CHANGES revision 162911
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 6162911Ssimon 7162911Ssimon *) Introduce limits to prevent malicious keys being able to 8162911Ssimon cause a denial of service. (CVE-2006-2940) 9162911Ssimon [Steve Henson, Bodo Moeller] 10162911Ssimon 11162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 12162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 13162911Ssimon 14162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 15162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 16162911Ssimon 17162911Ssimon *) Fix SSL client code which could crash if connecting to a 18162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 19162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 20162911Ssimon 21162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 22162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 23162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 24162911Ssimon the same strength classification in 0.9.7h) as we currently only 25162911Ssimon have a single AES bit in the ciphersuite description bitmap. 26162911Ssimon That change, however, also applied to ciphersuite strings such as 27162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 28162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 29162911Ssimon from SSL 3.0/TLS 1.0. 30162911Ssimon 31162911Ssimon So we change the selection algorithm again: Naming an explicit 32162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 33162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 34162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 35162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 36162911Ssimon 37162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 38162911Ssimon 128/256 bit distinction would be relevant, this works for now. 39162911Ssimon The proper fix will be to use different bits for AES128 and 40162911Ssimon AES256, which would have avoided the problems from the beginning; 41162911Ssimon however, bits are scarce, so we can only do this in a new release 42162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 43162911Ssimon definition to split the single 'unsigned long mask' bitmap into 44162911Ssimon multiple values to extend the available space. 45162911Ssimon 46162911Ssimon [Bodo Moeller] 47162911Ssimon 48162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 49162911Ssimon 50162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 51162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 52162911Ssimon 53162911Ssimon *) Add AES IGE and biIGE modes. 54162911Ssimon [Ben Laurie] 55162911Ssimon 56162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 57162911Ssimon possible instead of select(), since the latter has some 58162911Ssimon undesirable limitations. 59162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 60162911Ssimon 61162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 62162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 63162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 64162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 65162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 66162911Ssimon support, which is required for curve and point format negotiation 67162911Ssimon to avoid potential handshake problems. 68162911Ssimon [Bodo Moeller] 69162911Ssimon 70162911Ssimon *) Disable rogue ciphersuites: 71162911Ssimon 72162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 73162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 74162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 75162911Ssimon 76162911Ssimon The latter two were purportedly from 77162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 78162911Ssimon appear there. 79162911Ssimon 80162911Ssimon Also deactive the remaining ciphersuites from 81162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 82162911Ssimon unofficial, and the ID has long expired. 83162911Ssimon [Bodo Moeller] 84162911Ssimon 85162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 86162911Ssimon dual-core machines) and other potential thread-safety issues. 87162911Ssimon [Bodo Moeller] 88162911Ssimon 89162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 90162911Ssimon versions), which is now available for royalty-free use 91162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 92162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 93162911Ssimon 94162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 95162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 96162911Ssimon is configured with 'enable-camellia'. 97162911Ssimon [NTT] 98162911Ssimon 99162911Ssimon *) Disable the padding bug check when compression is in use. The padding 100162911Ssimon bug check assumes the first packet is of even length, this is not 101162911Ssimon necessarily true if compresssion is enabled and can result in false 102162911Ssimon positives causing handshake failure. The actual bug test is ancient 103162911Ssimon code so it is hoped that implementations will either have fixed it by 104162911Ssimon now or any which still have the bug do not support compression. 105162911Ssimon [Steve Henson] 106162911Ssimon 107160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 108160814Ssimon 109160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 110160814Ssimon cipher suite and only match that one cipher suite if it is. 111160814Ssimon [Steve Henson] 112160814Ssimon 113160814Ssimon *) Link in manifests for VC++ if needed. 114160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 115160814Ssimon 116160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 117160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 118160814Ssimon TLS extensions, which are supported starting with the 0.9.9 119160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 120160814Ssimon [Douglas Stebila] 121160814Ssimon 122160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 123160814Ssimon opaque EVP_CIPHER_CTX handling. 124160814Ssimon [Steve Henson] 125160814Ssimon 126160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 127160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 128160814Ssimon to conform with the standards mentioned here: 129160814Ssimon http://www.zlib.net/DLL_FAQ.txt 130160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 131160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 132160814Ssimon of the headers and library. Gracefully handle case where zlib library 133160814Ssimon can't be loaded. 134160814Ssimon [Steve Henson] 135160814Ssimon 136160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 137160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 138160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 139160814Ssimon non standard OBJ_obj2txt() behaviour. 