. For now, openssl will
765 (in its default mode of operation) continue to tolerate the inconsistent
766 forms that it has tolerated in the past, but authors and packagers should
767 consider trying openssl and their own applications when compiled with
768 these debugging symbols defined. It will help highlight potential bugs in
769 their own code, and will improve the test coverage for OpenSSL itself. At
770 some point, these tighter rules will become openssl's default to improve
771 maintainability, though the assert()s and other overheads will remain only
772 in debugging configurations. See bn.h for more details.
773 [Geoff Thorpe, Nils Larsch, Ulf Möller]
775 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
776 that can only be obtained through BN_CTX_new() (which implicitly
777 initialises it). The presence of this function only made it possible
778 to overwrite an existing structure (and cause memory leaks).
781 *) Because of the callback-based approach for implementing LHASH as a
782 template type, lh_insert() adds opaque objects to hash-tables and
783 lh_doall() or lh_doall_arg() are typically used with a destructor callback
784 to clean up those corresponding objects before destroying the hash table
785 (and losing the object pointers). So some over-zealous constifications in
786 LHASH have been relaxed so that lh_insert() does not take (nor store) the
787 objects as "const" and the lh_doall[_arg] callback wrappers are not
788 prototyped to have "const" restrictions on the object pointers they are
789 given (and so aren't required to cast them away any more).
792 *) The tmdiff.h API was so ugly and minimal that our own timing utility
793 (speed) prefers to use its own implementation. The two implementations
794 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
795 its object type properly exposed (MS_TM) instead of casting to/from "char
796 *". This may still change yet if someone realises MS_TM and "ms_time_***"
797 aren't necessarily the greatest nomenclatures - but this is what was used
798 internally to the implementation so I've used that for now.
801 *) Ensure that deprecated functions do not get compiled when
802 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
803 the self-tests were still using deprecated key-generation functions so
804 these have been updated also.
807 *) Reorganise PKCS#7 code to separate the digest location functionality
808 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
809 New function PKCS7_set_digest() to set the digest type for PKCS#7
810 digestedData type. Add additional code to correctly generate the
811 digestedData type and add support for this type in PKCS7 initialization
815 *) New function PKCS7_set0_type_other() this initializes a PKCS7
816 structure of type "other".
819 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
820 sure the loop does correctly stop and breaking ("division by zero")
821 modulus operations are not performed. The (pre-generated) prime
822 table crypto/bn/bn_prime.h was already correct, but it could not be
823 re-generated on some platforms because of the "division by zero"
824 situation in the script.
825 [Ralf S. Engelschall]
827 *) Update support for ECC-based TLS ciphersuites according to
828 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
829 SHA-1 now is only used for "small" curves (where the
830 representation of a field element takes up to 24 bytes); for
831 larger curves, the field element resulting from ECDH is directly
832 used as premaster secret.
833 [Douglas Stebila (Sun Microsystems Laboratories)]
835 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
836 curve secp160r1 to the tests.
837 [Douglas Stebila (Sun Microsystems Laboratories)]
839 *) Add the possibility to load symbols globally with DSO.
840 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
842 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
843 control of the error stack.
846 *) Add support for STORE in ENGINE.
849 *) Add the STORE type. The intention is to provide a common interface
850 to certificate and key stores, be they simple file-based stores, or
851 HSM-type store, or LDAP stores, or...
852 NOTE: The code is currently UNTESTED and isn't really used anywhere.
855 *) Add a generic structure called OPENSSL_ITEM. This can be used to
856 pass a list of arguments to any function as well as provide a way
857 for a function to pass data back to the caller.
860 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
861 works like BUF_strdup() but can be used to duplicate a portion of
862 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
866 *) Add the function sk_find_ex() which works like sk_find(), but will
867 return an index to an element even if an exact match couldn't be
868 found. The index is guaranteed to point at the element where the
869 searched-for key would be inserted to preserve sorting order.
872 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
873 takes an extra flags argument for optional functionality. Currently,
874 the following flags are defined:
876 OBJ_BSEARCH_VALUE_ON_NOMATCH
877 This one gets OBJ_bsearch_ex() to return a pointer to the first
878 element where the comparing function returns a negative or zero
881 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
882 This one gets OBJ_bsearch_ex() to return a pointer to the first
883 element where the comparing function returns zero. This is useful
884 if there are more than one element where the comparing function
888 *) Make it possible to create self-signed certificates with 'openssl ca'
889 in such a way that the self-signed certificate becomes part of the
890 CA database and uses the same mechanisms for serial number generation
891 as all other certificate signing. The new flag '-selfsign' enables
892 this functionality. Adapt CA.sh and CA.pl.in.
895 *) Add functionality to check the public key of a certificate request
896 against a given private. This is useful to check that a certificate
897 request can be signed by that key (self-signing).
900 *) Make it possible to have multiple active certificates with the same
901 subject in the CA index file. This is done only if the keyword
902 'unique_subject' is set to 'no' in the main CA section (default
903 if 'CA_default') of the configuration file. The value is saved
904 with the database itself in a separate index attribute file,
905 named like the index file with '.attr' appended to the name.
908 *) Generate muti valued AVAs using '+' notation in config files for
912 *) Support for nameConstraints certificate extension.
915 *) Support for policyConstraints certificate extension.
918 *) Support for policyMappings certificate extension.
921 *) Make sure the default DSA_METHOD implementation only uses its
922 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
923 and change its own handlers to be NULL so as to remove unnecessary
924 indirection. This lets alternative implementations fallback to the
925 default implementation more easily.
928 *) Support for directoryName in GeneralName related extensions
932 *) Make it possible to link applications using Makefile.shared.
933 Make that possible even when linking against static libraries!
936 *) Support for single pass processing for S/MIME signing. This now
937 means that S/MIME signing can be done from a pipe, in addition
938 cleartext signing (multipart/signed type) is effectively streaming
939 and the signed data does not need to be all held in memory.
941 This is done with a new flag PKCS7_STREAM. When this flag is set
942 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
943 is done after the data is output (and digests calculated) in
947 *) Add full support for -rpath/-R, both in shared libraries and
948 applications, at least on the platforms where it's known how
952 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
953 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
954 will now compute a table of multiples of the generator that
955 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
956 faster (notably in the case of a single point multiplication,
958 [Nils Larsch, Bodo Moeller]
960 *) IPv6 support for certificate extensions. The various extensions
961 which use the IP:a.b.c.d can now take IPv6 addresses using the
962 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
966 *) Added an ENGINE that implements RSA by performing private key
967 exponentiations with the GMP library. The conversions to and from
968 GMP's mpz_t format aren't optimised nor are any montgomery forms
969 cached, and on x86 it appears OpenSSL's own performance has caught up.
970 However there are likely to be other architectures where GMP could
971 provide a boost. This ENGINE is not built in by default, but it can be
972 specified at Configure time and should be accompanied by the necessary
973 linker additions, eg;
974 ./config -DOPENSSL_USE_GMP -lgmp
977 *) "openssl engine" will not display ENGINE/DSO load failure errors when
978 testing availability of engines with "-t" - the old behaviour is
979 produced by increasing the feature's verbosity with "-tt".
982 *) ECDSA routines: under certain error conditions uninitialized BN objects
983 could be freed. Solution: make sure initialization is performed early
984 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
988 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
989 and DH_METHOD (eg. by ENGINE implementations) to override the normal
990 software implementations. For DSA and DH, parameter generation can
991 also be overriden by providing the appropriate method callbacks.
994 *) Change the "progress" mechanism used in key-generation and
995 primality testing to functions that take a new BN_GENCB pointer in
996 place of callback/argument pairs. The new API functions have "_ex"
997 postfixes and the older functions are reimplemented as wrappers for
998 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
999 declarations of the old functions to help (graceful) attempts to
1000 migrate to the new functions. Also, the new key-generation API
1001 functions operate on a caller-supplied key-structure and return
1002 success/failure rather than returning a key or NULL - this is to
1003 help make "keygen" another member function of RSA_METHOD etc.
1005 Example for using the new callback interface:
1007 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1011 BN_GENCB_set(&my_cb, my_callback, my_arg);
1013 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1014 /* For the meaning of a, b in calls to my_callback(), see the
1015 * documentation of the function that calls the callback.
1016 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1017 * my_callback should return 1 if it wants BN_is_prime_ex()
1018 * to continue, or 0 to stop.
1023 *) Change the ZLIB compression method to be stateful, and make it
1024 available to TLS with the number defined in
1025 draft-ietf-tls-compression-04.txt.
1028 *) Add the ASN.1 structures and functions for CertificatePair, which
1029 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1031 CertificatePair ::= SEQUENCE {
1032 forward [0] Certificate OPTIONAL,
1033 reverse [1] Certificate OPTIONAL,
1034 -- at least one of the pair shall be present -- }
1036 Also implement the PEM functions to read and write certificate
1037 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1039 This needed to be defined, mostly for the sake of the LDAP
1040 attribute crossCertificatePair, but may prove useful elsewhere as
1044 *) Make it possible to inhibit symlinking of shared libraries in
1045 Makefile.shared, for Cygwin's sake.
1048 *) Extend the BIGNUM API by creating a function
1049 void BN_set_negative(BIGNUM *a, int neg);
1050 and a macro that behave like
1051 int BN_is_negative(const BIGNUM *a);
1053 to avoid the need to access 'a->neg' directly in applications.
1056 *) Implement fast modular reduction for pseudo-Mersenne primes
1057 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1058 EC_GROUP_new_curve_GFp() will now automatically use this
1060 [Nils Larsch <nla@trustcenter.de>]
1062 *) Add new lock type (CRYPTO_LOCK_BN).
1065 *) Change the ENGINE framework to automatically load engines
1066 dynamically from specific directories unless they could be
1067 found to already be built in or loaded. Move all the
1068 current engines except for the cryptodev one to a new
1070 The engines in engines/ are built as shared libraries if
1071 the "shared" options was given to ./Configure or ./config.
1072 Otherwise, they are inserted in libcrypto.a.
1073 /usr/local/ssl/engines is the default directory for dynamic
1074 engines, but that can be overriden at configure time through
1075 the usual use of --prefix and/or --openssldir, and at run
1076 time with the environment variable OPENSSL_ENGINES.
1077 [Geoff Thorpe and Richard Levitte]
1079 *) Add Makefile.shared, a helper makefile to build shared
1080 libraries. Addapt Makefile.org.
1083 *) Add version info to Win32 DLLs.
1084 [Peter 'Luna' Runestig" <peter@runestig.com>]
1086 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1087 can be added using this API to created arbitrary PKCS#12
1088 files while avoiding the low level API.
1090 New options to PKCS12_create(), key or cert can be NULL and
1091 will then be omitted from the output file. The encryption
1092 algorithm NIDs can be set to -1 for no encryption, the mac
1093 iteration count can be set to 0 to omit the mac.
1095 Enhance pkcs12 utility by making the -nokeys and -nocerts
1096 options work when creating a PKCS#12 file. New option -nomac
1097 to omit the mac, NONE can be set for an encryption algorithm.
1098 New code is modified to use the enhanced PKCS12_create()
1099 instead of the low level API.
1102 *) Extend ASN1 encoder to support indefinite length constructed
1103 encoding. This can output sequences tags and octet strings in
1104 this form. Modify pk7_asn1.c to support indefinite length
1105 encoding. This is experimental and needs additional code to
1106 be useful, such as an ASN1 bio and some enhanced streaming
1109 Extend template encode functionality so that tagging is passed
1110 down to the template encoder.
1113 *) Let 'openssl req' fail if an argument to '-newkey' is not
1114 recognized instead of using RSA as a default.
1117 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1118 As these are not official, they are not included in "ALL";
1119 the "ECCdraft" ciphersuite group alias can be used to select them.
1120 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1122 *) Add ECDH engine support.
1123 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1125 *) Add ECDH in new directory crypto/ecdh/.
1126 [Douglas Stebila (Sun Microsystems Laboratories)]
1128 *) Let BN_rand_range() abort with an error after 100 iterations
1129 without success (which indicates a broken PRNG).
1132 *) Change BN_mod_sqrt() so that it verifies that the input value
1133 is really the square of the return value. (Previously,
1134 BN_mod_sqrt would show GIGO behaviour.)
1137 *) Add named elliptic curves over binary fields from X9.62, SECG,
1138 and WAP/WTLS; add OIDs that were still missing.
1140 [Sheueling Chang Shantz and Douglas Stebila
1141 (Sun Microsystems Laboratories)]
1143 *) Extend the EC library for elliptic curves over binary fields
1144 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1147 EC_GF2m_simple_method
1151 EC_GROUP_new_curve_GF2m
1152 EC_GROUP_set_curve_GF2m
1153 EC_GROUP_get_curve_GF2m
1154 EC_POINT_set_affine_coordinates_GF2m
1155 EC_POINT_get_affine_coordinates_GF2m
1156 EC_POINT_set_compressed_coordinates_GF2m
1158 Point compression for binary fields is disabled by default for
1159 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1162 As binary polynomials are represented as BIGNUMs, various members
1163 of the EC_GROUP and EC_POINT data structures can be shared
1164 between the implementations for prime fields and binary fields;
1165 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1166 are essentially identical to their ..._GFp counterparts.
1167 (For simplicity, the '..._GFp' prefix has been dropped from
1168 various internal method names.)
1170 An internal 'field_div' method (similar to 'field_mul' and
1171 'field_sqr') has been added; this is used only for binary fields.
1173 [Sheueling Chang Shantz and Douglas Stebila
1174 (Sun Microsystems Laboratories)]
1176 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1177 through methods ('mul', 'precompute_mult').
1179 The generic implementations (now internally called 'ec_wNAF_mul'
1180 and 'ec_wNAF_precomputed_mult') remain the default if these
1181 methods are undefined.
1183 [Sheueling Chang Shantz and Douglas Stebila
1184 (Sun Microsystems Laboratories)]
1186 *) New function EC_GROUP_get_degree, which is defined through
1187 EC_METHOD. For curves over prime fields, this returns the bit
1188 length of the modulus.
1190 [Sheueling Chang Shantz and Douglas Stebila
1191 (Sun Microsystems Laboratories)]
1193 *) New functions EC_GROUP_dup, EC_POINT_dup.
1194 (These simply call ..._new and ..._copy).
1196 [Sheueling Chang Shantz and Douglas Stebila
1197 (Sun Microsystems Laboratories)]
1199 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1200 Polynomials are represented as BIGNUMs (where the sign bit is not
1201 used) in the following functions [macros]:
1204 BN_GF2m_sub [= BN_GF2m_add]
1205 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1206 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1207 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1209 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1210 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1211 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1212 BN_GF2m_cmp [= BN_ucmp]
1214 (Note that only the 'mod' functions are actually for fields GF(2^m).
1215 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1217 For some functions, an the irreducible polynomial defining a
1218 field can be given as an 'unsigned int[]' with strictly
1219 decreasing elements giving the indices of those bits that are set;
1220 i.e., p[] represents the polynomial
1221 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1223 p[0] > p[1] > ... > p[k] = 0.
1224 This applies to the following functions:
1229 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1230 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1232 BN_GF2m_mod_sqrt_arr
1233 BN_GF2m_mod_solve_quad_arr
1237 Conversion can be performed by the following functions:
1242 bntest.c has additional tests for binary polynomial arithmetic.
1244 Two implementations for BN_GF2m_mod_div() are available.
1245 The default algorithm simply uses BN_GF2m_mod_inv() and
1246 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1247 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1248 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1250 [Sheueling Chang Shantz and Douglas Stebila
1251 (Sun Microsystems Laboratories)]
1253 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1254 functionality is disabled at compile-time.
1255 [Douglas Stebila <douglas.stebila@sun.com>]
1257 *) Change default behaviour of 'openssl asn1parse' so that more
1258 information is visible when viewing, e.g., a certificate:
1260 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1261 mode the content of non-printable OCTET STRINGs is output in a
1262 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1263 avoid the appearance of a printable string.
1264 [Nils Larsch <nla@trustcenter.de>]
1266 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1268 EC_GROUP_set_asn1_flag()
1269 EC_GROUP_get_asn1_flag()
1270 EC_GROUP_set_point_conversion_form()
1271 EC_GROUP_get_point_conversion_form()
1272 These control ASN1 encoding details:
1273 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1274 has been set to OPENSSL_EC_NAMED_CURVE.
1275 - Points are encoded in uncompressed form by default; options for
1276 asn1_for are as for point2oct, namely
1277 POINT_CONVERSION_COMPRESSED
1278 POINT_CONVERSION_UNCOMPRESSED
1279 POINT_CONVERSION_HYBRID
1281 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1284 EC_GROUP_get0_seed()
1285 EC_GROUP_get_seed_len()
1286 This is used only for ASN1 purposes (so far).
1287 [Nils Larsch <nla@trustcenter.de>]
1289 *) Add 'field_type' member to EC_METHOD, which holds the NID
1290 of the appropriate field type OID. The new function
1291 EC_METHOD_get_field_type() returns this value.
1292 [Nils Larsch <nla@trustcenter.de>]
1297 EC_POINT_point2hex()
1298 EC_POINT_hex2point()
1299 providing useful interfaces to EC_POINT_point2oct() and
1300 EC_POINT_oct2point().
1301 [Nils Larsch <nla@trustcenter.de>]
1303 *) Change internals of the EC library so that the functions
1304 EC_GROUP_set_generator()
1305 EC_GROUP_get_generator()
1306 EC_GROUP_get_order()
1307 EC_GROUP_get_cofactor()
1308 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1309 to methods, which would lead to unnecessary code duplication when
1310 adding different types of curves.
1311 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1313 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1314 arithmetic, and such that modified wNAFs are generated
1315 (which avoid length expansion in many cases).
1318 *) Add a function EC_GROUP_check_discriminant() (defined via
1319 EC_METHOD) that verifies that the curve discriminant is non-zero.
1321 Add a function EC_GROUP_check() that makes some sanity tests
1322 on a EC_GROUP, its generator and order. This includes
1323 EC_GROUP_check_discriminant().
1324 [Nils Larsch <nla@trustcenter.de>]
1326 *) Add ECDSA in new directory crypto/ecdsa/.
1328 Add applications 'openssl ecparam' and 'openssl ecdsa'
1329 (these are based on 'openssl dsaparam' and 'openssl dsa').
1331 ECDSA support is also included in various other files across the
1332 library. Most notably,
1333 - 'openssl req' now has a '-newkey ecdsa:file' option;
1334 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1335 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1336 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1337 them suitable for ECDSA where domain parameters must be
1338 extracted before the specific public key;
1339 - ECDSA engine support has been added.
1340 [Nils Larsch <nla@trustcenter.de>]
1342 *) Include some named elliptic curves, and add OIDs from X9.62,
1343 SECG, and WAP/WTLS. Each curve can be obtained from the new
1345 EC_GROUP_new_by_curve_name(),
1346 and the list of available named curves can be obtained with
1347 EC_get_builtin_curves().
1348 Also add a 'curve_name' member to EC_GROUP objects, which can be
1350 EC_GROUP_set_curve_name()
1351 EC_GROUP_get_curve_name()
1352 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1354 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1355 was actually never needed) and in BN_mul(). The removal in BN_mul()
1356 required a small change in bn_mul_part_recursive() and the addition
1357 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1358 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1359 bn_sub_words() and bn_add_words() except they take arrays with
1363 Changes between 0.9.7m and 0.9.7n [xx XXX xxxx]
1365 *) In the SSL/TLS server implementation, be strict about session ID
1366 context matching (which matters if an application uses a single
1367 external cache for different purposes). Previously,
1368 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1369 set. This did ensure strict client verification, but meant that,
1370 with applications using a single external cache for quite
1371 different requirements, clients could circumvent ciphersuite
1372 restrictions for a given session ID context by starting a session
1373 in a different context.
1376 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1378 *) Cleanse PEM buffers before freeing them since they may contain
1380 [Benjamin Bennett <ben@psc.edu>]
1382 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1383 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1384 authentication-only ciphersuites.
1387 *) Since AES128 and AES256 share a single mask bit in the logic of
1388 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1389 kludge to work properly if AES128 is available and AES256 isn't.
1392 *) Expand security boundary to match 1.1.1 module.
1395 *) Remove redundant features: hash file source, editing of test vectors
1396 modify fipsld to use external fips_premain.c signature.
1399 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1400 run algorithm test programs.
1403 *) Make algorithm test programs more tolerant of whitespace.
1406 *) Have SSL/TLS server implementation tolerate "mismatched" record
1407 protocol version while receiving ClientHello even if the
1408 ClientHello is fragmented. (The server can't insist on the
1409 particular protocol version it has chosen before the ServerHello
1410 message has informed the client about his choice.)
1413 *) Load error codes if they are not already present instead of using a
1414 static variable. This allows them to be cleanly unloaded and reloaded.
1417 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1419 *) Introduce limits to prevent malicious keys being able to
1420 cause a denial of service. (CVE-2006-2940)
1421 [Steve Henson, Bodo Moeller]
1423 *) Fix ASN.1 parsing of certain invalid structures that can result
1424 in a denial of service. (CVE-2006-2937) [Steve Henson]
1426 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1427 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1429 *) Fix SSL client code which could crash if connecting to a
1430 malicious SSLv2 server. (CVE-2006-4343)
1431 [Tavis Ormandy and Will Drewry, Google Security Team]
1433 *) Change ciphersuite string processing so that an explicit
1434 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1435 will no longer include "AES128-SHA"), and any other similar
1436 ciphersuite (same bitmap) from *other* protocol versions (so that
1437 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1438 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1439 changes from 0.9.8b and 0.9.8d.
1442 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1444 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1445 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1447 *) Change the Unix randomness entropy gathering to use poll() when
1448 possible instead of select(), since the latter has some
1449 undesirable limitations.
1450 [Darryl Miles via Richard Levitte and Bodo Moeller]
1452 *) Disable rogue ciphersuites:
1454 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1455 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1456 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1458 The latter two were purportedly from
1459 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1462 Also deactive the remaining ciphersuites from
1463 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1464 unofficial, and the ID has long expired.
1467 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1468 dual-core machines) and other potential thread-safety issues.
1471 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1473 *) Adapt fipsld and the build system to link against the validated FIPS
1474 module in FIPS mode.
1477 *) Fixes for VC++ 2005 build under Windows.
1480 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1481 from a Windows bash shell such as MSYS. It is autodetected from the
1482 "config" script when run from a VC++ environment. Modify standard VC++
1483 build to use fipscanister.o from the GNU make build.
1486 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1488 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1489 The value now differs depending on if you build for FIPS or not.
1490 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1491 safely run with a non-FIPSed libcrypto, as it may crash because of
1492 the difference induced by this change.
1495 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1497 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1498 (part of SSL_OP_ALL). This option used to disable the
1499 countermeasure against man-in-the-middle protocol-version
1500 rollback in the SSL 2.0 server implementation, which is a bad
1501 idea. (CVE-2005-2969)
1503 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1504 for Information Security, National Institute of Advanced Industrial
1505 Science and Technology [AIST], Japan)]
1507 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1508 mainly for FIPS compliance and not fully integrated at this stage.
1511 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1512 the exponentiation using a fixed-length exponent. (Otherwise,
1513 the information leaked through timing could expose the secret key
1514 after many signatures; cf. Bleichenbacher's attack on DSA with
1518 *) Make a new fixed-window mod_exp implementation the default for
1519 RSA, DSA, and DH private-key operations so that the sequence of
1520 squares and multiplies and the memory access pattern are
1521 independent of the particular secret key. This will mitigate
1522 cache-timing and potential related attacks.
1524 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1525 and this is automatically used by BN_mod_exp_mont() if the new flag
1526 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1527 will use this BN flag for private exponents unless the flag
1528 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1529 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1531 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1533 *) Change the client implementation for SSLv23_method() and
1534 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1535 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1536 (Previously, the SSL 2.0 backwards compatible Client Hello
1537 message format would be used even with SSL_OP_NO_SSLv2.)
1540 *) Add support for smime-type MIME parameter in S/MIME messages which some
1544 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1545 a threadsafe manner. Modify rsa code to use new function and add calls
1546 to dsa and dh code (which had race conditions before).
1549 *) Include the fixed error library code in the C error file definitions
1550 instead of fixing them up at runtime. This keeps the error code
1551 structures constant.
1554 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1556 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1559 *) Fixes for newer kerberos headers. NB: the casts are needed because
1560 the 'length' field is signed on one version and unsigned on another
1561 with no (?) obvious way to tell the difference, without these VC++
1562 complains. Also the "definition" of FAR (blank) is no longer included
1563 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1564 some needed definitions.
1567 *) Undo Cygwin change.
1570 *) Added support for proxy certificates according to RFC 3820.
1571 Because they may be a security thread to unaware applications,
1572 they must be explicitely allowed in run-time. See
1573 docs/HOWTO/proxy_certificates.txt for further information.
1576 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1578 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1579 server and client random values. Previously
1580 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1581 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1583 This change has negligible security impact because:
1585 1. Server and client random values still have 24 bytes of pseudo random
1588 2. Server and client random values are sent in the clear in the initial
1591 3. The master secret is derived using the premaster secret (48 bytes in
1592 size for static RSA ciphersuites) as well as client server and random
1595 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1598 [Stephen Henson, reported by UK NISCC]
1600 *) Use Windows randomness collection on Cygwin.
1603 *) Fix hang in EGD/PRNGD query when communication socket is closed
1604 prematurely by EGD/PRNGD.
1605 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1607 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1610 *) Back-port of selected performance improvements from development
1611 branch, as well as improved support for PowerPC platforms.
1614 *) Add lots of checks for memory allocation failure, error codes to indicate
1615 failure and freeing up memory if a failure occurs.
1616 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1618 *) Add new -passin argument to dgst.
1621 *) Perform some character comparisons of different types in X509_NAME_cmp:
1622 this is needed for some certificates that reencode DNs into UTF8Strings
1623 (in violation of RFC3280) and can't or wont issue name rollover
1627 *) Make an explicit check during certificate validation to see that
1628 the CA setting in each certificate on the chain is correct. As a
1629 side effect always do the following basic checks on extensions,
1630 not just when there's an associated purpose to the check:
1632 - if there is an unhandled critical extension (unless the user
1633 has chosen to ignore this fault)
1634 - if the path length has been exceeded (if one is set at all)
1635 - that certain extensions fit the associated purpose (if one has
1639 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1641 *) Avoid a race condition when CRLs are checked in a multi threaded
1642 environment. This would happen due to the reordering of the revoked
1643 entries during signature checking and serial number lookup. Now the
1644 encoding is cached and the serial number sort performed under a lock.
