5 Changes between 0.9.8g and 0.9.8h [28 May 2008]
7 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
8 handshake which could lead to a cilent crash as found using the
9 Codenomicon TLS test suite (CVE-2008-1672)
10 [Steve Henson, Mark Cox]
12 *) Fix double free in TLS server name extensions which could lead to
13 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
16 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
18 Clear the error queue to ensure that error entries left from
19 older function calls do not interfere with the correct operation.
20 [Lutz Jaenicke, Erik de Castro Lopo]
22 *) Remove root CA certificates of commercial CAs:
24 The OpenSSL project does not recommend any specific CA and does not
25 have any policy with respect to including or excluding any CA.
26 Therefore it does not make any sense to ship an arbitrary selection
27 of root CA certificates with the OpenSSL software.
30 *) RSA OAEP patches to fix two separate invalid memory reads.
31 The first one involves inputs when 'lzero' is greater than
32 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
33 before the beginning of from). The second one involves inputs where
34 the 'db' section contains nothing but zeroes (there is a one-byte
35 invalid read after the end of 'db').
36 [Ivan Nestlerode <inestlerode@us.ibm.com>]
38 *) Partial backport from 0.9.9-dev:
40 Introduce bn_mul_mont (dedicated Montgomery multiplication
41 procedure) as a candidate for BIGNUM assembler implementation.
42 While 0.9.9-dev uses assembler for various architectures, only
43 x86_64 is available by default here in the 0.9.8 branch, and
44 32-bit x86 is available through a compile-time setting.
46 To try the 32-bit x86 assembler implementation, use Configure
47 option "enable-montasm" (which exists only for this backport).
49 As "enable-montasm" for 32-bit x86 disclaims code stability
50 anyway, in this constellation we activate additional code
51 backported from 0.9.9-dev for further performance improvements,
52 namely BN_from_montgomery_word. (To enable this otherwise,
53 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
55 [Andy Polyakov (backport partially by Bodo Moeller)]
57 *) Add TLS session ticket callback. This allows an application to set
58 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
59 values. This is useful for key rollover for example where several key
60 sets may exist with different names.
63 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
64 This was broken until now in 0.9.8 releases, such that the only way
65 a registered ENGINE could be used (assuming it initialises
66 successfully on the host) was to explicitly set it as the default
67 for the relevant algorithms. This is in contradiction with 0.9.7
68 behaviour and the documentation. With this fix, when an ENGINE is
69 registered into a given algorithm's table of implementations, the
70 'uptodate' flag is reset so that auto-discovery will be used next
71 time a new context for that algorithm attempts to select an
73 [Ian Lister (tweaked by Geoff Thorpe)]
75 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
76 implemention in the following ways:
78 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
81 Lack of BER streaming support means one pass streaming processing is
82 only supported if data is detached: setting the streaming flag is
83 ignored for embedded content.
85 CMS support is disabled by default and must be explicitly enabled
86 with the enable-cms configuration option.
89 *) Update the GMP engine glue to do direct copies between BIGNUM and
90 mpz_t when openssl and GMP use the same limb size. Otherwise the
91 existing "conversion via a text string export" trick is still used.
92 [Paul Sheer <paulsheer@gmail.com>]
94 *) Zlib compression BIO. This is a filter BIO which compressed and
95 uncompresses any data passed through it.
98 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
99 RFC3394 compatible AES key wrapping.
102 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
103 sets string data without copying. X509_ALGOR_set0() and
104 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
105 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
106 from an X509_ATTRIBUTE structure optionally checking it occurs only
107 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
111 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
112 to get the expected BN_FLG_CONSTTIME behavior.
113 [Bodo Moeller (Google)]
117 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
118 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
119 - added some more tests to do_tests.pl
120 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
121 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
122 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
123 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
124 - various changes to netware.pl to enable gcc-cross builds on Win32
126 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
127 - various changes to fix missing prototype warnings
128 - fixed x86nasm.pl to create correct asm files for NASM COFF output
129 - added AES, WHIRLPOOL and CPUID assembler code to build files
130 - added missing AES assembler make rules to mk1mf.pl
131 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
132 [Guenter Knauf <eflash@gmx.net>]
134 *) Implement certificate status request TLS extension defined in RFC3546.
135 A client can set the appropriate parameters and receive the encoded
136 OCSP response via a callback. A server can query the supplied parameters
137 and set the encoded OCSP response in the callback. Add simplified examples
138 to s_client and s_server.
141 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
144 + Binary incompatibility of ssl_ctx_st structure
145 + DTLS interoperation with non-compliant servers
146 + Don't call get_session_cb() without proposed session
147 + Fix ia64 assembler code
148 [Andy Polyakov, Steve Henson]
150 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
152 *) DTLS Handshake overhaul. There were longstanding issues with
153 OpenSSL DTLS implementation, which were making it impossible for
154 RFC 4347 compliant client to communicate with OpenSSL server.
155 Unfortunately just fixing these incompatibilities would "cut off"
156 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
157 server keeps tolerating non RFC compliant syntax. The opposite is
158 not true, 0.9.8f client can not communicate with earlier server.
159 This update even addresses CVE-2007-4995.
162 *) Changes to avoid need for function casts in OpenSSL: some compilers
163 (gcc 4.2 and later) reject their use.
164 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
167 *) Add RFC4507 support to OpenSSL. This includes the corrections in
168 RFC4507bis. The encrypted ticket format is an encrypted encoded
169 SSL_SESSION structure, that way new session features are automatically
172 If a client application caches session in an SSL_SESSION structure
173 support is transparent because tickets are now stored in the encoded
176 The SSL_CTX structure automatically generates keys for ticket
177 protection in servers so again support should be possible
178 with no application modification.
180 If a client or server wishes to disable RFC4507 support then the option
181 SSL_OP_NO_TICKET can be set.
183 Add a TLS extension debugging callback to allow the contents of any client
184 or server extensions to be examined.
186 This work was sponsored by Google.
189 *) Add initial support for TLS extensions, specifically for the server_name
190 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
191 have new members for a host name. The SSL data structure has an
192 additional member SSL_CTX *initial_ctx so that new sessions can be
193 stored in that context to allow for session resumption, even after the
194 SSL has been switched to a new SSL_CTX in reaction to a client's
195 server_name extension.
197 New functions (subject to change):
200 SSL_get_servername_type()
203 New CTRL codes and macros (subject to change):
205 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
206 - SSL_CTX_set_tlsext_servername_callback()
207 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
208 - SSL_CTX_set_tlsext_servername_arg()
209 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
211 openssl s_client has a new '-servername ...' option.
213 openssl s_server has new options '-servername_host ...', '-cert2 ...',
214 '-key2 ...', '-servername_fatal' (subject to change). This allows
215 testing the HostName extension for a specific single host name ('-cert'
216 and '-key' remain fallbacks for handshakes without HostName
217 negotiation). If the unrecogninzed_name alert has to be sent, this by
218 default is a warning; it becomes fatal with the '-servername_fatal'
221 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
223 *) Add AES and SSE2 assembly language support to VC++ build.
226 *) Mitigate attack on final subtraction in Montgomery reduction.
229 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
230 (which previously caused an internal error).
233 *) Squeeze another 10% out of IGE mode when in != out.
236 *) AES IGE mode speedup.
237 [Dean Gaudet (Google)]
239 *) Add the Korean symmetric 128-bit cipher SEED (see
240 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
241 add SEED ciphersuites from RFC 4162:
243 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
244 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
245 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
246 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
248 To minimize changes between patchlevels in the OpenSSL 0.9.8
249 series, SEED remains excluded from compilation unless OpenSSL
250 is configured with 'enable-seed'.
253 *) Mitigate branch prediction attacks, which can be practical if a
254 single processor is shared, allowing a spy process to extract
255 information. For detailed background information, see
256 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
257 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
258 and Necessary Software Countermeasures"). The core of the change
259 are new versions BN_div_no_branch() and
260 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
261 respectively, which are slower, but avoid the security-relevant
262 conditional branches. These are automatically called by BN_div()
263 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
264 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
265 remove a conditional branch.
267 BN_FLG_CONSTTIME is the new name for the previous
268 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
269 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
270 in the exponent causes BN_mod_exp_mont() to use the alternative
271 implementation in BN_mod_exp_mont_consttime().) The old name
272 remains as a deprecated alias.
274 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
275 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
276 constant-time implementations for more than just exponentiation.
277 Here too the old name is kept as a deprecated alias.
279 BN_BLINDING_new() will now use BN_dup() for the modulus so that
280 the BN_BLINDING structure gets an independent copy of the
281 modulus. This means that the previous "BIGNUM *m" argument to
282 BN_BLINDING_new() and to BN_BLINDING_create_param() now
283 essentially becomes "const BIGNUM *m", although we can't actually
284 change this in the header file before 0.9.9. It allows
285 RSA_setup_blinding() to use BN_with_flags() on the modulus to
286 enable BN_FLG_CONSTTIME.
288 [Matthew D Wood (Intel Corp)]
290 *) In the SSL/TLS server implementation, be strict about session ID
291 context matching (which matters if an application uses a single
292 external cache for different purposes). Previously,
293 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
294 set. This did ensure strict client verification, but meant that,
295 with applications using a single external cache for quite
296 different requirements, clients could circumvent ciphersuite
297 restrictions for a given session ID context by starting a session
298 in a different context.
301 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
302 a ciphersuite string such as "DEFAULT:RSA" cannot enable
303 authentication-only ciphersuites.
306 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
307 not complete and could lead to a possible single byte overflow
308 (CVE-2007-5135) [Ben Laurie]
310 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
312 *) Since AES128 and AES256 (and similarly Camellia128 and
313 Camellia256) share a single mask bit in the logic of
314 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
315 kludge to work properly if AES128 is available and AES256 isn't
316 (or if Camellia128 is available and Camellia256 isn't).
319 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
320 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
321 When a point or a seed is encoded in a BIT STRING, we need to
322 prevent the removal of trailing zero bits to get the proper DER
323 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
324 of a NamedBitList, for which trailing 0 bits need to be removed.)
327 *) Have SSL/TLS server implementation tolerate "mismatched" record
328 protocol version while receiving ClientHello even if the
329 ClientHello is fragmented. (The server can't insist on the
330 particular protocol version it has chosen before the ServerHello
331 message has informed the client about his choice.)
334 *) Add RFC 3779 support.
335 [Rob Austein for ARIN, Ben Laurie]
337 *) Load error codes if they are not already present instead of using a
338 static variable. This allows them to be cleanly unloaded and reloaded.
339 Improve header file function name parsing.
342 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
343 or CAPABILITY handshake as required by RFCs.
346 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
348 *) Introduce limits to prevent malicious keys being able to
349 cause a denial of service. (CVE-2006-2940)
350 [Steve Henson, Bodo Moeller]
352 *) Fix ASN.1 parsing of certain invalid structures that can result
353 in a denial of service. (CVE-2006-2937) [Steve Henson]
355 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
356 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
358 *) Fix SSL client code which could crash if connecting to a
359 malicious SSLv2 server. (CVE-2006-4343)
360 [Tavis Ormandy and Will Drewry, Google Security Team]
362 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
363 match only those. Before that, "AES256-SHA" would be interpreted
364 as a pattern and match "AES128-SHA" too (since AES128-SHA got
365 the same strength classification in 0.9.7h) as we currently only
366 have a single AES bit in the ciphersuite description bitmap.
367 That change, however, also applied to ciphersuite strings such as
368 "RC4-MD5" that intentionally matched multiple ciphersuites --
369 namely, SSL 2.0 ciphersuites in addition to the more common ones
370 from SSL 3.0/TLS 1.0.
372 So we change the selection algorithm again: Naming an explicit
373 ciphersuite selects this one ciphersuite, and any other similar
374 ciphersuite (same bitmap) from *other* protocol versions.
375 Thus, "RC4-MD5" again will properly select both the SSL 2.0
376 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
378 Since SSL 2.0 does not have any ciphersuites for which the
379 128/256 bit distinction would be relevant, this works for now.
380 The proper fix will be to use different bits for AES128 and
381 AES256, which would have avoided the problems from the beginning;
382 however, bits are scarce, so we can only do this in a new release
383 (not just a patchlevel) when we can change the SSL_CIPHER
384 definition to split the single 'unsigned long mask' bitmap into
385 multiple values to extend the available space.
389 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
391 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
392 (CVE-2006-4339) [Ben Laurie and Google Security Team]
394 *) Add AES IGE and biIGE modes.
397 *) Change the Unix randomness entropy gathering to use poll() when
398 possible instead of select(), since the latter has some
399 undesirable limitations.
400 [Darryl Miles via Richard Levitte and Bodo Moeller]
402 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
403 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
404 cannot be implicitly activated as part of, e.g., the "AES" alias.
405 However, please upgrade to OpenSSL 0.9.9[-dev] for
406 non-experimental use of the ECC ciphersuites to get TLS extension
407 support, which is required for curve and point format negotiation
408 to avoid potential handshake problems.
411 *) Disable rogue ciphersuites:
413 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
414 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
415 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
417 The latter two were purportedly from
418 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
421 Also deactivate the remaining ciphersuites from
422 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
423 unofficial, and the ID has long expired.
426 *) Fix RSA blinding Heisenbug (problems sometimes occured on
427 dual-core machines) and other potential thread-safety issues.
430 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
431 versions), which is now available for royalty-free use
432 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
433 Also, add Camellia TLS ciphersuites from RFC 4132.
435 To minimize changes between patchlevels in the OpenSSL 0.9.8
436 series, Camellia remains excluded from compilation unless OpenSSL
437 is configured with 'enable-camellia'.
440 *) Disable the padding bug check when compression is in use. The padding
441 bug check assumes the first packet is of even length, this is not
442 necessarily true if compresssion is enabled and can result in false
443 positives causing handshake failure. The actual bug test is ancient
444 code so it is hoped that implementations will either have fixed it by
445 now or any which still have the bug do not support compression.
448 Changes between 0.9.8a and 0.9.8b [04 May 2006]
450 *) When applying a cipher rule check to see if string match is an explicit
451 cipher suite and only match that one cipher suite if it is.
454 *) Link in manifests for VC++ if needed.
455 [Austin Ziegler <halostatue@gmail.com>]
457 *) Update support for ECC-based TLS ciphersuites according to
458 draft-ietf-tls-ecc-12.txt with proposed changes (but without
459 TLS extensions, which are supported starting with the 0.9.9
460 branch, not in the OpenSSL 0.9.8 branch).
463 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
464 opaque EVP_CIPHER_CTX handling.
467 *) Fixes and enhancements to zlib compression code. We now only use
468 "zlib1.dll" and use the default __cdecl calling convention on Win32
469 to conform with the standards mentioned here:
470 http://www.zlib.net/DLL_FAQ.txt
471 Static zlib linking now works on Windows and the new --with-zlib-include
472 --with-zlib-lib options to Configure can be used to supply the location
473 of the headers and library. Gracefully handle case where zlib library
477 *) Several fixes and enhancements to the OID generation code. The old code
478 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
479 handle numbers larger than ULONG_MAX, truncated printing and had a
480 non standard OBJ_obj2txt() behaviour.
483 *) Add support for building of engines under engine/ as shared libraries
484 under VC++ build system.
487 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
488 Hopefully, we will not see any false combination of paths any more.
491 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
493 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
494 (part of SSL_OP_ALL). This option used to disable the
495 countermeasure against man-in-the-middle protocol-version
496 rollback in the SSL 2.0 server implementation, which is a bad
497 idea. (CVE-2005-2969)
499 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
500 for Information Security, National Institute of Advanced Industrial
501 Science and Technology [AIST], Japan)]
503 *) Add two function to clear and return the verify parameter flags.
506 *) Keep cipherlists sorted in the source instead of sorting them at
507 runtime, thus removing the need for a lock.
510 *) Avoid some small subgroup attacks in Diffie-Hellman.
511 [Nick Mathewson and Ben Laurie]
513 *) Add functions for well-known primes.
516 *) Extended Windows CE support.
517 [Satoshi Nakamura and Andy Polyakov]
519 *) Initialize SSL_METHOD structures at compile time instead of during
520 runtime, thus removing the need for a lock.
523 *) Make PKCS7_decrypt() work even if no certificate is supplied by
524 attempting to decrypt each encrypted key in turn. Add support to
528 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
530 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
533 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
536 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
537 key into the same file any more.
540 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
543 *) Add -utf8 command line and config file option to 'ca'.
544 [Stefan <stf@udoma.org]
546 *) Removed the macro des_crypt(), as it seems to conflict with some
547 libraries. Use DES_crypt().
550 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
551 involves renaming the source and generated shared-libs for
552 both. The engines will accept the corrected or legacy ids
553 ('ncipher' and '4758_cca' respectively) when binding. NB,
554 this only applies when building 'shared'.
555 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
557 *) Add attribute functions to EVP_PKEY structure. Modify
558 PKCS12_create() to recognize a CSP name attribute and
559 use it. Make -CSP option work again in pkcs12 utility.
562 *) Add new functionality to the bn blinding code:
563 - automatic re-creation of the BN_BLINDING parameters after
564 a fixed number of uses (currently 32)
565 - add new function for parameter creation
566 - introduce flags to control the update behaviour of the
567 BN_BLINDING parameters
568 - hide BN_BLINDING structure
569 Add a second BN_BLINDING slot to the RSA structure to improve
570 performance when a single RSA object is shared among several
574 *) Add support for DTLS.
575 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
577 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
578 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
581 *) Remove buggy and incompletet DH cert support from
582 ssl/ssl_rsa.c and ssl/s3_both.c
585 *) Use SHA-1 instead of MD5 as the default digest algorithm for
586 the apps/openssl applications.
589 *) Compile clean with "-Wall -Wmissing-prototypes
590 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
591 DEBUG_SAFESTACK must also be set.
594 *) Change ./Configure so that certain algorithms can be disabled by default.
595 The new counterpiece to "no-xxx" is "enable-xxx".
597 The patented RC5 and MDC2 algorithms will now be disabled unless
598 "enable-rc5" and "enable-mdc2", respectively, are specified.
600 (IDEA remains enabled despite being patented. This is because IDEA
601 is frequently required for interoperability, and there is no license
602 fee for non-commercial use. As before, "no-idea" can be used to
603 avoid this algorithm.)
607 *) Add processing of proxy certificates (see RFC 3820). This work was
608 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
609 EGEE (Enabling Grids for E-science in Europe).
612 *) RC4 performance overhaul on modern architectures/implementations, such
613 as Intel P4, IA-64 and AMD64.
616 *) New utility extract-section.pl. This can be used specify an alternative
617 section number in a pod file instead of having to treat each file as
618 a separate case in Makefile. This can be done by adding two lines to the
621 =for comment openssl_section:XXX
623 The blank line is mandatory.
627 *) New arguments -certform, -keyform and -pass for s_client and s_server
628 to allow alternative format key and certificate files and passphrase
632 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
633 update associated structures and add various utility functions.
635 Add new policy related verify parameters, include policy checking in
636 standard verify code. Enhance 'smime' application with extra parameters
637 to support policy checking and print out.
640 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
641 Nehemiah processors. These extensions support AES encryption in hardware
642 as well as RNG (though RNG support is currently disabled).
643 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
645 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
648 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
649 [Andy Polyakov and a number of other people]
651 *) Improved PowerPC platform support. Most notably BIGNUM assembler
652 implementation contributed by IBM.
653 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
655 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
656 exponent rather than 'unsigned long'. There is a corresponding change to
657 the new 'rsa_keygen' element of the RSA_METHOD structure.
658 [Jelte Jansen, Geoff Thorpe]
660 *) Functionality for creating the initial serial number file is now
661 moved from CA.pl to the 'ca' utility with a new option -create_serial.
663 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
664 number file to 1, which is bound to cause problems. To avoid
665 the problems while respecting compatibility between different 0.9.7
666 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
667 CA.pl for serial number initialization. With the new release 0.9.8,
668 we can fix the problem directly in the 'ca' utility.)
671 *) Reduced header interdepencies by declaring more opaque objects in
672 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
673 give fewer recursive includes, which could break lazy source code - so
674 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
675 developers should define this symbol when building and using openssl to
676 ensure they track the recommended behaviour, interfaces, [etc], but
677 backwards-compatible behaviour prevails when this isn't defined.
680 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
683 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
684 This will generate a random key of the appropriate length based on the
685 cipher context. The EVP_CIPHER can provide its own random key generation
686 routine to support keys of a specific form. This is used in the des and
687 3des routines to generate a key of the correct parity. Update S/MIME
688 code to use new functions and hence generate correct parity DES keys.
689 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
690 valid (weak or incorrect parity).
693 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
694 as looking them up. This is useful when the verified structure may contain
695 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
696 present unless the new PKCS7_NO_CRL flag is asserted.
699 *) Extend ASN1 oid configuration module. It now additionally accepts the
702 shortName = some long name, 1.2.3.4
705 *) Reimplemented the BN_CTX implementation. There is now no more static
706 limitation on the number of variables it can handle nor the depth of the
707 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
708 information can now expand as required, and rather than having a single
709 static array of bignums, BN_CTX now uses a linked-list of such arrays
710 allowing it to expand on demand whilst maintaining the usefulness of
714 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
715 to allow all RSA operations to function using a single BN_CTX.
718 *) Preliminary support for certificate policy evaluation and checking. This
719 is initially intended to pass the tests outlined in "Conformance Testing
720 of Relying Party Client Certificate Path Processing Logic" v1.07.
723 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
724 remained unused and not that useful. A variety of other little bignum
725 tweaks and fixes have also been made continuing on from the audit (see
729 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
730 associated ASN1, EVP and SSL functions and old ASN1 macros.
733 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
734 and this should never fail. So the return value from the use of
735 BN_set_word() (which can fail due to needless expansion) is now deprecated;
736 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
739 *) BN_CTX_get() should return zero-valued bignums, providing the same
740 initialised value as BN_new().
741 [Geoff Thorpe, suggested by Ulf Möller]
743 *) Support for inhibitAnyPolicy certificate extension.
746 *) An audit of the BIGNUM code is underway, for which debugging code is
747 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
748 is considered valid when processing BIGNUMs, and causes execution to
749 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
750 further steps are taken to deliberately pollute unused data in BIGNUM
751 structures to try and expose faulty code further on. For now, openssl will
752 (in its default mode of operation) continue to tolerate the inconsistent
753 forms that it has tolerated in the past, but authors and packagers should
754 consider trying openssl and their own applications when compiled with
755 these debugging symbols defined. It will help highlight potential bugs in
756 their own code, and will improve the test coverage for OpenSSL itself. At
757 some point, these tighter rules will become openssl's default to improve
758 maintainability, though the assert()s and other overheads will remain only
759 in debugging configurations. See bn.h for more details.
760 [Geoff Thorpe, Nils Larsch, Ulf Möller]
762 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
763 that can only be obtained through BN_CTX_new() (which implicitly
764 initialises it). The presence of this function only made it possible
765 to overwrite an existing structure (and cause memory leaks).
768 *) Because of the callback-based approach for implementing LHASH as a
769 template type, lh_insert() adds opaque objects to hash-tables and
770 lh_doall() or lh_doall_arg() are typically used with a destructor callback
771 to clean up those corresponding objects before destroying the hash table
772 (and losing the object pointers). So some over-zealous constifications in
773 LHASH have been relaxed so that lh_insert() does not take (nor store) the
774 objects as "const" and the lh_doall[_arg] callback wrappers are not
775 prototyped to have "const" restrictions on the object pointers they are
776 given (and so aren't required to cast them away any more).
779 *) The tmdiff.h API was so ugly and minimal that our own timing utility
780 (speed) prefers to use its own implementation. The two implementations
781 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
782 its object type properly exposed (MS_TM) instead of casting to/from "char
783 *". This may still change yet if someone realises MS_TM and "ms_time_***"
784 aren't necessarily the greatest nomenclatures - but this is what was used
785 internally to the implementation so I've used that for now.
788 *) Ensure that deprecated functions do not get compiled when
789 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
790 the self-tests were still using deprecated key-generation functions so
791 these have been updated also.
794 *) Reorganise PKCS#7 code to separate the digest location functionality
795 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
796 New function PKCS7_set_digest() to set the digest type for PKCS#7
797 digestedData type. Add additional code to correctly generate the
798 digestedData type and add support for this type in PKCS7 initialization
802 *) New function PKCS7_set0_type_other() this initializes a PKCS7
803 structure of type "other".
806 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
807 sure the loop does correctly stop and breaking ("division by zero")
808 modulus operations are not performed. The (pre-generated) prime
809 table crypto/bn/bn_prime.h was already correct, but it could not be
810 re-generated on some platforms because of the "division by zero"
811 situation in the script.
812 [Ralf S. Engelschall]
814 *) Update support for ECC-based TLS ciphersuites according to
815 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
816 SHA-1 now is only used for "small" curves (where the
817 representation of a field element takes up to 24 bytes); for
818 larger curves, the field element resulting from ECDH is directly
819 used as premaster secret.
820 [Douglas Stebila (Sun Microsystems Laboratories)]
822 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
823 curve secp160r1 to the tests.
824 [Douglas Stebila (Sun Microsystems Laboratories)]
826 *) Add the possibility to load symbols globally with DSO.
827 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
829 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
830 control of the error stack.
833 *) Add support for STORE in ENGINE.
836 *) Add the STORE type. The intention is to provide a common interface
837 to certificate and key stores, be they simple file-based stores, or
838 HSM-type store, or LDAP stores, or...
839 NOTE: The code is currently UNTESTED and isn't really used anywhere.
842 *) Add a generic structure called OPENSSL_ITEM. This can be used to
843 pass a list of arguments to any function as well as provide a way
844 for a function to pass data back to the caller.
847 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
848 works like BUF_strdup() but can be used to duplicate a portion of
849 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
853 *) Add the function sk_find_ex() which works like sk_find(), but will
854 return an index to an element even if an exact match couldn't be
855 found. The index is guaranteed to point at the element where the
856 searched-for key would be inserted to preserve sorting order.
859 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
860 takes an extra flags argument for optional functionality. Currently,
861 the following flags are defined:
863 OBJ_BSEARCH_VALUE_ON_NOMATCH
864 This one gets OBJ_bsearch_ex() to return a pointer to the first
865 element where the comparing function returns a negative or zero
868 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
869 This one gets OBJ_bsearch_ex() to return a pointer to the first
870 element where the comparing function returns zero. This is useful
871 if there are more than one element where the comparing function
875 *) Make it possible to create self-signed certificates with 'openssl ca'
876 in such a way that the self-signed certificate becomes part of the
877 CA database and uses the same mechanisms for serial number generation
878 as all other certificate signing. The new flag '-selfsign' enables
879 this functionality. Adapt CA.sh and CA.pl.in.
882 *) Add functionality to check the public key of a certificate request
883 against a given private. This is useful to check that a certificate
884 request can be signed by that key (self-signing).
887 *) Make it possible to have multiple active certificates with the same
888 subject in the CA index file. This is done only if the keyword
889 'unique_subject' is set to 'no' in the main CA section (default
890 if 'CA_default') of the configuration file. The value is saved
891 with the database itself in a separate index attribute file,
892 named like the index file with '.attr' appended to the name.
895 *) Generate muti valued AVAs using '+' notation in config files for
899 *) Support for nameConstraints certificate extension.
902 *) Support for policyConstraints certificate extension.
905 *) Support for policyMappings certificate extension.
908 *) Make sure the default DSA_METHOD implementation only uses its
909 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
910 and change its own handlers to be NULL so as to remove unnecessary
911 indirection. This lets alternative implementations fallback to the
912 default implementation more easily.
915 *) Support for directoryName in GeneralName related extensions
919 *) Make it possible to link applications using Makefile.shared.
920 Make that possible even when linking against static libraries!
923 *) Support for single pass processing for S/MIME signing. This now
924 means that S/MIME signing can be done from a pipe, in addition
925 cleartext signing (multipart/signed type) is effectively streaming
926 and the signed data does not need to be all held in memory.
928 This is done with a new flag PKCS7_STREAM. When this flag is set
929 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
930 is done after the data is output (and digests calculated) in
934 *) Add full support for -rpath/-R, both in shared libraries and
935 applications, at least on the platforms where it's known how
939 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
940 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
941 will now compute a table of multiples of the generator that
942 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
943 faster (notably in the case of a single point multiplication,
945 [Nils Larsch, Bodo Moeller]
947 *) IPv6 support for certificate extensions. The various extensions
948 which use the IP:a.b.c.d can now take IPv6 addresses using the
949 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
953 *) Added an ENGINE that implements RSA by performing private key
954 exponentiations with the GMP library. The conversions to and from
955 GMP's mpz_t format aren't optimised nor are any montgomery forms
956 cached, and on x86 it appears OpenSSL's own performance has caught up.
957 However there are likely to be other architectures where GMP could
958 provide a boost. This ENGINE is not built in by default, but it can be
959 specified at Configure time and should be accompanied by the necessary
960 linker additions, eg;
961 ./config -DOPENSSL_USE_GMP -lgmp
964 *) "openssl engine" will not display ENGINE/DSO load failure errors when
965 testing availability of engines with "-t" - the old behaviour is
966 produced by increasing the feature's verbosity with "-tt".