140160814Ssimon [Steve Henson] 141160814Ssimon 142160814Ssimon *) Add support for building of engines under engine/ as shared libraries 143160814Ssimon under VC++ build system. 144160814Ssimon [Steve Henson] 145160814Ssimon 146160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 147160814Ssimon Hopefully, we will not see any false combination of paths any more. 148160814Ssimon [Richard Levitte] 149160814Ssimon 150160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 151160814Ssimon 152160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 153160814Ssimon (part of SSL_OP_ALL). This option used to disable the 154160814Ssimon countermeasure against man-in-the-middle protocol-version 155160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 156160814Ssimon idea. (CVE-2005-2969) 157160814Ssimon 158160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 159160814Ssimon for Information Security, National Institute of Advanced Industrial 160160814Ssimon Science and Technology [AIST], Japan)] 161160814Ssimon 162160814Ssimon *) Add two function to clear and return the verify parameter flags. 163160814Ssimon [Steve Henson] 164160814Ssimon 165160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 166160814Ssimon runtime, thus removing the need for a lock. 167160814Ssimon [Nils Larsch] 168160814Ssimon 169160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 170160814Ssimon [Nick Mathewson and Ben Laurie] 171160814Ssimon 172160814Ssimon *) Add functions for well-known primes. 173160814Ssimon [Nick Mathewson] 174160814Ssimon 175160814Ssimon *) Extended Windows CE support. 176160814Ssimon [Satoshi Nakamura and Andy Polyakov] 177160814Ssimon 178160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 179160814Ssimon runtime, thus removing the need for a lock. 180160814Ssimon [Steve Henson] 181160814Ssimon 182160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 183160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 184160814Ssimon smime utility. 185160814Ssimon [Steve Henson] 186160814Ssimon 187160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 188160814Ssimon 189162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 190162911Ssimon OpenSSL 0.9.8.] 191162911Ssimon 192160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 193160814Ssimon [Richard Levitte] 194160814Ssimon 195160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 196160814Ssimon key into the same file any more. 197160814Ssimon [Richard Levitte] 198160814Ssimon 199160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 200160814Ssimon [Andy Polyakov] 201160814Ssimon 202160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 203160814Ssimon [Stefan <stf@udoma.org] 204160814Ssimon 205160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 206160814Ssimon libraries. Use DES_crypt(). 207160814Ssimon [Richard Levitte] 208160814Ssimon 209160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 210160814Ssimon involves renaming the source and generated shared-libs for 211160814Ssimon both. The engines will accept the corrected or legacy ids 212160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 213160814Ssimon this only applies when building 'shared'. 214160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 215160814Ssimon 216160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 217160814Ssimon PKCS12_create() to recognize a CSP name attribute and 218160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 219160814Ssimon [Steve Henson] 220160814Ssimon 221160814Ssimon *) Add new functionality to the bn blinding code: 222160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 223160814Ssimon a fixed number of uses (currently 32) 224160814Ssimon - add new function for parameter creation 225160814Ssimon - introduce flags to control the update behaviour of the 226160814Ssimon BN_BLINDING parameters 227160814Ssimon - hide BN_BLINDING structure 228160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 229160814Ssimon performance when a single RSA object is shared among several 230160814Ssimon threads. 231160814Ssimon [Nils Larsch] 232160814Ssimon 233160814Ssimon *) Add support for DTLS. 234160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 235160814Ssimon 236160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 237160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 238160814Ssimon [Walter Goulet] 239160814Ssimon 240160814Ssimon *) Remove buggy and incompletet DH cert support from 241160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 242160814Ssimon [Nils Larsch] 243160814Ssimon 244160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 245160814Ssimon the apps/openssl applications. 246160814Ssimon [Nils Larsch] 247160814Ssimon 248160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 249160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 250160814Ssimon DEBUG_SAFESTACK must also be set. 251160814Ssimon [Ben Laurie] 252160814Ssimon 253160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 254160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 255160814Ssimon 256160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 257160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 258160814Ssimon 259160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 260160814Ssimon is frequently required for interoperability, and there is no license 261160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 262160814Ssimon avoid this algorithm.) 263160814Ssimon 264160814Ssimon [Bodo Moeller] 265160814Ssimon 266160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 267160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 268160814Ssimon EGEE (Enabling Grids for E-science in Europe). 269160814Ssimon [Richard Levitte] 270160814Ssimon 271160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 272160814Ssimon as Intel P4, IA-64 and AMD64. 273160814Ssimon [Andy Polyakov] 274160814Ssimon 275160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 276160814Ssimon section number in a pod file instead of having to treat each file as 277160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 278160814Ssimon pod file: 279160814Ssimon 280160814Ssimon =for comment openssl_section:XXX 281160814Ssimon 282160814Ssimon The blank line is mandatory. 283160814Ssimon 284160814Ssimon [Steve Henson] 285160814Ssimon 286160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 287160814Ssimon to allow alternative format key and certificate files and passphrase 288160814Ssimon sources. 289160814Ssimon [Steve Henson] 290160814Ssimon 291160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 292160814Ssimon update associated structures and add various utility functions. 