5 Changes between 0.9.8e and 0.9.8f-fips [xx XXX xxxx]
7 *) New flag EVP_MD_FLAG_SVCTX which passes EVP_MD_CTX and key to underlying
8 sign/verify method. This permits the method to perform finalization
9 and signing itself and have access to the EVP_MD_CTX structure in case
10 additional parameters are needed. Modify fips_{dsa,rsa}_{sign,verify}
11 to use EVP_MD_FLAG_SVCTX and support PSS and X9.31 RSA modes.
12 Modify RSA algorithm test programs to use new parameters.
15 *) Add small standalone ASN1 encoder/decoder to handle DSA signature format.
16 Modify test, algorithm test and selftest routines to use EVP for DSA.
17 Move FIPS implementation of EVP_sha*() and EVP_dss1() under fips-1.0.
20 *) Modify VC++ build system to rename .text and .rdata segments in
21 FIPS sources to .fipst${a,b,c}, and $fipsr${a,b,c} and place them
22 in a static library fipscanister.lib using a perl script. These are
23 then combined by the VC++ linker into a single segment in suffix
24 order but without the suffix (i.e. .fipstx, .fipsrd and .fipsda).
25 This serves the same purpose as fipscanister.o on other platforms
26 but has the advantage that it can be created using only standard VC++
28 [Steve Henson and Andy Polyakov]
30 *) Modify WIN32 build system to forward references functions implemented
34 *) Move error library so that all lhash dependencies are in a separate
35 file. Include a simplified ERR_get_state() function for stand alone
36 FIPS applications. Include a initialization function OPENSSL_init()
37 to set all callbacks, automatically call OPENSSL_init() once when
38 a cipher or digest is added. This should mean that almost all applications
39 set the callbacks automatically. Exceptional cases can call OPENSSL_init()
42 #ifdef OPENSSL_HAVE_INIT
45 before starting any threads.
48 *) Collect common functions into header file "fips_utl.h".
51 *) Only enable dynamic lock functionality in CRYPTO_lock() when it is really
52 needed. Move some lock functionality into new file dyn_lck.c .
53 This further reduces FIPS dependencies allowing the complete removal
54 of STACK and OBJ_bsearch().
57 *) Reduce FIPS test program dependencies by providing stand alone
58 versions of some existing functions in libcrypto. Avoid use
59 of BIOs by converting to system stdio. Move some functions in FIPS
60 files: e.g. all use of BIO_printf().
63 *) Modify build of libcrypto in FIPS mode by using a perl
64 script "arx.pl" which calls the archiver specifically
65 excluding any FIPS dependencies in libcrypto.
68 *) Port OpenSSL 0.9.7 FIPS code to 0.9.8. Convert to new
69 Makefile form. Update Configure. Convert and update
70 FIPS source files. Update libcrypto, libssl and apps
71 with additional functionality from 0.9.7 FIPS code.
72 Update Windows build system.
75 Changes between 0.9.8e and 0.9.8f [23 Feb 2007]
77 *) Mitigate branch prediction attacks, which can be practical if a
78 single processor is shared, allowing a spy process to extract
79 information. For detailed background information, see
80 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
81 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
82 and Necessary Software Countermeasures"). The core of the change
83 are new versions BN_div_no_branch() and
84 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
85 respectively, which are slower, but avoid the security-relevant
86 conditional branches. These are automatically called by BN_div()
87 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for the
88 modulus. Also, BN_is_bit_set() has been changed to remove a
91 BN_FLG_CONSTTIME is the new name for the previous
92 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
93 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
94 in the exponent causes BN_mod_exp_mont() to use the alternative
95 implementation in BN_mod_exp_mont_consttime().) The old name
96 remains as a deprecated alias.
98 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
99 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
100 constant-time implementations for more than just exponentiation.
101 Here too the old name is kept as a deprecated alias.
103 BN_BLINDING_new() will now use BN_dup() for the modulus so that
104 the BN_BLINDING structure gets an independent copy of the
105 modulus. This means that the previous "BIGNUM *m" argument to
106 BN_BLINDING_new() and to BN_BLINDING_create_param() now
107 essentially becomes "const BIGNUM *m", although we can't actually
108 change this in the header file before 0.9.9. It allows
109 RSA_setup_blinding() to use BN_with_flags() on the modulus to
110 enable BN_FLG_CONSTTIME.
112 [Matthew D Wood (Intel Corp)]
114 *) Mitigate branch prediction attacks, which can be practical if a
115 single processor is shared, allowing a spy process to extract
116 information. For detailed background information, see
117 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
118 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
119 and Necessary Software Countermeasures"). The core of the change
120 are new versions BN_div_no_branch() and
121 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
122 respectively, which are slower, but avoid the security-relevant
123 conditional branches. These are automatically called by BN_div()
124 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
125 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
126 remove a conditional branch.
128 BN_FLG_CONSTTIME is the new name for the previous
129 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
130 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
131 in the exponent causes BN_mod_exp_mont() to use the alternative
132 implementation in BN_mod_exp_mont_consttime().) The old name
133 remains as a deprecated alias.
135 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
136 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
137 constant-time implementations for more than just exponentiation.
138 Here too the old name is kept as a deprecated alias.
140 BN_BLINDING_new() will now use BN_dup() for the modulus so that
141 the BN_BLINDING structure gets an independent copy of the
142 modulus. This means that the previous "BIGNUM *m" argument to
143 BN_BLINDING_new() and to BN_BLINDING_create_param() now
144 essentially becomes "const BIGNUM *m", although we can't actually
145 change this in the header file before 0.9.9. It allows
146 RSA_setup_blinding() to use BN_with_flags() on the modulus to
147 enable BN_FLG_CONSTTIME.
149 [Matthew D Wood (Intel Corp)]
151 *) In the SSL/TLS server implementation, be strict about session ID
152 context matching (which matters if an application uses a single
153 external cache for different purposes). Previously,
154 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
155 set. This did ensure strict client verification, but meant that,
156 with applications using a single external cache for quite
157 different requirements, clients could circumvent ciphersuite
158 restrictions for a given session ID context by starting a session
159 in a different context.
162 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
163 a ciphersuite string such as "DEFAULT:RSA" cannot enable
164 authentication-only ciphersuites.
167 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
169 *) Since AES128 and AES256 (and similarly Camellia128 and
170 Camellia256) share a single mask bit in the logic of
171 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
172 kludge to work properly if AES128 is available and AES256 isn't
173 (or if Camellia128 is available and Camellia256 isn't).
176 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
177 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
178 When a point or a seed is encoded in a BIT STRING, we need to
179 prevent the removal of trailing zero bits to get the proper DER
180 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
181 of a NamedBitList, for which trailing 0 bits need to be removed.)
184 *) Have SSL/TLS server implementation tolerate "mismatched" record
185 protocol version while receiving ClientHello even if the
186 ClientHello is fragmented. (The server can't insist on the
187 particular protocol version it has chosen before the ServerHello
188 message has informed the client about his choice.)
191 *) Add RFC 3779 support.
192 [Rob Austein for ARIN, Ben Laurie]
194 *) Load error codes if they are not already present instead of using a
195 static variable. This allows them to be cleanly unloaded and reloaded.
196 Improve header file function name parsing.
199 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
200 or CAPABILITY handshake as required by RFCs.
203 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
205 *) Introduce limits to prevent malicious keys being able to
206 cause a denial of service. (CVE-2006-2940)
207 [Steve Henson, Bodo Moeller]
209 *) Fix ASN.1 parsing of certain invalid structures that can result
210 in a denial of service. (CVE-2006-2937) [Steve Henson]
212 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
213 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
215 *) Fix SSL client code which could crash if connecting to a
216 malicious SSLv2 server. (CVE-2006-4343)
217 [Tavis Ormandy and Will Drewry, Google Security Team]
219 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
220 match only those. Before that, "AES256-SHA" would be interpreted
221 as a pattern and match "AES128-SHA" too (since AES128-SHA got
222 the same strength classification in 0.9.7h) as we currently only
223 have a single AES bit in the ciphersuite description bitmap.
224 That change, however, also applied to ciphersuite strings such as
225 "RC4-MD5" that intentionally matched multiple ciphersuites --
226 namely, SSL 2.0 ciphersuites in addition to the more common ones
227 from SSL 3.0/TLS 1.0.
229 So we change the selection algorithm again: Naming an explicit
230 ciphersuite selects this one ciphersuite, and any other similar
231 ciphersuite (same bitmap) from *other* protocol versions.
232 Thus, "RC4-MD5" again will properly select both the SSL 2.0
233 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
235 Since SSL 2.0 does not have any ciphersuites for which the
236 128/256 bit distinction would be relevant, this works for now.
237 The proper fix will be to use different bits for AES128 and
238 AES256, which would have avoided the problems from the beginning;
239 however, bits are scarce, so we can only do this in a new release
240 (not just a patchlevel) when we can change the SSL_CIPHER
241 definition to split the single 'unsigned long mask' bitmap into
242 multiple values to extend the available space.
246 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
248 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
249 (CVE-2006-4339) [Ben Laurie and Google Security Team]
251 *) Add AES IGE and biIGE modes.
254 *) Change the Unix randomness entropy gathering to use poll() when
255 possible instead of select(), since the latter has some
256 undesirable limitations.
257 [Darryl Miles via Richard Levitte and Bodo Moeller]
259 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
260 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
261 cannot be implicitly activated as part of, e.g., the "AES" alias.
262 However, please upgrade to OpenSSL 0.9.9[-dev] for
263 non-experimental use of the ECC ciphersuites to get TLS extension
264 support, which is required for curve and point format negotiation
265 to avoid potential handshake problems.
268 *) Disable rogue ciphersuites:
270 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
271 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
272 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
274 The latter two were purportedly from
275 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
278 Also deactivate the remaining ciphersuites from
279 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
280 unofficial, and the ID has long expired.
283 *) Fix RSA blinding Heisenbug (problems sometimes occured on
284 dual-core machines) and other potential thread-safety issues.
287 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
288 versions), which is now available for royalty-free use
289 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
290 Also, add Camellia TLS ciphersuites from RFC 4132.
292 To minimize changes between patchlevels in the OpenSSL 0.9.8
293 series, Camellia remains excluded from compilation unless OpenSSL
294 is configured with 'enable-camellia'.
297 *) Disable the padding bug check when compression is in use. The padding
298 bug check assumes the first packet is of even length, this is not
299 necessarily true if compresssion is enabled and can result in false
300 positives causing handshake failure. The actual bug test is ancient
301 code so it is hoped that implementations will either have fixed it by
302 now or any which still have the bug do not support compression.
305 Changes between 0.9.8a and 0.9.8b [04 May 2006]
307 *) When applying a cipher rule check to see if string match is an explicit
308 cipher suite and only match that one cipher suite if it is.
311 *) Link in manifests for VC++ if needed.
312 [Austin Ziegler <halostatue@gmail.com>]
314 *) Update support for ECC-based TLS ciphersuites according to
315 draft-ietf-tls-ecc-12.txt with proposed changes (but without
316 TLS extensions, which are supported starting with the 0.9.9
317 branch, not in the OpenSSL 0.9.8 branch).
320 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
321 opaque EVP_CIPHER_CTX handling.
324 *) Fixes and enhancements to zlib compression code. We now only use
325 "zlib1.dll" and use the default __cdecl calling convention on Win32
326 to conform with the standards mentioned here:
327 http://www.zlib.net/DLL_FAQ.txt
328 Static zlib linking now works on Windows and the new --with-zlib-include
329 --with-zlib-lib options to Configure can be used to supply the location
330 of the headers and library. Gracefully handle case where zlib library
334 *) Several fixes and enhancements to the OID generation code. The old code
335 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
336 handle numbers larger than ULONG_MAX, truncated printing and had a
337 non standard OBJ_obj2txt() behaviour.
340 *) Add support for building of engines under engine/ as shared libraries
341 under VC++ build system.
344 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
345 Hopefully, we will not see any false combination of paths any more.
348 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
350 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
351 (part of SSL_OP_ALL). This option used to disable the
352 countermeasure against man-in-the-middle protocol-version
353 rollback in the SSL 2.0 server implementation, which is a bad
354 idea. (CVE-2005-2969)
356 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
357 for Information Security, National Institute of Advanced Industrial
358 Science and Technology [AIST], Japan)]
360 *) Add two function to clear and return the verify parameter flags.
363 *) Keep cipherlists sorted in the source instead of sorting them at
364 runtime, thus removing the need for a lock.
367 *) Avoid some small subgroup attacks in Diffie-Hellman.
368 [Nick Mathewson and Ben Laurie]
370 *) Add functions for well-known primes.
373 *) Extended Windows CE support.
374 [Satoshi Nakamura and Andy Polyakov]
376 *) Initialize SSL_METHOD structures at compile time instead of during
377 runtime, thus removing the need for a lock.
380 *) Make PKCS7_decrypt() work even if no certificate is supplied by
381 attempting to decrypt each encrypted key in turn. Add support to
385 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
387 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
390 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
393 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
394 key into the same file any more.
397 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
400 *) Add -utf8 command line and config file option to 'ca'.
401 [Stefan <stf@udoma.org]
403 *) Removed the macro des_crypt(), as it seems to conflict with some
404 libraries. Use DES_crypt().
407 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
408 involves renaming the source and generated shared-libs for
409 both. The engines will accept the corrected or legacy ids
410 ('ncipher' and '4758_cca' respectively) when binding. NB,
411 this only applies when building 'shared'.
412 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
414 *) Add attribute functions to EVP_PKEY structure. Modify
415 PKCS12_create() to recognize a CSP name attribute and
416 use it. Make -CSP option work again in pkcs12 utility.
419 *) Add new functionality to the bn blinding code:
420 - automatic re-creation of the BN_BLINDING parameters after
421 a fixed number of uses (currently 32)
422 - add new function for parameter creation
423 - introduce flags to control the update behaviour of the
424 BN_BLINDING parameters
425 - hide BN_BLINDING structure
426 Add a second BN_BLINDING slot to the RSA structure to improve
427 performance when a single RSA object is shared among several
431 *) Add support for DTLS.
432 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
434 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
435 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
438 *) Remove buggy and incompletet DH cert support from
439 ssl/ssl_rsa.c and ssl/s3_both.c
442 *) Use SHA-1 instead of MD5 as the default digest algorithm for
443 the apps/openssl applications.
446 *) Compile clean with "-Wall -Wmissing-prototypes
447 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
448 DEBUG_SAFESTACK must also be set.
451 *) Change ./Configure so that certain algorithms can be disabled by default.
452 The new counterpiece to "no-xxx" is "enable-xxx".
454 The patented RC5 and MDC2 algorithms will now be disabled unless
455 "enable-rc5" and "enable-mdc2", respectively, are specified.
457 (IDEA remains enabled despite being patented. This is because IDEA
458 is frequently required for interoperability, and there is no license
459 fee for non-commercial use. As before, "no-idea" can be used to
460 avoid this algorithm.)
464 *) Add processing of proxy certificates (see RFC 3820). This work was
465 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
466 EGEE (Enabling Grids for E-science in Europe).
469 *) RC4 performance overhaul on modern architectures/implementations, such
470 as Intel P4, IA-64 and AMD64.
473 *) New utility extract-section.pl. This can be used specify an alternative
474 section number in a pod file instead of having to treat each file as
475 a separate case in Makefile. This can be done by adding two lines to the
478 =for comment openssl_section:XXX
480 The blank line is mandatory.
484 *) New arguments -certform, -keyform and -pass for s_client and s_server
485 to allow alternative format key and certificate files and passphrase
489 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
490 update associated structures and add various utility functions.
492 Add new policy related verify parameters, include policy checking in
493 standard verify code. Enhance 'smime' application with extra parameters
494 to support policy checking and print out.
497 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
498 Nehemiah processors. These extensions support AES encryption in hardware
499 as well as RNG (though RNG support is currently disabled).
500 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
502 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
505 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
506 [Andy Polyakov and a number of other people]
508 *) Improved PowerPC platform support. Most notably BIGNUM assembler
509 implementation contributed by IBM.
510 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
512 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
513 exponent rather than 'unsigned long'. There is a corresponding change to
514 the new 'rsa_keygen' element of the RSA_METHOD structure.
515 [Jelte Jansen, Geoff Thorpe]
517 *) Functionality for creating the initial serial number file is now
518 moved from CA.pl to the 'ca' utility with a new option -create_serial.
520 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
521 number file to 1, which is bound to cause problems. To avoid
522 the problems while respecting compatibility between different 0.9.7
523 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
524 CA.pl for serial number initialization. With the new release 0.9.8,
525 we can fix the problem directly in the 'ca' utility.)
528 *) Reduced header interdepencies by declaring more opaque objects in
529 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
530 give fewer recursive includes, which could break lazy source code - so
531 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
532 developers should define this symbol when building and using openssl to
533 ensure they track the recommended behaviour, interfaces, [etc], but
534 backwards-compatible behaviour prevails when this isn't defined.
537 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
540 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
541 This will generate a random key of the appropriate length based on the
542 cipher context. The EVP_CIPHER can provide its own random key generation
543 routine to support keys of a specific form. This is used in the des and
544 3des routines to generate a key of the correct parity. Update S/MIME
545 code to use new functions and hence generate correct parity DES keys.
546 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
547 valid (weak or incorrect parity).
550 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
551 as looking them up. This is useful when the verified structure may contain
552 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
553 present unless the new PKCS7_NO_CRL flag is asserted.
556 *) Extend ASN1 oid configuration module. It now additionally accepts the
559 shortName = some long name, 1.2.3.4
562 *) Reimplemented the BN_CTX implementation. There is now no more static
563 limitation on the number of variables it can handle nor the depth of the
564 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
565 information can now expand as required, and rather than having a single
566 static array of bignums, BN_CTX now uses a linked-list of such arrays
567 allowing it to expand on demand whilst maintaining the usefulness of
571 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
572 to allow all RSA operations to function using a single BN_CTX.
575 *) Preliminary support for certificate policy evaluation and checking. This
576 is initially intended to pass the tests outlined in "Conformance Testing
577 of Relying Party Client Certificate Path Processing Logic" v1.07.
580 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
581 remained unused and not that useful. A variety of other little bignum
582 tweaks and fixes have also been made continuing on from the audit (see
586 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
587 associated ASN1, EVP and SSL functions and old ASN1 macros.
590 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
591 and this should never fail. So the return value from the use of
592 BN_set_word() (which can fail due to needless expansion) is now deprecated;
593 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
596 *) BN_CTX_get() should return zero-valued bignums, providing the same
597 initialised value as BN_new().
598 [Geoff Thorpe, suggested by Ulf Möller]
600 *) Support for inhibitAnyPolicy certificate extension.
603 *) An audit of the BIGNUM code is underway, for which debugging code is
604 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
605 is considered valid when processing BIGNUMs, and causes execution to
606 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
607 further steps are taken to deliberately pollute unused data in BIGNUM
608 structures to try and expose faulty code further on. For now, openssl will
609 (in its default mode of operation) continue to tolerate the inconsistent
610 forms that it has tolerated in the past, but authors and packagers should
611 consider trying openssl and their own applications when compiled with
612 these debugging symbols defined. It will help highlight potential bugs in
613 their own code, and will improve the test coverage for OpenSSL itself. At
614 some point, these tighter rules will become openssl's default to improve
615 maintainability, though the assert()s and other overheads will remain only
616 in debugging configurations. See bn.h for more details.
617 [Geoff Thorpe, Nils Larsch, Ulf Möller]
619 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
620 that can only be obtained through BN_CTX_new() (which implicitly
621 initialises it). The presence of this function only made it possible
622 to overwrite an existing structure (and cause memory leaks).
625 *) Because of the callback-based approach for implementing LHASH as a
626 template type, lh_insert() adds opaque objects to hash-tables and
627 lh_doall() or lh_doall_arg() are typically used with a destructor callback
628 to clean up those corresponding objects before destroying the hash table
629 (and losing the object pointers). So some over-zealous constifications in
630 LHASH have been relaxed so that lh_insert() does not take (nor store) the
631 objects as "const" and the lh_doall[_arg] callback wrappers are not
632 prototyped to have "const" restrictions on the object pointers they are
633 given (and so aren't required to cast them away any more).
636 *) The tmdiff.h API was so ugly and minimal that our own timing utility
637 (speed) prefers to use its own implementation. The two implementations
638 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
639 its object type properly exposed (MS_TM) instead of casting to/from "char
640 *". This may still change yet if someone realises MS_TM and "ms_time_***"
641 aren't necessarily the greatest nomenclatures - but this is what was used
642 internally to the implementation so I've used that for now.
645 *) Ensure that deprecated functions do not get compiled when
646 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
647 the self-tests were still using deprecated key-generation functions so
648 these have been updated also.
651 *) Reorganise PKCS#7 code to separate the digest location functionality
652 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
653 New function PKCS7_set_digest() to set the digest type for PKCS#7
654 digestedData type. Add additional code to correctly generate the
655 digestedData type and add support for this type in PKCS7 initialization
659 *) New function PKCS7_set0_type_other() this initializes a PKCS7
660 structure of type "other".
663 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
664 sure the loop does correctly stop and breaking ("division by zero")
665 modulus operations are not performed. The (pre-generated) prime
666 table crypto/bn/bn_prime.h was already correct, but it could not be
667 re-generated on some platforms because of the "division by zero"
668 situation in the script.
669 [Ralf S. Engelschall]
671 *) Update support for ECC-based TLS ciphersuites according to
672 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
673 SHA-1 now is only used for "small" curves (where the
674 representation of a field element takes up to 24 bytes); for
675 larger curves, the field element resulting from ECDH is directly
676 used as premaster secret.
677 [Douglas Stebila (Sun Microsystems Laboratories)]
679 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
680 curve secp160r1 to the tests.
681 [Douglas Stebila (Sun Microsystems Laboratories)]
683 *) Add the possibility to load symbols globally with DSO.
684 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
686 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
687 control of the error stack.
690 *) Add support for STORE in ENGINE.
693 *) Add the STORE type. The intention is to provide a common interface
694 to certificate and key stores, be they simple file-based stores, or
695 HSM-type store, or LDAP stores, or...
696 NOTE: The code is currently UNTESTED and isn't really used anywhere.
699 *) Add a generic structure called OPENSSL_ITEM. This can be used to
700 pass a list of arguments to any function as well as provide a way
701 for a function to pass data back to the caller.
704 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
705 works like BUF_strdup() but can be used to duplicate a portion of
706 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
710 *) Add the function sk_find_ex() which works like sk_find(), but will
711 return an index to an element even if an exact match couldn't be
712 found. The index is guaranteed to point at the element where the
713 searched-for key would be inserted to preserve sorting order.
716 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
717 takes an extra flags argument for optional functionality. Currently,
718 the following flags are defined:
720 OBJ_BSEARCH_VALUE_ON_NOMATCH
721 This one gets OBJ_bsearch_ex() to return a pointer to the first
722 element where the comparing function returns a negative or zero
725 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
726 This one gets OBJ_bsearch_ex() to return a pointer to the first
727 element where the comparing function returns zero. This is useful
728 if there are more than one element where the comparing function
732 *) Make it possible to create self-signed certificates with 'openssl ca'
733 in such a way that the self-signed certificate becomes part of the
734 CA database and uses the same mechanisms for serial number generation
735 as all other certificate signing. The new flag '-selfsign' enables
736 this functionality. Adapt CA.sh and CA.pl.in.
739 *) Add functionality to check the public key of a certificate request
740 against a given private. This is useful to check that a certificate
741 request can be signed by that key (self-signing).
744 *) Make it possible to have multiple active certificates with the same
745 subject in the CA index file. This is done only if the keyword
746 'unique_subject' is set to 'no' in the main CA section (default
747 if 'CA_default') of the configuration file. The value is saved
748 with the database itself in a separate index attribute file,
749 named like the index file with '.attr' appended to the name.
752 *) Generate muti valued AVAs using '+' notation in config files for
756 *) Support for nameConstraints certificate extension.
759 *) Support for policyConstraints certificate extension.
762 *) Support for policyMappings certificate extension.
765 *) Make sure the default DSA_METHOD implementation only uses its
766 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
767 and change its own handlers to be NULL so as to remove unnecessary
768 indirection. This lets alternative implementations fallback to the
769 default implementation more easily.
772 *) Support for directoryName in GeneralName related extensions
776 *) Make it possible to link applications using Makefile.shared.
777 Make that possible even when linking against static libraries!
780 *) Support for single pass processing for S/MIME signing. This now
781 means that S/MIME signing can be done from a pipe, in addition
782 cleartext signing (multipart/signed type) is effectively streaming
783 and the signed data does not need to be all held in memory.
785 This is done with a new flag PKCS7_STREAM. When this flag is set
786 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
787 is done after the data is output (and digests calculated) in
791 *) Add full support for -rpath/-R, both in shared libraries and
792 applications, at least on the platforms where it's known how
796 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
797 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
798 will now compute a table of multiples of the generator that
799 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
800 faster (notably in the case of a single point multiplication,
802 [Nils Larsch, Bodo Moeller]
804 *) IPv6 support for certificate extensions. The various extensions
805 which use the IP:a.b.c.d can now take IPv6 addresses using the
806 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
810 *) Added an ENGINE that implements RSA by performing private key
811 exponentiations with the GMP library. The conversions to and from
812 GMP's mpz_t format aren't optimised nor are any montgomery forms
813 cached, and on x86 it appears OpenSSL's own performance has caught up.
814 However there are likely to be other architectures where GMP could
815 provide a boost. This ENGINE is not built in by default, but it can be
816 specified at Configure time and should be accompanied by the necessary
817 linker additions, eg;
818 ./config -DOPENSSL_USE_GMP -lgmp
821 *) "openssl engine" will not display ENGINE/DSO load failure errors when
822 testing availability of engines with "-t" - the old behaviour is
823 produced by increasing the feature's verbosity with "-tt".
826 *) ECDSA routines: under certain error conditions uninitialized BN objects
827 could be freed. Solution: make sure initialization is performed early
828 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
832 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
833 and DH_METHOD (eg. by ENGINE implementations) to override the normal
834 software implementations. For DSA and DH, parameter generation can
835 also be overriden by providing the appropriate method callbacks.
838 *) Change the "progress" mechanism used in key-generation and
839 primality testing to functions that take a new BN_GENCB pointer in
840 place of callback/argument pairs. The new API functions have "_ex"
841 postfixes and the older functions are reimplemented as wrappers for
842 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
843 declarations of the old functions to help (graceful) attempts to
844 migrate to the new functions. Also, the new key-generation API
845 functions operate on a caller-supplied key-structure and return
846 success/failure rather than returning a key or NULL - this is to
847 help make "keygen" another member function of RSA_METHOD etc.
849 Example for using the new callback interface:
851 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
855 BN_GENCB_set(&my_cb, my_callback, my_arg);
857 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
858 /* For the meaning of a, b in calls to my_callback(), see the
859 * documentation of the function that calls the callback.
860 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
861 * my_callback should return 1 if it wants BN_is_prime_ex()
862 * to continue, or 0 to stop.
867 *) Change the ZLIB compression method to be stateful, and make it
868 available to TLS with the number defined in
869 draft-ietf-tls-compression-04.txt.
872 *) Add the ASN.1 structures and functions for CertificatePair, which
873 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
875 CertificatePair ::= SEQUENCE {
876 forward [0] Certificate OPTIONAL,
877 reverse [1] Certificate OPTIONAL,
878 -- at least one of the pair shall be present -- }
880 Also implement the PEM functions to read and write certificate
881 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
883 This needed to be defined, mostly for the sake of the LDAP
884 attribute crossCertificatePair, but may prove useful elsewhere as
888 *) Make it possible to inhibit symlinking of shared libraries in
889 Makefile.shared, for Cygwin's sake.
892 *) Extend the BIGNUM API by creating a function
893 void BN_set_negative(BIGNUM *a, int neg);
894 and a macro that behave like
895 int BN_is_negative(const BIGNUM *a);
897 to avoid the need to access 'a->neg' directly in applications.
900 *) Implement fast modular reduction for pseudo-Mersenne primes
901 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
902 EC_GROUP_new_curve_GFp() will now automatically use this
904 [Nils Larsch <nla@trustcenter.de>]
906 *) Add new lock type (CRYPTO_LOCK_BN).
909 *) Change the ENGINE framework to automatically load engines
910 dynamically from specific directories unless they could be
911 found to already be built in or loaded. Move all the
912 current engines except for the cryptodev one to a new
914 The engines in engines/ are built as shared libraries if
915 the "shared" options was given to ./Configure or ./config.
916 Otherwise, they are inserted in libcrypto.a.
917 /usr/local/ssl/engines is the default directory for dynamic
918 engines, but that can be overriden at configure time through
919 the usual use of --prefix and/or --openssldir, and at run
920 time with the environment variable OPENSSL_ENGINES.