969 *) ECDSA routines: under certain error conditions uninitialized BN objects
970 could be freed. Solution: make sure initialization is performed early
971 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
975 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
976 and DH_METHOD (eg. by ENGINE implementations) to override the normal
977 software implementations. For DSA and DH, parameter generation can
978 also be overriden by providing the appropriate method callbacks.
981 *) Change the "progress" mechanism used in key-generation and
982 primality testing to functions that take a new BN_GENCB pointer in
983 place of callback/argument pairs. The new API functions have "_ex"
984 postfixes and the older functions are reimplemented as wrappers for
985 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
986 declarations of the old functions to help (graceful) attempts to
987 migrate to the new functions. Also, the new key-generation API
988 functions operate on a caller-supplied key-structure and return
989 success/failure rather than returning a key or NULL - this is to
990 help make "keygen" another member function of RSA_METHOD etc.
992 Example for using the new callback interface:
994 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
998 BN_GENCB_set(&my_cb, my_callback, my_arg);
1000 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1001 /* For the meaning of a, b in calls to my_callback(), see the
1002 * documentation of the function that calls the callback.
1003 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1004 * my_callback should return 1 if it wants BN_is_prime_ex()
1005 * to continue, or 0 to stop.
1010 *) Change the ZLIB compression method to be stateful, and make it
1011 available to TLS with the number defined in
1012 draft-ietf-tls-compression-04.txt.
1015 *) Add the ASN.1 structures and functions for CertificatePair, which
1016 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1018 CertificatePair ::= SEQUENCE {
1019 forward [0] Certificate OPTIONAL,
1020 reverse [1] Certificate OPTIONAL,
1021 -- at least one of the pair shall be present -- }
1023 Also implement the PEM functions to read and write certificate
1024 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1026 This needed to be defined, mostly for the sake of the LDAP
1027 attribute crossCertificatePair, but may prove useful elsewhere as
1031 *) Make it possible to inhibit symlinking of shared libraries in
1032 Makefile.shared, for Cygwin's sake.
1035 *) Extend the BIGNUM API by creating a function
1036 void BN_set_negative(BIGNUM *a, int neg);
1037 and a macro that behave like
1038 int BN_is_negative(const BIGNUM *a);
1040 to avoid the need to access 'a->neg' directly in applications.
1043 *) Implement fast modular reduction for pseudo-Mersenne primes
1044 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1045 EC_GROUP_new_curve_GFp() will now automatically use this
1047 [Nils Larsch <nla@trustcenter.de>]
1049 *) Add new lock type (CRYPTO_LOCK_BN).
1052 *) Change the ENGINE framework to automatically load engines
1053 dynamically from specific directories unless they could be
1054 found to already be built in or loaded. Move all the
1055 current engines except for the cryptodev one to a new
1057 The engines in engines/ are built as shared libraries if
1058 the "shared" options was given to ./Configure or ./config.
1059 Otherwise, they are inserted in libcrypto.a.
1060 /usr/local/ssl/engines is the default directory for dynamic
1061 engines, but that can be overriden at configure time through
1062 the usual use of --prefix and/or --openssldir, and at run
1063 time with the environment variable OPENSSL_ENGINES.
1064 [Geoff Thorpe and Richard Levitte]
1066 *) Add Makefile.shared, a helper makefile to build shared
1067 libraries. Addapt Makefile.org.
1070 *) Add version info to Win32 DLLs.
1071 [Peter 'Luna' Runestig" <peter@runestig.com>]
1073 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1074 can be added using this API to created arbitrary PKCS#12
1075 files while avoiding the low level API.
1077 New options to PKCS12_create(), key or cert can be NULL and
1078 will then be omitted from the output file. The encryption
1079 algorithm NIDs can be set to -1 for no encryption, the mac
1080 iteration count can be set to 0 to omit the mac.
1082 Enhance pkcs12 utility by making the -nokeys and -nocerts
1083 options work when creating a PKCS#12 file. New option -nomac
1084 to omit the mac, NONE can be set for an encryption algorithm.
1085 New code is modified to use the enhanced PKCS12_create()
1086 instead of the low level API.
1089 *) Extend ASN1 encoder to support indefinite length constructed
1090 encoding. This can output sequences tags and octet strings in
1091 this form. Modify pk7_asn1.c to support indefinite length
1092 encoding. This is experimental and needs additional code to
1093 be useful, such as an ASN1 bio and some enhanced streaming
1096 Extend template encode functionality so that tagging is passed
1097 down to the template encoder.
1100 *) Let 'openssl req' fail if an argument to '-newkey' is not
1101 recognized instead of using RSA as a default.
1104 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1105 As these are not official, they are not included in "ALL";
1106 the "ECCdraft" ciphersuite group alias can be used to select them.
1107 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1109 *) Add ECDH engine support.
1110 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1112 *) Add ECDH in new directory crypto/ecdh/.
1113 [Douglas Stebila (Sun Microsystems Laboratories)]
1115 *) Let BN_rand_range() abort with an error after 100 iterations
1116 without success (which indicates a broken PRNG).
1119 *) Change BN_mod_sqrt() so that it verifies that the input value
1120 is really the square of the return value. (Previously,
1121 BN_mod_sqrt would show GIGO behaviour.)
1124 *) Add named elliptic curves over binary fields from X9.62, SECG,
1125 and WAP/WTLS; add OIDs that were still missing.
1127 [Sheueling Chang Shantz and Douglas Stebila
1128 (Sun Microsystems Laboratories)]
1130 *) Extend the EC library for elliptic curves over binary fields
1131 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1134 EC_GF2m_simple_method
1138 EC_GROUP_new_curve_GF2m
1139 EC_GROUP_set_curve_GF2m
1140 EC_GROUP_get_curve_GF2m
1141 EC_POINT_set_affine_coordinates_GF2m
1142 EC_POINT_get_affine_coordinates_GF2m
1143 EC_POINT_set_compressed_coordinates_GF2m
1145 Point compression for binary fields is disabled by default for
1146 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1149 As binary polynomials are represented as BIGNUMs, various members
1150 of the EC_GROUP and EC_POINT data structures can be shared
1151 between the implementations for prime fields and binary fields;
1152 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1153 are essentially identical to their ..._GFp counterparts.
1154 (For simplicity, the '..._GFp' prefix has been dropped from
1155 various internal method names.)
1157 An internal 'field_div' method (similar to 'field_mul' and
1158 'field_sqr') has been added; this is used only for binary fields.
1160 [Sheueling Chang Shantz and Douglas Stebila
1161 (Sun Microsystems Laboratories)]
1163 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1164 through methods ('mul', 'precompute_mult').
1166 The generic implementations (now internally called 'ec_wNAF_mul'
1167 and 'ec_wNAF_precomputed_mult') remain the default if these
1168 methods are undefined.
1170 [Sheueling Chang Shantz and Douglas Stebila
1171 (Sun Microsystems Laboratories)]
1173 *) New function EC_GROUP_get_degree, which is defined through
1174 EC_METHOD. For curves over prime fields, this returns the bit
1175 length of the modulus.
1177 [Sheueling Chang Shantz and Douglas Stebila
1178 (Sun Microsystems Laboratories)]
1180 *) New functions EC_GROUP_dup, EC_POINT_dup.
1181 (These simply call ..._new and ..._copy).
1183 [Sheueling Chang Shantz and Douglas Stebila
1184 (Sun Microsystems Laboratories)]
1186 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1187 Polynomials are represented as BIGNUMs (where the sign bit is not
1188 used) in the following functions [macros]:
1191 BN_GF2m_sub [= BN_GF2m_add]
1192 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1193 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1194 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1196 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1197 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1198 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1199 BN_GF2m_cmp [= BN_ucmp]
1201 (Note that only the 'mod' functions are actually for fields GF(2^m).
1202 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1204 For some functions, an the irreducible polynomial defining a
1205 field can be given as an 'unsigned int[]' with strictly
1206 decreasing elements giving the indices of those bits that are set;
1207 i.e., p[] represents the polynomial
1208 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1210 p[0] > p[1] > ... > p[k] = 0.
1211 This applies to the following functions:
1216 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1217 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1219 BN_GF2m_mod_sqrt_arr
1220 BN_GF2m_mod_solve_quad_arr
1224 Conversion can be performed by the following functions:
1229 bntest.c has additional tests for binary polynomial arithmetic.
1231 Two implementations for BN_GF2m_mod_div() are available.
1232 The default algorithm simply uses BN_GF2m_mod_inv() and
1233 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1234 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1235 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1237 [Sheueling Chang Shantz and Douglas Stebila
1238 (Sun Microsystems Laboratories)]
1240 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1241 functionality is disabled at compile-time.
1242 [Douglas Stebila <douglas.stebila@sun.com>]
1244 *) Change default behaviour of 'openssl asn1parse' so that more
1245 information is visible when viewing, e.g., a certificate:
1247 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1248 mode the content of non-printable OCTET STRINGs is output in a
1249 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1250 avoid the appearance of a printable string.
1251 [Nils Larsch <nla@trustcenter.de>]
1253 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1255 EC_GROUP_set_asn1_flag()
1256 EC_GROUP_get_asn1_flag()
1257 EC_GROUP_set_point_conversion_form()
1258 EC_GROUP_get_point_conversion_form()
1259 These control ASN1 encoding details:
1260 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1261 has been set to OPENSSL_EC_NAMED_CURVE.
1262 - Points are encoded in uncompressed form by default; options for
1263 asn1_for are as for point2oct, namely
1264 POINT_CONVERSION_COMPRESSED
1265 POINT_CONVERSION_UNCOMPRESSED
1266 POINT_CONVERSION_HYBRID
1268 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1271 EC_GROUP_get0_seed()
1272 EC_GROUP_get_seed_len()
1273 This is used only for ASN1 purposes (so far).
1274 [Nils Larsch <nla@trustcenter.de>]
1276 *) Add 'field_type' member to EC_METHOD, which holds the NID
1277 of the appropriate field type OID. The new function
1278 EC_METHOD_get_field_type() returns this value.
1279 [Nils Larsch <nla@trustcenter.de>]
1284 EC_POINT_point2hex()
1285 EC_POINT_hex2point()
1286 providing useful interfaces to EC_POINT_point2oct() and
1287 EC_POINT_oct2point().
1288 [Nils Larsch <nla@trustcenter.de>]
1290 *) Change internals of the EC library so that the functions
1291 EC_GROUP_set_generator()
1292 EC_GROUP_get_generator()
1293 EC_GROUP_get_order()
1294 EC_GROUP_get_cofactor()
1295 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1296 to methods, which would lead to unnecessary code duplication when
1297 adding different types of curves.
1298 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1300 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1301 arithmetic, and such that modified wNAFs are generated
1302 (which avoid length expansion in many cases).
1305 *) Add a function EC_GROUP_check_discriminant() (defined via
1306 EC_METHOD) that verifies that the curve discriminant is non-zero.
1308 Add a function EC_GROUP_check() that makes some sanity tests
1309 on a EC_GROUP, its generator and order. This includes
1310 EC_GROUP_check_discriminant().
1311 [Nils Larsch <nla@trustcenter.de>]
1313 *) Add ECDSA in new directory crypto/ecdsa/.
1315 Add applications 'openssl ecparam' and 'openssl ecdsa'
1316 (these are based on 'openssl dsaparam' and 'openssl dsa').
1318 ECDSA support is also included in various other files across the
1319 library. Most notably,
1320 - 'openssl req' now has a '-newkey ecdsa:file' option;
1321 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1322 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1323 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1324 them suitable for ECDSA where domain parameters must be
1325 extracted before the specific public key;
1326 - ECDSA engine support has been added.
1327 [Nils Larsch <nla@trustcenter.de>]
1329 *) Include some named elliptic curves, and add OIDs from X9.62,
1330 SECG, and WAP/WTLS. Each curve can be obtained from the new
1332 EC_GROUP_new_by_curve_name(),
1333 and the list of available named curves can be obtained with
1334 EC_get_builtin_curves().
1335 Also add a 'curve_name' member to EC_GROUP objects, which can be
1337 EC_GROUP_set_curve_name()
1338 EC_GROUP_get_curve_name()
1339 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1341 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1342 was actually never needed) and in BN_mul(). The removal in BN_mul()
1343 required a small change in bn_mul_part_recursive() and the addition
1344 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1345 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1346 bn_sub_words() and bn_add_words() except they take arrays with
1350 Changes between 0.9.7m and 0.9.7n [xx XXX xxxx]
1352 *) In the SSL/TLS server implementation, be strict about session ID
1353 context matching (which matters if an application uses a single
1354 external cache for different purposes). Previously,
1355 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1356 set. This did ensure strict client verification, but meant that,
1357 with applications using a single external cache for quite
1358 different requirements, clients could circumvent ciphersuite
1359 restrictions for a given session ID context by starting a session
1360 in a different context.
1363 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1365 *) Cleanse PEM buffers before freeing them since they may contain
1367 [Benjamin Bennett <ben@psc.edu>]
1369 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1370 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1371 authentication-only ciphersuites.
1374 *) Since AES128 and AES256 share a single mask bit in the logic of
1375 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1376 kludge to work properly if AES128 is available and AES256 isn't.
1379 *) Expand security boundary to match 1.1.1 module.
1382 *) Remove redundant features: hash file source, editing of test vectors
1383 modify fipsld to use external fips_premain.c signature.
1386 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1387 run algorithm test programs.
1390 *) Make algorithm test programs more tolerant of whitespace.
1393 *) Have SSL/TLS server implementation tolerate "mismatched" record
1394 protocol version while receiving ClientHello even if the
1395 ClientHello is fragmented. (The server can't insist on the
1396 particular protocol version it has chosen before the ServerHello
1397 message has informed the client about his choice.)
1400 *) Load error codes if they are not already present instead of using a
1401 static variable. This allows them to be cleanly unloaded and reloaded.
1404 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1406 *) Introduce limits to prevent malicious keys being able to
1407 cause a denial of service. (CVE-2006-2940)
1408 [Steve Henson, Bodo Moeller]
1410 *) Fix ASN.1 parsing of certain invalid structures that can result
1411 in a denial of service. (CVE-2006-2937) [Steve Henson]
1413 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1414 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1416 *) Fix SSL client code which could crash if connecting to a
1417 malicious SSLv2 server. (CVE-2006-4343)
1418 [Tavis Ormandy and Will Drewry, Google Security Team]
1420 *) Change ciphersuite string processing so that an explicit
1421 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1422 will no longer include "AES128-SHA"), and any other similar
1423 ciphersuite (same bitmap) from *other* protocol versions (so that
1424 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1425 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1426 changes from 0.9.8b and 0.9.8d.
1429 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1431 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1432 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1434 *) Change the Unix randomness entropy gathering to use poll() when
1435 possible instead of select(), since the latter has some
1436 undesirable limitations.
1437 [Darryl Miles via Richard Levitte and Bodo Moeller]
1439 *) Disable rogue ciphersuites:
1441 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1442 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1443 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1445 The latter two were purportedly from
1446 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1449 Also deactive the remaining ciphersuites from
1450 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1451 unofficial, and the ID has long expired.
1454 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1455 dual-core machines) and other potential thread-safety issues.
1458 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1460 *) Adapt fipsld and the build system to link against the validated FIPS
1461 module in FIPS mode.
1464 *) Fixes for VC++ 2005 build under Windows.
1467 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1468 from a Windows bash shell such as MSYS. It is autodetected from the
1469 "config" script when run from a VC++ environment. Modify standard VC++
1470 build to use fipscanister.o from the GNU make build.
1473 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1475 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1476 The value now differs depending on if you build for FIPS or not.
1477 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1478 safely run with a non-FIPSed libcrypto, as it may crash because of
1479 the difference induced by this change.
1482 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1484 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1485 (part of SSL_OP_ALL). This option used to disable the
1486 countermeasure against man-in-the-middle protocol-version
1487 rollback in the SSL 2.0 server implementation, which is a bad
1488 idea. (CVE-2005-2969)
1490 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1491 for Information Security, National Institute of Advanced Industrial
1492 Science and Technology [AIST], Japan)]
1494 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1495 mainly for FIPS compliance and not fully integrated at this stage.
1498 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1499 the exponentiation using a fixed-length exponent. (Otherwise,
1500 the information leaked through timing could expose the secret key
1501 after many signatures; cf. Bleichenbacher's attack on DSA with
1505 *) Make a new fixed-window mod_exp implementation the default for
1506 RSA, DSA, and DH private-key operations so that the sequence of
1507 squares and multiplies and the memory access pattern are
1508 independent of the particular secret key. This will mitigate
1509 cache-timing and potential related attacks.
1511 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1512 and this is automatically used by BN_mod_exp_mont() if the new flag
1513 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1514 will use this BN flag for private exponents unless the flag
1515 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1516 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1518 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1520 *) Change the client implementation for SSLv23_method() and
1521 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1522 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1523 (Previously, the SSL 2.0 backwards compatible Client Hello
1524 message format would be used even with SSL_OP_NO_SSLv2.)
1527 *) Add support for smime-type MIME parameter in S/MIME messages which some
1531 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1532 a threadsafe manner. Modify rsa code to use new function and add calls
1533 to dsa and dh code (which had race conditions before).
1536 *) Include the fixed error library code in the C error file definitions
1537 instead of fixing them up at runtime. This keeps the error code
1538 structures constant.
1541 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1543 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1546 *) Fixes for newer kerberos headers. NB: the casts are needed because
1547 the 'length' field is signed on one version and unsigned on another
1548 with no (?) obvious way to tell the difference, without these VC++
1549 complains. Also the "definition" of FAR (blank) is no longer included
1550 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1551 some needed definitions.
1554 *) Undo Cygwin change.
1557 *) Added support for proxy certificates according to RFC 3820.
1558 Because they may be a security thread to unaware applications,
1559 they must be explicitely allowed in run-time. See
1560 docs/HOWTO/proxy_certificates.txt for further information.
1563 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1565 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1566 server and client random values. Previously
1567 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1568 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1570 This change has negligible security impact because:
1572 1. Server and client random values still have 24 bytes of pseudo random
1575 2. Server and client random values are sent in the clear in the initial
1578 3. The master secret is derived using the premaster secret (48 bytes in
1579 size for static RSA ciphersuites) as well as client server and random
1582 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1585 [Stephen Henson, reported by UK NISCC]
1587 *) Use Windows randomness collection on Cygwin.
1590 *) Fix hang in EGD/PRNGD query when communication socket is closed
1591 prematurely by EGD/PRNGD.
1592 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1594 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1597 *) Back-port of selected performance improvements from development
1598 branch, as well as improved support for PowerPC platforms.
1601 *) Add lots of checks for memory allocation failure, error codes to indicate
1602 failure and freeing up memory if a failure occurs.
1603 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1605 *) Add new -passin argument to dgst.
1608 *) Perform some character comparisons of different types in X509_NAME_cmp:
1609 this is needed for some certificates that reencode DNs into UTF8Strings
1610 (in violation of RFC3280) and can't or wont issue name rollover
1614 *) Make an explicit check during certificate validation to see that
1615 the CA setting in each certificate on the chain is correct. As a
1616 side effect always do the following basic checks on extensions,
1617 not just when there's an associated purpose to the check:
1619 - if there is an unhandled critical extension (unless the user
1620 has chosen to ignore this fault)
1621 - if the path length has been exceeded (if one is set at all)
1622 - that certain extensions fit the associated purpose (if one has
1626 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1628 *) Avoid a race condition when CRLs are checked in a multi threaded
1629 environment. This would happen due to the reordering of the revoked
1630 entries during signature checking and serial number lookup. Now the
1631 encoding is cached and the serial number sort performed under a lock.
1632 Add new STACK function sk_is_sorted().
1635 *) Add Delta CRL to the extension code.
1638 *) Various fixes to s3_pkt.c so alerts are sent properly.
1639 [David Holmes <d.holmes@f5.com>]
1641 *) Reduce the chances of duplicate issuer name and serial numbers (in
1642 violation of RFC3280) using the OpenSSL certificate creation utilities.
1643 This is done by creating a random 64 bit value for the initial serial
1644 number when a serial number file is created or when a self signed
1645 certificate is created using 'openssl req -x509'. The initial serial
1646 number file is created using 'openssl x509 -next_serial' in CA.pl
1647 rather than being initialized to 1.
1650 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1652 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1653 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1654 [Joe Orton, Steve Henson]
1656 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1658 [Joe Orton, Steve Henson]
1660 *) Make it possible to have multiple active certificates with the same
1661 subject in the CA index file. This is done only if the keyword
1662 'unique_subject' is set to 'no' in the main CA section (default
1663 if 'CA_default') of the configuration file. The value is saved
1664 with the database itself in a separate index attribute file,
1665 named like the index file with '.attr' appended to the name.
1668 *) X509 verify fixes. Disable broken certificate workarounds when
1669 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1670 keyUsage extension present. Don't accept CRLs with unhandled critical
1671 extensions: since verify currently doesn't process CRL extensions this
1672 rejects a CRL with *any* critical extensions. Add new verify error codes
1676 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1677 A clarification of RFC2560 will require the use of OCTET STRINGs and
1678 some implementations cannot handle the current raw format. Since OpenSSL
1679 copies and compares OCSP nonces as opaque blobs without any attempt at
1680 parsing them this should not create any compatibility issues.
1683 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1684 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1685 this HMAC (and other) operations are several times slower than OpenSSL
1689 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1690 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1692 *) Use the correct content when signing type "other".
1695 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1697 *) Fix various bugs revealed by running the NISCC test suite:
1699 Stop out of bounds reads in the ASN1 code when presented with
1700 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1702 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1704 If verify callback ignores invalid public key errors don't try to check
1705 certificate signature with the NULL public key.
1709 *) New -ignore_err option in ocsp application to stop the server
1710 exiting on the first error in a request.
1713 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1714 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1718 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1719 extra data after the compression methods not only for TLS 1.0
1720 but also for SSL 3.0 (as required by the specification).
1721 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1723 *) Change X509_certificate_type() to mark the key as exported/exportable
1724 when it's 512 *bits* long, not 512 bytes.
1727 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1728 blocks during encryption.
1731 *) Various fixes to base64 BIO and non blocking I/O. On write
1732 flushes were not handled properly if the BIO retried. On read
1733 data was not being buffered properly and had various logic bugs.
1734 This also affects blocking I/O when the data being decoded is a
1738 *) Various S/MIME bugfixes and compatibility changes:
1739 output correct application/pkcs7 MIME type if
1740 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1741 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1742 of files as .eml work). Correctly handle very long lines in MIME
1746 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1748 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1749 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1750 a protocol version number mismatch like a decryption error
1751 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1754 *) Turn on RSA blinding by default in the default implementation
1755 to avoid a timing attack. Applications that don't want it can call
1756 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1757 They would be ill-advised to do so in most cases.
1758 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1760 *) Change RSA blinding code so that it works when the PRNG is not
1761 seeded (in this case, the secret RSA exponent is abused as
1762 an unpredictable seed -- if it is not unpredictable, there
1763 is no point in blinding anyway). Make RSA blinding thread-safe
1764 by remembering the creator's thread ID in rsa->blinding and
1765 having all other threads use local one-time blinding factors
1766 (this requires more computation than sharing rsa->blinding, but
1767 avoids excessive locking; and if an RSA object is not shared
1768 between threads, blinding will still be very fast).
1771 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1772 ENGINE as defaults for all supported algorithms irrespective of
1773 the 'flags' parameter. 'flags' is now honoured, so applications
1774 should make sure they are passing it correctly.
1777 *) Target "mingw" now allows native Windows code to be generated in
1778 the Cygwin environment as well as with the MinGW compiler.
1781 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1783 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1784 via timing by performing a MAC computation even if incorrrect
1785 block cipher padding has been found. This is a countermeasure
1786 against active attacks where the attacker has to distinguish
1787 between bad padding and a MAC verification error. (CVE-2003-0078)
1789 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1790 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1791 Martin Vuagnoux (EPFL, Ilion)]
1793 *) Make the no-err option work as intended. The intention with no-err
1794 is not to have the whole error stack handling routines removed from
1795 libcrypto, it's only intended to remove all the function name and
1796 reason texts, thereby removing some of the footprint that may not
1797 be interesting if those errors aren't displayed anyway.
1799 NOTE: it's still possible for any application or module to have it's
1800 own set of error texts inserted. The routines are there, just not
1801 used by default when no-err is given.
1804 *) Add support for FreeBSD on IA64.
1805 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1807 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1808 Kerberos function mit_des_cbc_cksum(). Before this change,
1809 the value returned by DES_cbc_cksum() was like the one from
1810 mit_des_cbc_cksum(), except the bytes were swapped.
1811 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1813 *) Allow an application to disable the automatic SSL chain building.
1814 Before this a rather primitive chain build was always performed in
1815 ssl3_output_cert_chain(): an application had no way to send the
1816 correct chain if the automatic operation produced an incorrect result.
1818 Now the chain builder is disabled if either:
1820 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1822 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1824 The reasoning behind this is that an application would not want the
1825 auto chain building to take place if extra chain certificates are
1826 present and it might also want a means of sending no additional
1827 certificates (for example the chain has two certificates and the
1831 *) Add the possibility to build without the ENGINE framework.
1832 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1834 *) Under Win32 gmtime() can return NULL: check return value in
1835 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1838 *) DSA routines: under certain error conditions uninitialized BN objects
1839 could be freed. Solution: make sure initialization is performed early
1840 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1841 Nils Larsch <nla@trustcenter.de> via PR#459)
1844 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1845 checked on reconnect on the client side, therefore session resumption
1846 could still fail with a "ssl session id is different" error. This
1847 behaviour is masked when SSL_OP_ALL is used due to
1848 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1849 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1850 followup to PR #377.
1853 *) IA-32 assembler support enhancements: unified ELF targets, support
1854 for SCO/Caldera platforms, fix for Cygwin shared build.
1857 *) Add support for FreeBSD on sparc64. As a consequence, support for
1858 FreeBSD on non-x86 processors is separate from x86 processors on
1859 the config script, much like the NetBSD support.
1860 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1862 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1864 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1867 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1868 code (06) was taken as the first octet of the session ID and the last
1869 octet was ignored consequently. As a result SSLv2 client side session
1870 caching could not have worked due to the session ID mismatch between
1872 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1876 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1877 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1881 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1882 seems that in spite of existing for more than a year, many application
1883 author have done nothing to provide the necessary callbacks, which
1884 means that this particular engine will not work properly anywhere.
1885 This is a very unfortunate situation which forces us, in the name
1886 of usability, to give the hw_ncipher.c a static lock, which is part
1888 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1889 appear in 0.9.8 or later. We EXPECT application authors to have
1890 dealt properly with this when 0.9.8 is released (unless we actually
1891 make such changes in the libcrypto locking code that changes will
1892 have to be made anyway).
1895 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1896 octets have been read, EOF or an error occurs. Without this change
1897 some truncated ASN1 structures will not produce an error.
1900 *) Disable Heimdal support, since it hasn't been fully implemented.
1901 Still give the possibility to force the use of Heimdal, but with
1902 warnings and a request that patches get sent to openssl-dev.
1905 *) Add the VC-CE target, introduce the WINCE sysname, and add
1906 INSTALL.WCE and appropriate conditionals to make it build.
1907 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1909 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1910 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1911 edit numbers of the version.
1912 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1914 *) Introduce safe string copy and catenation functions
1915 (BUF_strlcpy() and BUF_strlcat()).
1916 [Ben Laurie (CHATS) and Richard Levitte]
1918 *) Avoid using fixed-size buffers for one-line DNs.
1919 [Ben Laurie (CHATS)]
1921 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1922 resizing buffers containing secrets, and use where appropriate.
1923 [Ben Laurie (CHATS)]
1925 *) Avoid using fixed size buffers for configuration file location.
1926 [Ben Laurie (CHATS)]
1928 *) Avoid filename truncation for various CA files.
1929 [Ben Laurie (CHATS)]
1931 *) Use sizeof in preference to magic numbers.
1932 [Ben Laurie (CHATS)]
1934 *) Avoid filename truncation in cert requests.
1935 [Ben Laurie (CHATS)]
1937 *) Add assertions to check for (supposedly impossible) buffer
1939 [Ben Laurie (CHATS)]
1941 *) Don't cache truncated DNS entries in the local cache (this could
1942 potentially lead to a spoofing attack).
1943 [Ben Laurie (CHATS)]
1945 *) Fix various buffers to be large enough for hex/decimal
1946 representations in a platform independent manner.
1947 [Ben Laurie (CHATS)]
1949 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1950 resizing buffers containing secrets, and use where appropriate.
1951 [Ben Laurie (CHATS)]
1953 *) Add BIO_indent() to avoid much slightly worrying code to do
1955 [Ben Laurie (CHATS)]
1957 *) Convert sprintf()/BIO_puts() to BIO_printf().
1958 [Ben Laurie (CHATS)]
1960 *) buffer_gets() could terminate with the buffer only half
1962 [Ben Laurie (CHATS)]
1964 *) Add assertions to prevent user-supplied crypto functions from
1965 overflowing internal buffers by having large block sizes, etc.
1966 [Ben Laurie (CHATS)]
1968 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1970 [Ben Laurie (CHATS)]
1972 *) Eliminate unused copy of key in RC4.
1973 [Ben Laurie (CHATS)]
1975 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1976 [Ben Laurie (CHATS)]
1978 *) Fix off-by-one error in EGD path.