293160814Ssimon 294160814Ssimon Add new policy related verify parameters, include policy checking in 295160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 296160814Ssimon to support policy checking and print out. 297160814Ssimon [Steve Henson] 298160814Ssimon 299160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 300160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 301160814Ssimon as well as RNG (though RNG support is currently disabled). 302160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 303160814Ssimon 304160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 305160814Ssimon [Geoff Thorpe] 306160814Ssimon 307160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 308160814Ssimon [Andy Polyakov and a number of other people] 309160814Ssimon 310160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 311160814Ssimon implementation contributed by IBM. 312160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 313160814Ssimon 314160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 315160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 316160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 317160814Ssimon [Jelte Jansen, Geoff Thorpe] 318160814Ssimon 319160814Ssimon *) Functionality for creating the initial serial number file is now 320160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 321160814Ssimon 322160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 323160814Ssimon number file to 1, which is bound to cause problems. To avoid 324160814Ssimon the problems while respecting compatibility between different 0.9.7 325160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 326160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 327160814Ssimon we can fix the problem directly in the 'ca' utility.) 328160814Ssimon [Steve Henson] 329160814Ssimon 330160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 331160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 332160814Ssimon give fewer recursive includes, which could break lazy source code - so 333160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 334160814Ssimon developers should define this symbol when building and using openssl to 335160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 336160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 337160814Ssimon [Geoff Thorpe] 338160814Ssimon 339160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 340160814Ssimon [Steve Henson] 341160814Ssimon 342160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 343160814Ssimon This will generate a random key of the appropriate length based on the 344160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 345160814Ssimon routine to support keys of a specific form. This is used in the des and 346160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 347160814Ssimon code to use new functions and hence generate correct parity DES keys. 348160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 349160814Ssimon valid (weak or incorrect parity). 350160814Ssimon [Steve Henson] 351160814Ssimon 352160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 353160814Ssimon as looking them up. This is useful when the verified structure may contain 354160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 355160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 356160814Ssimon [Steve Henson] 357160814Ssimon 358160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 359160814Ssimon syntax: 360160814Ssimon 361160814Ssimon shortName = some long name, 1.2.3.4 362160814Ssimon [Steve Henson] 363160814Ssimon 364160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 365160814Ssimon limitation on the number of variables it can handle nor the depth of the 366160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 367160814Ssimon information can now expand as required, and rather than having a single 368160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 369160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 370160814Ssimon BN_CTX's "bundling". 371160814Ssimon [Geoff Thorpe] 372160814Ssimon 373160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 374160814Ssimon to allow all RSA operations to function using a single BN_CTX. 375160814Ssimon [Geoff Thorpe] 376160814Ssimon 377160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 378160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 379160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 380160814Ssimon [Steve Henson] 381160814Ssimon 382160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 383160814Ssimon remained unused and not that useful. A variety of other little bignum 384160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 385160814Ssimon below). 386160814Ssimon [Geoff Thorpe] 387160814Ssimon 388160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 389160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 390160814Ssimon [Richard Levitte] 391160814Ssimon 392160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 393160814Ssimon and this should never fail. So the return value from the use of 394160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 395160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 396160814Ssimon [Geoff Thorpe] 397160814Ssimon 398160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 399160814Ssimon initialised value as BN_new(). 400160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 401160814Ssimon 402160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 403160814Ssimon [Steve Henson] 404160814Ssimon 405160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 406160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 407160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 408160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 409160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 410160814Ssimon structures to try and expose faulty code further on. For now, openssl will 411160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 412160814Ssimon forms that it has tolerated in the past, but authors and packagers should 413160814Ssimon consider trying openssl and their own applications when compiled with 414160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 415160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 416160814Ssimon some point, these tighter rules will become openssl's default to improve 417160814Ssimon maintainability, though the assert()s and other overheads will remain only 418160814Ssimon in debugging configurations. See bn.h for more details. 419160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 420160814Ssimon 421160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 422160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 423160814Ssimon initialises it). The presence of this function only made it possible 424160814Ssimon to overwrite an existing structure (and cause memory leaks). 