921 [Geoff Thorpe and Richard Levitte]
923 *) Add Makefile.shared, a helper makefile to build shared
924 libraries. Addapt Makefile.org.
927 *) Add version info to Win32 DLLs.
928 [Peter 'Luna' Runestig" <peter@runestig.com>]
930 *) Add new 'medium level' PKCS#12 API. Certificates and keys
931 can be added using this API to created arbitrary PKCS#12
932 files while avoiding the low level API.
934 New options to PKCS12_create(), key or cert can be NULL and
935 will then be omitted from the output file. The encryption
936 algorithm NIDs can be set to -1 for no encryption, the mac
937 iteration count can be set to 0 to omit the mac.
939 Enhance pkcs12 utility by making the -nokeys and -nocerts
940 options work when creating a PKCS#12 file. New option -nomac
941 to omit the mac, NONE can be set for an encryption algorithm.
942 New code is modified to use the enhanced PKCS12_create()
943 instead of the low level API.
946 *) Extend ASN1 encoder to support indefinite length constructed
947 encoding. This can output sequences tags and octet strings in
948 this form. Modify pk7_asn1.c to support indefinite length
949 encoding. This is experimental and needs additional code to
950 be useful, such as an ASN1 bio and some enhanced streaming
953 Extend template encode functionality so that tagging is passed
954 down to the template encoder.
957 *) Let 'openssl req' fail if an argument to '-newkey' is not
958 recognized instead of using RSA as a default.
961 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
962 As these are not official, they are not included in "ALL";
963 the "ECCdraft" ciphersuite group alias can be used to select them.
964 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
966 *) Add ECDH engine support.
967 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
969 *) Add ECDH in new directory crypto/ecdh/.
970 [Douglas Stebila (Sun Microsystems Laboratories)]
972 *) Let BN_rand_range() abort with an error after 100 iterations
973 without success (which indicates a broken PRNG).
976 *) Change BN_mod_sqrt() so that it verifies that the input value
977 is really the square of the return value. (Previously,
978 BN_mod_sqrt would show GIGO behaviour.)
981 *) Add named elliptic curves over binary fields from X9.62, SECG,
982 and WAP/WTLS; add OIDs that were still missing.
984 [Sheueling Chang Shantz and Douglas Stebila
985 (Sun Microsystems Laboratories)]
987 *) Extend the EC library for elliptic curves over binary fields
988 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
991 EC_GF2m_simple_method
995 EC_GROUP_new_curve_GF2m
996 EC_GROUP_set_curve_GF2m
997 EC_GROUP_get_curve_GF2m
998 EC_POINT_set_affine_coordinates_GF2m
999 EC_POINT_get_affine_coordinates_GF2m
1000 EC_POINT_set_compressed_coordinates_GF2m
1002 Point compression for binary fields is disabled by default for
1003 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1006 As binary polynomials are represented as BIGNUMs, various members
1007 of the EC_GROUP and EC_POINT data structures can be shared
1008 between the implementations for prime fields and binary fields;
1009 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1010 are essentially identical to their ..._GFp counterparts.
1011 (For simplicity, the '..._GFp' prefix has been dropped from
1012 various internal method names.)
1014 An internal 'field_div' method (similar to 'field_mul' and
1015 'field_sqr') has been added; this is used only for binary fields.
1017 [Sheueling Chang Shantz and Douglas Stebila
1018 (Sun Microsystems Laboratories)]
1020 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1021 through methods ('mul', 'precompute_mult').
1023 The generic implementations (now internally called 'ec_wNAF_mul'
1024 and 'ec_wNAF_precomputed_mult') remain the default if these
1025 methods are undefined.
1027 [Sheueling Chang Shantz and Douglas Stebila
1028 (Sun Microsystems Laboratories)]
1030 *) New function EC_GROUP_get_degree, which is defined through
1031 EC_METHOD. For curves over prime fields, this returns the bit
1032 length of the modulus.
1034 [Sheueling Chang Shantz and Douglas Stebila
1035 (Sun Microsystems Laboratories)]
1037 *) New functions EC_GROUP_dup, EC_POINT_dup.
1038 (These simply call ..._new and ..._copy).
1040 [Sheueling Chang Shantz and Douglas Stebila
1041 (Sun Microsystems Laboratories)]
1043 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1044 Polynomials are represented as BIGNUMs (where the sign bit is not
1045 used) in the following functions [macros]:
1048 BN_GF2m_sub [= BN_GF2m_add]
1049 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1050 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1051 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1053 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1054 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1055 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1056 BN_GF2m_cmp [= BN_ucmp]
1058 (Note that only the 'mod' functions are actually for fields GF(2^m).
1059 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1061 For some functions, an the irreducible polynomial defining a
1062 field can be given as an 'unsigned int[]' with strictly
1063 decreasing elements giving the indices of those bits that are set;
1064 i.e., p[] represents the polynomial
1065 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1067 p[0] > p[1] > ... > p[k] = 0.
1068 This applies to the following functions:
1073 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1074 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1076 BN_GF2m_mod_sqrt_arr
1077 BN_GF2m_mod_solve_quad_arr
1081 Conversion can be performed by the following functions:
1086 bntest.c has additional tests for binary polynomial arithmetic.
1088 Two implementations for BN_GF2m_mod_div() are available.
1089 The default algorithm simply uses BN_GF2m_mod_inv() and
1090 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1091 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1092 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1094 [Sheueling Chang Shantz and Douglas Stebila
1095 (Sun Microsystems Laboratories)]
1097 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1098 functionality is disabled at compile-time.
1099 [Douglas Stebila <douglas.stebila@sun.com>]
1101 *) Change default behaviour of 'openssl asn1parse' so that more
1102 information is visible when viewing, e.g., a certificate:
1104 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1105 mode the content of non-printable OCTET STRINGs is output in a
1106 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1107 avoid the appearance of a printable string.
1108 [Nils Larsch <nla@trustcenter.de>]
1110 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1112 EC_GROUP_set_asn1_flag()
1113 EC_GROUP_get_asn1_flag()
1114 EC_GROUP_set_point_conversion_form()
1115 EC_GROUP_get_point_conversion_form()
1116 These control ASN1 encoding details:
1117 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1118 has been set to OPENSSL_EC_NAMED_CURVE.
1119 - Points are encoded in uncompressed form by default; options for
1120 asn1_for are as for point2oct, namely
1121 POINT_CONVERSION_COMPRESSED
1122 POINT_CONVERSION_UNCOMPRESSED
1123 POINT_CONVERSION_HYBRID
1125 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1128 EC_GROUP_get0_seed()
1129 EC_GROUP_get_seed_len()
1130 This is used only for ASN1 purposes (so far).
1131 [Nils Larsch <nla@trustcenter.de>]
1133 *) Add 'field_type' member to EC_METHOD, which holds the NID
1134 of the appropriate field type OID. The new function
1135 EC_METHOD_get_field_type() returns this value.
1136 [Nils Larsch <nla@trustcenter.de>]
1141 EC_POINT_point2hex()
1142 EC_POINT_hex2point()
1143 providing useful interfaces to EC_POINT_point2oct() and
1144 EC_POINT_oct2point().
1145 [Nils Larsch <nla@trustcenter.de>]
1147 *) Change internals of the EC library so that the functions
1148 EC_GROUP_set_generator()
1149 EC_GROUP_get_generator()
1150 EC_GROUP_get_order()
1151 EC_GROUP_get_cofactor()
1152 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1153 to methods, which would lead to unnecessary code duplication when
1154 adding different types of curves.
1155 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1157 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1158 arithmetic, and such that modified wNAFs are generated
1159 (which avoid length expansion in many cases).
1162 *) Add a function EC_GROUP_check_discriminant() (defined via
1163 EC_METHOD) that verifies that the curve discriminant is non-zero.
1165 Add a function EC_GROUP_check() that makes some sanity tests
1166 on a EC_GROUP, its generator and order. This includes
1167 EC_GROUP_check_discriminant().
1168 [Nils Larsch <nla@trustcenter.de>]
1170 *) Add ECDSA in new directory crypto/ecdsa/.
1172 Add applications 'openssl ecparam' and 'openssl ecdsa'
1173 (these are based on 'openssl dsaparam' and 'openssl dsa').
1175 ECDSA support is also included in various other files across the
1176 library. Most notably,
1177 - 'openssl req' now has a '-newkey ecdsa:file' option;
1178 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1179 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1180 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1181 them suitable for ECDSA where domain parameters must be
1182 extracted before the specific public key;
1183 - ECDSA engine support has been added.
1184 [Nils Larsch <nla@trustcenter.de>]
1186 *) Include some named elliptic curves, and add OIDs from X9.62,
1187 SECG, and WAP/WTLS. Each curve can be obtained from the new
1189 EC_GROUP_new_by_curve_name(),
1190 and the list of available named curves can be obtained with
1191 EC_get_builtin_curves().
1192 Also add a 'curve_name' member to EC_GROUP objects, which can be
1194 EC_GROUP_set_curve_name()
1195 EC_GROUP_get_curve_name()
1196 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1198 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1199 was actually never needed) and in BN_mul(). The removal in BN_mul()
1200 required a small change in bn_mul_part_recursive() and the addition
1201 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1202 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1203 bn_sub_words() and bn_add_words() except they take arrays with
1207 Changes between 0.9.7m and 0.9.7n [xx XXX xxxx]
1209 *) In the SSL/TLS server implementation, be strict about session ID
1210 context matching (which matters if an application uses a single
1211 external cache for different purposes). Previously,
1212 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1213 set. This did ensure strict client verification, but meant that,
1214 with applications using a single external cache for quite
1215 different requirements, clients could circumvent ciphersuite
1216 restrictions for a given session ID context by starting a session
1217 in a different context.
1220 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1222 *) Cleanse PEM buffers before freeing them since they may contain
1224 [Benjamin Bennett <ben@psc.edu>]
1226 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1227 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1228 authentication-only ciphersuites.
1231 *) Since AES128 and AES256 share a single mask bit in the logic of
1232 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1233 kludge to work properly if AES128 is available and AES256 isn't.
1236 *) Expand security boundary to match 1.1.1 module.
1239 *) Remove redundant features: hash file source, editing of test vectors
1240 modify fipsld to use external fips_premain.c signature.
1243 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1244 run algorithm test programs.
1247 *) Make algorithm test programs more tolerant of whitespace.
1250 *) Have SSL/TLS server implementation tolerate "mismatched" record
1251 protocol version while receiving ClientHello even if the
1252 ClientHello is fragmented. (The server can't insist on the
1253 particular protocol version it has chosen before the ServerHello
1254 message has informed the client about his choice.)
1257 *) Load error codes if they are not already present instead of using a
1258 static variable. This allows them to be cleanly unloaded and reloaded.
1261 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1263 *) Introduce limits to prevent malicious keys being able to
1264 cause a denial of service. (CVE-2006-2940)
1265 [Steve Henson, Bodo Moeller]
1267 *) Fix ASN.1 parsing of certain invalid structures that can result
1268 in a denial of service. (CVE-2006-2937) [Steve Henson]
1270 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1271 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1273 *) Fix SSL client code which could crash if connecting to a
1274 malicious SSLv2 server. (CVE-2006-4343)
1275 [Tavis Ormandy and Will Drewry, Google Security Team]
1277 *) Change ciphersuite string processing so that an explicit
1278 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1279 will no longer include "AES128-SHA"), and any other similar
1280 ciphersuite (same bitmap) from *other* protocol versions (so that
1281 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1282 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1283 changes from 0.9.8b and 0.9.8d.
1286 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1288 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1289 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1291 *) Change the Unix randomness entropy gathering to use poll() when
1292 possible instead of select(), since the latter has some
1293 undesirable limitations.
1294 [Darryl Miles via Richard Levitte and Bodo Moeller]
1296 *) Disable rogue ciphersuites:
1298 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1299 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1300 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1302 The latter two were purportedly from
1303 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1306 Also deactive the remaining ciphersuites from
1307 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1308 unofficial, and the ID has long expired.
1311 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1312 dual-core machines) and other potential thread-safety issues.
1315 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1317 *) Adapt fipsld and the build system to link against the validated FIPS
1318 module in FIPS mode.
1321 *) Fixes for VC++ 2005 build under Windows.
1324 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1325 from a Windows bash shell such as MSYS. It is autodetected from the
1326 "config" script when run from a VC++ environment. Modify standard VC++
1327 build to use fipscanister.o from the GNU make build.
1330 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1332 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1333 The value now differs depending on if you build for FIPS or not.
1334 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1335 safely run with a non-FIPSed libcrypto, as it may crash because of
1336 the difference induced by this change.
1339 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1341 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1342 (part of SSL_OP_ALL). This option used to disable the
1343 countermeasure against man-in-the-middle protocol-version
1344 rollback in the SSL 2.0 server implementation, which is a bad
1345 idea. (CVE-2005-2969)
1347 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1348 for Information Security, National Institute of Advanced Industrial
1349 Science and Technology [AIST], Japan)]
1351 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1352 mainly for FIPS compliance and not fully integrated at this stage.
1355 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1356 the exponentiation using a fixed-length exponent. (Otherwise,
1357 the information leaked through timing could expose the secret key
1358 after many signatures; cf. Bleichenbacher's attack on DSA with
1362 *) Make a new fixed-window mod_exp implementation the default for
1363 RSA, DSA, and DH private-key operations so that the sequence of
1364 squares and multiplies and the memory access pattern are
1365 independent of the particular secret key. This will mitigate
1366 cache-timing and potential related attacks.
1368 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1369 and this is automatically used by BN_mod_exp_mont() if the new flag
1370 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1371 will use this BN flag for private exponents unless the flag
1372 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1373 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1375 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1377 *) Change the client implementation for SSLv23_method() and
1378 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1379 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1380 (Previously, the SSL 2.0 backwards compatible Client Hello
1381 message format would be used even with SSL_OP_NO_SSLv2.)
1384 *) Add support for smime-type MIME parameter in S/MIME messages which some
1388 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1389 a threadsafe manner. Modify rsa code to use new function and add calls
1390 to dsa and dh code (which had race conditions before).
1393 *) Include the fixed error library code in the C error file definitions
1394 instead of fixing them up at runtime. This keeps the error code
1395 structures constant.
1398 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1400 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1403 *) Fixes for newer kerberos headers. NB: the casts are needed because
1404 the 'length' field is signed on one version and unsigned on another
1405 with no (?) obvious way to tell the difference, without these VC++
1406 complains. Also the "definition" of FAR (blank) is no longer included
1407 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1408 some needed definitions.
1411 *) Undo Cygwin change.
1414 *) Added support for proxy certificates according to RFC 3820.
1415 Because they may be a security thread to unaware applications,
1416 they must be explicitely allowed in run-time. See
1417 docs/HOWTO/proxy_certificates.txt for further information.
1420 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1422 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1423 server and client random values. Previously
1424 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1425 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1427 This change has negligible security impact because:
1429 1. Server and client random values still have 24 bytes of pseudo random
1432 2. Server and client random values are sent in the clear in the initial
1435 3. The master secret is derived using the premaster secret (48 bytes in
1436 size for static RSA ciphersuites) as well as client server and random
1439 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1442 [Stephen Henson, reported by UK NISCC]
1444 *) Use Windows randomness collection on Cygwin.
1447 *) Fix hang in EGD/PRNGD query when communication socket is closed
1448 prematurely by EGD/PRNGD.
1449 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1451 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1454 *) Back-port of selected performance improvements from development
1455 branch, as well as improved support for PowerPC platforms.
1458 *) Add lots of checks for memory allocation failure, error codes to indicate
1459 failure and freeing up memory if a failure occurs.
1460 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1462 *) Add new -passin argument to dgst.
1465 *) Perform some character comparisons of different types in X509_NAME_cmp:
1466 this is needed for some certificates that reencode DNs into UTF8Strings
1467 (in violation of RFC3280) and can't or wont issue name rollover
1471 *) Make an explicit check during certificate validation to see that
1472 the CA setting in each certificate on the chain is correct. As a
1473 side effect always do the following basic checks on extensions,
1474 not just when there's an associated purpose to the check:
1476 - if there is an unhandled critical extension (unless the user
1477 has chosen to ignore this fault)
1478 - if the path length has been exceeded (if one is set at all)
1479 - that certain extensions fit the associated purpose (if one has
1483 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1485 *) Avoid a race condition when CRLs are checked in a multi threaded
1486 environment. This would happen due to the reordering of the revoked
1487 entries during signature checking and serial number lookup. Now the
1488 encoding is cached and the serial number sort performed under a lock.
1489 Add new STACK function sk_is_sorted().
1492 *) Add Delta CRL to the extension code.
1495 *) Various fixes to s3_pkt.c so alerts are sent properly.
1496 [David Holmes <d.holmes@f5.com>]
1498 *) Reduce the chances of duplicate issuer name and serial numbers (in
1499 violation of RFC3280) using the OpenSSL certificate creation utilities.
1500 This is done by creating a random 64 bit value for the initial serial
1501 number when a serial number file is created or when a self signed
1502 certificate is created using 'openssl req -x509'. The initial serial
1503 number file is created using 'openssl x509 -next_serial' in CA.pl
1504 rather than being initialized to 1.
1507 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1509 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1510 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1511 [Joe Orton, Steve Henson]
1513 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1515 [Joe Orton, Steve Henson]
1517 *) Make it possible to have multiple active certificates with the same
1518 subject in the CA index file. This is done only if the keyword
1519 'unique_subject' is set to 'no' in the main CA section (default
1520 if 'CA_default') of the configuration file. The value is saved
1521 with the database itself in a separate index attribute file,
1522 named like the index file with '.attr' appended to the name.
1525 *) X509 verify fixes. Disable broken certificate workarounds when
1526 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1527 keyUsage extension present. Don't accept CRLs with unhandled critical
1528 extensions: since verify currently doesn't process CRL extensions this
1529 rejects a CRL with *any* critical extensions. Add new verify error codes
1533 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1534 A clarification of RFC2560 will require the use of OCTET STRINGs and
1535 some implementations cannot handle the current raw format. Since OpenSSL
1536 copies and compares OCSP nonces as opaque blobs without any attempt at
1537 parsing them this should not create any compatibility issues.
1540 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1541 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1542 this HMAC (and other) operations are several times slower than OpenSSL
1546 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1547 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1549 *) Use the correct content when signing type "other".
1552 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1554 *) Fix various bugs revealed by running the NISCC test suite:
1556 Stop out of bounds reads in the ASN1 code when presented with
1557 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1559 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1561 If verify callback ignores invalid public key errors don't try to check
1562 certificate signature with the NULL public key.
1566 *) New -ignore_err option in ocsp application to stop the server
1567 exiting on the first error in a request.
1570 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1571 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1575 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1576 extra data after the compression methods not only for TLS 1.0
1577 but also for SSL 3.0 (as required by the specification).
1578 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1580 *) Change X509_certificate_type() to mark the key as exported/exportable
1581 when it's 512 *bits* long, not 512 bytes.
1584 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1585 blocks during encryption.
1588 *) Various fixes to base64 BIO and non blocking I/O. On write
1589 flushes were not handled properly if the BIO retried. On read
1590 data was not being buffered properly and had various logic bugs.
1591 This also affects blocking I/O when the data being decoded is a
1595 *) Various S/MIME bugfixes and compatibility changes:
1596 output correct application/pkcs7 MIME type if
1597 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1598 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1599 of files as .eml work). Correctly handle very long lines in MIME
1603 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1605 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1606 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1607 a protocol version number mismatch like a decryption error
1608 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1611 *) Turn on RSA blinding by default in the default implementation
1612 to avoid a timing attack. Applications that don't want it can call
1613 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1614 They would be ill-advised to do so in most cases.
1615 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1617 *) Change RSA blinding code so that it works when the PRNG is not
1618 seeded (in this case, the secret RSA exponent is abused as
1619 an unpredictable seed -- if it is not unpredictable, there
1620 is no point in blinding anyway). Make RSA blinding thread-safe
1621 by remembering the creator's thread ID in rsa->blinding and
1622 having all other threads use local one-time blinding factors
1623 (this requires more computation than sharing rsa->blinding, but
1624 avoids excessive locking; and if an RSA object is not shared
1625 between threads, blinding will still be very fast).
1628 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1629 ENGINE as defaults for all supported algorithms irrespective of
1630 the 'flags' parameter. 'flags' is now honoured, so applications
1631 should make sure they are passing it correctly.
1634 *) Target "mingw" now allows native Windows code to be generated in
1635 the Cygwin environment as well as with the MinGW compiler.
1638 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1640 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1641 via timing by performing a MAC computation even if incorrrect
1642 block cipher padding has been found. This is a countermeasure
1643 against active attacks where the attacker has to distinguish
1644 between bad padding and a MAC verification error. (CVE-2003-0078)
1646 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1647 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1648 Martin Vuagnoux (EPFL, Ilion)]
1650 *) Make the no-err option work as intended. The intention with no-err
1651 is not to have the whole error stack handling routines removed from
1652 libcrypto, it's only intended to remove all the function name and
1653 reason texts, thereby removing some of the footprint that may not
1654 be interesting if those errors aren't displayed anyway.
1656 NOTE: it's still possible for any application or module to have it's
1657 own set of error texts inserted. The routines are there, just not
1658 used by default when no-err is given.
1661 *) Add support for FreeBSD on IA64.
1662 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1664 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1665 Kerberos function mit_des_cbc_cksum(). Before this change,
1666 the value returned by DES_cbc_cksum() was like the one from
1667 mit_des_cbc_cksum(), except the bytes were swapped.
1668 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1670 *) Allow an application to disable the automatic SSL chain building.
1671 Before this a rather primitive chain build was always performed in
1672 ssl3_output_cert_chain(): an application had no way to send the
1673 correct chain if the automatic operation produced an incorrect result.
1675 Now the chain builder is disabled if either:
1677 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1679 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1681 The reasoning behind this is that an application would not want the
1682 auto chain building to take place if extra chain certificates are
1683 present and it might also want a means of sending no additional
1684 certificates (for example the chain has two certificates and the
1688 *) Add the possibility to build without the ENGINE framework.
1689 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1691 *) Under Win32 gmtime() can return NULL: check return value in
1692 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1695 *) DSA routines: under certain error conditions uninitialized BN objects
1696 could be freed. Solution: make sure initialization is performed early
1697 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1698 Nils Larsch <nla@trustcenter.de> via PR#459)
1701 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1702 checked on reconnect on the client side, therefore session resumption
1703 could still fail with a "ssl session id is different" error. This
1704 behaviour is masked when SSL_OP_ALL is used due to
1705 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1706 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1707 followup to PR #377.
1710 *) IA-32 assembler support enhancements: unified ELF targets, support
1711 for SCO/Caldera platforms, fix for Cygwin shared build.
1714 *) Add support for FreeBSD on sparc64. As a consequence, support for
1715 FreeBSD on non-x86 processors is separate from x86 processors on
1716 the config script, much like the NetBSD support.
1717 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1719 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1721 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1724 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1725 code (06) was taken as the first octet of the session ID and the last
1726 octet was ignored consequently. As a result SSLv2 client side session
1727 caching could not have worked due to the session ID mismatch between
1729 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1733 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1734 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1738 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1739 seems that in spite of existing for more than a year, many application
1740 author have done nothing to provide the necessary callbacks, which
1741 means that this particular engine will not work properly anywhere.
1742 This is a very unfortunate situation which forces us, in the name
1743 of usability, to give the hw_ncipher.c a static lock, which is part
1745 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1746 appear in 0.9.8 or later. We EXPECT application authors to have
1747 dealt properly with this when 0.9.8 is released (unless we actually
1748 make such changes in the libcrypto locking code that changes will
1749 have to be made anyway).
1752 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1753 octets have been read, EOF or an error occurs. Without this change
1754 some truncated ASN1 structures will not produce an error.
1757 *) Disable Heimdal support, since it hasn't been fully implemented.
1758 Still give the possibility to force the use of Heimdal, but with
1759 warnings and a request that patches get sent to openssl-dev.
1762 *) Add the VC-CE target, introduce the WINCE sysname, and add
1763 INSTALL.WCE and appropriate conditionals to make it build.
1764 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1766 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1767 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1768 edit numbers of the version.
1769 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1771 *) Introduce safe string copy and catenation functions
1772 (BUF_strlcpy() and BUF_strlcat()).
1773 [Ben Laurie (CHATS) and Richard Levitte]
1775 *) Avoid using fixed-size buffers for one-line DNs.
1776 [Ben Laurie (CHATS)]
1778 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1779 resizing buffers containing secrets, and use where appropriate.
1780 [Ben Laurie (CHATS)]
1782 *) Avoid using fixed size buffers for configuration file location.
1783 [Ben Laurie (CHATS)]
1785 *) Avoid filename truncation for various CA files.
1786 [Ben Laurie (CHATS)]
1788 *) Use sizeof in preference to magic numbers.
1789 [Ben Laurie (CHATS)]
1791 *) Avoid filename truncation in cert requests.
1792 [Ben Laurie (CHATS)]
1794 *) Add assertions to check for (supposedly impossible) buffer
1796 [Ben Laurie (CHATS)]
1798 *) Don't cache truncated DNS entries in the local cache (this could
1799 potentially lead to a spoofing attack).
1800 [Ben Laurie (CHATS)]
1802 *) Fix various buffers to be large enough for hex/decimal
1803 representations in a platform independent manner.
1804 [Ben Laurie (CHATS)]
1806 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1807 resizing buffers containing secrets, and use where appropriate.
1808 [Ben Laurie (CHATS)]
1810 *) Add BIO_indent() to avoid much slightly worrying code to do
1812 [Ben Laurie (CHATS)]
1814 *) Convert sprintf()/BIO_puts() to BIO_printf().
1815 [Ben Laurie (CHATS)]
1817 *) buffer_gets() could terminate with the buffer only half
1819 [Ben Laurie (CHATS)]
1821 *) Add assertions to prevent user-supplied crypto functions from
1822 overflowing internal buffers by having large block sizes, etc.
1823 [Ben Laurie (CHATS)]
1825 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1827 [Ben Laurie (CHATS)]
1829 *) Eliminate unused copy of key in RC4.
1830 [Ben Laurie (CHATS)]
1832 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1833 [Ben Laurie (CHATS)]
1835 *) Fix off-by-one error in EGD path.
1836 [Ben Laurie (CHATS)]
1838 *) If RANDFILE path is too long, ignore instead of truncating.
1839 [Ben Laurie (CHATS)]
1841 *) Eliminate unused and incorrectly sized X.509 structure
1843 [Ben Laurie (CHATS)]
1845 *) Eliminate unused and dangerous function knumber().
1846 [Ben Laurie (CHATS)]
1848 *) Eliminate unused and dangerous structure, KSSL_ERR.
1849 [Ben Laurie (CHATS)]
1851 *) Protect against overlong session ID context length in an encoded
1852 session object. Since these are local, this does not appear to be
1854 [Ben Laurie (CHATS)]
1856 *) Change from security patch (see 0.9.6e below) that did not affect
1857 the 0.9.6 release series:
1859 Remote buffer overflow in SSL3 protocol - an attacker could
1860 supply an oversized master key in Kerberos-enabled versions.
1862 [Ben Laurie (CHATS)]
1864 *) Change the SSL kerb5 codes to match RFC 2712.
1867 *) Make -nameopt work fully for req and add -reqopt switch.
1868 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1870 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1871 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1873 *) Make sure tests can be performed even if the corresponding algorithms
1874 have been removed entirely. This was also the last step to make
1875 OpenSSL compilable with DJGPP under all reasonable conditions.
1876 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1878 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1879 to allow version independent disabling of normally unselected ciphers,
1880 which may be activated as a side-effect of selecting a single cipher.
1882 (E.g., cipher list string "RSA" enables ciphersuites that are left
1883 out of "ALL" because they do not provide symmetric encryption.
1884 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1885 [Lutz Jaenicke, Bodo Moeller]
1887 *) Add appropriate support for separate platform-dependent build
1888 directories. The recommended way to make a platform-dependent
1889 build directory is the following (tested on Linux), maybe with
1892 # Place yourself outside of the OpenSSL source tree. In
1893 # this example, the environment variable OPENSSL_SOURCE
1894 # is assumed to contain the absolute OpenSSL source directory.
1895 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1896 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1897 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1898 mkdir -p `dirname $F`
1899 ln -s $OPENSSL_SOURCE/$F $F
1902 To be absolutely sure not to disturb the source tree, a "make clean"
1903 is a good thing. If it isn't successfull, don't worry about it,
1904 it probably means the source directory is very clean.
1907 *) Make sure any ENGINE control commands make local copies of string
1908 pointers passed to them whenever necessary. Otherwise it is possible
1909 the caller may have overwritten (or deallocated) the original string
1910 data when a later ENGINE operation tries to use the stored values.
1911 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1913 *) Improve diagnostics in file reading and command-line digests.