1979 [Ben Laurie (CHATS)]
1981 *) If RANDFILE path is too long, ignore instead of truncating.
1982 [Ben Laurie (CHATS)]
1984 *) Eliminate unused and incorrectly sized X.509 structure
1986 [Ben Laurie (CHATS)]
1988 *) Eliminate unused and dangerous function knumber().
1989 [Ben Laurie (CHATS)]
1991 *) Eliminate unused and dangerous structure, KSSL_ERR.
1992 [Ben Laurie (CHATS)]
1994 *) Protect against overlong session ID context length in an encoded
1995 session object. Since these are local, this does not appear to be
1997 [Ben Laurie (CHATS)]
1999 *) Change from security patch (see 0.9.6e below) that did not affect
2000 the 0.9.6 release series:
2002 Remote buffer overflow in SSL3 protocol - an attacker could
2003 supply an oversized master key in Kerberos-enabled versions.
2005 [Ben Laurie (CHATS)]
2007 *) Change the SSL kerb5 codes to match RFC 2712.
2010 *) Make -nameopt work fully for req and add -reqopt switch.
2011 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2013 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2014 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2016 *) Make sure tests can be performed even if the corresponding algorithms
2017 have been removed entirely. This was also the last step to make
2018 OpenSSL compilable with DJGPP under all reasonable conditions.
2019 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2021 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2022 to allow version independent disabling of normally unselected ciphers,
2023 which may be activated as a side-effect of selecting a single cipher.
2025 (E.g., cipher list string "RSA" enables ciphersuites that are left
2026 out of "ALL" because they do not provide symmetric encryption.
2027 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2028 [Lutz Jaenicke, Bodo Moeller]
2030 *) Add appropriate support for separate platform-dependent build
2031 directories. The recommended way to make a platform-dependent
2032 build directory is the following (tested on Linux), maybe with
2035 # Place yourself outside of the OpenSSL source tree. In
2036 # this example, the environment variable OPENSSL_SOURCE
2037 # is assumed to contain the absolute OpenSSL source directory.
2038 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2039 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2040 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2041 mkdir -p `dirname $F`
2042 ln -s $OPENSSL_SOURCE/$F $F
2045 To be absolutely sure not to disturb the source tree, a "make clean"
2046 is a good thing. If it isn't successfull, don't worry about it,
2047 it probably means the source directory is very clean.
2050 *) Make sure any ENGINE control commands make local copies of string
2051 pointers passed to them whenever necessary. Otherwise it is possible
2052 the caller may have overwritten (or deallocated) the original string
2053 data when a later ENGINE operation tries to use the stored values.
2054 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2056 *) Improve diagnostics in file reading and command-line digests.
2057 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2059 *) Add AES modes CFB and OFB to the object database. Correct an
2060 error in AES-CFB decryption.
2063 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2064 allows existing EVP_CIPHER_CTX structures to be reused after
2065 calling EVP_*Final(). This behaviour is used by encryption
2066 BIOs and some applications. This has the side effect that
2067 applications must explicitly clean up cipher contexts with
2068 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2071 *) Check the values of dna and dnb in bn_mul_recursive before calling
2072 bn_mul_comba (a non zero value means the a or b arrays do not contain
2073 n2 elements) and fallback to bn_mul_normal if either is not zero.
2076 *) Fix escaping of non-ASCII characters when using the -subj option
2077 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2080 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2081 form for "surname", serialNumber has no short form.
2082 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2083 therefore remove "mail" short name for "internet 7".
2084 The OID for unique identifiers in X509 certificates is
2085 x500UniqueIdentifier, not uniqueIdentifier.
2086 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2089 *) Add an "init" command to the ENGINE config module and auto initialize
2090 ENGINEs. Without any "init" command the ENGINE will be initialized
2091 after all ctrl commands have been executed on it. If init=1 the
2092 ENGINE is initailized at that point (ctrls before that point are run
2093 on the uninitialized ENGINE and after on the initialized one). If
2094 init=0 then the ENGINE will not be iniatialized at all.
2097 *) Fix the 'app_verify_callback' interface so that the user-defined
2098 argument is actually passed to the callback: In the
2099 SSL_CTX_set_cert_verify_callback() prototype, the callback
2100 declaration has been changed from
2103 int (*cb)(X509_STORE_CTX *,void *);
2104 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2105 i=s->ctx->app_verify_callback(&ctx)
2106 has been changed into
2107 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2109 To update applications using SSL_CTX_set_cert_verify_callback(),
2110 a dummy argument can be added to their callback functions.
2111 [D. K. Smetters <smetters@parc.xerox.com>]
2113 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2114 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2116 *) Add and OPENSSL_LOAD_CONF define which will cause
2117 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2118 This allows older applications to transparently support certain
2119 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2120 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2121 load the config file and OPENSSL_add_all_algorithms_conf() which will
2122 always load it have also been added.
2125 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2126 Adjust NIDs and EVP layer.
2127 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2129 *) Config modules support in openssl utility.
2131 Most commands now load modules from the config file,
2132 though in a few (such as version) this isn't done
2133 because it couldn't be used for anything.
2135 In the case of ca and req the config file used is
2136 the same as the utility itself: that is the -config
2137 command line option can be used to specify an
2141 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2142 use "openssl_conf" if filename is NULL use default openssl config file.
2145 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2146 config section name. Add a new flag to tolerate a missing config file
2147 and move code to CONF_modules_load_file().
2150 *) Support for crypto accelerator cards from Accelerated Encryption
2151 Processing, www.aep.ie. (Use engine 'aep')
2152 The support was copied from 0.9.6c [engine] and adapted/corrected
2153 to work with the new engine framework.
2154 [AEP Inc. and Richard Levitte]
2156 *) Support for SureWare crypto accelerator cards from Baltimore
2157 Technologies. (Use engine 'sureware')
2158 The support was copied from 0.9.6c [engine] and adapted
2159 to work with the new engine framework.
2162 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2163 make the newer ENGINE framework commands for the CHIL engine work.
2164 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2166 *) Make it possible to produce shared libraries on ReliantUNIX.
2167 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2169 *) Add the configuration target debug-linux-ppro.
2170 Make 'openssl rsa' use the general key loading routines
2171 implemented in apps.c, and make those routines able to
2172 handle the key format FORMAT_NETSCAPE and the variant
2174 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2176 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2177 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2179 *) Add -keyform to rsautl, and document -engine.
2180 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2182 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2183 BIO_R_NO_SUCH_FILE error code rather than the generic
2184 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2187 *) Add new functions
2189 ERR_peek_last_error_line
2190 ERR_peek_last_error_line_data.
2191 These are similar to
2194 ERR_peek_error_line_data,
2195 but report on the latest error recorded rather than the first one
2196 still in the error queue.
2197 [Ben Laurie, Bodo Moeller]
2199 *) default_algorithms option in ENGINE config module. This allows things
2201 default_algorithms = ALL
2202 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2205 *) Prelminary ENGINE config module.
2208 *) New experimental application configuration code.
2211 *) Change the AES code to follow the same name structure as all other
2212 symmetric ciphers, and behave the same way. Move everything to
2213 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2214 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2216 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2217 [Ben Laurie and Theo de Raadt]
2219 *) Add option to output public keys in req command.
2220 [Massimiliano Pala madwolf@openca.org]
2222 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2223 (up to about 10% better than before for P-192 and P-224).
2226 *) New functions/macros
2228 SSL_CTX_set_msg_callback(ctx, cb)
2229 SSL_CTX_set_msg_callback_arg(ctx, arg)
2230 SSL_set_msg_callback(ssl, cb)
2231 SSL_set_msg_callback_arg(ssl, arg)
2233 to request calling a callback function
2235 void cb(int write_p, int version, int content_type,
2236 const void *buf, size_t len, SSL *ssl, void *arg)
2238 whenever a protocol message has been completely received
2239 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2240 protocol version according to which the SSL library interprets
2241 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2242 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2243 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2244 specification (change_cipher_spec(20), alert(21), handshake(22)).
2245 'buf' and 'len' point to the actual message, 'ssl' to the
2246 SSL object, and 'arg' is the application-defined value set by
2247 SSL[_CTX]_set_msg_callback_arg().
2249 'openssl s_client' and 'openssl s_server' have new '-msg' options
2250 to enable a callback that displays all protocol messages.
2253 *) Change the shared library support so shared libraries are built as
2254 soon as the corresponding static library is finished, and thereby get
2255 openssl and the test programs linked against the shared library.
2256 This still only happens when the keyword "shard" has been given to
2257 the configuration scripts.
2259 NOTE: shared library support is still an experimental thing, and
2260 backward binary compatibility is still not guaranteed.
2261 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2263 *) Add support for Subject Information Access extension.
2264 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2266 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2267 additional bytes when new memory had to be allocated, not just
2268 when reusing an existing buffer.
2271 *) New command line and configuration option 'utf8' for the req command.
2272 This allows field values to be specified as UTF8 strings.
2275 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2276 runs for the former and machine-readable output for the latter.
2279 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2280 of the e-mail address in the DN (i.e., it will go into a certificate
2281 extension only). The new configuration file option 'email_in_dn = no'
2282 has the same effect.
2283 [Massimiliano Pala madwolf@openca.org]
2285 *) Change all functions with names starting with des_ to be starting
2286 with DES_ instead. Add wrappers that are compatible with libdes,
2287 but are named _ossl_old_des_*. Finally, add macros that map the
2288 des_* symbols to the corresponding _ossl_old_des_* if libdes
2289 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2290 desired, the des_* symbols will be mapped to DES_*, with one
2293 Since we provide two compatibility mappings, the user needs to
2294 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2295 compatibility is desired. The default (i.e., when that macro
2296 isn't defined) is OpenSSL 0.9.6c compatibility.
2298 There are also macros that enable and disable the support of old
2299 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2300 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2301 are defined, the default will apply: to support the old des routines.
2303 In either case, one must include openssl/des.h to get the correct
2304 definitions. Do not try to just include openssl/des_old.h, that
2307 NOTE: This is a major break of an old API into a new one. Software
2308 authors are encouraged to switch to the DES_ style functions. Some
2309 time in the future, des_old.h and the libdes compatibility functions
2310 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2311 default), and then completely removed.
2314 *) Test for certificates which contain unsupported critical extensions.
2315 If such a certificate is found during a verify operation it is
2316 rejected by default: this behaviour can be overridden by either
2317 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2318 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2319 X509_supported_extension() has also been added which returns 1 if a
2320 particular extension is supported.
2323 *) Modify the behaviour of EVP cipher functions in similar way to digests
2324 to retain compatibility with existing code.
2327 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2328 compatibility with existing code. In particular the 'ctx' parameter does
2329 not have to be to be initialized before the call to EVP_DigestInit() and
2330 it is tidied up after a call to EVP_DigestFinal(). New function
2331 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2332 EVP_MD_CTX_copy() changed to not require the destination to be
2333 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2334 requires the destination to be valid.
2336 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2337 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2340 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2341 so that complete 'Handshake' protocol structures are kept in memory
2342 instead of overwriting 'msg_type' and 'length' with 'body' data.
2345 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2346 [Massimo Santin via Richard Levitte]
2348 *) Major restructuring to the underlying ENGINE code. This includes
2349 reduction of linker bloat, separation of pure "ENGINE" manipulation
2350 (initialisation, etc) from functionality dealing with implementations
2351 of specific crypto iterfaces. This change also introduces integrated
2352 support for symmetric ciphers and digest implementations - so ENGINEs
2353 can now accelerate these by providing EVP_CIPHER and EVP_MD
2354 implementations of their own. This is detailed in crypto/engine/README
2355 as it couldn't be adequately described here. However, there are a few
2356 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2357 were changed in the original introduction of ENGINE code have now
2358 reverted back - the hooking from this code to ENGINE is now a good
2359 deal more passive and at run-time, operations deal directly with
2360 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2361 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2362 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2363 they were not being used by the framework as there is no concept of a
2364 BIGNUM_METHOD and they could not be generalised to the new
2365 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2366 ENGINE_cpy() has been removed as it cannot be consistently defined in
2370 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2373 *) Change mkdef.pl to sort symbols that get the same entry number,
2374 and make sure the automatically generated functions ERR_load_*
2375 become part of libeay.num as well.
2378 *) New function SSL_renegotiate_pending(). This returns true once
2379 renegotiation has been requested (either SSL_renegotiate() call
2380 or HelloRequest/ClientHello receveived from the peer) and becomes
2381 false once a handshake has been completed.
2382 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2383 sends a HelloRequest, but does not ensure that a handshake takes
2384 place. SSL_renegotiate_pending() is useful for checking if the
2385 client has followed the request.)
2388 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2389 By default, clients may request session resumption even during
2390 renegotiation (if session ID contexts permit); with this option,
2391 session resumption is possible only in the first handshake.
2393 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2394 more bits available for options that should not be part of
2395 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2398 *) Add some demos for certificate and certificate request creation.
2401 *) Make maximum certificate chain size accepted from the peer application
2402 settable (SSL*_get/set_max_cert_list()), as proposed by
2403 "Douglas E. Engert" <deengert@anl.gov>.
2406 *) Add support for shared libraries for Unixware-7
2407 (Boyd Lynn Gerber <gerberb@zenez.com>).
2410 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2411 be done prior to destruction. Use this to unload error strings from
2412 ENGINEs that load their own error strings. NB: This adds two new API
2413 functions to "get" and "set" this destroy handler in an ENGINE.
2416 *) Alter all existing ENGINE implementations (except "openssl" and
2417 "openbsd") to dynamically instantiate their own error strings. This
2418 makes them more flexible to be built both as statically-linked ENGINEs
2419 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2420 Also, add stub code to each that makes building them as self-contained
2421 shared-libraries easier (see README.ENGINE).
2424 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2425 implementations into applications that are completely implemented in
2426 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2427 commands that can be used to configure what shared-library to load and
2428 to control aspects of the way it is handled. Also, made an update to
2429 the README.ENGINE file that brings its information up-to-date and
2430 provides some information and instructions on the "dynamic" ENGINE
2431 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2434 *) Make it possible to unload ranges of ERR strings with a new
2435 "ERR_unload_strings" function.
2438 *) Add a copy() function to EVP_MD.
2441 *) Make EVP_MD routines take a context pointer instead of just the
2442 md_data void pointer.
2445 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2446 that the digest can only process a single chunk of data
2447 (typically because it is provided by a piece of
2448 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2449 is only going to provide a single chunk of data, and hence the
2450 framework needn't accumulate the data for oneshot drivers.
2453 *) As with "ERR", make it possible to replace the underlying "ex_data"
2454 functions. This change also alters the storage and management of global
2455 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2456 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2457 index counters. The API functions that use this state have been changed
2458 to take a "class_index" rather than pointers to the class's local STACK
2459 and counter, and there is now an API function to dynamically create new
2460 classes. This centralisation allows us to (a) plug a lot of the
2461 thread-safety problems that existed, and (b) makes it possible to clean
2462 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2463 such data would previously have always leaked in application code and
2464 workarounds were in place to make the memory debugging turn a blind eye
2465 to it. Application code that doesn't use this new function will still
2466 leak as before, but their memory debugging output will announce it now
2467 rather than letting it slide.
2469 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2470 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2471 has a return value to indicate success or failure.
2474 *) Make it possible to replace the underlying "ERR" functions such that the
2475 global state (2 LHASH tables and 2 locks) is only used by the "default"
2476 implementation. This change also adds two functions to "get" and "set"
2477 the implementation prior to it being automatically set the first time
2478 any other ERR function takes place. Ie. an application can call "get",
2479 pass the return value to a module it has just loaded, and that module
2480 can call its own "set" function using that value. This means the
2481 module's "ERR" operations will use (and modify) the error state in the
2482 application and not in its own statically linked copy of OpenSSL code.
2485 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2486 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2487 the operation, and provides a more encapsulated way for external code
2488 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2489 to use these functions rather than manually incrementing the counts.
2491 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2494 *) Add EVP test program.
2497 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2500 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2501 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2502 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2503 These allow a CRL to be built without having to access X509_CRL fields
2504 directly. Modify 'ca' application to use new functions.
2507 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2508 bug workarounds. Rollback attack detection is a security feature.
2509 The problem will only arise on OpenSSL servers when TLSv1 is not
2510 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2511 Software authors not wanting to support TLSv1 will have special reasons
2512 for their choice and can explicitly enable this option.
2513 [Bodo Moeller, Lutz Jaenicke]
2515 *) Rationalise EVP so it can be extended: don't include a union of
2516 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2517 (similar to those existing for EVP_CIPHER_CTX).
2522 EVP_MD_CTX_init(&md); /* new function call */
2523 EVP_DigestInit(&md, EVP_sha1());
2524 EVP_DigestUpdate(&md, in, len);
2525 EVP_DigestFinal(&md, out, NULL);
2526 EVP_MD_CTX_cleanup(&md); /* new function call */
2530 *) Make DES key schedule conform to the usual scheme, as well as
2531 correcting its structure. This means that calls to DES functions
2532 now have to pass a pointer to a des_key_schedule instead of a
2533 plain des_key_schedule (which was actually always a pointer
2536 des_key_schedule ks;
2538 des_set_key_checked(..., &ks);
2539 des_ncbc_encrypt(..., &ks, ...);
2541 (Note that a later change renames 'des_...' into 'DES_...'.)
2544 *) Initial reduction of linker bloat: the use of some functions, such as
2545 PEM causes large amounts of unused functions to be linked in due to
2546 poor organisation. For example pem_all.c contains every PEM function
2547 which has a knock on effect of linking in large amounts of (unused)
2548 ASN1 code. Grouping together similar functions and splitting unrelated
2549 functions prevents this.
2552 *) Cleanup of EVP macros.
2555 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2556 correct _ecb suffix.
2559 *) Add initial OCSP responder support to ocsp application. The
2560 revocation information is handled using the text based index
2561 use by the ca application. The responder can either handle
2562 requests generated internally, supplied in files (for example
2563 via a CGI script) or using an internal minimal server.
2566 *) Add configuration choices to get zlib compression for TLS.
2569 *) Changes to Kerberos SSL for RFC 2712 compliance:
2570 1. Implemented real KerberosWrapper, instead of just using
2571 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2572 2. Implemented optional authenticator field of KerberosWrapper.
2574 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2575 and authenticator structs; see crypto/krb5/.
2577 Generalized Kerberos calls to support multiple Kerberos libraries.
2578 [Vern Staats <staatsvr@asc.hpc.mil>,
2579 Jeffrey Altman <jaltman@columbia.edu>
2580 via Richard Levitte]
2582 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2583 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2584 values for each of the key sizes rather than having just
2585 parameters (and 'speed' generating keys each time).
2588 *) Speed up EVP routines.
2591 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2592 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2593 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2594 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2596 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2597 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2598 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2601 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2603 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2606 *) Added the OS2-EMX target.
2607 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2609 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2610 to support NCONF routines in extension code. New function CONF_set_nconf()
2611 to allow functions which take an NCONF to also handle the old LHASH
2612 structure: this means that the old CONF compatible routines can be
2613 retained (in particular wrt extensions) without having to duplicate the
2614 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2617 *) Enhance the general user interface with mechanisms for inner control
2618 and with possibilities to have yes/no kind of prompts.
2621 *) Change all calls to low level digest routines in the library and
2622 applications to use EVP. Add missing calls to HMAC_cleanup() and
2623 don't assume HMAC_CTX can be copied using memcpy().
2624 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2626 *) Add the possibility to control engines through control names but with
2627 arbitrary arguments instead of just a string.
2628 Change the key loaders to take a UI_METHOD instead of a callback
2629 function pointer. NOTE: this breaks binary compatibility with earlier
2630 versions of OpenSSL [engine].
2631 Adapt the nCipher code for these new conditions and add a card insertion
2635 *) Enhance the general user interface with mechanisms to better support
2636 dialog box interfaces, application-defined prompts, the possibility
2637 to use defaults (for example default passwords from somewhere else)
2638 and interrupts/cancellations.
2641 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2642 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2645 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2646 tidy up some unnecessarily weird code in 'sk_new()').
2647 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2649 *) Change the key loading routines for ENGINEs to use the same kind
2650 callback (pem_password_cb) as all other routines that need this
2654 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2655 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2656 than this minimum value is recommended.
2659 *) New random seeder for OpenVMS, using the system process statistics
2660 that are easily reachable.
2663 *) Windows apparently can't transparently handle global
2664 variables defined in DLLs. Initialisations such as:
2666 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2668 wont compile. This is used by the any applications that need to
2669 declare their own ASN1 modules. This was fixed by adding the option
2670 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2671 needed for static libraries under Win32.
2674 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2675 setting of purpose and trust fields. New X509_STORE trust and
2676 purpose functions and tidy up setting in other SSL functions.
2679 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2680 structure. These are inherited by X509_STORE_CTX when it is
2681 initialised. This allows various defaults to be set in the
2682 X509_STORE structure (such as flags for CRL checking and custom
2683 purpose or trust settings) for functions which only use X509_STORE_CTX
2684 internally such as S/MIME.
2686 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2687 trust settings if they are not set in X509_STORE. This allows X509_STORE
2688 purposes and trust (in S/MIME for example) to override any set by default.
2690 Add command line options for CRL checking to smime, s_client and s_server
2694 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2695 are set then the CRL is looked up in the X509_STORE structure and
2696 its validity and signature checked, then if the certificate is found
2697 in the CRL the verify fails with a revoked error.
2699 Various new CRL related callbacks added to X509_STORE_CTX structure.
2701 Command line options added to 'verify' application to support this.
2703 This needs some additional work, such as being able to handle multiple
2704 CRLs with different times, extension based lookup (rather than just
2705 by subject name) and ultimately more complete V2 CRL extension
2709 *) Add a general user interface API (crypto/ui/). This is designed
2710 to replace things like des_read_password and friends (backward
2711 compatibility functions using this new API are provided).
2712 The purpose is to remove prompting functions from the DES code
2713 section as well as provide for prompting through dialog boxes in
2714 a window system and the like.
2717 *) Add "ex_data" support to ENGINE so implementations can add state at a
2718 per-structure level rather than having to store it globally.
2721 *) Make it possible for ENGINE structures to be copied when retrieved by
2722 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2723 This causes the "original" ENGINE structure to act like a template,
2724 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2725 operational state can be localised to each ENGINE structure, despite the
2726 fact they all share the same "methods". New ENGINE structures returned in
2727 this case have no functional references and the return value is the single
2728 structural reference. This matches the single structural reference returned
2729 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2733 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2734 needs to match any other type at all we need to manually clear the
2738 *) Changes to the "openssl engine" utility to include;
2739 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2740 about an ENGINE's available control commands.
2741 - executing control commands from command line arguments using the
2742 '-pre' and '-post' switches. '-post' is only used if '-t' is
2743 specified and the ENGINE is successfully initialised. The syntax for
2744 the individual commands are colon-separated, for example;
2745 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2748 *) New dynamic control command support for ENGINEs. ENGINEs can now
2749 declare their own commands (numbers), names (strings), descriptions,
2750 and input types for run-time discovery by calling applications. A
2751 subset of these commands are implicitly classed as "executable"
2752 depending on their input type, and only these can be invoked through
2753 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2754 can be based on user input, config files, etc). The distinction is
2755 that "executable" commands cannot return anything other than a boolean
2756 result and can only support numeric or string input, whereas some
2757 discoverable commands may only be for direct use through
2758 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2759 pointers, or other custom uses. The "executable" commands are to
2760 support parameterisations of ENGINE behaviour that can be
2761 unambiguously defined by ENGINEs and used consistently across any
2762 OpenSSL-based application. Commands have been added to all the
2763 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2764 control over shared-library paths without source code alterations.
2767 *) Changed all ENGINE implementations to dynamically allocate their
2768 ENGINEs rather than declaring them statically. Apart from this being
2769 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2770 this also allows the implementations to compile without using the
2771 internal engine_int.h header.
2774 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2775 'const' value. Any code that should be able to modify a RAND_METHOD
2776 should already have non-const pointers to it (ie. they should only
2777 modify their own ones).
2780 *) Made a variety of little tweaks to the ENGINE code.
2781 - "atalla" and "ubsec" string definitions were moved from header files
2782 to C code. "nuron" string definitions were placed in variables
2783 rather than hard-coded - allowing parameterisation of these values
2784 later on via ctrl() commands.
2785 - Removed unused "#if 0"'d code.
2786 - Fixed engine list iteration code so it uses ENGINE_free() to release
2787 structural references.
2788 - Constified the RAND_METHOD element of ENGINE structures.
2789 - Constified various get/set functions as appropriate and added
2790 missing functions (including a catch-all ENGINE_cpy that duplicates
2791 all ENGINE values onto a new ENGINE except reference counts/state).
2792 - Removed NULL parameter checks in get/set functions. Setting a method
2793 or function to NULL is a way of cancelling out a previously set
2794 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2795 and doesn't justify the extra error symbols and code.
2796 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2797 flags from engine_int.h to engine.h.
2798 - Changed prototypes for ENGINE handler functions (init(), finish(),
2799 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2802 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2803 to the algorithm using long division. The binary algorithm can be
2804 used only if the modulus is odd. On 32-bit systems, it is faster
2805 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2806 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2807 up to 450 bits. In 64-bit environments, the binary algorithm
2808 appears to be advantageous for much longer moduli; here we use it
2809 for moduli up to 2048 bits.
2812 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2813 could not support the combine flag in choice fields.
2816 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2817 extensions from a certificate request to the certificate.
2820 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2821 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2822 file: this allows the display of the certificate about to be
2823 signed to be customised, to allow certain fields to be included
2824 or excluded and extension details. The old system didn't display
2825 multicharacter strings properly, omitted fields not in the policy
2826 and couldn't display additional details such as extensions.
2829 *) Function EC_POINTs_mul for multiple scalar multiplication
2830 of an arbitrary number of elliptic curve points
2831 \sum scalars[i]*points[i],
2832 optionally including the generator defined for the EC_GROUP:
2833 scalar*generator + \sum scalars[i]*points[i].
2835 EC_POINT_mul is a simple wrapper function for the typical case
2836 that the point list has just one item (besides the optional
2840 *) First EC_METHODs for curves over GF(p):
2842 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2843 operations and provides various method functions that can also
2844 operate with faster implementations of modular arithmetic.
2846 EC_GFp_mont_method() reuses most functions that are part of
2847 EC_GFp_simple_method, but uses Montgomery arithmetic.
2849 [Bodo Moeller; point addition and point doubling
2850 implementation directly derived from source code provided by
2851 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2853 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2854 crypto/ec/ec_lib.c):
2856 Curves are EC_GROUP objects (with an optional group generator)
2857 based on EC_METHODs that are built into the library.
2859 Points are EC_POINT objects based on EC_GROUP objects.
2861 Most of the framework would be able to handle curves over arbitrary
2862 finite fields, but as there are no obvious types for fields other
2863 than GF(p), some functions are limited to that for now.
2866 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2867 that the file contains a complete HTTP response.
2870 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2871 change the def and num file printf format specifier from "%-40sXXX"
2872 to "%-39s XXX". The latter will always guarantee a space after the
2873 field while the former will cause them to run together if the field
2874 is 40 of more characters long.
2877 *) Constify the cipher and digest 'method' functions and structures
2878 and modify related functions to take constant EVP_MD and EVP_CIPHER
2882 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2883 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2886 *) Modify EVP_Digest*() routines so they now return values. Although the
2887 internal software routines can never fail additional hardware versions
2891 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2893 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2894 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2897 ERR_R_NESTED_ASN1_ERROR
2899 ERR_R_MISSING_ASN1_EOS
2900 were 4 .. 9, conflicting with
2901 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2903 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2904 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2906 Add new error code 'ERR_R_INTERNAL_ERROR'.
2909 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2913 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2914 sets the subject name for a new request or supersedes the
2915 subject name in a given request. Formats that can be parsed are
2916 'CN=Some Name, OU=myOU, C=IT'
2918 'CN=Some Name/OU=myOU/C=IT'.
2920 Add options '-batch' and '-verbose' to 'openssl req'.
2921 [Massimiliano Pala <madwolf@hackmasters.net>]
2923 *) Introduce the possibility to access global variables through
2924 functions on platform were that's the best way to handle exporting
2925 global variables in shared libraries. To enable this functionality,
2926 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2927 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2928 is normally done by Configure or something similar).
2930 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2931 in the source file (foo.c) like this:
2933 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2934 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2936 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2937 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2939 OPENSSL_DECLARE_GLOBAL(int,foo);
2940 #define foo OPENSSL_GLOBAL_REF(foo)
2941 OPENSSL_DECLARE_GLOBAL(double,bar);
2942 #define bar OPENSSL_GLOBAL_REF(bar)
2944 The #defines are very important, and therefore so is including the
2945 header file everywhere where the defined globals are used.
2947 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2948 of ASN.1 items, but that structure is a bit different.
2950 The largest change is in util/mkdef.pl which has been enhanced with
2951 better and easier to understand logic to choose which symbols should
2952 go into the Windows .def files as well as a number of fixes and code
2953 cleanup (among others, algorithm keywords are now sorted
2954 lexicographically to avoid constant rewrites).
2957 *) In BN_div() keep a copy of the sign of 'num' before writing the
2958 result to 'rm' because if rm==num the value will be overwritten
2959 and produce the wrong result if 'num' is negative: this caused
2960 problems with BN_mod() and BN_nnmod().