425160814Ssimon [Geoff Thorpe] 426160814Ssimon 427160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 428160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 429160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 430160814Ssimon to clean up those corresponding objects before destroying the hash table 431160814Ssimon (and losing the object pointers). So some over-zealous constifications in 432160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 433160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 434160814Ssimon prototyped to have "const" restrictions on the object pointers they are 435160814Ssimon given (and so aren't required to cast them away any more). 436160814Ssimon [Geoff Thorpe] 437160814Ssimon 438160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 439160814Ssimon (speed) prefers to use its own implementation. The two implementations 440160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 441160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 442160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 443160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 444160814Ssimon internally to the implementation so I've used that for now. 445160814Ssimon [Geoff Thorpe] 446160814Ssimon 447160814Ssimon *) Ensure that deprecated functions do not get compiled when 448160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 449160814Ssimon the self-tests were still using deprecated key-generation functions so 450160814Ssimon these have been updated also. 451160814Ssimon [Geoff Thorpe] 452160814Ssimon 453160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 454160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 455160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 456160814Ssimon digestedData type. Add additional code to correctly generate the 457160814Ssimon digestedData type and add support for this type in PKCS7 initialization 458160814Ssimon functions. 459160814Ssimon [Steve Henson] 460160814Ssimon 461160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 462160814Ssimon structure of type "other". 463160814Ssimon [Steve Henson] 464160814Ssimon 465160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 466160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 467160814Ssimon modulus operations are not performed. The (pre-generated) prime 468160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 469160814Ssimon re-generated on some platforms because of the "division by zero" 470160814Ssimon situation in the script. 471160814Ssimon [Ralf S. Engelschall] 472160814Ssimon 473160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 474160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 475160814Ssimon SHA-1 now is only used for "small" curves (where the 476160814Ssimon representation of a field element takes up to 24 bytes); for 477160814Ssimon larger curves, the field element resulting from ECDH is directly 478160814Ssimon used as premaster secret. 479160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 480160814Ssimon 481160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 482160814Ssimon curve secp160r1 to the tests. 483160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 484160814Ssimon 485160814Ssimon *) Add the possibility to load symbols globally with DSO. 486160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 487160814Ssimon 488160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 489160814Ssimon control of the error stack. 490160814Ssimon [Richard Levitte] 491160814Ssimon 492160814Ssimon *) Add support for STORE in ENGINE. 493160814Ssimon [Richard Levitte] 494160814Ssimon 495160814Ssimon *) Add the STORE type. The intention is to provide a common interface 496160814Ssimon to certificate and key stores, be they simple file-based stores, or 497160814Ssimon HSM-type store, or LDAP stores, or... 498160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 499160814Ssimon [Richard Levitte] 500160814Ssimon 501160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 502160814Ssimon pass a list of arguments to any function as well as provide a way 503160814Ssimon for a function to pass data back to the caller. 504160814Ssimon [Richard Levitte] 505160814Ssimon 506160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 507160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 508160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 509160814Ssimon a memory area. 510160814Ssimon [Richard Levitte] 511160814Ssimon 512160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 513160814Ssimon return an index to an element even if an exact match couldn't be 514160814Ssimon found. The index is guaranteed to point at the element where the 515160814Ssimon searched-for key would be inserted to preserve sorting order. 516160814Ssimon [Richard Levitte] 517160814Ssimon 518160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 519160814Ssimon takes an extra flags argument for optional functionality. Currently, 520160814Ssimon the following flags are defined: 521160814Ssimon 522160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 523160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 524160814Ssimon element where the comparing function returns a negative or zero 525160814Ssimon number. 526160814Ssimon 527160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 528160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 529160814Ssimon element where the comparing function returns zero. This is useful 530160814Ssimon if there are more than one element where the comparing function 531160814Ssimon returns zero. 532160814Ssimon [Richard Levitte] 533160814Ssimon 534160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 535160814Ssimon in such a way that the self-signed certificate becomes part of the 536160814Ssimon CA database and uses the same mechanisms for serial number generation 537160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 538160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 539160814Ssimon [Richard Levitte] 540160814Ssimon 541160814Ssimon *) Add functionality to check the public key of a certificate request 542160814Ssimon against a given private. This is useful to check that a certificate 543160814Ssimon request can be signed by that key (self-signing). 544160814Ssimon [Richard Levitte] 545160814Ssimon 546160814Ssimon *) Make it possible to have multiple active certificates with the same 547160814Ssimon subject in the CA index file. This is done only if the keyword 548160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 549160814Ssimon if 'CA_default') of the configuration file. The value is saved 550160814Ssimon with the database itself in a separate index attribute file, 551160814Ssimon named like the index file with '.attr' appended to the name. 552160814Ssimon [Richard Levitte] 553160814Ssimon 554160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 555160814Ssimon req and dirName. 