1645 Add new STACK function sk_is_sorted().
1648 *) Add Delta CRL to the extension code.
1651 *) Various fixes to s3_pkt.c so alerts are sent properly.
1652 [David Holmes <d.holmes@f5.com>]
1654 *) Reduce the chances of duplicate issuer name and serial numbers (in
1655 violation of RFC3280) using the OpenSSL certificate creation utilities.
1656 This is done by creating a random 64 bit value for the initial serial
1657 number when a serial number file is created or when a self signed
1658 certificate is created using 'openssl req -x509'. The initial serial
1659 number file is created using 'openssl x509 -next_serial' in CA.pl
1660 rather than being initialized to 1.
1663 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1665 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1666 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1667 [Joe Orton, Steve Henson]
1669 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1671 [Joe Orton, Steve Henson]
1673 *) Make it possible to have multiple active certificates with the same
1674 subject in the CA index file. This is done only if the keyword
1675 'unique_subject' is set to 'no' in the main CA section (default
1676 if 'CA_default') of the configuration file. The value is saved
1677 with the database itself in a separate index attribute file,
1678 named like the index file with '.attr' appended to the name.
1681 *) X509 verify fixes. Disable broken certificate workarounds when
1682 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1683 keyUsage extension present. Don't accept CRLs with unhandled critical
1684 extensions: since verify currently doesn't process CRL extensions this
1685 rejects a CRL with *any* critical extensions. Add new verify error codes
1689 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1690 A clarification of RFC2560 will require the use of OCTET STRINGs and
1691 some implementations cannot handle the current raw format. Since OpenSSL
1692 copies and compares OCSP nonces as opaque blobs without any attempt at
1693 parsing them this should not create any compatibility issues.
1696 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1697 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1698 this HMAC (and other) operations are several times slower than OpenSSL
1702 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1703 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1705 *) Use the correct content when signing type "other".
1708 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1710 *) Fix various bugs revealed by running the NISCC test suite:
1712 Stop out of bounds reads in the ASN1 code when presented with
1713 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1715 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1717 If verify callback ignores invalid public key errors don't try to check
1718 certificate signature with the NULL public key.
1722 *) New -ignore_err option in ocsp application to stop the server
1723 exiting on the first error in a request.
1726 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1727 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1731 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1732 extra data after the compression methods not only for TLS 1.0
1733 but also for SSL 3.0 (as required by the specification).
1734 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1736 *) Change X509_certificate_type() to mark the key as exported/exportable
1737 when it's 512 *bits* long, not 512 bytes.
1740 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1741 blocks during encryption.
1744 *) Various fixes to base64 BIO and non blocking I/O. On write
1745 flushes were not handled properly if the BIO retried. On read
1746 data was not being buffered properly and had various logic bugs.
1747 This also affects blocking I/O when the data being decoded is a
1751 *) Various S/MIME bugfixes and compatibility changes:
1752 output correct application/pkcs7 MIME type if
1753 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1754 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1755 of files as .eml work). Correctly handle very long lines in MIME
1759 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1761 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1762 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1763 a protocol version number mismatch like a decryption error
1764 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1767 *) Turn on RSA blinding by default in the default implementation
1768 to avoid a timing attack. Applications that don't want it can call
1769 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1770 They would be ill-advised to do so in most cases.
1771 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1773 *) Change RSA blinding code so that it works when the PRNG is not
1774 seeded (in this case, the secret RSA exponent is abused as
1775 an unpredictable seed -- if it is not unpredictable, there
1776 is no point in blinding anyway). Make RSA blinding thread-safe
1777 by remembering the creator's thread ID in rsa->blinding and
1778 having all other threads use local one-time blinding factors
1779 (this requires more computation than sharing rsa->blinding, but
1780 avoids excessive locking; and if an RSA object is not shared
1781 between threads, blinding will still be very fast).
1784 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1785 ENGINE as defaults for all supported algorithms irrespective of
1786 the 'flags' parameter. 'flags' is now honoured, so applications
1787 should make sure they are passing it correctly.
1790 *) Target "mingw" now allows native Windows code to be generated in
1791 the Cygwin environment as well as with the MinGW compiler.
1794 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1796 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1797 via timing by performing a MAC computation even if incorrrect
1798 block cipher padding has been found. This is a countermeasure
1799 against active attacks where the attacker has to distinguish
1800 between bad padding and a MAC verification error. (CVE-2003-0078)
1802 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1803 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1804 Martin Vuagnoux (EPFL, Ilion)]
1806 *) Make the no-err option work as intended. The intention with no-err
1807 is not to have the whole error stack handling routines removed from
1808 libcrypto, it's only intended to remove all the function name and
1809 reason texts, thereby removing some of the footprint that may not
1810 be interesting if those errors aren't displayed anyway.
1812 NOTE: it's still possible for any application or module to have it's
1813 own set of error texts inserted. The routines are there, just not
1814 used by default when no-err is given.
1817 *) Add support for FreeBSD on IA64.
1818 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1820 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1821 Kerberos function mit_des_cbc_cksum(). Before this change,
1822 the value returned by DES_cbc_cksum() was like the one from
1823 mit_des_cbc_cksum(), except the bytes were swapped.
1824 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1826 *) Allow an application to disable the automatic SSL chain building.
1827 Before this a rather primitive chain build was always performed in
1828 ssl3_output_cert_chain(): an application had no way to send the
1829 correct chain if the automatic operation produced an incorrect result.
1831 Now the chain builder is disabled if either:
1833 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1835 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1837 The reasoning behind this is that an application would not want the
1838 auto chain building to take place if extra chain certificates are
1839 present and it might also want a means of sending no additional
1840 certificates (for example the chain has two certificates and the
1844 *) Add the possibility to build without the ENGINE framework.
1845 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1847 *) Under Win32 gmtime() can return NULL: check return value in
1848 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1851 *) DSA routines: under certain error conditions uninitialized BN objects
1852 could be freed. Solution: make sure initialization is performed early
1853 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1854 Nils Larsch <nla@trustcenter.de> via PR#459)
1857 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1858 checked on reconnect on the client side, therefore session resumption
1859 could still fail with a "ssl session id is different" error. This
1860 behaviour is masked when SSL_OP_ALL is used due to
1861 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1862 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1863 followup to PR #377.
1866 *) IA-32 assembler support enhancements: unified ELF targets, support
1867 for SCO/Caldera platforms, fix for Cygwin shared build.
1870 *) Add support for FreeBSD on sparc64. As a consequence, support for
1871 FreeBSD on non-x86 processors is separate from x86 processors on
1872 the config script, much like the NetBSD support.
1873 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1875 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1877 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1880 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1881 code (06) was taken as the first octet of the session ID and the last
1882 octet was ignored consequently. As a result SSLv2 client side session
1883 caching could not have worked due to the session ID mismatch between
1885 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1889 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1890 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1894 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1895 seems that in spite of existing for more than a year, many application
1896 author have done nothing to provide the necessary callbacks, which
1897 means that this particular engine will not work properly anywhere.
1898 This is a very unfortunate situation which forces us, in the name
1899 of usability, to give the hw_ncipher.c a static lock, which is part
1901 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1902 appear in 0.9.8 or later. We EXPECT application authors to have
1903 dealt properly with this when 0.9.8 is released (unless we actually
1904 make such changes in the libcrypto locking code that changes will
1905 have to be made anyway).
1908 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1909 octets have been read, EOF or an error occurs. Without this change
1910 some truncated ASN1 structures will not produce an error.
1913 *) Disable Heimdal support, since it hasn't been fully implemented.
1914 Still give the possibility to force the use of Heimdal, but with
1915 warnings and a request that patches get sent to openssl-dev.
1918 *) Add the VC-CE target, introduce the WINCE sysname, and add
1919 INSTALL.WCE and appropriate conditionals to make it build.
1920 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1922 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1923 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1924 edit numbers of the version.
1925 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1927 *) Introduce safe string copy and catenation functions
1928 (BUF_strlcpy() and BUF_strlcat()).
1929 [Ben Laurie (CHATS) and Richard Levitte]
1931 *) Avoid using fixed-size buffers for one-line DNs.
1932 [Ben Laurie (CHATS)]
1934 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1935 resizing buffers containing secrets, and use where appropriate.
1936 [Ben Laurie (CHATS)]
1938 *) Avoid using fixed size buffers for configuration file location.
1939 [Ben Laurie (CHATS)]
1941 *) Avoid filename truncation for various CA files.
1942 [Ben Laurie (CHATS)]
1944 *) Use sizeof in preference to magic numbers.
1945 [Ben Laurie (CHATS)]
1947 *) Avoid filename truncation in cert requests.
1948 [Ben Laurie (CHATS)]
1950 *) Add assertions to check for (supposedly impossible) buffer
1952 [Ben Laurie (CHATS)]
1954 *) Don't cache truncated DNS entries in the local cache (this could
1955 potentially lead to a spoofing attack).
1956 [Ben Laurie (CHATS)]
1958 *) Fix various buffers to be large enough for hex/decimal
1959 representations in a platform independent manner.
1960 [Ben Laurie (CHATS)]
1962 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1963 resizing buffers containing secrets, and use where appropriate.
1964 [Ben Laurie (CHATS)]
1966 *) Add BIO_indent() to avoid much slightly worrying code to do
1968 [Ben Laurie (CHATS)]
1970 *) Convert sprintf()/BIO_puts() to BIO_printf().
1971 [Ben Laurie (CHATS)]
1973 *) buffer_gets() could terminate with the buffer only half
1975 [Ben Laurie (CHATS)]
1977 *) Add assertions to prevent user-supplied crypto functions from
1978 overflowing internal buffers by having large block sizes, etc.
1979 [Ben Laurie (CHATS)]
1981 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1983 [Ben Laurie (CHATS)]
1985 *) Eliminate unused copy of key in RC4.
1986 [Ben Laurie (CHATS)]
1988 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1989 [Ben Laurie (CHATS)]
1991 *) Fix off-by-one error in EGD path.
1992 [Ben Laurie (CHATS)]
1994 *) If RANDFILE path is too long, ignore instead of truncating.
1995 [Ben Laurie (CHATS)]
1997 *) Eliminate unused and incorrectly sized X.509 structure
1999 [Ben Laurie (CHATS)]
2001 *) Eliminate unused and dangerous function knumber().
2002 [Ben Laurie (CHATS)]
2004 *) Eliminate unused and dangerous structure, KSSL_ERR.
2005 [Ben Laurie (CHATS)]
2007 *) Protect against overlong session ID context length in an encoded
2008 session object. Since these are local, this does not appear to be
2010 [Ben Laurie (CHATS)]
2012 *) Change from security patch (see 0.9.6e below) that did not affect
2013 the 0.9.6 release series:
2015 Remote buffer overflow in SSL3 protocol - an attacker could
2016 supply an oversized master key in Kerberos-enabled versions.
2018 [Ben Laurie (CHATS)]
2020 *) Change the SSL kerb5 codes to match RFC 2712.
2023 *) Make -nameopt work fully for req and add -reqopt switch.
2024 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2026 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2027 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2029 *) Make sure tests can be performed even if the corresponding algorithms
2030 have been removed entirely. This was also the last step to make
2031 OpenSSL compilable with DJGPP under all reasonable conditions.
2032 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2034 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2035 to allow version independent disabling of normally unselected ciphers,
2036 which may be activated as a side-effect of selecting a single cipher.
2038 (E.g., cipher list string "RSA" enables ciphersuites that are left
2039 out of "ALL" because they do not provide symmetric encryption.
2040 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2041 [Lutz Jaenicke, Bodo Moeller]
2043 *) Add appropriate support for separate platform-dependent build
2044 directories. The recommended way to make a platform-dependent
2045 build directory is the following (tested on Linux), maybe with
2048 # Place yourself outside of the OpenSSL source tree. In
2049 # this example, the environment variable OPENSSL_SOURCE
2050 # is assumed to contain the absolute OpenSSL source directory.
2051 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2052 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2053 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2054 mkdir -p `dirname $F`
2055 ln -s $OPENSSL_SOURCE/$F $F
2058 To be absolutely sure not to disturb the source tree, a "make clean"
2059 is a good thing. If it isn't successfull, don't worry about it,
2060 it probably means the source directory is very clean.
2063 *) Make sure any ENGINE control commands make local copies of string
2064 pointers passed to them whenever necessary. Otherwise it is possible
2065 the caller may have overwritten (or deallocated) the original string
2066 data when a later ENGINE operation tries to use the stored values.
2067 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2069 *) Improve diagnostics in file reading and command-line digests.
2070 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2072 *) Add AES modes CFB and OFB to the object database. Correct an
2073 error in AES-CFB decryption.
2076 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2077 allows existing EVP_CIPHER_CTX structures to be reused after
2078 calling EVP_*Final(). This behaviour is used by encryption
2079 BIOs and some applications. This has the side effect that
2080 applications must explicitly clean up cipher contexts with
2081 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2084 *) Check the values of dna and dnb in bn_mul_recursive before calling
2085 bn_mul_comba (a non zero value means the a or b arrays do not contain
2086 n2 elements) and fallback to bn_mul_normal if either is not zero.
2089 *) Fix escaping of non-ASCII characters when using the -subj option
2090 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2093 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2094 form for "surname", serialNumber has no short form.
2095 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2096 therefore remove "mail" short name for "internet 7".
2097 The OID for unique identifiers in X509 certificates is
2098 x500UniqueIdentifier, not uniqueIdentifier.
2099 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2102 *) Add an "init" command to the ENGINE config module and auto initialize
2103 ENGINEs. Without any "init" command the ENGINE will be initialized
2104 after all ctrl commands have been executed on it. If init=1 the
2105 ENGINE is initailized at that point (ctrls before that point are run
2106 on the uninitialized ENGINE and after on the initialized one). If
2107 init=0 then the ENGINE will not be iniatialized at all.
2110 *) Fix the 'app_verify_callback' interface so that the user-defined
2111 argument is actually passed to the callback: In the
2112 SSL_CTX_set_cert_verify_callback() prototype, the callback
2113 declaration has been changed from
2116 int (*cb)(X509_STORE_CTX *,void *);
2117 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2118 i=s->ctx->app_verify_callback(&ctx)
2119 has been changed into
2120 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2122 To update applications using SSL_CTX_set_cert_verify_callback(),
2123 a dummy argument can be added to their callback functions.
2124 [D. K. Smetters <smetters@parc.xerox.com>]
2126 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2127 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2129 *) Add and OPENSSL_LOAD_CONF define which will cause
2130 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2131 This allows older applications to transparently support certain
2132 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2133 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2134 load the config file and OPENSSL_add_all_algorithms_conf() which will
2135 always load it have also been added.
2138 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2139 Adjust NIDs and EVP layer.
2140 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2142 *) Config modules support in openssl utility.
2144 Most commands now load modules from the config file,
2145 though in a few (such as version) this isn't done
2146 because it couldn't be used for anything.
2148 In the case of ca and req the config file used is
2149 the same as the utility itself: that is the -config
2150 command line option can be used to specify an
2154 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2155 use "openssl_conf" if filename is NULL use default openssl config file.
2158 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2159 config section name. Add a new flag to tolerate a missing config file
2160 and move code to CONF_modules_load_file().
2163 *) Support for crypto accelerator cards from Accelerated Encryption
2164 Processing, www.aep.ie. (Use engine 'aep')
2165 The support was copied from 0.9.6c [engine] and adapted/corrected
2166 to work with the new engine framework.
2167 [AEP Inc. and Richard Levitte]
2169 *) Support for SureWare crypto accelerator cards from Baltimore
2170 Technologies. (Use engine 'sureware')
2171 The support was copied from 0.9.6c [engine] and adapted
2172 to work with the new engine framework.
2175 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2176 make the newer ENGINE framework commands for the CHIL engine work.
2177 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2179 *) Make it possible to produce shared libraries on ReliantUNIX.
2180 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2182 *) Add the configuration target debug-linux-ppro.
2183 Make 'openssl rsa' use the general key loading routines
2184 implemented in apps.c, and make those routines able to
2185 handle the key format FORMAT_NETSCAPE and the variant
2187 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2189 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2190 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2192 *) Add -keyform to rsautl, and document -engine.
2193 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2195 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2196 BIO_R_NO_SUCH_FILE error code rather than the generic
2197 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2200 *) Add new functions
2202 ERR_peek_last_error_line
2203 ERR_peek_last_error_line_data.
2204 These are similar to
2207 ERR_peek_error_line_data,
2208 but report on the latest error recorded rather than the first one
2209 still in the error queue.
2210 [Ben Laurie, Bodo Moeller]
2212 *) default_algorithms option in ENGINE config module. This allows things
2214 default_algorithms = ALL
2215 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2218 *) Prelminary ENGINE config module.
2221 *) New experimental application configuration code.
2224 *) Change the AES code to follow the same name structure as all other
2225 symmetric ciphers, and behave the same way. Move everything to
2226 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2227 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2229 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2230 [Ben Laurie and Theo de Raadt]
2232 *) Add option to output public keys in req command.
2233 [Massimiliano Pala madwolf@openca.org]
2235 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2236 (up to about 10% better than before for P-192 and P-224).
2239 *) New functions/macros
2241 SSL_CTX_set_msg_callback(ctx, cb)
2242 SSL_CTX_set_msg_callback_arg(ctx, arg)
2243 SSL_set_msg_callback(ssl, cb)
2244 SSL_set_msg_callback_arg(ssl, arg)
2246 to request calling a callback function
2248 void cb(int write_p, int version, int content_type,
2249 const void *buf, size_t len, SSL *ssl, void *arg)
2251 whenever a protocol message has been completely received
2252 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2253 protocol version according to which the SSL library interprets
2254 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2255 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2256 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2257 specification (change_cipher_spec(20), alert(21), handshake(22)).
2258 'buf' and 'len' point to the actual message, 'ssl' to the
2259 SSL object, and 'arg' is the application-defined value set by
2260 SSL[_CTX]_set_msg_callback_arg().
2262 'openssl s_client' and 'openssl s_server' have new '-msg' options
2263 to enable a callback that displays all protocol messages.
2266 *) Change the shared library support so shared libraries are built as
2267 soon as the corresponding static library is finished, and thereby get
2268 openssl and the test programs linked against the shared library.
2269 This still only happens when the keyword "shard" has been given to
2270 the configuration scripts.
2272 NOTE: shared library support is still an experimental thing, and
2273 backward binary compatibility is still not guaranteed.
2274 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2276 *) Add support for Subject Information Access extension.
2277 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2279 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2280 additional bytes when new memory had to be allocated, not just
2281 when reusing an existing buffer.
2284 *) New command line and configuration option 'utf8' for the req command.
2285 This allows field values to be specified as UTF8 strings.
2288 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2289 runs for the former and machine-readable output for the latter.
2292 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2293 of the e-mail address in the DN (i.e., it will go into a certificate
2294 extension only). The new configuration file option 'email_in_dn = no'
2295 has the same effect.
2296 [Massimiliano Pala madwolf@openca.org]
2298 *) Change all functions with names starting with des_ to be starting
2299 with DES_ instead. Add wrappers that are compatible with libdes,
2300 but are named _ossl_old_des_*. Finally, add macros that map the
2301 des_* symbols to the corresponding _ossl_old_des_* if libdes
2302 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2303 desired, the des_* symbols will be mapped to DES_*, with one
2306 Since we provide two compatibility mappings, the user needs to
2307 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2308 compatibility is desired. The default (i.e., when that macro
2309 isn't defined) is OpenSSL 0.9.6c compatibility.
2311 There are also macros that enable and disable the support of old
2312 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2313 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2314 are defined, the default will apply: to support the old des routines.
2316 In either case, one must include openssl/des.h to get the correct
2317 definitions. Do not try to just include openssl/des_old.h, that
2320 NOTE: This is a major break of an old API into a new one. Software
2321 authors are encouraged to switch to the DES_ style functions. Some
2322 time in the future, des_old.h and the libdes compatibility functions
2323 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2324 default), and then completely removed.
2327 *) Test for certificates which contain unsupported critical extensions.
2328 If such a certificate is found during a verify operation it is
2329 rejected by default: this behaviour can be overridden by either
2330 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2331 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2332 X509_supported_extension() has also been added which returns 1 if a
2333 particular extension is supported.
2336 *) Modify the behaviour of EVP cipher functions in similar way to digests
2337 to retain compatibility with existing code.
2340 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2341 compatibility with existing code. In particular the 'ctx' parameter does
2342 not have to be to be initialized before the call to EVP_DigestInit() and
2343 it is tidied up after a call to EVP_DigestFinal(). New function
2344 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2345 EVP_MD_CTX_copy() changed to not require the destination to be
2346 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2347 requires the destination to be valid.
2349 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2350 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2353 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2354 so that complete 'Handshake' protocol structures are kept in memory
2355 instead of overwriting 'msg_type' and 'length' with 'body' data.
2358 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2359 [Massimo Santin via Richard Levitte]
2361 *) Major restructuring to the underlying ENGINE code. This includes
2362 reduction of linker bloat, separation of pure "ENGINE" manipulation
2363 (initialisation, etc) from functionality dealing with implementations
2364 of specific crypto iterfaces. This change also introduces integrated
2365 support for symmetric ciphers and digest implementations - so ENGINEs
2366 can now accelerate these by providing EVP_CIPHER and EVP_MD
2367 implementations of their own. This is detailed in crypto/engine/README
2368 as it couldn't be adequately described here. However, there are a few
2369 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2370 were changed in the original introduction of ENGINE code have now
2371 reverted back - the hooking from this code to ENGINE is now a good
2372 deal more passive and at run-time, operations deal directly with
2373 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2374 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2375 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2376 they were not being used by the framework as there is no concept of a
2377 BIGNUM_METHOD and they could not be generalised to the new
2378 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2379 ENGINE_cpy() has been removed as it cannot be consistently defined in
2383 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2386 *) Change mkdef.pl to sort symbols that get the same entry number,
2387 and make sure the automatically generated functions ERR_load_*
2388 become part of libeay.num as well.
2391 *) New function SSL_renegotiate_pending(). This returns true once
2392 renegotiation has been requested (either SSL_renegotiate() call
2393 or HelloRequest/ClientHello receveived from the peer) and becomes
2394 false once a handshake has been completed.
2395 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2396 sends a HelloRequest, but does not ensure that a handshake takes
2397 place. SSL_renegotiate_pending() is useful for checking if the
2398 client has followed the request.)
2401 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2402 By default, clients may request session resumption even during
2403 renegotiation (if session ID contexts permit); with this option,
2404 session resumption is possible only in the first handshake.
2406 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2407 more bits available for options that should not be part of
2408 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2411 *) Add some demos for certificate and certificate request creation.
2414 *) Make maximum certificate chain size accepted from the peer application
2415 settable (SSL*_get/set_max_cert_list()), as proposed by
2416 "Douglas E. Engert" <deengert@anl.gov>.
2419 *) Add support for shared libraries for Unixware-7
2420 (Boyd Lynn Gerber <gerberb@zenez.com>).
2423 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2424 be done prior to destruction. Use this to unload error strings from
2425 ENGINEs that load their own error strings. NB: This adds two new API
2426 functions to "get" and "set" this destroy handler in an ENGINE.
2429 *) Alter all existing ENGINE implementations (except "openssl" and
2430 "openbsd") to dynamically instantiate their own error strings. This
2431 makes them more flexible to be built both as statically-linked ENGINEs
2432 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2433 Also, add stub code to each that makes building them as self-contained
2434 shared-libraries easier (see README.ENGINE).
2437 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2438 implementations into applications that are completely implemented in
2439 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2440 commands that can be used to configure what shared-library to load and
2441 to control aspects of the way it is handled. Also, made an update to
2442 the README.ENGINE file that brings its information up-to-date and
2443 provides some information and instructions on the "dynamic" ENGINE
2444 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2447 *) Make it possible to unload ranges of ERR strings with a new
2448 "ERR_unload_strings" function.
2451 *) Add a copy() function to EVP_MD.
2454 *) Make EVP_MD routines take a context pointer instead of just the
2455 md_data void pointer.
2458 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2459 that the digest can only process a single chunk of data
2460 (typically because it is provided by a piece of
2461 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2462 is only going to provide a single chunk of data, and hence the
2463 framework needn't accumulate the data for oneshot drivers.
2466 *) As with "ERR", make it possible to replace the underlying "ex_data"
2467 functions. This change also alters the storage and management of global
2468 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2469 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2470 index counters. The API functions that use this state have been changed
2471 to take a "class_index" rather than pointers to the class's local STACK
2472 and counter, and there is now an API function to dynamically create new
2473 classes. This centralisation allows us to (a) plug a lot of the
2474 thread-safety problems that existed, and (b) makes it possible to clean
2475 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2476 such data would previously have always leaked in application code and
2477 workarounds were in place to make the memory debugging turn a blind eye
2478 to it. Application code that doesn't use this new function will still
2479 leak as before, but their memory debugging output will announce it now
2480 rather than letting it slide.
2482 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2483 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2484 has a return value to indicate success or failure.
2487 *) Make it possible to replace the underlying "ERR" functions such that the
2488 global state (2 LHASH tables and 2 locks) is only used by the "default"
2489 implementation. This change also adds two functions to "get" and "set"
2490 the implementation prior to it being automatically set the first time
2491 any other ERR function takes place. Ie. an application can call "get",
2492 pass the return value to a module it has just loaded, and that module
2493 can call its own "set" function using that value. This means the
2494 module's "ERR" operations will use (and modify) the error state in the
2495 application and not in its own statically linked copy of OpenSSL code.
2498 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2499 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2500 the operation, and provides a more encapsulated way for external code
2501 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2502 to use these functions rather than manually incrementing the counts.
2504 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2507 *) Add EVP test program.
2510 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2513 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2514 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2515 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2516 These allow a CRL to be built without having to access X509_CRL fields
2517 directly. Modify 'ca' application to use new functions.
2520 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2521 bug workarounds. Rollback attack detection is a security feature.
2522 The problem will only arise on OpenSSL servers when TLSv1 is not
2523 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2524 Software authors not wanting to support TLSv1 will have special reasons
2525 for their choice and can explicitly enable this option.
2526 [Bodo Moeller, Lutz Jaenicke]
2528 *) Rationalise EVP so it can be extended: don't include a union of
2529 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2530 (similar to those existing for EVP_CIPHER_CTX).