1914 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1916 *) Add AES modes CFB and OFB to the object database. Correct an
1917 error in AES-CFB decryption.
1920 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1921 allows existing EVP_CIPHER_CTX structures to be reused after
1922 calling EVP_*Final(). This behaviour is used by encryption
1923 BIOs and some applications. This has the side effect that
1924 applications must explicitly clean up cipher contexts with
1925 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1928 *) Check the values of dna and dnb in bn_mul_recursive before calling
1929 bn_mul_comba (a non zero value means the a or b arrays do not contain
1930 n2 elements) and fallback to bn_mul_normal if either is not zero.
1933 *) Fix escaping of non-ASCII characters when using the -subj option
1934 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1937 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1938 form for "surname", serialNumber has no short form.
1939 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1940 therefore remove "mail" short name for "internet 7".
1941 The OID for unique identifiers in X509 certificates is
1942 x500UniqueIdentifier, not uniqueIdentifier.
1943 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1946 *) Add an "init" command to the ENGINE config module and auto initialize
1947 ENGINEs. Without any "init" command the ENGINE will be initialized
1948 after all ctrl commands have been executed on it. If init=1 the
1949 ENGINE is initailized at that point (ctrls before that point are run
1950 on the uninitialized ENGINE and after on the initialized one). If
1951 init=0 then the ENGINE will not be iniatialized at all.
1954 *) Fix the 'app_verify_callback' interface so that the user-defined
1955 argument is actually passed to the callback: In the
1956 SSL_CTX_set_cert_verify_callback() prototype, the callback
1957 declaration has been changed from
1960 int (*cb)(X509_STORE_CTX *,void *);
1961 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1962 i=s->ctx->app_verify_callback(&ctx)
1963 has been changed into
1964 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1966 To update applications using SSL_CTX_set_cert_verify_callback(),
1967 a dummy argument can be added to their callback functions.
1968 [D. K. Smetters <smetters@parc.xerox.com>]
1970 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1971 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1973 *) Add and OPENSSL_LOAD_CONF define which will cause
1974 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1975 This allows older applications to transparently support certain
1976 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1977 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1978 load the config file and OPENSSL_add_all_algorithms_conf() which will
1979 always load it have also been added.
1982 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1983 Adjust NIDs and EVP layer.
1984 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1986 *) Config modules support in openssl utility.
1988 Most commands now load modules from the config file,
1989 though in a few (such as version) this isn't done
1990 because it couldn't be used for anything.
1992 In the case of ca and req the config file used is
1993 the same as the utility itself: that is the -config
1994 command line option can be used to specify an
1998 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1999 use "openssl_conf" if filename is NULL use default openssl config file.
2002 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2003 config section name. Add a new flag to tolerate a missing config file
2004 and move code to CONF_modules_load_file().
2007 *) Support for crypto accelerator cards from Accelerated Encryption
2008 Processing, www.aep.ie. (Use engine 'aep')
2009 The support was copied from 0.9.6c [engine] and adapted/corrected
2010 to work with the new engine framework.
2011 [AEP Inc. and Richard Levitte]
2013 *) Support for SureWare crypto accelerator cards from Baltimore
2014 Technologies. (Use engine 'sureware')
2015 The support was copied from 0.9.6c [engine] and adapted
2016 to work with the new engine framework.
2019 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2020 make the newer ENGINE framework commands for the CHIL engine work.
2021 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2023 *) Make it possible to produce shared libraries on ReliantUNIX.
2024 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2026 *) Add the configuration target debug-linux-ppro.
2027 Make 'openssl rsa' use the general key loading routines
2028 implemented in apps.c, and make those routines able to
2029 handle the key format FORMAT_NETSCAPE and the variant
2031 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2033 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2034 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2036 *) Add -keyform to rsautl, and document -engine.
2037 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2039 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2040 BIO_R_NO_SUCH_FILE error code rather than the generic
2041 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2044 *) Add new functions
2046 ERR_peek_last_error_line
2047 ERR_peek_last_error_line_data.
2048 These are similar to
2051 ERR_peek_error_line_data,
2052 but report on the latest error recorded rather than the first one
2053 still in the error queue.
2054 [Ben Laurie, Bodo Moeller]
2056 *) default_algorithms option in ENGINE config module. This allows things
2058 default_algorithms = ALL
2059 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2062 *) Prelminary ENGINE config module.
2065 *) New experimental application configuration code.
2068 *) Change the AES code to follow the same name structure as all other
2069 symmetric ciphers, and behave the same way. Move everything to
2070 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2071 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2073 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2074 [Ben Laurie and Theo de Raadt]
2076 *) Add option to output public keys in req command.
2077 [Massimiliano Pala madwolf@openca.org]
2079 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2080 (up to about 10% better than before for P-192 and P-224).
2083 *) New functions/macros
2085 SSL_CTX_set_msg_callback(ctx, cb)
2086 SSL_CTX_set_msg_callback_arg(ctx, arg)
2087 SSL_set_msg_callback(ssl, cb)
2088 SSL_set_msg_callback_arg(ssl, arg)
2090 to request calling a callback function
2092 void cb(int write_p, int version, int content_type,
2093 const void *buf, size_t len, SSL *ssl, void *arg)
2095 whenever a protocol message has been completely received
2096 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2097 protocol version according to which the SSL library interprets
2098 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2099 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2100 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2101 specification (change_cipher_spec(20), alert(21), handshake(22)).
2102 'buf' and 'len' point to the actual message, 'ssl' to the
2103 SSL object, and 'arg' is the application-defined value set by
2104 SSL[_CTX]_set_msg_callback_arg().
2106 'openssl s_client' and 'openssl s_server' have new '-msg' options
2107 to enable a callback that displays all protocol messages.
2110 *) Change the shared library support so shared libraries are built as
2111 soon as the corresponding static library is finished, and thereby get
2112 openssl and the test programs linked against the shared library.
2113 This still only happens when the keyword "shard" has been given to
2114 the configuration scripts.
2116 NOTE: shared library support is still an experimental thing, and
2117 backward binary compatibility is still not guaranteed.
2118 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2120 *) Add support for Subject Information Access extension.
2121 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2123 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2124 additional bytes when new memory had to be allocated, not just
2125 when reusing an existing buffer.
2128 *) New command line and configuration option 'utf8' for the req command.
2129 This allows field values to be specified as UTF8 strings.
2132 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2133 runs for the former and machine-readable output for the latter.
2136 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2137 of the e-mail address in the DN (i.e., it will go into a certificate
2138 extension only). The new configuration file option 'email_in_dn = no'
2139 has the same effect.
2140 [Massimiliano Pala madwolf@openca.org]
2142 *) Change all functions with names starting with des_ to be starting
2143 with DES_ instead. Add wrappers that are compatible with libdes,
2144 but are named _ossl_old_des_*. Finally, add macros that map the
2145 des_* symbols to the corresponding _ossl_old_des_* if libdes
2146 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2147 desired, the des_* symbols will be mapped to DES_*, with one
2150 Since we provide two compatibility mappings, the user needs to
2151 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2152 compatibility is desired. The default (i.e., when that macro
2153 isn't defined) is OpenSSL 0.9.6c compatibility.
2155 There are also macros that enable and disable the support of old
2156 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2157 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2158 are defined, the default will apply: to support the old des routines.
2160 In either case, one must include openssl/des.h to get the correct
2161 definitions. Do not try to just include openssl/des_old.h, that
2164 NOTE: This is a major break of an old API into a new one. Software
2165 authors are encouraged to switch to the DES_ style functions. Some
2166 time in the future, des_old.h and the libdes compatibility functions
2167 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2168 default), and then completely removed.
2171 *) Test for certificates which contain unsupported critical extensions.
2172 If such a certificate is found during a verify operation it is
2173 rejected by default: this behaviour can be overridden by either
2174 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2175 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2176 X509_supported_extension() has also been added which returns 1 if a
2177 particular extension is supported.
2180 *) Modify the behaviour of EVP cipher functions in similar way to digests
2181 to retain compatibility with existing code.
2184 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2185 compatibility with existing code. In particular the 'ctx' parameter does
2186 not have to be to be initialized before the call to EVP_DigestInit() and
2187 it is tidied up after a call to EVP_DigestFinal(). New function
2188 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2189 EVP_MD_CTX_copy() changed to not require the destination to be
2190 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2191 requires the destination to be valid.
2193 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2194 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2197 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2198 so that complete 'Handshake' protocol structures are kept in memory
2199 instead of overwriting 'msg_type' and 'length' with 'body' data.
2202 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2203 [Massimo Santin via Richard Levitte]
2205 *) Major restructuring to the underlying ENGINE code. This includes
2206 reduction of linker bloat, separation of pure "ENGINE" manipulation
2207 (initialisation, etc) from functionality dealing with implementations
2208 of specific crypto iterfaces. This change also introduces integrated
2209 support for symmetric ciphers and digest implementations - so ENGINEs
2210 can now accelerate these by providing EVP_CIPHER and EVP_MD
2211 implementations of their own. This is detailed in crypto/engine/README
2212 as it couldn't be adequately described here. However, there are a few
2213 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2214 were changed in the original introduction of ENGINE code have now
2215 reverted back - the hooking from this code to ENGINE is now a good
2216 deal more passive and at run-time, operations deal directly with
2217 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2218 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2219 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2220 they were not being used by the framework as there is no concept of a
2221 BIGNUM_METHOD and they could not be generalised to the new
2222 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2223 ENGINE_cpy() has been removed as it cannot be consistently defined in
2227 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2230 *) Change mkdef.pl to sort symbols that get the same entry number,
2231 and make sure the automatically generated functions ERR_load_*
2232 become part of libeay.num as well.
2235 *) New function SSL_renegotiate_pending(). This returns true once
2236 renegotiation has been requested (either SSL_renegotiate() call
2237 or HelloRequest/ClientHello receveived from the peer) and becomes
2238 false once a handshake has been completed.
2239 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2240 sends a HelloRequest, but does not ensure that a handshake takes
2241 place. SSL_renegotiate_pending() is useful for checking if the
2242 client has followed the request.)
2245 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2246 By default, clients may request session resumption even during
2247 renegotiation (if session ID contexts permit); with this option,
2248 session resumption is possible only in the first handshake.
2250 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2251 more bits available for options that should not be part of
2252 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2255 *) Add some demos for certificate and certificate request creation.
2258 *) Make maximum certificate chain size accepted from the peer application
2259 settable (SSL*_get/set_max_cert_list()), as proposed by
2260 "Douglas E. Engert" <deengert@anl.gov>.
2263 *) Add support for shared libraries for Unixware-7
2264 (Boyd Lynn Gerber <gerberb@zenez.com>).
2267 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2268 be done prior to destruction. Use this to unload error strings from
2269 ENGINEs that load their own error strings. NB: This adds two new API
2270 functions to "get" and "set" this destroy handler in an ENGINE.
2273 *) Alter all existing ENGINE implementations (except "openssl" and
2274 "openbsd") to dynamically instantiate their own error strings. This
2275 makes them more flexible to be built both as statically-linked ENGINEs
2276 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2277 Also, add stub code to each that makes building them as self-contained
2278 shared-libraries easier (see README.ENGINE).
2281 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2282 implementations into applications that are completely implemented in
2283 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2284 commands that can be used to configure what shared-library to load and
2285 to control aspects of the way it is handled. Also, made an update to
2286 the README.ENGINE file that brings its information up-to-date and
2287 provides some information and instructions on the "dynamic" ENGINE
2288 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2291 *) Make it possible to unload ranges of ERR strings with a new
2292 "ERR_unload_strings" function.
2295 *) Add a copy() function to EVP_MD.
2298 *) Make EVP_MD routines take a context pointer instead of just the
2299 md_data void pointer.
2302 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2303 that the digest can only process a single chunk of data
2304 (typically because it is provided by a piece of
2305 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2306 is only going to provide a single chunk of data, and hence the
2307 framework needn't accumulate the data for oneshot drivers.
2310 *) As with "ERR", make it possible to replace the underlying "ex_data"
2311 functions. This change also alters the storage and management of global
2312 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2313 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2314 index counters. The API functions that use this state have been changed
2315 to take a "class_index" rather than pointers to the class's local STACK
2316 and counter, and there is now an API function to dynamically create new
2317 classes. This centralisation allows us to (a) plug a lot of the
2318 thread-safety problems that existed, and (b) makes it possible to clean
2319 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2320 such data would previously have always leaked in application code and
2321 workarounds were in place to make the memory debugging turn a blind eye
2322 to it. Application code that doesn't use this new function will still
2323 leak as before, but their memory debugging output will announce it now
2324 rather than letting it slide.
2326 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2327 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2328 has a return value to indicate success or failure.
2331 *) Make it possible to replace the underlying "ERR" functions such that the
2332 global state (2 LHASH tables and 2 locks) is only used by the "default"
2333 implementation. This change also adds two functions to "get" and "set"
2334 the implementation prior to it being automatically set the first time
2335 any other ERR function takes place. Ie. an application can call "get",
2336 pass the return value to a module it has just loaded, and that module
2337 can call its own "set" function using that value. This means the
2338 module's "ERR" operations will use (and modify) the error state in the
2339 application and not in its own statically linked copy of OpenSSL code.
2342 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2343 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2344 the operation, and provides a more encapsulated way for external code
2345 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2346 to use these functions rather than manually incrementing the counts.
2348 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2351 *) Add EVP test program.
2354 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2357 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2358 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2359 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2360 These allow a CRL to be built without having to access X509_CRL fields
2361 directly. Modify 'ca' application to use new functions.
2364 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2365 bug workarounds. Rollback attack detection is a security feature.
2366 The problem will only arise on OpenSSL servers when TLSv1 is not
2367 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2368 Software authors not wanting to support TLSv1 will have special reasons
2369 for their choice and can explicitly enable this option.
2370 [Bodo Moeller, Lutz Jaenicke]
2372 *) Rationalise EVP so it can be extended: don't include a union of
2373 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2374 (similar to those existing for EVP_CIPHER_CTX).
2379 EVP_MD_CTX_init(&md); /* new function call */
2380 EVP_DigestInit(&md, EVP_sha1());
2381 EVP_DigestUpdate(&md, in, len);
2382 EVP_DigestFinal(&md, out, NULL);
2383 EVP_MD_CTX_cleanup(&md); /* new function call */
2387 *) Make DES key schedule conform to the usual scheme, as well as
2388 correcting its structure. This means that calls to DES functions
2389 now have to pass a pointer to a des_key_schedule instead of a
2390 plain des_key_schedule (which was actually always a pointer
2393 des_key_schedule ks;
2395 des_set_key_checked(..., &ks);
2396 des_ncbc_encrypt(..., &ks, ...);
2398 (Note that a later change renames 'des_...' into 'DES_...'.)
2401 *) Initial reduction of linker bloat: the use of some functions, such as
2402 PEM causes large amounts of unused functions to be linked in due to
2403 poor organisation. For example pem_all.c contains every PEM function
2404 which has a knock on effect of linking in large amounts of (unused)
2405 ASN1 code. Grouping together similar functions and splitting unrelated
2406 functions prevents this.
2409 *) Cleanup of EVP macros.
2412 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2413 correct _ecb suffix.
2416 *) Add initial OCSP responder support to ocsp application. The
2417 revocation information is handled using the text based index
2418 use by the ca application. The responder can either handle
2419 requests generated internally, supplied in files (for example
2420 via a CGI script) or using an internal minimal server.
2423 *) Add configuration choices to get zlib compression for TLS.
2426 *) Changes to Kerberos SSL for RFC 2712 compliance:
2427 1. Implemented real KerberosWrapper, instead of just using
2428 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2429 2. Implemented optional authenticator field of KerberosWrapper.
2431 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2432 and authenticator structs; see crypto/krb5/.
2434 Generalized Kerberos calls to support multiple Kerberos libraries.
2435 [Vern Staats <staatsvr@asc.hpc.mil>,
2436 Jeffrey Altman <jaltman@columbia.edu>
2437 via Richard Levitte]
2439 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2440 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2441 values for each of the key sizes rather than having just
2442 parameters (and 'speed' generating keys each time).
2445 *) Speed up EVP routines.
2448 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2449 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2450 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2451 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2453 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2454 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2455 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2458 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2460 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2463 *) Added the OS2-EMX target.
2464 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2466 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2467 to support NCONF routines in extension code. New function CONF_set_nconf()
2468 to allow functions which take an NCONF to also handle the old LHASH
2469 structure: this means that the old CONF compatible routines can be
2470 retained (in particular wrt extensions) without having to duplicate the
2471 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2474 *) Enhance the general user interface with mechanisms for inner control
2475 and with possibilities to have yes/no kind of prompts.
2478 *) Change all calls to low level digest routines in the library and
2479 applications to use EVP. Add missing calls to HMAC_cleanup() and
2480 don't assume HMAC_CTX can be copied using memcpy().
2481 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2483 *) Add the possibility to control engines through control names but with
2484 arbitrary arguments instead of just a string.
2485 Change the key loaders to take a UI_METHOD instead of a callback
2486 function pointer. NOTE: this breaks binary compatibility with earlier
2487 versions of OpenSSL [engine].
2488 Adapt the nCipher code for these new conditions and add a card insertion
2492 *) Enhance the general user interface with mechanisms to better support
2493 dialog box interfaces, application-defined prompts, the possibility
2494 to use defaults (for example default passwords from somewhere else)
2495 and interrupts/cancellations.
2498 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2499 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2502 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2503 tidy up some unnecessarily weird code in 'sk_new()').
2504 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2506 *) Change the key loading routines for ENGINEs to use the same kind
2507 callback (pem_password_cb) as all other routines that need this
2511 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2512 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2513 than this minimum value is recommended.
2516 *) New random seeder for OpenVMS, using the system process statistics
2517 that are easily reachable.
2520 *) Windows apparently can't transparently handle global
2521 variables defined in DLLs. Initialisations such as:
2523 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2525 wont compile. This is used by the any applications that need to
2526 declare their own ASN1 modules. This was fixed by adding the option
2527 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2528 needed for static libraries under Win32.
2531 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2532 setting of purpose and trust fields. New X509_STORE trust and
2533 purpose functions and tidy up setting in other SSL functions.
2536 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2537 structure. These are inherited by X509_STORE_CTX when it is
2538 initialised. This allows various defaults to be set in the
2539 X509_STORE structure (such as flags for CRL checking and custom
2540 purpose or trust settings) for functions which only use X509_STORE_CTX
2541 internally such as S/MIME.
2543 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2544 trust settings if they are not set in X509_STORE. This allows X509_STORE
2545 purposes and trust (in S/MIME for example) to override any set by default.
2547 Add command line options for CRL checking to smime, s_client and s_server
2551 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2552 are set then the CRL is looked up in the X509_STORE structure and
2553 its validity and signature checked, then if the certificate is found
2554 in the CRL the verify fails with a revoked error.
2556 Various new CRL related callbacks added to X509_STORE_CTX structure.
2558 Command line options added to 'verify' application to support this.
2560 This needs some additional work, such as being able to handle multiple
2561 CRLs with different times, extension based lookup (rather than just
2562 by subject name) and ultimately more complete V2 CRL extension
2566 *) Add a general user interface API (crypto/ui/). This is designed
2567 to replace things like des_read_password and friends (backward
2568 compatibility functions using this new API are provided).
2569 The purpose is to remove prompting functions from the DES code
2570 section as well as provide for prompting through dialog boxes in
2571 a window system and the like.
2574 *) Add "ex_data" support to ENGINE so implementations can add state at a
2575 per-structure level rather than having to store it globally.
2578 *) Make it possible for ENGINE structures to be copied when retrieved by
2579 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2580 This causes the "original" ENGINE structure to act like a template,
2581 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2582 operational state can be localised to each ENGINE structure, despite the
2583 fact they all share the same "methods". New ENGINE structures returned in
2584 this case have no functional references and the return value is the single
2585 structural reference. This matches the single structural reference returned
2586 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2590 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2591 needs to match any other type at all we need to manually clear the
2595 *) Changes to the "openssl engine" utility to include;
2596 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2597 about an ENGINE's available control commands.
2598 - executing control commands from command line arguments using the
2599 '-pre' and '-post' switches. '-post' is only used if '-t' is
2600 specified and the ENGINE is successfully initialised. The syntax for
2601 the individual commands are colon-separated, for example;
2602 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2605 *) New dynamic control command support for ENGINEs. ENGINEs can now
2606 declare their own commands (numbers), names (strings), descriptions,
2607 and input types for run-time discovery by calling applications. A
2608 subset of these commands are implicitly classed as "executable"
2609 depending on their input type, and only these can be invoked through
2610 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2611 can be based on user input, config files, etc). The distinction is
2612 that "executable" commands cannot return anything other than a boolean
2613 result and can only support numeric or string input, whereas some
2614 discoverable commands may only be for direct use through
2615 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2616 pointers, or other custom uses. The "executable" commands are to
2617 support parameterisations of ENGINE behaviour that can be
2618 unambiguously defined by ENGINEs and used consistently across any
2619 OpenSSL-based application. Commands have been added to all the
2620 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2621 control over shared-library paths without source code alterations.
2624 *) Changed all ENGINE implementations to dynamically allocate their
2625 ENGINEs rather than declaring them statically. Apart from this being
2626 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2627 this also allows the implementations to compile without using the
2628 internal engine_int.h header.
2631 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2632 'const' value. Any code that should be able to modify a RAND_METHOD
2633 should already have non-const pointers to it (ie. they should only
2634 modify their own ones).
2637 *) Made a variety of little tweaks to the ENGINE code.
2638 - "atalla" and "ubsec" string definitions were moved from header files
2639 to C code. "nuron" string definitions were placed in variables
2640 rather than hard-coded - allowing parameterisation of these values
2641 later on via ctrl() commands.
2642 - Removed unused "#if 0"'d code.
2643 - Fixed engine list iteration code so it uses ENGINE_free() to release
2644 structural references.
2645 - Constified the RAND_METHOD element of ENGINE structures.
2646 - Constified various get/set functions as appropriate and added
2647 missing functions (including a catch-all ENGINE_cpy that duplicates
2648 all ENGINE values onto a new ENGINE except reference counts/state).
2649 - Removed NULL parameter checks in get/set functions. Setting a method
2650 or function to NULL is a way of cancelling out a previously set
2651 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2652 and doesn't justify the extra error symbols and code.
2653 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2654 flags from engine_int.h to engine.h.
2655 - Changed prototypes for ENGINE handler functions (init(), finish(),
2656 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2659 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2660 to the algorithm using long division. The binary algorithm can be
2661 used only if the modulus is odd. On 32-bit systems, it is faster
2662 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2663 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2664 up to 450 bits. In 64-bit environments, the binary algorithm
2665 appears to be advantageous for much longer moduli; here we use it
2666 for moduli up to 2048 bits.
2669 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2670 could not support the combine flag in choice fields.
2673 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2674 extensions from a certificate request to the certificate.
2677 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2678 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2679 file: this allows the display of the certificate about to be
2680 signed to be customised, to allow certain fields to be included
2681 or excluded and extension details. The old system didn't display
2682 multicharacter strings properly, omitted fields not in the policy
2683 and couldn't display additional details such as extensions.
2686 *) Function EC_POINTs_mul for multiple scalar multiplication
2687 of an arbitrary number of elliptic curve points
2688 \sum scalars[i]*points[i],
2689 optionally including the generator defined for the EC_GROUP:
2690 scalar*generator + \sum scalars[i]*points[i].
2692 EC_POINT_mul is a simple wrapper function for the typical case
2693 that the point list has just one item (besides the optional
2697 *) First EC_METHODs for curves over GF(p):
2699 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2700 operations and provides various method functions that can also
2701 operate with faster implementations of modular arithmetic.
2703 EC_GFp_mont_method() reuses most functions that are part of
2704 EC_GFp_simple_method, but uses Montgomery arithmetic.
2706 [Bodo Moeller; point addition and point doubling
2707 implementation directly derived from source code provided by
2708 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2710 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2711 crypto/ec/ec_lib.c):
2713 Curves are EC_GROUP objects (with an optional group generator)
2714 based on EC_METHODs that are built into the library.
2716 Points are EC_POINT objects based on EC_GROUP objects.
2718 Most of the framework would be able to handle curves over arbitrary
2719 finite fields, but as there are no obvious types for fields other
2720 than GF(p), some functions are limited to that for now.
2723 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2724 that the file contains a complete HTTP response.
2727 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2728 change the def and num file printf format specifier from "%-40sXXX"
2729 to "%-39s XXX". The latter will always guarantee a space after the
2730 field while the former will cause them to run together if the field
2731 is 40 of more characters long.
2734 *) Constify the cipher and digest 'method' functions and structures
2735 and modify related functions to take constant EVP_MD and EVP_CIPHER
2739 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2740 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2743 *) Modify EVP_Digest*() routines so they now return values. Although the
2744 internal software routines can never fail additional hardware versions
2748 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2750 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2751 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2754 ERR_R_NESTED_ASN1_ERROR
2756 ERR_R_MISSING_ASN1_EOS
2757 were 4 .. 9, conflicting with
2758 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2760 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2761 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2763 Add new error code 'ERR_R_INTERNAL_ERROR'.
2766 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2770 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2771 sets the subject name for a new request or supersedes the
2772 subject name in a given request. Formats that can be parsed are
2773 'CN=Some Name, OU=myOU, C=IT'
2775 'CN=Some Name/OU=myOU/C=IT'.
2777 Add options '-batch' and '-verbose' to 'openssl req'.
2778 [Massimiliano Pala <madwolf@hackmasters.net>]
2780 *) Introduce the possibility to access global variables through
2781 functions on platform were that's the best way to handle exporting
2782 global variables in shared libraries. To enable this functionality,
2783 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2784 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2785 is normally done by Configure or something similar).
2787 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2788 in the source file (foo.c) like this:
2790 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2791 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2793 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2794 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2796 OPENSSL_DECLARE_GLOBAL(int,foo);
2797 #define foo OPENSSL_GLOBAL_REF(foo)
2798 OPENSSL_DECLARE_GLOBAL(double,bar);
2799 #define bar OPENSSL_GLOBAL_REF(bar)
2801 The #defines are very important, and therefore so is including the
2802 header file everywhere where the defined globals are used.
2804 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2805 of ASN.1 items, but that structure is a bit different.
2807 The largest change is in util/mkdef.pl which has been enhanced with
2808 better and easier to understand logic to choose which symbols should
2809 go into the Windows .def files as well as a number of fixes and code
2810 cleanup (among others, algorithm keywords are now sorted
2811 lexicographically to avoid constant rewrites).
2814 *) In BN_div() keep a copy of the sign of 'num' before writing the
2815 result to 'rm' because if rm==num the value will be overwritten
2816 and produce the wrong result if 'num' is negative: this caused
2817 problems with BN_mod() and BN_nnmod().
2820 *) Function OCSP_request_verify(). This checks the signature on an
2821 OCSP request and verifies the signer certificate. The signer
2822 certificate is just checked for a generic purpose and OCSP request
2826 *) Add OCSP_check_validity() function to check the validity of OCSP
2827 responses. OCSP responses are prepared in real time and may only
2828 be a few seconds old. Simply checking that the current time lies
2829 between thisUpdate and nextUpdate max reject otherwise valid responses
2830 caused by either OCSP responder or client clock inaccuracy. Instead
2831 we allow thisUpdate and nextUpdate to fall within a certain period of
2832 the current time. The age of the response can also optionally be
2833 checked. Two new options -validity_period and -status_age added to
2837 *) If signature or public key algorithm is unrecognized print out its
2838 OID rather that just UNKNOWN.
2841 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2842 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2843 ID to be generated from the issuer certificate alone which can then be
2844 passed to OCSP_id_issuer_cmp().
2847 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2848 ASN1 modules to export functions returning ASN1_ITEM pointers
2849 instead of the ASN1_ITEM structures themselves. This adds several
2850 new macros which allow the underlying ASN1 function/structure to
2851 be accessed transparently. As a result code should not use ASN1_ITEM
2852 references directly (such as &X509_it) but instead use the relevant
2853 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2854 use of the new ASN1 code on platforms where exporting structures
2855 is problematical (for example in shared libraries) but exporting
2856 functions returning pointers to structures is not.
2859 *) Add support for overriding the generation of SSL/TLS session IDs.
2860 These callbacks can be registered either in an SSL_CTX or per SSL.
2861 The purpose of this is to allow applications to control, if they wish,
2862 the arbitrary values chosen for use as session IDs, particularly as it
2863 can be useful for session caching in multiple-server environments. A
2864 command-line switch for testing this (and any client code that wishes
2865 to use such a feature) has been added to "s_server".
2866 [Geoff Thorpe, Lutz Jaenicke]
2868 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2869 of the form '#if defined(...) || defined(...) || ...' and
2870 '#if !defined(...) && !defined(...) && ...'. This also avoids
2871 the growing number of special cases it was previously handling.