556160814Ssimon [Steve Henson] 557160814Ssimon 558160814Ssimon *) Support for nameConstraints certificate extension. 559160814Ssimon [Steve Henson] 560160814Ssimon 561160814Ssimon *) Support for policyConstraints certificate extension. 562160814Ssimon [Steve Henson] 563160814Ssimon 564160814Ssimon *) Support for policyMappings certificate extension. 565160814Ssimon [Steve Henson] 566160814Ssimon 567160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 568160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 569160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 570160814Ssimon indirection. This lets alternative implementations fallback to the 571160814Ssimon default implementation more easily. 572160814Ssimon [Geoff Thorpe] 573160814Ssimon 574160814Ssimon *) Support for directoryName in GeneralName related extensions 575160814Ssimon in config files. 576160814Ssimon [Steve Henson] 577160814Ssimon 578160814Ssimon *) Make it possible to link applications using Makefile.shared. 579160814Ssimon Make that possible even when linking against static libraries! 580160814Ssimon [Richard Levitte] 581160814Ssimon 582160814Ssimon *) Support for single pass processing for S/MIME signing. This now 583160814Ssimon means that S/MIME signing can be done from a pipe, in addition 584160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 585160814Ssimon and the signed data does not need to be all held in memory. 586160814Ssimon 587160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 588160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 589160814Ssimon is done after the data is output (and digests calculated) in 590160814Ssimon SMIME_write_PKCS7(). 591160814Ssimon [Steve Henson] 592160814Ssimon 593160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 594160814Ssimon applications, at least on the platforms where it's known how 595160814Ssimon to do it. 596160814Ssimon [Richard Levitte] 597160814Ssimon 598160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 599160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 600160814Ssimon will now compute a table of multiples of the generator that 601160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 602160814Ssimon faster (notably in the case of a single point multiplication, 603160814Ssimon scalar * generator). 604160814Ssimon [Nils Larsch, Bodo Moeller] 605160814Ssimon 606160814Ssimon *) IPv6 support for certificate extensions. The various extensions 607160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 608160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 609160814Ssimon correctly. 610160814Ssimon [Steve Henson] 611160814Ssimon 612160814Ssimon *) Added an ENGINE that implements RSA by performing private key 613160814Ssimon exponentiations with the GMP library. The conversions to and from 614160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 615160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 616160814Ssimon However there are likely to be other architectures where GMP could 617160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 618160814Ssimon specified at Configure time and should be accompanied by the necessary 619160814Ssimon linker additions, eg; 620160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 621160814Ssimon [Geoff Thorpe] 622160814Ssimon 623160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 624160814Ssimon testing availability of engines with "-t" - the old behaviour is 625160814Ssimon produced by increasing the feature's verbosity with "-tt". 626160814Ssimon [Geoff Thorpe] 627160814Ssimon 628160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 629160814Ssimon could be freed. Solution: make sure initialization is performed early 630160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 631160814Ssimon via PR#459) 632160814Ssimon [Lutz Jaenicke] 633160814Ssimon 634160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 635160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 636160814Ssimon software implementations. For DSA and DH, parameter generation can 637160814Ssimon also be overriden by providing the appropriate method callbacks. 638160814Ssimon [Geoff Thorpe] 639160814Ssimon 640160814Ssimon *) Change the "progress" mechanism used in key-generation and 641160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 642160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 643160814Ssimon postfixes and the older functions are reimplemented as wrappers for 644160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 645160814Ssimon declarations of the old functions to help (graceful) attempts to 646160814Ssimon migrate to the new functions. Also, the new key-generation API 647160814Ssimon functions operate on a caller-supplied key-structure and return 648160814Ssimon success/failure rather than returning a key or NULL - this is to 649160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 650160814Ssimon 651160814Ssimon Example for using the new callback interface: 652160814Ssimon 653160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 654160814Ssimon void *my_arg = ...; 655160814Ssimon BN_GENCB my_cb; 656160814Ssimon 657160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 658160814Ssimon 659160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 660160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 661160814Ssimon * documentation of the function that calls the callback. 662160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 663160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 664160814Ssimon * to continue, or 0 to stop. 665160814Ssimon */ 666160814Ssimon 667160814Ssimon [Geoff Thorpe] 668160814Ssimon 669160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 670160814Ssimon available to TLS with the number defined in 671160814Ssimon draft-ietf-tls-compression-04.txt. 672160814Ssimon [Richard Levitte] 673160814Ssimon 674160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 675160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 676160814Ssimon 677160814Ssimon CertificatePair ::= SEQUENCE { 678160814Ssimon forward [0] Certificate OPTIONAL, 679160814Ssimon reverse [1] Certificate OPTIONAL, 680160814Ssimon -- at least one of the pair shall be present -- } 681160814Ssimon 682160814Ssimon Also implement the PEM functions to read and write certificate 683160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 684160814Ssimon 685160814Ssimon This needed to be defined, mostly for the sake of the LDAP 686160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 687160814Ssimon well. 688160814Ssimon [Richard Levitte] 689160814Ssimon 690160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 691160814Ssimon Makefile.shared, for Cygwin's sake. 692160814Ssimon [Richard Levitte] 693160814Ssimon 694160814Ssimon *) Extend the BIGNUM API by creating a function 695160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 696160814Ssimon and a macro that behave like 697160814Ssimon int BN_is_negative(const BIGNUM *a); 698160814Ssimon 699160814Ssimon to avoid the need to access 'a->neg' directly in applications. 