2535 EVP_MD_CTX_init(&md); /* new function call */
2536 EVP_DigestInit(&md, EVP_sha1());
2537 EVP_DigestUpdate(&md, in, len);
2538 EVP_DigestFinal(&md, out, NULL);
2539 EVP_MD_CTX_cleanup(&md); /* new function call */
2543 *) Make DES key schedule conform to the usual scheme, as well as
2544 correcting its structure. This means that calls to DES functions
2545 now have to pass a pointer to a des_key_schedule instead of a
2546 plain des_key_schedule (which was actually always a pointer
2549 des_key_schedule ks;
2551 des_set_key_checked(..., &ks);
2552 des_ncbc_encrypt(..., &ks, ...);
2554 (Note that a later change renames 'des_...' into 'DES_...'.)
2557 *) Initial reduction of linker bloat: the use of some functions, such as
2558 PEM causes large amounts of unused functions to be linked in due to
2559 poor organisation. For example pem_all.c contains every PEM function
2560 which has a knock on effect of linking in large amounts of (unused)
2561 ASN1 code. Grouping together similar functions and splitting unrelated
2562 functions prevents this.
2565 *) Cleanup of EVP macros.
2568 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2569 correct _ecb suffix.
2572 *) Add initial OCSP responder support to ocsp application. The
2573 revocation information is handled using the text based index
2574 use by the ca application. The responder can either handle
2575 requests generated internally, supplied in files (for example
2576 via a CGI script) or using an internal minimal server.
2579 *) Add configuration choices to get zlib compression for TLS.
2582 *) Changes to Kerberos SSL for RFC 2712 compliance:
2583 1. Implemented real KerberosWrapper, instead of just using
2584 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2585 2. Implemented optional authenticator field of KerberosWrapper.
2587 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2588 and authenticator structs; see crypto/krb5/.
2590 Generalized Kerberos calls to support multiple Kerberos libraries.
2591 [Vern Staats <staatsvr@asc.hpc.mil>,
2592 Jeffrey Altman <jaltman@columbia.edu>
2593 via Richard Levitte]
2595 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2596 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2597 values for each of the key sizes rather than having just
2598 parameters (and 'speed' generating keys each time).
2601 *) Speed up EVP routines.
2604 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2605 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2606 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2607 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2609 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2610 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2611 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2614 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2616 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2619 *) Added the OS2-EMX target.
2620 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2622 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2623 to support NCONF routines in extension code. New function CONF_set_nconf()
2624 to allow functions which take an NCONF to also handle the old LHASH
2625 structure: this means that the old CONF compatible routines can be
2626 retained (in particular wrt extensions) without having to duplicate the
2627 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2630 *) Enhance the general user interface with mechanisms for inner control
2631 and with possibilities to have yes/no kind of prompts.
2634 *) Change all calls to low level digest routines in the library and
2635 applications to use EVP. Add missing calls to HMAC_cleanup() and
2636 don't assume HMAC_CTX can be copied using memcpy().
2637 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2639 *) Add the possibility to control engines through control names but with
2640 arbitrary arguments instead of just a string.
2641 Change the key loaders to take a UI_METHOD instead of a callback
2642 function pointer. NOTE: this breaks binary compatibility with earlier
2643 versions of OpenSSL [engine].
2644 Adapt the nCipher code for these new conditions and add a card insertion
2648 *) Enhance the general user interface with mechanisms to better support
2649 dialog box interfaces, application-defined prompts, the possibility
2650 to use defaults (for example default passwords from somewhere else)
2651 and interrupts/cancellations.
2654 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2655 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2658 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2659 tidy up some unnecessarily weird code in 'sk_new()').
2660 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2662 *) Change the key loading routines for ENGINEs to use the same kind
2663 callback (pem_password_cb) as all other routines that need this
2667 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2668 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2669 than this minimum value is recommended.
2672 *) New random seeder for OpenVMS, using the system process statistics
2673 that are easily reachable.
2676 *) Windows apparently can't transparently handle global
2677 variables defined in DLLs. Initialisations such as:
2679 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2681 wont compile. This is used by the any applications that need to
2682 declare their own ASN1 modules. This was fixed by adding the option
2683 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2684 needed for static libraries under Win32.
2687 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2688 setting of purpose and trust fields. New X509_STORE trust and
2689 purpose functions and tidy up setting in other SSL functions.
2692 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2693 structure. These are inherited by X509_STORE_CTX when it is
2694 initialised. This allows various defaults to be set in the
2695 X509_STORE structure (such as flags for CRL checking and custom
2696 purpose or trust settings) for functions which only use X509_STORE_CTX
2697 internally such as S/MIME.
2699 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2700 trust settings if they are not set in X509_STORE. This allows X509_STORE
2701 purposes and trust (in S/MIME for example) to override any set by default.
2703 Add command line options for CRL checking to smime, s_client and s_server
2707 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2708 are set then the CRL is looked up in the X509_STORE structure and
2709 its validity and signature checked, then if the certificate is found
2710 in the CRL the verify fails with a revoked error.
2712 Various new CRL related callbacks added to X509_STORE_CTX structure.
2714 Command line options added to 'verify' application to support this.
2716 This needs some additional work, such as being able to handle multiple
2717 CRLs with different times, extension based lookup (rather than just
2718 by subject name) and ultimately more complete V2 CRL extension
2722 *) Add a general user interface API (crypto/ui/). This is designed
2723 to replace things like des_read_password and friends (backward
2724 compatibility functions using this new API are provided).
2725 The purpose is to remove prompting functions from the DES code
2726 section as well as provide for prompting through dialog boxes in
2727 a window system and the like.
2730 *) Add "ex_data" support to ENGINE so implementations can add state at a
2731 per-structure level rather than having to store it globally.
2734 *) Make it possible for ENGINE structures to be copied when retrieved by
2735 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2736 This causes the "original" ENGINE structure to act like a template,
2737 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2738 operational state can be localised to each ENGINE structure, despite the
2739 fact they all share the same "methods". New ENGINE structures returned in
2740 this case have no functional references and the return value is the single
2741 structural reference. This matches the single structural reference returned
2742 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2746 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2747 needs to match any other type at all we need to manually clear the
2751 *) Changes to the "openssl engine" utility to include;
2752 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2753 about an ENGINE's available control commands.
2754 - executing control commands from command line arguments using the
2755 '-pre' and '-post' switches. '-post' is only used if '-t' is
2756 specified and the ENGINE is successfully initialised. The syntax for
2757 the individual commands are colon-separated, for example;
2758 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2761 *) New dynamic control command support for ENGINEs. ENGINEs can now
2762 declare their own commands (numbers), names (strings), descriptions,
2763 and input types for run-time discovery by calling applications. A
2764 subset of these commands are implicitly classed as "executable"
2765 depending on their input type, and only these can be invoked through
2766 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2767 can be based on user input, config files, etc). The distinction is
2768 that "executable" commands cannot return anything other than a boolean
2769 result and can only support numeric or string input, whereas some
2770 discoverable commands may only be for direct use through
2771 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2772 pointers, or other custom uses. The "executable" commands are to
2773 support parameterisations of ENGINE behaviour that can be
2774 unambiguously defined by ENGINEs and used consistently across any
2775 OpenSSL-based application. Commands have been added to all the
2776 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2777 control over shared-library paths without source code alterations.
2780 *) Changed all ENGINE implementations to dynamically allocate their
2781 ENGINEs rather than declaring them statically. Apart from this being
2782 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2783 this also allows the implementations to compile without using the
2784 internal engine_int.h header.
2787 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2788 'const' value. Any code that should be able to modify a RAND_METHOD
2789 should already have non-const pointers to it (ie. they should only
2790 modify their own ones).
2793 *) Made a variety of little tweaks to the ENGINE code.
2794 - "atalla" and "ubsec" string definitions were moved from header files
2795 to C code. "nuron" string definitions were placed in variables
2796 rather than hard-coded - allowing parameterisation of these values
2797 later on via ctrl() commands.
2798 - Removed unused "#if 0"'d code.
2799 - Fixed engine list iteration code so it uses ENGINE_free() to release
2800 structural references.
2801 - Constified the RAND_METHOD element of ENGINE structures.
2802 - Constified various get/set functions as appropriate and added
2803 missing functions (including a catch-all ENGINE_cpy that duplicates
2804 all ENGINE values onto a new ENGINE except reference counts/state).
2805 - Removed NULL parameter checks in get/set functions. Setting a method
2806 or function to NULL is a way of cancelling out a previously set
2807 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2808 and doesn't justify the extra error symbols and code.
2809 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2810 flags from engine_int.h to engine.h.
2811 - Changed prototypes for ENGINE handler functions (init(), finish(),
2812 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2815 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2816 to the algorithm using long division. The binary algorithm can be
2817 used only if the modulus is odd. On 32-bit systems, it is faster
2818 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2819 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2820 up to 450 bits. In 64-bit environments, the binary algorithm
2821 appears to be advantageous for much longer moduli; here we use it
2822 for moduli up to 2048 bits.
2825 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2826 could not support the combine flag in choice fields.
2829 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2830 extensions from a certificate request to the certificate.
2833 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2834 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2835 file: this allows the display of the certificate about to be
2836 signed to be customised, to allow certain fields to be included
2837 or excluded and extension details. The old system didn't display
2838 multicharacter strings properly, omitted fields not in the policy
2839 and couldn't display additional details such as extensions.
2842 *) Function EC_POINTs_mul for multiple scalar multiplication
2843 of an arbitrary number of elliptic curve points
2844 \sum scalars[i]*points[i],
2845 optionally including the generator defined for the EC_GROUP:
2846 scalar*generator + \sum scalars[i]*points[i].
2848 EC_POINT_mul is a simple wrapper function for the typical case
2849 that the point list has just one item (besides the optional
2853 *) First EC_METHODs for curves over GF(p):
2855 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2856 operations and provides various method functions that can also
2857 operate with faster implementations of modular arithmetic.
2859 EC_GFp_mont_method() reuses most functions that are part of
2860 EC_GFp_simple_method, but uses Montgomery arithmetic.
2862 [Bodo Moeller; point addition and point doubling
2863 implementation directly derived from source code provided by
2864 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2866 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2867 crypto/ec/ec_lib.c):
2869 Curves are EC_GROUP objects (with an optional group generator)
2870 based on EC_METHODs that are built into the library.
2872 Points are EC_POINT objects based on EC_GROUP objects.
2874 Most of the framework would be able to handle curves over arbitrary
2875 finite fields, but as there are no obvious types for fields other
2876 than GF(p), some functions are limited to that for now.
2879 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2880 that the file contains a complete HTTP response.
2883 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2884 change the def and num file printf format specifier from "%-40sXXX"
2885 to "%-39s XXX". The latter will always guarantee a space after the
2886 field while the former will cause them to run together if the field
2887 is 40 of more characters long.
2890 *) Constify the cipher and digest 'method' functions and structures
2891 and modify related functions to take constant EVP_MD and EVP_CIPHER
2895 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2896 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2899 *) Modify EVP_Digest*() routines so they now return values. Although the
2900 internal software routines can never fail additional hardware versions
2904 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2906 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2907 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2910 ERR_R_NESTED_ASN1_ERROR
2912 ERR_R_MISSING_ASN1_EOS
2913 were 4 .. 9, conflicting with
2914 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2916 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2917 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2919 Add new error code 'ERR_R_INTERNAL_ERROR'.
2922 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2926 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2927 sets the subject name for a new request or supersedes the
2928 subject name in a given request. Formats that can be parsed are
2929 'CN=Some Name, OU=myOU, C=IT'
2931 'CN=Some Name/OU=myOU/C=IT'.
2933 Add options '-batch' and '-verbose' to 'openssl req'.
2934 [Massimiliano Pala <madwolf@hackmasters.net>]
2936 *) Introduce the possibility to access global variables through
2937 functions on platform were that's the best way to handle exporting
2938 global variables in shared libraries. To enable this functionality,
2939 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2940 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2941 is normally done by Configure or something similar).
2943 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2944 in the source file (foo.c) like this:
2946 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2947 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2949 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2950 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2952 OPENSSL_DECLARE_GLOBAL(int,foo);
2953 #define foo OPENSSL_GLOBAL_REF(foo)
2954 OPENSSL_DECLARE_GLOBAL(double,bar);
2955 #define bar OPENSSL_GLOBAL_REF(bar)
2957 The #defines are very important, and therefore so is including the
2958 header file everywhere where the defined globals are used.
2960 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2961 of ASN.1 items, but that structure is a bit different.
2963 The largest change is in util/mkdef.pl which has been enhanced with
2964 better and easier to understand logic to choose which symbols should
2965 go into the Windows .def files as well as a number of fixes and code
2966 cleanup (among others, algorithm keywords are now sorted
2967 lexicographically to avoid constant rewrites).
2970 *) In BN_div() keep a copy of the sign of 'num' before writing the
2971 result to 'rm' because if rm==num the value will be overwritten
2972 and produce the wrong result if 'num' is negative: this caused
2973 problems with BN_mod() and BN_nnmod().
2976 *) Function OCSP_request_verify(). This checks the signature on an
2977 OCSP request and verifies the signer certificate. The signer
2978 certificate is just checked for a generic purpose and OCSP request
2982 *) Add OCSP_check_validity() function to check the validity of OCSP
2983 responses. OCSP responses are prepared in real time and may only
2984 be a few seconds old. Simply checking that the current time lies
2985 between thisUpdate and nextUpdate max reject otherwise valid responses
2986 caused by either OCSP responder or client clock inaccuracy. Instead
2987 we allow thisUpdate and nextUpdate to fall within a certain period of
2988 the current time. The age of the response can also optionally be
2989 checked. Two new options -validity_period and -status_age added to
2993 *) If signature or public key algorithm is unrecognized print out its
2994 OID rather that just UNKNOWN.
2997 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2998 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2999 ID to be generated from the issuer certificate alone which can then be
3000 passed to OCSP_id_issuer_cmp().
3003 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3004 ASN1 modules to export functions returning ASN1_ITEM pointers
3005 instead of the ASN1_ITEM structures themselves. This adds several
3006 new macros which allow the underlying ASN1 function/structure to
3007 be accessed transparently. As a result code should not use ASN1_ITEM
3008 references directly (such as &X509_it) but instead use the relevant
3009 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3010 use of the new ASN1 code on platforms where exporting structures
3011 is problematical (for example in shared libraries) but exporting
3012 functions returning pointers to structures is not.
3015 *) Add support for overriding the generation of SSL/TLS session IDs.
3016 These callbacks can be registered either in an SSL_CTX or per SSL.
3017 The purpose of this is to allow applications to control, if they wish,
3018 the arbitrary values chosen for use as session IDs, particularly as it
3019 can be useful for session caching in multiple-server environments. A
3020 command-line switch for testing this (and any client code that wishes
3021 to use such a feature) has been added to "s_server".
3022 [Geoff Thorpe, Lutz Jaenicke]
3024 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3025 of the form '#if defined(...) || defined(...) || ...' and
3026 '#if !defined(...) && !defined(...) && ...'. This also avoids
3027 the growing number of special cases it was previously handling.
3030 *) Make all configuration macros available for application by making
3031 sure they are available in opensslconf.h, by giving them names starting
3032 with "OPENSSL_" to avoid conflicts with other packages and by making
3033 sure e_os2.h will cover all platform-specific cases together with
3035 Additionally, it is now possible to define configuration/platform-
3036 specific names (called "system identities"). In the C code, these
3037 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3038 macro with the name beginning with "OPENSSL_SYS_", which is determined
3039 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3043 *) New option -set_serial to 'req' and 'x509' this allows the serial
3044 number to use to be specified on the command line. Previously self
3045 signed certificates were hard coded with serial number 0 and the
3046 CA options of 'x509' had to use a serial number in a file which was
3050 *) New options to 'ca' utility to support V2 CRL entry extensions.
3051 Currently CRL reason, invalidity date and hold instruction are
3052 supported. Add new CRL extensions to V3 code and some new objects.
3055 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3056 disable standard block padding (aka PKCS#5 padding) in the EVP
3057 API, which was previously mandatory. This means that the data is
3058 not padded in any way and so the total length much be a multiple
3059 of the block size, otherwise an error occurs.
3062 *) Initial (incomplete) OCSP SSL support.
3065 *) New function OCSP_parse_url(). This splits up a URL into its host,
3066 port and path components: primarily to parse OCSP URLs. New -url
3067 option to ocsp utility.
3070 *) New nonce behavior. The return value of OCSP_check_nonce() now
3071 reflects the various checks performed. Applications can decide
3072 whether to tolerate certain situations such as an absent nonce
3073 in a response when one was present in a request: the ocsp application
3074 just prints out a warning. New function OCSP_add1_basic_nonce()
3075 this is to allow responders to include a nonce in a response even if
3076 the request is nonce-less.
3079 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3080 skipped when using openssl x509 multiple times on a single input file,
3081 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3084 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3085 set string type: to handle setting ASN1_TIME structures. Fix ca
3086 utility to correctly initialize revocation date of CRLs.
3089 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3090 the clients preferred ciphersuites and rather use its own preferences.
3091 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3092 Internet Explorer by ensuring unchanged hash method during stepup.
3093 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3096 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3097 to aes and add a new 'exist' option to print out symbols that don't
3101 *) Additional options to ocsp utility to allow flags to be set and
3102 additional certificates supplied.
3105 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3106 OCSP client a number of certificate to only verify the response
3110 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3111 handle the new API. Currently only ECB, CBC modes supported. Add new
3114 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3115 Encryption Standard (AES) Ciphersuites for Transport Layer
3116 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3117 not enabled by default and were not part of the "ALL" ciphersuite
3118 alias because they were not yet official; they could be
3119 explicitly requested by specifying the "AESdraft" ciphersuite
3120 group alias. In the final release of OpenSSL 0.9.7, the group
3121 alias is called "AES" and is part of "ALL".)
3122 [Ben Laurie, Steve Henson, Bodo Moeller]
3124 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3125 request to response.
3128 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3129 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3130 extract information from a certificate request. OCSP_response_create()
3131 creates a response and optionally adds a basic response structure.
3132 OCSP_basic_add1_status() adds a complete single response to a basic
3133 response and returns the OCSP_SINGLERESP structure just added (to allow
3134 extensions to be included for example). OCSP_basic_add1_cert() adds a
3135 certificate to a basic response and OCSP_basic_sign() signs a basic
3136 response with various flags. New helper functions ASN1_TIME_check()
3137 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3138 (converts ASN1_TIME to GeneralizedTime).
3141 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3142 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3143 structure from a certificate. X509_pubkey_digest() digests the public_key
3144 contents: this is used in various key identifiers.
3147 *) Make sk_sort() tolerate a NULL argument.
3148 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3150 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3151 passed by the function are trusted implicitly. If any of them signed the
3152 response then it is assumed to be valid and is not verified.
3155 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3156 to data. This was previously part of the PKCS7 ASN1 code. This
3157 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3158 [Steve Henson, reported by Kenneth R. Robinette
3159 <support@securenetterm.com>]
3161 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3162 routines: without these tracing memory leaks is very painful.
3163 Fix leaks in PKCS12 and PKCS7 routines.
3166 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3167 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3168 effectively meant GeneralizedTime would never be used. Now it
3169 is initialised to -1 but X509_time_adj() now has to check the value
3170 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3171 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3172 [Steve Henson, reported by Kenneth R. Robinette
3173 <support@securenetterm.com>]
3175 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3176 result in a zero length in the ASN1_INTEGER structure which was
3177 not consistent with the structure when d2i_ASN1_INTEGER() was used
3178 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3179 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3180 where it did not print out a minus for negative ASN1_INTEGER.
3183 *) Add summary printout to ocsp utility. The various functions which
3184 convert status values to strings have been renamed to:
3185 OCSP_response_status_str(), OCSP_cert_status_str() and
3186 OCSP_crl_reason_str() and are no longer static. New options
3187 to verify nonce values and to disable verification. OCSP response
3188 printout format cleaned up.
3191 *) Add additional OCSP certificate checks. These are those specified
3192 in RFC2560. This consists of two separate checks: the CA of the
3193 certificate being checked must either be the OCSP signer certificate
3194 or the issuer of the OCSP signer certificate. In the latter case the
3195 OCSP signer certificate must contain the OCSP signing extended key
3196 usage. This check is performed by attempting to match the OCSP
3197 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3198 in the OCSP_CERTID structures of the response.
3201 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3202 and related routines. This uses the standard OpenSSL certificate
3203 verify routines to perform initial checks (just CA validity) and
3204 to obtain the certificate chain. Then additional checks will be
3205 performed on the chain. Currently the root CA is checked to see
3206 if it is explicitly trusted for OCSP signing. This is used to set
3207 a root CA as a global signing root: that is any certificate that
3208 chains to that CA is an acceptable OCSP signing certificate.
3211 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3212 extensions from a separate configuration file.
3213 As when reading extensions from the main configuration file,
3214 the '-extensions ...' option may be used for specifying the
3216 [Massimiliano Pala <madwolf@comune.modena.it>]
3218 *) New OCSP utility. Allows OCSP requests to be generated or
3219 read. The request can be sent to a responder and the output
3220 parsed, outputed or printed in text form. Not complete yet:
3221 still needs to check the OCSP response validity.
3224 *) New subcommands for 'openssl ca':
3225 'openssl ca -status <serial>' prints the status of the cert with
3226 the given serial number (according to the index file).
3227 'openssl ca -updatedb' updates the expiry status of certificates
3229 [Massimiliano Pala <madwolf@comune.modena.it>]
3231 *) New '-newreq-nodes' command option to CA.pl. This is like
3232 '-newreq', but calls 'openssl req' with the '-nodes' option
3233 so that the resulting key is not encrypted.
3234 [Damien Miller <djm@mindrot.org>]
3236 *) New configuration for the GNU Hurd.
3237 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3239 *) Initial code to implement OCSP basic response verify. This
3240 is currently incomplete. Currently just finds the signer's
3241 certificate and verifies the signature on the response.
3244 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3245 value of OPENSSLDIR. This is available via the new '-d' option
3246 to 'openssl version', and is also included in 'openssl version -a'.
3249 *) Allowing defining memory allocation callbacks that will be given
3250 file name and line number information in additional arguments
3251 (a const char* and an int). The basic functionality remains, as
3252 well as the original possibility to just replace malloc(),
3253 realloc() and free() by functions that do not know about these
3254 additional arguments. To register and find out the current
3255 settings for extended allocation functions, the following
3256 functions are provided:
3258 CRYPTO_set_mem_ex_functions
3259 CRYPTO_set_locked_mem_ex_functions
3260 CRYPTO_get_mem_ex_functions
3261 CRYPTO_get_locked_mem_ex_functions
3263 These work the same way as CRYPTO_set_mem_functions and friends.
3264 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3265 extended allocation function is enabled.
3266 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3267 a conventional allocation function is enabled.
3268 [Richard Levitte, Bodo Moeller]
3270 *) Finish off removing the remaining LHASH function pointer casts.
3271 There should no longer be any prototype-casting required when using
3272 the LHASH abstraction, and any casts that remain are "bugs". See
3273 the callback types and macros at the head of lhash.h for details
3274 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3277 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3278 If /dev/[u]random devices are not available or do not return enough
3279 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3281 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3282 /etc/entropy will be queried once each in this sequence, quering stops
3283 when enough entropy was collected without querying more sockets.
3286 *) Change the Unix RAND_poll() variant to be able to poll several
3287 random devices, as specified by DEVRANDOM, until a sufficient amount
3288 of data has been collected. We spend at most 10 ms on each file
3289 (select timeout) and read in non-blocking mode. DEVRANDOM now
3290 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3291 (previously it was just the string "/dev/urandom"), so on typical
3292 platforms the 10 ms delay will never occur.
3293 Also separate out the Unix variant to its own file, rand_unix.c.
3294 For VMS, there's a currently-empty rand_vms.c.
3297 *) Move OCSP client related routines to ocsp_cl.c. These
3298 provide utility functions which an application needing
3299 to issue a request to an OCSP responder and analyse the
3300 response will typically need: as opposed to those which an
3301 OCSP responder itself would need which will be added later.
3303 OCSP_request_sign() signs an OCSP request with an API similar
3304 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3305 response. OCSP_response_get1_basic() extracts basic response
3306 from response. OCSP_resp_find_status(): finds and extracts status
3307 information from an OCSP_CERTID structure (which will be created
3308 when the request structure is built). These are built from lower
3309 level functions which work on OCSP_SINGLERESP structures but
3310 wont normally be used unless the application wishes to examine
3311 extensions in the OCSP response for example.
3313 Replace nonce routines with a pair of functions.
3314 OCSP_request_add1_nonce() adds a nonce value and optionally
3315 generates a random value. OCSP_check_nonce() checks the
3316 validity of the nonce in an OCSP response.
3319 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3320 This doesn't copy the supplied OCSP_CERTID and avoids the
3321 need to free up the newly created id. Change return type
3322 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3323 This can then be used to add extensions to the request.
3324 Deleted OCSP_request_new(), since most of its functionality
3325 is now in OCSP_REQUEST_new() (and the case insensitive name
3326 clash) apart from the ability to set the request name which
3327 will be added elsewhere.
3330 *) Update OCSP API. Remove obsolete extensions argument from
3331 various functions. Extensions are now handled using the new
3332 OCSP extension code. New simple OCSP HTTP function which
3333 can be used to send requests and parse the response.
3336 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3337 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3338 uses the special reorder version of SET OF to sort the attributes
3339 and reorder them to match the encoded order. This resolves a long
3340 standing problem: a verify on a PKCS7 structure just after signing
3341 it used to fail because the attribute order did not match the
3342 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3343 it uses the received order. This is necessary to tolerate some broken
3344 software that does not order SET OF. This is handled by encoding
3345 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3346 to produce the required SET OF.
3349 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3350 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3351 files to get correct declarations of the ASN.1 item variables.
3354 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3355 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3356 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3357 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3358 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3359 ASN1_ITEM and no wrapper functions.
3362 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3363 replace the old function pointer based I/O routines. Change most of
3364 the *_d2i_bio() and *_d2i_fp() functions to use these.
3367 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3368 lines, recognice more "algorithms" that can be deselected, and make
3369 it complain about algorithm deselection that isn't recognised.
3372 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3373 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3374 to use new functions. Add NO_ASN1_OLD which can be set to remove
3375 some old style ASN1 functions: this can be used to determine if old
3376 code will still work when these eventually go away.
3379 *) New extension functions for OCSP structures, these follow the
3380 same conventions as certificates and CRLs.
3383 *) New function X509V3_add1_i2d(). This automatically encodes and
3384 adds an extension. Its behaviour can be customised with various
3385 flags to append, replace or delete. Various wrappers added for
3386 certifcates and CRLs.