2963 *) Function OCSP_request_verify(). This checks the signature on an
2964 OCSP request and verifies the signer certificate. The signer
2965 certificate is just checked for a generic purpose and OCSP request
2969 *) Add OCSP_check_validity() function to check the validity of OCSP
2970 responses. OCSP responses are prepared in real time and may only
2971 be a few seconds old. Simply checking that the current time lies
2972 between thisUpdate and nextUpdate max reject otherwise valid responses
2973 caused by either OCSP responder or client clock inaccuracy. Instead
2974 we allow thisUpdate and nextUpdate to fall within a certain period of
2975 the current time. The age of the response can also optionally be
2976 checked. Two new options -validity_period and -status_age added to
2980 *) If signature or public key algorithm is unrecognized print out its
2981 OID rather that just UNKNOWN.
2984 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2985 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2986 ID to be generated from the issuer certificate alone which can then be
2987 passed to OCSP_id_issuer_cmp().
2990 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2991 ASN1 modules to export functions returning ASN1_ITEM pointers
2992 instead of the ASN1_ITEM structures themselves. This adds several
2993 new macros which allow the underlying ASN1 function/structure to
2994 be accessed transparently. As a result code should not use ASN1_ITEM
2995 references directly (such as &X509_it) but instead use the relevant
2996 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2997 use of the new ASN1 code on platforms where exporting structures
2998 is problematical (for example in shared libraries) but exporting
2999 functions returning pointers to structures is not.
3002 *) Add support for overriding the generation of SSL/TLS session IDs.
3003 These callbacks can be registered either in an SSL_CTX or per SSL.
3004 The purpose of this is to allow applications to control, if they wish,
3005 the arbitrary values chosen for use as session IDs, particularly as it
3006 can be useful for session caching in multiple-server environments. A
3007 command-line switch for testing this (and any client code that wishes
3008 to use such a feature) has been added to "s_server".
3009 [Geoff Thorpe, Lutz Jaenicke]
3011 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3012 of the form '#if defined(...) || defined(...) || ...' and
3013 '#if !defined(...) && !defined(...) && ...'. This also avoids
3014 the growing number of special cases it was previously handling.
3017 *) Make all configuration macros available for application by making
3018 sure they are available in opensslconf.h, by giving them names starting
3019 with "OPENSSL_" to avoid conflicts with other packages and by making
3020 sure e_os2.h will cover all platform-specific cases together with
3022 Additionally, it is now possible to define configuration/platform-
3023 specific names (called "system identities"). In the C code, these
3024 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3025 macro with the name beginning with "OPENSSL_SYS_", which is determined
3026 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3030 *) New option -set_serial to 'req' and 'x509' this allows the serial
3031 number to use to be specified on the command line. Previously self
3032 signed certificates were hard coded with serial number 0 and the
3033 CA options of 'x509' had to use a serial number in a file which was
3037 *) New options to 'ca' utility to support V2 CRL entry extensions.
3038 Currently CRL reason, invalidity date and hold instruction are
3039 supported. Add new CRL extensions to V3 code and some new objects.
3042 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3043 disable standard block padding (aka PKCS#5 padding) in the EVP
3044 API, which was previously mandatory. This means that the data is
3045 not padded in any way and so the total length much be a multiple
3046 of the block size, otherwise an error occurs.
3049 *) Initial (incomplete) OCSP SSL support.
3052 *) New function OCSP_parse_url(). This splits up a URL into its host,
3053 port and path components: primarily to parse OCSP URLs. New -url
3054 option to ocsp utility.
3057 *) New nonce behavior. The return value of OCSP_check_nonce() now
3058 reflects the various checks performed. Applications can decide
3059 whether to tolerate certain situations such as an absent nonce
3060 in a response when one was present in a request: the ocsp application
3061 just prints out a warning. New function OCSP_add1_basic_nonce()
3062 this is to allow responders to include a nonce in a response even if
3063 the request is nonce-less.
3066 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3067 skipped when using openssl x509 multiple times on a single input file,
3068 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3071 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3072 set string type: to handle setting ASN1_TIME structures. Fix ca
3073 utility to correctly initialize revocation date of CRLs.
3076 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3077 the clients preferred ciphersuites and rather use its own preferences.
3078 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3079 Internet Explorer by ensuring unchanged hash method during stepup.
3080 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3083 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3084 to aes and add a new 'exist' option to print out symbols that don't
3088 *) Additional options to ocsp utility to allow flags to be set and
3089 additional certificates supplied.
3092 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3093 OCSP client a number of certificate to only verify the response
3097 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3098 handle the new API. Currently only ECB, CBC modes supported. Add new
3101 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3102 Encryption Standard (AES) Ciphersuites for Transport Layer
3103 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3104 not enabled by default and were not part of the "ALL" ciphersuite
3105 alias because they were not yet official; they could be
3106 explicitly requested by specifying the "AESdraft" ciphersuite
3107 group alias. In the final release of OpenSSL 0.9.7, the group
3108 alias is called "AES" and is part of "ALL".)
3109 [Ben Laurie, Steve Henson, Bodo Moeller]
3111 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3112 request to response.
3115 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3116 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3117 extract information from a certificate request. OCSP_response_create()
3118 creates a response and optionally adds a basic response structure.
3119 OCSP_basic_add1_status() adds a complete single response to a basic
3120 response and returns the OCSP_SINGLERESP structure just added (to allow
3121 extensions to be included for example). OCSP_basic_add1_cert() adds a
3122 certificate to a basic response and OCSP_basic_sign() signs a basic
3123 response with various flags. New helper functions ASN1_TIME_check()
3124 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3125 (converts ASN1_TIME to GeneralizedTime).
3128 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3129 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3130 structure from a certificate. X509_pubkey_digest() digests the public_key
3131 contents: this is used in various key identifiers.
3134 *) Make sk_sort() tolerate a NULL argument.
3135 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3137 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3138 passed by the function are trusted implicitly. If any of them signed the
3139 response then it is assumed to be valid and is not verified.
3142 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3143 to data. This was previously part of the PKCS7 ASN1 code. This
3144 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3145 [Steve Henson, reported by Kenneth R. Robinette
3146 <support@securenetterm.com>]
3148 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3149 routines: without these tracing memory leaks is very painful.
3150 Fix leaks in PKCS12 and PKCS7 routines.
3153 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3154 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3155 effectively meant GeneralizedTime would never be used. Now it
3156 is initialised to -1 but X509_time_adj() now has to check the value
3157 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3158 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3159 [Steve Henson, reported by Kenneth R. Robinette
3160 <support@securenetterm.com>]
3162 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3163 result in a zero length in the ASN1_INTEGER structure which was
3164 not consistent with the structure when d2i_ASN1_INTEGER() was used
3165 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3166 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3167 where it did not print out a minus for negative ASN1_INTEGER.
3170 *) Add summary printout to ocsp utility. The various functions which
3171 convert status values to strings have been renamed to:
3172 OCSP_response_status_str(), OCSP_cert_status_str() and
3173 OCSP_crl_reason_str() and are no longer static. New options
3174 to verify nonce values and to disable verification. OCSP response
3175 printout format cleaned up.
3178 *) Add additional OCSP certificate checks. These are those specified
3179 in RFC2560. This consists of two separate checks: the CA of the
3180 certificate being checked must either be the OCSP signer certificate
3181 or the issuer of the OCSP signer certificate. In the latter case the
3182 OCSP signer certificate must contain the OCSP signing extended key
3183 usage. This check is performed by attempting to match the OCSP
3184 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3185 in the OCSP_CERTID structures of the response.
3188 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3189 and related routines. This uses the standard OpenSSL certificate
3190 verify routines to perform initial checks (just CA validity) and
3191 to obtain the certificate chain. Then additional checks will be
3192 performed on the chain. Currently the root CA is checked to see
3193 if it is explicitly trusted for OCSP signing. This is used to set
3194 a root CA as a global signing root: that is any certificate that
3195 chains to that CA is an acceptable OCSP signing certificate.
3198 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3199 extensions from a separate configuration file.
3200 As when reading extensions from the main configuration file,
3201 the '-extensions ...' option may be used for specifying the
3203 [Massimiliano Pala <madwolf@comune.modena.it>]
3205 *) New OCSP utility. Allows OCSP requests to be generated or
3206 read. The request can be sent to a responder and the output
3207 parsed, outputed or printed in text form. Not complete yet:
3208 still needs to check the OCSP response validity.
3211 *) New subcommands for 'openssl ca':
3212 'openssl ca -status <serial>' prints the status of the cert with
3213 the given serial number (according to the index file).
3214 'openssl ca -updatedb' updates the expiry status of certificates
3216 [Massimiliano Pala <madwolf@comune.modena.it>]
3218 *) New '-newreq-nodes' command option to CA.pl. This is like
3219 '-newreq', but calls 'openssl req' with the '-nodes' option
3220 so that the resulting key is not encrypted.
3221 [Damien Miller <djm@mindrot.org>]
3223 *) New configuration for the GNU Hurd.
3224 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3226 *) Initial code to implement OCSP basic response verify. This
3227 is currently incomplete. Currently just finds the signer's
3228 certificate and verifies the signature on the response.
3231 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3232 value of OPENSSLDIR. This is available via the new '-d' option
3233 to 'openssl version', and is also included in 'openssl version -a'.
3236 *) Allowing defining memory allocation callbacks that will be given
3237 file name and line number information in additional arguments
3238 (a const char* and an int). The basic functionality remains, as
3239 well as the original possibility to just replace malloc(),
3240 realloc() and free() by functions that do not know about these
3241 additional arguments. To register and find out the current
3242 settings for extended allocation functions, the following
3243 functions are provided:
3245 CRYPTO_set_mem_ex_functions
3246 CRYPTO_set_locked_mem_ex_functions
3247 CRYPTO_get_mem_ex_functions
3248 CRYPTO_get_locked_mem_ex_functions
3250 These work the same way as CRYPTO_set_mem_functions and friends.
3251 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3252 extended allocation function is enabled.
3253 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3254 a conventional allocation function is enabled.
3255 [Richard Levitte, Bodo Moeller]
3257 *) Finish off removing the remaining LHASH function pointer casts.
3258 There should no longer be any prototype-casting required when using
3259 the LHASH abstraction, and any casts that remain are "bugs". See
3260 the callback types and macros at the head of lhash.h for details
3261 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3264 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3265 If /dev/[u]random devices are not available or do not return enough
3266 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3268 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3269 /etc/entropy will be queried once each in this sequence, quering stops
3270 when enough entropy was collected without querying more sockets.
3273 *) Change the Unix RAND_poll() variant to be able to poll several
3274 random devices, as specified by DEVRANDOM, until a sufficient amount
3275 of data has been collected. We spend at most 10 ms on each file
3276 (select timeout) and read in non-blocking mode. DEVRANDOM now
3277 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3278 (previously it was just the string "/dev/urandom"), so on typical
3279 platforms the 10 ms delay will never occur.
3280 Also separate out the Unix variant to its own file, rand_unix.c.
3281 For VMS, there's a currently-empty rand_vms.c.
3284 *) Move OCSP client related routines to ocsp_cl.c. These
3285 provide utility functions which an application needing
3286 to issue a request to an OCSP responder and analyse the
3287 response will typically need: as opposed to those which an
3288 OCSP responder itself would need which will be added later.
3290 OCSP_request_sign() signs an OCSP request with an API similar
3291 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3292 response. OCSP_response_get1_basic() extracts basic response
3293 from response. OCSP_resp_find_status(): finds and extracts status
3294 information from an OCSP_CERTID structure (which will be created
3295 when the request structure is built). These are built from lower
3296 level functions which work on OCSP_SINGLERESP structures but
3297 wont normally be used unless the application wishes to examine
3298 extensions in the OCSP response for example.
3300 Replace nonce routines with a pair of functions.
3301 OCSP_request_add1_nonce() adds a nonce value and optionally
3302 generates a random value. OCSP_check_nonce() checks the
3303 validity of the nonce in an OCSP response.
3306 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3307 This doesn't copy the supplied OCSP_CERTID and avoids the
3308 need to free up the newly created id. Change return type
3309 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3310 This can then be used to add extensions to the request.
3311 Deleted OCSP_request_new(), since most of its functionality
3312 is now in OCSP_REQUEST_new() (and the case insensitive name
3313 clash) apart from the ability to set the request name which
3314 will be added elsewhere.
3317 *) Update OCSP API. Remove obsolete extensions argument from
3318 various functions. Extensions are now handled using the new
3319 OCSP extension code. New simple OCSP HTTP function which
3320 can be used to send requests and parse the response.
3323 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3324 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3325 uses the special reorder version of SET OF to sort the attributes
3326 and reorder them to match the encoded order. This resolves a long
3327 standing problem: a verify on a PKCS7 structure just after signing
3328 it used to fail because the attribute order did not match the
3329 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3330 it uses the received order. This is necessary to tolerate some broken
3331 software that does not order SET OF. This is handled by encoding
3332 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3333 to produce the required SET OF.
3336 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3337 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3338 files to get correct declarations of the ASN.1 item variables.
3341 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3342 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3343 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3344 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3345 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3346 ASN1_ITEM and no wrapper functions.
3349 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3350 replace the old function pointer based I/O routines. Change most of
3351 the *_d2i_bio() and *_d2i_fp() functions to use these.
3354 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3355 lines, recognice more "algorithms" that can be deselected, and make
3356 it complain about algorithm deselection that isn't recognised.
3359 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3360 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3361 to use new functions. Add NO_ASN1_OLD which can be set to remove
3362 some old style ASN1 functions: this can be used to determine if old
3363 code will still work when these eventually go away.
3366 *) New extension functions for OCSP structures, these follow the
3367 same conventions as certificates and CRLs.
3370 *) New function X509V3_add1_i2d(). This automatically encodes and
3371 adds an extension. Its behaviour can be customised with various
3372 flags to append, replace or delete. Various wrappers added for
3373 certifcates and CRLs.
3376 *) Fix to avoid calling the underlying ASN1 print routine when
3377 an extension cannot be parsed. Correct a typo in the
3378 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3381 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3382 entries for variables.
3385 *) Add functionality to apps/openssl.c for detecting locking
3386 problems: As the program is single-threaded, all we have
3387 to do is register a locking callback using an array for
3388 storing which locks are currently held by the program.
3391 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3392 SSL_get_ex_data_X509_STORE_idx(), which is used in
3393 ssl_verify_cert_chain() and thus can be called at any time
3394 during TLS/SSL handshakes so that thread-safety is essential.
3395 Unfortunately, the ex_data design is not at all suited
3396 for multi-threaded use, so it probably should be abolished.
3399 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3400 [Broadcom, tweaked and integrated by Geoff Thorpe]
3402 *) Move common extension printing code to new function
3403 X509V3_print_extensions(). Reorganise OCSP print routines and
3404 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3407 *) New function X509_signature_print() to remove duplication in some
3411 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3412 set (this was treated exactly the same as SET OF previously). This
3413 is used to reorder the STACK representing the structure to match the
3414 encoding. This will be used to get round a problem where a PKCS7
3415 structure which was signed could not be verified because the STACK
3416 order did not reflect the encoded order.
3419 *) Reimplement the OCSP ASN1 module using the new code.
3422 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3423 for its ASN1 operations. The old style function pointers still exist
3424 for now but they will eventually go away.
3427 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3428 completely replaces the old ASN1 functionality with a table driven
3429 encoder and decoder which interprets an ASN1_ITEM structure describing
3430 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3431 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3432 has also been converted to the new form.
3435 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3436 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3437 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3438 for negative moduli.
3441 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3442 of not touching the result's sign bit.
3445 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3449 *) Changed the LHASH code to use prototypes for callbacks, and created
3450 macros to declare and implement thin (optionally static) functions
3451 that provide type-safety and avoid function pointer casting for the
3452 type-specific callbacks.
3455 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3457 [Veers Staats <staatsvr@asc.hpc.mil>,
3458 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3460 *) Reformat the FAQ so the different questions and answers can be divided
3461 in sections depending on the subject.
3464 *) Have the zlib compression code load ZLIB.DLL dynamically under
3468 *) New function BN_mod_sqrt for computing square roots modulo a prime
3469 (using the probabilistic Tonelli-Shanks algorithm unless
3470 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3471 be handled deterministically).
3472 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3474 *) Make BN_mod_inverse faster by explicitly handling small quotients
3475 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3476 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3479 *) New function BN_kronecker.
3482 *) Fix BN_gcd so that it works on negative inputs; the result is
3483 positive unless both parameters are zero.
3484 Previously something reasonably close to an infinite loop was
3485 possible because numbers could be growing instead of shrinking
3486 in the implementation of Euclid's algorithm.
3489 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3490 sign of the number in question.
3492 Fix BN_is_word(a,w) to work correctly for w == 0.
3494 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3495 because its test if the absolute value of 'a' equals 'w'.
3496 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3497 it exists mostly for use in the implementations of BN_is_zero(),
3498 BN_is_one(), and BN_is_word().
3501 *) New function BN_swap.
3504 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3505 the exponentiation functions are more likely to produce reasonable
3506 results on negative inputs.
3509 *) Change BN_mod_mul so that the result is always non-negative.
3510 Previously, it could be negative if one of the factors was negative;
3511 I don't think anyone really wanted that behaviour.
3514 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3515 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3516 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3517 and add new functions:
3526 BN_mod_lshift1_quick
3530 These functions always generate non-negative results.
3532 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3533 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3535 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3536 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3537 be reduced modulo m.
3538 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3541 The following entry accidentily appeared in the CHANGES file
3542 distributed with OpenSSL 0.9.7. The modifications described in
3543 it do *not* apply to OpenSSL 0.9.7.
3545 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3546 was actually never needed) and in BN_mul(). The removal in BN_mul()
3547 required a small change in bn_mul_part_recursive() and the addition
3548 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3549 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3550 bn_sub_words() and bn_add_words() except they take arrays with
3555 *) In 'openssl passwd', verify passwords read from the terminal
3556 unless the '-salt' option is used (which usually means that
3557 verification would just waste user's time since the resulting
3558 hash is going to be compared with some given password hash)
3559 or the new '-noverify' option is used.
3561 This is an incompatible change, but it does not affect
3562 non-interactive use of 'openssl passwd' (passwords on the command
3563 line, '-stdin' option, '-in ...' option) and thus should not
3567 *) Remove all references to RSAref, since there's no more need for it.
3570 *) Make DSO load along a path given through an environment variable
3571 (SHLIB_PATH) with shl_load().
3574 *) Constify the ENGINE code as a result of BIGNUM constification.
3575 Also constify the RSA code and most things related to it. In a
3576 few places, most notable in the depth of the ASN.1 code, ugly
3577 casts back to non-const were required (to be solved at a later
3581 *) Make it so the openssl application has all engines loaded by default.
3584 *) Constify the BIGNUM routines a little more.
3587 *) Add the following functions:
3589 ENGINE_load_cswift()
3591 ENGINE_load_atalla()
3593 ENGINE_load_builtin_engines()
3595 That way, an application can itself choose if external engines that
3596 are built-in in OpenSSL shall ever be used or not. The benefit is
3597 that applications won't have to be linked with libdl or other dso
3598 libraries unless it's really needed.
3600 Changed 'openssl engine' to load all engines on demand.
3601 Changed the engine header files to avoid the duplication of some
3602 declarations (they differed!).
3605 *) 'openssl engine' can now list capabilities.
3608 *) Better error reporting in 'openssl engine'.
3611 *) Never call load_dh_param(NULL) in s_server.
3614 *) Add engine application. It can currently list engines by name and
3615 identity, and test if they are actually available.
3618 *) Improve RPM specification file by forcing symbolic linking and making
3619 sure the installed documentation is also owned by root.root.
3620 [Damien Miller <djm@mindrot.org>]
3622 *) Give the OpenSSL applications more possibilities to make use of
3623 keys (public as well as private) handled by engines.
3626 *) Add OCSP code that comes from CertCo.
3629 *) Add VMS support for the Rijndael code.
3632 *) Added untested support for Nuron crypto accelerator.
3635 *) Add support for external cryptographic devices. This code was
3636 previously distributed separately as the "engine" branch.
3637 [Geoff Thorpe, Richard Levitte]
3639 *) Rework the filename-translation in the DSO code. It is now possible to
3640 have far greater control over how a "name" is turned into a filename
3641 depending on the operating environment and any oddities about the
3642 different shared library filenames on each system.
3645 *) Support threads on FreeBSD-elf in Configure.
3648 *) Fix for SHA1 assembly problem with MASM: it produces
3649 warnings about corrupt line number information when assembling
3650 with debugging information. This is caused by the overlapping
3652 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3655 NCONF_get_number() has no error checking at all. As a replacement,
3656 NCONF_get_number_e() is defined (_e for "error checking") and is
3657 promoted strongly. The old NCONF_get_number is kept around for
3658 binary backward compatibility.
3659 Make it possible for methods to load from something other than a BIO,
3660 by providing a function pointer that is given a name instead of a BIO.
3661 For example, this could be used to load configuration data from an
3665 *) Fix for non blocking accept BIOs. Added new I/O special reason
3666 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3667 with non blocking I/O was not possible because no retry code was
3668 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3672 *) Added the beginnings of Rijndael support.
3675 *) Fix for bug in DirectoryString mask setting. Add support for
3676 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3677 to allow certificate printing to more controllable, additional
3678 'certopt' option to 'x509' to allow new printing options to be
3682 *) Clean old EAY MD5 hack from e_os.h.
3685 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3687 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3688 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3689 [Joe Orton, Steve Henson]
3691 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3693 *) Fix additional bug revealed by the NISCC test suite:
3695 Stop bug triggering large recursion when presented with
3696 certain ASN.1 tags (CVE-2003-0851)
3699 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3701 *) Fix various bugs revealed by running the NISCC test suite:
3703 Stop out of bounds reads in the ASN1 code when presented with
3704 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3706 If verify callback ignores invalid public key errors don't try to check
3707 certificate signature with the NULL public key.
3711 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3712 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3716 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3717 extra data after the compression methods not only for TLS 1.0
3718 but also for SSL 3.0 (as required by the specification).
3719 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3721 *) Change X509_certificate_type() to mark the key as exported/exportable
3722 when it's 512 *bits* long, not 512 bytes.
3725 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3727 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3728 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3729 a protocol version number mismatch like a decryption error
3730 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3733 *) Turn on RSA blinding by default in the default implementation
3734 to avoid a timing attack. Applications that don't want it can call
3735 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3736 They would be ill-advised to do so in most cases.
3737 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3739 *) Change RSA blinding code so that it works when the PRNG is not
3740 seeded (in this case, the secret RSA exponent is abused as
3741 an unpredictable seed -- if it is not unpredictable, there
3742 is no point in blinding anyway). Make RSA blinding thread-safe
3743 by remembering the creator's thread ID in rsa->blinding and
3744 having all other threads use local one-time blinding factors
3745 (this requires more computation than sharing rsa->blinding, but
3746 avoids excessive locking; and if an RSA object is not shared
3747 between threads, blinding will still be very fast).
3750 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3752 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3753 via timing by performing a MAC computation even if incorrrect
3754 block cipher padding has been found. This is a countermeasure
3755 against active attacks where the attacker has to distinguish
3756 between bad padding and a MAC verification error. (CVE-2003-0078)
3758 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3759 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3760 Martin Vuagnoux (EPFL, Ilion)]
3762 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3764 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3765 memory from it's contents. This is done with a counter that will
3766 place alternating values in each byte. This can be used to solve
3767 two issues: 1) the removal of calls to memset() by highly optimizing
3768 compilers, and 2) cleansing with other values than 0, since those can
3769 be read through on certain media, for example a swap space on disk.
3772 *) Bugfix: client side session caching did not work with external caching,
3773 because the session->cipher setting was not restored when reloading
3774 from the external cache. This problem was masked, when
3775 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3776 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3779 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3780 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3781 [Zeev Lieber <zeev-l@yahoo.com>]
3783 *) Undo an undocumented change introduced in 0.9.6e which caused
3784 repeated calls to OpenSSL_add_all_ciphers() and
3785 OpenSSL_add_all_digests() to be ignored, even after calling
3789 *) Change the default configuration reader to deal with last line not
3790 being properly terminated.
3793 *) Change X509_NAME_cmp() so it applies the special rules on handling
3794 DN values that are of type PrintableString, as well as RDNs of type
3795 emailAddress where the value has the type ia5String.
3796 [stefank@valicert.com via Richard Levitte]
3798 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3799 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3800 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3801 the bitwise-OR of the two for use by the majority of applications
3802 wanting this behaviour, and update the docs. The documented
3803 behaviour and actual behaviour were inconsistent and had been
3804 changing anyway, so this is more a bug-fix than a behavioural
3806 [Geoff Thorpe, diagnosed by Nadav Har'El]
3808 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3809 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3812 *) Fix initialization code race conditions in
3813 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3814 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3815 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3816 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3817 ssl2_get_cipher_by_char(),
3818 ssl3_get_cipher_by_char().
3819 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3821 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3822 the cached sessions are flushed, as the remove_cb() might use ex_data
3823 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3824 (see [openssl.org #212]).
3825 [Geoff Thorpe, Lutz Jaenicke]
3827 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3828 length, instead of the encoding length to d2i_ASN1_OBJECT.
3831 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3833 *) [In 0.9.6g-engine release:]
3834 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3835 [Lynn Gazis <lgazis@rainbow.com>]
3837 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3839 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3840 and get fix the header length calculation.
3841 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3842 Alon Kantor <alonk@checkpoint.com> (and others),
3845 *) Use proper error handling instead of 'assertions' in buffer
3846 overflow checks added in 0.9.6e. This prevents DoS (the
3847 assertions could call abort()).
3848 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3850 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3852 *) Add various sanity checks to asn1_get_length() to reject
3853 the ASN1 length bytes if they exceed sizeof(long), will appear
3854 negative or the content length exceeds the length of the
3856 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3858 *) Fix cipher selection routines: ciphers without encryption had no flags
3859 for the cipher strength set and where therefore not handled correctly
3860 by the selection routines (PR #130).
3863 *) Fix EVP_dsa_sha macro.
3867 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3868 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3869 that was added in OpenSSL 0.9.6d.
3871 As the countermeasure turned out to be incompatible with some
3872 broken SSL implementations, the new option is part of SSL_OP_ALL.
3873 SSL_OP_ALL is usually employed when compatibility with weird SSL
3874 implementations is desired (e.g. '-bugs' option to 's_client' and
3875 's_server'), so the new option is automatically set in many
3879 *) Changes in security patch:
3881 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3882 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3883 Air Force Materiel Command, USAF, under agreement number
3886 *) Add various sanity checks to asn1_get_length() to reject
3887 the ASN1 length bytes if they exceed sizeof(long), will appear
3888 negative or the content length exceeds the length of the
3889 supplied buffer. (CVE-2002-0659)
3890 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3892 *) Assertions for various potential buffer overflows, not known to
3894 [Ben Laurie (CHATS)]
3896 *) Various temporary buffers to hold ASCII versions of integers were
3897 too small for 64 bit platforms. (CVE-2002-0655)
3898 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3900 *) Remote buffer overflow in SSL3 protocol - an attacker could
3901 supply an oversized session ID to a client. (CVE-2002-0656)
3902 [Ben Laurie (CHATS)]
3904 *) Remote buffer overflow in SSL2 protocol - an attacker could
3905 supply an oversized client master key. (CVE-2002-0656)
3906 [Ben Laurie (CHATS)]
3908 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3910 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3911 encoded as NULL) with id-dsa-with-sha1.
3912 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3914 *) Check various X509_...() return values in apps/req.c.
3915 [Nils Larsch <nla@trustcenter.de>]
3917 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3918 an end-of-file condition would erronously be flagged, when the CRLF
3919 was just at the end of a processed block. The bug was discovered when
3920 processing data through a buffering memory BIO handing the data to a
3921 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3922 <ptsekov@syntrex.com> and Nedelcho Stanev.
3925 *) Implement a countermeasure against a vulnerability recently found
3926 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3927 before application data chunks to avoid the use of known IVs
3928 with data potentially chosen by the attacker.
3931 *) Fix length checks in ssl3_get_client_hello().
3934 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3935 to prevent ssl3_read_internal() from incorrectly assuming that
3936 ssl3_read_bytes() found application data while handshake
3937 processing was enabled when in fact s->s3->in_read_app_data was
3938 merely automatically cleared during the initial handshake.
3939 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3941 *) Fix object definitions for Private and Enterprise: they were not
3942 recognized in their shortname (=lowercase) representation. Extend
3943 obj_dat.pl to issue an error when using undefined keywords instead
3944 of silently ignoring the problem (Svenning Sorensen
3945 <sss@sss.dnsalias.net>).
3948 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3949 generators, i.e. generators other than 2 and 5. (Previously, the
3950 code did not properly initialise the 'add' and 'rem' values to
3951 BN_generate_prime().)
3953 In the new general case, we do not insist that 'generator' is
3954 actually a primitive root: This requirement is rather pointless;
3955 a generator of the order-q subgroup is just as good, if not
3959 *) Map new X509 verification errors to alerts. Discovered and submitted by
3960 Tom Wu <tom@arcot.com>.
3963 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3964 returning non-zero before the data has been completely received
3965 when using non-blocking I/O.