700160814Ssimon [Nils Larsch] 701160814Ssimon 702160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 703160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 704160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 705160814Ssimon if applicable. 706160814Ssimon [Nils Larsch <nla@trustcenter.de>] 707160814Ssimon 708160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 709160814Ssimon [Bodo Moeller] 710160814Ssimon 711160814Ssimon *) Change the ENGINE framework to automatically load engines 712160814Ssimon dynamically from specific directories unless they could be 713160814Ssimon found to already be built in or loaded. Move all the 714160814Ssimon current engines except for the cryptodev one to a new 715160814Ssimon directory engines/. 716160814Ssimon The engines in engines/ are built as shared libraries if 717160814Ssimon the "shared" options was given to ./Configure or ./config. 718160814Ssimon Otherwise, they are inserted in libcrypto.a. 719160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 720160814Ssimon engines, but that can be overriden at configure time through 721160814Ssimon the usual use of --prefix and/or --openssldir, and at run 722160814Ssimon time with the environment variable OPENSSL_ENGINES. 723160814Ssimon [Geoff Thorpe and Richard Levitte] 724160814Ssimon 725160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 726160814Ssimon libraries. Addapt Makefile.org. 727160814Ssimon [Richard Levitte] 728160814Ssimon 729160814Ssimon *) Add version info to Win32 DLLs. 730160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 731160814Ssimon 732160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 733160814Ssimon can be added using this API to created arbitrary PKCS#12 734160814Ssimon files while avoiding the low level API. 735160814Ssimon 736160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 737160814Ssimon will then be omitted from the output file. The encryption 738160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 739160814Ssimon iteration count can be set to 0 to omit the mac. 740160814Ssimon 741160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 742160814Ssimon options work when creating a PKCS#12 file. New option -nomac 743160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 744160814Ssimon New code is modified to use the enhanced PKCS12_create() 745160814Ssimon instead of the low level API. 746160814Ssimon [Steve Henson] 747160814Ssimon 748160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 749160814Ssimon encoding. This can output sequences tags and octet strings in 750160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 751160814Ssimon encoding. This is experimental and needs additional code to 752160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 753160814Ssimon PKCS#7 code. 754160814Ssimon 755160814Ssimon Extend template encode functionality so that tagging is passed 756160814Ssimon down to the template encoder. 757160814Ssimon [Steve Henson] 758160814Ssimon 759160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 760160814Ssimon recognized instead of using RSA as a default. 761160814Ssimon [Bodo Moeller] 762160814Ssimon 763160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 764160814Ssimon As these are not official, they are not included in "ALL"; 765160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 766160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 767160814Ssimon 768160814Ssimon *) Add ECDH engine support. 769160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 770160814Ssimon 771160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 772160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 773160814Ssimon 774160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 775160814Ssimon without success (which indicates a broken PRNG). 776160814Ssimon [Bodo Moeller] 777160814Ssimon 778160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 779160814Ssimon is really the square of the return value. (Previously, 780160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 781160814Ssimon [Bodo Moeller] 782160814Ssimon 783160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 784160814Ssimon and WAP/WTLS; add OIDs that were still missing. 785160814Ssimon 786160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 787160814Ssimon (Sun Microsystems Laboratories)] 788160814Ssimon 789160814Ssimon *) Extend the EC library for elliptic curves over binary fields 790160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 791160814Ssimon New EC_METHOD: 792160814Ssimon 793160814Ssimon EC_GF2m_simple_method 794160814Ssimon 795160814Ssimon New API functions: 796160814Ssimon 797160814Ssimon EC_GROUP_new_curve_GF2m 798160814Ssimon EC_GROUP_set_curve_GF2m 799160814Ssimon EC_GROUP_get_curve_GF2m 800160814Ssimon EC_POINT_set_affine_coordinates_GF2m 801160814Ssimon EC_POINT_get_affine_coordinates_GF2m 802160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 803160814Ssimon 804160814Ssimon Point compression for binary fields is disabled by default for 805160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 806160814Ssimon enable it). 807160814Ssimon 808160814Ssimon As binary polynomials are represented as BIGNUMs, various members 809160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 810160814Ssimon between the implementations for prime fields and binary fields; 811160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 812160814Ssimon are essentially identical to their ..._GFp counterparts. 813160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 814160814Ssimon various internal method names.) 815160814Ssimon 816160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 817160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 818160814Ssimon 819160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 820160814Ssimon (Sun Microsystems Laboratories)] 821160814Ssimon 822160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 823160814Ssimon through methods ('mul', 'precompute_mult'). 824160814Ssimon 825160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 826160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 827160814Ssimon methods are undefined. 828160814Ssimon 829160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 830160814Ssimon (Sun Microsystems Laboratories)] 831160814Ssimon 832160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 833160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 834160814Ssimon length of the modulus. 