3389 *) Fix to avoid calling the underlying ASN1 print routine when
3390 an extension cannot be parsed. Correct a typo in the
3391 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3394 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3395 entries for variables.
3398 *) Add functionality to apps/openssl.c for detecting locking
3399 problems: As the program is single-threaded, all we have
3400 to do is register a locking callback using an array for
3401 storing which locks are currently held by the program.
3404 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3405 SSL_get_ex_data_X509_STORE_idx(), which is used in
3406 ssl_verify_cert_chain() and thus can be called at any time
3407 during TLS/SSL handshakes so that thread-safety is essential.
3408 Unfortunately, the ex_data design is not at all suited
3409 for multi-threaded use, so it probably should be abolished.
3412 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3413 [Broadcom, tweaked and integrated by Geoff Thorpe]
3415 *) Move common extension printing code to new function
3416 X509V3_print_extensions(). Reorganise OCSP print routines and
3417 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3420 *) New function X509_signature_print() to remove duplication in some
3424 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3425 set (this was treated exactly the same as SET OF previously). This
3426 is used to reorder the STACK representing the structure to match the
3427 encoding. This will be used to get round a problem where a PKCS7
3428 structure which was signed could not be verified because the STACK
3429 order did not reflect the encoded order.
3432 *) Reimplement the OCSP ASN1 module using the new code.
3435 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3436 for its ASN1 operations. The old style function pointers still exist
3437 for now but they will eventually go away.
3440 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3441 completely replaces the old ASN1 functionality with a table driven
3442 encoder and decoder which interprets an ASN1_ITEM structure describing
3443 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3444 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3445 has also been converted to the new form.
3448 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3449 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3450 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3451 for negative moduli.
3454 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3455 of not touching the result's sign bit.
3458 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3462 *) Changed the LHASH code to use prototypes for callbacks, and created
3463 macros to declare and implement thin (optionally static) functions
3464 that provide type-safety and avoid function pointer casting for the
3465 type-specific callbacks.
3468 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3470 [Veers Staats <staatsvr@asc.hpc.mil>,
3471 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3473 *) Reformat the FAQ so the different questions and answers can be divided
3474 in sections depending on the subject.
3477 *) Have the zlib compression code load ZLIB.DLL dynamically under
3481 *) New function BN_mod_sqrt for computing square roots modulo a prime
3482 (using the probabilistic Tonelli-Shanks algorithm unless
3483 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3484 be handled deterministically).
3485 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3487 *) Make BN_mod_inverse faster by explicitly handling small quotients
3488 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3489 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3492 *) New function BN_kronecker.
3495 *) Fix BN_gcd so that it works on negative inputs; the result is
3496 positive unless both parameters are zero.
3497 Previously something reasonably close to an infinite loop was
3498 possible because numbers could be growing instead of shrinking
3499 in the implementation of Euclid's algorithm.
3502 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3503 sign of the number in question.
3505 Fix BN_is_word(a,w) to work correctly for w == 0.
3507 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3508 because its test if the absolute value of 'a' equals 'w'.
3509 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3510 it exists mostly for use in the implementations of BN_is_zero(),
3511 BN_is_one(), and BN_is_word().
3514 *) New function BN_swap.
3517 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3518 the exponentiation functions are more likely to produce reasonable
3519 results on negative inputs.
3522 *) Change BN_mod_mul so that the result is always non-negative.
3523 Previously, it could be negative if one of the factors was negative;
3524 I don't think anyone really wanted that behaviour.
3527 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3528 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3529 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3530 and add new functions:
3539 BN_mod_lshift1_quick
3543 These functions always generate non-negative results.
3545 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3546 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3548 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3549 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3550 be reduced modulo m.
3551 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3554 The following entry accidentily appeared in the CHANGES file
3555 distributed with OpenSSL 0.9.7. The modifications described in
3556 it do *not* apply to OpenSSL 0.9.7.
3558 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3559 was actually never needed) and in BN_mul(). The removal in BN_mul()
3560 required a small change in bn_mul_part_recursive() and the addition
3561 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3562 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3563 bn_sub_words() and bn_add_words() except they take arrays with
3568 *) In 'openssl passwd', verify passwords read from the terminal
3569 unless the '-salt' option is used (which usually means that
3570 verification would just waste user's time since the resulting
3571 hash is going to be compared with some given password hash)
3572 or the new '-noverify' option is used.
3574 This is an incompatible change, but it does not affect
3575 non-interactive use of 'openssl passwd' (passwords on the command
3576 line, '-stdin' option, '-in ...' option) and thus should not
3580 *) Remove all references to RSAref, since there's no more need for it.
3583 *) Make DSO load along a path given through an environment variable
3584 (SHLIB_PATH) with shl_load().
3587 *) Constify the ENGINE code as a result of BIGNUM constification.
3588 Also constify the RSA code and most things related to it. In a
3589 few places, most notable in the depth of the ASN.1 code, ugly
3590 casts back to non-const were required (to be solved at a later
3594 *) Make it so the openssl application has all engines loaded by default.
3597 *) Constify the BIGNUM routines a little more.
3600 *) Add the following functions:
3602 ENGINE_load_cswift()
3604 ENGINE_load_atalla()
3606 ENGINE_load_builtin_engines()
3608 That way, an application can itself choose if external engines that
3609 are built-in in OpenSSL shall ever be used or not. The benefit is
3610 that applications won't have to be linked with libdl or other dso
3611 libraries unless it's really needed.
3613 Changed 'openssl engine' to load all engines on demand.
3614 Changed the engine header files to avoid the duplication of some
3615 declarations (they differed!).
3618 *) 'openssl engine' can now list capabilities.
3621 *) Better error reporting in 'openssl engine'.
3624 *) Never call load_dh_param(NULL) in s_server.
3627 *) Add engine application. It can currently list engines by name and
3628 identity, and test if they are actually available.
3631 *) Improve RPM specification file by forcing symbolic linking and making
3632 sure the installed documentation is also owned by root.root.
3633 [Damien Miller <djm@mindrot.org>]
3635 *) Give the OpenSSL applications more possibilities to make use of
3636 keys (public as well as private) handled by engines.
3639 *) Add OCSP code that comes from CertCo.
3642 *) Add VMS support for the Rijndael code.
3645 *) Added untested support for Nuron crypto accelerator.
3648 *) Add support for external cryptographic devices. This code was
3649 previously distributed separately as the "engine" branch.
3650 [Geoff Thorpe, Richard Levitte]
3652 *) Rework the filename-translation in the DSO code. It is now possible to
3653 have far greater control over how a "name" is turned into a filename
3654 depending on the operating environment and any oddities about the
3655 different shared library filenames on each system.
3658 *) Support threads on FreeBSD-elf in Configure.
3661 *) Fix for SHA1 assembly problem with MASM: it produces
3662 warnings about corrupt line number information when assembling
3663 with debugging information. This is caused by the overlapping
3665 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3668 NCONF_get_number() has no error checking at all. As a replacement,
3669 NCONF_get_number_e() is defined (_e for "error checking") and is
3670 promoted strongly. The old NCONF_get_number is kept around for
3671 binary backward compatibility.
3672 Make it possible for methods to load from something other than a BIO,
3673 by providing a function pointer that is given a name instead of a BIO.
3674 For example, this could be used to load configuration data from an
3678 *) Fix for non blocking accept BIOs. Added new I/O special reason
3679 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3680 with non blocking I/O was not possible because no retry code was
3681 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3685 *) Added the beginnings of Rijndael support.
3688 *) Fix for bug in DirectoryString mask setting. Add support for
3689 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3690 to allow certificate printing to more controllable, additional
3691 'certopt' option to 'x509' to allow new printing options to be
3695 *) Clean old EAY MD5 hack from e_os.h.
3698 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3700 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3701 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3702 [Joe Orton, Steve Henson]
3704 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3706 *) Fix additional bug revealed by the NISCC test suite:
3708 Stop bug triggering large recursion when presented with
3709 certain ASN.1 tags (CVE-2003-0851)
3712 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3714 *) Fix various bugs revealed by running the NISCC test suite:
3716 Stop out of bounds reads in the ASN1 code when presented with
3717 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3719 If verify callback ignores invalid public key errors don't try to check
3720 certificate signature with the NULL public key.
3724 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3725 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3729 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3730 extra data after the compression methods not only for TLS 1.0
3731 but also for SSL 3.0 (as required by the specification).
3732 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3734 *) Change X509_certificate_type() to mark the key as exported/exportable
3735 when it's 512 *bits* long, not 512 bytes.
3738 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3740 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3741 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3742 a protocol version number mismatch like a decryption error
3743 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3746 *) Turn on RSA blinding by default in the default implementation
3747 to avoid a timing attack. Applications that don't want it can call
3748 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3749 They would be ill-advised to do so in most cases.
3750 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3752 *) Change RSA blinding code so that it works when the PRNG is not
3753 seeded (in this case, the secret RSA exponent is abused as
3754 an unpredictable seed -- if it is not unpredictable, there
3755 is no point in blinding anyway). Make RSA blinding thread-safe
3756 by remembering the creator's thread ID in rsa->blinding and
3757 having all other threads use local one-time blinding factors
3758 (this requires more computation than sharing rsa->blinding, but
3759 avoids excessive locking; and if an RSA object is not shared
3760 between threads, blinding will still be very fast).
3763 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3765 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3766 via timing by performing a MAC computation even if incorrrect
3767 block cipher padding has been found. This is a countermeasure
3768 against active attacks where the attacker has to distinguish
3769 between bad padding and a MAC verification error. (CVE-2003-0078)
3771 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3772 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3773 Martin Vuagnoux (EPFL, Ilion)]
3775 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3777 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3778 memory from it's contents. This is done with a counter that will
3779 place alternating values in each byte. This can be used to solve
3780 two issues: 1) the removal of calls to memset() by highly optimizing
3781 compilers, and 2) cleansing with other values than 0, since those can
3782 be read through on certain media, for example a swap space on disk.
3785 *) Bugfix: client side session caching did not work with external caching,
3786 because the session->cipher setting was not restored when reloading
3787 from the external cache. This problem was masked, when
3788 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3789 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3792 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3793 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3794 [Zeev Lieber <zeev-l@yahoo.com>]
3796 *) Undo an undocumented change introduced in 0.9.6e which caused
3797 repeated calls to OpenSSL_add_all_ciphers() and
3798 OpenSSL_add_all_digests() to be ignored, even after calling
3802 *) Change the default configuration reader to deal with last line not
3803 being properly terminated.
3806 *) Change X509_NAME_cmp() so it applies the special rules on handling
3807 DN values that are of type PrintableString, as well as RDNs of type
3808 emailAddress where the value has the type ia5String.
3809 [stefank@valicert.com via Richard Levitte]
3811 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3812 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3813 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3814 the bitwise-OR of the two for use by the majority of applications
3815 wanting this behaviour, and update the docs. The documented
3816 behaviour and actual behaviour were inconsistent and had been
3817 changing anyway, so this is more a bug-fix than a behavioural
3819 [Geoff Thorpe, diagnosed by Nadav Har'El]
3821 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3822 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3825 *) Fix initialization code race conditions in
3826 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3827 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3828 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3829 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3830 ssl2_get_cipher_by_char(),
3831 ssl3_get_cipher_by_char().
3832 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3834 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3835 the cached sessions are flushed, as the remove_cb() might use ex_data
3836 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3837 (see [openssl.org #212]).
3838 [Geoff Thorpe, Lutz Jaenicke]
3840 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3841 length, instead of the encoding length to d2i_ASN1_OBJECT.
3844 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3846 *) [In 0.9.6g-engine release:]
3847 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3848 [Lynn Gazis <lgazis@rainbow.com>]
3850 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3852 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3853 and get fix the header length calculation.
3854 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3855 Alon Kantor <alonk@checkpoint.com> (and others),
3858 *) Use proper error handling instead of 'assertions' in buffer
3859 overflow checks added in 0.9.6e. This prevents DoS (the
3860 assertions could call abort()).
3861 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3863 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3865 *) Add various sanity checks to asn1_get_length() to reject
3866 the ASN1 length bytes if they exceed sizeof(long), will appear
3867 negative or the content length exceeds the length of the
3869 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3871 *) Fix cipher selection routines: ciphers without encryption had no flags
3872 for the cipher strength set and where therefore not handled correctly
3873 by the selection routines (PR #130).
3876 *) Fix EVP_dsa_sha macro.
3880 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3881 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3882 that was added in OpenSSL 0.9.6d.
3884 As the countermeasure turned out to be incompatible with some
3885 broken SSL implementations, the new option is part of SSL_OP_ALL.
3886 SSL_OP_ALL is usually employed when compatibility with weird SSL
3887 implementations is desired (e.g. '-bugs' option to 's_client' and
3888 's_server'), so the new option is automatically set in many
3892 *) Changes in security patch:
3894 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3895 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3896 Air Force Materiel Command, USAF, under agreement number
3899 *) Add various sanity checks to asn1_get_length() to reject
3900 the ASN1 length bytes if they exceed sizeof(long), will appear
3901 negative or the content length exceeds the length of the
3902 supplied buffer. (CVE-2002-0659)
3903 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3905 *) Assertions for various potential buffer overflows, not known to
3907 [Ben Laurie (CHATS)]
3909 *) Various temporary buffers to hold ASCII versions of integers were
3910 too small for 64 bit platforms. (CVE-2002-0655)
3911 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3913 *) Remote buffer overflow in SSL3 protocol - an attacker could
3914 supply an oversized session ID to a client. (CVE-2002-0656)
3915 [Ben Laurie (CHATS)]
3917 *) Remote buffer overflow in SSL2 protocol - an attacker could
3918 supply an oversized client master key. (CVE-2002-0656)
3919 [Ben Laurie (CHATS)]
3921 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3923 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3924 encoded as NULL) with id-dsa-with-sha1.
3925 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3927 *) Check various X509_...() return values in apps/req.c.
3928 [Nils Larsch <nla@trustcenter.de>]
3930 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3931 an end-of-file condition would erronously be flagged, when the CRLF
3932 was just at the end of a processed block. The bug was discovered when
3933 processing data through a buffering memory BIO handing the data to a
3934 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3935 <ptsekov@syntrex.com> and Nedelcho Stanev.
3938 *) Implement a countermeasure against a vulnerability recently found
3939 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3940 before application data chunks to avoid the use of known IVs
3941 with data potentially chosen by the attacker.
3944 *) Fix length checks in ssl3_get_client_hello().
3947 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3948 to prevent ssl3_read_internal() from incorrectly assuming that
3949 ssl3_read_bytes() found application data while handshake
3950 processing was enabled when in fact s->s3->in_read_app_data was
3951 merely automatically cleared during the initial handshake.
3952 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3954 *) Fix object definitions for Private and Enterprise: they were not
3955 recognized in their shortname (=lowercase) representation. Extend
3956 obj_dat.pl to issue an error when using undefined keywords instead
3957 of silently ignoring the problem (Svenning Sorensen
3958 <sss@sss.dnsalias.net>).
3961 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3962 generators, i.e. generators other than 2 and 5. (Previously, the
3963 code did not properly initialise the 'add' and 'rem' values to
3964 BN_generate_prime().)
3966 In the new general case, we do not insist that 'generator' is
3967 actually a primitive root: This requirement is rather pointless;
3968 a generator of the order-q subgroup is just as good, if not
3972 *) Map new X509 verification errors to alerts. Discovered and submitted by
3973 Tom Wu <tom@arcot.com>.
3976 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3977 returning non-zero before the data has been completely received
3978 when using non-blocking I/O.
3979 [Bodo Moeller; problem pointed out by John Hughes]
3981 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3982 [Ben Laurie, Lutz Jaenicke]
3984 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3985 Yoram Zahavi <YoramZ@gilian.com>).
3988 *) Add information about CygWin 1.3 and on, and preserve proper
3989 configuration for the versions before that.
3990 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3992 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3993 check whether we deal with a copy of a session and do not delete from
3994 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3995 <izhar@checkpoint.com>.
3998 *) Do not store session data into the internal session cache, if it
3999 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4000 flag is set). Proposed by Aslam <aslam@funk.com>.
4003 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4007 *) [In 0.9.6d-engine release:]
4008 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4009 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4011 *) Add the configuration target linux-s390x.
4012 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4014 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4015 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4016 variable as an indication that a ClientHello message has been
4017 received. As the flag value will be lost between multiple
4018 invocations of ssl3_accept when using non-blocking I/O, the
4019 function may not be aware that a handshake has actually taken
4020 place, thus preventing a new session from being added to the
4023 To avoid this problem, we now set s->new_session to 2 instead of
4024 using a local variable.
4025 [Lutz Jaenicke, Bodo Moeller]
4027 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4028 if the SSL_R_LENGTH_MISMATCH error is detected.
4029 [Geoff Thorpe, Bodo Moeller]
4031 *) New 'shared_ldflag' column in Configure platform table.
4034 *) Fix EVP_CIPHER_mode macro.
4035 ["Dan S. Camper" <dan@bti.net>]
4037 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4038 type, we must throw them away by setting rr->length to 0.
4039 [D P Chang <dpc@qualys.com>]
4041 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4043 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4044 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4045 worked incorrectly for those cases where range = 10..._2 and
4046 3*range is two bits longer than range.)
4049 *) Only add signing time to PKCS7 structures if it is not already
4053 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4054 OBJ_ld_ce should be OBJ_id_ce.
4055 Also some ip-pda OIDs in crypto/objects/objects.txt were
4056 incorrect (cf. RFC 3039).
4057 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4059 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4060 returns early because it has nothing to do.
4061 [Andy Schneider <andy.schneider@bjss.co.uk>]
4063 *) [In 0.9.6c-engine release:]
4064 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4065 [Andy Schneider <andy.schneider@bjss.co.uk>]
4067 *) [In 0.9.6c-engine release:]
4068 Add support for Cryptographic Appliance's keyserver technology.
4069 (Use engine 'keyclient')
4070 [Cryptographic Appliances and Geoff Thorpe]
4072 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4073 is called via tools/c89.sh because arguments have to be
4074 rearranged (all '-L' options must appear before the first object
4076 [Richard Shapiro <rshapiro@abinitio.com>]
4078 *) [In 0.9.6c-engine release:]
4079 Add support for Broadcom crypto accelerator cards, backported
4081 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4083 *) [In 0.9.6c-engine release:]
4084 Add support for SureWare crypto accelerator cards from
4085 Baltimore Technologies. (Use engine 'sureware')
4086 [Baltimore Technologies and Mark Cox]
4088 *) [In 0.9.6c-engine release:]
4089 Add support for crypto accelerator cards from Accelerated
4090 Encryption Processing, www.aep.ie. (Use engine 'aep')
4091 [AEP Inc. and Mark Cox]
4093 *) Add a configuration entry for gcc on UnixWare.
4094 [Gary Benson <gbenson@redhat.com>]
4096 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4097 messages are stored in a single piece (fixed-length part and
4098 variable-length part combined) and fix various bugs found on the way.
4101 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4102 instead. BIO_gethostbyname() does not know what timeouts are
4103 appropriate, so entries would stay in cache even when they have
4105 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4107 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4108 faced with a pathologically small ClientHello fragment that does
4109 not contain client_version: Instead of aborting with an error,
4110 simply choose the highest available protocol version (i.e.,
4111 TLS 1.0 unless it is disabled). In practice, ClientHello
4112 messages are never sent like this, but this change gives us
4113 strictly correct behaviour at least for TLS.
4116 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4117 never resets s->method to s->ctx->method when called from within
4118 one of the SSL handshake functions.
4119 [Bodo Moeller; problem pointed out by Niko Baric]
4121 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4122 (sent using the client's version number) if client_version is
4123 smaller than the protocol version in use. Also change
4124 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4125 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4126 the client will at least see that alert.
4129 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4133 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4134 client receives HelloRequest while in a handshake.
4135 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4137 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4138 should end in 'break', not 'goto end' which circuments various
4139 cleanups done in state SSL_ST_OK. But session related stuff
4140 must be disabled for SSL_ST_OK in the case that we just sent a
4143 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4144 before just sending a HelloRequest.
4145 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4147 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4148 reveal whether illegal block cipher padding was found or a MAC
4149 verification error occured. (Neither SSLerr() codes nor alerts
4150 are directly visible to potential attackers, but the information
4151 may leak via logfiles.)
4153 Similar changes are not required for the SSL 2.0 implementation
4154 because the number of padding bytes is sent in clear for SSL 2.0,
4155 and the extra bytes are just ignored. However ssl/s2_pkt.c
4156 failed to verify that the purported number of padding bytes is in
4160 *) Add OpenUNIX-8 support including shared libraries
4161 (Boyd Lynn Gerber <gerberb@zenez.com>).
4164 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4165 'wristwatch attack' using huge encoding parameters (cf.
4166 James H. Manger's CRYPTO 2001 paper). Note that the
4167 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4168 encoding parameters and hence was not vulnerable.
4171 *) BN_sqr() bug fix.
4172 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4174 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4175 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4176 followed by modular reduction.
4177 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4179 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4180 equivalent based on BN_pseudo_rand() instead of BN_rand().
4183 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4184 This function was broken, as the check for a new client hello message
4185 to handle SGC did not allow these large messages.
4186 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4189 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4192 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4193 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4196 *) Rework the configuration and shared library support for Tru64 Unix.
4197 The configuration part makes use of modern compiler features and
4198 still retains old compiler behavior for those that run older versions
4199 of the OS. The shared library support part includes a variant that
4200 uses the RPATH feature, and is available through the special
4201 configuration target "alpha-cc-rpath", which will never be selected
4203 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4205 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4206 with the same message size as in ssl3_get_certificate_request().
4207 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4208 messages might inadvertently be reject as too long.
4209 [Petr Lampa <lampa@fee.vutbr.cz>]
4211 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4214 *) Modified SSL library such that the verify_callback that has been set
4215 specificly for an SSL object with SSL_set_verify() is actually being
4216 used. Before the change, a verify_callback set with this function was
4217 ignored and the verify_callback() set in the SSL_CTX at the time of
4218 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4219 to allow the necessary settings.
4222 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4223 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4224 done automatically (in contradiction to the requirements of the C
4225 standard). This made problems when used from OpenSSH.
4228 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4229 dh->length and always used
4231 BN_rand_range(priv_key, dh->p).
4233 BN_rand_range() is not necessary for Diffie-Hellman, and this
4234 specific range makes Diffie-Hellman unnecessarily inefficient if
4235 dh->length (recommended exponent length) is much smaller than the
4236 length of dh->p. We could use BN_rand_range() if the order of
4237 the subgroup was stored in the DH structure, but we only have
4242 BN_rand(priv_key, l, ...)
4244 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4250 RSA_eay_public_encrypt
4251 RSA_eay_private_decrypt
4252 RSA_eay_private_encrypt (signing)
4253 RSA_eay_public_decrypt (signature verification)
4255 (default implementations for RSA_public_encrypt,
4256 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4257 always reject numbers >= n.
4260 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4261 to synchronize access to 'locking_thread'. This is necessary on
4262 systems where access to 'locking_thread' (an 'unsigned long'
4263 variable) is not atomic.
4266 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4267 *before* setting the 'crypto_lock_rand' flag. The previous code had
4268 a race condition if 0 is a valid thread ID.
4269 [Travis Vitek <vitek@roguewave.com>]
4271 *) Add support for shared libraries under Irix.
4272 [Albert Chin-A-Young <china@thewrittenword.com>]
4274 *) Add configuration option to build on Linux on both big-endian and
4276 [Ralf Baechle <ralf@uni-koblenz.de>]
4278 *) Add the possibility to create shared libraries on HP-UX.
4281 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4283 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4284 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4285 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4286 PRNG state recovery was possible based on the output of
4287 one PRNG request appropriately sized to gain knowledge on
4288 'md' followed by enough consecutive 1-byte PRNG requests
4289 to traverse all of 'state'.
4291 1. When updating 'md_local' (the current thread's copy of 'md')
4292 during PRNG output generation, hash all of the previous
4293 'md_local' value, not just the half used for PRNG output.
4295 2. Make the number of bytes from 'state' included into the hash
4296 independent from the number of PRNG bytes requested.
4298 The first measure alone would be sufficient to avoid
4299 Markku-Juhani's attack. (Actually it had never occurred
4300 to me that the half of 'md_local' used for chaining was the
4301 half from which PRNG output bytes were taken -- I had always
4302 assumed that the secret half would be used.) The second
4303 measure makes sure that additional data from 'state' is never
4304 mixed into 'md_local' in small portions; this heuristically
4305 further strengthens the PRNG.
4308 *) Fix crypto/bn/asm/mips3.s.
4311 *) When only the key is given to "enc", the IV is undefined. Print out
4312 an error message in this case.
4315 *) Handle special case when X509_NAME is empty in X509 printing routines.
4318 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4319 positive and less than q.
4322 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4323 used: it isn't thread safe and the add_lock_callback should handle
4325 [Paul Rose <Paul.Rose@bridge.com>]
4327 *) Verify that incoming data obeys the block size in
4328 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4332 [Ulf Möller, Bodo Möller]
4334 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4335 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4336 when fixing the server behaviour for backwards-compatible 'client
4337 hello' messages. (Note that the attack is impractical against
4338 SSL 3.0 and TLS 1.0 anyway because length and version checking
4339 means that the probability of guessing a valid ciphertext is
4340 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4343 Before 0.9.5, the countermeasure (hide the error by generating a
4344 random 'decryption result') did not work properly because
4345 ERR_clear_error() was missing, meaning that SSL_get_error() would
4346 detect the supposedly ignored error.
4348 Both problems are now fixed.
4351 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4352 (previously it was 1024).
4355 *) Fix for compatibility mode trust settings: ignore trust settings
4356 unless some valid trust or reject settings are present.
4359 *) Fix for blowfish EVP: its a variable length cipher.
4362 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4363 parameters in DSA public key structures and return an error in the
4364 DSA routines if parameters are absent.
4367 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4368 in the current directory if neither $RANDFILE nor $HOME was set.
4369 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4370 caused some confusion to Windows users who haven't defined $HOME.
4371 Thus RAND_file_name() is changed again: e_os.h can define a
4372 DEFAULT_HOME, which will be used if $HOME is not set.
4373 For Windows, we use "C:"; on other platforms, we still require
4374 environment variables.
4376 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4377 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4378 having multiple threads call RAND_poll() concurrently.
4381 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4382 combination of a flag and a thread ID variable.
4383 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4384 flag), *other* threads can enter ssleay_add_bytes without obeying
4385 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4386 that they do not hold after the first thread unsets add_do_not_lock).