3966 [Bodo Moeller; problem pointed out by John Hughes]
3968 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3969 [Ben Laurie, Lutz Jaenicke]
3971 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3972 Yoram Zahavi <YoramZ@gilian.com>).
3975 *) Add information about CygWin 1.3 and on, and preserve proper
3976 configuration for the versions before that.
3977 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3979 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3980 check whether we deal with a copy of a session and do not delete from
3981 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3982 <izhar@checkpoint.com>.
3985 *) Do not store session data into the internal session cache, if it
3986 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3987 flag is set). Proposed by Aslam <aslam@funk.com>.
3990 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3994 *) [In 0.9.6d-engine release:]
3995 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3996 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3998 *) Add the configuration target linux-s390x.
3999 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4001 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4002 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4003 variable as an indication that a ClientHello message has been
4004 received. As the flag value will be lost between multiple
4005 invocations of ssl3_accept when using non-blocking I/O, the
4006 function may not be aware that a handshake has actually taken
4007 place, thus preventing a new session from being added to the
4010 To avoid this problem, we now set s->new_session to 2 instead of
4011 using a local variable.
4012 [Lutz Jaenicke, Bodo Moeller]
4014 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4015 if the SSL_R_LENGTH_MISMATCH error is detected.
4016 [Geoff Thorpe, Bodo Moeller]
4018 *) New 'shared_ldflag' column in Configure platform table.
4021 *) Fix EVP_CIPHER_mode macro.
4022 ["Dan S. Camper" <dan@bti.net>]
4024 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4025 type, we must throw them away by setting rr->length to 0.
4026 [D P Chang <dpc@qualys.com>]
4028 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4030 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4031 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4032 worked incorrectly for those cases where range = 10..._2 and
4033 3*range is two bits longer than range.)
4036 *) Only add signing time to PKCS7 structures if it is not already
4040 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4041 OBJ_ld_ce should be OBJ_id_ce.
4042 Also some ip-pda OIDs in crypto/objects/objects.txt were
4043 incorrect (cf. RFC 3039).
4044 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4046 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4047 returns early because it has nothing to do.
4048 [Andy Schneider <andy.schneider@bjss.co.uk>]
4050 *) [In 0.9.6c-engine release:]
4051 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4052 [Andy Schneider <andy.schneider@bjss.co.uk>]
4054 *) [In 0.9.6c-engine release:]
4055 Add support for Cryptographic Appliance's keyserver technology.
4056 (Use engine 'keyclient')
4057 [Cryptographic Appliances and Geoff Thorpe]
4059 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4060 is called via tools/c89.sh because arguments have to be
4061 rearranged (all '-L' options must appear before the first object
4063 [Richard Shapiro <rshapiro@abinitio.com>]
4065 *) [In 0.9.6c-engine release:]
4066 Add support for Broadcom crypto accelerator cards, backported
4068 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4070 *) [In 0.9.6c-engine release:]
4071 Add support for SureWare crypto accelerator cards from
4072 Baltimore Technologies. (Use engine 'sureware')
4073 [Baltimore Technologies and Mark Cox]
4075 *) [In 0.9.6c-engine release:]
4076 Add support for crypto accelerator cards from Accelerated
4077 Encryption Processing, www.aep.ie. (Use engine 'aep')
4078 [AEP Inc. and Mark Cox]
4080 *) Add a configuration entry for gcc on UnixWare.
4081 [Gary Benson <gbenson@redhat.com>]
4083 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4084 messages are stored in a single piece (fixed-length part and
4085 variable-length part combined) and fix various bugs found on the way.
4088 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4089 instead. BIO_gethostbyname() does not know what timeouts are
4090 appropriate, so entries would stay in cache even when they have
4092 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4094 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4095 faced with a pathologically small ClientHello fragment that does
4096 not contain client_version: Instead of aborting with an error,
4097 simply choose the highest available protocol version (i.e.,
4098 TLS 1.0 unless it is disabled). In practice, ClientHello
4099 messages are never sent like this, but this change gives us
4100 strictly correct behaviour at least for TLS.
4103 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4104 never resets s->method to s->ctx->method when called from within
4105 one of the SSL handshake functions.
4106 [Bodo Moeller; problem pointed out by Niko Baric]
4108 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4109 (sent using the client's version number) if client_version is
4110 smaller than the protocol version in use. Also change
4111 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4112 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4113 the client will at least see that alert.
4116 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4120 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4121 client receives HelloRequest while in a handshake.
4122 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4124 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4125 should end in 'break', not 'goto end' which circuments various
4126 cleanups done in state SSL_ST_OK. But session related stuff
4127 must be disabled for SSL_ST_OK in the case that we just sent a
4130 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4131 before just sending a HelloRequest.
4132 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4134 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4135 reveal whether illegal block cipher padding was found or a MAC
4136 verification error occured. (Neither SSLerr() codes nor alerts
4137 are directly visible to potential attackers, but the information
4138 may leak via logfiles.)
4140 Similar changes are not required for the SSL 2.0 implementation
4141 because the number of padding bytes is sent in clear for SSL 2.0,
4142 and the extra bytes are just ignored. However ssl/s2_pkt.c
4143 failed to verify that the purported number of padding bytes is in
4147 *) Add OpenUNIX-8 support including shared libraries
4148 (Boyd Lynn Gerber <gerberb@zenez.com>).
4151 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4152 'wristwatch attack' using huge encoding parameters (cf.
4153 James H. Manger's CRYPTO 2001 paper). Note that the
4154 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4155 encoding parameters and hence was not vulnerable.
4158 *) BN_sqr() bug fix.
4159 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4161 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4162 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4163 followed by modular reduction.
4164 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4166 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4167 equivalent based on BN_pseudo_rand() instead of BN_rand().
4170 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4171 This function was broken, as the check for a new client hello message
4172 to handle SGC did not allow these large messages.
4173 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4176 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4179 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4180 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4183 *) Rework the configuration and shared library support for Tru64 Unix.
4184 The configuration part makes use of modern compiler features and
4185 still retains old compiler behavior for those that run older versions
4186 of the OS. The shared library support part includes a variant that
4187 uses the RPATH feature, and is available through the special
4188 configuration target "alpha-cc-rpath", which will never be selected
4190 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4192 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4193 with the same message size as in ssl3_get_certificate_request().
4194 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4195 messages might inadvertently be reject as too long.
4196 [Petr Lampa <lampa@fee.vutbr.cz>]
4198 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4201 *) Modified SSL library such that the verify_callback that has been set
4202 specificly for an SSL object with SSL_set_verify() is actually being
4203 used. Before the change, a verify_callback set with this function was
4204 ignored and the verify_callback() set in the SSL_CTX at the time of
4205 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4206 to allow the necessary settings.
4209 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4210 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4211 done automatically (in contradiction to the requirements of the C
4212 standard). This made problems when used from OpenSSH.
4215 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4216 dh->length and always used
4218 BN_rand_range(priv_key, dh->p).
4220 BN_rand_range() is not necessary for Diffie-Hellman, and this
4221 specific range makes Diffie-Hellman unnecessarily inefficient if
4222 dh->length (recommended exponent length) is much smaller than the
4223 length of dh->p. We could use BN_rand_range() if the order of
4224 the subgroup was stored in the DH structure, but we only have
4229 BN_rand(priv_key, l, ...)
4231 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4237 RSA_eay_public_encrypt
4238 RSA_eay_private_decrypt
4239 RSA_eay_private_encrypt (signing)
4240 RSA_eay_public_decrypt (signature verification)
4242 (default implementations for RSA_public_encrypt,
4243 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4244 always reject numbers >= n.
4247 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4248 to synchronize access to 'locking_thread'. This is necessary on
4249 systems where access to 'locking_thread' (an 'unsigned long'
4250 variable) is not atomic.
4253 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4254 *before* setting the 'crypto_lock_rand' flag. The previous code had
4255 a race condition if 0 is a valid thread ID.
4256 [Travis Vitek <vitek@roguewave.com>]
4258 *) Add support for shared libraries under Irix.
4259 [Albert Chin-A-Young <china@thewrittenword.com>]
4261 *) Add configuration option to build on Linux on both big-endian and
4263 [Ralf Baechle <ralf@uni-koblenz.de>]
4265 *) Add the possibility to create shared libraries on HP-UX.
4268 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4270 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4271 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4272 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4273 PRNG state recovery was possible based on the output of
4274 one PRNG request appropriately sized to gain knowledge on
4275 'md' followed by enough consecutive 1-byte PRNG requests
4276 to traverse all of 'state'.
4278 1. When updating 'md_local' (the current thread's copy of 'md')
4279 during PRNG output generation, hash all of the previous
4280 'md_local' value, not just the half used for PRNG output.
4282 2. Make the number of bytes from 'state' included into the hash
4283 independent from the number of PRNG bytes requested.
4285 The first measure alone would be sufficient to avoid
4286 Markku-Juhani's attack. (Actually it had never occurred
4287 to me that the half of 'md_local' used for chaining was the
4288 half from which PRNG output bytes were taken -- I had always
4289 assumed that the secret half would be used.) The second
4290 measure makes sure that additional data from 'state' is never
4291 mixed into 'md_local' in small portions; this heuristically
4292 further strengthens the PRNG.
4295 *) Fix crypto/bn/asm/mips3.s.
4298 *) When only the key is given to "enc", the IV is undefined. Print out
4299 an error message in this case.
4302 *) Handle special case when X509_NAME is empty in X509 printing routines.
4305 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4306 positive and less than q.
4309 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4310 used: it isn't thread safe and the add_lock_callback should handle
4312 [Paul Rose <Paul.Rose@bridge.com>]
4314 *) Verify that incoming data obeys the block size in
4315 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4319 [Ulf Möller, Bodo Möller]
4321 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4322 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4323 when fixing the server behaviour for backwards-compatible 'client
4324 hello' messages. (Note that the attack is impractical against
4325 SSL 3.0 and TLS 1.0 anyway because length and version checking
4326 means that the probability of guessing a valid ciphertext is
4327 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4330 Before 0.9.5, the countermeasure (hide the error by generating a
4331 random 'decryption result') did not work properly because
4332 ERR_clear_error() was missing, meaning that SSL_get_error() would
4333 detect the supposedly ignored error.
4335 Both problems are now fixed.
4338 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4339 (previously it was 1024).
4342 *) Fix for compatibility mode trust settings: ignore trust settings
4343 unless some valid trust or reject settings are present.
4346 *) Fix for blowfish EVP: its a variable length cipher.
4349 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4350 parameters in DSA public key structures and return an error in the
4351 DSA routines if parameters are absent.
4354 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4355 in the current directory if neither $RANDFILE nor $HOME was set.
4356 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4357 caused some confusion to Windows users who haven't defined $HOME.
4358 Thus RAND_file_name() is changed again: e_os.h can define a
4359 DEFAULT_HOME, which will be used if $HOME is not set.
4360 For Windows, we use "C:"; on other platforms, we still require
4361 environment variables.
4363 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4364 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4365 having multiple threads call RAND_poll() concurrently.
4368 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4369 combination of a flag and a thread ID variable.
4370 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4371 flag), *other* threads can enter ssleay_add_bytes without obeying
4372 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4373 that they do not hold after the first thread unsets add_do_not_lock).
4376 *) Change bctest again: '-x' expressions are not available in all
4380 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4382 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4383 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4385 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4386 the default extension for executables, if any. Also, make the perl
4387 scripts that use symlink() to test if it really exists and use "cp"
4388 if it doesn't. All this made OpenSSL compilable and installable in
4392 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4393 If SEQUENCE is length is indefinite just set c->slen to the total
4394 amount of data available.
4395 [Steve Henson, reported by shige@FreeBSD.org]
4396 [This change does not apply to 0.9.7.]
4398 *) Change bctest to avoid here-documents inside command substitution
4399 (workaround for FreeBSD /bin/sh bug).
4400 For compatibility with Ultrix, avoid shell functions (introduced
4401 in the bctest version that searches along $PATH).
4404 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4405 with des_encrypt() defined on some operating systems, like Solaris
4409 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4410 On the Importance of Eliminating Errors in Cryptographic
4411 Computations, J. Cryptology 14 (2001) 2, 101-119,
4412 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4415 *) MIPS assembler BIGNUM division bug fix.
4418 *) Disabled incorrect Alpha assembler code.
4421 *) Fix PKCS#7 decode routines so they correctly update the length
4422 after reading an EOC for the EXPLICIT tag.
4424 [This change does not apply to 0.9.7.]
4426 *) Fix bug in PKCS#12 key generation routines. This was triggered
4427 if a 3DES key was generated with a 0 initial byte. Include
4428 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4429 (but broken) behaviour.
4432 *) Enhance bctest to search for a working bc along $PATH and print
4434 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4436 *) Fix memory leaks in err.c: free err_data string if necessary;
4437 don't write to the wrong index in ERR_set_error_data.
4440 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4444 *) Replace rdtsc with _emit statements for VC++ version 5.
4445 [Jeremy Cooper <jeremy@baymoo.org>]
4447 *) Make it possible to reuse SSLv2 sessions.
4450 *) In copy_email() check for >= 0 as a return value for
4451 X509_NAME_get_index_by_NID() since 0 is a valid index.
4452 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4454 *) Avoid coredump with unsupported or invalid public keys by checking if
4455 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4456 PKCS7_verify() fails with non detached data.
4459 *) Don't use getenv in library functions when run as setuid/setgid.
4460 New function OPENSSL_issetugid().
4463 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4464 due to incorrect handling of multi-threading:
4466 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4468 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4470 3. Count how many times MemCheck_off() has been called so that
4471 nested use can be treated correctly. This also avoids
4472 inband-signalling in the previous code (which relied on the
4473 assumption that thread ID 0 is impossible).
4476 *) Add "-rand" option also to s_client and s_server.
4479 *) Fix CPU detection on Irix 6.x.
4480 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4481 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4483 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4486 [This change does not apply to 0.9.7.]
4488 *) Use the cached encoding of an X509_NAME structure rather than
4489 copying it. This is apparently the reason for the libsafe "errors"
4490 but the code is actually correct.
4493 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4494 Bleichenbacher's DSA attack.
4495 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4496 to be set and top=0 forces the highest bit to be set; top=-1 is new
4497 and leaves the highest bit random.
4498 [Ulf Moeller, Bodo Moeller]
4500 *) In the NCONF_...-based implementations for CONF_... queries
4501 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4502 a temporary CONF structure with the data component set to NULL
4503 (which gives segmentation faults in lh_retrieve).
4504 Instead, use NULL for the CONF pointer in CONF_get_string and
4505 CONF_get_number (which may use environment variables) and directly
4506 return NULL from CONF_get_section.
4509 *) Fix potential buffer overrun for EBCDIC.
4512 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4513 keyUsage if basicConstraints absent for a CA.
4516 *) Make SMIME_write_PKCS7() write mail header values with a format that
4517 is more generally accepted (no spaces before the semicolon), since
4518 some programs can't parse those values properly otherwise. Also make
4519 sure BIO's that break lines after each write do not create invalid
4523 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4524 macros previously used would not encode an empty SEQUENCE OF
4525 and break the signature.
4527 [This change does not apply to 0.9.7.]
4529 *) Zero the premaster secret after deriving the master secret in
4533 *) Add some EVP_add_digest_alias registrations (as found in
4534 OpenSSL_add_all_digests()) to SSL_library_init()
4535 aka OpenSSL_add_ssl_algorithms(). This provides improved
4536 compatibility with peers using X.509 certificates
4537 with unconventional AlgorithmIdentifier OIDs.
4540 *) Fix for Irix with NO_ASM.
4541 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4543 *) ./config script fixes.
4544 [Ulf Moeller, Richard Levitte]
4546 *) Fix 'openssl passwd -1'.
4549 *) Change PKCS12_key_gen_asc() so it can cope with non null
4550 terminated strings whose length is passed in the passlen
4551 parameter, for example from PEM callbacks. This was done
4552 by adding an extra length parameter to asc2uni().
4553 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4555 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4556 call failed, free the DSA structure.
4559 *) Fix to uni2asc() to cope with zero length Unicode strings.
4560 These are present in some PKCS#12 files.
4563 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4564 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4565 when writing a 32767 byte record.
4566 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4568 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4569 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4571 (RSA objects have a reference count access to which is protected
4572 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4573 so they are meant to be shared between threads.)
4574 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4575 "Reddie, Steven" <Steven.Reddie@ca.com>]
4577 *) Fix a deadlock in CRYPTO_mem_leaks().
4580 *) Use better test patterns in bntest.
4583 *) rand_win.c fix for Borland C.
4586 *) BN_rshift bugfix for n == 0.
4589 *) Add a 'bctest' script that checks for some known 'bc' bugs
4590 so that 'make test' does not abort just because 'bc' is broken.
4593 *) Store verify_result within SSL_SESSION also for client side to
4594 avoid potential security hole. (Re-used sessions on the client side
4595 always resulted in verify_result==X509_V_OK, not using the original
4596 result of the server certificate verification.)
4599 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4600 SSL3_RT_APPLICATION_DATA, return 0.
4601 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4605 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4606 releases, have been re-implemented by renaming the previous
4607 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4608 and ssl3_read_internal, respectively, and adding 'peek' parameters
4609 to them. The new ssl[23]_{read,peek} functions are calls to
4610 ssl[23]_read_internal with the 'peek' flag set appropriately.
4611 A 'peek' parameter has also been added to ssl3_read_bytes, which
4612 does the actual work for ssl3_read_internal.
4615 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4616 the method-specific "init()" handler. Also clean up ex_data after
4617 calling the method-specific "finish()" handler. Previously, this was
4618 happening the other way round.
4621 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4622 The previous value, 12, was not always sufficient for BN_mod_exp().
4625 *) Make sure that shared libraries get the internal name engine with
4626 the full version number and not just 0. This should mark the
4627 shared libraries as not backward compatible. Of course, this should
4628 be changed again when we can guarantee backward binary compatibility.
4631 *) Fix typo in get_cert_by_subject() in by_dir.c
4632 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4634 *) Rework the system to generate shared libraries:
4636 - Make note of the expected extension for the shared libraries and
4637 if there is a need for symbolic links from for example libcrypto.so.0
4638 to libcrypto.so.0.9.7. There is extended info in Configure for
4641 - Make as few rebuilds of the shared libraries as possible.
4643 - Still avoid linking the OpenSSL programs with the shared libraries.
4645 - When installing, install the shared libraries separately from the
4649 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4651 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4652 and not in SSL_clear because the latter is also used by the
4653 accept/connect functions; previously, the settings made by
4654 SSL_set_read_ahead would be lost during the handshake.
4655 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4657 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4658 Previously, it would create entries for disableed algorithms no
4662 *) Added several new manual pages for SSL_* function.
4665 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4667 *) In ssl23_get_client_hello, generate an error message when faced
4668 with an initial SSL 3.0/TLS record that is too small to contain the
4669 first two bytes of the ClientHello message, i.e. client_version.
4670 (Note that this is a pathologic case that probably has never happened
4671 in real life.) The previous approach was to use the version number
4672 from the record header as a substitute; but our protocol choice
4673 should not depend on that one because it is not authenticated
4674 by the Finished messages.
4677 *) More robust randomness gathering functions for Windows.
4678 [Jeffrey Altman <jaltman@columbia.edu>]
4680 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4681 not set then we don't setup the error code for issuer check errors
4682 to avoid possibly overwriting other errors which the callback does
4683 handle. If an application does set the flag then we assume it knows
4684 what it is doing and can handle the new informational codes
4688 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4689 a general "ANY" type, as such it should be able to decode anything
4690 including tagged types. However it didn't check the class so it would
4691 wrongly interpret tagged types in the same way as their universal
4692 counterpart and unknown types were just rejected. Changed so that the
4693 tagged and unknown types are handled in the same way as a SEQUENCE:
4694 that is the encoding is stored intact. There is also a new type
4695 "V_ASN1_OTHER" which is used when the class is not universal, in this
4696 case we have no idea what the actual type is so we just lump them all
4700 *) On VMS, stdout may very well lead to a file that is written to
4701 in a record-oriented fashion. That means that every write() will
4702 write a separate record, which will be read separately by the
4703 programs trying to read from it. This can be very confusing.
4705 The solution is to put a BIO filter in the way that will buffer
4706 text until a linefeed is reached, and then write everything a
4707 line at a time, so every record written will be an actual line,
4708 not chunks of lines and not (usually doesn't happen, but I've
4709 seen it once) several lines in one record. BIO_f_linebuffer() is
4712 Currently, it's a VMS-only method, because that's where it has
4713 been tested well enough.
4716 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4717 it can return incorrect results.
4718 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4719 but it was in 0.9.6-beta[12].)
4722 *) Disable the check for content being present when verifying detached
4723 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4724 include zero length content when signing messages.
4727 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4728 BIO_ctrl (for BIO pairs).
4731 *) Add DSO method for VMS.
4734 *) Bug fix: Montgomery multiplication could produce results with the
4738 *) Add RPM specification openssl.spec and modify it to build three
4739 packages. The default package contains applications, application
4740 documentation and run-time libraries. The devel package contains
4741 include files, static libraries and function documentation. The
4742 doc package contains the contents of the doc directory. The original
4743 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4746 *) Add a large number of documentation files for many SSL routines.
4747 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4749 *) Add a configuration entry for Sony News 4.
4750 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4752 *) Don't set the two most significant bits to one when generating a
4753 random number < q in the DSA library.
4756 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4757 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4758 the underlying transport is blocking) if a handshake took place.
4759 (The default behaviour is needed by applications such as s_client
4760 and s_server that use select() to determine when to use SSL_read;
4761 but for applications that know in advance when to expect data, it
4762 just makes things more complicated.)
4765 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4769 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4770 work better on such systems.
4771 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4773 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4774 Update PKCS12_parse() so it copies the friendlyName and the
4775 keyid to the certificates aux info.
4778 *) Fix bug in PKCS7_verify() which caused an infinite loop
4779 if there was more than one signature.
4780 [Sven Uszpelkat <su@celocom.de>]
4782 *) Major change in util/mkdef.pl to include extra information
4783 about each symbol, as well as presentig variables as well
4784 as functions. This change means that there's n more need
4785 to rebuild the .num files when some algorithms are excluded.
4788 *) Allow the verify time to be set by an application,
4789 rather than always using the current time.
4792 *) Phase 2 verify code reorganisation. The certificate
4793 verify code now looks up an issuer certificate by a
4794 number of criteria: subject name, authority key id
4795 and key usage. It also verifies self signed certificates
4796 by the same criteria. The main comparison function is
4797 X509_check_issued() which performs these checks.
4799 Lot of changes were necessary in order to support this
4800 without completely rewriting the lookup code.
4802 Authority and subject key identifier are now cached.
4804 The LHASH 'certs' is X509_STORE has now been replaced
4805 by a STACK_OF(X509_OBJECT). This is mainly because an
4806 LHASH can't store or retrieve multiple objects with
4807 the same hash value.
4809 As a result various functions (which were all internal
4810 use only) have changed to handle the new X509_STORE
4811 structure. This will break anything that messed round
4812 with X509_STORE internally.
4814 The functions X509_STORE_add_cert() now checks for an
4815 exact match, rather than just subject name.
4817 The X509_STORE API doesn't directly support the retrieval
4818 of multiple certificates matching a given criteria, however
4819 this can be worked round by performing a lookup first
4820 (which will fill the cache with candidate certificates)
4821 and then examining the cache for matches. This is probably
4822 the best we can do without throwing out X509_LOOKUP
4823 entirely (maybe later...).
4825 The X509_VERIFY_CTX structure has been enhanced considerably.
4827 All certificate lookup operations now go via a get_issuer()
4828 callback. Although this currently uses an X509_STORE it
4829 can be replaced by custom lookups. This is a simple way
4830 to bypass the X509_STORE hackery necessary to make this
4831 work and makes it possible to use more efficient techniques
4832 in future. A very simple version which uses a simple
4833 STACK for its trusted certificate store is also provided
4834 using X509_STORE_CTX_trusted_stack().
4836 The verify_cb() and verify() callbacks now have equivalents
4837 in the X509_STORE_CTX structure.
4839 X509_STORE_CTX also has a 'flags' field which can be used
4840 to customise the verify behaviour.
4843 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4844 excludes S/MIME capabilities.
4847 *) When a certificate request is read in keep a copy of the
4848 original encoding of the signed data and use it when outputing
4849 again. Signatures then use the original encoding rather than
4850 a decoded, encoded version which may cause problems if the
4851 request is improperly encoded.
4854 *) For consistency with other BIO_puts implementations, call
4855 buffer_write(b, ...) directly in buffer_puts instead of calling
4858 In BIO_puts, increment b->num_write as in BIO_write.
4859 [Peter.Sylvester@EdelWeb.fr]
4861 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4862 BN_zero, we may not return a BIGNUM with an array consisting of
4866 *) Avoid calling abort() from within the library when problems are
4867 detected, except if preprocessor symbols have been defined
4868 (such as REF_CHECK, BN_DEBUG etc.).
4871 *) New openssl application 'rsautl'. This utility can be
4872 used for low level RSA operations. DER public key
4873 BIO/fp routines also added.
4876 *) New Configure entry and patches for compiling on QNX 4.
4877 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4879 *) A demo state-machine implementation was sponsored by
4880 Nuron (http://www.nuron.com/) and is now available in
4881 demos/state_machine.
4884 *) New options added to the 'dgst' utility for signature
4885 generation and verification.
4888 *) Unrecognized PKCS#7 content types are now handled via a
4889 catch all ASN1_TYPE structure. This allows unsupported
4890 types to be stored as a "blob" and an application can
4891 encode and decode it manually.
4894 *) Fix various signed/unsigned issues to make a_strex.c
4896 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4898 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4899 length if passed a buffer. ASN1_INTEGER_to_BN failed
4900 if passed a NULL BN and its argument was negative.
4901 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4903 *) Modification to PKCS#7 encoding routines to output definite
4904 length encoding. Since currently the whole structures are in
4905 memory there's not real point in using indefinite length
4906 constructed encoding. However if OpenSSL is compiled with
4907 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4910 *) Added BIO_vprintf() and BIO_vsnprintf().
4913 *) Added more prefixes to parse for in the the strings written
4914 through a logging bio, to cover all the levels that are available
4915 through syslog. The prefixes are now:
4917 PANIC, EMERG, EMR => LOG_EMERG
4918 ALERT, ALR => LOG_ALERT
4919 CRIT, CRI => LOG_CRIT
4920 ERROR, ERR => LOG_ERR
4921 WARNING, WARN, WAR => LOG_WARNING
4922 NOTICE, NOTE, NOT => LOG_NOTICE
4923 INFO, INF => LOG_INFO
4924 DEBUG, DBG => LOG_DEBUG
4926 and as before, if none of those prefixes are present at the
4927 beginning of the string, LOG_ERR is chosen.
4929 On Win32, the LOG_* levels are mapped according to this:
4931 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4932 LOG_WARNING => EVENTLOG_WARNING_TYPE
4933 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4937 *) Made it possible to reconfigure with just the configuration
4938 argument "reconf" or "reconfigure". The command line arguments
4939 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4940 and are retrieved from there when reconfiguring.
4944 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4946 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4949 *) The obj_dat.pl script was messing up the sorting of object
4950 names. The reason was that it compared the quoted version
4951 of strings as a result "OCSP" > "OCSP Signing" because
4952 " > SPACE. Changed script to store unquoted versions of
4953 names and add quotes on output. It was also omitting some
4954 names from the lookup table if they were given a default
4955 value (that is if SN is missing it is given the same
4956 value as LN and vice versa), these are now added on the
4957 grounds that if an object has a name we should be able to
4958 look it up. Finally added warning output when duplicate
4959 short or long names are found.
4962 *) Changes needed for Tandem NSK.
4963 [Scott Uroff <scott@xypro.com>]
4965 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4966 RSA_padding_check_SSLv23(), special padding was never detected
4967 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4968 version rollback attacks was not effective.
4970 In s23_clnt.c, don't use special rollback-attack detection padding
4971 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4972 client; similarly, in s23_srvr.c, don't do the rollback check if
4973 SSL 2.0 is the only protocol enabled in the server.
4976 *) Make it possible to get hexdumps of unprintable data with 'openssl
4977 asn1parse'. By implication, the functions ASN1_parse_dump() and
4978 BIO_dump_indent() are added.
4981 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4982 these print out strings and name structures based on various
4983 flags including RFC2253 support and proper handling of
4984 multibyte characters. Added options to the 'x509' utility
4985 to allow the various flags to be set.
4988 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4989 Also change the functions X509_cmp_current_time() and
4990 X509_gmtime_adj() work with an ASN1_TIME structure,
4991 this will enable certificates using GeneralizedTime in validity
4992 dates to be checked.
4995 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4996 negative public key encodings) on by default,
4997 NO_NEG_PUBKEY_BUG can be set to disable it.
5000 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5001 content octets. An i2c_ASN1_OBJECT is unnecessary because
5002 the encoding can be trivially obtained from the structure.
5005 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5006 not read locks (CRYPTO_r_[un]lock).
5009 *) A first attempt at creating official support for shared
5010 libraries through configuration. I've kept it so the
5011 default is static libraries only, and the OpenSSL programs
5012 are always statically linked for now, but there are
5013 preparations for dynamic linking in place.