835160814Ssimon 836160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 837160814Ssimon (Sun Microsystems Laboratories)] 838160814Ssimon 839160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 840160814Ssimon (These simply call ..._new and ..._copy). 841160814Ssimon 842160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 843160814Ssimon (Sun Microsystems Laboratories)] 844160814Ssimon 845160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 846160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 847160814Ssimon used) in the following functions [macros]: 848160814Ssimon 849160814Ssimon BN_GF2m_add 850160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 851160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 852160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 853160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 854160814Ssimon BN_GF2m_mod_inv 855160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 856160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 857160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 858160814Ssimon BN_GF2m_cmp [= BN_ucmp] 859160814Ssimon 860160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 861160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 862160814Ssimon 863160814Ssimon For some functions, an the irreducible polynomial defining a 864160814Ssimon field can be given as an 'unsigned int[]' with strictly 865160814Ssimon decreasing elements giving the indices of those bits that are set; 866160814Ssimon i.e., p[] represents the polynomial 867160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 868160814Ssimon where 869160814Ssimon p[0] > p[1] > ... > p[k] = 0. 870160814Ssimon This applies to the following functions: 871160814Ssimon 872160814Ssimon BN_GF2m_mod_arr 873160814Ssimon BN_GF2m_mod_mul_arr 874160814Ssimon BN_GF2m_mod_sqr_arr 875160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 876160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 877160814Ssimon BN_GF2m_mod_exp_arr 878160814Ssimon BN_GF2m_mod_sqrt_arr 879160814Ssimon BN_GF2m_mod_solve_quad_arr 880160814Ssimon BN_GF2m_poly2arr 881160814Ssimon BN_GF2m_arr2poly 882160814Ssimon 883160814Ssimon Conversion can be performed by the following functions: 884160814Ssimon 885160814Ssimon BN_GF2m_poly2arr 886160814Ssimon BN_GF2m_arr2poly 887160814Ssimon 888160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 889160814Ssimon 890160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 891160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 892160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 893160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 894160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 895160814Ssimon 896160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 897160814Ssimon (Sun Microsystems Laboratories)] 898160814Ssimon 899160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 900160814Ssimon functionality is disabled at compile-time. 901160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 902160814Ssimon 903160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 904160814Ssimon information is visible when viewing, e.g., a certificate: 905160814Ssimon 906160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 907160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 908160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 909160814Ssimon avoid the appearance of a printable string. 910160814Ssimon [Nils Larsch <nla@trustcenter.de>] 911160814Ssimon 912160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 913160814Ssimon functions 914160814Ssimon EC_GROUP_set_asn1_flag() 915160814Ssimon EC_GROUP_get_asn1_flag() 916160814Ssimon EC_GROUP_set_point_conversion_form() 917160814Ssimon EC_GROUP_get_point_conversion_form() 918160814Ssimon These control ASN1 encoding details: 919160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 920160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 921160814Ssimon - Points are encoded in uncompressed form by default; options for 922160814Ssimon asn1_for are as for point2oct, namely 923160814Ssimon POINT_CONVERSION_COMPRESSED 924160814Ssimon POINT_CONVERSION_UNCOMPRESSED 925160814Ssimon POINT_CONVERSION_HYBRID 926160814Ssimon 927160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 928160814Ssimon functions 929160814Ssimon EC_GROUP_set_seed() 930160814Ssimon EC_GROUP_get0_seed() 931160814Ssimon EC_GROUP_get_seed_len() 932160814Ssimon This is used only for ASN1 purposes (so far). 933160814Ssimon [Nils Larsch <nla@trustcenter.de>] 934160814Ssimon 935160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 936160814Ssimon of the appropriate field type OID. The new function 937160814Ssimon EC_METHOD_get_field_type() returns this value. 938160814Ssimon [Nils Larsch <nla@trustcenter.de>] 939160814Ssimon 940160814Ssimon *) Add functions 941160814Ssimon EC_POINT_point2bn() 942160814Ssimon EC_POINT_bn2point() 943160814Ssimon EC_POINT_point2hex() 944160814Ssimon EC_POINT_hex2point() 945160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 946160814Ssimon EC_POINT_oct2point(). 947160814Ssimon [Nils Larsch <nla@trustcenter.de>] 948160814Ssimon 949160814Ssimon *) Change internals of the EC library so that the functions 950160814Ssimon EC_GROUP_set_generator() 951160814Ssimon EC_GROUP_get_generator() 952160814Ssimon EC_GROUP_get_order() 953160814Ssimon EC_GROUP_get_cofactor() 954160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 955160814Ssimon to methods, which would lead to unnecessary code duplication when 956160814Ssimon adding different types of curves. 957160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 958160814Ssimon 959160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 960160814Ssimon arithmetic, and such that modified wNAFs are generated 961160814Ssimon (which avoid length expansion in many cases). 962160814Ssimon [Bodo Moeller] 963160814Ssimon 964160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 965160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 966160814Ssimon 967160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 968160814Ssimon on a EC_GROUP, its generator and order. This includes 969160814Ssimon EC_GROUP_check_discriminant(). 970160814Ssimon [Nils Larsch <nla@trustcenter.de>] 971160814Ssimon 972160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 973160814Ssimon 974160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 975160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 976160814Ssimon 977160814Ssimon ECDSA support is also included in various other files across the 978160814Ssimon library. Most notably, 979160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 980160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 981160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 982160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 983160814Ssimon them suitable for ECDSA where domain parameters must be 984160814Ssimon extracted before the specific public key; 985160814Ssimon - ECDSA engine support has been added. 