4389 *) Change bctest again: '-x' expressions are not available in all
4393 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4395 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4396 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4398 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4399 the default extension for executables, if any. Also, make the perl
4400 scripts that use symlink() to test if it really exists and use "cp"
4401 if it doesn't. All this made OpenSSL compilable and installable in
4405 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4406 If SEQUENCE is length is indefinite just set c->slen to the total
4407 amount of data available.
4408 [Steve Henson, reported by shige@FreeBSD.org]
4409 [This change does not apply to 0.9.7.]
4411 *) Change bctest to avoid here-documents inside command substitution
4412 (workaround for FreeBSD /bin/sh bug).
4413 For compatibility with Ultrix, avoid shell functions (introduced
4414 in the bctest version that searches along $PATH).
4417 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4418 with des_encrypt() defined on some operating systems, like Solaris
4422 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4423 On the Importance of Eliminating Errors in Cryptographic
4424 Computations, J. Cryptology 14 (2001) 2, 101-119,
4425 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4428 *) MIPS assembler BIGNUM division bug fix.
4431 *) Disabled incorrect Alpha assembler code.
4434 *) Fix PKCS#7 decode routines so they correctly update the length
4435 after reading an EOC for the EXPLICIT tag.
4437 [This change does not apply to 0.9.7.]
4439 *) Fix bug in PKCS#12 key generation routines. This was triggered
4440 if a 3DES key was generated with a 0 initial byte. Include
4441 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4442 (but broken) behaviour.
4445 *) Enhance bctest to search for a working bc along $PATH and print
4447 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4449 *) Fix memory leaks in err.c: free err_data string if necessary;
4450 don't write to the wrong index in ERR_set_error_data.
4453 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4457 *) Replace rdtsc with _emit statements for VC++ version 5.
4458 [Jeremy Cooper <jeremy@baymoo.org>]
4460 *) Make it possible to reuse SSLv2 sessions.
4463 *) In copy_email() check for >= 0 as a return value for
4464 X509_NAME_get_index_by_NID() since 0 is a valid index.
4465 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4467 *) Avoid coredump with unsupported or invalid public keys by checking if
4468 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4469 PKCS7_verify() fails with non detached data.
4472 *) Don't use getenv in library functions when run as setuid/setgid.
4473 New function OPENSSL_issetugid().
4476 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4477 due to incorrect handling of multi-threading:
4479 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4481 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4483 3. Count how many times MemCheck_off() has been called so that
4484 nested use can be treated correctly. This also avoids
4485 inband-signalling in the previous code (which relied on the
4486 assumption that thread ID 0 is impossible).
4489 *) Add "-rand" option also to s_client and s_server.
4492 *) Fix CPU detection on Irix 6.x.
4493 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4494 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4496 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4499 [This change does not apply to 0.9.7.]
4501 *) Use the cached encoding of an X509_NAME structure rather than
4502 copying it. This is apparently the reason for the libsafe "errors"
4503 but the code is actually correct.
4506 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4507 Bleichenbacher's DSA attack.
4508 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4509 to be set and top=0 forces the highest bit to be set; top=-1 is new
4510 and leaves the highest bit random.
4511 [Ulf Moeller, Bodo Moeller]
4513 *) In the NCONF_...-based implementations for CONF_... queries
4514 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4515 a temporary CONF structure with the data component set to NULL
4516 (which gives segmentation faults in lh_retrieve).
4517 Instead, use NULL for the CONF pointer in CONF_get_string and
4518 CONF_get_number (which may use environment variables) and directly
4519 return NULL from CONF_get_section.
4522 *) Fix potential buffer overrun for EBCDIC.
4525 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4526 keyUsage if basicConstraints absent for a CA.
4529 *) Make SMIME_write_PKCS7() write mail header values with a format that
4530 is more generally accepted (no spaces before the semicolon), since
4531 some programs can't parse those values properly otherwise. Also make
4532 sure BIO's that break lines after each write do not create invalid
4536 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4537 macros previously used would not encode an empty SEQUENCE OF
4538 and break the signature.
4540 [This change does not apply to 0.9.7.]
4542 *) Zero the premaster secret after deriving the master secret in
4546 *) Add some EVP_add_digest_alias registrations (as found in
4547 OpenSSL_add_all_digests()) to SSL_library_init()
4548 aka OpenSSL_add_ssl_algorithms(). This provides improved
4549 compatibility with peers using X.509 certificates
4550 with unconventional AlgorithmIdentifier OIDs.
4553 *) Fix for Irix with NO_ASM.
4554 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4556 *) ./config script fixes.
4557 [Ulf Moeller, Richard Levitte]
4559 *) Fix 'openssl passwd -1'.
4562 *) Change PKCS12_key_gen_asc() so it can cope with non null
4563 terminated strings whose length is passed in the passlen
4564 parameter, for example from PEM callbacks. This was done
4565 by adding an extra length parameter to asc2uni().
4566 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4568 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4569 call failed, free the DSA structure.
4572 *) Fix to uni2asc() to cope with zero length Unicode strings.
4573 These are present in some PKCS#12 files.
4576 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4577 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4578 when writing a 32767 byte record.
4579 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4581 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4582 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4584 (RSA objects have a reference count access to which is protected
4585 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4586 so they are meant to be shared between threads.)
4587 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4588 "Reddie, Steven" <Steven.Reddie@ca.com>]
4590 *) Fix a deadlock in CRYPTO_mem_leaks().
4593 *) Use better test patterns in bntest.
4596 *) rand_win.c fix for Borland C.
4599 *) BN_rshift bugfix for n == 0.
4602 *) Add a 'bctest' script that checks for some known 'bc' bugs
4603 so that 'make test' does not abort just because 'bc' is broken.
4606 *) Store verify_result within SSL_SESSION also for client side to
4607 avoid potential security hole. (Re-used sessions on the client side
4608 always resulted in verify_result==X509_V_OK, not using the original
4609 result of the server certificate verification.)
4612 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4613 SSL3_RT_APPLICATION_DATA, return 0.
4614 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4618 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4619 releases, have been re-implemented by renaming the previous
4620 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4621 and ssl3_read_internal, respectively, and adding 'peek' parameters
4622 to them. The new ssl[23]_{read,peek} functions are calls to
4623 ssl[23]_read_internal with the 'peek' flag set appropriately.
4624 A 'peek' parameter has also been added to ssl3_read_bytes, which
4625 does the actual work for ssl3_read_internal.
4628 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4629 the method-specific "init()" handler. Also clean up ex_data after
4630 calling the method-specific "finish()" handler. Previously, this was
4631 happening the other way round.
4634 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4635 The previous value, 12, was not always sufficient for BN_mod_exp().
4638 *) Make sure that shared libraries get the internal name engine with
4639 the full version number and not just 0. This should mark the
4640 shared libraries as not backward compatible. Of course, this should
4641 be changed again when we can guarantee backward binary compatibility.
4644 *) Fix typo in get_cert_by_subject() in by_dir.c
4645 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4647 *) Rework the system to generate shared libraries:
4649 - Make note of the expected extension for the shared libraries and
4650 if there is a need for symbolic links from for example libcrypto.so.0
4651 to libcrypto.so.0.9.7. There is extended info in Configure for
4654 - Make as few rebuilds of the shared libraries as possible.
4656 - Still avoid linking the OpenSSL programs with the shared libraries.
4658 - When installing, install the shared libraries separately from the
4662 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4664 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4665 and not in SSL_clear because the latter is also used by the
4666 accept/connect functions; previously, the settings made by
4667 SSL_set_read_ahead would be lost during the handshake.
4668 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4670 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4671 Previously, it would create entries for disableed algorithms no
4675 *) Added several new manual pages for SSL_* function.
4678 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4680 *) In ssl23_get_client_hello, generate an error message when faced
4681 with an initial SSL 3.0/TLS record that is too small to contain the
4682 first two bytes of the ClientHello message, i.e. client_version.
4683 (Note that this is a pathologic case that probably has never happened
4684 in real life.) The previous approach was to use the version number
4685 from the record header as a substitute; but our protocol choice
4686 should not depend on that one because it is not authenticated
4687 by the Finished messages.
4690 *) More robust randomness gathering functions for Windows.
4691 [Jeffrey Altman <jaltman@columbia.edu>]
4693 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4694 not set then we don't setup the error code for issuer check errors
4695 to avoid possibly overwriting other errors which the callback does
4696 handle. If an application does set the flag then we assume it knows
4697 what it is doing and can handle the new informational codes
4701 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4702 a general "ANY" type, as such it should be able to decode anything
4703 including tagged types. However it didn't check the class so it would
4704 wrongly interpret tagged types in the same way as their universal
4705 counterpart and unknown types were just rejected. Changed so that the
4706 tagged and unknown types are handled in the same way as a SEQUENCE:
4707 that is the encoding is stored intact. There is also a new type
4708 "V_ASN1_OTHER" which is used when the class is not universal, in this
4709 case we have no idea what the actual type is so we just lump them all
4713 *) On VMS, stdout may very well lead to a file that is written to
4714 in a record-oriented fashion. That means that every write() will
4715 write a separate record, which will be read separately by the
4716 programs trying to read from it. This can be very confusing.
4718 The solution is to put a BIO filter in the way that will buffer
4719 text until a linefeed is reached, and then write everything a
4720 line at a time, so every record written will be an actual line,
4721 not chunks of lines and not (usually doesn't happen, but I've
4722 seen it once) several lines in one record. BIO_f_linebuffer() is
4725 Currently, it's a VMS-only method, because that's where it has
4726 been tested well enough.
4729 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4730 it can return incorrect results.
4731 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4732 but it was in 0.9.6-beta[12].)
4735 *) Disable the check for content being present when verifying detached
4736 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4737 include zero length content when signing messages.
4740 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4741 BIO_ctrl (for BIO pairs).
4744 *) Add DSO method for VMS.
4747 *) Bug fix: Montgomery multiplication could produce results with the
4751 *) Add RPM specification openssl.spec and modify it to build three
4752 packages. The default package contains applications, application
4753 documentation and run-time libraries. The devel package contains
4754 include files, static libraries and function documentation. The
4755 doc package contains the contents of the doc directory. The original
4756 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4759 *) Add a large number of documentation files for many SSL routines.
4760 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4762 *) Add a configuration entry for Sony News 4.
4763 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4765 *) Don't set the two most significant bits to one when generating a
4766 random number < q in the DSA library.
4769 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4770 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4771 the underlying transport is blocking) if a handshake took place.
4772 (The default behaviour is needed by applications such as s_client
4773 and s_server that use select() to determine when to use SSL_read;
4774 but for applications that know in advance when to expect data, it
4775 just makes things more complicated.)
4778 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4782 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4783 work better on such systems.
4784 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4786 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4787 Update PKCS12_parse() so it copies the friendlyName and the
4788 keyid to the certificates aux info.
4791 *) Fix bug in PKCS7_verify() which caused an infinite loop
4792 if there was more than one signature.
4793 [Sven Uszpelkat <su@celocom.de>]
4795 *) Major change in util/mkdef.pl to include extra information
4796 about each symbol, as well as presentig variables as well
4797 as functions. This change means that there's n more need
4798 to rebuild the .num files when some algorithms are excluded.
4801 *) Allow the verify time to be set by an application,
4802 rather than always using the current time.
4805 *) Phase 2 verify code reorganisation. The certificate
4806 verify code now looks up an issuer certificate by a
4807 number of criteria: subject name, authority key id
4808 and key usage. It also verifies self signed certificates
4809 by the same criteria. The main comparison function is
4810 X509_check_issued() which performs these checks.
4812 Lot of changes were necessary in order to support this
4813 without completely rewriting the lookup code.
4815 Authority and subject key identifier are now cached.
4817 The LHASH 'certs' is X509_STORE has now been replaced
4818 by a STACK_OF(X509_OBJECT). This is mainly because an
4819 LHASH can't store or retrieve multiple objects with
4820 the same hash value.
4822 As a result various functions (which were all internal
4823 use only) have changed to handle the new X509_STORE
4824 structure. This will break anything that messed round
4825 with X509_STORE internally.
4827 The functions X509_STORE_add_cert() now checks for an
4828 exact match, rather than just subject name.
4830 The X509_STORE API doesn't directly support the retrieval
4831 of multiple certificates matching a given criteria, however
4832 this can be worked round by performing a lookup first
4833 (which will fill the cache with candidate certificates)
4834 and then examining the cache for matches. This is probably
4835 the best we can do without throwing out X509_LOOKUP
4836 entirely (maybe later...).
4838 The X509_VERIFY_CTX structure has been enhanced considerably.
4840 All certificate lookup operations now go via a get_issuer()
4841 callback. Although this currently uses an X509_STORE it
4842 can be replaced by custom lookups. This is a simple way
4843 to bypass the X509_STORE hackery necessary to make this
4844 work and makes it possible to use more efficient techniques
4845 in future. A very simple version which uses a simple
4846 STACK for its trusted certificate store is also provided
4847 using X509_STORE_CTX_trusted_stack().
4849 The verify_cb() and verify() callbacks now have equivalents
4850 in the X509_STORE_CTX structure.
4852 X509_STORE_CTX also has a 'flags' field which can be used
4853 to customise the verify behaviour.
4856 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4857 excludes S/MIME capabilities.
4860 *) When a certificate request is read in keep a copy of the
4861 original encoding of the signed data and use it when outputing
4862 again. Signatures then use the original encoding rather than
4863 a decoded, encoded version which may cause problems if the
4864 request is improperly encoded.
4867 *) For consistency with other BIO_puts implementations, call
4868 buffer_write(b, ...) directly in buffer_puts instead of calling
4871 In BIO_puts, increment b->num_write as in BIO_write.
4872 [Peter.Sylvester@EdelWeb.fr]
4874 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4875 BN_zero, we may not return a BIGNUM with an array consisting of
4879 *) Avoid calling abort() from within the library when problems are
4880 detected, except if preprocessor symbols have been defined
4881 (such as REF_CHECK, BN_DEBUG etc.).
4884 *) New openssl application 'rsautl'. This utility can be
4885 used for low level RSA operations. DER public key
4886 BIO/fp routines also added.
4889 *) New Configure entry and patches for compiling on QNX 4.
4890 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4892 *) A demo state-machine implementation was sponsored by
4893 Nuron (http://www.nuron.com/) and is now available in
4894 demos/state_machine.
4897 *) New options added to the 'dgst' utility for signature
4898 generation and verification.
4901 *) Unrecognized PKCS#7 content types are now handled via a
4902 catch all ASN1_TYPE structure. This allows unsupported
4903 types to be stored as a "blob" and an application can
4904 encode and decode it manually.
4907 *) Fix various signed/unsigned issues to make a_strex.c
4909 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4911 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4912 length if passed a buffer. ASN1_INTEGER_to_BN failed
4913 if passed a NULL BN and its argument was negative.
4914 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4916 *) Modification to PKCS#7 encoding routines to output definite
4917 length encoding. Since currently the whole structures are in
4918 memory there's not real point in using indefinite length
4919 constructed encoding. However if OpenSSL is compiled with
4920 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4923 *) Added BIO_vprintf() and BIO_vsnprintf().
4926 *) Added more prefixes to parse for in the the strings written
4927 through a logging bio, to cover all the levels that are available
4928 through syslog. The prefixes are now:
4930 PANIC, EMERG, EMR => LOG_EMERG
4931 ALERT, ALR => LOG_ALERT
4932 CRIT, CRI => LOG_CRIT
4933 ERROR, ERR => LOG_ERR
4934 WARNING, WARN, WAR => LOG_WARNING
4935 NOTICE, NOTE, NOT => LOG_NOTICE
4936 INFO, INF => LOG_INFO
4937 DEBUG, DBG => LOG_DEBUG
4939 and as before, if none of those prefixes are present at the
4940 beginning of the string, LOG_ERR is chosen.
4942 On Win32, the LOG_* levels are mapped according to this:
4944 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4945 LOG_WARNING => EVENTLOG_WARNING_TYPE
4946 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4950 *) Made it possible to reconfigure with just the configuration
4951 argument "reconf" or "reconfigure". The command line arguments
4952 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4953 and are retrieved from there when reconfiguring.
4957 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4959 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4962 *) The obj_dat.pl script was messing up the sorting of object
4963 names. The reason was that it compared the quoted version
4964 of strings as a result "OCSP" > "OCSP Signing" because
4965 " > SPACE. Changed script to store unquoted versions of
4966 names and add quotes on output. It was also omitting some
4967 names from the lookup table if they were given a default
4968 value (that is if SN is missing it is given the same
4969 value as LN and vice versa), these are now added on the
4970 grounds that if an object has a name we should be able to
4971 look it up. Finally added warning output when duplicate
4972 short or long names are found.
4975 *) Changes needed for Tandem NSK.
4976 [Scott Uroff <scott@xypro.com>]
4978 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4979 RSA_padding_check_SSLv23(), special padding was never detected
4980 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4981 version rollback attacks was not effective.
4983 In s23_clnt.c, don't use special rollback-attack detection padding
4984 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4985 client; similarly, in s23_srvr.c, don't do the rollback check if
4986 SSL 2.0 is the only protocol enabled in the server.
4989 *) Make it possible to get hexdumps of unprintable data with 'openssl
4990 asn1parse'. By implication, the functions ASN1_parse_dump() and
4991 BIO_dump_indent() are added.
4994 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4995 these print out strings and name structures based on various
4996 flags including RFC2253 support and proper handling of
4997 multibyte characters. Added options to the 'x509' utility
4998 to allow the various flags to be set.
5001 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5002 Also change the functions X509_cmp_current_time() and
5003 X509_gmtime_adj() work with an ASN1_TIME structure,
5004 this will enable certificates using GeneralizedTime in validity
5005 dates to be checked.
5008 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5009 negative public key encodings) on by default,
5010 NO_NEG_PUBKEY_BUG can be set to disable it.
5013 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5014 content octets. An i2c_ASN1_OBJECT is unnecessary because
5015 the encoding can be trivially obtained from the structure.
5018 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5019 not read locks (CRYPTO_r_[un]lock).
5022 *) A first attempt at creating official support for shared
5023 libraries through configuration. I've kept it so the
5024 default is static libraries only, and the OpenSSL programs
5025 are always statically linked for now, but there are
5026 preparations for dynamic linking in place.
5027 This has been tested on Linux and Tru64.
5030 *) Randomness polling function for Win9x, as described in:
5031 Peter Gutmann, Software Generation of Practically Strong
5035 *) Fix so PRNG is seeded in req if using an already existing
5039 *) New options to smime application. -inform and -outform
5040 allow alternative formats for the S/MIME message including
5041 PEM and DER. The -content option allows the content to be
5042 specified separately. This should allow things like Netscape
5043 form signing output easier to verify.
5046 *) Fix the ASN1 encoding of tags using the 'long form'.
5049 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5050 STRING types. These convert content octets to and from the
5051 underlying type. The actual tag and length octets are
5052 already assumed to have been read in and checked. These
5053 are needed because all other string types have virtually
5054 identical handling apart from the tag. By having versions
5055 of the ASN1 functions that just operate on content octets
5056 IMPLICIT tagging can be handled properly. It also allows
5057 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5058 and ASN1_INTEGER are identical apart from the tag.
5061 *) Change the handling of OID objects as follows:
5063 - New object identifiers are inserted in objects.txt, following
5064 the syntax given in objects.README.
5065 - objects.pl is used to process obj_mac.num and create a new
5067 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5070 This is currently kind of a hack, and the perl code in objects.pl
5071 isn't very elegant, but it works as I intended. The simplest way
5072 to check that it worked correctly is to look in obj_dat.h and
5073 check the array nid_objs and make sure the objects haven't moved
5074 around (this is important!). Additions are OK, as well as
5075 consistent name changes.
5078 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5081 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5082 The given file adds to whatever has already been seeded into the
5083 random pool through the RANDFILE configuration file option or
5084 environment variable, or the default random state file.
5087 *) mkstack.pl now sorts each macro group into lexical order.
5088 Previously the output order depended on the order the files
5089 appeared in the directory, resulting in needless rewriting
5093 *) Patches to make OpenSSL compile under Win32 again. Mostly
5094 work arounds for the VC++ problem that it treats func() as
5095 func(void). Also stripped out the parts of mkdef.pl that
5096 added extra typesafe functions: these no longer exist.
5099 *) Reorganisation of the stack code. The macros are now all
5100 collected in safestack.h . Each macro is defined in terms of
5101 a "stack macro" of the form SKM_<name>(type, a, b). The
5102 DEBUG_SAFESTACK is now handled in terms of function casts,
5103 this has the advantage of retaining type safety without the
5104 use of additional functions. If DEBUG_SAFESTACK is not defined
5105 then the non typesafe macros are used instead. Also modified the
5106 mkstack.pl script to handle the new form. Needs testing to see
5107 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5108 the default if no major problems. Similar behaviour for ASN1_SET_OF
5109 and PKCS12_STACK_OF.
5112 *) When some versions of IIS use the 'NET' form of private key the
5113 key derivation algorithm is different. Normally MD5(password) is
5114 used as a 128 bit RC4 key. In the modified case
5115 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5116 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5117 as the old Netscape_RSA functions except they have an additional
5118 'sgckey' parameter which uses the modified algorithm. Also added
5119 an -sgckey command line option to the rsa utility. Thanks to
5120 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5121 algorithm to openssl-dev.
5124 *) The evp_local.h macros were using 'c.##kname' which resulted in
5125 invalid expansion on some systems (SCO 5.0.5 for example).
5126 Corrected to 'c.kname'.
5127 [Phillip Porch <root@theporch.com>]
5129 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5130 a STACK of email addresses from a certificate or request, these look
5131 in the subject name and the subject alternative name extensions and
5132 omit any duplicate addresses.
5135 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5136 This makes DSA verification about 2 % faster.
5139 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5140 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5141 plus overhead for 1024 bit moduli).
5142 This makes exponentiations about 0.5 % faster for 1024 bit
5143 exponents (as measured by "openssl speed rsa2048").
5146 *) Rename memory handling macros to avoid conflicts with other
5148 Malloc => OPENSSL_malloc
5149 Malloc_locked => OPENSSL_malloc_locked
5150 Realloc => OPENSSL_realloc
5151 Free => OPENSSL_free
5154 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5155 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5158 *) CygWin32 support.
5159 [John Jarvie <jjarvie@newsguy.com>]
5161 *) The type-safe stack code has been rejigged. It is now only compiled
5162 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5163 by default all type-specific stack functions are "#define"d back to
5164 standard stack functions. This results in more streamlined output
5165 but retains the type-safety checking possibilities of the original
5169 *) The STACK code has been cleaned up, and certain type declarations
5170 that didn't make a lot of sense have been brought in line. This has
5171 also involved a cleanup of sorts in safestack.h to more correctly
5172 map type-safe stack functions onto their plain stack counterparts.
5173 This work has also resulted in a variety of "const"ifications of
5174 lots of the code, especially "_cmp" operations which should normally
5175 be prototyped with "const" parameters anyway.
5178 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5179 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5180 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5181 where all of 'md' is used each time the PRNG is used, but 'state'
5182 is used only indexed by a cyclic counter. As entropy may not be
5183 well distributed from the beginning, 'md' is important as a
5184 chaining variable. However, the output function chains only half
5185 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5186 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5187 in all of 'state' being rewritten, with the new values depending
5188 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5191 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5192 the handshake is continued after ssl_verify_cert_chain();
5193 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5194 can lead to 'unexplainable' connection aborts later.
5195 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5197 *) Major EVP API cipher revision.
5198 Add hooks for extra EVP features. This allows various cipher
5199 parameters to be set in the EVP interface. Support added for variable
5200 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5201 setting of RC2 and RC5 parameters.
5203 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5206 Remove lots of duplicated code from the EVP library. For example *every*
5207 cipher init() function handles the 'iv' in the same way according to the
5208 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5209 for CFB and OFB modes they zero ctx->num.
5211 New functionality allows removal of S/MIME code RC2 hack.
5213 Most of the routines have the same form and so can be declared in terms
5216 By shifting this to the top level EVP_CipherInit() it can be removed from
5217 all individual ciphers. If the cipher wants to handle IVs or keys
5218 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5221 Change lots of functions like EVP_EncryptUpdate() to now return a
5222 value: although software versions of the algorithms cannot fail
5223 any installed hardware versions can.
5226 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5227 this option is set, tolerate broken clients that send the negotiated
5228 protocol version number instead of the requested protocol version
5232 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5233 i.e. non-zero for export ciphersuites, zero otherwise.
5234 Previous versions had this flag inverted, inconsistent with
5235 rsa_tmp_cb (..._TMP_RSA_CB).
5236 [Bodo Moeller; problem reported by Amit Chopra]
5238 *) Add missing DSA library text string. Work around for some IIS
5239 key files with invalid SEQUENCE encoding.
5242 *) Add a document (doc/standards.txt) that list all kinds of standards
5243 and so on that are implemented in OpenSSL.
5246 *) Enhance c_rehash script. Old version would mishandle certificates
5247 with the same subject name hash and wouldn't handle CRLs at all.
5248 Added -fingerprint option to crl utility, to support new c_rehash
5252 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5255 *) Fix for SSL server purpose checking. Server checking was
5256 rejecting certificates which had extended key usage present
5257 but no ssl client purpose.
5258 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5260 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5261 is a little unclear about how a blank password is handled.
5262 Since the password in encoded as a BMPString with terminating
5263 double NULL a zero length password would end up as just the
5264 double NULL. However no password at all is different and is
5265 handled differently in the PKCS#12 key generation code. NS
5266 treats a blank password as zero length. MSIE treats it as no
5267 password on export: but it will try both on import. We now do
5268 the same: PKCS12_parse() tries zero length and no password if
5269 the password is set to "" or NULL (NULL is now a valid password:
5270 it wasn't before) as does the pkcs12 application.
5273 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5274 perror when PEM_read_bio_X509_REQ fails, the error message must
5275 be obtained from the error queue.
5278 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5279 it in ERR_remove_state if appropriate, and change ERR_get_state
5280 accordingly to avoid race conditions (this is necessary because
5281 thread_hash is no longer constant once set).
5284 *) Bugfix for linux-elf makefile.one.
5287 *) RSA_get_default_method() will now cause a default
5288 RSA_METHOD to be chosen if one doesn't exist already.
5289 Previously this was only set during a call to RSA_new()
5290 or RSA_new_method(NULL) meaning it was possible for
5291 RSA_get_default_method() to return NULL.
5294 *) Added native name translation to the existing DSO code
5295 that will convert (if the flag to do so is set) filenames
5296 that are sufficiently small and have no path information
5297 into a canonical native form. Eg. "blah" converted to
5298 "libblah.so" or "blah.dll" etc.