2874 *) Make all configuration macros available for application by making
2875 sure they are available in opensslconf.h, by giving them names starting
2876 with "OPENSSL_" to avoid conflicts with other packages and by making
2877 sure e_os2.h will cover all platform-specific cases together with
2879 Additionally, it is now possible to define configuration/platform-
2880 specific names (called "system identities"). In the C code, these
2881 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2882 macro with the name beginning with "OPENSSL_SYS_", which is determined
2883 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2887 *) New option -set_serial to 'req' and 'x509' this allows the serial
2888 number to use to be specified on the command line. Previously self
2889 signed certificates were hard coded with serial number 0 and the
2890 CA options of 'x509' had to use a serial number in a file which was
2894 *) New options to 'ca' utility to support V2 CRL entry extensions.
2895 Currently CRL reason, invalidity date and hold instruction are
2896 supported. Add new CRL extensions to V3 code and some new objects.
2899 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2900 disable standard block padding (aka PKCS#5 padding) in the EVP
2901 API, which was previously mandatory. This means that the data is
2902 not padded in any way and so the total length much be a multiple
2903 of the block size, otherwise an error occurs.
2906 *) Initial (incomplete) OCSP SSL support.
2909 *) New function OCSP_parse_url(). This splits up a URL into its host,
2910 port and path components: primarily to parse OCSP URLs. New -url
2911 option to ocsp utility.
2914 *) New nonce behavior. The return value of OCSP_check_nonce() now
2915 reflects the various checks performed. Applications can decide
2916 whether to tolerate certain situations such as an absent nonce
2917 in a response when one was present in a request: the ocsp application
2918 just prints out a warning. New function OCSP_add1_basic_nonce()
2919 this is to allow responders to include a nonce in a response even if
2920 the request is nonce-less.
2923 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2924 skipped when using openssl x509 multiple times on a single input file,
2925 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2928 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2929 set string type: to handle setting ASN1_TIME structures. Fix ca
2930 utility to correctly initialize revocation date of CRLs.
2933 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2934 the clients preferred ciphersuites and rather use its own preferences.
2935 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2936 Internet Explorer by ensuring unchanged hash method during stepup.
2937 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2940 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2941 to aes and add a new 'exist' option to print out symbols that don't
2945 *) Additional options to ocsp utility to allow flags to be set and
2946 additional certificates supplied.
2949 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2950 OCSP client a number of certificate to only verify the response
2954 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2955 handle the new API. Currently only ECB, CBC modes supported. Add new
2958 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2959 Encryption Standard (AES) Ciphersuites for Transport Layer
2960 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2961 not enabled by default and were not part of the "ALL" ciphersuite
2962 alias because they were not yet official; they could be
2963 explicitly requested by specifying the "AESdraft" ciphersuite
2964 group alias. In the final release of OpenSSL 0.9.7, the group
2965 alias is called "AES" and is part of "ALL".)
2966 [Ben Laurie, Steve Henson, Bodo Moeller]
2968 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2969 request to response.
2972 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2973 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2974 extract information from a certificate request. OCSP_response_create()
2975 creates a response and optionally adds a basic response structure.
2976 OCSP_basic_add1_status() adds a complete single response to a basic
2977 response and returns the OCSP_SINGLERESP structure just added (to allow
2978 extensions to be included for example). OCSP_basic_add1_cert() adds a
2979 certificate to a basic response and OCSP_basic_sign() signs a basic
2980 response with various flags. New helper functions ASN1_TIME_check()
2981 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2982 (converts ASN1_TIME to GeneralizedTime).
2985 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2986 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2987 structure from a certificate. X509_pubkey_digest() digests the public_key
2988 contents: this is used in various key identifiers.
2991 *) Make sk_sort() tolerate a NULL argument.
2992 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2994 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2995 passed by the function are trusted implicitly. If any of them signed the
2996 response then it is assumed to be valid and is not verified.
2999 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3000 to data. This was previously part of the PKCS7 ASN1 code. This
3001 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3002 [Steve Henson, reported by Kenneth R. Robinette
3003 <support@securenetterm.com>]
3005 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3006 routines: without these tracing memory leaks is very painful.
3007 Fix leaks in PKCS12 and PKCS7 routines.
3010 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3011 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3012 effectively meant GeneralizedTime would never be used. Now it
3013 is initialised to -1 but X509_time_adj() now has to check the value
3014 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3015 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3016 [Steve Henson, reported by Kenneth R. Robinette
3017 <support@securenetterm.com>]
3019 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3020 result in a zero length in the ASN1_INTEGER structure which was
3021 not consistent with the structure when d2i_ASN1_INTEGER() was used
3022 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3023 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3024 where it did not print out a minus for negative ASN1_INTEGER.
3027 *) Add summary printout to ocsp utility. The various functions which
3028 convert status values to strings have been renamed to:
3029 OCSP_response_status_str(), OCSP_cert_status_str() and
3030 OCSP_crl_reason_str() and are no longer static. New options
3031 to verify nonce values and to disable verification. OCSP response
3032 printout format cleaned up.
3035 *) Add additional OCSP certificate checks. These are those specified
3036 in RFC2560. This consists of two separate checks: the CA of the
3037 certificate being checked must either be the OCSP signer certificate
3038 or the issuer of the OCSP signer certificate. In the latter case the
3039 OCSP signer certificate must contain the OCSP signing extended key
3040 usage. This check is performed by attempting to match the OCSP
3041 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3042 in the OCSP_CERTID structures of the response.
3045 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3046 and related routines. This uses the standard OpenSSL certificate
3047 verify routines to perform initial checks (just CA validity) and
3048 to obtain the certificate chain. Then additional checks will be
3049 performed on the chain. Currently the root CA is checked to see
3050 if it is explicitly trusted for OCSP signing. This is used to set
3051 a root CA as a global signing root: that is any certificate that
3052 chains to that CA is an acceptable OCSP signing certificate.
3055 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3056 extensions from a separate configuration file.
3057 As when reading extensions from the main configuration file,
3058 the '-extensions ...' option may be used for specifying the
3060 [Massimiliano Pala <madwolf@comune.modena.it>]
3062 *) New OCSP utility. Allows OCSP requests to be generated or
3063 read. The request can be sent to a responder and the output
3064 parsed, outputed or printed in text form. Not complete yet:
3065 still needs to check the OCSP response validity.
3068 *) New subcommands for 'openssl ca':
3069 'openssl ca -status <serial>' prints the status of the cert with
3070 the given serial number (according to the index file).
3071 'openssl ca -updatedb' updates the expiry status of certificates
3073 [Massimiliano Pala <madwolf@comune.modena.it>]
3075 *) New '-newreq-nodes' command option to CA.pl. This is like
3076 '-newreq', but calls 'openssl req' with the '-nodes' option
3077 so that the resulting key is not encrypted.
3078 [Damien Miller <djm@mindrot.org>]
3080 *) New configuration for the GNU Hurd.
3081 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3083 *) Initial code to implement OCSP basic response verify. This
3084 is currently incomplete. Currently just finds the signer's
3085 certificate and verifies the signature on the response.
3088 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3089 value of OPENSSLDIR. This is available via the new '-d' option
3090 to 'openssl version', and is also included in 'openssl version -a'.
3093 *) Allowing defining memory allocation callbacks that will be given
3094 file name and line number information in additional arguments
3095 (a const char* and an int). The basic functionality remains, as
3096 well as the original possibility to just replace malloc(),
3097 realloc() and free() by functions that do not know about these
3098 additional arguments. To register and find out the current
3099 settings for extended allocation functions, the following
3100 functions are provided:
3102 CRYPTO_set_mem_ex_functions
3103 CRYPTO_set_locked_mem_ex_functions
3104 CRYPTO_get_mem_ex_functions
3105 CRYPTO_get_locked_mem_ex_functions
3107 These work the same way as CRYPTO_set_mem_functions and friends.
3108 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3109 extended allocation function is enabled.
3110 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3111 a conventional allocation function is enabled.
3112 [Richard Levitte, Bodo Moeller]
3114 *) Finish off removing the remaining LHASH function pointer casts.
3115 There should no longer be any prototype-casting required when using
3116 the LHASH abstraction, and any casts that remain are "bugs". See
3117 the callback types and macros at the head of lhash.h for details
3118 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3121 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3122 If /dev/[u]random devices are not available or do not return enough
3123 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3125 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3126 /etc/entropy will be queried once each in this sequence, quering stops
3127 when enough entropy was collected without querying more sockets.
3130 *) Change the Unix RAND_poll() variant to be able to poll several
3131 random devices, as specified by DEVRANDOM, until a sufficient amount
3132 of data has been collected. We spend at most 10 ms on each file
3133 (select timeout) and read in non-blocking mode. DEVRANDOM now
3134 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3135 (previously it was just the string "/dev/urandom"), so on typical
3136 platforms the 10 ms delay will never occur.
3137 Also separate out the Unix variant to its own file, rand_unix.c.
3138 For VMS, there's a currently-empty rand_vms.c.
3141 *) Move OCSP client related routines to ocsp_cl.c. These
3142 provide utility functions which an application needing
3143 to issue a request to an OCSP responder and analyse the
3144 response will typically need: as opposed to those which an
3145 OCSP responder itself would need which will be added later.
3147 OCSP_request_sign() signs an OCSP request with an API similar
3148 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3149 response. OCSP_response_get1_basic() extracts basic response
3150 from response. OCSP_resp_find_status(): finds and extracts status
3151 information from an OCSP_CERTID structure (which will be created
3152 when the request structure is built). These are built from lower
3153 level functions which work on OCSP_SINGLERESP structures but
3154 wont normally be used unless the application wishes to examine
3155 extensions in the OCSP response for example.
3157 Replace nonce routines with a pair of functions.
3158 OCSP_request_add1_nonce() adds a nonce value and optionally
3159 generates a random value. OCSP_check_nonce() checks the
3160 validity of the nonce in an OCSP response.
3163 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3164 This doesn't copy the supplied OCSP_CERTID and avoids the
3165 need to free up the newly created id. Change return type
3166 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3167 This can then be used to add extensions to the request.
3168 Deleted OCSP_request_new(), since most of its functionality
3169 is now in OCSP_REQUEST_new() (and the case insensitive name
3170 clash) apart from the ability to set the request name which
3171 will be added elsewhere.
3174 *) Update OCSP API. Remove obsolete extensions argument from
3175 various functions. Extensions are now handled using the new
3176 OCSP extension code. New simple OCSP HTTP function which
3177 can be used to send requests and parse the response.
3180 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3181 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3182 uses the special reorder version of SET OF to sort the attributes
3183 and reorder them to match the encoded order. This resolves a long
3184 standing problem: a verify on a PKCS7 structure just after signing
3185 it used to fail because the attribute order did not match the
3186 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3187 it uses the received order. This is necessary to tolerate some broken
3188 software that does not order SET OF. This is handled by encoding
3189 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3190 to produce the required SET OF.
3193 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3194 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3195 files to get correct declarations of the ASN.1 item variables.
3198 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3199 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3200 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3201 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3202 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3203 ASN1_ITEM and no wrapper functions.
3206 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3207 replace the old function pointer based I/O routines. Change most of
3208 the *_d2i_bio() and *_d2i_fp() functions to use these.
3211 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3212 lines, recognice more "algorithms" that can be deselected, and make
3213 it complain about algorithm deselection that isn't recognised.
3216 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3217 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3218 to use new functions. Add NO_ASN1_OLD which can be set to remove
3219 some old style ASN1 functions: this can be used to determine if old
3220 code will still work when these eventually go away.
3223 *) New extension functions for OCSP structures, these follow the
3224 same conventions as certificates and CRLs.
3227 *) New function X509V3_add1_i2d(). This automatically encodes and
3228 adds an extension. Its behaviour can be customised with various
3229 flags to append, replace or delete. Various wrappers added for
3230 certifcates and CRLs.
3233 *) Fix to avoid calling the underlying ASN1 print routine when
3234 an extension cannot be parsed. Correct a typo in the
3235 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3238 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3239 entries for variables.
3242 *) Add functionality to apps/openssl.c for detecting locking
3243 problems: As the program is single-threaded, all we have
3244 to do is register a locking callback using an array for
3245 storing which locks are currently held by the program.
3248 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3249 SSL_get_ex_data_X509_STORE_idx(), which is used in
3250 ssl_verify_cert_chain() and thus can be called at any time
3251 during TLS/SSL handshakes so that thread-safety is essential.
3252 Unfortunately, the ex_data design is not at all suited
3253 for multi-threaded use, so it probably should be abolished.
3256 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3257 [Broadcom, tweaked and integrated by Geoff Thorpe]
3259 *) Move common extension printing code to new function
3260 X509V3_print_extensions(). Reorganise OCSP print routines and
3261 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3264 *) New function X509_signature_print() to remove duplication in some
3268 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3269 set (this was treated exactly the same as SET OF previously). This
3270 is used to reorder the STACK representing the structure to match the
3271 encoding. This will be used to get round a problem where a PKCS7
3272 structure which was signed could not be verified because the STACK
3273 order did not reflect the encoded order.
3276 *) Reimplement the OCSP ASN1 module using the new code.
3279 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3280 for its ASN1 operations. The old style function pointers still exist
3281 for now but they will eventually go away.
3284 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3285 completely replaces the old ASN1 functionality with a table driven
3286 encoder and decoder which interprets an ASN1_ITEM structure describing
3287 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3288 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3289 has also been converted to the new form.
3292 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3293 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3294 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3295 for negative moduli.
3298 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3299 of not touching the result's sign bit.
3302 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3306 *) Changed the LHASH code to use prototypes for callbacks, and created
3307 macros to declare and implement thin (optionally static) functions
3308 that provide type-safety and avoid function pointer casting for the
3309 type-specific callbacks.
3312 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3314 [Veers Staats <staatsvr@asc.hpc.mil>,
3315 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3317 *) Reformat the FAQ so the different questions and answers can be divided
3318 in sections depending on the subject.
3321 *) Have the zlib compression code load ZLIB.DLL dynamically under
3325 *) New function BN_mod_sqrt for computing square roots modulo a prime
3326 (using the probabilistic Tonelli-Shanks algorithm unless
3327 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3328 be handled deterministically).
3329 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3331 *) Make BN_mod_inverse faster by explicitly handling small quotients
3332 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3333 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3336 *) New function BN_kronecker.
3339 *) Fix BN_gcd so that it works on negative inputs; the result is
3340 positive unless both parameters are zero.
3341 Previously something reasonably close to an infinite loop was
3342 possible because numbers could be growing instead of shrinking
3343 in the implementation of Euclid's algorithm.
3346 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3347 sign of the number in question.
3349 Fix BN_is_word(a,w) to work correctly for w == 0.
3351 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3352 because its test if the absolute value of 'a' equals 'w'.
3353 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3354 it exists mostly for use in the implementations of BN_is_zero(),
3355 BN_is_one(), and BN_is_word().
3358 *) New function BN_swap.
3361 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3362 the exponentiation functions are more likely to produce reasonable
3363 results on negative inputs.
3366 *) Change BN_mod_mul so that the result is always non-negative.
3367 Previously, it could be negative if one of the factors was negative;
3368 I don't think anyone really wanted that behaviour.
3371 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3372 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3373 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3374 and add new functions:
3383 BN_mod_lshift1_quick
3387 These functions always generate non-negative results.
3389 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3390 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3392 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3393 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3394 be reduced modulo m.
3395 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3398 The following entry accidentily appeared in the CHANGES file
3399 distributed with OpenSSL 0.9.7. The modifications described in
3400 it do *not* apply to OpenSSL 0.9.7.
3402 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3403 was actually never needed) and in BN_mul(). The removal in BN_mul()
3404 required a small change in bn_mul_part_recursive() and the addition
3405 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3406 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3407 bn_sub_words() and bn_add_words() except they take arrays with
3412 *) In 'openssl passwd', verify passwords read from the terminal
3413 unless the '-salt' option is used (which usually means that
3414 verification would just waste user's time since the resulting
3415 hash is going to be compared with some given password hash)
3416 or the new '-noverify' option is used.
3418 This is an incompatible change, but it does not affect
3419 non-interactive use of 'openssl passwd' (passwords on the command
3420 line, '-stdin' option, '-in ...' option) and thus should not
3424 *) Remove all references to RSAref, since there's no more need for it.
3427 *) Make DSO load along a path given through an environment variable
3428 (SHLIB_PATH) with shl_load().
3431 *) Constify the ENGINE code as a result of BIGNUM constification.
3432 Also constify the RSA code and most things related to it. In a
3433 few places, most notable in the depth of the ASN.1 code, ugly
3434 casts back to non-const were required (to be solved at a later
3438 *) Make it so the openssl application has all engines loaded by default.
3441 *) Constify the BIGNUM routines a little more.
3444 *) Add the following functions:
3446 ENGINE_load_cswift()
3448 ENGINE_load_atalla()
3450 ENGINE_load_builtin_engines()
3452 That way, an application can itself choose if external engines that
3453 are built-in in OpenSSL shall ever be used or not. The benefit is
3454 that applications won't have to be linked with libdl or other dso
3455 libraries unless it's really needed.
3457 Changed 'openssl engine' to load all engines on demand.
3458 Changed the engine header files to avoid the duplication of some
3459 declarations (they differed!).
3462 *) 'openssl engine' can now list capabilities.
3465 *) Better error reporting in 'openssl engine'.
3468 *) Never call load_dh_param(NULL) in s_server.
3471 *) Add engine application. It can currently list engines by name and
3472 identity, and test if they are actually available.
3475 *) Improve RPM specification file by forcing symbolic linking and making
3476 sure the installed documentation is also owned by root.root.
3477 [Damien Miller <djm@mindrot.org>]
3479 *) Give the OpenSSL applications more possibilities to make use of
3480 keys (public as well as private) handled by engines.
3483 *) Add OCSP code that comes from CertCo.
3486 *) Add VMS support for the Rijndael code.
3489 *) Added untested support for Nuron crypto accelerator.
3492 *) Add support for external cryptographic devices. This code was
3493 previously distributed separately as the "engine" branch.
3494 [Geoff Thorpe, Richard Levitte]
3496 *) Rework the filename-translation in the DSO code. It is now possible to
3497 have far greater control over how a "name" is turned into a filename
3498 depending on the operating environment and any oddities about the
3499 different shared library filenames on each system.
3502 *) Support threads on FreeBSD-elf in Configure.
3505 *) Fix for SHA1 assembly problem with MASM: it produces
3506 warnings about corrupt line number information when assembling
3507 with debugging information. This is caused by the overlapping
3509 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3512 NCONF_get_number() has no error checking at all. As a replacement,
3513 NCONF_get_number_e() is defined (_e for "error checking") and is
3514 promoted strongly. The old NCONF_get_number is kept around for
3515 binary backward compatibility.
3516 Make it possible for methods to load from something other than a BIO,
3517 by providing a function pointer that is given a name instead of a BIO.
3518 For example, this could be used to load configuration data from an
3522 *) Fix for non blocking accept BIOs. Added new I/O special reason
3523 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3524 with non blocking I/O was not possible because no retry code was
3525 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3529 *) Added the beginnings of Rijndael support.
3532 *) Fix for bug in DirectoryString mask setting. Add support for
3533 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3534 to allow certificate printing to more controllable, additional
3535 'certopt' option to 'x509' to allow new printing options to be
3539 *) Clean old EAY MD5 hack from e_os.h.
3542 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3544 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3545 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3546 [Joe Orton, Steve Henson]
3548 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3550 *) Fix additional bug revealed by the NISCC test suite:
3552 Stop bug triggering large recursion when presented with
3553 certain ASN.1 tags (CVE-2003-0851)
3556 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3558 *) Fix various bugs revealed by running the NISCC test suite:
3560 Stop out of bounds reads in the ASN1 code when presented with
3561 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3563 If verify callback ignores invalid public key errors don't try to check
3564 certificate signature with the NULL public key.
3568 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3569 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3573 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3574 extra data after the compression methods not only for TLS 1.0
3575 but also for SSL 3.0 (as required by the specification).
3576 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3578 *) Change X509_certificate_type() to mark the key as exported/exportable
3579 when it's 512 *bits* long, not 512 bytes.
3582 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3584 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3585 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3586 a protocol version number mismatch like a decryption error
3587 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3590 *) Turn on RSA blinding by default in the default implementation
3591 to avoid a timing attack. Applications that don't want it can call
3592 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3593 They would be ill-advised to do so in most cases.
3594 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3596 *) Change RSA blinding code so that it works when the PRNG is not
3597 seeded (in this case, the secret RSA exponent is abused as
3598 an unpredictable seed -- if it is not unpredictable, there
3599 is no point in blinding anyway). Make RSA blinding thread-safe
3600 by remembering the creator's thread ID in rsa->blinding and
3601 having all other threads use local one-time blinding factors
3602 (this requires more computation than sharing rsa->blinding, but
3603 avoids excessive locking; and if an RSA object is not shared
3604 between threads, blinding will still be very fast).
3607 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3609 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3610 via timing by performing a MAC computation even if incorrrect
3611 block cipher padding has been found. This is a countermeasure
3612 against active attacks where the attacker has to distinguish
3613 between bad padding and a MAC verification error. (CVE-2003-0078)
3615 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3616 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3617 Martin Vuagnoux (EPFL, Ilion)]
3619 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3621 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3622 memory from it's contents. This is done with a counter that will
3623 place alternating values in each byte. This can be used to solve
3624 two issues: 1) the removal of calls to memset() by highly optimizing
3625 compilers, and 2) cleansing with other values than 0, since those can
3626 be read through on certain media, for example a swap space on disk.
3629 *) Bugfix: client side session caching did not work with external caching,
3630 because the session->cipher setting was not restored when reloading
3631 from the external cache. This problem was masked, when
3632 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3633 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3636 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3637 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3638 [Zeev Lieber <zeev-l@yahoo.com>]
3640 *) Undo an undocumented change introduced in 0.9.6e which caused
3641 repeated calls to OpenSSL_add_all_ciphers() and
3642 OpenSSL_add_all_digests() to be ignored, even after calling
3646 *) Change the default configuration reader to deal with last line not
3647 being properly terminated.
3650 *) Change X509_NAME_cmp() so it applies the special rules on handling
3651 DN values that are of type PrintableString, as well as RDNs of type
3652 emailAddress where the value has the type ia5String.
3653 [stefank@valicert.com via Richard Levitte]
3655 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3656 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3657 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3658 the bitwise-OR of the two for use by the majority of applications
3659 wanting this behaviour, and update the docs. The documented
3660 behaviour and actual behaviour were inconsistent and had been
3661 changing anyway, so this is more a bug-fix than a behavioural
3663 [Geoff Thorpe, diagnosed by Nadav Har'El]
3665 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3666 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3669 *) Fix initialization code race conditions in
3670 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3671 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3672 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3673 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3674 ssl2_get_cipher_by_char(),
3675 ssl3_get_cipher_by_char().
3676 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3678 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3679 the cached sessions are flushed, as the remove_cb() might use ex_data
3680 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3681 (see [openssl.org #212]).
3682 [Geoff Thorpe, Lutz Jaenicke]
3684 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3685 length, instead of the encoding length to d2i_ASN1_OBJECT.
3688 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3690 *) [In 0.9.6g-engine release:]
3691 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3692 [Lynn Gazis <lgazis@rainbow.com>]
3694 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3696 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3697 and get fix the header length calculation.
3698 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3699 Alon Kantor <alonk@checkpoint.com> (and others),
3702 *) Use proper error handling instead of 'assertions' in buffer
3703 overflow checks added in 0.9.6e. This prevents DoS (the
3704 assertions could call abort()).
3705 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3707 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3709 *) Add various sanity checks to asn1_get_length() to reject
3710 the ASN1 length bytes if they exceed sizeof(long), will appear
3711 negative or the content length exceeds the length of the
3713 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3715 *) Fix cipher selection routines: ciphers without encryption had no flags
3716 for the cipher strength set and where therefore not handled correctly
3717 by the selection routines (PR #130).
3720 *) Fix EVP_dsa_sha macro.
3724 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3725 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3726 that was added in OpenSSL 0.9.6d.
3728 As the countermeasure turned out to be incompatible with some
3729 broken SSL implementations, the new option is part of SSL_OP_ALL.
3730 SSL_OP_ALL is usually employed when compatibility with weird SSL
3731 implementations is desired (e.g. '-bugs' option to 's_client' and
3732 's_server'), so the new option is automatically set in many
3736 *) Changes in security patch:
3738 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3739 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3740 Air Force Materiel Command, USAF, under agreement number
3743 *) Add various sanity checks to asn1_get_length() to reject
3744 the ASN1 length bytes if they exceed sizeof(long), will appear
3745 negative or the content length exceeds the length of the
3746 supplied buffer. (CVE-2002-0659)
3747 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3749 *) Assertions for various potential buffer overflows, not known to
3751 [Ben Laurie (CHATS)]
3753 *) Various temporary buffers to hold ASCII versions of integers were
3754 too small for 64 bit platforms. (CVE-2002-0655)
3755 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3757 *) Remote buffer overflow in SSL3 protocol - an attacker could
3758 supply an oversized session ID to a client. (CVE-2002-0656)
3759 [Ben Laurie (CHATS)]
3761 *) Remote buffer overflow in SSL2 protocol - an attacker could
3762 supply an oversized client master key. (CVE-2002-0656)
3763 [Ben Laurie (CHATS)]
3765 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3767 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3768 encoded as NULL) with id-dsa-with-sha1.
3769 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3771 *) Check various X509_...() return values in apps/req.c.
3772 [Nils Larsch <nla@trustcenter.de>]
3774 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3775 an end-of-file condition would erronously be flagged, when the CRLF
3776 was just at the end of a processed block. The bug was discovered when
3777 processing data through a buffering memory BIO handing the data to a
3778 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3779 <ptsekov@syntrex.com> and Nedelcho Stanev.
3782 *) Implement a countermeasure against a vulnerability recently found
3783 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3784 before application data chunks to avoid the use of known IVs
3785 with data potentially chosen by the attacker.
3788 *) Fix length checks in ssl3_get_client_hello().
3791 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3792 to prevent ssl3_read_internal() from incorrectly assuming that
3793 ssl3_read_bytes() found application data while handshake
3794 processing was enabled when in fact s->s3->in_read_app_data was
3795 merely automatically cleared during the initial handshake.
3796 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3798 *) Fix object definitions for Private and Enterprise: they were not
3799 recognized in their shortname (=lowercase) representation. Extend
3800 obj_dat.pl to issue an error when using undefined keywords instead
3801 of silently ignoring the problem (Svenning Sorensen
3802 <sss@sss.dnsalias.net>).
3805 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3806 generators, i.e. generators other than 2 and 5. (Previously, the
3807 code did not properly initialise the 'add' and 'rem' values to
3808 BN_generate_prime().)
3810 In the new general case, we do not insist that 'generator' is
3811 actually a primitive root: This requirement is rather pointless;
3812 a generator of the order-q subgroup is just as good, if not
3816 *) Map new X509 verification errors to alerts. Discovered and submitted by
3817 Tom Wu <tom@arcot.com>.
3820 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3821 returning non-zero before the data has been completely received
3822 when using non-blocking I/O.
3823 [Bodo Moeller; problem pointed out by John Hughes]
3825 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3826 [Ben Laurie, Lutz Jaenicke]
3828 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3829 Yoram Zahavi <YoramZ@gilian.com>).
3832 *) Add information about CygWin 1.3 and on, and preserve proper
3833 configuration for the versions before that.
3834 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3836 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3837 check whether we deal with a copy of a session and do not delete from
3838 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3839 <izhar@checkpoint.com>.
3842 *) Do not store session data into the internal session cache, if it
3843 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3844 flag is set). Proposed by Aslam <aslam@funk.com>.
3847 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3851 *) [In 0.9.6d-engine release:]
3852 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3853 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3855 *) Add the configuration target linux-s390x.
3856 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3858 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3859 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3860 variable as an indication that a ClientHello message has been
3861 received. As the flag value will be lost between multiple
3862 invocations of ssl3_accept when using non-blocking I/O, the
3863 function may not be aware that a handshake has actually taken
3864 place, thus preventing a new session from being added to the
3867 To avoid this problem, we now set s->new_session to 2 instead of
3868 using a local variable.
3869 [Lutz Jaenicke, Bodo Moeller]
3871 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3872 if the SSL_R_LENGTH_MISMATCH error is detected.
3873 [Geoff Thorpe, Bodo Moeller]
3875 *) New 'shared_ldflag' column in Configure platform table.
3878 *) Fix EVP_CIPHER_mode macro.
3879 ["Dan S. Camper" <dan@bti.net>]
3881 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3882 type, we must throw them away by setting rr->length to 0.
3883 [D P Chang <dpc@qualys.com>]
3885 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3887 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3888 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3889 worked incorrectly for those cases where range = 10..._2 and
3890 3*range is two bits longer than range.)