5014 This has been tested on Linux and Tru64.
5017 *) Randomness polling function for Win9x, as described in:
5018 Peter Gutmann, Software Generation of Practically Strong
5022 *) Fix so PRNG is seeded in req if using an already existing
5026 *) New options to smime application. -inform and -outform
5027 allow alternative formats for the S/MIME message including
5028 PEM and DER. The -content option allows the content to be
5029 specified separately. This should allow things like Netscape
5030 form signing output easier to verify.
5033 *) Fix the ASN1 encoding of tags using the 'long form'.
5036 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5037 STRING types. These convert content octets to and from the
5038 underlying type. The actual tag and length octets are
5039 already assumed to have been read in and checked. These
5040 are needed because all other string types have virtually
5041 identical handling apart from the tag. By having versions
5042 of the ASN1 functions that just operate on content octets
5043 IMPLICIT tagging can be handled properly. It also allows
5044 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5045 and ASN1_INTEGER are identical apart from the tag.
5048 *) Change the handling of OID objects as follows:
5050 - New object identifiers are inserted in objects.txt, following
5051 the syntax given in objects.README.
5052 - objects.pl is used to process obj_mac.num and create a new
5054 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5057 This is currently kind of a hack, and the perl code in objects.pl
5058 isn't very elegant, but it works as I intended. The simplest way
5059 to check that it worked correctly is to look in obj_dat.h and
5060 check the array nid_objs and make sure the objects haven't moved
5061 around (this is important!). Additions are OK, as well as
5062 consistent name changes.
5065 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5068 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5069 The given file adds to whatever has already been seeded into the
5070 random pool through the RANDFILE configuration file option or
5071 environment variable, or the default random state file.
5074 *) mkstack.pl now sorts each macro group into lexical order.
5075 Previously the output order depended on the order the files
5076 appeared in the directory, resulting in needless rewriting
5080 *) Patches to make OpenSSL compile under Win32 again. Mostly
5081 work arounds for the VC++ problem that it treats func() as
5082 func(void). Also stripped out the parts of mkdef.pl that
5083 added extra typesafe functions: these no longer exist.
5086 *) Reorganisation of the stack code. The macros are now all
5087 collected in safestack.h . Each macro is defined in terms of
5088 a "stack macro" of the form SKM_<name>(type, a, b). The
5089 DEBUG_SAFESTACK is now handled in terms of function casts,
5090 this has the advantage of retaining type safety without the
5091 use of additional functions. If DEBUG_SAFESTACK is not defined
5092 then the non typesafe macros are used instead. Also modified the
5093 mkstack.pl script to handle the new form. Needs testing to see
5094 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5095 the default if no major problems. Similar behaviour for ASN1_SET_OF
5096 and PKCS12_STACK_OF.
5099 *) When some versions of IIS use the 'NET' form of private key the
5100 key derivation algorithm is different. Normally MD5(password) is
5101 used as a 128 bit RC4 key. In the modified case
5102 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5103 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5104 as the old Netscape_RSA functions except they have an additional
5105 'sgckey' parameter which uses the modified algorithm. Also added
5106 an -sgckey command line option to the rsa utility. Thanks to
5107 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5108 algorithm to openssl-dev.
5111 *) The evp_local.h macros were using 'c.##kname' which resulted in
5112 invalid expansion on some systems (SCO 5.0.5 for example).
5113 Corrected to 'c.kname'.
5114 [Phillip Porch <root@theporch.com>]
5116 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5117 a STACK of email addresses from a certificate or request, these look
5118 in the subject name and the subject alternative name extensions and
5119 omit any duplicate addresses.
5122 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5123 This makes DSA verification about 2 % faster.
5126 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5127 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5128 plus overhead for 1024 bit moduli).
5129 This makes exponentiations about 0.5 % faster for 1024 bit
5130 exponents (as measured by "openssl speed rsa2048").
5133 *) Rename memory handling macros to avoid conflicts with other
5135 Malloc => OPENSSL_malloc
5136 Malloc_locked => OPENSSL_malloc_locked
5137 Realloc => OPENSSL_realloc
5138 Free => OPENSSL_free
5141 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5142 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5145 *) CygWin32 support.
5146 [John Jarvie <jjarvie@newsguy.com>]
5148 *) The type-safe stack code has been rejigged. It is now only compiled
5149 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5150 by default all type-specific stack functions are "#define"d back to
5151 standard stack functions. This results in more streamlined output
5152 but retains the type-safety checking possibilities of the original
5156 *) The STACK code has been cleaned up, and certain type declarations
5157 that didn't make a lot of sense have been brought in line. This has
5158 also involved a cleanup of sorts in safestack.h to more correctly
5159 map type-safe stack functions onto their plain stack counterparts.
5160 This work has also resulted in a variety of "const"ifications of
5161 lots of the code, especially "_cmp" operations which should normally
5162 be prototyped with "const" parameters anyway.
5165 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5166 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5167 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5168 where all of 'md' is used each time the PRNG is used, but 'state'
5169 is used only indexed by a cyclic counter. As entropy may not be
5170 well distributed from the beginning, 'md' is important as a
5171 chaining variable. However, the output function chains only half
5172 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5173 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5174 in all of 'state' being rewritten, with the new values depending
5175 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5178 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5179 the handshake is continued after ssl_verify_cert_chain();
5180 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5181 can lead to 'unexplainable' connection aborts later.
5182 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5184 *) Major EVP API cipher revision.
5185 Add hooks for extra EVP features. This allows various cipher
5186 parameters to be set in the EVP interface. Support added for variable
5187 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5188 setting of RC2 and RC5 parameters.
5190 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5193 Remove lots of duplicated code from the EVP library. For example *every*
5194 cipher init() function handles the 'iv' in the same way according to the
5195 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5196 for CFB and OFB modes they zero ctx->num.
5198 New functionality allows removal of S/MIME code RC2 hack.
5200 Most of the routines have the same form and so can be declared in terms
5203 By shifting this to the top level EVP_CipherInit() it can be removed from
5204 all individual ciphers. If the cipher wants to handle IVs or keys
5205 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5208 Change lots of functions like EVP_EncryptUpdate() to now return a
5209 value: although software versions of the algorithms cannot fail
5210 any installed hardware versions can.
5213 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5214 this option is set, tolerate broken clients that send the negotiated
5215 protocol version number instead of the requested protocol version
5219 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5220 i.e. non-zero for export ciphersuites, zero otherwise.
5221 Previous versions had this flag inverted, inconsistent with
5222 rsa_tmp_cb (..._TMP_RSA_CB).
5223 [Bodo Moeller; problem reported by Amit Chopra]
5225 *) Add missing DSA library text string. Work around for some IIS
5226 key files with invalid SEQUENCE encoding.
5229 *) Add a document (doc/standards.txt) that list all kinds of standards
5230 and so on that are implemented in OpenSSL.
5233 *) Enhance c_rehash script. Old version would mishandle certificates
5234 with the same subject name hash and wouldn't handle CRLs at all.
5235 Added -fingerprint option to crl utility, to support new c_rehash
5239 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5242 *) Fix for SSL server purpose checking. Server checking was
5243 rejecting certificates which had extended key usage present
5244 but no ssl client purpose.
5245 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5247 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5248 is a little unclear about how a blank password is handled.
5249 Since the password in encoded as a BMPString with terminating
5250 double NULL a zero length password would end up as just the
5251 double NULL. However no password at all is different and is
5252 handled differently in the PKCS#12 key generation code. NS
5253 treats a blank password as zero length. MSIE treats it as no
5254 password on export: but it will try both on import. We now do
5255 the same: PKCS12_parse() tries zero length and no password if
5256 the password is set to "" or NULL (NULL is now a valid password:
5257 it wasn't before) as does the pkcs12 application.
5260 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5261 perror when PEM_read_bio_X509_REQ fails, the error message must
5262 be obtained from the error queue.
5265 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5266 it in ERR_remove_state if appropriate, and change ERR_get_state
5267 accordingly to avoid race conditions (this is necessary because
5268 thread_hash is no longer constant once set).
5271 *) Bugfix for linux-elf makefile.one.
5274 *) RSA_get_default_method() will now cause a default
5275 RSA_METHOD to be chosen if one doesn't exist already.
5276 Previously this was only set during a call to RSA_new()
5277 or RSA_new_method(NULL) meaning it was possible for
5278 RSA_get_default_method() to return NULL.
5281 *) Added native name translation to the existing DSO code
5282 that will convert (if the flag to do so is set) filenames
5283 that are sufficiently small and have no path information
5284 into a canonical native form. Eg. "blah" converted to
5285 "libblah.so" or "blah.dll" etc.
5288 *) New function ERR_error_string_n(e, buf, len) which is like
5289 ERR_error_string(e, buf), but writes at most 'len' bytes
5290 including the 0 terminator. For ERR_error_string_n, 'buf'
5292 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5294 *) CONF library reworked to become more general. A new CONF
5295 configuration file reader "class" is implemented as well as a
5296 new functions (NCONF_*, for "New CONF") to handle it. The now
5297 old CONF_* functions are still there, but are reimplemented to
5298 work in terms of the new functions. Also, a set of functions
5299 to handle the internal storage of the configuration data is
5300 provided to make it easier to write new configuration file
5301 reader "classes" (I can definitely see something reading a
5302 configuration file in XML format, for example), called _CONF_*,
5303 or "the configuration storage API"...
5305 The new configuration file reading functions are:
5307 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5308 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5310 NCONF_default, NCONF_WIN32
5312 NCONF_dump_fp, NCONF_dump_bio
5314 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5315 NCONF_new creates a new CONF object. This works in the same way
5316 as other interfaces in OpenSSL, like the BIO interface.
5317 NCONF_dump_* dump the internal storage of the configuration file,
5318 which is useful for debugging. All other functions take the same
5319 arguments as the old CONF_* functions wth the exception of the
5320 first that must be a `CONF *' instead of a `LHASH *'.
5322 To make it easer to use the new classes with the old CONF_* functions,
5323 the function CONF_set_default_method is provided.
5326 *) Add '-tls1' option to 'openssl ciphers', which was already
5327 mentioned in the documentation but had not been implemented.
5328 (This option is not yet really useful because even the additional
5329 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5332 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5333 OpenSSL-based applications) load shared libraries and bind to
5334 them in a portable way.
5335 [Geoff Thorpe, with contributions from Richard Levitte]
5337 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5339 *) Make sure _lrotl and _lrotr are only used with MSVC.
5341 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5342 (the default implementation of RAND_status).
5344 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5345 to '-clrext' (= clear extensions), as intended and documented.
5346 [Bodo Moeller; inconsistency pointed out by Michael Attili
5349 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5350 was larger than the MD block size.
5351 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5353 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5354 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5355 using the passed key: if the passed key was a private key the result
5356 of X509_print(), for example, would be to print out all the private key
5360 *) des_quad_cksum() byte order bug fix.
5361 [Ulf Möller, using the problem description in krb4-0.9.7, where
5362 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5364 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5366 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5368 *) For easily testing in shell scripts whether some command
5369 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5370 returns with exit code 0 iff no command of the given name is available.
5371 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5372 the output goes to stdout and nothing is printed to stderr.
5373 Additional arguments are always ignored.
5375 Since for each cipher there is a command of the same name,
5376 the 'no-cipher' compilation switches can be tested this way.
5378 ('openssl no-XXX' is not able to detect pseudo-commands such
5379 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5382 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5385 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5386 is set; it will be thrown away anyway because each handshake creates
5388 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5389 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5390 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5391 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5394 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5395 'Q' and 'R' lose their special meanings (quit/renegotiate).
5396 This is part of what -quiet does; unlike -quiet, -ign_eof
5397 does not suppress any output.
5400 *) Add compatibility options to the purpose and trust code. The
5401 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5402 accepts a certificate or CA, this was the previous behaviour,
5403 with all the associated security issues.
5405 X509_TRUST_COMPAT is the old trust behaviour: only and
5406 automatically trust self signed roots in certificate store. A
5407 new trust setting X509_TRUST_DEFAULT is used to specify that
5408 a purpose has no associated trust setting and it should instead
5409 use the value in the default purpose.
5412 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5413 and fix a memory leak.
5416 *) In util/mkerr.pl (which implements 'make errors'), preserve
5417 reason strings from the previous version of the .c file, as
5418 the default to have only downcase letters (and digits) in
5419 automatically generated reasons codes is not always appropriate.
5422 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5423 using strerror. Previously, ERR_reason_error_string() returned
5424 library names as reason strings for SYSerr; but SYSerr is a special
5425 case where small numbers are errno values, not library numbers.
5428 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5429 converts DSA parameters into DH parameters. (When creating parameters,
5430 DSA_generate_parameters is used.)
5433 *) Include 'length' (recommended exponent length) in C code generated
5434 by 'openssl dhparam -C'.
5437 *) The second argument to set_label in perlasm was already being used
5438 so couldn't be used as a "file scope" flag. Moved to third argument
5442 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5443 instead of RAND_bytes for encryption IVs and salts.
5446 *) Include RAND_status() into RAND_METHOD instead of implementing
5447 it only for md_rand.c Otherwise replacing the PRNG by calling
5448 RAND_set_rand_method would be impossible.
5451 *) Don't let DSA_generate_key() enter an infinite loop if the random
5452 number generation fails.
5455 *) New 'rand' application for creating pseudo-random output.
5458 *) Added configuration support for Linux/IA64
5459 [Rolf Haberrecker <rolf@suse.de>]
5461 *) Assembler module support for Mingw32.
5464 *) Shared library support for HPUX (in shlib/).
5465 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5467 *) Shared library support for Solaris gcc.
5468 [Lutz Behnke <behnke@trustcenter.de>]
5470 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5472 *) PKCS7_encrypt() was adding text MIME headers twice because they
5473 were added manually and by SMIME_crlf_copy().
5476 *) In bntest.c don't call BN_rand with zero bits argument.
5477 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5479 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5480 case was implemented. This caused BN_div_recp() to fail occasionally.
5483 *) Add an optional second argument to the set_label() in the perl
5484 assembly language builder. If this argument exists and is set
5485 to 1 it signals that the assembler should use a symbol whose
5486 scope is the entire file, not just the current function. This
5487 is needed with MASM which uses the format label:: for this scope.
5488 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5490 *) Change the ASN1 types so they are typedefs by default. Before
5491 almost all types were #define'd to ASN1_STRING which was causing
5492 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5496 *) Change names of new functions to the new get1/get0 naming
5497 convention: After 'get1', the caller owns a reference count
5498 and has to call ..._free; 'get0' returns a pointer to some
5499 data structure without incrementing reference counters.
5500 (Some of the existing 'get' functions increment a reference
5501 counter, some don't.)
5502 Similarly, 'set1' and 'add1' functions increase reference
5503 counters or duplicate objects.
5506 *) Allow for the possibility of temp RSA key generation failure:
5507 the code used to assume it always worked and crashed on failure.
5510 *) Fix potential buffer overrun problem in BIO_printf().
5511 [Ulf Möller, using public domain code by Patrick Powell; problem
5512 pointed out by David Sacerdote <das33@cornell.edu>]
5514 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5515 RAND_egd() and RAND_status(). In the command line application,
5516 the EGD socket can be specified like a seed file using RANDFILE
5520 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5521 Some CAs (e.g. Verisign) distribute certificates in this form.
5524 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5525 list to exclude them. This means that no special compilation option
5526 is needed to use anonymous DH: it just needs to be included in the
5530 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5531 EVP_MD_type. The old functionality is available in a new macro called
5532 EVP_MD_md(). Change code that uses it and update docs.
5535 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5536 where the 'void *' argument is replaced by a function pointer argument.
5537 Previously 'void *' was abused to point to functions, which works on
5538 many platforms, but is not correct. As these functions are usually
5539 called by macros defined in OpenSSL header files, most source code
5540 should work without changes.
5543 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5544 sections with information on -D... compiler switches used for
5545 compiling the library so that applications can see them. To enable
5546 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5547 must be defined. E.g.,
5548 #define OPENSSL_ALGORITHM_DEFINES
5549 #include <openssl/opensslconf.h>
5550 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5551 [Richard Levitte, Ulf and Bodo Möller]
5553 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5557 *) Change the 'other' type in certificate aux info to a STACK_OF
5558 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5559 the required ASN1 format: arbitrary types determined by an OID.
5562 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5563 argument to 'req'. This is not because the function is newer or
5564 better than others it just uses the work 'NEW' in the certificate
5565 request header lines. Some software needs this.
5568 *) Reorganise password command line arguments: now passwords can be
5569 obtained from various sources. Delete the PEM_cb function and make
5570 it the default behaviour: i.e. if the callback is NULL and the
5571 usrdata argument is not NULL interpret it as a null terminated pass
5572 phrase. If usrdata and the callback are NULL then the pass phrase
5573 is prompted for as usual.
5576 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5577 the support is automatically enabled. The resulting binaries will
5578 autodetect the card and use it if present.
5579 [Ben Laurie and Compaq Inc.]
5581 *) Work around for Netscape hang bug. This sends certificate request
5582 and server done in one record. Since this is perfectly legal in the
5583 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5584 the bugs/SSLv3 entry for more info.
5587 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5590 *) Add -rand argument to smime and pkcs12 applications and read/write
5594 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5597 *) Add command line password options to the remaining applications.
5600 *) Bug fix for BN_div_recp() for numerators with an even number of
5604 *) More tests in bntest.c, and changed test_bn output.
5607 *) ./config recognizes MacOS X now.
5610 *) Bug fix for BN_div() when the first words of num and divsor are
5611 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5614 *) Add support for various broken PKCS#8 formats, and command line
5615 options to produce them.
5618 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5619 get temporary BIGNUMs from a BN_CTX.
5622 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5626 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5627 include a #define from the old name to the new. The original intent
5628 was that statically linked binaries could for example just call
5629 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5630 link with digests. This never worked becayse SSLeay_add_all_digests()
5631 and SSLeay_add_all_ciphers() were in the same source file so calling
5632 one would link with the other. They are now in separate source files.
5635 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5638 *) Use a less unusual form of the Miller-Rabin primality test (it used
5639 a binary algorithm for exponentiation integrated into the Miller-Rabin
5640 loop, our standard modexp algorithms are faster).
5643 *) Support for the EBCDIC character set completed.
5644 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5646 *) Source code cleanups: use const where appropriate, eliminate casts,
5647 use void * instead of char * in lhash.
5650 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5651 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5652 this the server could overwrite ephemeral keys that the client
5656 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5657 using 50 iterations of the Rabin-Miller test.
5659 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5660 iterations of the Rabin-Miller test as required by the appendix
5661 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5662 As BN_is_prime_fasttest includes trial division, DSA parameter
5663 generation becomes much faster.
5665 This implies a change for the callback functions in DSA_is_prime
5666 and DSA_generate_parameters: The callback function is called once
5667 for each positive witness in the Rabin-Miller test, not just
5668 occasionally in the inner loop; and the parameters to the
5669 callback function now provide an iteration count for the outer
5670 loop rather than for the current invocation of the inner loop.
5671 DSA_generate_parameters additionally can call the callback
5672 function with an 'iteration count' of -1, meaning that a
5673 candidate has passed the trial division test (when q is generated
5674 from an application-provided seed, trial division is skipped).
5677 *) New function BN_is_prime_fasttest that optionally does trial
5678 division before starting the Rabin-Miller test and has
5679 an additional BN_CTX * argument (whereas BN_is_prime always
5680 has to allocate at least one BN_CTX).
5681 'callback(1, -1, cb_arg)' is called when a number has passed the
5682 trial division stage.
5685 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5689 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5692 *) New function BN_pseudo_rand().
5695 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5696 bignum version of BN_from_montgomery() with the working code from
5697 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5701 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5702 made it impossible to use the same SSL_SESSION data structure in
5703 SSL2 clients in multiple threads.
5706 *) The return value of RAND_load_file() no longer counts bytes obtained
5707 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5708 to seed the PRNG (previously an explicit byte count was required).
5709 [Ulf Möller, Bodo Möller]
5711 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5712 used (char *) instead of (void *) and had casts all over the place.
5715 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5718 *) Retain source code compatibility for BN_prime_checks macro:
5719 BN_is_prime(..., BN_prime_checks, ...) now uses
5720 BN_prime_checks_for_size to determine the appropriate number of
5721 Rabin-Miller iterations.
5724 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5725 DH_CHECK_P_NOT_SAFE_PRIME.
5726 (Check if this is true? OpenPGP calls them "strong".)
5729 *) Merge the functionality of "dh" and "gendh" programs into a new program
5730 "dhparam". The old programs are retained for now but will handle DH keys
5731 (instead of parameters) in future.
5734 *) Make the ciphers, s_server and s_client programs check the return values
5735 when a new cipher list is set.
5738 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5739 ciphers. Before when the 56bit ciphers were enabled the sorting was
5742 The syntax for the cipher sorting has been extended to support sorting by
5743 cipher-strength (using the strength_bits hard coded in the tables).
5744 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5746 Fix a bug in the cipher-command parser: when supplying a cipher command
5747 string with an "undefined" symbol (neither command nor alphanumeric
5748 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5749 an error is flagged.
5751 Due to the strength-sorting extension, the code of the
5752 ssl_create_cipher_list() function was completely rearranged. I hope that
5753 the readability was also increased :-)
5754 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5756 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5757 for the first serial number and places 2 in the serial number file. This
5758 avoids problems when the root CA is created with serial number zero and
5759 the first user certificate has the same issuer name and serial number
5763 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5764 the new code. Add documentation for this stuff.
5767 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5768 X509_*() to X509at_*() on the grounds that they don't handle X509
5769 structures and behave in an analagous way to the X509v3 functions:
5770 they shouldn't be called directly but wrapper functions should be used
5773 So we also now have some wrapper functions that call the X509at functions
5774 when passed certificate requests. (TO DO: similar things can be done with
5775 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5776 things. Some of these need some d2i or i2d and print functionality
5777 because they handle more complex structures.)
5780 *) Add missing #ifndefs that caused missing symbols when building libssl
5781 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5782 NO_RSA in ssl/s2*.c.
5783 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5785 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5786 has a return value which indicates the quality of the random data
5787 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5788 error queue. New function RAND_pseudo_bytes() generates output that is
5789 guaranteed to be unique but not unpredictable. RAND_add is like
5790 RAND_seed, but takes an extra argument for an entropy estimate
5791 (RAND_seed always assumes full entropy).
5794 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5795 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5796 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5797 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5798 false-positive rate of at most 2^-80 for random input.
5801 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5804 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5805 in the 0.9.5 release), this returns the chain
5806 from an X509_CTX structure with a dup of the stack and all
5807 the X509 reference counts upped: so the stack will exist
5808 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5811 Also make SSL_SESSION_print() print out the verify return
5815 *) Add manpage for the pkcs12 command. Also change the default
5816 behaviour so MAC iteration counts are used unless the new
5817 -nomaciter option is used. This improves file security and
5818 only older versions of MSIE (4.0 for example) need it.
5821 *) Honor the no-xxx Configure options when creating .DEF files.
5824 *) Add PKCS#10 attributes to field table: challengePassword,
5825 unstructuredName and unstructuredAddress. These are taken from
5826 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5827 international characters are used.
5829 More changes to X509_ATTRIBUTE code: allow the setting of types
5830 based on strings. Remove the 'loc' parameter when adding
5831 attributes because these will be a SET OF encoding which is sorted
5835 *) Initial changes to the 'req' utility to allow request generation
5836 automation. This will allow an application to just generate a template
5837 file containing all the field values and have req construct the
5840 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5841 used all over the place including certificate requests and PKCS#7
5842 structures. They are currently handled manually where necessary with
5843 some primitive wrappers for PKCS#7. The new functions behave in a
5844 manner analogous to the X509 extension functions: they allow
5845 attributes to be looked up by NID and added.
5847 Later something similar to the X509V3 code would be desirable to
5848 automatically handle the encoding, decoding and printing of the
5849 more complex types. The string types like challengePassword can
5850 be handled by the string table functions.
5852 Also modified the multi byte string table handling. Now there is
5853 a 'global mask' which masks out certain types. The table itself
5854 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5855 is useful when for example there is only one permissible type
5856 (as in countryName) and using the mask might result in no valid
5860 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5861 SSL_get_peer_finished to allow applications to obtain the latest
5862 Finished messages sent to the peer or expected from the peer,
5863 respectively. (SSL_get_peer_finished is usually the Finished message
5864 actually received from the peer, otherwise the protocol will be aborted.)
5866 As the Finished message are message digests of the complete handshake
5867 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5868 be used for external authentication procedures when the authentication
5869 provided by SSL/TLS is not desired or is not enough.
5872 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5873 the host supports BWX extension and if Compaq C is present on the
5874 $PATH. Just exploiting of the BWX extension results in 20-30%
5875 performance kick for some algorithms, e.g. DES and RC4 to mention
5876 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5880 *) Add support for MS "fast SGC". This is arguably a violation of the
5881 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5882 weak crypto and after checking the certificate is SGC a second one
5883 with strong crypto. MS SGC stops the first handshake after receiving
5884 the server certificate message and sends a second client hello. Since
5885 a server will typically do all the time consuming operations before
5886 expecting any further messages from the client (server key exchange
5887 is the most expensive) there is little difference between the two.
5889 To get OpenSSL to support MS SGC we have to permit a second client
5890 hello message after we have sent server done. In addition we have to
5891 reset the MAC if we do get this second client hello.
5894 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5895 if a DER encoded private key is RSA or DSA traditional format. Changed
5896 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5897 format DER encoded private key. Newer code should use PKCS#8 format which
5898 has the key type encoded in the ASN1 structure. Added DER private key
5899 support to pkcs8 application.
5902 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5903 ciphersuites has been selected (as required by the SSL 3/TLS 1
5904 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5905 is set, we interpret this as a request to violate the specification
5906 (the worst that can happen is a handshake failure, and 'correct'
5907 behaviour would result in a handshake failure anyway).
5910 *) In SSL_CTX_add_session, take into account that there might be multiple
5911 SSL_SESSION structures with the same session ID (e.g. when two threads
5912 concurrently obtain them from an external cache).
5913 The internal cache can handle only one SSL_SESSION with a given ID,
5914 so if there's a conflict, we now throw out the old one to achieve
5918 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5919 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5920 some routines that use cipher OIDs: some ciphers do not have OIDs
5921 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5925 *) Simplify the trust setting structure and code. Now we just have
5926 two sequences of OIDs for trusted and rejected settings. These will
5927 typically have values the same as the extended key usage extension
5928 and any application specific purposes.
5930 The trust checking code now has a default behaviour: it will just
5931 check for an object with the same NID as the passed id. Functions can
5932 be provided to override either the default behaviour or the behaviour
5933 for a given id. SSL client, server and email already have functions
5934 in place for compatibility: they check the NID and also return "trusted"
5935 if the certificate is self signed.
5938 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5939 traditional format into an EVP_PKEY structure.
5942 *) Add a password callback function PEM_cb() which either prompts for
5943 a password if usr_data is NULL or otherwise assumes it is a null
5944 terminated password. Allow passwords to be passed on command line
5945 environment or config files in a few more utilities.
5948 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5949 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5950 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5951 Update documentation.
5954 *) Support for ASN1 "NULL" type. This could be handled before by using
5955 ASN1_TYPE but there wasn't any function that would try to read a NULL
5956 and produce an error if it couldn't. For compatibility we also have
5957 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5958 don't allocate anything because they don't need to.
5961 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5963 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5965 *) Rebuild of the memory allocation routines used by OpenSSL code and
5966 possibly others as well. The purpose is to make an interface that
5967 provide hooks so anyone can build a separate set of allocation and
5968 deallocation routines to be used by OpenSSL, for example memory
5969 pool implementations, or something else, which was previously hard
5970 since Malloc(), Realloc() and Free() were defined as macros having
5971 the values malloc, realloc and free, respectively (except for Win32
5972 compilations). The same is provided for memory debugging code.
5973 OpenSSL already comes with functionality to find memory leaks, but
5974 this gives people a chance to debug other memory problems.
5976 With these changes, a new set of functions and macros have appeared:
5978 CRYPTO_set_mem_debug_functions() [F]
5979 CRYPTO_get_mem_debug_functions() [F]
5980 CRYPTO_dbg_set_options() [F]
5981 CRYPTO_dbg_get_options() [F]
5982 CRYPTO_malloc_debug_init() [M]
5984 The memory debug functions are NULL by default, unless the library
5985 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5986 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5987 gives the standard debugging functions that come with OpenSSL) or
5988 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5989 provided by the library user) must be used. When the standard
5990 debugging functions are used, CRYPTO_dbg_set_options can be used to
5991 request additional information:
5992 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5993 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5995 Also, things like CRYPTO_set_mem_functions will always give the
5996 expected result (the new set of functions is used for allocation
5997 and deallocation) at all times, regardless of platform and compiler
6000 To finish it up, some functions that were never use in any other
6001 way than through macros have a new API and new semantic:
6004 CRYPTO_dbg_realloc()
6007 All macros of value have retained their old syntax.
6008 [Richard Levitte and Bodo Moeller]
6010 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6011 ordering of SMIMECapabilities wasn't in "strength order" and there
6012 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6016 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6017 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6018 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6020 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6021 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6022 functionality to handle multipart/signed properly) and a utility
6023 called 'smime' to call all this stuff. This is based on code I
6024 originally wrote for Celo who have kindly allowed it to be
6025 included in OpenSSL.