5301 *) New function ERR_error_string_n(e, buf, len) which is like
5302 ERR_error_string(e, buf), but writes at most 'len' bytes
5303 including the 0 terminator. For ERR_error_string_n, 'buf'
5305 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5307 *) CONF library reworked to become more general. A new CONF
5308 configuration file reader "class" is implemented as well as a
5309 new functions (NCONF_*, for "New CONF") to handle it. The now
5310 old CONF_* functions are still there, but are reimplemented to
5311 work in terms of the new functions. Also, a set of functions
5312 to handle the internal storage of the configuration data is
5313 provided to make it easier to write new configuration file
5314 reader "classes" (I can definitely see something reading a
5315 configuration file in XML format, for example), called _CONF_*,
5316 or "the configuration storage API"...
5318 The new configuration file reading functions are:
5320 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5321 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5323 NCONF_default, NCONF_WIN32
5325 NCONF_dump_fp, NCONF_dump_bio
5327 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5328 NCONF_new creates a new CONF object. This works in the same way
5329 as other interfaces in OpenSSL, like the BIO interface.
5330 NCONF_dump_* dump the internal storage of the configuration file,
5331 which is useful for debugging. All other functions take the same
5332 arguments as the old CONF_* functions wth the exception of the
5333 first that must be a `CONF *' instead of a `LHASH *'.
5335 To make it easer to use the new classes with the old CONF_* functions,
5336 the function CONF_set_default_method is provided.
5339 *) Add '-tls1' option to 'openssl ciphers', which was already
5340 mentioned in the documentation but had not been implemented.
5341 (This option is not yet really useful because even the additional
5342 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5345 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5346 OpenSSL-based applications) load shared libraries and bind to
5347 them in a portable way.
5348 [Geoff Thorpe, with contributions from Richard Levitte]
5350 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5352 *) Make sure _lrotl and _lrotr are only used with MSVC.
5354 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5355 (the default implementation of RAND_status).
5357 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5358 to '-clrext' (= clear extensions), as intended and documented.
5359 [Bodo Moeller; inconsistency pointed out by Michael Attili
5362 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5363 was larger than the MD block size.
5364 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5366 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5367 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5368 using the passed key: if the passed key was a private key the result
5369 of X509_print(), for example, would be to print out all the private key
5373 *) des_quad_cksum() byte order bug fix.
5374 [Ulf Möller, using the problem description in krb4-0.9.7, where
5375 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5377 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5379 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5381 *) For easily testing in shell scripts whether some command
5382 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5383 returns with exit code 0 iff no command of the given name is available.
5384 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5385 the output goes to stdout and nothing is printed to stderr.
5386 Additional arguments are always ignored.
5388 Since for each cipher there is a command of the same name,
5389 the 'no-cipher' compilation switches can be tested this way.
5391 ('openssl no-XXX' is not able to detect pseudo-commands such
5392 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5395 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5398 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5399 is set; it will be thrown away anyway because each handshake creates
5401 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5402 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5403 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5404 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5407 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5408 'Q' and 'R' lose their special meanings (quit/renegotiate).
5409 This is part of what -quiet does; unlike -quiet, -ign_eof
5410 does not suppress any output.
5413 *) Add compatibility options to the purpose and trust code. The
5414 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5415 accepts a certificate or CA, this was the previous behaviour,
5416 with all the associated security issues.
5418 X509_TRUST_COMPAT is the old trust behaviour: only and
5419 automatically trust self signed roots in certificate store. A
5420 new trust setting X509_TRUST_DEFAULT is used to specify that
5421 a purpose has no associated trust setting and it should instead
5422 use the value in the default purpose.
5425 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5426 and fix a memory leak.
5429 *) In util/mkerr.pl (which implements 'make errors'), preserve
5430 reason strings from the previous version of the .c file, as
5431 the default to have only downcase letters (and digits) in
5432 automatically generated reasons codes is not always appropriate.
5435 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5436 using strerror. Previously, ERR_reason_error_string() returned
5437 library names as reason strings for SYSerr; but SYSerr is a special
5438 case where small numbers are errno values, not library numbers.
5441 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5442 converts DSA parameters into DH parameters. (When creating parameters,
5443 DSA_generate_parameters is used.)
5446 *) Include 'length' (recommended exponent length) in C code generated
5447 by 'openssl dhparam -C'.
5450 *) The second argument to set_label in perlasm was already being used
5451 so couldn't be used as a "file scope" flag. Moved to third argument
5455 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5456 instead of RAND_bytes for encryption IVs and salts.
5459 *) Include RAND_status() into RAND_METHOD instead of implementing
5460 it only for md_rand.c Otherwise replacing the PRNG by calling
5461 RAND_set_rand_method would be impossible.
5464 *) Don't let DSA_generate_key() enter an infinite loop if the random
5465 number generation fails.
5468 *) New 'rand' application for creating pseudo-random output.
5471 *) Added configuration support for Linux/IA64
5472 [Rolf Haberrecker <rolf@suse.de>]
5474 *) Assembler module support for Mingw32.
5477 *) Shared library support for HPUX (in shlib/).
5478 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5480 *) Shared library support for Solaris gcc.
5481 [Lutz Behnke <behnke@trustcenter.de>]
5483 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5485 *) PKCS7_encrypt() was adding text MIME headers twice because they
5486 were added manually and by SMIME_crlf_copy().
5489 *) In bntest.c don't call BN_rand with zero bits argument.
5490 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5492 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5493 case was implemented. This caused BN_div_recp() to fail occasionally.
5496 *) Add an optional second argument to the set_label() in the perl
5497 assembly language builder. If this argument exists and is set
5498 to 1 it signals that the assembler should use a symbol whose
5499 scope is the entire file, not just the current function. This
5500 is needed with MASM which uses the format label:: for this scope.
5501 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5503 *) Change the ASN1 types so they are typedefs by default. Before
5504 almost all types were #define'd to ASN1_STRING which was causing
5505 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5509 *) Change names of new functions to the new get1/get0 naming
5510 convention: After 'get1', the caller owns a reference count
5511 and has to call ..._free; 'get0' returns a pointer to some
5512 data structure without incrementing reference counters.
5513 (Some of the existing 'get' functions increment a reference
5514 counter, some don't.)
5515 Similarly, 'set1' and 'add1' functions increase reference
5516 counters or duplicate objects.
5519 *) Allow for the possibility of temp RSA key generation failure:
5520 the code used to assume it always worked and crashed on failure.
5523 *) Fix potential buffer overrun problem in BIO_printf().
5524 [Ulf Möller, using public domain code by Patrick Powell; problem
5525 pointed out by David Sacerdote <das33@cornell.edu>]
5527 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5528 RAND_egd() and RAND_status(). In the command line application,
5529 the EGD socket can be specified like a seed file using RANDFILE
5533 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5534 Some CAs (e.g. Verisign) distribute certificates in this form.
5537 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5538 list to exclude them. This means that no special compilation option
5539 is needed to use anonymous DH: it just needs to be included in the
5543 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5544 EVP_MD_type. The old functionality is available in a new macro called
5545 EVP_MD_md(). Change code that uses it and update docs.
5548 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5549 where the 'void *' argument is replaced by a function pointer argument.
5550 Previously 'void *' was abused to point to functions, which works on
5551 many platforms, but is not correct. As these functions are usually
5552 called by macros defined in OpenSSL header files, most source code
5553 should work without changes.
5556 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5557 sections with information on -D... compiler switches used for
5558 compiling the library so that applications can see them. To enable
5559 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5560 must be defined. E.g.,
5561 #define OPENSSL_ALGORITHM_DEFINES
5562 #include <openssl/opensslconf.h>
5563 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5564 [Richard Levitte, Ulf and Bodo Möller]
5566 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5570 *) Change the 'other' type in certificate aux info to a STACK_OF
5571 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5572 the required ASN1 format: arbitrary types determined by an OID.
5575 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5576 argument to 'req'. This is not because the function is newer or
5577 better than others it just uses the work 'NEW' in the certificate
5578 request header lines. Some software needs this.
5581 *) Reorganise password command line arguments: now passwords can be
5582 obtained from various sources. Delete the PEM_cb function and make
5583 it the default behaviour: i.e. if the callback is NULL and the
5584 usrdata argument is not NULL interpret it as a null terminated pass
5585 phrase. If usrdata and the callback are NULL then the pass phrase
5586 is prompted for as usual.
5589 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5590 the support is automatically enabled. The resulting binaries will
5591 autodetect the card and use it if present.
5592 [Ben Laurie and Compaq Inc.]
5594 *) Work around for Netscape hang bug. This sends certificate request
5595 and server done in one record. Since this is perfectly legal in the
5596 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5597 the bugs/SSLv3 entry for more info.
5600 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5603 *) Add -rand argument to smime and pkcs12 applications and read/write
5607 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5610 *) Add command line password options to the remaining applications.
5613 *) Bug fix for BN_div_recp() for numerators with an even number of
5617 *) More tests in bntest.c, and changed test_bn output.
5620 *) ./config recognizes MacOS X now.
5623 *) Bug fix for BN_div() when the first words of num and divsor are
5624 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5627 *) Add support for various broken PKCS#8 formats, and command line
5628 options to produce them.
5631 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5632 get temporary BIGNUMs from a BN_CTX.
5635 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5639 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5640 include a #define from the old name to the new. The original intent
5641 was that statically linked binaries could for example just call
5642 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5643 link with digests. This never worked becayse SSLeay_add_all_digests()
5644 and SSLeay_add_all_ciphers() were in the same source file so calling
5645 one would link with the other. They are now in separate source files.
5648 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5651 *) Use a less unusual form of the Miller-Rabin primality test (it used
5652 a binary algorithm for exponentiation integrated into the Miller-Rabin
5653 loop, our standard modexp algorithms are faster).
5656 *) Support for the EBCDIC character set completed.
5657 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5659 *) Source code cleanups: use const where appropriate, eliminate casts,
5660 use void * instead of char * in lhash.
5663 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5664 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5665 this the server could overwrite ephemeral keys that the client
5669 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5670 using 50 iterations of the Rabin-Miller test.
5672 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5673 iterations of the Rabin-Miller test as required by the appendix
5674 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5675 As BN_is_prime_fasttest includes trial division, DSA parameter
5676 generation becomes much faster.
5678 This implies a change for the callback functions in DSA_is_prime
5679 and DSA_generate_parameters: The callback function is called once
5680 for each positive witness in the Rabin-Miller test, not just
5681 occasionally in the inner loop; and the parameters to the
5682 callback function now provide an iteration count for the outer
5683 loop rather than for the current invocation of the inner loop.
5684 DSA_generate_parameters additionally can call the callback
5685 function with an 'iteration count' of -1, meaning that a
5686 candidate has passed the trial division test (when q is generated
5687 from an application-provided seed, trial division is skipped).
5690 *) New function BN_is_prime_fasttest that optionally does trial
5691 division before starting the Rabin-Miller test and has
5692 an additional BN_CTX * argument (whereas BN_is_prime always
5693 has to allocate at least one BN_CTX).
5694 'callback(1, -1, cb_arg)' is called when a number has passed the
5695 trial division stage.
5698 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5702 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5705 *) New function BN_pseudo_rand().
5708 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5709 bignum version of BN_from_montgomery() with the working code from
5710 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5714 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5715 made it impossible to use the same SSL_SESSION data structure in
5716 SSL2 clients in multiple threads.
5719 *) The return value of RAND_load_file() no longer counts bytes obtained
5720 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5721 to seed the PRNG (previously an explicit byte count was required).
5722 [Ulf Möller, Bodo Möller]
5724 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5725 used (char *) instead of (void *) and had casts all over the place.
5728 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5731 *) Retain source code compatibility for BN_prime_checks macro:
5732 BN_is_prime(..., BN_prime_checks, ...) now uses
5733 BN_prime_checks_for_size to determine the appropriate number of
5734 Rabin-Miller iterations.
5737 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5738 DH_CHECK_P_NOT_SAFE_PRIME.
5739 (Check if this is true? OpenPGP calls them "strong".)
5742 *) Merge the functionality of "dh" and "gendh" programs into a new program
5743 "dhparam". The old programs are retained for now but will handle DH keys
5744 (instead of parameters) in future.
5747 *) Make the ciphers, s_server and s_client programs check the return values
5748 when a new cipher list is set.
5751 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5752 ciphers. Before when the 56bit ciphers were enabled the sorting was
5755 The syntax for the cipher sorting has been extended to support sorting by
5756 cipher-strength (using the strength_bits hard coded in the tables).
5757 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5759 Fix a bug in the cipher-command parser: when supplying a cipher command
5760 string with an "undefined" symbol (neither command nor alphanumeric
5761 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5762 an error is flagged.
5764 Due to the strength-sorting extension, the code of the
5765 ssl_create_cipher_list() function was completely rearranged. I hope that
5766 the readability was also increased :-)
5767 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5769 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5770 for the first serial number and places 2 in the serial number file. This
5771 avoids problems when the root CA is created with serial number zero and
5772 the first user certificate has the same issuer name and serial number
5776 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5777 the new code. Add documentation for this stuff.
5780 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5781 X509_*() to X509at_*() on the grounds that they don't handle X509
5782 structures and behave in an analagous way to the X509v3 functions:
5783 they shouldn't be called directly but wrapper functions should be used
5786 So we also now have some wrapper functions that call the X509at functions
5787 when passed certificate requests. (TO DO: similar things can be done with
5788 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5789 things. Some of these need some d2i or i2d and print functionality
5790 because they handle more complex structures.)
5793 *) Add missing #ifndefs that caused missing symbols when building libssl
5794 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5795 NO_RSA in ssl/s2*.c.
5796 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5798 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5799 has a return value which indicates the quality of the random data
5800 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5801 error queue. New function RAND_pseudo_bytes() generates output that is
5802 guaranteed to be unique but not unpredictable. RAND_add is like
5803 RAND_seed, but takes an extra argument for an entropy estimate
5804 (RAND_seed always assumes full entropy).
5807 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5808 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5809 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5810 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5811 false-positive rate of at most 2^-80 for random input.
5814 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5817 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5818 in the 0.9.5 release), this returns the chain
5819 from an X509_CTX structure with a dup of the stack and all
5820 the X509 reference counts upped: so the stack will exist
5821 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5824 Also make SSL_SESSION_print() print out the verify return
5828 *) Add manpage for the pkcs12 command. Also change the default
5829 behaviour so MAC iteration counts are used unless the new
5830 -nomaciter option is used. This improves file security and
5831 only older versions of MSIE (4.0 for example) need it.
5834 *) Honor the no-xxx Configure options when creating .DEF files.
5837 *) Add PKCS#10 attributes to field table: challengePassword,
5838 unstructuredName and unstructuredAddress. These are taken from
5839 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5840 international characters are used.
5842 More changes to X509_ATTRIBUTE code: allow the setting of types
5843 based on strings. Remove the 'loc' parameter when adding
5844 attributes because these will be a SET OF encoding which is sorted
5848 *) Initial changes to the 'req' utility to allow request generation
5849 automation. This will allow an application to just generate a template
5850 file containing all the field values and have req construct the
5853 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5854 used all over the place including certificate requests and PKCS#7
5855 structures. They are currently handled manually where necessary with
5856 some primitive wrappers for PKCS#7. The new functions behave in a
5857 manner analogous to the X509 extension functions: they allow
5858 attributes to be looked up by NID and added.
5860 Later something similar to the X509V3 code would be desirable to
5861 automatically handle the encoding, decoding and printing of the
5862 more complex types. The string types like challengePassword can
5863 be handled by the string table functions.
5865 Also modified the multi byte string table handling. Now there is
5866 a 'global mask' which masks out certain types. The table itself
5867 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5868 is useful when for example there is only one permissible type
5869 (as in countryName) and using the mask might result in no valid
5873 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5874 SSL_get_peer_finished to allow applications to obtain the latest
5875 Finished messages sent to the peer or expected from the peer,
5876 respectively. (SSL_get_peer_finished is usually the Finished message
5877 actually received from the peer, otherwise the protocol will be aborted.)
5879 As the Finished message are message digests of the complete handshake
5880 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5881 be used for external authentication procedures when the authentication
5882 provided by SSL/TLS is not desired or is not enough.
5885 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5886 the host supports BWX extension and if Compaq C is present on the
5887 $PATH. Just exploiting of the BWX extension results in 20-30%
5888 performance kick for some algorithms, e.g. DES and RC4 to mention
5889 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5893 *) Add support for MS "fast SGC". This is arguably a violation of the
5894 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5895 weak crypto and after checking the certificate is SGC a second one
5896 with strong crypto. MS SGC stops the first handshake after receiving
5897 the server certificate message and sends a second client hello. Since
5898 a server will typically do all the time consuming operations before
5899 expecting any further messages from the client (server key exchange
5900 is the most expensive) there is little difference between the two.
5902 To get OpenSSL to support MS SGC we have to permit a second client
5903 hello message after we have sent server done. In addition we have to
5904 reset the MAC if we do get this second client hello.
5907 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5908 if a DER encoded private key is RSA or DSA traditional format. Changed
5909 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5910 format DER encoded private key. Newer code should use PKCS#8 format which
5911 has the key type encoded in the ASN1 structure. Added DER private key
5912 support to pkcs8 application.
5915 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5916 ciphersuites has been selected (as required by the SSL 3/TLS 1
5917 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5918 is set, we interpret this as a request to violate the specification
5919 (the worst that can happen is a handshake failure, and 'correct'
5920 behaviour would result in a handshake failure anyway).
5923 *) In SSL_CTX_add_session, take into account that there might be multiple
5924 SSL_SESSION structures with the same session ID (e.g. when two threads
5925 concurrently obtain them from an external cache).
5926 The internal cache can handle only one SSL_SESSION with a given ID,
5927 so if there's a conflict, we now throw out the old one to achieve
5931 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5932 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5933 some routines that use cipher OIDs: some ciphers do not have OIDs
5934 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5938 *) Simplify the trust setting structure and code. Now we just have
5939 two sequences of OIDs for trusted and rejected settings. These will
5940 typically have values the same as the extended key usage extension
5941 and any application specific purposes.
5943 The trust checking code now has a default behaviour: it will just
5944 check for an object with the same NID as the passed id. Functions can
5945 be provided to override either the default behaviour or the behaviour
5946 for a given id. SSL client, server and email already have functions
5947 in place for compatibility: they check the NID and also return "trusted"
5948 if the certificate is self signed.
5951 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5952 traditional format into an EVP_PKEY structure.
5955 *) Add a password callback function PEM_cb() which either prompts for
5956 a password if usr_data is NULL or otherwise assumes it is a null
5957 terminated password. Allow passwords to be passed on command line
5958 environment or config files in a few more utilities.
5961 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5962 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5963 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5964 Update documentation.
5967 *) Support for ASN1 "NULL" type. This could be handled before by using
5968 ASN1_TYPE but there wasn't any function that would try to read a NULL
5969 and produce an error if it couldn't. For compatibility we also have
5970 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5971 don't allocate anything because they don't need to.
5974 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5976 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5978 *) Rebuild of the memory allocation routines used by OpenSSL code and
5979 possibly others as well. The purpose is to make an interface that
5980 provide hooks so anyone can build a separate set of allocation and
5981 deallocation routines to be used by OpenSSL, for example memory
5982 pool implementations, or something else, which was previously hard
5983 since Malloc(), Realloc() and Free() were defined as macros having
5984 the values malloc, realloc and free, respectively (except for Win32
5985 compilations). The same is provided for memory debugging code.
5986 OpenSSL already comes with functionality to find memory leaks, but
5987 this gives people a chance to debug other memory problems.
5989 With these changes, a new set of functions and macros have appeared:
5991 CRYPTO_set_mem_debug_functions() [F]
5992 CRYPTO_get_mem_debug_functions() [F]
5993 CRYPTO_dbg_set_options() [F]
5994 CRYPTO_dbg_get_options() [F]
5995 CRYPTO_malloc_debug_init() [M]
5997 The memory debug functions are NULL by default, unless the library
5998 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5999 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6000 gives the standard debugging functions that come with OpenSSL) or
6001 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6002 provided by the library user) must be used. When the standard
6003 debugging functions are used, CRYPTO_dbg_set_options can be used to
6004 request additional information:
6005 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6006 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6008 Also, things like CRYPTO_set_mem_functions will always give the
6009 expected result (the new set of functions is used for allocation
6010 and deallocation) at all times, regardless of platform and compiler
6013 To finish it up, some functions that were never use in any other
6014 way than through macros have a new API and new semantic:
6017 CRYPTO_dbg_realloc()
6020 All macros of value have retained their old syntax.
6021 [Richard Levitte and Bodo Moeller]
6023 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6024 ordering of SMIMECapabilities wasn't in "strength order" and there
6025 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6029 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6030 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6031 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6033 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6034 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6035 functionality to handle multipart/signed properly) and a utility
6036 called 'smime' to call all this stuff. This is based on code I
6037 originally wrote for Celo who have kindly allowed it to be
6038 included in OpenSSL.
6041 *) Add variants des_set_key_checked and des_set_key_unchecked of
6042 des_set_key (aka des_key_sched). Global variable des_check_key
6043 decides which of these is called by des_set_key; this way
6044 des_check_key behaves as it always did, but applications and
6045 the library itself, which was buggy for des_check_key == 1,
6046 have a cleaner way to pick the version they need.
6049 *) New function PKCS12_newpass() which changes the password of a
6053 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6054 dynamic mix. In both cases the ids can be used as an index into the
6055 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6056 functions so they accept a list of the field values and the
6057 application doesn't need to directly manipulate the X509_TRUST
6061 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6065 *) Modify the way the V3 extension code looks up extensions. This now
6066 works in a similar way to the object code: we have some "standard"
6067 extensions in a static table which is searched with OBJ_bsearch()
6068 and the application can add dynamic ones if needed. The file
6069 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6070 updated whenever a new extension is added to the core code and kept
6071 in ext_nid order. There is a simple program 'tabtest.c' which checks
6072 this. New extensions are not added too often so this file can readily
6073 be maintained manually.
6075 There are two big advantages in doing things this way. The extensions
6076 can be looked up immediately and no longer need to be "added" using
6077 X509V3_add_standard_extensions(): this function now does nothing.
6078 [Side note: I get *lots* of email saying the extension code doesn't
6079 work because people forget to call this function]
6080 Also no dynamic allocation is done unless new extensions are added:
6081 so if we don't add custom extensions there is no need to call
6082 X509V3_EXT_cleanup().
6085 *) Modify enc utility's salting as follows: make salting the default. Add a
6086 magic header, so unsalted files fail gracefully instead of just decrypting
6087 to garbage. This is because not salting is a big security hole, so people
6088 should be discouraged from doing it.
6091 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6092 digest to be passed on the command line but it only used this
6093 parameter when signing a certificate. Modified so all relevant
6094 operations are affected by the digest parameter including the
6095 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6096 DSA key was used because it didn't fix the digest.
6099 *) Initial certificate chain verify code. Currently tests the untrusted
6100 certificates for consistency with the verify purpose (which is set
6101 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6103 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6104 this is because it will reject chains with invalid extensions whereas
6105 every previous version of OpenSSL and SSLeay made no checks at all.
6107 Trust code: checks the root CA for the relevant trust settings. Trust
6108 settings have an initial value consistent with the verify purpose: e.g.
6109 if the verify purpose is for SSL client use it expects the CA to be
6110 trusted for SSL client use. However the default value can be changed to
6111 permit custom trust settings: one example of this would be to only trust
6112 certificates from a specific "secure" set of CAs.
6114 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6115 which should be used for version portability: especially since the
6116 verify structure is likely to change more often now.
6118 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6119 to set them. If not set then assume SSL clients will verify SSL servers
6122 Two new options to the verify program: -untrusted allows a set of
6123 untrusted certificates to be passed in and -purpose which sets the
6124 intended purpose of the certificate. If a purpose is set then the
6125 new chain verify code is used to check extension consistency.
6128 *) Support for the authority information access extension.
6131 *) Modify RSA and DSA PEM read routines to transparently handle
6132 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6133 public keys in a format compatible with certificate
6134 SubjectPublicKeyInfo structures. Unfortunately there were already
6135 functions called *_PublicKey_* which used various odd formats so
6136 these are retained for compatibility: however the DSA variants were
6137 never in a public release so they have been deleted. Changed dsa/rsa
6138 utilities to handle the new format: note no releases ever handled public
6139 keys so we should be OK.
6141 The primary motivation for this change is to avoid the same fiasco
6142 that dogs private keys: there are several incompatible private key
6143 formats some of which are standard and some OpenSSL specific and
6144 require various evil hacks to allow partial transparent handling and
6145 even then it doesn't work with DER formats. Given the option anything
6146 other than PKCS#8 should be dumped: but the other formats have to
6147 stay in the name of compatibility.
6149 With public keys and the benefit of hindsight one standard format
6150 is used which works with EVP_PKEY, RSA or DSA structures: though
6151 it clearly returns an error if you try to read the wrong kind of key.
6153 Added a -pubkey option to the 'x509' utility to output the public key.
6154 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6155 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6156 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6157 that do the same as the EVP_PKEY_assign_*() except they up the
6158 reference count of the added key (they don't "swallow" the
6162 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6163 CRLs would fail if the file contained no certificates or no CRLs:
6164 added a new function to read in both types and return the number
6165 read: this means that if none are read it will be an error. The
6166 DER versions of the certificate and CRL reader would always fail
6167 because it isn't possible to mix certificates and CRLs in DER format
6168 without choking one or the other routine. Changed this to just read
6169 a certificate: this is the best we can do. Also modified the code
6170 in apps/verify.c to take notice of return codes: it was previously
6171 attempting to read in certificates from NULL pointers and ignoring
6172 any errors: this is one reason why the cert and CRL reader seemed
6173 to work. It doesn't check return codes from the default certificate
6174 routines: these may well fail if the certificates aren't installed.
6177 *) Code to support otherName option in GeneralName.
6180 *) First update to verify code. Change the verify utility
6181 so it warns if it is passed a self signed certificate:
6182 for consistency with the normal behaviour. X509_verify
6183 has been modified to it will now verify a self signed
6184 certificate if *exactly* the same certificate appears
6185 in the store: it was previously impossible to trust a
6186 single self signed certificate. This means that:
6187 openssl verify ss.pem
6188 now gives a warning about a self signed certificate but
6189 openssl verify -CAfile ss.pem ss.pem
6193 *) For servers, store verify_result in SSL_SESSION data structure
6194 (and add it to external session representation).