3893 *) Only add signing time to PKCS7 structures if it is not already
3897 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3898 OBJ_ld_ce should be OBJ_id_ce.
3899 Also some ip-pda OIDs in crypto/objects/objects.txt were
3900 incorrect (cf. RFC 3039).
3901 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3903 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3904 returns early because it has nothing to do.
3905 [Andy Schneider <andy.schneider@bjss.co.uk>]
3907 *) [In 0.9.6c-engine release:]
3908 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3909 [Andy Schneider <andy.schneider@bjss.co.uk>]
3911 *) [In 0.9.6c-engine release:]
3912 Add support for Cryptographic Appliance's keyserver technology.
3913 (Use engine 'keyclient')
3914 [Cryptographic Appliances and Geoff Thorpe]
3916 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3917 is called via tools/c89.sh because arguments have to be
3918 rearranged (all '-L' options must appear before the first object
3920 [Richard Shapiro <rshapiro@abinitio.com>]
3922 *) [In 0.9.6c-engine release:]
3923 Add support for Broadcom crypto accelerator cards, backported
3925 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3927 *) [In 0.9.6c-engine release:]
3928 Add support for SureWare crypto accelerator cards from
3929 Baltimore Technologies. (Use engine 'sureware')
3930 [Baltimore Technologies and Mark Cox]
3932 *) [In 0.9.6c-engine release:]
3933 Add support for crypto accelerator cards from Accelerated
3934 Encryption Processing, www.aep.ie. (Use engine 'aep')
3935 [AEP Inc. and Mark Cox]
3937 *) Add a configuration entry for gcc on UnixWare.
3938 [Gary Benson <gbenson@redhat.com>]
3940 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3941 messages are stored in a single piece (fixed-length part and
3942 variable-length part combined) and fix various bugs found on the way.
3945 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3946 instead. BIO_gethostbyname() does not know what timeouts are
3947 appropriate, so entries would stay in cache even when they have
3949 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3951 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3952 faced with a pathologically small ClientHello fragment that does
3953 not contain client_version: Instead of aborting with an error,
3954 simply choose the highest available protocol version (i.e.,
3955 TLS 1.0 unless it is disabled). In practice, ClientHello
3956 messages are never sent like this, but this change gives us
3957 strictly correct behaviour at least for TLS.
3960 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3961 never resets s->method to s->ctx->method when called from within
3962 one of the SSL handshake functions.
3963 [Bodo Moeller; problem pointed out by Niko Baric]
3965 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3966 (sent using the client's version number) if client_version is
3967 smaller than the protocol version in use. Also change
3968 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3969 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3970 the client will at least see that alert.
3973 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3977 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3978 client receives HelloRequest while in a handshake.
3979 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3981 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3982 should end in 'break', not 'goto end' which circuments various
3983 cleanups done in state SSL_ST_OK. But session related stuff
3984 must be disabled for SSL_ST_OK in the case that we just sent a
3987 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3988 before just sending a HelloRequest.
3989 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3991 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3992 reveal whether illegal block cipher padding was found or a MAC
3993 verification error occured. (Neither SSLerr() codes nor alerts
3994 are directly visible to potential attackers, but the information
3995 may leak via logfiles.)
3997 Similar changes are not required for the SSL 2.0 implementation
3998 because the number of padding bytes is sent in clear for SSL 2.0,
3999 and the extra bytes are just ignored. However ssl/s2_pkt.c
4000 failed to verify that the purported number of padding bytes is in
4004 *) Add OpenUNIX-8 support including shared libraries
4005 (Boyd Lynn Gerber <gerberb@zenez.com>).
4008 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4009 'wristwatch attack' using huge encoding parameters (cf.
4010 James H. Manger's CRYPTO 2001 paper). Note that the
4011 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4012 encoding parameters and hence was not vulnerable.
4015 *) BN_sqr() bug fix.
4016 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4018 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4019 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4020 followed by modular reduction.
4021 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4023 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4024 equivalent based on BN_pseudo_rand() instead of BN_rand().
4027 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4028 This function was broken, as the check for a new client hello message
4029 to handle SGC did not allow these large messages.
4030 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4033 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4036 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4037 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4040 *) Rework the configuration and shared library support for Tru64 Unix.
4041 The configuration part makes use of modern compiler features and
4042 still retains old compiler behavior for those that run older versions
4043 of the OS. The shared library support part includes a variant that
4044 uses the RPATH feature, and is available through the special
4045 configuration target "alpha-cc-rpath", which will never be selected
4047 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4049 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4050 with the same message size as in ssl3_get_certificate_request().
4051 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4052 messages might inadvertently be reject as too long.
4053 [Petr Lampa <lampa@fee.vutbr.cz>]
4055 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4058 *) Modified SSL library such that the verify_callback that has been set
4059 specificly for an SSL object with SSL_set_verify() is actually being
4060 used. Before the change, a verify_callback set with this function was
4061 ignored and the verify_callback() set in the SSL_CTX at the time of
4062 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4063 to allow the necessary settings.
4066 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4067 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4068 done automatically (in contradiction to the requirements of the C
4069 standard). This made problems when used from OpenSSH.
4072 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4073 dh->length and always used
4075 BN_rand_range(priv_key, dh->p).
4077 BN_rand_range() is not necessary for Diffie-Hellman, and this
4078 specific range makes Diffie-Hellman unnecessarily inefficient if
4079 dh->length (recommended exponent length) is much smaller than the
4080 length of dh->p. We could use BN_rand_range() if the order of
4081 the subgroup was stored in the DH structure, but we only have
4086 BN_rand(priv_key, l, ...)
4088 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4094 RSA_eay_public_encrypt
4095 RSA_eay_private_decrypt
4096 RSA_eay_private_encrypt (signing)
4097 RSA_eay_public_decrypt (signature verification)
4099 (default implementations for RSA_public_encrypt,
4100 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4101 always reject numbers >= n.
4104 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4105 to synchronize access to 'locking_thread'. This is necessary on
4106 systems where access to 'locking_thread' (an 'unsigned long'
4107 variable) is not atomic.
4110 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4111 *before* setting the 'crypto_lock_rand' flag. The previous code had
4112 a race condition if 0 is a valid thread ID.
4113 [Travis Vitek <vitek@roguewave.com>]
4115 *) Add support for shared libraries under Irix.
4116 [Albert Chin-A-Young <china@thewrittenword.com>]
4118 *) Add configuration option to build on Linux on both big-endian and
4120 [Ralf Baechle <ralf@uni-koblenz.de>]
4122 *) Add the possibility to create shared libraries on HP-UX.
4125 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4127 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4128 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4129 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4130 PRNG state recovery was possible based on the output of
4131 one PRNG request appropriately sized to gain knowledge on
4132 'md' followed by enough consecutive 1-byte PRNG requests
4133 to traverse all of 'state'.
4135 1. When updating 'md_local' (the current thread's copy of 'md')
4136 during PRNG output generation, hash all of the previous
4137 'md_local' value, not just the half used for PRNG output.
4139 2. Make the number of bytes from 'state' included into the hash
4140 independent from the number of PRNG bytes requested.
4142 The first measure alone would be sufficient to avoid
4143 Markku-Juhani's attack. (Actually it had never occurred
4144 to me that the half of 'md_local' used for chaining was the
4145 half from which PRNG output bytes were taken -- I had always
4146 assumed that the secret half would be used.) The second
4147 measure makes sure that additional data from 'state' is never
4148 mixed into 'md_local' in small portions; this heuristically
4149 further strengthens the PRNG.
4152 *) Fix crypto/bn/asm/mips3.s.
4155 *) When only the key is given to "enc", the IV is undefined. Print out
4156 an error message in this case.
4159 *) Handle special case when X509_NAME is empty in X509 printing routines.
4162 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4163 positive and less than q.
4166 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4167 used: it isn't thread safe and the add_lock_callback should handle
4169 [Paul Rose <Paul.Rose@bridge.com>]
4171 *) Verify that incoming data obeys the block size in
4172 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4176 [Ulf Möller, Bodo Möller]
4178 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4179 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4180 when fixing the server behaviour for backwards-compatible 'client
4181 hello' messages. (Note that the attack is impractical against
4182 SSL 3.0 and TLS 1.0 anyway because length and version checking
4183 means that the probability of guessing a valid ciphertext is
4184 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4187 Before 0.9.5, the countermeasure (hide the error by generating a
4188 random 'decryption result') did not work properly because
4189 ERR_clear_error() was missing, meaning that SSL_get_error() would
4190 detect the supposedly ignored error.
4192 Both problems are now fixed.
4195 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4196 (previously it was 1024).
4199 *) Fix for compatibility mode trust settings: ignore trust settings
4200 unless some valid trust or reject settings are present.
4203 *) Fix for blowfish EVP: its a variable length cipher.
4206 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4207 parameters in DSA public key structures and return an error in the
4208 DSA routines if parameters are absent.
4211 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4212 in the current directory if neither $RANDFILE nor $HOME was set.
4213 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4214 caused some confusion to Windows users who haven't defined $HOME.
4215 Thus RAND_file_name() is changed again: e_os.h can define a
4216 DEFAULT_HOME, which will be used if $HOME is not set.
4217 For Windows, we use "C:"; on other platforms, we still require
4218 environment variables.
4220 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4221 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4222 having multiple threads call RAND_poll() concurrently.
4225 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4226 combination of a flag and a thread ID variable.
4227 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4228 flag), *other* threads can enter ssleay_add_bytes without obeying
4229 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4230 that they do not hold after the first thread unsets add_do_not_lock).
4233 *) Change bctest again: '-x' expressions are not available in all
4237 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4239 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4240 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4242 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4243 the default extension for executables, if any. Also, make the perl
4244 scripts that use symlink() to test if it really exists and use "cp"
4245 if it doesn't. All this made OpenSSL compilable and installable in
4249 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4250 If SEQUENCE is length is indefinite just set c->slen to the total
4251 amount of data available.
4252 [Steve Henson, reported by shige@FreeBSD.org]
4253 [This change does not apply to 0.9.7.]
4255 *) Change bctest to avoid here-documents inside command substitution
4256 (workaround for FreeBSD /bin/sh bug).
4257 For compatibility with Ultrix, avoid shell functions (introduced
4258 in the bctest version that searches along $PATH).
4261 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4262 with des_encrypt() defined on some operating systems, like Solaris
4266 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4267 On the Importance of Eliminating Errors in Cryptographic
4268 Computations, J. Cryptology 14 (2001) 2, 101-119,
4269 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4272 *) MIPS assembler BIGNUM division bug fix.
4275 *) Disabled incorrect Alpha assembler code.
4278 *) Fix PKCS#7 decode routines so they correctly update the length
4279 after reading an EOC for the EXPLICIT tag.
4281 [This change does not apply to 0.9.7.]
4283 *) Fix bug in PKCS#12 key generation routines. This was triggered
4284 if a 3DES key was generated with a 0 initial byte. Include
4285 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4286 (but broken) behaviour.
4289 *) Enhance bctest to search for a working bc along $PATH and print
4291 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4293 *) Fix memory leaks in err.c: free err_data string if necessary;
4294 don't write to the wrong index in ERR_set_error_data.
4297 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4301 *) Replace rdtsc with _emit statements for VC++ version 5.
4302 [Jeremy Cooper <jeremy@baymoo.org>]
4304 *) Make it possible to reuse SSLv2 sessions.
4307 *) In copy_email() check for >= 0 as a return value for
4308 X509_NAME_get_index_by_NID() since 0 is a valid index.
4309 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4311 *) Avoid coredump with unsupported or invalid public keys by checking if
4312 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4313 PKCS7_verify() fails with non detached data.
4316 *) Don't use getenv in library functions when run as setuid/setgid.
4317 New function OPENSSL_issetugid().
4320 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4321 due to incorrect handling of multi-threading:
4323 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4325 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4327 3. Count how many times MemCheck_off() has been called so that
4328 nested use can be treated correctly. This also avoids
4329 inband-signalling in the previous code (which relied on the
4330 assumption that thread ID 0 is impossible).
4333 *) Add "-rand" option also to s_client and s_server.
4336 *) Fix CPU detection on Irix 6.x.
4337 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4338 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4340 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4343 [This change does not apply to 0.9.7.]
4345 *) Use the cached encoding of an X509_NAME structure rather than
4346 copying it. This is apparently the reason for the libsafe "errors"
4347 but the code is actually correct.
4350 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4351 Bleichenbacher's DSA attack.
4352 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4353 to be set and top=0 forces the highest bit to be set; top=-1 is new
4354 and leaves the highest bit random.
4355 [Ulf Moeller, Bodo Moeller]
4357 *) In the NCONF_...-based implementations for CONF_... queries
4358 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4359 a temporary CONF structure with the data component set to NULL
4360 (which gives segmentation faults in lh_retrieve).
4361 Instead, use NULL for the CONF pointer in CONF_get_string and
4362 CONF_get_number (which may use environment variables) and directly
4363 return NULL from CONF_get_section.
4366 *) Fix potential buffer overrun for EBCDIC.
4369 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4370 keyUsage if basicConstraints absent for a CA.
4373 *) Make SMIME_write_PKCS7() write mail header values with a format that
4374 is more generally accepted (no spaces before the semicolon), since
4375 some programs can't parse those values properly otherwise. Also make
4376 sure BIO's that break lines after each write do not create invalid
4380 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4381 macros previously used would not encode an empty SEQUENCE OF
4382 and break the signature.
4384 [This change does not apply to 0.9.7.]
4386 *) Zero the premaster secret after deriving the master secret in
4390 *) Add some EVP_add_digest_alias registrations (as found in
4391 OpenSSL_add_all_digests()) to SSL_library_init()
4392 aka OpenSSL_add_ssl_algorithms(). This provides improved
4393 compatibility with peers using X.509 certificates
4394 with unconventional AlgorithmIdentifier OIDs.
4397 *) Fix for Irix with NO_ASM.
4398 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4400 *) ./config script fixes.
4401 [Ulf Moeller, Richard Levitte]
4403 *) Fix 'openssl passwd -1'.
4406 *) Change PKCS12_key_gen_asc() so it can cope with non null
4407 terminated strings whose length is passed in the passlen
4408 parameter, for example from PEM callbacks. This was done
4409 by adding an extra length parameter to asc2uni().
4410 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4412 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4413 call failed, free the DSA structure.
4416 *) Fix to uni2asc() to cope with zero length Unicode strings.
4417 These are present in some PKCS#12 files.
4420 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4421 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4422 when writing a 32767 byte record.
4423 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4425 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4426 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4428 (RSA objects have a reference count access to which is protected
4429 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4430 so they are meant to be shared between threads.)
4431 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4432 "Reddie, Steven" <Steven.Reddie@ca.com>]
4434 *) Fix a deadlock in CRYPTO_mem_leaks().
4437 *) Use better test patterns in bntest.
4440 *) rand_win.c fix for Borland C.
4443 *) BN_rshift bugfix for n == 0.
4446 *) Add a 'bctest' script that checks for some known 'bc' bugs
4447 so that 'make test' does not abort just because 'bc' is broken.
4450 *) Store verify_result within SSL_SESSION also for client side to
4451 avoid potential security hole. (Re-used sessions on the client side
4452 always resulted in verify_result==X509_V_OK, not using the original
4453 result of the server certificate verification.)
4456 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4457 SSL3_RT_APPLICATION_DATA, return 0.
4458 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4462 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4463 releases, have been re-implemented by renaming the previous
4464 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4465 and ssl3_read_internal, respectively, and adding 'peek' parameters
4466 to them. The new ssl[23]_{read,peek} functions are calls to
4467 ssl[23]_read_internal with the 'peek' flag set appropriately.
4468 A 'peek' parameter has also been added to ssl3_read_bytes, which
4469 does the actual work for ssl3_read_internal.
4472 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4473 the method-specific "init()" handler. Also clean up ex_data after
4474 calling the method-specific "finish()" handler. Previously, this was
4475 happening the other way round.
4478 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4479 The previous value, 12, was not always sufficient for BN_mod_exp().
4482 *) Make sure that shared libraries get the internal name engine with
4483 the full version number and not just 0. This should mark the
4484 shared libraries as not backward compatible. Of course, this should
4485 be changed again when we can guarantee backward binary compatibility.
4488 *) Fix typo in get_cert_by_subject() in by_dir.c
4489 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4491 *) Rework the system to generate shared libraries:
4493 - Make note of the expected extension for the shared libraries and
4494 if there is a need for symbolic links from for example libcrypto.so.0
4495 to libcrypto.so.0.9.7. There is extended info in Configure for
4498 - Make as few rebuilds of the shared libraries as possible.
4500 - Still avoid linking the OpenSSL programs with the shared libraries.
4502 - When installing, install the shared libraries separately from the
4506 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4508 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4509 and not in SSL_clear because the latter is also used by the
4510 accept/connect functions; previously, the settings made by
4511 SSL_set_read_ahead would be lost during the handshake.
4512 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4514 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4515 Previously, it would create entries for disableed algorithms no
4519 *) Added several new manual pages for SSL_* function.
4522 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4524 *) In ssl23_get_client_hello, generate an error message when faced
4525 with an initial SSL 3.0/TLS record that is too small to contain the
4526 first two bytes of the ClientHello message, i.e. client_version.
4527 (Note that this is a pathologic case that probably has never happened
4528 in real life.) The previous approach was to use the version number
4529 from the record header as a substitute; but our protocol choice
4530 should not depend on that one because it is not authenticated
4531 by the Finished messages.
4534 *) More robust randomness gathering functions for Windows.
4535 [Jeffrey Altman <jaltman@columbia.edu>]
4537 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4538 not set then we don't setup the error code for issuer check errors
4539 to avoid possibly overwriting other errors which the callback does
4540 handle. If an application does set the flag then we assume it knows
4541 what it is doing and can handle the new informational codes
4545 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4546 a general "ANY" type, as such it should be able to decode anything
4547 including tagged types. However it didn't check the class so it would
4548 wrongly interpret tagged types in the same way as their universal
4549 counterpart and unknown types were just rejected. Changed so that the
4550 tagged and unknown types are handled in the same way as a SEQUENCE:
4551 that is the encoding is stored intact. There is also a new type
4552 "V_ASN1_OTHER" which is used when the class is not universal, in this
4553 case we have no idea what the actual type is so we just lump them all
4557 *) On VMS, stdout may very well lead to a file that is written to
4558 in a record-oriented fashion. That means that every write() will
4559 write a separate record, which will be read separately by the
4560 programs trying to read from it. This can be very confusing.
4562 The solution is to put a BIO filter in the way that will buffer
4563 text until a linefeed is reached, and then write everything a
4564 line at a time, so every record written will be an actual line,
4565 not chunks of lines and not (usually doesn't happen, but I've
4566 seen it once) several lines in one record. BIO_f_linebuffer() is
4569 Currently, it's a VMS-only method, because that's where it has
4570 been tested well enough.
4573 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4574 it can return incorrect results.
4575 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4576 but it was in 0.9.6-beta[12].)
4579 *) Disable the check for content being present when verifying detached
4580 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4581 include zero length content when signing messages.
4584 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4585 BIO_ctrl (for BIO pairs).
4588 *) Add DSO method for VMS.
4591 *) Bug fix: Montgomery multiplication could produce results with the
4595 *) Add RPM specification openssl.spec and modify it to build three
4596 packages. The default package contains applications, application
4597 documentation and run-time libraries. The devel package contains
4598 include files, static libraries and function documentation. The
4599 doc package contains the contents of the doc directory. The original
4600 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4603 *) Add a large number of documentation files for many SSL routines.
4604 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4606 *) Add a configuration entry for Sony News 4.
4607 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4609 *) Don't set the two most significant bits to one when generating a
4610 random number < q in the DSA library.
4613 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4614 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4615 the underlying transport is blocking) if a handshake took place.
4616 (The default behaviour is needed by applications such as s_client
4617 and s_server that use select() to determine when to use SSL_read;
4618 but for applications that know in advance when to expect data, it
4619 just makes things more complicated.)
4622 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4626 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4627 work better on such systems.
4628 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4630 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4631 Update PKCS12_parse() so it copies the friendlyName and the
4632 keyid to the certificates aux info.
4635 *) Fix bug in PKCS7_verify() which caused an infinite loop
4636 if there was more than one signature.
4637 [Sven Uszpelkat <su@celocom.de>]
4639 *) Major change in util/mkdef.pl to include extra information
4640 about each symbol, as well as presentig variables as well
4641 as functions. This change means that there's n more need
4642 to rebuild the .num files when some algorithms are excluded.
4645 *) Allow the verify time to be set by an application,
4646 rather than always using the current time.
4649 *) Phase 2 verify code reorganisation. The certificate
4650 verify code now looks up an issuer certificate by a
4651 number of criteria: subject name, authority key id
4652 and key usage. It also verifies self signed certificates
4653 by the same criteria. The main comparison function is
4654 X509_check_issued() which performs these checks.
4656 Lot of changes were necessary in order to support this
4657 without completely rewriting the lookup code.
4659 Authority and subject key identifier are now cached.
4661 The LHASH 'certs' is X509_STORE has now been replaced
4662 by a STACK_OF(X509_OBJECT). This is mainly because an
4663 LHASH can't store or retrieve multiple objects with
4664 the same hash value.
4666 As a result various functions (which were all internal
4667 use only) have changed to handle the new X509_STORE
4668 structure. This will break anything that messed round
4669 with X509_STORE internally.
4671 The functions X509_STORE_add_cert() now checks for an
4672 exact match, rather than just subject name.
4674 The X509_STORE API doesn't directly support the retrieval
4675 of multiple certificates matching a given criteria, however
4676 this can be worked round by performing a lookup first
4677 (which will fill the cache with candidate certificates)
4678 and then examining the cache for matches. This is probably
4679 the best we can do without throwing out X509_LOOKUP
4680 entirely (maybe later...).
4682 The X509_VERIFY_CTX structure has been enhanced considerably.
4684 All certificate lookup operations now go via a get_issuer()
4685 callback. Although this currently uses an X509_STORE it
4686 can be replaced by custom lookups. This is a simple way
4687 to bypass the X509_STORE hackery necessary to make this
4688 work and makes it possible to use more efficient techniques
4689 in future. A very simple version which uses a simple
4690 STACK for its trusted certificate store is also provided
4691 using X509_STORE_CTX_trusted_stack().
4693 The verify_cb() and verify() callbacks now have equivalents
4694 in the X509_STORE_CTX structure.
4696 X509_STORE_CTX also has a 'flags' field which can be used
4697 to customise the verify behaviour.
4700 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4701 excludes S/MIME capabilities.
4704 *) When a certificate request is read in keep a copy of the
4705 original encoding of the signed data and use it when outputing
4706 again. Signatures then use the original encoding rather than
4707 a decoded, encoded version which may cause problems if the
4708 request is improperly encoded.
4711 *) For consistency with other BIO_puts implementations, call
4712 buffer_write(b, ...) directly in buffer_puts instead of calling
4715 In BIO_puts, increment b->num_write as in BIO_write.
4716 [Peter.Sylvester@EdelWeb.fr]
4718 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4719 BN_zero, we may not return a BIGNUM with an array consisting of
4723 *) Avoid calling abort() from within the library when problems are
4724 detected, except if preprocessor symbols have been defined
4725 (such as REF_CHECK, BN_DEBUG etc.).
4728 *) New openssl application 'rsautl'. This utility can be
4729 used for low level RSA operations. DER public key
4730 BIO/fp routines also added.
4733 *) New Configure entry and patches for compiling on QNX 4.
4734 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4736 *) A demo state-machine implementation was sponsored by
4737 Nuron (http://www.nuron.com/) and is now available in
4738 demos/state_machine.
4741 *) New options added to the 'dgst' utility for signature
4742 generation and verification.
4745 *) Unrecognized PKCS#7 content types are now handled via a
4746 catch all ASN1_TYPE structure. This allows unsupported
4747 types to be stored as a "blob" and an application can
4748 encode and decode it manually.
4751 *) Fix various signed/unsigned issues to make a_strex.c
4753 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4755 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4756 length if passed a buffer. ASN1_INTEGER_to_BN failed
4757 if passed a NULL BN and its argument was negative.
4758 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4760 *) Modification to PKCS#7 encoding routines to output definite
4761 length encoding. Since currently the whole structures are in
4762 memory there's not real point in using indefinite length
4763 constructed encoding. However if OpenSSL is compiled with
4764 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4767 *) Added BIO_vprintf() and BIO_vsnprintf().
4770 *) Added more prefixes to parse for in the the strings written
4771 through a logging bio, to cover all the levels that are available
4772 through syslog. The prefixes are now:
4774 PANIC, EMERG, EMR => LOG_EMERG
4775 ALERT, ALR => LOG_ALERT
4776 CRIT, CRI => LOG_CRIT
4777 ERROR, ERR => LOG_ERR
4778 WARNING, WARN, WAR => LOG_WARNING
4779 NOTICE, NOTE, NOT => LOG_NOTICE
4780 INFO, INF => LOG_INFO
4781 DEBUG, DBG => LOG_DEBUG
4783 and as before, if none of those prefixes are present at the
4784 beginning of the string, LOG_ERR is chosen.
4786 On Win32, the LOG_* levels are mapped according to this:
4788 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4789 LOG_WARNING => EVENTLOG_WARNING_TYPE
4790 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4794 *) Made it possible to reconfigure with just the configuration
4795 argument "reconf" or "reconfigure". The command line arguments
4796 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4797 and are retrieved from there when reconfiguring.
4801 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4803 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4806 *) The obj_dat.pl script was messing up the sorting of object
4807 names. The reason was that it compared the quoted version
4808 of strings as a result "OCSP" > "OCSP Signing" because
4809 " > SPACE. Changed script to store unquoted versions of
4810 names and add quotes on output. It was also omitting some
4811 names from the lookup table if they were given a default
4812 value (that is if SN is missing it is given the same
4813 value as LN and vice versa), these are now added on the
4814 grounds that if an object has a name we should be able to
4815 look it up. Finally added warning output when duplicate
4816 short or long names are found.
4819 *) Changes needed for Tandem NSK.
4820 [Scott Uroff <scott@xypro.com>]
4822 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4823 RSA_padding_check_SSLv23(), special padding was never detected
4824 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4825 version rollback attacks was not effective.
4827 In s23_clnt.c, don't use special rollback-attack detection padding
4828 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4829 client; similarly, in s23_srvr.c, don't do the rollback check if
4830 SSL 2.0 is the only protocol enabled in the server.
4833 *) Make it possible to get hexdumps of unprintable data with 'openssl
4834 asn1parse'. By implication, the functions ASN1_parse_dump() and
4835 BIO_dump_indent() are added.
4838 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4839 these print out strings and name structures based on various
4840 flags including RFC2253 support and proper handling of
4841 multibyte characters. Added options to the 'x509' utility
4842 to allow the various flags to be set.
4845 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4846 Also change the functions X509_cmp_current_time() and
4847 X509_gmtime_adj() work with an ASN1_TIME structure,
4848 this will enable certificates using GeneralizedTime in validity
4849 dates to be checked.
4852 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4853 negative public key encodings) on by default,
4854 NO_NEG_PUBKEY_BUG can be set to disable it.
4857 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4858 content octets. An i2c_ASN1_OBJECT is unnecessary because
4859 the encoding can be trivially obtained from the structure.
4862 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4863 not read locks (CRYPTO_r_[un]lock).
4866 *) A first attempt at creating official support for shared
4867 libraries through configuration. I've kept it so the
4868 default is static libraries only, and the OpenSSL programs
4869 are always statically linked for now, but there are
4870 preparations for dynamic linking in place.
4871 This has been tested on Linux and Tru64.
4874 *) Randomness polling function for Win9x, as described in:
4875 Peter Gutmann, Software Generation of Practically Strong
4879 *) Fix so PRNG is seeded in req if using an already existing
4883 *) New options to smime application. -inform and -outform
4884 allow alternative formats for the S/MIME message including
4885 PEM and DER. The -content option allows the content to be
4886 specified separately. This should allow things like Netscape
4887 form signing output easier to verify.
4890 *) Fix the ASN1 encoding of tags using the 'long form'.
4893 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4894 STRING types. These convert content octets to and from the
4895 underlying type. The actual tag and length octets are
4896 already assumed to have been read in and checked. These
4897 are needed because all other string types have virtually
4898 identical handling apart from the tag. By having versions
4899 of the ASN1 functions that just operate on content octets
4900 IMPLICIT tagging can be handled properly. It also allows
4901 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4902 and ASN1_INTEGER are identical apart from the tag.
4905 *) Change the handling of OID objects as follows:
4907 - New object identifiers are inserted in objects.txt, following
4908 the syntax given in objects.README.