6028 *) Add variants des_set_key_checked and des_set_key_unchecked of
6029 des_set_key (aka des_key_sched). Global variable des_check_key
6030 decides which of these is called by des_set_key; this way
6031 des_check_key behaves as it always did, but applications and
6032 the library itself, which was buggy for des_check_key == 1,
6033 have a cleaner way to pick the version they need.
6036 *) New function PKCS12_newpass() which changes the password of a
6040 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6041 dynamic mix. In both cases the ids can be used as an index into the
6042 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6043 functions so they accept a list of the field values and the
6044 application doesn't need to directly manipulate the X509_TRUST
6048 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6052 *) Modify the way the V3 extension code looks up extensions. This now
6053 works in a similar way to the object code: we have some "standard"
6054 extensions in a static table which is searched with OBJ_bsearch()
6055 and the application can add dynamic ones if needed. The file
6056 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6057 updated whenever a new extension is added to the core code and kept
6058 in ext_nid order. There is a simple program 'tabtest.c' which checks
6059 this. New extensions are not added too often so this file can readily
6060 be maintained manually.
6062 There are two big advantages in doing things this way. The extensions
6063 can be looked up immediately and no longer need to be "added" using
6064 X509V3_add_standard_extensions(): this function now does nothing.
6065 [Side note: I get *lots* of email saying the extension code doesn't
6066 work because people forget to call this function]
6067 Also no dynamic allocation is done unless new extensions are added:
6068 so if we don't add custom extensions there is no need to call
6069 X509V3_EXT_cleanup().
6072 *) Modify enc utility's salting as follows: make salting the default. Add a
6073 magic header, so unsalted files fail gracefully instead of just decrypting
6074 to garbage. This is because not salting is a big security hole, so people
6075 should be discouraged from doing it.
6078 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6079 digest to be passed on the command line but it only used this
6080 parameter when signing a certificate. Modified so all relevant
6081 operations are affected by the digest parameter including the
6082 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6083 DSA key was used because it didn't fix the digest.
6086 *) Initial certificate chain verify code. Currently tests the untrusted
6087 certificates for consistency with the verify purpose (which is set
6088 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6090 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6091 this is because it will reject chains with invalid extensions whereas
6092 every previous version of OpenSSL and SSLeay made no checks at all.
6094 Trust code: checks the root CA for the relevant trust settings. Trust
6095 settings have an initial value consistent with the verify purpose: e.g.
6096 if the verify purpose is for SSL client use it expects the CA to be
6097 trusted for SSL client use. However the default value can be changed to
6098 permit custom trust settings: one example of this would be to only trust
6099 certificates from a specific "secure" set of CAs.
6101 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6102 which should be used for version portability: especially since the
6103 verify structure is likely to change more often now.
6105 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6106 to set them. If not set then assume SSL clients will verify SSL servers
6109 Two new options to the verify program: -untrusted allows a set of
6110 untrusted certificates to be passed in and -purpose which sets the
6111 intended purpose of the certificate. If a purpose is set then the
6112 new chain verify code is used to check extension consistency.
6115 *) Support for the authority information access extension.
6118 *) Modify RSA and DSA PEM read routines to transparently handle
6119 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6120 public keys in a format compatible with certificate
6121 SubjectPublicKeyInfo structures. Unfortunately there were already
6122 functions called *_PublicKey_* which used various odd formats so
6123 these are retained for compatibility: however the DSA variants were
6124 never in a public release so they have been deleted. Changed dsa/rsa
6125 utilities to handle the new format: note no releases ever handled public
6126 keys so we should be OK.
6128 The primary motivation for this change is to avoid the same fiasco
6129 that dogs private keys: there are several incompatible private key
6130 formats some of which are standard and some OpenSSL specific and
6131 require various evil hacks to allow partial transparent handling and
6132 even then it doesn't work with DER formats. Given the option anything
6133 other than PKCS#8 should be dumped: but the other formats have to
6134 stay in the name of compatibility.
6136 With public keys and the benefit of hindsight one standard format
6137 is used which works with EVP_PKEY, RSA or DSA structures: though
6138 it clearly returns an error if you try to read the wrong kind of key.
6140 Added a -pubkey option to the 'x509' utility to output the public key.
6141 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6142 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6143 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6144 that do the same as the EVP_PKEY_assign_*() except they up the
6145 reference count of the added key (they don't "swallow" the
6149 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6150 CRLs would fail if the file contained no certificates or no CRLs:
6151 added a new function to read in both types and return the number
6152 read: this means that if none are read it will be an error. The
6153 DER versions of the certificate and CRL reader would always fail
6154 because it isn't possible to mix certificates and CRLs in DER format
6155 without choking one or the other routine. Changed this to just read
6156 a certificate: this is the best we can do. Also modified the code
6157 in apps/verify.c to take notice of return codes: it was previously
6158 attempting to read in certificates from NULL pointers and ignoring
6159 any errors: this is one reason why the cert and CRL reader seemed
6160 to work. It doesn't check return codes from the default certificate
6161 routines: these may well fail if the certificates aren't installed.
6164 *) Code to support otherName option in GeneralName.
6167 *) First update to verify code. Change the verify utility
6168 so it warns if it is passed a self signed certificate:
6169 for consistency with the normal behaviour. X509_verify
6170 has been modified to it will now verify a self signed
6171 certificate if *exactly* the same certificate appears
6172 in the store: it was previously impossible to trust a
6173 single self signed certificate. This means that:
6174 openssl verify ss.pem
6175 now gives a warning about a self signed certificate but
6176 openssl verify -CAfile ss.pem ss.pem
6180 *) For servers, store verify_result in SSL_SESSION data structure
6181 (and add it to external session representation).
6182 This is needed when client certificate verifications fails,
6183 but an application-provided verification callback (set by
6184 SSL_CTX_set_cert_verify_callback) allows accepting the session
6185 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6186 but returns 1): When the session is reused, we have to set
6187 ssl->verify_result to the appropriate error code to avoid
6189 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6191 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6192 case in PKCS7_dataInit() where the signed PKCS7 structure
6193 didn't contain any existing data because it was being created.
6194 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6196 *) Add a salt to the key derivation routines in enc.c. This
6197 forms the first 8 bytes of the encrypted file. Also add a
6198 -S option to allow a salt to be input on the command line.
6201 *) New function X509_cmp(). Oddly enough there wasn't a function
6202 to compare two certificates. We do this by working out the SHA1
6203 hash and comparing that. X509_cmp() will be needed by the trust
6207 *) SSL_get1_session() is like SSL_get_session(), but increments
6208 the reference count in the SSL_SESSION returned.
6209 [Geoff Thorpe <geoff@eu.c2.net>]
6211 *) Fix for 'req': it was adding a null to request attributes.
6212 Also change the X509_LOOKUP and X509_INFO code to handle
6213 certificate auxiliary information.
6216 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6220 *) Add the possibility to add extra information to the memory leak
6221 detecting output, to form tracebacks, showing from where each
6222 allocation was originated: CRYPTO_push_info("constant string") adds
6223 the string plus current file name and line number to a per-thread
6224 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6225 is like calling CYRPTO_pop_info() until the stack is empty.
6226 Also updated memory leak detection code to be multi-thread-safe.
6229 *) Add options -text and -noout to pkcs7 utility and delete the
6230 encryption options which never did anything. Update docs.
6233 *) Add options to some of the utilities to allow the pass phrase
6234 to be included on either the command line (not recommended on
6235 OSes like Unix) or read from the environment. Update the
6236 manpages and fix a few bugs.
6239 *) Add a few manpages for some of the openssl commands.
6242 *) Fix the -revoke option in ca. It was freeing up memory twice,
6243 leaking and not finding already revoked certificates.
6246 *) Extensive changes to support certificate auxiliary information.
6247 This involves the use of X509_CERT_AUX structure and X509_AUX
6248 functions. An X509_AUX function such as PEM_read_X509_AUX()
6249 can still read in a certificate file in the usual way but it
6250 will also read in any additional "auxiliary information". By
6251 doing things this way a fair degree of compatibility can be
6252 retained: existing certificates can have this information added
6253 using the new 'x509' options.
6255 Current auxiliary information includes an "alias" and some trust
6256 settings. The trust settings will ultimately be used in enhanced
6257 certificate chain verification routines: currently a certificate
6258 can only be trusted if it is self signed and then it is trusted
6262 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6263 The problem was that one of the replacement routines had not been working
6264 since SSLeay releases. For now the offending routine has been replaced
6265 with non-optimised assembler. Even so, this now gives around 95%
6266 performance improvement for 1024 bit RSA signs.
6269 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6270 handling. Most clients have the effective key size in bits equal to
6271 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6272 A few however don't do this and instead use the size of the decrypted key
6273 to determine the RC2 key length and the AlgorithmIdentifier to determine
6274 the effective key length. In this case the effective key length can still
6275 be 40 bits but the key length can be 168 bits for example. This is fixed
6276 by manually forcing an RC2 key into the EVP_PKEY structure because the
6277 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6278 the key length and effective key length are equal.
6281 *) Add a bunch of functions that should simplify the creation of
6282 X509_NAME structures. Now you should be able to do:
6283 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6284 and have it automatically work out the correct field type and fill in
6285 the structures. The more adventurous can try:
6286 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6287 and it will (hopefully) work out the correct multibyte encoding.
6290 *) Change the 'req' utility to use the new field handling and multibyte
6291 copy routines. Before the DN field creation was handled in an ad hoc
6292 way in req, ca, and x509 which was rather broken and didn't support
6293 BMPStrings or UTF8Strings. Since some software doesn't implement
6294 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6295 using the dirstring_type option. See the new comment in the default
6296 openssl.cnf for more info.
6299 *) Make crypto/rand/md_rand.c more robust:
6300 - Assure unique random numbers after fork().
6301 - Make sure that concurrent threads access the global counter and
6302 md serializably so that we never lose entropy in them
6303 or use exactly the same state in multiple threads.
6304 Access to the large state is not always serializable because
6305 the additional locking could be a performance killer, and
6306 md should be large enough anyway.
6309 *) New file apps/app_rand.c with commonly needed functionality
6310 for handling the random seed file.
6312 Use the random seed file in some applications that previously did not:
6314 dsaparam -genkey (which also ignored its '-rand' option),
6317 x509 (when signing).
6318 Except on systems with /dev/urandom, it is crucial to have a random
6319 seed file at least for key creation, DSA signing, and for DH exchanges;
6320 for RSA signatures we could do without one.
6322 gendh and gendsa (unlike genrsa) used to read only the first byte
6323 of each file listed in the '-rand' option. The function as previously
6324 found in genrsa is now in app_rand.c and is used by all programs
6325 that support '-rand'.
6328 *) In RAND_write_file, use mode 0600 for creating files;
6329 don't just chmod when it may be too late.
6332 *) Report an error from X509_STORE_load_locations
6333 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6336 *) New function ASN1_mbstring_copy() this copies a string in either
6337 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6338 into an ASN1_STRING type. A mask of permissible types is passed
6339 and it chooses the "minimal" type to use or an error if not type
6343 *) Add function equivalents to the various macros in asn1.h. The old
6344 macros are retained with an M_ prefix. Code inside the library can
6345 use the M_ macros. External code (including the openssl utility)
6346 should *NOT* in order to be "shared library friendly".
6349 *) Add various functions that can check a certificate's extensions
6350 to see if it usable for various purposes such as SSL client,
6351 server or S/MIME and CAs of these types. This is currently
6352 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6353 verification. Also added a -purpose flag to x509 utility to
6354 print out all the purposes.
6357 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6361 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6362 for, obtain and decode and extension and obtain its critical flag.
6363 This allows all the necessary extension code to be handled in a
6364 single function call.
6367 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6368 platforms. See crypto/rc4/rc4_enc.c for further details.
6371 *) New -noout option to asn1parse. This causes no output to be produced
6372 its main use is when combined with -strparse and -out to extract data
6373 from a file (which may not be in ASN.1 format).
6376 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6377 when producing the local key id.
6378 [Richard Levitte <levitte@stacken.kth.se>]
6380 *) New option -dhparam in s_server. This allows a DH parameter file to be
6381 stated explicitly. If it is not stated then it tries the first server
6382 certificate file. The previous behaviour hard coded the filename
6386 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6387 a public key to be input or output. For example:
6388 openssl rsa -in key.pem -pubout -out pubkey.pem
6389 Also added necessary DSA public key functions to handle this.
6392 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6393 in the message. This was handled by allowing
6394 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6395 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6397 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6398 to the end of the strings whereas this didn't. This would cause problems
6399 if strings read with d2i_ASN1_bytes() were later modified.
6400 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6402 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6403 data and it contains EOF it will end up returning an error. This is
6404 caused by input 46 bytes long. The cause is due to the way base64
6405 BIOs find the start of base64 encoded data. They do this by trying a
6406 trial decode on each line until they find one that works. When they
6407 do a flag is set and it starts again knowing it can pass all the
6408 data directly through the decoder. Unfortunately it doesn't reset
6409 the context it uses. This means that if EOF is reached an attempt
6410 is made to pass two EOFs through the context and this causes the
6411 resulting error. This can also cause other problems as well. As is
6412 usual with these problems it takes *ages* to find and the fix is
6413 trivial: move one line.
6414 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6416 *) Ugly workaround to get s_client and s_server working under Windows. The
6417 old code wouldn't work because it needed to select() on sockets and the
6418 tty (for keypresses and to see if data could be written). Win32 only
6419 supports select() on sockets so we select() with a 1s timeout on the
6420 sockets and then see if any characters are waiting to be read, if none
6421 are present then we retry, we also assume we can always write data to
6422 the tty. This isn't nice because the code then blocks until we've
6423 received a complete line of data and it is effectively polling the
6424 keyboard at 1s intervals: however it's quite a bit better than not
6425 working at all :-) A dedicated Windows application might handle this
6426 with an event loop for example.
6429 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6430 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6431 will be called when RSA_sign() and RSA_verify() are used. This is useful
6432 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6433 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6434 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6435 This necessitated the support of an extra signature type NID_md5_sha1
6436 for SSL signatures and modifications to the SSL library to use it instead
6437 of calling RSA_public_decrypt() and RSA_private_encrypt().
6440 *) Add new -verify -CAfile and -CApath options to the crl program, these
6441 will lookup a CRL issuers certificate and verify the signature in a
6442 similar way to the verify program. Tidy up the crl program so it
6443 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6444 less strict. It will now permit CRL extensions even if it is not
6445 a V2 CRL: this will allow it to tolerate some broken CRLs.
6448 *) Initialize all non-automatic variables each time one of the openssl
6449 sub-programs is started (this is necessary as they may be started
6450 multiple times from the "OpenSSL>" prompt).
6451 [Lennart Bang, Bodo Moeller]
6453 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6454 removing all other RSA functionality (this is what NO_RSA does). This
6455 is so (for example) those in the US can disable those operations covered
6456 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6460 *) Non-copying interface to BIO pairs.
6461 (still largely untested)
6464 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6465 ASCII string. This was handled independently in various places before.
6468 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6469 UTF8 strings a character at a time.
6472 *) Use client_version from client hello to select the protocol
6473 (s23_srvr.c) and for RSA client key exchange verification
6474 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6477 *) Add various utility functions to handle SPKACs, these were previously
6478 handled by poking round in the structure internals. Added new function
6479 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6480 print, verify and generate SPKACs. Based on an original idea from
6481 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6484 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6487 *) Allow the config file extension section to be overwritten on the
6488 command line. Based on an original idea from Massimiliano Pala
6489 <madwolf@comune.modena.it>. The new option is called -extensions
6490 and can be applied to ca, req and x509. Also -reqexts to override
6491 the request extensions in req and -crlexts to override the crl extensions
6495 *) Add new feature to the SPKAC handling in ca. Now you can include
6496 the same field multiple times by preceding it by "XXXX." for example:
6499 this is the same syntax as used in the req config file.
6502 *) Allow certificate extensions to be added to certificate requests. These
6503 are specified in a 'req_extensions' option of the req section of the
6504 config file. They can be printed out with the -text option to req but
6505 are otherwise ignored at present.
6508 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6509 data read consists of only the final block it would not decrypted because
6510 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6511 A misplaced 'break' also meant the decrypted final block might not be
6512 copied until the next read.
6515 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6516 a few extra parameters to the DH structure: these will be useful if
6517 for example we want the value of 'q' or implement X9.42 DH.
6520 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6521 provides hooks that allow the default DSA functions or functions on a
6522 "per key" basis to be replaced. This allows hardware acceleration and
6523 hardware key storage to be handled without major modification to the
6524 library. Also added low level modexp hooks and CRYPTO_EX structure and
6525 associated functions.
6528 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6529 as "read only": it can't be written to and the buffer it points to will
6530 not be freed. Reading from a read only BIO is much more efficient than
6531 a normal memory BIO. This was added because there are several times when
6532 an area of memory needs to be read from a BIO. The previous method was
6533 to create a memory BIO and write the data to it, this results in two
6534 copies of the data and an O(n^2) reading algorithm. There is a new
6535 function BIO_new_mem_buf() which creates a read only memory BIO from
6536 an area of memory. Also modified the PKCS#7 routines to use read only
6540 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6541 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6542 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6543 but a retry condition occured while trying to read the rest.
6546 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6547 NID_pkcs7_encrypted by default: this was wrong since this should almost
6548 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6549 the encrypted data type: this is a more sensible place to put it and it
6550 allows the PKCS#12 code to be tidied up that duplicated this
6554 *) Changed obj_dat.pl script so it takes its input and output files on
6555 the command line. This should avoid shell escape redirection problems
6559 *) Initial support for certificate extension requests, these are included
6560 in things like Xenroll certificate requests. Included functions to allow
6561 extensions to be obtained and added.
6564 *) -crlf option to s_client and s_server for sending newlines as
6565 CRLF (as required by many protocols).
6568 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6570 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6571 [Ralf S. Engelschall]
6573 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6574 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6576 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6580 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6581 DH parameters/keys (q is lost during that conversion, but the resulting
6582 DH parameters contain its length).
6584 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6585 much faster than DH_generate_parameters (which creates parameters
6586 where p = 2*q + 1), and also the smaller q makes DH computations
6587 much more efficient (160-bit exponentiation instead of 1024-bit
6588 exponentiation); so this provides a convenient way to support DHE
6589 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6590 utter importance to use
6591 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6593 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6594 when such DH parameters are used, because otherwise small subgroup
6595 attacks may become possible!
6598 *) Avoid memory leak in i2d_DHparams.
6601 *) Allow the -k option to be used more than once in the enc program:
6602 this allows the same encrypted message to be read by multiple recipients.
6605 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6606 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6607 it will always use the numerical form of the OID, even if it has a short
6611 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6612 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6613 otherwise bn_mod_exp was called. In the case of hardware keys for example
6614 no private key components need be present and it might store extra data
6615 in the RSA structure, which cannot be accessed from bn_mod_exp.
6616 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6617 private key operations.
6620 *) Added support for SPARC Linux.
6623 *) pem_password_cb function type incompatibly changed from
6624 typedef int pem_password_cb(char *buf, int size, int rwflag);
6626 ....(char *buf, int size, int rwflag, void *userdata);
6627 so that applications can pass data to their callbacks:
6628 The PEM[_ASN1]_{read,write}... functions and macros now take an
6629 additional void * argument, which is just handed through whenever
6630 the password callback is called.
6631 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6633 New function SSL_CTX_set_default_passwd_cb_userdata.
6635 Compatibility note: As many C implementations push function arguments
6636 onto the stack in reverse order, the new library version is likely to
6637 interoperate with programs that have been compiled with the old
6638 pem_password_cb definition (PEM_whatever takes some data that
6639 happens to be on the stack as its last argument, and the callback
6640 just ignores this garbage); but there is no guarantee whatsoever that
6643 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6644 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6645 problems not only on Windows, but also on some Unix platforms.
6646 To avoid problematic command lines, these definitions are now in an
6647 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6648 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6651 *) MIPS III/IV assembler module is reimplemented.
6654 *) More DES library cleanups: remove references to srand/rand and
6655 delete an unused file.
6658 *) Add support for the the free Netwide assembler (NASM) under Win32,
6659 since not many people have MASM (ml) and it can be hard to obtain.
6660 This is currently experimental but it seems to work OK and pass all
6661 the tests. Check out INSTALL.W32 for info.
6664 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6665 without temporary keys kept an extra copy of the server key,
6666 and connections with temporary keys did not free everything in case
6670 *) New function RSA_check_key and new openssl rsa option -check
6671 for verifying the consistency of RSA keys.
6672 [Ulf Moeller, Bodo Moeller]
6674 *) Various changes to make Win32 compile work:
6675 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6676 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6677 comparison" warnings.
6678 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6681 *) Add a debugging option to PKCS#5 v2 key generation function: when
6682 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6683 derived keys are printed to stderr.
6686 *) Copy the flags in ASN1_STRING_dup().
6687 [Roman E. Pavlov <pre@mo.msk.ru>]
6689 *) The x509 application mishandled signing requests containing DSA
6690 keys when the signing key was also DSA and the parameters didn't match.
6692 It was supposed to omit the parameters when they matched the signing key:
6693 the verifying software was then supposed to automatically use the CA's
6694 parameters if they were absent from the end user certificate.
6696 Omitting parameters is no longer recommended. The test was also
6697 the wrong way round! This was probably due to unusual behaviour in
6698 EVP_cmp_parameters() which returns 1 if the parameters match.
6699 This meant that parameters were omitted when they *didn't* match and
6700 the certificate was useless. Certificates signed with 'ca' didn't have
6702 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6704 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6705 The interface is as follows:
6706 Applications can use
6707 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6708 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6709 "off" is now the default.
6710 The library internally uses
6711 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6712 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6713 to disable memory-checking temporarily.
6715 Some inconsistent states that previously were possible (and were
6716 even the default) are now avoided.
6718 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6719 with each memory chunk allocated; this is occasionally more helpful
6720 than just having a counter.
6722 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6724 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6728 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6729 which largely parallels "options", but is for changing API behaviour,
6730 whereas "options" are about protocol behaviour.
6731 Initial "mode" flags are:
6733 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6734 a single record has been written.
6735 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6736 retries use the same buffer location.
6737 (But all of the contents must be
6741 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6744 *) Fix problems with no-hmac etc.
6745 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6747 *) New functions RSA_get_default_method(), RSA_set_method() and
6748 RSA_get_method(). These allows replacement of RSA_METHODs without having
6749 to mess around with the internals of an RSA structure.
6752 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6753 Also really enable memory leak checks in openssl.c and in some
6755 [Chad C. Mulligan, Bodo Moeller]
6757 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6758 up the length of negative integers. This has now been simplified to just
6759 store the length when it is first determined and use it later, rather
6760 than trying to keep track of where data is copied and updating it to
6762 [Steve Henson, reported by Brien Wheeler
6763 <bwheeler@authentica-security.com>]
6765 *) Add a new function PKCS7_signatureVerify. This allows the verification
6766 of a PKCS#7 signature but with the signing certificate passed to the
6767 function itself. This contrasts with PKCS7_dataVerify which assumes the
6768 certificate is present in the PKCS#7 structure. This isn't always the
6769 case: certificates can be omitted from a PKCS#7 structure and be
6770 distributed by "out of band" means (such as a certificate database).
6773 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6774 function prototypes in pem.h, also change util/mkdef.pl to add the
6775 necessary function names.
6778 *) mk1mf.pl (used by Windows builds) did not properly read the
6779 options set by Configure in the top level Makefile, and Configure
6780 was not even able to write more than one option correctly.
6781 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6784 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6785 file to be loaded from a BIO or FILE pointer. The BIO version will
6786 for example allow memory BIOs to contain config info.
6789 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6790 Whoever hopes to achieve shared-library compatibility across versions
6791 must use this, not the compile-time macro.
6792 (Exercise 0.9.4: Which is the minimum library version required by
6794 Note: All this applies only to multi-threaded programs, others don't
6798 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6799 through a BIO pair triggered the default case, i.e.
6800 SSLerr(...,SSL_R_UNKNOWN_STATE).
6803 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6804 can use the SSL library even if none of the specific BIOs is
6808 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6809 for the encoded length.
6810 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6812 *) Add initial documentation of the X509V3 functions.
6815 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6816 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6817 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6818 secure PKCS#8 private key format with a high iteration count.
6821 *) Fix determination of Perl interpreter: A perl or perl5
6822 _directory_ in $PATH was also accepted as the interpreter.
6823 [Ralf S. Engelschall]
6825 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6826 wrong with it but it was very old and did things like calling
6827 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6831 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6832 to use the new extension code.
6835 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6836 with macros. This should make it easier to change their form, add extra
6837 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6841 *) Add to configuration table a new entry that can specify an alternative
6842 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6843 according to Mark Crispin <MRC@Panda.COM>.
6847 *) DES CBC did not update the IV. Weird.
6850 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6851 Changing the behaviour of the former might break existing programs --
6852 where IV updating is needed, des_ncbc_encrypt can be used.
6855 *) When bntest is run from "make test" it drives bc to check its
6856 calculations, as well as internally checking them. If an internal check
6857 fails, it needs to cause bc to give a non-zero result or make test carries
6858 on without noticing the failure. Fixed.
6861 *) DES library cleanups.
6864 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6865 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6866 ciphers. NOTE: although the key derivation function has been verified
6867 against some published test vectors it has not been extensively tested
6868 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6872 *) Instead of "mkdir -p", which is not fully portable, use new
6873 Perl script "util/mkdir-p.pl".
6876 *) Rewrite the way password based encryption (PBE) is handled. It used to
6877 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6878 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6879 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6880 the 'parameter' field of the AlgorithmIdentifier is passed to the
6881 underlying key generation function so it must do its own ASN1 parsing.
6882 This has also changed the EVP_PBE_CipherInit() function which now has a
6883 'parameter' argument instead of literal salt and iteration count values
6884 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6887 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6888 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6889 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6890 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6891 value was just used as a "magic string" and not used directly its
6892 value doesn't matter.
6895 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6899 *) "linux-sparc64" configuration (ultrapenguin).
6900 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6901 "linux-sparc" configuration.
6902 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6904 *) config now generates no-xxx options for missing ciphers.
6907 *) Support the EBCDIC character set (work in progress).
6908 File ebcdic.c not yet included because it has a different license.
6909 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6911 *) Support BS2000/OSD-POSIX.
6912 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6914 *) Make callbacks for key generation use void * instead of char *.
6917 *) Make S/MIME samples compile (not yet tested).
6920 *) Additional typesafe stacks.
6923 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6927 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6929 *) New configuration variant "sco5-gcc".
6931 *) Updated some demos.
6932 [Sean O Riordain, Wade Scholine]
6934 *) Add missing BIO_free at exit of pkcs12 application.
6937 *) Fix memory leak in conf.c.
6940 *) Updates for Win32 to assembler version of MD5.
6943 *) Set #! path to perl in apps/der_chop to where we found it
6944 instead of using a fixed path.
6947 *) SHA library changes for irix64-mips4-cc.
6950 *) Improvements for VMS support.
6954 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6956 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6957 This also avoids the problems with SC4.2 and unpatched SC5.
6958 [Andy Polyakov <appro@fy.chalmers.se>]
6960 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6961 These are required because of the typesafe stack would otherwise break
6962 existing code. If old code used a structure member which used to be STACK
6963 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6964 sk_num or sk_value it would produce an error because the num, data members
6965 are not present in STACK_OF. Now it just produces a warning. sk_set
6966 replaces the old method of assigning a value to sk_value
6967 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6968 that does this will no longer work (and should use sk_set instead) but
6969 this could be regarded as a "questionable" behaviour anyway.
6972 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6973 correctly handle encrypted S/MIME data.
6976 *) Change type of various DES function arguments from des_cblock
6977 (which means, in function argument declarations, pointer to char)
6978 to des_cblock * (meaning pointer to array with 8 char elements),
6979 which allows the compiler to do more typechecking; it was like
6980 that back in SSLeay, but with lots of ugly casts.
6982 Introduce new type const_des_cblock.
6985 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6986 problems: find RecipientInfo structure that matches recipient certificate
6987 and initialise the ASN1 structures properly based on passed cipher.
6990 *) Belatedly make the BN tests actually check the results.
6993 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6994 to and from BNs: it was completely broken. New compilation option
6995 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6996 key elements as negative integers.
6999 *) Reorganize and speed up MD5.
7000 [Andy Polyakov <appro@fy.chalmers.se>]
7003 [Richard Levitte <richard@levitte.org>]
7005 *) New option -out to asn1parse to allow the parsed structure to be
7006 output to a file. This is most useful when combined with the -strparse
7007 option to examine the output of things like OCTET STRINGS.
7010 *) Make SSL library a little more fool-proof by not requiring any longer
7011 that SSL_set_{accept,connect}_state be called before
7012 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7013 in many applications because usually everything *appeared* to work as
7014 intended anyway -- now it really works as intended).
7017 *) Move openssl.cnf out of lib/.
7020 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7021 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7022 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7023 [Ralf S. Engelschall]
7025 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7026 handle PKCS#7 enveloped data properly.