6195 This is needed when client certificate verifications fails,
6196 but an application-provided verification callback (set by
6197 SSL_CTX_set_cert_verify_callback) allows accepting the session
6198 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6199 but returns 1): When the session is reused, we have to set
6200 ssl->verify_result to the appropriate error code to avoid
6202 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6204 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6205 case in PKCS7_dataInit() where the signed PKCS7 structure
6206 didn't contain any existing data because it was being created.
6207 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6209 *) Add a salt to the key derivation routines in enc.c. This
6210 forms the first 8 bytes of the encrypted file. Also add a
6211 -S option to allow a salt to be input on the command line.
6214 *) New function X509_cmp(). Oddly enough there wasn't a function
6215 to compare two certificates. We do this by working out the SHA1
6216 hash and comparing that. X509_cmp() will be needed by the trust
6220 *) SSL_get1_session() is like SSL_get_session(), but increments
6221 the reference count in the SSL_SESSION returned.
6222 [Geoff Thorpe <geoff@eu.c2.net>]
6224 *) Fix for 'req': it was adding a null to request attributes.
6225 Also change the X509_LOOKUP and X509_INFO code to handle
6226 certificate auxiliary information.
6229 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6233 *) Add the possibility to add extra information to the memory leak
6234 detecting output, to form tracebacks, showing from where each
6235 allocation was originated: CRYPTO_push_info("constant string") adds
6236 the string plus current file name and line number to a per-thread
6237 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6238 is like calling CYRPTO_pop_info() until the stack is empty.
6239 Also updated memory leak detection code to be multi-thread-safe.
6242 *) Add options -text and -noout to pkcs7 utility and delete the
6243 encryption options which never did anything. Update docs.
6246 *) Add options to some of the utilities to allow the pass phrase
6247 to be included on either the command line (not recommended on
6248 OSes like Unix) or read from the environment. Update the
6249 manpages and fix a few bugs.
6252 *) Add a few manpages for some of the openssl commands.
6255 *) Fix the -revoke option in ca. It was freeing up memory twice,
6256 leaking and not finding already revoked certificates.
6259 *) Extensive changes to support certificate auxiliary information.
6260 This involves the use of X509_CERT_AUX structure and X509_AUX
6261 functions. An X509_AUX function such as PEM_read_X509_AUX()
6262 can still read in a certificate file in the usual way but it
6263 will also read in any additional "auxiliary information". By
6264 doing things this way a fair degree of compatibility can be
6265 retained: existing certificates can have this information added
6266 using the new 'x509' options.
6268 Current auxiliary information includes an "alias" and some trust
6269 settings. The trust settings will ultimately be used in enhanced
6270 certificate chain verification routines: currently a certificate
6271 can only be trusted if it is self signed and then it is trusted
6275 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6276 The problem was that one of the replacement routines had not been working
6277 since SSLeay releases. For now the offending routine has been replaced
6278 with non-optimised assembler. Even so, this now gives around 95%
6279 performance improvement for 1024 bit RSA signs.
6282 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6283 handling. Most clients have the effective key size in bits equal to
6284 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6285 A few however don't do this and instead use the size of the decrypted key
6286 to determine the RC2 key length and the AlgorithmIdentifier to determine
6287 the effective key length. In this case the effective key length can still
6288 be 40 bits but the key length can be 168 bits for example. This is fixed
6289 by manually forcing an RC2 key into the EVP_PKEY structure because the
6290 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6291 the key length and effective key length are equal.
6294 *) Add a bunch of functions that should simplify the creation of
6295 X509_NAME structures. Now you should be able to do:
6296 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6297 and have it automatically work out the correct field type and fill in
6298 the structures. The more adventurous can try:
6299 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6300 and it will (hopefully) work out the correct multibyte encoding.
6303 *) Change the 'req' utility to use the new field handling and multibyte
6304 copy routines. Before the DN field creation was handled in an ad hoc
6305 way in req, ca, and x509 which was rather broken and didn't support
6306 BMPStrings or UTF8Strings. Since some software doesn't implement
6307 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6308 using the dirstring_type option. See the new comment in the default
6309 openssl.cnf for more info.
6312 *) Make crypto/rand/md_rand.c more robust:
6313 - Assure unique random numbers after fork().
6314 - Make sure that concurrent threads access the global counter and
6315 md serializably so that we never lose entropy in them
6316 or use exactly the same state in multiple threads.
6317 Access to the large state is not always serializable because
6318 the additional locking could be a performance killer, and
6319 md should be large enough anyway.
6322 *) New file apps/app_rand.c with commonly needed functionality
6323 for handling the random seed file.
6325 Use the random seed file in some applications that previously did not:
6327 dsaparam -genkey (which also ignored its '-rand' option),
6330 x509 (when signing).
6331 Except on systems with /dev/urandom, it is crucial to have a random
6332 seed file at least for key creation, DSA signing, and for DH exchanges;
6333 for RSA signatures we could do without one.
6335 gendh and gendsa (unlike genrsa) used to read only the first byte
6336 of each file listed in the '-rand' option. The function as previously
6337 found in genrsa is now in app_rand.c and is used by all programs
6338 that support '-rand'.
6341 *) In RAND_write_file, use mode 0600 for creating files;
6342 don't just chmod when it may be too late.
6345 *) Report an error from X509_STORE_load_locations
6346 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6349 *) New function ASN1_mbstring_copy() this copies a string in either
6350 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6351 into an ASN1_STRING type. A mask of permissible types is passed
6352 and it chooses the "minimal" type to use or an error if not type
6356 *) Add function equivalents to the various macros in asn1.h. The old
6357 macros are retained with an M_ prefix. Code inside the library can
6358 use the M_ macros. External code (including the openssl utility)
6359 should *NOT* in order to be "shared library friendly".
6362 *) Add various functions that can check a certificate's extensions
6363 to see if it usable for various purposes such as SSL client,
6364 server or S/MIME and CAs of these types. This is currently
6365 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6366 verification. Also added a -purpose flag to x509 utility to
6367 print out all the purposes.
6370 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6374 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6375 for, obtain and decode and extension and obtain its critical flag.
6376 This allows all the necessary extension code to be handled in a
6377 single function call.
6380 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6381 platforms. See crypto/rc4/rc4_enc.c for further details.
6384 *) New -noout option to asn1parse. This causes no output to be produced
6385 its main use is when combined with -strparse and -out to extract data
6386 from a file (which may not be in ASN.1 format).
6389 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6390 when producing the local key id.
6391 [Richard Levitte <levitte@stacken.kth.se>]
6393 *) New option -dhparam in s_server. This allows a DH parameter file to be
6394 stated explicitly. If it is not stated then it tries the first server
6395 certificate file. The previous behaviour hard coded the filename
6399 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6400 a public key to be input or output. For example:
6401 openssl rsa -in key.pem -pubout -out pubkey.pem
6402 Also added necessary DSA public key functions to handle this.
6405 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6406 in the message. This was handled by allowing
6407 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6408 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6410 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6411 to the end of the strings whereas this didn't. This would cause problems
6412 if strings read with d2i_ASN1_bytes() were later modified.
6413 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6415 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6416 data and it contains EOF it will end up returning an error. This is
6417 caused by input 46 bytes long. The cause is due to the way base64
6418 BIOs find the start of base64 encoded data. They do this by trying a
6419 trial decode on each line until they find one that works. When they
6420 do a flag is set and it starts again knowing it can pass all the
6421 data directly through the decoder. Unfortunately it doesn't reset
6422 the context it uses. This means that if EOF is reached an attempt
6423 is made to pass two EOFs through the context and this causes the
6424 resulting error. This can also cause other problems as well. As is
6425 usual with these problems it takes *ages* to find and the fix is
6426 trivial: move one line.
6427 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6429 *) Ugly workaround to get s_client and s_server working under Windows. The
6430 old code wouldn't work because it needed to select() on sockets and the
6431 tty (for keypresses and to see if data could be written). Win32 only
6432 supports select() on sockets so we select() with a 1s timeout on the
6433 sockets and then see if any characters are waiting to be read, if none
6434 are present then we retry, we also assume we can always write data to
6435 the tty. This isn't nice because the code then blocks until we've
6436 received a complete line of data and it is effectively polling the
6437 keyboard at 1s intervals: however it's quite a bit better than not
6438 working at all :-) A dedicated Windows application might handle this
6439 with an event loop for example.
6442 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6443 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6444 will be called when RSA_sign() and RSA_verify() are used. This is useful
6445 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6446 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6447 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6448 This necessitated the support of an extra signature type NID_md5_sha1
6449 for SSL signatures and modifications to the SSL library to use it instead
6450 of calling RSA_public_decrypt() and RSA_private_encrypt().
6453 *) Add new -verify -CAfile and -CApath options to the crl program, these
6454 will lookup a CRL issuers certificate and verify the signature in a
6455 similar way to the verify program. Tidy up the crl program so it
6456 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6457 less strict. It will now permit CRL extensions even if it is not
6458 a V2 CRL: this will allow it to tolerate some broken CRLs.
6461 *) Initialize all non-automatic variables each time one of the openssl
6462 sub-programs is started (this is necessary as they may be started
6463 multiple times from the "OpenSSL>" prompt).
6464 [Lennart Bang, Bodo Moeller]
6466 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6467 removing all other RSA functionality (this is what NO_RSA does). This
6468 is so (for example) those in the US can disable those operations covered
6469 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6473 *) Non-copying interface to BIO pairs.
6474 (still largely untested)
6477 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6478 ASCII string. This was handled independently in various places before.
6481 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6482 UTF8 strings a character at a time.
6485 *) Use client_version from client hello to select the protocol
6486 (s23_srvr.c) and for RSA client key exchange verification
6487 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6490 *) Add various utility functions to handle SPKACs, these were previously
6491 handled by poking round in the structure internals. Added new function
6492 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6493 print, verify and generate SPKACs. Based on an original idea from
6494 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6497 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6500 *) Allow the config file extension section to be overwritten on the
6501 command line. Based on an original idea from Massimiliano Pala
6502 <madwolf@comune.modena.it>. The new option is called -extensions
6503 and can be applied to ca, req and x509. Also -reqexts to override
6504 the request extensions in req and -crlexts to override the crl extensions
6508 *) Add new feature to the SPKAC handling in ca. Now you can include
6509 the same field multiple times by preceding it by "XXXX." for example:
6512 this is the same syntax as used in the req config file.
6515 *) Allow certificate extensions to be added to certificate requests. These
6516 are specified in a 'req_extensions' option of the req section of the
6517 config file. They can be printed out with the -text option to req but
6518 are otherwise ignored at present.
6521 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6522 data read consists of only the final block it would not decrypted because
6523 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6524 A misplaced 'break' also meant the decrypted final block might not be
6525 copied until the next read.
6528 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6529 a few extra parameters to the DH structure: these will be useful if
6530 for example we want the value of 'q' or implement X9.42 DH.
6533 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6534 provides hooks that allow the default DSA functions or functions on a
6535 "per key" basis to be replaced. This allows hardware acceleration and
6536 hardware key storage to be handled without major modification to the
6537 library. Also added low level modexp hooks and CRYPTO_EX structure and
6538 associated functions.
6541 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6542 as "read only": it can't be written to and the buffer it points to will
6543 not be freed. Reading from a read only BIO is much more efficient than
6544 a normal memory BIO. This was added because there are several times when
6545 an area of memory needs to be read from a BIO. The previous method was
6546 to create a memory BIO and write the data to it, this results in two
6547 copies of the data and an O(n^2) reading algorithm. There is a new
6548 function BIO_new_mem_buf() which creates a read only memory BIO from
6549 an area of memory. Also modified the PKCS#7 routines to use read only
6553 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6554 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6555 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6556 but a retry condition occured while trying to read the rest.
6559 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6560 NID_pkcs7_encrypted by default: this was wrong since this should almost
6561 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6562 the encrypted data type: this is a more sensible place to put it and it
6563 allows the PKCS#12 code to be tidied up that duplicated this
6567 *) Changed obj_dat.pl script so it takes its input and output files on
6568 the command line. This should avoid shell escape redirection problems
6572 *) Initial support for certificate extension requests, these are included
6573 in things like Xenroll certificate requests. Included functions to allow
6574 extensions to be obtained and added.
6577 *) -crlf option to s_client and s_server for sending newlines as
6578 CRLF (as required by many protocols).
6581 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6583 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6584 [Ralf S. Engelschall]
6586 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6587 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6589 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6593 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6594 DH parameters/keys (q is lost during that conversion, but the resulting
6595 DH parameters contain its length).
6597 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6598 much faster than DH_generate_parameters (which creates parameters
6599 where p = 2*q + 1), and also the smaller q makes DH computations
6600 much more efficient (160-bit exponentiation instead of 1024-bit
6601 exponentiation); so this provides a convenient way to support DHE
6602 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6603 utter importance to use
6604 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6606 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6607 when such DH parameters are used, because otherwise small subgroup
6608 attacks may become possible!
6611 *) Avoid memory leak in i2d_DHparams.
6614 *) Allow the -k option to be used more than once in the enc program:
6615 this allows the same encrypted message to be read by multiple recipients.
6618 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6619 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6620 it will always use the numerical form of the OID, even if it has a short
6624 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6625 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6626 otherwise bn_mod_exp was called. In the case of hardware keys for example
6627 no private key components need be present and it might store extra data
6628 in the RSA structure, which cannot be accessed from bn_mod_exp.
6629 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6630 private key operations.
6633 *) Added support for SPARC Linux.
6636 *) pem_password_cb function type incompatibly changed from
6637 typedef int pem_password_cb(char *buf, int size, int rwflag);
6639 ....(char *buf, int size, int rwflag, void *userdata);
6640 so that applications can pass data to their callbacks:
6641 The PEM[_ASN1]_{read,write}... functions and macros now take an
6642 additional void * argument, which is just handed through whenever
6643 the password callback is called.
6644 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6646 New function SSL_CTX_set_default_passwd_cb_userdata.
6648 Compatibility note: As many C implementations push function arguments
6649 onto the stack in reverse order, the new library version is likely to
6650 interoperate with programs that have been compiled with the old
6651 pem_password_cb definition (PEM_whatever takes some data that
6652 happens to be on the stack as its last argument, and the callback
6653 just ignores this garbage); but there is no guarantee whatsoever that
6656 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6657 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6658 problems not only on Windows, but also on some Unix platforms.
6659 To avoid problematic command lines, these definitions are now in an
6660 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6661 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6664 *) MIPS III/IV assembler module is reimplemented.
6667 *) More DES library cleanups: remove references to srand/rand and
6668 delete an unused file.
6671 *) Add support for the the free Netwide assembler (NASM) under Win32,
6672 since not many people have MASM (ml) and it can be hard to obtain.
6673 This is currently experimental but it seems to work OK and pass all
6674 the tests. Check out INSTALL.W32 for info.
6677 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6678 without temporary keys kept an extra copy of the server key,
6679 and connections with temporary keys did not free everything in case
6683 *) New function RSA_check_key and new openssl rsa option -check
6684 for verifying the consistency of RSA keys.
6685 [Ulf Moeller, Bodo Moeller]
6687 *) Various changes to make Win32 compile work:
6688 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6689 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6690 comparison" warnings.
6691 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6694 *) Add a debugging option to PKCS#5 v2 key generation function: when
6695 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6696 derived keys are printed to stderr.
6699 *) Copy the flags in ASN1_STRING_dup().
6700 [Roman E. Pavlov <pre@mo.msk.ru>]
6702 *) The x509 application mishandled signing requests containing DSA
6703 keys when the signing key was also DSA and the parameters didn't match.
6705 It was supposed to omit the parameters when they matched the signing key:
6706 the verifying software was then supposed to automatically use the CA's
6707 parameters if they were absent from the end user certificate.
6709 Omitting parameters is no longer recommended. The test was also
6710 the wrong way round! This was probably due to unusual behaviour in
6711 EVP_cmp_parameters() which returns 1 if the parameters match.
6712 This meant that parameters were omitted when they *didn't* match and
6713 the certificate was useless. Certificates signed with 'ca' didn't have
6715 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6717 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6718 The interface is as follows:
6719 Applications can use
6720 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6721 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6722 "off" is now the default.
6723 The library internally uses
6724 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6725 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6726 to disable memory-checking temporarily.
6728 Some inconsistent states that previously were possible (and were
6729 even the default) are now avoided.
6731 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6732 with each memory chunk allocated; this is occasionally more helpful
6733 than just having a counter.
6735 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6737 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6741 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6742 which largely parallels "options", but is for changing API behaviour,
6743 whereas "options" are about protocol behaviour.
6744 Initial "mode" flags are:
6746 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6747 a single record has been written.
6748 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6749 retries use the same buffer location.
6750 (But all of the contents must be
6754 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6757 *) Fix problems with no-hmac etc.
6758 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6760 *) New functions RSA_get_default_method(), RSA_set_method() and
6761 RSA_get_method(). These allows replacement of RSA_METHODs without having
6762 to mess around with the internals of an RSA structure.
6765 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6766 Also really enable memory leak checks in openssl.c and in some
6768 [Chad C. Mulligan, Bodo Moeller]
6770 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6771 up the length of negative integers. This has now been simplified to just
6772 store the length when it is first determined and use it later, rather
6773 than trying to keep track of where data is copied and updating it to
6775 [Steve Henson, reported by Brien Wheeler
6776 <bwheeler@authentica-security.com>]
6778 *) Add a new function PKCS7_signatureVerify. This allows the verification
6779 of a PKCS#7 signature but with the signing certificate passed to the
6780 function itself. This contrasts with PKCS7_dataVerify which assumes the
6781 certificate is present in the PKCS#7 structure. This isn't always the
6782 case: certificates can be omitted from a PKCS#7 structure and be
6783 distributed by "out of band" means (such as a certificate database).
6786 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6787 function prototypes in pem.h, also change util/mkdef.pl to add the
6788 necessary function names.
6791 *) mk1mf.pl (used by Windows builds) did not properly read the
6792 options set by Configure in the top level Makefile, and Configure
6793 was not even able to write more than one option correctly.
6794 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6797 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6798 file to be loaded from a BIO or FILE pointer. The BIO version will
6799 for example allow memory BIOs to contain config info.
6802 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6803 Whoever hopes to achieve shared-library compatibility across versions
6804 must use this, not the compile-time macro.
6805 (Exercise 0.9.4: Which is the minimum library version required by
6807 Note: All this applies only to multi-threaded programs, others don't
6811 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6812 through a BIO pair triggered the default case, i.e.
6813 SSLerr(...,SSL_R_UNKNOWN_STATE).
6816 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6817 can use the SSL library even if none of the specific BIOs is
6821 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6822 for the encoded length.
6823 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6825 *) Add initial documentation of the X509V3 functions.
6828 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6829 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6830 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6831 secure PKCS#8 private key format with a high iteration count.
6834 *) Fix determination of Perl interpreter: A perl or perl5
6835 _directory_ in $PATH was also accepted as the interpreter.
6836 [Ralf S. Engelschall]
6838 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6839 wrong with it but it was very old and did things like calling
6840 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6844 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6845 to use the new extension code.
6848 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6849 with macros. This should make it easier to change their form, add extra
6850 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6854 *) Add to configuration table a new entry that can specify an alternative
6855 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6856 according to Mark Crispin <MRC@Panda.COM>.
6860 *) DES CBC did not update the IV. Weird.
6863 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6864 Changing the behaviour of the former might break existing programs --
6865 where IV updating is needed, des_ncbc_encrypt can be used.
6868 *) When bntest is run from "make test" it drives bc to check its
6869 calculations, as well as internally checking them. If an internal check
6870 fails, it needs to cause bc to give a non-zero result or make test carries
6871 on without noticing the failure. Fixed.
6874 *) DES library cleanups.
6877 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6878 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6879 ciphers. NOTE: although the key derivation function has been verified
6880 against some published test vectors it has not been extensively tested
6881 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6885 *) Instead of "mkdir -p", which is not fully portable, use new
6886 Perl script "util/mkdir-p.pl".
6889 *) Rewrite the way password based encryption (PBE) is handled. It used to
6890 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6891 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6892 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6893 the 'parameter' field of the AlgorithmIdentifier is passed to the
6894 underlying key generation function so it must do its own ASN1 parsing.
6895 This has also changed the EVP_PBE_CipherInit() function which now has a
6896 'parameter' argument instead of literal salt and iteration count values
6897 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6900 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6901 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6902 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6903 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6904 value was just used as a "magic string" and not used directly its
6905 value doesn't matter.
6908 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6912 *) "linux-sparc64" configuration (ultrapenguin).
6913 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6914 "linux-sparc" configuration.
6915 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6917 *) config now generates no-xxx options for missing ciphers.
6920 *) Support the EBCDIC character set (work in progress).
6921 File ebcdic.c not yet included because it has a different license.
6922 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6924 *) Support BS2000/OSD-POSIX.
6925 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6927 *) Make callbacks for key generation use void * instead of char *.
6930 *) Make S/MIME samples compile (not yet tested).
6933 *) Additional typesafe stacks.
6936 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6940 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6942 *) New configuration variant "sco5-gcc".
6944 *) Updated some demos.
6945 [Sean O Riordain, Wade Scholine]
6947 *) Add missing BIO_free at exit of pkcs12 application.
6950 *) Fix memory leak in conf.c.
6953 *) Updates for Win32 to assembler version of MD5.
6956 *) Set #! path to perl in apps/der_chop to where we found it
6957 instead of using a fixed path.
6960 *) SHA library changes for irix64-mips4-cc.
6963 *) Improvements for VMS support.
6967 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6969 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6970 This also avoids the problems with SC4.2 and unpatched SC5.
6971 [Andy Polyakov <appro@fy.chalmers.se>]
6973 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6974 These are required because of the typesafe stack would otherwise break
6975 existing code. If old code used a structure member which used to be STACK
6976 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6977 sk_num or sk_value it would produce an error because the num, data members
6978 are not present in STACK_OF. Now it just produces a warning. sk_set
6979 replaces the old method of assigning a value to sk_value
6980 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6981 that does this will no longer work (and should use sk_set instead) but
6982 this could be regarded as a "questionable" behaviour anyway.
6985 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6986 correctly handle encrypted S/MIME data.
6989 *) Change type of various DES function arguments from des_cblock
6990 (which means, in function argument declarations, pointer to char)
6991 to des_cblock * (meaning pointer to array with 8 char elements),
6992 which allows the compiler to do more typechecking; it was like
6993 that back in SSLeay, but with lots of ugly casts.
6995 Introduce new type const_des_cblock.
6998 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6999 problems: find RecipientInfo structure that matches recipient certificate
7000 and initialise the ASN1 structures properly based on passed cipher.
7003 *) Belatedly make the BN tests actually check the results.
7006 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7007 to and from BNs: it was completely broken. New compilation option
7008 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7009 key elements as negative integers.
7012 *) Reorganize and speed up MD5.
7013 [Andy Polyakov <appro@fy.chalmers.se>]
7016 [Richard Levitte <richard@levitte.org>]
7018 *) New option -out to asn1parse to allow the parsed structure to be
7019 output to a file. This is most useful when combined with the -strparse
7020 option to examine the output of things like OCTET STRINGS.
7023 *) Make SSL library a little more fool-proof by not requiring any longer
7024 that SSL_set_{accept,connect}_state be called before
7025 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7026 in many applications because usually everything *appeared* to work as
7027 intended anyway -- now it really works as intended).
7030 *) Move openssl.cnf out of lib/.
7033 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7034 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7035 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7036 [Ralf S. Engelschall]
7038 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7039 handle PKCS#7 enveloped data properly.
7040 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7042 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7043 copying pointers. The cert_st handling is changed by this in
7044 various ways (and thus what used to be known as ctx->default_cert
7045 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7046 any longer when s->cert does not give us what we need).
7047 ssl_cert_instantiate becomes obsolete by this change.
7048 As soon as we've got the new code right (possibly it already is?),
7049 we have solved a couple of bugs of the earlier code where s->cert
7050 was used as if it could not have been shared with other SSL structures.
7052 Note that using the SSL API in certain dirty ways now will result
7053 in different behaviour than observed with earlier library versions:
7054 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7055 does not influence s as it used to.
7057 In order to clean up things more thoroughly, inside SSL_SESSION
7058 we don't use CERT any longer, but a new structure SESS_CERT
7059 that holds per-session data (if available); currently, this is
7060 the peer's certificate chain and, for clients, the server's certificate
7061 and temporary key. CERT holds only those values that can have
7062 meaningful defaults in an SSL_CTX.
7065 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7066 from the internal representation. Various PKCS#7 fixes: remove some
7067 evil casts and set the enc_dig_alg field properly based on the signing
7071 *) Allow PKCS#12 password to be set from the command line or the
7072 environment. Let 'ca' get its config file name from the environment
7073 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7077 *) Allow certificate policies extension to use an IA5STRING for the
7078 organization field. This is contrary to the PKIX definition but
7079 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7083 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7084 without disallowing inline assembler and the like for non-pedantic builds.
7087 *) Support Borland C++ builder.
7088 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7093 *) SHA-1 cleanups and performance enhancements.
7094 [Andy Polyakov <appro@fy.chalmers.se>]
7096 *) Sparc v8plus assembler for the bignum library.
7097 [Andy Polyakov <appro@fy.chalmers.se>]
7099 *) Accept any -xxx and +xxx compiler options in Configure.
7102 *) Update HPUX configuration.
7105 *) Add missing sk_<type>_unshift() function to safestack.h
7106 [Ralf S. Engelschall]
7108 *) New function SSL_CTX_use_certificate_chain_file that sets the
7109 "extra_cert"s in addition to the certificate. (This makes sense
7110 only for "PEM" format files, as chains as a whole are not
7114 *) Support verify_depth from the SSL API.
7115 x509_vfy.c had what can be considered an off-by-one-error:
7116 Its depth (which was not part of the external interface)
7117 was actually counting the number of certificates in a chain;
7118 now it really counts the depth.
7121 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7122 instead of X509err, which often resulted in confusing error
7123 messages since the error codes are not globally unique
7124 (e.g. an alleged error in ssl3_accept when a certificate
7125 didn't match the private key).
7127 *) New function SSL_CTX_set_session_id_context that allows to set a default
7128 value (so that you don't need SSL_set_session_id_context for each
7129 connection using the SSL_CTX).
7132 *) OAEP decoding bug fix.