4909 - objects.pl is used to process obj_mac.num and create a new
4911 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4914 This is currently kind of a hack, and the perl code in objects.pl
4915 isn't very elegant, but it works as I intended. The simplest way
4916 to check that it worked correctly is to look in obj_dat.h and
4917 check the array nid_objs and make sure the objects haven't moved
4918 around (this is important!). Additions are OK, as well as
4919 consistent name changes.
4922 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4925 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4926 The given file adds to whatever has already been seeded into the
4927 random pool through the RANDFILE configuration file option or
4928 environment variable, or the default random state file.
4931 *) mkstack.pl now sorts each macro group into lexical order.
4932 Previously the output order depended on the order the files
4933 appeared in the directory, resulting in needless rewriting
4937 *) Patches to make OpenSSL compile under Win32 again. Mostly
4938 work arounds for the VC++ problem that it treats func() as
4939 func(void). Also stripped out the parts of mkdef.pl that
4940 added extra typesafe functions: these no longer exist.
4943 *) Reorganisation of the stack code. The macros are now all
4944 collected in safestack.h . Each macro is defined in terms of
4945 a "stack macro" of the form SKM_<name>(type, a, b). The
4946 DEBUG_SAFESTACK is now handled in terms of function casts,
4947 this has the advantage of retaining type safety without the
4948 use of additional functions. If DEBUG_SAFESTACK is not defined
4949 then the non typesafe macros are used instead. Also modified the
4950 mkstack.pl script to handle the new form. Needs testing to see
4951 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4952 the default if no major problems. Similar behaviour for ASN1_SET_OF
4953 and PKCS12_STACK_OF.
4956 *) When some versions of IIS use the 'NET' form of private key the
4957 key derivation algorithm is different. Normally MD5(password) is
4958 used as a 128 bit RC4 key. In the modified case
4959 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4960 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4961 as the old Netscape_RSA functions except they have an additional
4962 'sgckey' parameter which uses the modified algorithm. Also added
4963 an -sgckey command line option to the rsa utility. Thanks to
4964 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4965 algorithm to openssl-dev.
4968 *) The evp_local.h macros were using 'c.##kname' which resulted in
4969 invalid expansion on some systems (SCO 5.0.5 for example).
4970 Corrected to 'c.kname'.
4971 [Phillip Porch <root@theporch.com>]
4973 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4974 a STACK of email addresses from a certificate or request, these look
4975 in the subject name and the subject alternative name extensions and
4976 omit any duplicate addresses.
4979 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4980 This makes DSA verification about 2 % faster.
4983 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4984 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4985 plus overhead for 1024 bit moduli).
4986 This makes exponentiations about 0.5 % faster for 1024 bit
4987 exponents (as measured by "openssl speed rsa2048").
4990 *) Rename memory handling macros to avoid conflicts with other
4992 Malloc => OPENSSL_malloc
4993 Malloc_locked => OPENSSL_malloc_locked
4994 Realloc => OPENSSL_realloc
4995 Free => OPENSSL_free
4998 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4999 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5002 *) CygWin32 support.
5003 [John Jarvie <jjarvie@newsguy.com>]
5005 *) The type-safe stack code has been rejigged. It is now only compiled
5006 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5007 by default all type-specific stack functions are "#define"d back to
5008 standard stack functions. This results in more streamlined output
5009 but retains the type-safety checking possibilities of the original
5013 *) The STACK code has been cleaned up, and certain type declarations
5014 that didn't make a lot of sense have been brought in line. This has
5015 also involved a cleanup of sorts in safestack.h to more correctly
5016 map type-safe stack functions onto their plain stack counterparts.
5017 This work has also resulted in a variety of "const"ifications of
5018 lots of the code, especially "_cmp" operations which should normally
5019 be prototyped with "const" parameters anyway.
5022 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5023 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5024 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5025 where all of 'md' is used each time the PRNG is used, but 'state'
5026 is used only indexed by a cyclic counter. As entropy may not be
5027 well distributed from the beginning, 'md' is important as a
5028 chaining variable. However, the output function chains only half
5029 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5030 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5031 in all of 'state' being rewritten, with the new values depending
5032 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5035 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5036 the handshake is continued after ssl_verify_cert_chain();
5037 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5038 can lead to 'unexplainable' connection aborts later.
5039 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5041 *) Major EVP API cipher revision.
5042 Add hooks for extra EVP features. This allows various cipher
5043 parameters to be set in the EVP interface. Support added for variable
5044 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5045 setting of RC2 and RC5 parameters.
5047 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5050 Remove lots of duplicated code from the EVP library. For example *every*
5051 cipher init() function handles the 'iv' in the same way according to the
5052 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5053 for CFB and OFB modes they zero ctx->num.
5055 New functionality allows removal of S/MIME code RC2 hack.
5057 Most of the routines have the same form and so can be declared in terms
5060 By shifting this to the top level EVP_CipherInit() it can be removed from
5061 all individual ciphers. If the cipher wants to handle IVs or keys
5062 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5065 Change lots of functions like EVP_EncryptUpdate() to now return a
5066 value: although software versions of the algorithms cannot fail
5067 any installed hardware versions can.
5070 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5071 this option is set, tolerate broken clients that send the negotiated
5072 protocol version number instead of the requested protocol version
5076 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5077 i.e. non-zero for export ciphersuites, zero otherwise.
5078 Previous versions had this flag inverted, inconsistent with
5079 rsa_tmp_cb (..._TMP_RSA_CB).
5080 [Bodo Moeller; problem reported by Amit Chopra]
5082 *) Add missing DSA library text string. Work around for some IIS
5083 key files with invalid SEQUENCE encoding.
5086 *) Add a document (doc/standards.txt) that list all kinds of standards
5087 and so on that are implemented in OpenSSL.
5090 *) Enhance c_rehash script. Old version would mishandle certificates
5091 with the same subject name hash and wouldn't handle CRLs at all.
5092 Added -fingerprint option to crl utility, to support new c_rehash
5096 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5099 *) Fix for SSL server purpose checking. Server checking was
5100 rejecting certificates which had extended key usage present
5101 but no ssl client purpose.
5102 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5104 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5105 is a little unclear about how a blank password is handled.
5106 Since the password in encoded as a BMPString with terminating
5107 double NULL a zero length password would end up as just the
5108 double NULL. However no password at all is different and is
5109 handled differently in the PKCS#12 key generation code. NS
5110 treats a blank password as zero length. MSIE treats it as no
5111 password on export: but it will try both on import. We now do
5112 the same: PKCS12_parse() tries zero length and no password if
5113 the password is set to "" or NULL (NULL is now a valid password:
5114 it wasn't before) as does the pkcs12 application.
5117 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5118 perror when PEM_read_bio_X509_REQ fails, the error message must
5119 be obtained from the error queue.
5122 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5123 it in ERR_remove_state if appropriate, and change ERR_get_state
5124 accordingly to avoid race conditions (this is necessary because
5125 thread_hash is no longer constant once set).
5128 *) Bugfix for linux-elf makefile.one.
5131 *) RSA_get_default_method() will now cause a default
5132 RSA_METHOD to be chosen if one doesn't exist already.
5133 Previously this was only set during a call to RSA_new()
5134 or RSA_new_method(NULL) meaning it was possible for
5135 RSA_get_default_method() to return NULL.
5138 *) Added native name translation to the existing DSO code
5139 that will convert (if the flag to do so is set) filenames
5140 that are sufficiently small and have no path information
5141 into a canonical native form. Eg. "blah" converted to
5142 "libblah.so" or "blah.dll" etc.
5145 *) New function ERR_error_string_n(e, buf, len) which is like
5146 ERR_error_string(e, buf), but writes at most 'len' bytes
5147 including the 0 terminator. For ERR_error_string_n, 'buf'
5149 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5151 *) CONF library reworked to become more general. A new CONF
5152 configuration file reader "class" is implemented as well as a
5153 new functions (NCONF_*, for "New CONF") to handle it. The now
5154 old CONF_* functions are still there, but are reimplemented to
5155 work in terms of the new functions. Also, a set of functions
5156 to handle the internal storage of the configuration data is
5157 provided to make it easier to write new configuration file
5158 reader "classes" (I can definitely see something reading a
5159 configuration file in XML format, for example), called _CONF_*,
5160 or "the configuration storage API"...
5162 The new configuration file reading functions are:
5164 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5165 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5167 NCONF_default, NCONF_WIN32
5169 NCONF_dump_fp, NCONF_dump_bio
5171 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5172 NCONF_new creates a new CONF object. This works in the same way
5173 as other interfaces in OpenSSL, like the BIO interface.
5174 NCONF_dump_* dump the internal storage of the configuration file,
5175 which is useful for debugging. All other functions take the same
5176 arguments as the old CONF_* functions wth the exception of the
5177 first that must be a `CONF *' instead of a `LHASH *'.
5179 To make it easer to use the new classes with the old CONF_* functions,
5180 the function CONF_set_default_method is provided.
5183 *) Add '-tls1' option to 'openssl ciphers', which was already
5184 mentioned in the documentation but had not been implemented.
5185 (This option is not yet really useful because even the additional
5186 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5189 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5190 OpenSSL-based applications) load shared libraries and bind to
5191 them in a portable way.
5192 [Geoff Thorpe, with contributions from Richard Levitte]
5194 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5196 *) Make sure _lrotl and _lrotr are only used with MSVC.
5198 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5199 (the default implementation of RAND_status).
5201 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5202 to '-clrext' (= clear extensions), as intended and documented.
5203 [Bodo Moeller; inconsistency pointed out by Michael Attili
5206 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5207 was larger than the MD block size.
5208 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5210 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5211 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5212 using the passed key: if the passed key was a private key the result
5213 of X509_print(), for example, would be to print out all the private key
5217 *) des_quad_cksum() byte order bug fix.
5218 [Ulf Möller, using the problem description in krb4-0.9.7, where
5219 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5221 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5223 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5225 *) For easily testing in shell scripts whether some command
5226 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5227 returns with exit code 0 iff no command of the given name is available.
5228 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5229 the output goes to stdout and nothing is printed to stderr.
5230 Additional arguments are always ignored.
5232 Since for each cipher there is a command of the same name,
5233 the 'no-cipher' compilation switches can be tested this way.
5235 ('openssl no-XXX' is not able to detect pseudo-commands such
5236 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5239 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5242 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5243 is set; it will be thrown away anyway because each handshake creates
5245 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5246 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5247 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5248 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5251 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5252 'Q' and 'R' lose their special meanings (quit/renegotiate).
5253 This is part of what -quiet does; unlike -quiet, -ign_eof
5254 does not suppress any output.
5257 *) Add compatibility options to the purpose and trust code. The
5258 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5259 accepts a certificate or CA, this was the previous behaviour,
5260 with all the associated security issues.
5262 X509_TRUST_COMPAT is the old trust behaviour: only and
5263 automatically trust self signed roots in certificate store. A
5264 new trust setting X509_TRUST_DEFAULT is used to specify that
5265 a purpose has no associated trust setting and it should instead
5266 use the value in the default purpose.
5269 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5270 and fix a memory leak.
5273 *) In util/mkerr.pl (which implements 'make errors'), preserve
5274 reason strings from the previous version of the .c file, as
5275 the default to have only downcase letters (and digits) in
5276 automatically generated reasons codes is not always appropriate.
5279 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5280 using strerror. Previously, ERR_reason_error_string() returned
5281 library names as reason strings for SYSerr; but SYSerr is a special
5282 case where small numbers are errno values, not library numbers.
5285 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5286 converts DSA parameters into DH parameters. (When creating parameters,
5287 DSA_generate_parameters is used.)
5290 *) Include 'length' (recommended exponent length) in C code generated
5291 by 'openssl dhparam -C'.
5294 *) The second argument to set_label in perlasm was already being used
5295 so couldn't be used as a "file scope" flag. Moved to third argument
5299 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5300 instead of RAND_bytes for encryption IVs and salts.
5303 *) Include RAND_status() into RAND_METHOD instead of implementing
5304 it only for md_rand.c Otherwise replacing the PRNG by calling
5305 RAND_set_rand_method would be impossible.
5308 *) Don't let DSA_generate_key() enter an infinite loop if the random
5309 number generation fails.
5312 *) New 'rand' application for creating pseudo-random output.
5315 *) Added configuration support for Linux/IA64
5316 [Rolf Haberrecker <rolf@suse.de>]
5318 *) Assembler module support for Mingw32.
5321 *) Shared library support for HPUX (in shlib/).
5322 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5324 *) Shared library support for Solaris gcc.
5325 [Lutz Behnke <behnke@trustcenter.de>]
5327 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5329 *) PKCS7_encrypt() was adding text MIME headers twice because they
5330 were added manually and by SMIME_crlf_copy().
5333 *) In bntest.c don't call BN_rand with zero bits argument.
5334 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5336 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5337 case was implemented. This caused BN_div_recp() to fail occasionally.
5340 *) Add an optional second argument to the set_label() in the perl
5341 assembly language builder. If this argument exists and is set
5342 to 1 it signals that the assembler should use a symbol whose
5343 scope is the entire file, not just the current function. This
5344 is needed with MASM which uses the format label:: for this scope.
5345 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5347 *) Change the ASN1 types so they are typedefs by default. Before
5348 almost all types were #define'd to ASN1_STRING which was causing
5349 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5353 *) Change names of new functions to the new get1/get0 naming
5354 convention: After 'get1', the caller owns a reference count
5355 and has to call ..._free; 'get0' returns a pointer to some
5356 data structure without incrementing reference counters.
5357 (Some of the existing 'get' functions increment a reference
5358 counter, some don't.)
5359 Similarly, 'set1' and 'add1' functions increase reference
5360 counters or duplicate objects.
5363 *) Allow for the possibility of temp RSA key generation failure:
5364 the code used to assume it always worked and crashed on failure.
5367 *) Fix potential buffer overrun problem in BIO_printf().
5368 [Ulf Möller, using public domain code by Patrick Powell; problem
5369 pointed out by David Sacerdote <das33@cornell.edu>]
5371 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5372 RAND_egd() and RAND_status(). In the command line application,
5373 the EGD socket can be specified like a seed file using RANDFILE
5377 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5378 Some CAs (e.g. Verisign) distribute certificates in this form.
5381 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5382 list to exclude them. This means that no special compilation option
5383 is needed to use anonymous DH: it just needs to be included in the
5387 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5388 EVP_MD_type. The old functionality is available in a new macro called
5389 EVP_MD_md(). Change code that uses it and update docs.
5392 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5393 where the 'void *' argument is replaced by a function pointer argument.
5394 Previously 'void *' was abused to point to functions, which works on
5395 many platforms, but is not correct. As these functions are usually
5396 called by macros defined in OpenSSL header files, most source code
5397 should work without changes.
5400 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5401 sections with information on -D... compiler switches used for
5402 compiling the library so that applications can see them. To enable
5403 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5404 must be defined. E.g.,
5405 #define OPENSSL_ALGORITHM_DEFINES
5406 #include <openssl/opensslconf.h>
5407 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5408 [Richard Levitte, Ulf and Bodo Möller]
5410 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5414 *) Change the 'other' type in certificate aux info to a STACK_OF
5415 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5416 the required ASN1 format: arbitrary types determined by an OID.
5419 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5420 argument to 'req'. This is not because the function is newer or
5421 better than others it just uses the work 'NEW' in the certificate
5422 request header lines. Some software needs this.
5425 *) Reorganise password command line arguments: now passwords can be
5426 obtained from various sources. Delete the PEM_cb function and make
5427 it the default behaviour: i.e. if the callback is NULL and the
5428 usrdata argument is not NULL interpret it as a null terminated pass
5429 phrase. If usrdata and the callback are NULL then the pass phrase
5430 is prompted for as usual.
5433 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5434 the support is automatically enabled. The resulting binaries will
5435 autodetect the card and use it if present.
5436 [Ben Laurie and Compaq Inc.]
5438 *) Work around for Netscape hang bug. This sends certificate request
5439 and server done in one record. Since this is perfectly legal in the
5440 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5441 the bugs/SSLv3 entry for more info.
5444 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5447 *) Add -rand argument to smime and pkcs12 applications and read/write
5451 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5454 *) Add command line password options to the remaining applications.
5457 *) Bug fix for BN_div_recp() for numerators with an even number of
5461 *) More tests in bntest.c, and changed test_bn output.
5464 *) ./config recognizes MacOS X now.
5467 *) Bug fix for BN_div() when the first words of num and divsor are
5468 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5471 *) Add support for various broken PKCS#8 formats, and command line
5472 options to produce them.
5475 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5476 get temporary BIGNUMs from a BN_CTX.
5479 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5483 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5484 include a #define from the old name to the new. The original intent
5485 was that statically linked binaries could for example just call
5486 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5487 link with digests. This never worked becayse SSLeay_add_all_digests()
5488 and SSLeay_add_all_ciphers() were in the same source file so calling
5489 one would link with the other. They are now in separate source files.
5492 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5495 *) Use a less unusual form of the Miller-Rabin primality test (it used
5496 a binary algorithm for exponentiation integrated into the Miller-Rabin
5497 loop, our standard modexp algorithms are faster).
5500 *) Support for the EBCDIC character set completed.
5501 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5503 *) Source code cleanups: use const where appropriate, eliminate casts,
5504 use void * instead of char * in lhash.
5507 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5508 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5509 this the server could overwrite ephemeral keys that the client
5513 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5514 using 50 iterations of the Rabin-Miller test.
5516 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5517 iterations of the Rabin-Miller test as required by the appendix
5518 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5519 As BN_is_prime_fasttest includes trial division, DSA parameter
5520 generation becomes much faster.
5522 This implies a change for the callback functions in DSA_is_prime
5523 and DSA_generate_parameters: The callback function is called once
5524 for each positive witness in the Rabin-Miller test, not just
5525 occasionally in the inner loop; and the parameters to the
5526 callback function now provide an iteration count for the outer
5527 loop rather than for the current invocation of the inner loop.
5528 DSA_generate_parameters additionally can call the callback
5529 function with an 'iteration count' of -1, meaning that a
5530 candidate has passed the trial division test (when q is generated
5531 from an application-provided seed, trial division is skipped).
5534 *) New function BN_is_prime_fasttest that optionally does trial
5535 division before starting the Rabin-Miller test and has
5536 an additional BN_CTX * argument (whereas BN_is_prime always
5537 has to allocate at least one BN_CTX).
5538 'callback(1, -1, cb_arg)' is called when a number has passed the
5539 trial division stage.
5542 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5546 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5549 *) New function BN_pseudo_rand().
5552 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5553 bignum version of BN_from_montgomery() with the working code from
5554 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5558 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5559 made it impossible to use the same SSL_SESSION data structure in
5560 SSL2 clients in multiple threads.
5563 *) The return value of RAND_load_file() no longer counts bytes obtained
5564 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5565 to seed the PRNG (previously an explicit byte count was required).
5566 [Ulf Möller, Bodo Möller]
5568 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5569 used (char *) instead of (void *) and had casts all over the place.
5572 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5575 *) Retain source code compatibility for BN_prime_checks macro:
5576 BN_is_prime(..., BN_prime_checks, ...) now uses
5577 BN_prime_checks_for_size to determine the appropriate number of
5578 Rabin-Miller iterations.
5581 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5582 DH_CHECK_P_NOT_SAFE_PRIME.
5583 (Check if this is true? OpenPGP calls them "strong".)
5586 *) Merge the functionality of "dh" and "gendh" programs into a new program
5587 "dhparam". The old programs are retained for now but will handle DH keys
5588 (instead of parameters) in future.
5591 *) Make the ciphers, s_server and s_client programs check the return values
5592 when a new cipher list is set.
5595 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5596 ciphers. Before when the 56bit ciphers were enabled the sorting was
5599 The syntax for the cipher sorting has been extended to support sorting by
5600 cipher-strength (using the strength_bits hard coded in the tables).
5601 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5603 Fix a bug in the cipher-command parser: when supplying a cipher command
5604 string with an "undefined" symbol (neither command nor alphanumeric
5605 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5606 an error is flagged.
5608 Due to the strength-sorting extension, the code of the
5609 ssl_create_cipher_list() function was completely rearranged. I hope that
5610 the readability was also increased :-)
5611 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5613 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5614 for the first serial number and places 2 in the serial number file. This
5615 avoids problems when the root CA is created with serial number zero and
5616 the first user certificate has the same issuer name and serial number
5620 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5621 the new code. Add documentation for this stuff.
5624 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5625 X509_*() to X509at_*() on the grounds that they don't handle X509
5626 structures and behave in an analagous way to the X509v3 functions:
5627 they shouldn't be called directly but wrapper functions should be used
5630 So we also now have some wrapper functions that call the X509at functions
5631 when passed certificate requests. (TO DO: similar things can be done with
5632 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5633 things. Some of these need some d2i or i2d and print functionality
5634 because they handle more complex structures.)
5637 *) Add missing #ifndefs that caused missing symbols when building libssl
5638 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5639 NO_RSA in ssl/s2*.c.
5640 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5642 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5643 has a return value which indicates the quality of the random data
5644 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5645 error queue. New function RAND_pseudo_bytes() generates output that is
5646 guaranteed to be unique but not unpredictable. RAND_add is like
5647 RAND_seed, but takes an extra argument for an entropy estimate
5648 (RAND_seed always assumes full entropy).
5651 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5652 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5653 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5654 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5655 false-positive rate of at most 2^-80 for random input.
5658 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5661 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5662 in the 0.9.5 release), this returns the chain
5663 from an X509_CTX structure with a dup of the stack and all
5664 the X509 reference counts upped: so the stack will exist
5665 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5668 Also make SSL_SESSION_print() print out the verify return
5672 *) Add manpage for the pkcs12 command. Also change the default
5673 behaviour so MAC iteration counts are used unless the new
5674 -nomaciter option is used. This improves file security and
5675 only older versions of MSIE (4.0 for example) need it.
5678 *) Honor the no-xxx Configure options when creating .DEF files.
5681 *) Add PKCS#10 attributes to field table: challengePassword,
5682 unstructuredName and unstructuredAddress. These are taken from
5683 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5684 international characters are used.
5686 More changes to X509_ATTRIBUTE code: allow the setting of types
5687 based on strings. Remove the 'loc' parameter when adding
5688 attributes because these will be a SET OF encoding which is sorted
5692 *) Initial changes to the 'req' utility to allow request generation
5693 automation. This will allow an application to just generate a template
5694 file containing all the field values and have req construct the
5697 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5698 used all over the place including certificate requests and PKCS#7
5699 structures. They are currently handled manually where necessary with
5700 some primitive wrappers for PKCS#7. The new functions behave in a
5701 manner analogous to the X509 extension functions: they allow
5702 attributes to be looked up by NID and added.
5704 Later something similar to the X509V3 code would be desirable to
5705 automatically handle the encoding, decoding and printing of the
5706 more complex types. The string types like challengePassword can
5707 be handled by the string table functions.
5709 Also modified the multi byte string table handling. Now there is
5710 a 'global mask' which masks out certain types. The table itself
5711 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5712 is useful when for example there is only one permissible type
5713 (as in countryName) and using the mask might result in no valid
5717 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5718 SSL_get_peer_finished to allow applications to obtain the latest
5719 Finished messages sent to the peer or expected from the peer,
5720 respectively. (SSL_get_peer_finished is usually the Finished message
5721 actually received from the peer, otherwise the protocol will be aborted.)
5723 As the Finished message are message digests of the complete handshake
5724 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5725 be used for external authentication procedures when the authentication
5726 provided by SSL/TLS is not desired or is not enough.
5729 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5730 the host supports BWX extension and if Compaq C is present on the
5731 $PATH. Just exploiting of the BWX extension results in 20-30%
5732 performance kick for some algorithms, e.g. DES and RC4 to mention
5733 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5737 *) Add support for MS "fast SGC". This is arguably a violation of the
5738 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5739 weak crypto and after checking the certificate is SGC a second one
5740 with strong crypto. MS SGC stops the first handshake after receiving
5741 the server certificate message and sends a second client hello. Since
5742 a server will typically do all the time consuming operations before
5743 expecting any further messages from the client (server key exchange
5744 is the most expensive) there is little difference between the two.
5746 To get OpenSSL to support MS SGC we have to permit a second client
5747 hello message after we have sent server done. In addition we have to
5748 reset the MAC if we do get this second client hello.
5751 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5752 if a DER encoded private key is RSA or DSA traditional format. Changed
5753 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5754 format DER encoded private key. Newer code should use PKCS#8 format which
5755 has the key type encoded in the ASN1 structure. Added DER private key
5756 support to pkcs8 application.
5759 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5760 ciphersuites has been selected (as required by the SSL 3/TLS 1
5761 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5762 is set, we interpret this as a request to violate the specification
5763 (the worst that can happen is a handshake failure, and 'correct'
5764 behaviour would result in a handshake failure anyway).
5767 *) In SSL_CTX_add_session, take into account that there might be multiple
5768 SSL_SESSION structures with the same session ID (e.g. when two threads
5769 concurrently obtain them from an external cache).
5770 The internal cache can handle only one SSL_SESSION with a given ID,
5771 so if there's a conflict, we now throw out the old one to achieve
5775 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5776 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5777 some routines that use cipher OIDs: some ciphers do not have OIDs
5778 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5782 *) Simplify the trust setting structure and code. Now we just have
5783 two sequences of OIDs for trusted and rejected settings. These will
5784 typically have values the same as the extended key usage extension
5785 and any application specific purposes.
5787 The trust checking code now has a default behaviour: it will just
5788 check for an object with the same NID as the passed id. Functions can
5789 be provided to override either the default behaviour or the behaviour
5790 for a given id. SSL client, server and email already have functions
5791 in place for compatibility: they check the NID and also return "trusted"
5792 if the certificate is self signed.
5795 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5796 traditional format into an EVP_PKEY structure.
5799 *) Add a password callback function PEM_cb() which either prompts for
5800 a password if usr_data is NULL or otherwise assumes it is a null
5801 terminated password. Allow passwords to be passed on command line
5802 environment or config files in a few more utilities.
5805 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5806 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5807 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5808 Update documentation.
5811 *) Support for ASN1 "NULL" type. This could be handled before by using
5812 ASN1_TYPE but there wasn't any function that would try to read a NULL
5813 and produce an error if it couldn't. For compatibility we also have
5814 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5815 don't allocate anything because they don't need to.
5818 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5820 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5822 *) Rebuild of the memory allocation routines used by OpenSSL code and
5823 possibly others as well. The purpose is to make an interface that
5824 provide hooks so anyone can build a separate set of allocation and
5825 deallocation routines to be used by OpenSSL, for example memory
5826 pool implementations, or something else, which was previously hard
5827 since Malloc(), Realloc() and Free() were defined as macros having
5828 the values malloc, realloc and free, respectively (except for Win32
5829 compilations). The same is provided for memory debugging code.
5830 OpenSSL already comes with functionality to find memory leaks, but
5831 this gives people a chance to debug other memory problems.
5833 With these changes, a new set of functions and macros have appeared:
5835 CRYPTO_set_mem_debug_functions() [F]
5836 CRYPTO_get_mem_debug_functions() [F]
5837 CRYPTO_dbg_set_options() [F]
5838 CRYPTO_dbg_get_options() [F]
5839 CRYPTO_malloc_debug_init() [M]
5841 The memory debug functions are NULL by default, unless the library
5842 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5843 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5844 gives the standard debugging functions that come with OpenSSL) or
5845 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5846 provided by the library user) must be used. When the standard
5847 debugging functions are used, CRYPTO_dbg_set_options can be used to
5848 request additional information:
5849 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5850 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5852 Also, things like CRYPTO_set_mem_functions will always give the
5853 expected result (the new set of functions is used for allocation
5854 and deallocation) at all times, regardless of platform and compiler
5857 To finish it up, some functions that were never use in any other
5858 way than through macros have a new API and new semantic:
5861 CRYPTO_dbg_realloc()
5864 All macros of value have retained their old syntax.
5865 [Richard Levitte and Bodo Moeller]
5867 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5868 ordering of SMIMECapabilities wasn't in "strength order" and there
5869 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5873 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5874 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5875 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5877 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5878 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5879 functionality to handle multipart/signed properly) and a utility
5880 called 'smime' to call all this stuff. This is based on code I
5881 originally wrote for Celo who have kindly allowed it to be
5882 included in OpenSSL.
5885 *) Add variants des_set_key_checked and des_set_key_unchecked of
5886 des_set_key (aka des_key_sched). Global variable des_check_key
5887 decides which of these is called by des_set_key; this way
5888 des_check_key behaves as it always did, but applications and
5889 the library itself, which was buggy for des_check_key == 1,
5890 have a cleaner way to pick the version they need.
5893 *) New function PKCS12_newpass() which changes the password of a
5897 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5898 dynamic mix. In both cases the ids can be used as an index into the
5899 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5900 functions so they accept a list of the field values and the
5901 application doesn't need to directly manipulate the X509_TRUST
5905 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5909 *) Modify the way the V3 extension code looks up extensions. This now
5910 works in a similar way to the object code: we have some "standard"
5911 extensions in a static table which is searched with OBJ_bsearch()
5912 and the application can add dynamic ones if needed. The file
5913 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5914 updated whenever a new extension is added to the core code and kept
5915 in ext_nid order. There is a simple program 'tabtest.c' which checks
5916 this. New extensions are not added too often so this file can readily
5917 be maintained manually.