7027 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7029 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7030 copying pointers. The cert_st handling is changed by this in
7031 various ways (and thus what used to be known as ctx->default_cert
7032 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7033 any longer when s->cert does not give us what we need).
7034 ssl_cert_instantiate becomes obsolete by this change.
7035 As soon as we've got the new code right (possibly it already is?),
7036 we have solved a couple of bugs of the earlier code where s->cert
7037 was used as if it could not have been shared with other SSL structures.
7039 Note that using the SSL API in certain dirty ways now will result
7040 in different behaviour than observed with earlier library versions:
7041 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7042 does not influence s as it used to.
7044 In order to clean up things more thoroughly, inside SSL_SESSION
7045 we don't use CERT any longer, but a new structure SESS_CERT
7046 that holds per-session data (if available); currently, this is
7047 the peer's certificate chain and, for clients, the server's certificate
7048 and temporary key. CERT holds only those values that can have
7049 meaningful defaults in an SSL_CTX.
7052 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7053 from the internal representation. Various PKCS#7 fixes: remove some
7054 evil casts and set the enc_dig_alg field properly based on the signing
7058 *) Allow PKCS#12 password to be set from the command line or the
7059 environment. Let 'ca' get its config file name from the environment
7060 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7064 *) Allow certificate policies extension to use an IA5STRING for the
7065 organization field. This is contrary to the PKIX definition but
7066 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7070 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7071 without disallowing inline assembler and the like for non-pedantic builds.
7074 *) Support Borland C++ builder.
7075 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7080 *) SHA-1 cleanups and performance enhancements.
7081 [Andy Polyakov <appro@fy.chalmers.se>]
7083 *) Sparc v8plus assembler for the bignum library.
7084 [Andy Polyakov <appro@fy.chalmers.se>]
7086 *) Accept any -xxx and +xxx compiler options in Configure.
7089 *) Update HPUX configuration.
7092 *) Add missing sk_<type>_unshift() function to safestack.h
7093 [Ralf S. Engelschall]
7095 *) New function SSL_CTX_use_certificate_chain_file that sets the
7096 "extra_cert"s in addition to the certificate. (This makes sense
7097 only for "PEM" format files, as chains as a whole are not
7101 *) Support verify_depth from the SSL API.
7102 x509_vfy.c had what can be considered an off-by-one-error:
7103 Its depth (which was not part of the external interface)
7104 was actually counting the number of certificates in a chain;
7105 now it really counts the depth.
7108 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7109 instead of X509err, which often resulted in confusing error
7110 messages since the error codes are not globally unique
7111 (e.g. an alleged error in ssl3_accept when a certificate
7112 didn't match the private key).
7114 *) New function SSL_CTX_set_session_id_context that allows to set a default
7115 value (so that you don't need SSL_set_session_id_context for each
7116 connection using the SSL_CTX).
7119 *) OAEP decoding bug fix.
7122 *) Support INSTALL_PREFIX for package builders, as proposed by
7126 *) New Configure options "threads" and "no-threads". For systems
7127 where the proper compiler options are known (currently Solaris
7128 and Linux), "threads" is the default.
7131 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7134 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7135 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7136 such as /usr/local/bin.
7139 *) "make linux-shared" to build shared libraries.
7140 [Niels Poppe <niels@netbox.org>]
7142 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7145 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7146 extension adding in x509 utility.
7149 *) Remove NOPROTO sections and error code comments.
7152 *) Partial rewrite of the DEF file generator to now parse the ANSI
7156 *) New Configure options --prefix=DIR and --openssldir=DIR.
7159 *) Complete rewrite of the error code script(s). It is all now handled
7160 by one script at the top level which handles error code gathering,
7161 header rewriting and C source file generation. It should be much better
7162 than the old method: it now uses a modified version of Ulf's parser to
7163 read the ANSI prototypes in all header files (thus the old K&R definitions
7164 aren't needed for error creation any more) and do a better job of
7165 translating function codes into names. The old 'ASN1 error code imbedded
7166 in a comment' is no longer necessary and it doesn't use .err files which
7167 have now been deleted. Also the error code call doesn't have to appear all
7168 on one line (which resulted in some large lines...).
7171 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7174 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7175 0 (which usually indicates a closed connection), but continue reading.
7178 *) Fix some race conditions.
7181 *) Add support for CRL distribution points extension. Add Certificate
7182 Policies and CRL distribution points documentation.
7185 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7188 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7189 8 of keying material. Merlin has also confirmed interop with this fix
7190 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7191 [Merlin Hughes <merlin@baltimore.ie>]
7193 *) Fix lots of warnings.
7194 [Richard Levitte <levitte@stacken.kth.se>]
7196 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7197 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7198 [Richard Levitte <levitte@stacken.kth.se>]
7200 *) Fix problems with sizeof(long) == 8.
7201 [Andy Polyakov <appro@fy.chalmers.se>]
7203 *) Change functions to ANSI C.
7206 *) Fix typos in error codes.
7207 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7209 *) Remove defunct assembler files from Configure.
7212 *) SPARC v8 assembler BIGNUM implementation.
7213 [Andy Polyakov <appro@fy.chalmers.se>]
7215 *) Support for Certificate Policies extension: both print and set.
7216 Various additions to support the r2i method this uses.
7219 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7220 return a const string when you are expecting an allocated buffer.
7223 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7224 types DirectoryString and DisplayText.
7227 *) Add code to allow r2i extensions to access the configuration database,
7228 add an LHASH database driver and add several ctx helper functions.
7231 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7232 fail when they extended the size of a BIGNUM.
7235 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7236 support typesafe stack.
7239 *) Fix typo in SSL_[gs]et_options().
7240 [Nils Frostberg <nils@medcom.se>]
7242 *) Delete various functions and files that belonged to the (now obsolete)
7243 old X509V3 handling code.
7246 *) New Configure option "rsaref".
7249 *) Don't auto-generate pem.h.
7252 *) Introduce type-safe ASN.1 SETs.
7255 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7256 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7258 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7259 that links with OpenSSL (well at least cause lots of warnings), but fear
7260 not: the conversion is trivial, and it eliminates loads of evil casts. A
7261 few STACKed things have been converted already. Feel free to convert more.
7262 In the fullness of time, I'll do away with the STACK type altogether.
7265 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7266 specified in <certfile> by updating the entry in the index.txt file.
7267 This way one no longer has to edit the index.txt file manually for
7268 revoking a certificate. The -revoke option does the gory details now.
7269 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7271 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7272 `-text' option at all and this way the `-noout -text' combination was
7273 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7274 [Ralf S. Engelschall]
7276 *) Make sure a corresponding plain text error message exists for the
7277 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7278 verify callback function determined that a certificate was revoked.
7279 [Ralf S. Engelschall]
7281 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7282 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7283 all available cipers including rc5, which was forgotten until now.
7284 In order to let the testing shell script know which algorithms
7285 are available, a new (up to now undocumented) command
7286 "openssl list-cipher-commands" is used.
7289 *) Bugfix: s_client occasionally would sleep in select() when
7290 it should have checked SSL_pending() first.
7293 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7294 the raw DSA values prior to ASN.1 encoding.
7297 *) Tweaks to Configure
7298 [Niels Poppe <niels@netbox.org>]
7300 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7304 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7307 *) New config option to avoid instructions that are illegal on the 80386.
7308 The default code is faster, but requires at least a 486.
7311 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7312 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7313 same as SSL2_VERSION anyway.
7316 *) New "-showcerts" option for s_client.
7319 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7320 application. Various cleanups and fixes.
7323 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7324 modify error routines to work internally. Add error codes and PBE init
7325 to library startup routines.
7328 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7329 packing functions to asn1 and evp. Changed function names and error
7330 codes along the way.
7333 *) PKCS12 integration: and so it begins... First of several patches to
7334 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7335 objects to objects.h
7338 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7339 and display support for Thawte strong extranet extension.
7342 *) Add LinuxPPC support.
7343 [Jeff Dubrule <igor@pobox.org>]
7345 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7346 bn_div_words in alpha.s.
7347 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7349 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7350 OAEP isn't supported when OpenSSL is built with RSAref.
7351 [Ulf Moeller <ulf@fitug.de>]
7353 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7354 so they no longer are missing under -DNOPROTO.
7355 [Soren S. Jorvang <soren@t.dk>]
7358 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7360 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7361 doesn't work when the session is reused. Coming soon!
7364 *) Fix a security hole, that allows sessions to be reused in the wrong
7365 context thus bypassing client cert protection! All software that uses
7366 client certs and session caches in multiple contexts NEEDS PATCHING to
7367 allow session reuse! A fuller solution is in the works.
7368 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7370 *) Some more source tree cleanups (removed obsolete files
7371 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7372 permission on "config" script to be executable) and a fix for the INSTALL
7374 [Ulf Moeller <ulf@fitug.de>]
7376 *) Remove some legacy and erroneous uses of malloc, free instead of
7378 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7380 *) Make rsa_oaep_test return non-zero on error.
7381 [Ulf Moeller <ulf@fitug.de>]
7383 *) Add support for native Solaris shared libraries. Configure
7384 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7385 if someone would make that last step automatic.
7386 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7388 *) ctx_size was not built with the right compiler during "make links". Fixed.
7391 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7392 except NULL ciphers". This means the default cipher list will no longer
7393 enable NULL ciphers. They need to be specifically enabled e.g. with
7394 the string "DEFAULT:eNULL".
7397 *) Fix to RSA private encryption routines: if p < q then it would
7398 occasionally produce an invalid result. This will only happen with
7399 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7402 *) Be less restrictive and allow also `perl util/perlpath.pl
7403 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7404 because this way one can also use an interpreter named `perl5' (which is
7405 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7406 installed as `perl').
7407 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7409 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7410 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7412 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7413 advapi32.lib to Win32 build and change the pem test comparision
7414 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7415 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7416 and crypto/des/ede_cbcm_enc.c.
7419 *) DES quad checksum was broken on big-endian architectures. Fixed.
7422 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7423 Win32 test batch file so it (might) work again. The Win32 test batch file
7424 is horrible: I feel ill....
7427 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7428 in e_os.h. Audit of header files to check ANSI and non ANSI
7429 sections: 10 functions were absent from non ANSI section and not exported
7430 from Windows DLLs. Fixed up libeay.num for new functions.
7433 *) Make `openssl version' output lines consistent.
7434 [Ralf S. Engelschall]
7436 *) Fix Win32 symbol export lists for BIO functions: Added
7437 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7438 to ms/libeay{16,32}.def.
7439 [Ralf S. Engelschall]
7441 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7442 fine under Unix and passes some trivial tests I've now added. But the
7443 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7444 added to make sure no one expects that this stuff really works in the
7445 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7446 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7448 [Ralf S. Engelschall]
7450 *) Fix the generation of two part addresses in perl.
7451 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7453 *) Add config entry for Linux on MIPS.
7454 [John Tobey <jtobey@channel1.com>]
7456 *) Make links whenever Configure is run, unless we are on Windoze.
7459 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7460 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7464 *) Add a useful kludge to allow package maintainers to specify compiler and
7465 other platforms details on the command line without having to patch the
7466 Configure script everytime: One now can use ``perl Configure
7467 <id>:<details>'', i.e. platform ids are allowed to have details appended
7468 to them (seperated by colons). This is treated as there would be a static
7469 pre-configured entry in Configure's %table under key <id> with value
7470 <details> and ``perl Configure <id>'' is called. So, when you want to
7471 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7472 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7473 now, which overrides the FreeBSD-elf entry on-the-fly.
7474 [Ralf S. Engelschall]
7476 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7479 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7480 on the `perl Configure ...' command line. This way one can compile
7481 OpenSSL libraries with Position Independent Code (PIC) which is needed
7482 for linking it into DSOs.
7483 [Ralf S. Engelschall]
7485 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7489 *) Cleaned up the LICENSE document: The official contact for any license
7490 questions now is the OpenSSL core team under openssl-core@openssl.org.
7491 And add a paragraph about the dual-license situation to make sure people
7492 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7493 to the OpenSSL toolkit.
7494 [Ralf S. Engelschall]
7496 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7497 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7498 Additonally cleaned up the `make links' target: Remove unnecessary
7499 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7500 to speed processing and no longer clutter the display with confusing
7501 stuff. Instead only the actually done links are displayed.
7502 [Ralf S. Engelschall]
7504 *) Permit null encryption ciphersuites, used for authentication only. It used
7505 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7506 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7510 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7511 signed attributes when verifying signatures (this would break them),
7512 the detached data encoding was wrong and public keys obtained using
7513 X509_get_pubkey() weren't freed.
7516 *) Add text documentation for the BUFFER functions. Also added a work around
7517 to a Win95 console bug. This was triggered by the password read stuff: the
7518 last character typed gets carried over to the next fread(). If you were
7519 generating a new cert request using 'req' for example then the last
7520 character of the passphrase would be CR which would then enter the first
7524 *) Added the new `Includes OpenSSL Cryptography Software' button as
7525 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7526 button and can be used by applications based on OpenSSL to show the
7527 relationship to the OpenSSL project.
7528 [Ralf S. Engelschall]
7530 *) Remove confusing variables in function signatures in files
7531 ssl/ssl_lib.c and ssl/ssl.h.
7532 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7534 *) Don't install bss_file.c under PREFIX/include/
7535 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7537 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7538 functions that return function pointers and has support for NT specific
7539 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7540 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7541 unsigned to signed types: this was killing the Win32 compile.
7544 *) Add new certificate file to stack functions,
7545 SSL_add_dir_cert_subjects_to_stack() and
7546 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7547 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7548 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7549 This means that Apache-SSL and similar packages don't have to mess around
7550 to add as many CAs as they want to the preferred list.
7553 *) Experiment with doxygen documentation. Currently only partially applied to
7555 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7556 openssl.doxy as the configuration file.
7559 *) Get rid of remaining C++-style comments which strict C compilers hate.
7560 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7562 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7563 compiled in by default: it has problems with large keys.
7566 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7567 DH private keys and/or callback functions which directly correspond to
7568 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7569 is needed for applications which have to configure certificates on a
7570 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7572 For the RSA certificate situation is makes no difference, but
7573 for the DSA certificate situation this fixes the "no shared cipher"
7574 problem where the OpenSSL cipher selection procedure failed because the
7575 temporary keys were not overtaken from the context and the API provided
7576 no way to reconfigure them.
7577 The new functions now let applications reconfigure the stuff and they
7578 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7579 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7580 non-public-API function ssl_cert_instantiate() is used as a helper
7581 function and also to reduce code redundancy inside ssl_rsa.c.
7582 [Ralf S. Engelschall]
7584 *) Move s_server -dcert and -dkey options out of the undocumented feature
7585 area because they are useful for the DSA situation and should be
7586 recognized by the users.
7587 [Ralf S. Engelschall]
7589 *) Fix the cipher decision scheme for export ciphers: the export bits are
7590 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7591 SSL_EXP_MASK. So, the original variable has to be used instead of the
7592 already masked variable.
7593 [Richard Levitte <levitte@stacken.kth.se>]
7595 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7596 [Richard Levitte <levitte@stacken.kth.se>]
7598 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7599 from `int' to `unsigned int' because it's a length and initialized by
7600 EVP_DigestFinal() which expects an `unsigned int *'.
7601 [Richard Levitte <levitte@stacken.kth.se>]
7603 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7604 script. Instead use the usual Shell->Perl transition trick.
7605 [Ralf S. Engelschall]
7607 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7608 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7609 -noout -modulus' as it's already the case for `openssl rsa -noout
7610 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7611 currently the public key is printed (a decision which was already done by
7612 `openssl dsa -modulus' in the past) which serves a similar purpose.
7613 Additionally the NO_RSA no longer completely removes the whole -modulus
7614 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7616 [Ralf S. Engelschall]
7618 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7619 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7620 [Arne Ansper <arne@ats.cyber.ee>]
7622 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7623 to be added. Now both 'req' and 'ca' can use new objects defined in the
7627 *) Add cool BIO that does syslog (or event log on NT).
7628 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7630 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7631 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7632 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7633 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7636 *) Add preliminary config info for new extension code.
7639 *) Make RSA_NO_PADDING really use no padding.
7640 [Ulf Moeller <ulf@fitug.de>]
7642 *) Generate errors when private/public key check is done.
7645 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7646 for some CRL extensions and new objects added.
7649 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7650 key usage extension and fuller support for authority key id.
7653 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7654 padding method for RSA, which is recommended for new applications in PKCS
7655 #1 v2.0 (RFC 2437, October 1998).
7656 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7657 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7658 against Bleichbacher's attack on RSA.
7659 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7662 *) Updates to the new SSL compression code
7663 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7665 *) Fix so that the version number in the master secret, when passed
7666 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7667 (because the server will not accept higher), that the version number
7668 is 0x03,0x01, not 0x03,0x00
7669 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7671 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7672 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7673 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7676 *) Support for RAW extensions where an arbitrary extension can be
7677 created by including its DER encoding. See apps/openssl.cnf for
7681 *) Make sure latest Perl versions don't interpret some generated C array
7682 code as Perl array code in the crypto/err/err_genc.pl script.
7683 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7685 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7686 not many people have the assembler. Various Win32 compilation fixes and
7687 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7691 *) Modify configure script 'Configure' to automatically create crypto/date.h
7692 file under Win32 and also build pem.h from pem.org. New script
7693 util/mkfiles.pl to create the MINFO file on environments that can't do a
7694 'make files': perl util/mkfiles.pl >MINFO should work.
7697 *) Major rework of DES function declarations, in the pursuit of correctness
7698 and purity. As a result, many evil casts evaporated, and some weirdness,
7699 too. You may find this causes warnings in your code. Zapping your evil
7700 casts will probably fix them. Mostly.
7703 *) Fix for a typo in asn1.h. Bug fix to object creation script
7704 obj_dat.pl. It considered a zero in an object definition to mean
7705 "end of object": none of the objects in objects.h have any zeros
7706 so it wasn't spotted.
7707 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7709 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7710 Masking (CBCM). In the absence of test vectors, the best I have been able
7711 to do is check that the decrypt undoes the encrypt, so far. Send me test
7712 vectors if you have them.
7715 *) Correct calculation of key length for export ciphers (too much space was
7716 allocated for null ciphers). This has not been tested!
7719 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7720 message is now correct (it understands "crypto" and "ssl" on its
7721 command line). There is also now an "update" option. This will update
7722 the util/ssleay.num and util/libeay.num files with any new functions.
7724 perl util/mkdef.pl crypto ssl update
7725 it will update them.
7728 *) Overhauled the Perl interface (perl/*):
7729 - ported BN stuff to OpenSSL's different BN library
7730 - made the perl/ source tree CVS-aware
7731 - renamed the package from SSLeay to OpenSSL (the files still contain
7732 their history because I've copied them in the repository)
7733 - removed obsolete files (the test scripts will be replaced
7734 by better Test::Harness variants in the future)
7735 [Ralf S. Engelschall]
7737 *) First cut for a very conservative source tree cleanup:
7738 1. merge various obsolete readme texts into doc/ssleay.txt
7739 where we collect the old documents and readme texts.
7740 2. remove the first part of files where I'm already sure that we no
7741 longer need them because of three reasons: either they are just temporary
7742 files which were left by Eric or they are preserved original files where
7743 I've verified that the diff is also available in the CVS via "cvs diff
7744 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7745 the crypto/md/ stuff).
7746 [Ralf S. Engelschall]
7748 *) More extension code. Incomplete support for subject and issuer alt
7749 name, issuer and authority key id. Change the i2v function parameters
7750 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7751 what that's for :-) Fix to ASN1 macro which messed up
7752 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7755 *) Preliminary support for ENUMERATED type. This is largely copied from the
7759 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7760 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7762 *) Make sure `make rehash' target really finds the `openssl' program.
7763 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7765 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7766 like to hear about it if this slows down other processors.
7769 *) Add CygWin32 platform information to Configure script.
7770 [Alan Batie <batie@aahz.jf.intel.com>]
7772 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7773 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7775 *) New program nseq to manipulate netscape certificate sequences
7778 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7782 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7783 but the BN code had some problems that would cause failures when
7784 doing certificate verification and some other functions.
7785 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7787 *) Add ASN1 and PEM code to support netscape certificate sequences.
7790 *) Add ASN1 and PEM code to support netscape certificate sequences.
7793 *) Add several PKIX and private extended key usage OIDs.
7796 *) Modify the 'ca' program to handle the new extension code. Modify
7797 openssl.cnf for new extension format, add comments.
7800 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7801 and add a sample to openssl.cnf so req -x509 now adds appropriate
7805 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7806 error code, add initial support to X509_print() and x509 application.
7809 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7810 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7811 stuff is currently isolated and isn't even compiled yet.
7814 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7815 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7816 Removed the versions check from X509 routines when loading extensions:
7817 this allows certain broken certificates that don't set the version
7818 properly to be processed.
7821 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7822 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7823 can still be regenerated with "make depend".
7826 *) Spelling mistake in C version of CAST-128.
7827 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7829 *) Changes to the error generation code. The perl script err-code.pl
7830 now reads in the old error codes and retains the old numbers, only
7831 adding new ones if necessary. It also only changes the .err files if new
7832 codes are added. The makefiles have been modified to only insert errors
7833 when needed (to avoid needlessly modifying header files). This is done
7834 by only inserting errors if the .err file is newer than the auto generated
7835 C file. To rebuild all the error codes from scratch (the old behaviour)
7836 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7837 or delete all the .err files.
7840 *) CAST-128 was incorrectly implemented for short keys. The C version has
7841 been fixed, but is untested. The assembler versions are also fixed, but
7842 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7843 to regenerate it if needed.
7844 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7845 Hagino <itojun@kame.net>]
7847 *) File was opened incorrectly in randfile.c.
7848 [Ulf Möller <ulf@fitug.de>]
7850 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7851 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7852 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7853 al: it's just almost always a UTCTime. Note this patch adds new error
7854 codes so do a "make errors" if there are problems.
7857 *) Correct Linux 1 recognition in config.
7858 [Ulf Möller <ulf@fitug.de>]
7860 *) Remove pointless MD5 hash when using DSA keys in ca.
7861 [Anonymous <nobody@replay.com>]
7863 *) Generate an error if given an empty string as a cert directory. Also
7864 generate an error if handed NULL (previously returned 0 to indicate an
7865 error, but didn't set one).
7866 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7868 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7871 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7872 parameters. This was causing a warning which killed off the Win32 compile.
7875 *) Remove C++ style comments from crypto/bn/bn_local.h.
7876 [Neil Costigan <neil.costigan@celocom.com>]
7878 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7879 based on a text string, looking up short and long names and finally
7880 "dot" format. The "dot" format stuff didn't work. Added new function
7881 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7882 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7883 OID is not part of the table.
7886 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7887 X509_LOOKUP_by_alias().
7890 *) Sort openssl functions by name.
7893 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7894 encryption from sample DSA keys (in case anyone is interested the password
7898 *) Make _all_ *_free functions accept a NULL pointer.
7899 [Frans Heymans <fheymans@isaserver.be>]
7901 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7903 [Anonymous <nobody@replay.com>]
7905 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7906 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7908 *) Don't blow it for numeric -newkey arguments to apps/req.
7909 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7911 *) Temp key "for export" tests were wrong in s3_srvr.c.
7912 [Anonymous <nobody@replay.com>]
7914 *) Add prototype for temp key callback functions
7915 SSL_CTX_set_tmp_{rsa,dh}_callback().
7918 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7919 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7922 *) X509_name_add_entry() freed the wrong thing after an error.
7923 [Arne Ansper <arne@ats.cyber.ee>]
7925 *) rsa_eay.c would attempt to free a NULL context.
7926 [Arne Ansper <arne@ats.cyber.ee>]
7928 *) BIO_s_socket() had a broken should_retry() on Windoze.
7929 [Arne Ansper <arne@ats.cyber.ee>]
7931 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7932 [Arne Ansper <arne@ats.cyber.ee>]
7934 *) Make sure the already existing X509_STORE->depth variable is initialized
7935 in X509_STORE_new(), but document the fact that this variable is still
7936 unused in the certificate verification process.
7937 [Ralf S. Engelschall]
7939 *) Fix the various library and apps files to free up pkeys obtained from
7940 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7943 *) Fix reference counting in X509_PUBKEY_get(). This makes
7944 demos/maurice/example2.c work, amongst others, probably.
7945 [Steve Henson and Ben Laurie]
7947 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7948 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7949 are no longer created. This way we have a single and consistent command
7950 line interface `openssl <command>', similar to `cvs <command>'.
7951 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7953 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7954 BIT STRING wrapper always have zero unused bits.
7957 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7960 *) Make the top-level INSTALL documentation easier to understand.
7963 *) Makefiles updated to exit if an error occurs in a sub-directory
7964 make (including if user presses ^C) [Paul Sutton]
7966 *) Make Montgomery context stuff explicit in RSA data structure.
7969 *) Fix build order of pem and err to allow for generated pem.h.
7972 *) Fix renumbering bug in X509_NAME_delete_entry().
7975 *) Enhanced the err-ins.pl script so it makes the error library number
7976 global and can add a library name. This is needed for external ASN1 and
7977 other error libraries.
7980 *) Fixed sk_insert which never worked properly.
7983 *) Fix ASN1 macros so they can handle indefinite length construted
7984 EXPLICIT tags. Some non standard certificates use these: they can now
7988 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7989 into a single doc/ssleay.txt bundle. This way the information is still
7990 preserved but no longer messes up this directory. Now it's new room for
7991 the new set of documenation files.
7992 [Ralf S. Engelschall]
7994 *) SETs were incorrectly DER encoded. This was a major pain, because they
7995 shared code with SEQUENCEs, which aren't coded the same. This means that
7996 almost everything to do with SETs or SEQUENCEs has either changed name or
7997 number of arguments.
7998 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8000 *) Fix test data to work with the above.
8003 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8004 was already fixed by Eric for 0.9.1 it seems.
8005 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8007 *) Autodetect FreeBSD3.
8010 *) Fix various bugs in Configure. This affects the following platforms:
8014 unixware-2.0-pentium
8018 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8019 before they are needed.
8022 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8026 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8028 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8029 changed SSLeay to OpenSSL in version strings.
8030 [Ralf S. Engelschall]
8032 *) Some fixups to the top-level documents.
8035 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8036 because the symlink to include/ was missing.
8037 [Ralf S. Engelschall]
8039 *) Incorporated the popular no-RSA/DSA-only patches
8040 which allow to compile a RSA-free SSLeay.
8041 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8043 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8044 when "ssleay" is still not found.
8045 [Ralf S. Engelschall]
8047 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8048 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8050 *) Updated the README file.
8051 [Ralf S. Engelschall]
8053 *) Added various .cvsignore files in the CVS repository subdirs
8054 to make a "cvs update" really silent.
8055 [Ralf S. Engelschall]
8057 *) Recompiled the error-definition header files and added
8058 missing symbols to the Win32 linker tables.
8059 [Ralf S. Engelschall]
8061 *) Cleaned up the top-level documents;
8062 o new files: CHANGES and LICENSE
8063 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8064 o merged COPYRIGHT into LICENSE
8065 o removed obsolete TODO file
8066 o renamed MICROSOFT to INSTALL.W32
8067 [Ralf S. Engelschall]
8069 *) Removed dummy files from the 0.9.1b source tree:
8070 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8071 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8072 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8073 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8074 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8075 [Ralf S. Engelschall]
8077 *) Added various platform portability fixes.
8080 *) The Genesis of the OpenSSL rpject:
8081 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8082 Young and Tim J. Hudson created while they were working for C2Net until
8084 [The OpenSSL Project]
8087 Changes between 0.9.0b and 0.9.1b [not released]
8089 *) Updated a few CA certificates under certs/
8092 *) Changed some BIGNUM api stuff.
8095 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8096 DGUX x86, Linux Alpha, etc.
8099 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8100 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8104 *) Add -strparse option to asn1pars program which parses nested
8106 [Dr Stephen Henson <shenson@bigfoot.com>]
8108 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8111 *) DSA fix for "ca" program.
8114 *) Added "-genkey" option to "dsaparam" program.
8117 *) Added RIPE MD160 (rmd160) message digest.
8120 *) Added -a (all) option to "ssleay version" command.
8123 *) Added PLATFORM define which is the id given to Configure.
8126 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8129 *) Extended the ASN.1 parser routines.
8132 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8135 *) Added a BN_CTX to the BN library.
8138 *) Fixed the weak key values in DES library
8141 *) Changed API in EVP library for cipher aliases.
8144 *) Added support for RC2/64bit cipher.
8147 *) Converted the lhash library to the crypto/mem.c functions.
8150 *) Added more recognized ASN.1 object ids.
8153 *) Added more RSA padding checks for SSL/TLS.
8156 *) Added BIO proxy/filter functionality.
8159 *) Added extra_certs to SSL_CTX which can be used
8160 send extra CA certificates to the client in the CA cert chain sending
8161 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8164 *) Now Fortezza is denied in the authentication phase because
8165 this is key exchange mechanism is not supported by SSLeay at all.
8168 *) Additional PKCS1 checks.
8171 *) Support the string "TLSv1" for all TLS v1 ciphers.
8174 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8175 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8178 *) Fixed a few memory leaks.
8181 *) Fixed various code and comment typos.
8184 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8185 bytes sent in the client random.
8186 [Edward Bishop <ebishop@spyglass.com>]