7135 *) Support INSTALL_PREFIX for package builders, as proposed by
7139 *) New Configure options "threads" and "no-threads". For systems
7140 where the proper compiler options are known (currently Solaris
7141 and Linux), "threads" is the default.
7144 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7147 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7148 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7149 such as /usr/local/bin.
7152 *) "make linux-shared" to build shared libraries.
7153 [Niels Poppe <niels@netbox.org>]
7155 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7158 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7159 extension adding in x509 utility.
7162 *) Remove NOPROTO sections and error code comments.
7165 *) Partial rewrite of the DEF file generator to now parse the ANSI
7169 *) New Configure options --prefix=DIR and --openssldir=DIR.
7172 *) Complete rewrite of the error code script(s). It is all now handled
7173 by one script at the top level which handles error code gathering,
7174 header rewriting and C source file generation. It should be much better
7175 than the old method: it now uses a modified version of Ulf's parser to
7176 read the ANSI prototypes in all header files (thus the old K&R definitions
7177 aren't needed for error creation any more) and do a better job of
7178 translating function codes into names. The old 'ASN1 error code imbedded
7179 in a comment' is no longer necessary and it doesn't use .err files which
7180 have now been deleted. Also the error code call doesn't have to appear all
7181 on one line (which resulted in some large lines...).
7184 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7187 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7188 0 (which usually indicates a closed connection), but continue reading.
7191 *) Fix some race conditions.
7194 *) Add support for CRL distribution points extension. Add Certificate
7195 Policies and CRL distribution points documentation.
7198 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7201 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7202 8 of keying material. Merlin has also confirmed interop with this fix
7203 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7204 [Merlin Hughes <merlin@baltimore.ie>]
7206 *) Fix lots of warnings.
7207 [Richard Levitte <levitte@stacken.kth.se>]
7209 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7210 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7211 [Richard Levitte <levitte@stacken.kth.se>]
7213 *) Fix problems with sizeof(long) == 8.
7214 [Andy Polyakov <appro@fy.chalmers.se>]
7216 *) Change functions to ANSI C.
7219 *) Fix typos in error codes.
7220 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7222 *) Remove defunct assembler files from Configure.
7225 *) SPARC v8 assembler BIGNUM implementation.
7226 [Andy Polyakov <appro@fy.chalmers.se>]
7228 *) Support for Certificate Policies extension: both print and set.
7229 Various additions to support the r2i method this uses.
7232 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7233 return a const string when you are expecting an allocated buffer.
7236 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7237 types DirectoryString and DisplayText.
7240 *) Add code to allow r2i extensions to access the configuration database,
7241 add an LHASH database driver and add several ctx helper functions.
7244 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7245 fail when they extended the size of a BIGNUM.
7248 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7249 support typesafe stack.
7252 *) Fix typo in SSL_[gs]et_options().
7253 [Nils Frostberg <nils@medcom.se>]
7255 *) Delete various functions and files that belonged to the (now obsolete)
7256 old X509V3 handling code.
7259 *) New Configure option "rsaref".
7262 *) Don't auto-generate pem.h.
7265 *) Introduce type-safe ASN.1 SETs.
7268 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7269 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7271 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7272 that links with OpenSSL (well at least cause lots of warnings), but fear
7273 not: the conversion is trivial, and it eliminates loads of evil casts. A
7274 few STACKed things have been converted already. Feel free to convert more.
7275 In the fullness of time, I'll do away with the STACK type altogether.
7278 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7279 specified in <certfile> by updating the entry in the index.txt file.
7280 This way one no longer has to edit the index.txt file manually for
7281 revoking a certificate. The -revoke option does the gory details now.
7282 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7284 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7285 `-text' option at all and this way the `-noout -text' combination was
7286 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7287 [Ralf S. Engelschall]
7289 *) Make sure a corresponding plain text error message exists for the
7290 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7291 verify callback function determined that a certificate was revoked.
7292 [Ralf S. Engelschall]
7294 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7295 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7296 all available cipers including rc5, which was forgotten until now.
7297 In order to let the testing shell script know which algorithms
7298 are available, a new (up to now undocumented) command
7299 "openssl list-cipher-commands" is used.
7302 *) Bugfix: s_client occasionally would sleep in select() when
7303 it should have checked SSL_pending() first.
7306 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7307 the raw DSA values prior to ASN.1 encoding.
7310 *) Tweaks to Configure
7311 [Niels Poppe <niels@netbox.org>]
7313 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7317 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7320 *) New config option to avoid instructions that are illegal on the 80386.
7321 The default code is faster, but requires at least a 486.
7324 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7325 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7326 same as SSL2_VERSION anyway.
7329 *) New "-showcerts" option for s_client.
7332 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7333 application. Various cleanups and fixes.
7336 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7337 modify error routines to work internally. Add error codes and PBE init
7338 to library startup routines.
7341 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7342 packing functions to asn1 and evp. Changed function names and error
7343 codes along the way.
7346 *) PKCS12 integration: and so it begins... First of several patches to
7347 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7348 objects to objects.h
7351 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7352 and display support for Thawte strong extranet extension.
7355 *) Add LinuxPPC support.
7356 [Jeff Dubrule <igor@pobox.org>]
7358 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7359 bn_div_words in alpha.s.
7360 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7362 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7363 OAEP isn't supported when OpenSSL is built with RSAref.
7364 [Ulf Moeller <ulf@fitug.de>]
7366 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7367 so they no longer are missing under -DNOPROTO.
7368 [Soren S. Jorvang <soren@t.dk>]
7371 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7373 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7374 doesn't work when the session is reused. Coming soon!
7377 *) Fix a security hole, that allows sessions to be reused in the wrong
7378 context thus bypassing client cert protection! All software that uses
7379 client certs and session caches in multiple contexts NEEDS PATCHING to
7380 allow session reuse! A fuller solution is in the works.
7381 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7383 *) Some more source tree cleanups (removed obsolete files
7384 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7385 permission on "config" script to be executable) and a fix for the INSTALL
7387 [Ulf Moeller <ulf@fitug.de>]
7389 *) Remove some legacy and erroneous uses of malloc, free instead of
7391 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7393 *) Make rsa_oaep_test return non-zero on error.
7394 [Ulf Moeller <ulf@fitug.de>]
7396 *) Add support for native Solaris shared libraries. Configure
7397 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7398 if someone would make that last step automatic.
7399 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7401 *) ctx_size was not built with the right compiler during "make links". Fixed.
7404 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7405 except NULL ciphers". This means the default cipher list will no longer
7406 enable NULL ciphers. They need to be specifically enabled e.g. with
7407 the string "DEFAULT:eNULL".
7410 *) Fix to RSA private encryption routines: if p < q then it would
7411 occasionally produce an invalid result. This will only happen with
7412 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7415 *) Be less restrictive and allow also `perl util/perlpath.pl
7416 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7417 because this way one can also use an interpreter named `perl5' (which is
7418 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7419 installed as `perl').
7420 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7422 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7423 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7425 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7426 advapi32.lib to Win32 build and change the pem test comparision
7427 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7428 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7429 and crypto/des/ede_cbcm_enc.c.
7432 *) DES quad checksum was broken on big-endian architectures. Fixed.
7435 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7436 Win32 test batch file so it (might) work again. The Win32 test batch file
7437 is horrible: I feel ill....
7440 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7441 in e_os.h. Audit of header files to check ANSI and non ANSI
7442 sections: 10 functions were absent from non ANSI section and not exported
7443 from Windows DLLs. Fixed up libeay.num for new functions.
7446 *) Make `openssl version' output lines consistent.
7447 [Ralf S. Engelschall]
7449 *) Fix Win32 symbol export lists for BIO functions: Added
7450 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7451 to ms/libeay{16,32}.def.
7452 [Ralf S. Engelschall]
7454 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7455 fine under Unix and passes some trivial tests I've now added. But the
7456 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7457 added to make sure no one expects that this stuff really works in the
7458 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7459 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7461 [Ralf S. Engelschall]
7463 *) Fix the generation of two part addresses in perl.
7464 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7466 *) Add config entry for Linux on MIPS.
7467 [John Tobey <jtobey@channel1.com>]
7469 *) Make links whenever Configure is run, unless we are on Windoze.
7472 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7473 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7477 *) Add a useful kludge to allow package maintainers to specify compiler and
7478 other platforms details on the command line without having to patch the
7479 Configure script everytime: One now can use ``perl Configure
7480 <id>:<details>'', i.e. platform ids are allowed to have details appended
7481 to them (seperated by colons). This is treated as there would be a static
7482 pre-configured entry in Configure's %table under key <id> with value
7483 <details> and ``perl Configure <id>'' is called. So, when you want to
7484 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7485 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7486 now, which overrides the FreeBSD-elf entry on-the-fly.
7487 [Ralf S. Engelschall]
7489 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7492 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7493 on the `perl Configure ...' command line. This way one can compile
7494 OpenSSL libraries with Position Independent Code (PIC) which is needed
7495 for linking it into DSOs.
7496 [Ralf S. Engelschall]
7498 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7502 *) Cleaned up the LICENSE document: The official contact for any license
7503 questions now is the OpenSSL core team under openssl-core@openssl.org.
7504 And add a paragraph about the dual-license situation to make sure people
7505 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7506 to the OpenSSL toolkit.
7507 [Ralf S. Engelschall]
7509 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7510 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7511 Additonally cleaned up the `make links' target: Remove unnecessary
7512 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7513 to speed processing and no longer clutter the display with confusing
7514 stuff. Instead only the actually done links are displayed.
7515 [Ralf S. Engelschall]
7517 *) Permit null encryption ciphersuites, used for authentication only. It used
7518 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7519 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7523 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7524 signed attributes when verifying signatures (this would break them),
7525 the detached data encoding was wrong and public keys obtained using
7526 X509_get_pubkey() weren't freed.
7529 *) Add text documentation for the BUFFER functions. Also added a work around
7530 to a Win95 console bug. This was triggered by the password read stuff: the
7531 last character typed gets carried over to the next fread(). If you were
7532 generating a new cert request using 'req' for example then the last
7533 character of the passphrase would be CR which would then enter the first
7537 *) Added the new `Includes OpenSSL Cryptography Software' button as
7538 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7539 button and can be used by applications based on OpenSSL to show the
7540 relationship to the OpenSSL project.
7541 [Ralf S. Engelschall]
7543 *) Remove confusing variables in function signatures in files
7544 ssl/ssl_lib.c and ssl/ssl.h.
7545 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7547 *) Don't install bss_file.c under PREFIX/include/
7548 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7550 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7551 functions that return function pointers and has support for NT specific
7552 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7553 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7554 unsigned to signed types: this was killing the Win32 compile.
7557 *) Add new certificate file to stack functions,
7558 SSL_add_dir_cert_subjects_to_stack() and
7559 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7560 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7561 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7562 This means that Apache-SSL and similar packages don't have to mess around
7563 to add as many CAs as they want to the preferred list.
7566 *) Experiment with doxygen documentation. Currently only partially applied to
7568 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7569 openssl.doxy as the configuration file.
7572 *) Get rid of remaining C++-style comments which strict C compilers hate.
7573 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7575 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7576 compiled in by default: it has problems with large keys.
7579 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7580 DH private keys and/or callback functions which directly correspond to
7581 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7582 is needed for applications which have to configure certificates on a
7583 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7585 For the RSA certificate situation is makes no difference, but
7586 for the DSA certificate situation this fixes the "no shared cipher"
7587 problem where the OpenSSL cipher selection procedure failed because the
7588 temporary keys were not overtaken from the context and the API provided
7589 no way to reconfigure them.
7590 The new functions now let applications reconfigure the stuff and they
7591 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7592 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7593 non-public-API function ssl_cert_instantiate() is used as a helper
7594 function and also to reduce code redundancy inside ssl_rsa.c.
7595 [Ralf S. Engelschall]
7597 *) Move s_server -dcert and -dkey options out of the undocumented feature
7598 area because they are useful for the DSA situation and should be
7599 recognized by the users.
7600 [Ralf S. Engelschall]
7602 *) Fix the cipher decision scheme for export ciphers: the export bits are
7603 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7604 SSL_EXP_MASK. So, the original variable has to be used instead of the
7605 already masked variable.
7606 [Richard Levitte <levitte@stacken.kth.se>]
7608 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7609 [Richard Levitte <levitte@stacken.kth.se>]
7611 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7612 from `int' to `unsigned int' because it's a length and initialized by
7613 EVP_DigestFinal() which expects an `unsigned int *'.
7614 [Richard Levitte <levitte@stacken.kth.se>]
7616 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7617 script. Instead use the usual Shell->Perl transition trick.
7618 [Ralf S. Engelschall]
7620 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7621 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7622 -noout -modulus' as it's already the case for `openssl rsa -noout
7623 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7624 currently the public key is printed (a decision which was already done by
7625 `openssl dsa -modulus' in the past) which serves a similar purpose.
7626 Additionally the NO_RSA no longer completely removes the whole -modulus
7627 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7629 [Ralf S. Engelschall]
7631 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7632 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7633 [Arne Ansper <arne@ats.cyber.ee>]
7635 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7636 to be added. Now both 'req' and 'ca' can use new objects defined in the
7640 *) Add cool BIO that does syslog (or event log on NT).
7641 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7643 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7644 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7645 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7646 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7649 *) Add preliminary config info for new extension code.
7652 *) Make RSA_NO_PADDING really use no padding.
7653 [Ulf Moeller <ulf@fitug.de>]
7655 *) Generate errors when private/public key check is done.
7658 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7659 for some CRL extensions and new objects added.
7662 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7663 key usage extension and fuller support for authority key id.
7666 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7667 padding method for RSA, which is recommended for new applications in PKCS
7668 #1 v2.0 (RFC 2437, October 1998).
7669 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7670 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7671 against Bleichbacher's attack on RSA.
7672 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7675 *) Updates to the new SSL compression code
7676 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7678 *) Fix so that the version number in the master secret, when passed
7679 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7680 (because the server will not accept higher), that the version number
7681 is 0x03,0x01, not 0x03,0x00
7682 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7684 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7685 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7686 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7689 *) Support for RAW extensions where an arbitrary extension can be
7690 created by including its DER encoding. See apps/openssl.cnf for
7694 *) Make sure latest Perl versions don't interpret some generated C array
7695 code as Perl array code in the crypto/err/err_genc.pl script.
7696 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7698 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7699 not many people have the assembler. Various Win32 compilation fixes and
7700 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7704 *) Modify configure script 'Configure' to automatically create crypto/date.h
7705 file under Win32 and also build pem.h from pem.org. New script
7706 util/mkfiles.pl to create the MINFO file on environments that can't do a
7707 'make files': perl util/mkfiles.pl >MINFO should work.
7710 *) Major rework of DES function declarations, in the pursuit of correctness
7711 and purity. As a result, many evil casts evaporated, and some weirdness,
7712 too. You may find this causes warnings in your code. Zapping your evil
7713 casts will probably fix them. Mostly.
7716 *) Fix for a typo in asn1.h. Bug fix to object creation script
7717 obj_dat.pl. It considered a zero in an object definition to mean
7718 "end of object": none of the objects in objects.h have any zeros
7719 so it wasn't spotted.
7720 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7722 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7723 Masking (CBCM). In the absence of test vectors, the best I have been able
7724 to do is check that the decrypt undoes the encrypt, so far. Send me test
7725 vectors if you have them.
7728 *) Correct calculation of key length for export ciphers (too much space was
7729 allocated for null ciphers). This has not been tested!
7732 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7733 message is now correct (it understands "crypto" and "ssl" on its
7734 command line). There is also now an "update" option. This will update
7735 the util/ssleay.num and util/libeay.num files with any new functions.
7737 perl util/mkdef.pl crypto ssl update
7738 it will update them.
7741 *) Overhauled the Perl interface (perl/*):
7742 - ported BN stuff to OpenSSL's different BN library
7743 - made the perl/ source tree CVS-aware
7744 - renamed the package from SSLeay to OpenSSL (the files still contain
7745 their history because I've copied them in the repository)
7746 - removed obsolete files (the test scripts will be replaced
7747 by better Test::Harness variants in the future)
7748 [Ralf S. Engelschall]
7750 *) First cut for a very conservative source tree cleanup:
7751 1. merge various obsolete readme texts into doc/ssleay.txt
7752 where we collect the old documents and readme texts.
7753 2. remove the first part of files where I'm already sure that we no
7754 longer need them because of three reasons: either they are just temporary
7755 files which were left by Eric or they are preserved original files where
7756 I've verified that the diff is also available in the CVS via "cvs diff
7757 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7758 the crypto/md/ stuff).
7759 [Ralf S. Engelschall]
7761 *) More extension code. Incomplete support for subject and issuer alt
7762 name, issuer and authority key id. Change the i2v function parameters
7763 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7764 what that's for :-) Fix to ASN1 macro which messed up
7765 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7768 *) Preliminary support for ENUMERATED type. This is largely copied from the
7772 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7773 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7775 *) Make sure `make rehash' target really finds the `openssl' program.
7776 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7778 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7779 like to hear about it if this slows down other processors.
7782 *) Add CygWin32 platform information to Configure script.
7783 [Alan Batie <batie@aahz.jf.intel.com>]
7785 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7786 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7788 *) New program nseq to manipulate netscape certificate sequences
7791 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7795 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7796 but the BN code had some problems that would cause failures when
7797 doing certificate verification and some other functions.
7798 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7800 *) Add ASN1 and PEM code to support netscape certificate sequences.
7803 *) Add ASN1 and PEM code to support netscape certificate sequences.
7806 *) Add several PKIX and private extended key usage OIDs.
7809 *) Modify the 'ca' program to handle the new extension code. Modify
7810 openssl.cnf for new extension format, add comments.
7813 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7814 and add a sample to openssl.cnf so req -x509 now adds appropriate
7818 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7819 error code, add initial support to X509_print() and x509 application.
7822 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7823 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7824 stuff is currently isolated and isn't even compiled yet.
7827 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7828 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7829 Removed the versions check from X509 routines when loading extensions:
7830 this allows certain broken certificates that don't set the version
7831 properly to be processed.
7834 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7835 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7836 can still be regenerated with "make depend".
7839 *) Spelling mistake in C version of CAST-128.
7840 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7842 *) Changes to the error generation code. The perl script err-code.pl
7843 now reads in the old error codes and retains the old numbers, only
7844 adding new ones if necessary. It also only changes the .err files if new
7845 codes are added. The makefiles have been modified to only insert errors
7846 when needed (to avoid needlessly modifying header files). This is done
7847 by only inserting errors if the .err file is newer than the auto generated
7848 C file. To rebuild all the error codes from scratch (the old behaviour)
7849 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7850 or delete all the .err files.
7853 *) CAST-128 was incorrectly implemented for short keys. The C version has
7854 been fixed, but is untested. The assembler versions are also fixed, but
7855 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7856 to regenerate it if needed.
7857 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7858 Hagino <itojun@kame.net>]
7860 *) File was opened incorrectly in randfile.c.
7861 [Ulf Möller <ulf@fitug.de>]
7863 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7864 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7865 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7866 al: it's just almost always a UTCTime. Note this patch adds new error
7867 codes so do a "make errors" if there are problems.
7870 *) Correct Linux 1 recognition in config.
7871 [Ulf Möller <ulf@fitug.de>]
7873 *) Remove pointless MD5 hash when using DSA keys in ca.
7874 [Anonymous <nobody@replay.com>]
7876 *) Generate an error if given an empty string as a cert directory. Also
7877 generate an error if handed NULL (previously returned 0 to indicate an
7878 error, but didn't set one).
7879 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7881 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7884 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7885 parameters. This was causing a warning which killed off the Win32 compile.
7888 *) Remove C++ style comments from crypto/bn/bn_local.h.
7889 [Neil Costigan <neil.costigan@celocom.com>]
7891 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7892 based on a text string, looking up short and long names and finally
7893 "dot" format. The "dot" format stuff didn't work. Added new function
7894 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7895 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7896 OID is not part of the table.
7899 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7900 X509_LOOKUP_by_alias().
7903 *) Sort openssl functions by name.
7906 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7907 encryption from sample DSA keys (in case anyone is interested the password
7911 *) Make _all_ *_free functions accept a NULL pointer.
7912 [Frans Heymans <fheymans@isaserver.be>]
7914 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7916 [Anonymous <nobody@replay.com>]
7918 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7919 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7921 *) Don't blow it for numeric -newkey arguments to apps/req.
7922 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7924 *) Temp key "for export" tests were wrong in s3_srvr.c.
7925 [Anonymous <nobody@replay.com>]
7927 *) Add prototype for temp key callback functions
7928 SSL_CTX_set_tmp_{rsa,dh}_callback().
7931 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7932 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7935 *) X509_name_add_entry() freed the wrong thing after an error.
7936 [Arne Ansper <arne@ats.cyber.ee>]
7938 *) rsa_eay.c would attempt to free a NULL context.
7939 [Arne Ansper <arne@ats.cyber.ee>]
7941 *) BIO_s_socket() had a broken should_retry() on Windoze.
7942 [Arne Ansper <arne@ats.cyber.ee>]
7944 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7945 [Arne Ansper <arne@ats.cyber.ee>]
7947 *) Make sure the already existing X509_STORE->depth variable is initialized
7948 in X509_STORE_new(), but document the fact that this variable is still
7949 unused in the certificate verification process.
7950 [Ralf S. Engelschall]
7952 *) Fix the various library and apps files to free up pkeys obtained from
7953 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7956 *) Fix reference counting in X509_PUBKEY_get(). This makes
7957 demos/maurice/example2.c work, amongst others, probably.
7958 [Steve Henson and Ben Laurie]
7960 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7961 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7962 are no longer created. This way we have a single and consistent command
7963 line interface `openssl <command>', similar to `cvs <command>'.
7964 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7966 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7967 BIT STRING wrapper always have zero unused bits.
7970 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7973 *) Make the top-level INSTALL documentation easier to understand.
7976 *) Makefiles updated to exit if an error occurs in a sub-directory
7977 make (including if user presses ^C) [Paul Sutton]
7979 *) Make Montgomery context stuff explicit in RSA data structure.
7982 *) Fix build order of pem and err to allow for generated pem.h.
7985 *) Fix renumbering bug in X509_NAME_delete_entry().
7988 *) Enhanced the err-ins.pl script so it makes the error library number
7989 global and can add a library name. This is needed for external ASN1 and
7990 other error libraries.
7993 *) Fixed sk_insert which never worked properly.
7996 *) Fix ASN1 macros so they can handle indefinite length construted
7997 EXPLICIT tags. Some non standard certificates use these: they can now
8001 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8002 into a single doc/ssleay.txt bundle. This way the information is still
8003 preserved but no longer messes up this directory. Now it's new room for
8004 the new set of documenation files.
8005 [Ralf S. Engelschall]
8007 *) SETs were incorrectly DER encoded. This was a major pain, because they
8008 shared code with SEQUENCEs, which aren't coded the same. This means that
8009 almost everything to do with SETs or SEQUENCEs has either changed name or
8010 number of arguments.
8011 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8013 *) Fix test data to work with the above.
8016 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8017 was already fixed by Eric for 0.9.1 it seems.
8018 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8020 *) Autodetect FreeBSD3.
8023 *) Fix various bugs in Configure. This affects the following platforms:
8027 unixware-2.0-pentium
8031 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8032 before they are needed.
8035 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8039 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8041 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8042 changed SSLeay to OpenSSL in version strings.
8043 [Ralf S. Engelschall]
8045 *) Some fixups to the top-level documents.
8048 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8049 because the symlink to include/ was missing.
8050 [Ralf S. Engelschall]
8052 *) Incorporated the popular no-RSA/DSA-only patches
8053 which allow to compile a RSA-free SSLeay.
8054 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8056 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8057 when "ssleay" is still not found.
8058 [Ralf S. Engelschall]
8060 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8061 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8063 *) Updated the README file.
8064 [Ralf S. Engelschall]
8066 *) Added various .cvsignore files in the CVS repository subdirs
8067 to make a "cvs update" really silent.
8068 [Ralf S. Engelschall]
8070 *) Recompiled the error-definition header files and added
8071 missing symbols to the Win32 linker tables.
8072 [Ralf S. Engelschall]
8074 *) Cleaned up the top-level documents;
8075 o new files: CHANGES and LICENSE
8076 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8077 o merged COPYRIGHT into LICENSE
8078 o removed obsolete TODO file
8079 o renamed MICROSOFT to INSTALL.W32
8080 [Ralf S. Engelschall]
8082 *) Removed dummy files from the 0.9.1b source tree:
8083 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8084 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8085 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8086 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8087 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8088 [Ralf S. Engelschall]
8090 *) Added various platform portability fixes.
8093 *) The Genesis of the OpenSSL rpject:
8094 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8095 Young and Tim J. Hudson created while they were working for C2Net until
8097 [The OpenSSL Project]
8100 Changes between 0.9.0b and 0.9.1b [not released]
8102 *) Updated a few CA certificates under certs/
8105 *) Changed some BIGNUM api stuff.
8108 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8109 DGUX x86, Linux Alpha, etc.
8112 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8113 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8117 *) Add -strparse option to asn1pars program which parses nested
8119 [Dr Stephen Henson <shenson@bigfoot.com>]
8121 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8124 *) DSA fix for "ca" program.
8127 *) Added "-genkey" option to "dsaparam" program.
8130 *) Added RIPE MD160 (rmd160) message digest.
8133 *) Added -a (all) option to "ssleay version" command.
8136 *) Added PLATFORM define which is the id given to Configure.
8139 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8142 *) Extended the ASN.1 parser routines.
8145 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8148 *) Added a BN_CTX to the BN library.
8151 *) Fixed the weak key values in DES library
8154 *) Changed API in EVP library for cipher aliases.
8157 *) Added support for RC2/64bit cipher.
8160 *) Converted the lhash library to the crypto/mem.c functions.
8163 *) Added more recognized ASN.1 object ids.
8166 *) Added more RSA padding checks for SSL/TLS.
8169 *) Added BIO proxy/filter functionality.
8172 *) Added extra_certs to SSL_CTX which can be used
8173 send extra CA certificates to the client in the CA cert chain sending
8174 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8177 *) Now Fortezza is denied in the authentication phase because
8178 this is key exchange mechanism is not supported by SSLeay at all.
8181 *) Additional PKCS1 checks.
8184 *) Support the string "TLSv1" for all TLS v1 ciphers.
8187 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8188 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8191 *) Fixed a few memory leaks.
8194 *) Fixed various code and comment typos.
8197 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8198 bytes sent in the client random.
8199 [Edward Bishop <ebishop@spyglass.com>]