5919 There are two big advantages in doing things this way. The extensions
5920 can be looked up immediately and no longer need to be "added" using
5921 X509V3_add_standard_extensions(): this function now does nothing.
5922 [Side note: I get *lots* of email saying the extension code doesn't
5923 work because people forget to call this function]
5924 Also no dynamic allocation is done unless new extensions are added:
5925 so if we don't add custom extensions there is no need to call
5926 X509V3_EXT_cleanup().
5929 *) Modify enc utility's salting as follows: make salting the default. Add a
5930 magic header, so unsalted files fail gracefully instead of just decrypting
5931 to garbage. This is because not salting is a big security hole, so people
5932 should be discouraged from doing it.
5935 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5936 digest to be passed on the command line but it only used this
5937 parameter when signing a certificate. Modified so all relevant
5938 operations are affected by the digest parameter including the
5939 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5940 DSA key was used because it didn't fix the digest.
5943 *) Initial certificate chain verify code. Currently tests the untrusted
5944 certificates for consistency with the verify purpose (which is set
5945 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5947 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5948 this is because it will reject chains with invalid extensions whereas
5949 every previous version of OpenSSL and SSLeay made no checks at all.
5951 Trust code: checks the root CA for the relevant trust settings. Trust
5952 settings have an initial value consistent with the verify purpose: e.g.
5953 if the verify purpose is for SSL client use it expects the CA to be
5954 trusted for SSL client use. However the default value can be changed to
5955 permit custom trust settings: one example of this would be to only trust
5956 certificates from a specific "secure" set of CAs.
5958 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5959 which should be used for version portability: especially since the
5960 verify structure is likely to change more often now.
5962 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5963 to set them. If not set then assume SSL clients will verify SSL servers
5966 Two new options to the verify program: -untrusted allows a set of
5967 untrusted certificates to be passed in and -purpose which sets the
5968 intended purpose of the certificate. If a purpose is set then the
5969 new chain verify code is used to check extension consistency.
5972 *) Support for the authority information access extension.
5975 *) Modify RSA and DSA PEM read routines to transparently handle
5976 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5977 public keys in a format compatible with certificate
5978 SubjectPublicKeyInfo structures. Unfortunately there were already
5979 functions called *_PublicKey_* which used various odd formats so
5980 these are retained for compatibility: however the DSA variants were
5981 never in a public release so they have been deleted. Changed dsa/rsa
5982 utilities to handle the new format: note no releases ever handled public
5983 keys so we should be OK.
5985 The primary motivation for this change is to avoid the same fiasco
5986 that dogs private keys: there are several incompatible private key
5987 formats some of which are standard and some OpenSSL specific and
5988 require various evil hacks to allow partial transparent handling and
5989 even then it doesn't work with DER formats. Given the option anything
5990 other than PKCS#8 should be dumped: but the other formats have to
5991 stay in the name of compatibility.
5993 With public keys and the benefit of hindsight one standard format
5994 is used which works with EVP_PKEY, RSA or DSA structures: though
5995 it clearly returns an error if you try to read the wrong kind of key.
5997 Added a -pubkey option to the 'x509' utility to output the public key.
5998 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5999 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6000 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6001 that do the same as the EVP_PKEY_assign_*() except they up the
6002 reference count of the added key (they don't "swallow" the
6006 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6007 CRLs would fail if the file contained no certificates or no CRLs:
6008 added a new function to read in both types and return the number
6009 read: this means that if none are read it will be an error. The
6010 DER versions of the certificate and CRL reader would always fail
6011 because it isn't possible to mix certificates and CRLs in DER format
6012 without choking one or the other routine. Changed this to just read
6013 a certificate: this is the best we can do. Also modified the code
6014 in apps/verify.c to take notice of return codes: it was previously
6015 attempting to read in certificates from NULL pointers and ignoring
6016 any errors: this is one reason why the cert and CRL reader seemed
6017 to work. It doesn't check return codes from the default certificate
6018 routines: these may well fail if the certificates aren't installed.
6021 *) Code to support otherName option in GeneralName.
6024 *) First update to verify code. Change the verify utility
6025 so it warns if it is passed a self signed certificate:
6026 for consistency with the normal behaviour. X509_verify
6027 has been modified to it will now verify a self signed
6028 certificate if *exactly* the same certificate appears
6029 in the store: it was previously impossible to trust a
6030 single self signed certificate. This means that:
6031 openssl verify ss.pem
6032 now gives a warning about a self signed certificate but
6033 openssl verify -CAfile ss.pem ss.pem
6037 *) For servers, store verify_result in SSL_SESSION data structure
6038 (and add it to external session representation).
6039 This is needed when client certificate verifications fails,
6040 but an application-provided verification callback (set by
6041 SSL_CTX_set_cert_verify_callback) allows accepting the session
6042 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6043 but returns 1): When the session is reused, we have to set
6044 ssl->verify_result to the appropriate error code to avoid
6046 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6048 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6049 case in PKCS7_dataInit() where the signed PKCS7 structure
6050 didn't contain any existing data because it was being created.
6051 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6053 *) Add a salt to the key derivation routines in enc.c. This
6054 forms the first 8 bytes of the encrypted file. Also add a
6055 -S option to allow a salt to be input on the command line.
6058 *) New function X509_cmp(). Oddly enough there wasn't a function
6059 to compare two certificates. We do this by working out the SHA1
6060 hash and comparing that. X509_cmp() will be needed by the trust
6064 *) SSL_get1_session() is like SSL_get_session(), but increments
6065 the reference count in the SSL_SESSION returned.
6066 [Geoff Thorpe <geoff@eu.c2.net>]
6068 *) Fix for 'req': it was adding a null to request attributes.
6069 Also change the X509_LOOKUP and X509_INFO code to handle
6070 certificate auxiliary information.
6073 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6077 *) Add the possibility to add extra information to the memory leak
6078 detecting output, to form tracebacks, showing from where each
6079 allocation was originated: CRYPTO_push_info("constant string") adds
6080 the string plus current file name and line number to a per-thread
6081 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6082 is like calling CYRPTO_pop_info() until the stack is empty.
6083 Also updated memory leak detection code to be multi-thread-safe.
6086 *) Add options -text and -noout to pkcs7 utility and delete the
6087 encryption options which never did anything. Update docs.
6090 *) Add options to some of the utilities to allow the pass phrase
6091 to be included on either the command line (not recommended on
6092 OSes like Unix) or read from the environment. Update the
6093 manpages and fix a few bugs.
6096 *) Add a few manpages for some of the openssl commands.
6099 *) Fix the -revoke option in ca. It was freeing up memory twice,
6100 leaking and not finding already revoked certificates.
6103 *) Extensive changes to support certificate auxiliary information.
6104 This involves the use of X509_CERT_AUX structure and X509_AUX
6105 functions. An X509_AUX function such as PEM_read_X509_AUX()
6106 can still read in a certificate file in the usual way but it
6107 will also read in any additional "auxiliary information". By
6108 doing things this way a fair degree of compatibility can be
6109 retained: existing certificates can have this information added
6110 using the new 'x509' options.
6112 Current auxiliary information includes an "alias" and some trust
6113 settings. The trust settings will ultimately be used in enhanced
6114 certificate chain verification routines: currently a certificate
6115 can only be trusted if it is self signed and then it is trusted
6119 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6120 The problem was that one of the replacement routines had not been working
6121 since SSLeay releases. For now the offending routine has been replaced
6122 with non-optimised assembler. Even so, this now gives around 95%
6123 performance improvement for 1024 bit RSA signs.
6126 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6127 handling. Most clients have the effective key size in bits equal to
6128 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6129 A few however don't do this and instead use the size of the decrypted key
6130 to determine the RC2 key length and the AlgorithmIdentifier to determine
6131 the effective key length. In this case the effective key length can still
6132 be 40 bits but the key length can be 168 bits for example. This is fixed
6133 by manually forcing an RC2 key into the EVP_PKEY structure because the
6134 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6135 the key length and effective key length are equal.
6138 *) Add a bunch of functions that should simplify the creation of
6139 X509_NAME structures. Now you should be able to do:
6140 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6141 and have it automatically work out the correct field type and fill in
6142 the structures. The more adventurous can try:
6143 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6144 and it will (hopefully) work out the correct multibyte encoding.
6147 *) Change the 'req' utility to use the new field handling and multibyte
6148 copy routines. Before the DN field creation was handled in an ad hoc
6149 way in req, ca, and x509 which was rather broken and didn't support
6150 BMPStrings or UTF8Strings. Since some software doesn't implement
6151 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6152 using the dirstring_type option. See the new comment in the default
6153 openssl.cnf for more info.
6156 *) Make crypto/rand/md_rand.c more robust:
6157 - Assure unique random numbers after fork().
6158 - Make sure that concurrent threads access the global counter and
6159 md serializably so that we never lose entropy in them
6160 or use exactly the same state in multiple threads.
6161 Access to the large state is not always serializable because
6162 the additional locking could be a performance killer, and
6163 md should be large enough anyway.
6166 *) New file apps/app_rand.c with commonly needed functionality
6167 for handling the random seed file.
6169 Use the random seed file in some applications that previously did not:
6171 dsaparam -genkey (which also ignored its '-rand' option),
6174 x509 (when signing).
6175 Except on systems with /dev/urandom, it is crucial to have a random
6176 seed file at least for key creation, DSA signing, and for DH exchanges;
6177 for RSA signatures we could do without one.
6179 gendh and gendsa (unlike genrsa) used to read only the first byte
6180 of each file listed in the '-rand' option. The function as previously
6181 found in genrsa is now in app_rand.c and is used by all programs
6182 that support '-rand'.
6185 *) In RAND_write_file, use mode 0600 for creating files;
6186 don't just chmod when it may be too late.
6189 *) Report an error from X509_STORE_load_locations
6190 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6193 *) New function ASN1_mbstring_copy() this copies a string in either
6194 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6195 into an ASN1_STRING type. A mask of permissible types is passed
6196 and it chooses the "minimal" type to use or an error if not type
6200 *) Add function equivalents to the various macros in asn1.h. The old
6201 macros are retained with an M_ prefix. Code inside the library can
6202 use the M_ macros. External code (including the openssl utility)
6203 should *NOT* in order to be "shared library friendly".
6206 *) Add various functions that can check a certificate's extensions
6207 to see if it usable for various purposes such as SSL client,
6208 server or S/MIME and CAs of these types. This is currently
6209 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6210 verification. Also added a -purpose flag to x509 utility to
6211 print out all the purposes.
6214 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6218 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6219 for, obtain and decode and extension and obtain its critical flag.
6220 This allows all the necessary extension code to be handled in a
6221 single function call.
6224 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6225 platforms. See crypto/rc4/rc4_enc.c for further details.
6228 *) New -noout option to asn1parse. This causes no output to be produced
6229 its main use is when combined with -strparse and -out to extract data
6230 from a file (which may not be in ASN.1 format).
6233 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6234 when producing the local key id.
6235 [Richard Levitte <levitte@stacken.kth.se>]
6237 *) New option -dhparam in s_server. This allows a DH parameter file to be
6238 stated explicitly. If it is not stated then it tries the first server
6239 certificate file. The previous behaviour hard coded the filename
6243 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6244 a public key to be input or output. For example:
6245 openssl rsa -in key.pem -pubout -out pubkey.pem
6246 Also added necessary DSA public key functions to handle this.
6249 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6250 in the message. This was handled by allowing
6251 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6252 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6254 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6255 to the end of the strings whereas this didn't. This would cause problems
6256 if strings read with d2i_ASN1_bytes() were later modified.
6257 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6259 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6260 data and it contains EOF it will end up returning an error. This is
6261 caused by input 46 bytes long. The cause is due to the way base64
6262 BIOs find the start of base64 encoded data. They do this by trying a
6263 trial decode on each line until they find one that works. When they
6264 do a flag is set and it starts again knowing it can pass all the
6265 data directly through the decoder. Unfortunately it doesn't reset
6266 the context it uses. This means that if EOF is reached an attempt
6267 is made to pass two EOFs through the context and this causes the
6268 resulting error. This can also cause other problems as well. As is
6269 usual with these problems it takes *ages* to find and the fix is
6270 trivial: move one line.
6271 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6273 *) Ugly workaround to get s_client and s_server working under Windows. The
6274 old code wouldn't work because it needed to select() on sockets and the
6275 tty (for keypresses and to see if data could be written). Win32 only
6276 supports select() on sockets so we select() with a 1s timeout on the
6277 sockets and then see if any characters are waiting to be read, if none
6278 are present then we retry, we also assume we can always write data to
6279 the tty. This isn't nice because the code then blocks until we've
6280 received a complete line of data and it is effectively polling the
6281 keyboard at 1s intervals: however it's quite a bit better than not
6282 working at all :-) A dedicated Windows application might handle this
6283 with an event loop for example.
6286 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6287 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6288 will be called when RSA_sign() and RSA_verify() are used. This is useful
6289 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6290 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6291 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6292 This necessitated the support of an extra signature type NID_md5_sha1
6293 for SSL signatures and modifications to the SSL library to use it instead
6294 of calling RSA_public_decrypt() and RSA_private_encrypt().
6297 *) Add new -verify -CAfile and -CApath options to the crl program, these
6298 will lookup a CRL issuers certificate and verify the signature in a
6299 similar way to the verify program. Tidy up the crl program so it
6300 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6301 less strict. It will now permit CRL extensions even if it is not
6302 a V2 CRL: this will allow it to tolerate some broken CRLs.
6305 *) Initialize all non-automatic variables each time one of the openssl
6306 sub-programs is started (this is necessary as they may be started
6307 multiple times from the "OpenSSL>" prompt).
6308 [Lennart Bang, Bodo Moeller]
6310 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6311 removing all other RSA functionality (this is what NO_RSA does). This
6312 is so (for example) those in the US can disable those operations covered
6313 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6317 *) Non-copying interface to BIO pairs.
6318 (still largely untested)
6321 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6322 ASCII string. This was handled independently in various places before.
6325 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6326 UTF8 strings a character at a time.
6329 *) Use client_version from client hello to select the protocol
6330 (s23_srvr.c) and for RSA client key exchange verification
6331 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6334 *) Add various utility functions to handle SPKACs, these were previously
6335 handled by poking round in the structure internals. Added new function
6336 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6337 print, verify and generate SPKACs. Based on an original idea from
6338 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6341 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6344 *) Allow the config file extension section to be overwritten on the
6345 command line. Based on an original idea from Massimiliano Pala
6346 <madwolf@comune.modena.it>. The new option is called -extensions
6347 and can be applied to ca, req and x509. Also -reqexts to override
6348 the request extensions in req and -crlexts to override the crl extensions
6352 *) Add new feature to the SPKAC handling in ca. Now you can include
6353 the same field multiple times by preceding it by "XXXX." for example:
6356 this is the same syntax as used in the req config file.
6359 *) Allow certificate extensions to be added to certificate requests. These
6360 are specified in a 'req_extensions' option of the req section of the
6361 config file. They can be printed out with the -text option to req but
6362 are otherwise ignored at present.
6365 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6366 data read consists of only the final block it would not decrypted because
6367 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6368 A misplaced 'break' also meant the decrypted final block might not be
6369 copied until the next read.
6372 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6373 a few extra parameters to the DH structure: these will be useful if
6374 for example we want the value of 'q' or implement X9.42 DH.
6377 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6378 provides hooks that allow the default DSA functions or functions on a
6379 "per key" basis to be replaced. This allows hardware acceleration and
6380 hardware key storage to be handled without major modification to the
6381 library. Also added low level modexp hooks and CRYPTO_EX structure and
6382 associated functions.
6385 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6386 as "read only": it can't be written to and the buffer it points to will
6387 not be freed. Reading from a read only BIO is much more efficient than
6388 a normal memory BIO. This was added because there are several times when
6389 an area of memory needs to be read from a BIO. The previous method was
6390 to create a memory BIO and write the data to it, this results in two
6391 copies of the data and an O(n^2) reading algorithm. There is a new
6392 function BIO_new_mem_buf() which creates a read only memory BIO from
6393 an area of memory. Also modified the PKCS#7 routines to use read only
6397 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6398 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6399 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6400 but a retry condition occured while trying to read the rest.
6403 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6404 NID_pkcs7_encrypted by default: this was wrong since this should almost
6405 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6406 the encrypted data type: this is a more sensible place to put it and it
6407 allows the PKCS#12 code to be tidied up that duplicated this
6411 *) Changed obj_dat.pl script so it takes its input and output files on
6412 the command line. This should avoid shell escape redirection problems
6416 *) Initial support for certificate extension requests, these are included
6417 in things like Xenroll certificate requests. Included functions to allow
6418 extensions to be obtained and added.
6421 *) -crlf option to s_client and s_server for sending newlines as
6422 CRLF (as required by many protocols).
6425 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6427 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6428 [Ralf S. Engelschall]
6430 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6431 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6433 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6437 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6438 DH parameters/keys (q is lost during that conversion, but the resulting
6439 DH parameters contain its length).
6441 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6442 much faster than DH_generate_parameters (which creates parameters
6443 where p = 2*q + 1), and also the smaller q makes DH computations
6444 much more efficient (160-bit exponentiation instead of 1024-bit
6445 exponentiation); so this provides a convenient way to support DHE
6446 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6447 utter importance to use
6448 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6450 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6451 when such DH parameters are used, because otherwise small subgroup
6452 attacks may become possible!
6455 *) Avoid memory leak in i2d_DHparams.
6458 *) Allow the -k option to be used more than once in the enc program:
6459 this allows the same encrypted message to be read by multiple recipients.
6462 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6463 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6464 it will always use the numerical form of the OID, even if it has a short
6468 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6469 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6470 otherwise bn_mod_exp was called. In the case of hardware keys for example
6471 no private key components need be present and it might store extra data
6472 in the RSA structure, which cannot be accessed from bn_mod_exp.
6473 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6474 private key operations.
6477 *) Added support for SPARC Linux.
6480 *) pem_password_cb function type incompatibly changed from
6481 typedef int pem_password_cb(char *buf, int size, int rwflag);
6483 ....(char *buf, int size, int rwflag, void *userdata);
6484 so that applications can pass data to their callbacks:
6485 The PEM[_ASN1]_{read,write}... functions and macros now take an
6486 additional void * argument, which is just handed through whenever
6487 the password callback is called.
6488 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6490 New function SSL_CTX_set_default_passwd_cb_userdata.
6492 Compatibility note: As many C implementations push function arguments
6493 onto the stack in reverse order, the new library version is likely to
6494 interoperate with programs that have been compiled with the old
6495 pem_password_cb definition (PEM_whatever takes some data that
6496 happens to be on the stack as its last argument, and the callback
6497 just ignores this garbage); but there is no guarantee whatsoever that
6500 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6501 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6502 problems not only on Windows, but also on some Unix platforms.
6503 To avoid problematic command lines, these definitions are now in an
6504 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6505 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6508 *) MIPS III/IV assembler module is reimplemented.
6511 *) More DES library cleanups: remove references to srand/rand and
6512 delete an unused file.
6515 *) Add support for the the free Netwide assembler (NASM) under Win32,
6516 since not many people have MASM (ml) and it can be hard to obtain.
6517 This is currently experimental but it seems to work OK and pass all
6518 the tests. Check out INSTALL.W32 for info.
6521 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6522 without temporary keys kept an extra copy of the server key,
6523 and connections with temporary keys did not free everything in case
6527 *) New function RSA_check_key and new openssl rsa option -check
6528 for verifying the consistency of RSA keys.
6529 [Ulf Moeller, Bodo Moeller]
6531 *) Various changes to make Win32 compile work:
6532 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6533 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6534 comparison" warnings.
6535 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6538 *) Add a debugging option to PKCS#5 v2 key generation function: when
6539 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6540 derived keys are printed to stderr.
6543 *) Copy the flags in ASN1_STRING_dup().
6544 [Roman E. Pavlov <pre@mo.msk.ru>]
6546 *) The x509 application mishandled signing requests containing DSA
6547 keys when the signing key was also DSA and the parameters didn't match.
6549 It was supposed to omit the parameters when they matched the signing key:
6550 the verifying software was then supposed to automatically use the CA's
6551 parameters if they were absent from the end user certificate.
6553 Omitting parameters is no longer recommended. The test was also
6554 the wrong way round! This was probably due to unusual behaviour in
6555 EVP_cmp_parameters() which returns 1 if the parameters match.
6556 This meant that parameters were omitted when they *didn't* match and
6557 the certificate was useless. Certificates signed with 'ca' didn't have
6559 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6561 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6562 The interface is as follows:
6563 Applications can use
6564 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6565 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6566 "off" is now the default.
6567 The library internally uses
6568 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6569 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6570 to disable memory-checking temporarily.
6572 Some inconsistent states that previously were possible (and were
6573 even the default) are now avoided.
6575 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6576 with each memory chunk allocated; this is occasionally more helpful
6577 than just having a counter.
6579 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6581 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6585 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6586 which largely parallels "options", but is for changing API behaviour,
6587 whereas "options" are about protocol behaviour.
6588 Initial "mode" flags are:
6590 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6591 a single record has been written.
6592 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6593 retries use the same buffer location.
6594 (But all of the contents must be
6598 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6601 *) Fix problems with no-hmac etc.
6602 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6604 *) New functions RSA_get_default_method(), RSA_set_method() and
6605 RSA_get_method(). These allows replacement of RSA_METHODs without having
6606 to mess around with the internals of an RSA structure.
6609 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6610 Also really enable memory leak checks in openssl.c and in some
6612 [Chad C. Mulligan, Bodo Moeller]
6614 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6615 up the length of negative integers. This has now been simplified to just
6616 store the length when it is first determined and use it later, rather
6617 than trying to keep track of where data is copied and updating it to
6619 [Steve Henson, reported by Brien Wheeler
6620 <bwheeler@authentica-security.com>]
6622 *) Add a new function PKCS7_signatureVerify. This allows the verification
6623 of a PKCS#7 signature but with the signing certificate passed to the
6624 function itself. This contrasts with PKCS7_dataVerify which assumes the
6625 certificate is present in the PKCS#7 structure. This isn't always the
6626 case: certificates can be omitted from a PKCS#7 structure and be
6627 distributed by "out of band" means (such as a certificate database).
6630 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6631 function prototypes in pem.h, also change util/mkdef.pl to add the
6632 necessary function names.
6635 *) mk1mf.pl (used by Windows builds) did not properly read the
6636 options set by Configure in the top level Makefile, and Configure
6637 was not even able to write more than one option correctly.
6638 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6641 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6642 file to be loaded from a BIO or FILE pointer. The BIO version will
6643 for example allow memory BIOs to contain config info.
6646 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6647 Whoever hopes to achieve shared-library compatibility across versions
6648 must use this, not the compile-time macro.
6649 (Exercise 0.9.4: Which is the minimum library version required by
6651 Note: All this applies only to multi-threaded programs, others don't
6655 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6656 through a BIO pair triggered the default case, i.e.
6657 SSLerr(...,SSL_R_UNKNOWN_STATE).
6660 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6661 can use the SSL library even if none of the specific BIOs is
6665 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6666 for the encoded length.
6667 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6669 *) Add initial documentation of the X509V3 functions.
6672 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6673 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6674 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6675 secure PKCS#8 private key format with a high iteration count.
6678 *) Fix determination of Perl interpreter: A perl or perl5
6679 _directory_ in $PATH was also accepted as the interpreter.
6680 [Ralf S. Engelschall]
6682 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6683 wrong with it but it was very old and did things like calling
6684 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6688 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6689 to use the new extension code.
6692 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6693 with macros. This should make it easier to change their form, add extra
6694 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6698 *) Add to configuration table a new entry that can specify an alternative
6699 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6700 according to Mark Crispin <MRC@Panda.COM>.
6704 *) DES CBC did not update the IV. Weird.
6707 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6708 Changing the behaviour of the former might break existing programs --
6709 where IV updating is needed, des_ncbc_encrypt can be used.
6712 *) When bntest is run from "make test" it drives bc to check its
6713 calculations, as well as internally checking them. If an internal check
6714 fails, it needs to cause bc to give a non-zero result or make test carries
6715 on without noticing the failure. Fixed.
6718 *) DES library cleanups.
6721 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6722 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6723 ciphers. NOTE: although the key derivation function has been verified
6724 against some published test vectors it has not been extensively tested
6725 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6729 *) Instead of "mkdir -p", which is not fully portable, use new
6730 Perl script "util/mkdir-p.pl".
6733 *) Rewrite the way password based encryption (PBE) is handled. It used to
6734 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6735 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6736 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6737 the 'parameter' field of the AlgorithmIdentifier is passed to the
6738 underlying key generation function so it must do its own ASN1 parsing.
6739 This has also changed the EVP_PBE_CipherInit() function which now has a
6740 'parameter' argument instead of literal salt and iteration count values
6741 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6744 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6745 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6746 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6747 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6748 value was just used as a "magic string" and not used directly its
6749 value doesn't matter.
6752 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6756 *) "linux-sparc64" configuration (ultrapenguin).
6757 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6758 "linux-sparc" configuration.
6759 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6761 *) config now generates no-xxx options for missing ciphers.
6764 *) Support the EBCDIC character set (work in progress).
6765 File ebcdic.c not yet included because it has a different license.
6766 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6768 *) Support BS2000/OSD-POSIX.
6769 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6771 *) Make callbacks for key generation use void * instead of char *.
6774 *) Make S/MIME samples compile (not yet tested).
6777 *) Additional typesafe stacks.
6780 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6784 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6786 *) New configuration variant "sco5-gcc".
6788 *) Updated some demos.
6789 [Sean O Riordain, Wade Scholine]
6791 *) Add missing BIO_free at exit of pkcs12 application.
6794 *) Fix memory leak in conf.c.
6797 *) Updates for Win32 to assembler version of MD5.
6800 *) Set #! path to perl in apps/der_chop to where we found it
6801 instead of using a fixed path.
6804 *) SHA library changes for irix64-mips4-cc.
6807 *) Improvements for VMS support.
6811 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6813 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6814 This also avoids the problems with SC4.2 and unpatched SC5.
6815 [Andy Polyakov <appro@fy.chalmers.se>]
6817 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6818 These are required because of the typesafe stack would otherwise break
6819 existing code. If old code used a structure member which used to be STACK
6820 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6821 sk_num or sk_value it would produce an error because the num, data members
6822 are not present in STACK_OF. Now it just produces a warning. sk_set
6823 replaces the old method of assigning a value to sk_value
6824 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6825 that does this will no longer work (and should use sk_set instead) but
6826 this could be regarded as a "questionable" behaviour anyway.
6829 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6830 correctly handle encrypted S/MIME data.
6833 *) Change type of various DES function arguments from des_cblock
6834 (which means, in function argument declarations, pointer to char)
6835 to des_cblock * (meaning pointer to array with 8 char elements),
6836 which allows the compiler to do more typechecking; it was like
6837 that back in SSLeay, but with lots of ugly casts.
6839 Introduce new type const_des_cblock.
6842 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6843 problems: find RecipientInfo structure that matches recipient certificate
6844 and initialise the ASN1 structures properly based on passed cipher.
6847 *) Belatedly make the BN tests actually check the results.
6850 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6851 to and from BNs: it was completely broken. New compilation option
6852 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6853 key elements as negative integers.
6856 *) Reorganize and speed up MD5.
6857 [Andy Polyakov <appro@fy.chalmers.se>]
6860 [Richard Levitte <richard@levitte.org>]
6862 *) New option -out to asn1parse to allow the parsed structure to be
6863 output to a file. This is most useful when combined with the -strparse
6864 option to examine the output of things like OCTET STRINGS.
6867 *) Make SSL library a little more fool-proof by not requiring any longer
6868 that SSL_set_{accept,connect}_state be called before
6869 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6870 in many applications because usually everything *appeared* to work as
6871 intended anyway -- now it really works as intended).
6874 *) Move openssl.cnf out of lib/.
6877 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6878 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6879 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6880 [Ralf S. Engelschall]
6882 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6883 handle PKCS#7 enveloped data properly.
6884 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6886 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6887 copying pointers. The cert_st handling is changed by this in
6888 various ways (and thus what used to be known as ctx->default_cert
6889 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6890 any longer when s->cert does not give us what we need).
6891 ssl_cert_instantiate becomes obsolete by this change.
6892 As soon as we've got the new code right (possibly it already is?),
6893 we have solved a couple of bugs of the earlier code where s->cert
6894 was used as if it could not have been shared with other SSL structures.
6896 Note that using the SSL API in certain dirty ways now will result
6897 in different behaviour than observed with earlier library versions:
6898 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6899 does not influence s as it used to.
6901 In order to clean up things more thoroughly, inside SSL_SESSION
6902 we don't use CERT any longer, but a new structure SESS_CERT
6903 that holds per-session data (if available); currently, this is
6904 the peer's certificate chain and, for clients, the server's certificate
6905 and temporary key. CERT holds only those values that can have
6906 meaningful defaults in an SSL_CTX.
6909 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6910 from the internal representation. Various PKCS#7 fixes: remove some
6911 evil casts and set the enc_dig_alg field properly based on the signing
6915 *) Allow PKCS#12 password to be set from the command line or the
6916 environment. Let 'ca' get its config file name from the environment
6917 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6921 *) Allow certificate policies extension to use an IA5STRING for the
6922 organization field. This is contrary to the PKIX definition but
6923 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6927 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6928 without disallowing inline assembler and the like for non-pedantic builds.
6931 *) Support Borland C++ builder.
6932 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6937 *) SHA-1 cleanups and performance enhancements.
6938 [Andy Polyakov <appro@fy.chalmers.se>]
6940 *) Sparc v8plus assembler for the bignum library.
6941 [Andy Polyakov <appro@fy.chalmers.se>]
6943 *) Accept any -xxx and +xxx compiler options in Configure.
6946 *) Update HPUX configuration.
6949 *) Add missing sk_<type>_unshift() function to safestack.h
6950 [Ralf S. Engelschall]
6952 *) New function SSL_CTX_use_certificate_chain_file that sets the
6953 "extra_cert"s in addition to the certificate. (This makes sense
6954 only for "PEM" format files, as chains as a whole are not
6958 *) Support verify_depth from the SSL API.
6959 x509_vfy.c had what can be considered an off-by-one-error:
6960 Its depth (which was not part of the external interface)
6961 was actually counting the number of certificates in a chain;
6962 now it really counts the depth.
6965 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6966 instead of X509err, which often resulted in confusing error
6967 messages since the error codes are not globally unique
6968 (e.g. an alleged error in ssl3_accept when a certificate
6969 didn't match the private key).
6971 *) New function SSL_CTX_set_session_id_context that allows to set a default
6972 value (so that you don't need SSL_set_session_id_context for each
6973 connection using the SSL_CTX).
6976 *) OAEP decoding bug fix.
6979 *) Support INSTALL_PREFIX for package builders, as proposed by
6983 *) New Configure options "threads" and "no-threads". For systems
6984 where the proper compiler options are known (currently Solaris
6985 and Linux), "threads" is the default.
6988 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6991 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6992 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6993 such as /usr/local/bin.
6996 *) "make linux-shared" to build shared libraries.
6997 [Niels Poppe <niels@netbox.org>]
6999 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7002 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7003 extension adding in x509 utility.
7006 *) Remove NOPROTO sections and error code comments.
7009 *) Partial rewrite of the DEF file generator to now parse the ANSI
7013 *) New Configure options --prefix=DIR and --openssldir=DIR.
7016 *) Complete rewrite of the error code script(s). It is all now handled
7017 by one script at the top level which handles error code gathering,
7018 header rewriting and C source file generation. It should be much better
7019 than the old method: it now uses a modified version of Ulf's parser to
7020 read the ANSI prototypes in all header files (thus the old K&R definitions
7021 aren't needed for error creation any more) and do a better job of
7022 translating function codes into names. The old 'ASN1 error code imbedded
7023 in a comment' is no longer necessary and it doesn't use .err files which
7024 have now been deleted. Also the error code call doesn't have to appear all
7025 on one line (which resulted in some large lines...).
7028 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7031 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7032 0 (which usually indicates a closed connection), but continue reading.
7035 *) Fix some race conditions.
7038 *) Add support for CRL distribution points extension. Add Certificate
7039 Policies and CRL distribution points documentation.
7042 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7045 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7046 8 of keying material. Merlin has also confirmed interop with this fix
7047 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7048 [Merlin Hughes <merlin@baltimore.ie>]
7050 *) Fix lots of warnings.
7051 [Richard Levitte <levitte@stacken.kth.se>]
7053 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7054 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7055 [Richard Levitte <levitte@stacken.kth.se>]
7057 *) Fix problems with sizeof(long) == 8.
7058 [Andy Polyakov <appro@fy.chalmers.se>]
7060 *) Change functions to ANSI C.
7063 *) Fix typos in error codes.
7064 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7066 *) Remove defunct assembler files from Configure.
7069 *) SPARC v8 assembler BIGNUM implementation.
7070 [Andy Polyakov <appro@fy.chalmers.se>]
7072 *) Support for Certificate Policies extension: both print and set.
7073 Various additions to support the r2i method this uses.
7076 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7077 return a const string when you are expecting an allocated buffer.
7080 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7081 types DirectoryString and DisplayText.
7084 *) Add code to allow r2i extensions to access the configuration database,
7085 add an LHASH database driver and add several ctx helper functions.
7088 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7089 fail when they extended the size of a BIGNUM.
7092 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7093 support typesafe stack.
7096 *) Fix typo in SSL_[gs]et_options().
7097 [Nils Frostberg <nils@medcom.se>]
7099 *) Delete various functions and files that belonged to the (now obsolete)
7100 old X509V3 handling code.
7103 *) New Configure option "rsaref".
7106 *) Don't auto-generate pem.h.
7109 *) Introduce type-safe ASN.1 SETs.
7112 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7113 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7115 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7116 that links with OpenSSL (well at least cause lots of warnings), but fear
7117 not: the conversion is trivial, and it eliminates loads of evil casts. A
7118 few STACKed things have been converted already. Feel free to convert more.
7119 In the fullness of time, I'll do away with the STACK type altogether.
7122 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7123 specified in <certfile> by updating the entry in the index.txt file.
7124 This way one no longer has to edit the index.txt file manually for
7125 revoking a certificate. The -revoke option does the gory details now.
7126 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7128 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7129 `-text' option at all and this way the `-noout -text' combination was
7130 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7131 [Ralf S. Engelschall]
7133 *) Make sure a corresponding plain text error message exists for the
7134 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7135 verify callback function determined that a certificate was revoked.
7136 [Ralf S. Engelschall]
7138 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7139 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7140 all available cipers including rc5, which was forgotten until now.
7141 In order to let the testing shell script know which algorithms
7142 are available, a new (up to now undocumented) command
7143 "openssl list-cipher-commands" is used.
7146 *) Bugfix: s_client occasionally would sleep in select() when
7147 it should have checked SSL_pending() first.
7150 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7151 the raw DSA values prior to ASN.1 encoding.
7154 *) Tweaks to Configure
7155 [Niels Poppe <niels@netbox.org>]
7157 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7161 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7164 *) New config option to avoid instructions that are illegal on the 80386.
7165 The default code is faster, but requires at least a 486.
7168 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7169 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7170 same as SSL2_VERSION anyway.
7173 *) New "-showcerts" option for s_client.
7176 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7177 application. Various cleanups and fixes.
7180 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7181 modify error routines to work internally. Add error codes and PBE init
7182 to library startup routines.
7185 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7186 packing functions to asn1 and evp. Changed function names and error
7187 codes along the way.
7190 *) PKCS12 integration: and so it begins... First of several patches to
7191 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7192 objects to objects.h
7195 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7196 and display support for Thawte strong extranet extension.
7199 *) Add LinuxPPC support.
7200 [Jeff Dubrule <igor@pobox.org>]
7202 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7203 bn_div_words in alpha.s.
7204 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7206 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7207 OAEP isn't supported when OpenSSL is built with RSAref.
7208 [Ulf Moeller <ulf@fitug.de>]
7210 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7211 so they no longer are missing under -DNOPROTO.
7212 [Soren S. Jorvang <soren@t.dk>]
7215 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7217 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7218 doesn't work when the session is reused. Coming soon!
7221 *) Fix a security hole, that allows sessions to be reused in the wrong
7222 context thus bypassing client cert protection! All software that uses
7223 client certs and session caches in multiple contexts NEEDS PATCHING to
7224 allow session reuse! A fuller solution is in the works.
7225 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7227 *) Some more source tree cleanups (removed obsolete files
7228 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7229 permission on "config" script to be executable) and a fix for the INSTALL
7231 [Ulf Moeller <ulf@fitug.de>]
7233 *) Remove some legacy and erroneous uses of malloc, free instead of
7235 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7237 *) Make rsa_oaep_test return non-zero on error.
7238 [Ulf Moeller <ulf@fitug.de>]
7240 *) Add support for native Solaris shared libraries. Configure
7241 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7242 if someone would make that last step automatic.
7243 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7245 *) ctx_size was not built with the right compiler during "make links". Fixed.
7248 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7249 except NULL ciphers". This means the default cipher list will no longer
7250 enable NULL ciphers. They need to be specifically enabled e.g. with
7251 the string "DEFAULT:eNULL".
7254 *) Fix to RSA private encryption routines: if p < q then it would
7255 occasionally produce an invalid result. This will only happen with
7256 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7259 *) Be less restrictive and allow also `perl util/perlpath.pl
7260 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7261 because this way one can also use an interpreter named `perl5' (which is
7262 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7263 installed as `perl').
7264 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7266 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7267 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7269 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7270 advapi32.lib to Win32 build and change the pem test comparision
7271 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7272 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7273 and crypto/des/ede_cbcm_enc.c.
7276 *) DES quad checksum was broken on big-endian architectures. Fixed.
7279 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7280 Win32 test batch file so it (might) work again. The Win32 test batch file
7281 is horrible: I feel ill....
7284 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7285 in e_os.h. Audit of header files to check ANSI and non ANSI
7286 sections: 10 functions were absent from non ANSI section and not exported
7287 from Windows DLLs. Fixed up libeay.num for new functions.
7290 *) Make `openssl version' output lines consistent.
7291 [Ralf S. Engelschall]
7293 *) Fix Win32 symbol export lists for BIO functions: Added
7294 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7295 to ms/libeay{16,32}.def.
7296 [Ralf S. Engelschall]
7298 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7299 fine under Unix and passes some trivial tests I've now added. But the
7300 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7301 added to make sure no one expects that this stuff really works in the
7302 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7303 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7305 [Ralf S. Engelschall]
7307 *) Fix the generation of two part addresses in perl.
7308 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7310 *) Add config entry for Linux on MIPS.
7311 [John Tobey <jtobey@channel1.com>]
7313 *) Make links whenever Configure is run, unless we are on Windoze.
7316 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7317 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7321 *) Add a useful kludge to allow package maintainers to specify compiler and
7322 other platforms details on the command line without having to patch the
7323 Configure script everytime: One now can use ``perl Configure
7324 <id>:<details>'', i.e. platform ids are allowed to have details appended
7325 to them (seperated by colons). This is treated as there would be a static
7326 pre-configured entry in Configure's %table under key <id> with value
7327 <details> and ``perl Configure <id>'' is called. So, when you want to
7328 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7329 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7330 now, which overrides the FreeBSD-elf entry on-the-fly.
7331 [Ralf S. Engelschall]
7333 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7336 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7337 on the `perl Configure ...' command line. This way one can compile
7338 OpenSSL libraries with Position Independent Code (PIC) which is needed
7339 for linking it into DSOs.
7340 [Ralf S. Engelschall]
7342 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7346 *) Cleaned up the LICENSE document: The official contact for any license
7347 questions now is the OpenSSL core team under openssl-core@openssl.org.
7348 And add a paragraph about the dual-license situation to make sure people
7349 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7350 to the OpenSSL toolkit.
7351 [Ralf S. Engelschall]
7353 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7354 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7355 Additonally cleaned up the `make links' target: Remove unnecessary
7356 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7357 to speed processing and no longer clutter the display with confusing
7358 stuff. Instead only the actually done links are displayed.
7359 [Ralf S. Engelschall]
7361 *) Permit null encryption ciphersuites, used for authentication only. It used
7362 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7363 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7367 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7368 signed attributes when verifying signatures (this would break them),
7369 the detached data encoding was wrong and public keys obtained using
7370 X509_get_pubkey() weren't freed.
7373 *) Add text documentation for the BUFFER functions. Also added a work around
7374 to a Win95 console bug. This was triggered by the password read stuff: the
7375 last character typed gets carried over to the next fread(). If you were
7376 generating a new cert request using 'req' for example then the last
7377 character of the passphrase would be CR which would then enter the first
7381 *) Added the new `Includes OpenSSL Cryptography Software' button as
7382 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7383 button and can be used by applications based on OpenSSL to show the
7384 relationship to the OpenSSL project.
7385 [Ralf S. Engelschall]
7387 *) Remove confusing variables in function signatures in files
7388 ssl/ssl_lib.c and ssl/ssl.h.
7389 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7391 *) Don't install bss_file.c under PREFIX/include/
7392 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7394 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7395 functions that return function pointers and has support for NT specific
7396 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7397 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7398 unsigned to signed types: this was killing the Win32 compile.
7401 *) Add new certificate file to stack functions,
7402 SSL_add_dir_cert_subjects_to_stack() and
7403 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7404 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7405 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7406 This means that Apache-SSL and similar packages don't have to mess around
7407 to add as many CAs as they want to the preferred list.
7410 *) Experiment with doxygen documentation. Currently only partially applied to
7412 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7413 openssl.doxy as the configuration file.
7416 *) Get rid of remaining C++-style comments which strict C compilers hate.
7417 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7419 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7420 compiled in by default: it has problems with large keys.
7423 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7424 DH private keys and/or callback functions which directly correspond to
7425 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7426 is needed for applications which have to configure certificates on a
7427 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7429 For the RSA certificate situation is makes no difference, but
7430 for the DSA certificate situation this fixes the "no shared cipher"
7431 problem where the OpenSSL cipher selection procedure failed because the
7432 temporary keys were not overtaken from the context and the API provided
7433 no way to reconfigure them.
7434 The new functions now let applications reconfigure the stuff and they
7435 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7436 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7437 non-public-API function ssl_cert_instantiate() is used as a helper
7438 function and also to reduce code redundancy inside ssl_rsa.c.
7439 [Ralf S. Engelschall]
7441 *) Move s_server -dcert and -dkey options out of the undocumented feature
7442 area because they are useful for the DSA situation and should be
7443 recognized by the users.
7444 [Ralf S. Engelschall]
7446 *) Fix the cipher decision scheme for export ciphers: the export bits are
7447 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7448 SSL_EXP_MASK. So, the original variable has to be used instead of the
7449 already masked variable.
7450 [Richard Levitte <levitte@stacken.kth.se>]
7452 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7453 [Richard Levitte <levitte@stacken.kth.se>]
7455 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7456 from `int' to `unsigned int' because it's a length and initialized by
7457 EVP_DigestFinal() which expects an `unsigned int *'.
7458 [Richard Levitte <levitte@stacken.kth.se>]
7460 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7461 script. Instead use the usual Shell->Perl transition trick.
7462 [Ralf S. Engelschall]
7464 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7465 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7466 -noout -modulus' as it's already the case for `openssl rsa -noout
7467 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7468 currently the public key is printed (a decision which was already done by
7469 `openssl dsa -modulus' in the past) which serves a similar purpose.
7470 Additionally the NO_RSA no longer completely removes the whole -modulus
7471 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7473 [Ralf S. Engelschall]
7475 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7476 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7477 [Arne Ansper <arne@ats.cyber.ee>]
7479 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7480 to be added. Now both 'req' and 'ca' can use new objects defined in the
7484 *) Add cool BIO that does syslog (or event log on NT).
7485 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7487 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7488 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7489 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7490 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7493 *) Add preliminary config info for new extension code.
7496 *) Make RSA_NO_PADDING really use no padding.
7497 [Ulf Moeller <ulf@fitug.de>]
7499 *) Generate errors when private/public key check is done.
7502 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7503 for some CRL extensions and new objects added.
7506 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7507 key usage extension and fuller support for authority key id.
7510 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7511 padding method for RSA, which is recommended for new applications in PKCS
7512 #1 v2.0 (RFC 2437, October 1998).
7513 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7514 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7515 against Bleichbacher's attack on RSA.
7516 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7519 *) Updates to the new SSL compression code
7520 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7522 *) Fix so that the version number in the master secret, when passed
7523 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7524 (because the server will not accept higher), that the version number
7525 is 0x03,0x01, not 0x03,0x00
7526 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7528 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7529 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7530 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7533 *) Support for RAW extensions where an arbitrary extension can be
7534 created by including its DER encoding. See apps/openssl.cnf for
7538 *) Make sure latest Perl versions don't interpret some generated C array
7539 code as Perl array code in the crypto/err/err_genc.pl script.
7540 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7542 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7543 not many people have the assembler. Various Win32 compilation fixes and
7544 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7548 *) Modify configure script 'Configure' to automatically create crypto/date.h
7549 file under Win32 and also build pem.h from pem.org. New script
7550 util/mkfiles.pl to create the MINFO file on environments that can't do a
7551 'make files': perl util/mkfiles.pl >MINFO should work.
7554 *) Major rework of DES function declarations, in the pursuit of correctness
7555 and purity. As a result, many evil casts evaporated, and some weirdness,
7556 too. You may find this causes warnings in your code. Zapping your evil
7557 casts will probably fix them. Mostly.
7560 *) Fix for a typo in asn1.h. Bug fix to object creation script
7561 obj_dat.pl. It considered a zero in an object definition to mean
7562 "end of object": none of the objects in objects.h have any zeros
7563 so it wasn't spotted.
7564 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7566 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7567 Masking (CBCM). In the absence of test vectors, the best I have been able
7568 to do is check that the decrypt undoes the encrypt, so far. Send me test
7569 vectors if you have them.
7572 *) Correct calculation of key length for export ciphers (too much space was
7573 allocated for null ciphers). This has not been tested!
7576 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7577 message is now correct (it understands "crypto" and "ssl" on its
7578 command line). There is also now an "update" option. This will update
7579 the util/ssleay.num and util/libeay.num files with any new functions.
7581 perl util/mkdef.pl crypto ssl update
7582 it will update them.
7585 *) Overhauled the Perl interface (perl/*):
7586 - ported BN stuff to OpenSSL's different BN library
7587 - made the perl/ source tree CVS-aware
7588 - renamed the package from SSLeay to OpenSSL (the files still contain
7589 their history because I've copied them in the repository)
7590 - removed obsolete files (the test scripts will be replaced
7591 by better Test::Harness variants in the future)
7592 [Ralf S. Engelschall]
7594 *) First cut for a very conservative source tree cleanup:
7595 1. merge various obsolete readme texts into doc/ssleay.txt
7596 where we collect the old documents and readme texts.
7597 2. remove the first part of files where I'm already sure that we no
7598 longer need them because of three reasons: either they are just temporary
7599 files which were left by Eric or they are preserved original files where
7600 I've verified that the diff is also available in the CVS via "cvs diff
7601 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7602 the crypto/md/ stuff).
7603 [Ralf S. Engelschall]
7605 *) More extension code. Incomplete support for subject and issuer alt
7606 name, issuer and authority key id. Change the i2v function parameters
7607 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7608 what that's for :-) Fix to ASN1 macro which messed up
7609 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7612 *) Preliminary support for ENUMERATED type. This is largely copied from the
7616 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7617 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7619 *) Make sure `make rehash' target really finds the `openssl' program.
7620 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7622 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7623 like to hear about it if this slows down other processors.
7626 *) Add CygWin32 platform information to Configure script.
7627 [Alan Batie <batie@aahz.jf.intel.com>]
7629 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7630 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7632 *) New program nseq to manipulate netscape certificate sequences
7635 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7639 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7640 but the BN code had some problems that would cause failures when
7641 doing certificate verification and some other functions.
7642 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7644 *) Add ASN1 and PEM code to support netscape certificate sequences.
7647 *) Add ASN1 and PEM code to support netscape certificate sequences.
7650 *) Add several PKIX and private extended key usage OIDs.
7653 *) Modify the 'ca' program to handle the new extension code. Modify
7654 openssl.cnf for new extension format, add comments.
7657 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7658 and add a sample to openssl.cnf so req -x509 now adds appropriate
7662 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7663 error code, add initial support to X509_print() and x509 application.
7666 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7667 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7668 stuff is currently isolated and isn't even compiled yet.
7671 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7672 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7673 Removed the versions check from X509 routines when loading extensions:
7674 this allows certain broken certificates that don't set the version
7675 properly to be processed.
7678 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7679 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7680 can still be regenerated with "make depend".
7683 *) Spelling mistake in C version of CAST-128.
7684 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7686 *) Changes to the error generation code. The perl script err-code.pl
7687 now reads in the old error codes and retains the old numbers, only
7688 adding new ones if necessary. It also only changes the .err files if new
7689 codes are added. The makefiles have been modified to only insert errors
7690 when needed (to avoid needlessly modifying header files). This is done
7691 by only inserting errors if the .err file is newer than the auto generated
7692 C file. To rebuild all the error codes from scratch (the old behaviour)
7693 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7694 or delete all the .err files.
7697 *) CAST-128 was incorrectly implemented for short keys. The C version has
7698 been fixed, but is untested. The assembler versions are also fixed, but
7699 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7700 to regenerate it if needed.
7701 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7702 Hagino <itojun@kame.net>]
7704 *) File was opened incorrectly in randfile.c.
7705 [Ulf Möller <ulf@fitug.de>]
7707 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7708 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7709 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7710 al: it's just almost always a UTCTime. Note this patch adds new error
7711 codes so do a "make errors" if there are problems.
7714 *) Correct Linux 1 recognition in config.
7715 [Ulf Möller <ulf@fitug.de>]
7717 *) Remove pointless MD5 hash when using DSA keys in ca.
7718 [Anonymous <nobody@replay.com>]
7720 *) Generate an error if given an empty string as a cert directory. Also
7721 generate an error if handed NULL (previously returned 0 to indicate an
7722 error, but didn't set one).
7723 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7725 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7728 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7729 parameters. This was causing a warning which killed off the Win32 compile.
7732 *) Remove C++ style comments from crypto/bn/bn_local.h.
7733 [Neil Costigan <neil.costigan@celocom.com>]
7735 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7736 based on a text string, looking up short and long names and finally
7737 "dot" format. The "dot" format stuff didn't work. Added new function
7738 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7739 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7740 OID is not part of the table.
7743 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7744 X509_LOOKUP_by_alias().
7747 *) Sort openssl functions by name.
7750 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7751 encryption from sample DSA keys (in case anyone is interested the password
7755 *) Make _all_ *_free functions accept a NULL pointer.
7756 [Frans Heymans <fheymans@isaserver.be>]
7758 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7760 [Anonymous <nobody@replay.com>]
7762 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7763 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7765 *) Don't blow it for numeric -newkey arguments to apps/req.
7766 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7768 *) Temp key "for export" tests were wrong in s3_srvr.c.
7769 [Anonymous <nobody@replay.com>]
7771 *) Add prototype for temp key callback functions
7772 SSL_CTX_set_tmp_{rsa,dh}_callback().
7775 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7776 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7779 *) X509_name_add_entry() freed the wrong thing after an error.
7780 [Arne Ansper <arne@ats.cyber.ee>]
7782 *) rsa_eay.c would attempt to free a NULL context.
7783 [Arne Ansper <arne@ats.cyber.ee>]
7785 *) BIO_s_socket() had a broken should_retry() on Windoze.
7786 [Arne Ansper <arne@ats.cyber.ee>]
7788 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7789 [Arne Ansper <arne@ats.cyber.ee>]
7791 *) Make sure the already existing X509_STORE->depth variable is initialized
7792 in X509_STORE_new(), but document the fact that this variable is still
7793 unused in the certificate verification process.
7794 [Ralf S. Engelschall]
7796 *) Fix the various library and apps files to free up pkeys obtained from
7797 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7800 *) Fix reference counting in X509_PUBKEY_get(). This makes
7801 demos/maurice/example2.c work, amongst others, probably.
7802 [Steve Henson and Ben Laurie]
7804 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7805 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7806 are no longer created. This way we have a single and consistent command
7807 line interface `openssl <command>', similar to `cvs <command>'.
7808 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7810 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7811 BIT STRING wrapper always have zero unused bits.
7814 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7817 *) Make the top-level INSTALL documentation easier to understand.
7820 *) Makefiles updated to exit if an error occurs in a sub-directory
7821 make (including if user presses ^C) [Paul Sutton]
7823 *) Make Montgomery context stuff explicit in RSA data structure.
7826 *) Fix build order of pem and err to allow for generated pem.h.
7829 *) Fix renumbering bug in X509_NAME_delete_entry().
7832 *) Enhanced the err-ins.pl script so it makes the error library number
7833 global and can add a library name. This is needed for external ASN1 and
7834 other error libraries.
7837 *) Fixed sk_insert which never worked properly.
7840 *) Fix ASN1 macros so they can handle indefinite length construted
7841 EXPLICIT tags. Some non standard certificates use these: they can now
7845 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7846 into a single doc/ssleay.txt bundle. This way the information is still
7847 preserved but no longer messes up this directory. Now it's new room for
7848 the new set of documenation files.
7849 [Ralf S. Engelschall]
7851 *) SETs were incorrectly DER encoded. This was a major pain, because they
7852 shared code with SEQUENCEs, which aren't coded the same. This means that
7853 almost everything to do with SETs or SEQUENCEs has either changed name or
7854 number of arguments.
7855 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7857 *) Fix test data to work with the above.
7860 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7861 was already fixed by Eric for 0.9.1 it seems.
7862 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7864 *) Autodetect FreeBSD3.
7867 *) Fix various bugs in Configure. This affects the following platforms:
7871 unixware-2.0-pentium
7875 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7876 before they are needed.
7879 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7883 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7885 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7886 changed SSLeay to OpenSSL in version strings.
7887 [Ralf S. Engelschall]
7889 *) Some fixups to the top-level documents.
7892 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7893 because the symlink to include/ was missing.
7894 [Ralf S. Engelschall]
7896 *) Incorporated the popular no-RSA/DSA-only patches
7897 which allow to compile a RSA-free SSLeay.
7898 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7900 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7901 when "ssleay" is still not found.
7902 [Ralf S. Engelschall]
7904 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7905 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7907 *) Updated the README file.
7908 [Ralf S. Engelschall]
7910 *) Added various .cvsignore files in the CVS repository subdirs
7911 to make a "cvs update" really silent.
7912 [Ralf S. Engelschall]
7914 *) Recompiled the error-definition header files and added
7915 missing symbols to the Win32 linker tables.
7916 [Ralf S. Engelschall]
7918 *) Cleaned up the top-level documents;
7919 o new files: CHANGES and LICENSE
7920 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7921 o merged COPYRIGHT into LICENSE
7922 o removed obsolete TODO file
7923 o renamed MICROSOFT to INSTALL.W32
7924 [Ralf S. Engelschall]
7926 *) Removed dummy files from the 0.9.1b source tree:
7927 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7928 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7929 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7930 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7931 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7932 [Ralf S. Engelschall]
7934 *) Added various platform portability fixes.
7937 *) The Genesis of the OpenSSL rpject:
7938 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7939 Young and Tim J. Hudson created while they were working for C2Net until
7941 [The OpenSSL Project]
7944 Changes between 0.9.0b and 0.9.1b [not released]
7946 *) Updated a few CA certificates under certs/
7949 *) Changed some BIGNUM api stuff.
7952 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7953 DGUX x86, Linux Alpha, etc.
7956 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7957 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7961 *) Add -strparse option to asn1pars program which parses nested
7963 [Dr Stephen Henson <shenson@bigfoot.com>]
7965 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7968 *) DSA fix for "ca" program.
7971 *) Added "-genkey" option to "dsaparam" program.
7974 *) Added RIPE MD160 (rmd160) message digest.
7977 *) Added -a (all) option to "ssleay version" command.
7980 *) Added PLATFORM define which is the id given to Configure.
7983 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7986 *) Extended the ASN.1 parser routines.
7989 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7992 *) Added a BN_CTX to the BN library.
7995 *) Fixed the weak key values in DES library
7998 *) Changed API in EVP library for cipher aliases.
8001 *) Added support for RC2/64bit cipher.
8004 *) Converted the lhash library to the crypto/mem.c functions.
8007 *) Added more recognized ASN.1 object ids.
8010 *) Added more RSA padding checks for SSL/TLS.
8013 *) Added BIO proxy/filter functionality.
8016 *) Added extra_certs to SSL_CTX which can be used
8017 send extra CA certificates to the client in the CA cert chain sending
8018 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8021 *) Now Fortezza is denied in the authentication phase because
8022 this is key exchange mechanism is not supported by SSLeay at all.
8025 *) Additional PKCS1 checks.
8028 *) Support the string "TLSv1" for all TLS v1 ciphers.
8031 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8032 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8035 *) Fixed a few memory leaks.
8038 *) Fixed various code and comment typos.
8041 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8042 bytes sent in the client random.
8043 [Edward Bishop <ebishop@spyglass.com>]