986160814Ssimon [Nils Larsch <nla@trustcenter.de>] 987160814Ssimon 988160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 989160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 990160814Ssimon function 991160814Ssimon EC_GROUP_new_by_curve_name(), 992160814Ssimon and the list of available named curves can be obtained with 993160814Ssimon EC_get_builtin_curves(). 994160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 995160814Ssimon accessed via 996160814Ssimon EC_GROUP_set_curve_name() 997160814Ssimon EC_GROUP_get_curve_name() 998160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 999160814Ssimon 1000160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1001160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 1002160814Ssimon required a small change in bn_mul_part_recursive() and the addition 1003160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 1004160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 1005160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 1006160814Ssimon differing sizes. 1007160814Ssimon [Richard Levitte] 1008160814Ssimon 1009162911Ssimon Changes between 0.9.7k and 0.9.7l [xx XXX xxxx] 1010162911Ssimon 1011162911Ssimon *) Change ciphersuite string processing so that an explicit 1012162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 1013162911Ssimon will no longer include "AES128-SHA"), and any other similar 1014162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 1015162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 1016162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 1017162911Ssimon changes from 0.9.8b and 0.9.8d. 1018162911Ssimon [Bodo Moeller] 1019162911Ssimon 1020162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 1021162911Ssimon 1022162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 1023162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 1024162911Ssimon 1025162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 1026162911Ssimon possible instead of select(), since the latter has some 1027162911Ssimon undesirable limitations. 1028162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 1029162911Ssimon 1030162911Ssimon *) Disable rogue ciphersuites: 1031162911Ssimon 1032162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 1033162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 1034162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 1035162911Ssimon 1036162911Ssimon The latter two were purportedly from 1037162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 1038162911Ssimon appear there. 1039162911Ssimon 1040162911Ssimon Also deactive the remaining ciphersuites from 1041162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 1042162911Ssimon unofficial, and the ID has long expired. 1043162911Ssimon [Bodo Moeller] 1044162911Ssimon 1045162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 1046162911Ssimon dual-core machines) and other potential thread-safety issues. 1047162911Ssimon [Bodo Moeller] 1048162911Ssimon 1049162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 1050162911Ssimon 1051162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 1052162911Ssimon module in FIPS mode. 1053162911Ssimon [Steve Henson] 1054162911Ssimon 1055162911Ssimon *) Fixes for VC++ 2005 build under Windows. 1056162911Ssimon [Steve Henson] 1057162911Ssimon 1058162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 1059162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 1060162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 1061162911Ssimon build to use fipscanister.o from the GNU make build. 1062162911Ssimon [Steve Henson] 1063162911Ssimon 1064160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 1065160814Ssimon 1066160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 1067160814Ssimon The value now differs depending on if you build for FIPS or not. 1068160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 1069160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 1070160814Ssimon the difference induced by this change. 1071160814Ssimon [Andy Polyakov] 1072160814Ssimon 1073160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 1074160814Ssimon 1075160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 1076160814Ssimon (part of SSL_OP_ALL). This option used to disable the 1077160814Ssimon countermeasure against man-in-the-middle protocol-version 1078160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 1079160814Ssimon idea. (CVE-2005-2969) 1080160814Ssimon 1081160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 1082160814Ssimon for Information Security, National Institute of Advanced Industrial 1083160814Ssimon Science and Technology [AIST], Japan)] 1084160814Ssimon 1085160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 1086160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 1087160814Ssimon [Steve Henson] 1088160814Ssimon 1089160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 1090160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 1091160814Ssimon the information leaked through timing could expose the secret key 1092160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 1093160814Ssimon biased k.) 1094160814Ssimon [Bodo Moeller] 1095160814Ssimon 1096160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 1097160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 1098160814Ssimon squares and multiplies and the memory access pattern are 1099160814Ssimon independent of the particular secret key. This will mitigate 1100160814Ssimon cache-timing and potential related attacks. 1101160814Ssimon 1102160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 1103160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 1104160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 1105160814Ssimon will use this BN flag for private exponents unless the flag 1106160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 1107160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 1108160814Ssimon 1109160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 1110160814Ssimon 1111160814Ssimon *) Change the client implementation for SSLv23_method() and 1112160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 1113160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 1114160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 1115160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 1116160814Ssimon [Bodo Moeller] 1117160814Ssimon 1118160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 1119160814Ssimon clients need. 1120160814Ssimon [Steve Henson] 1121160814Ssimon 1122160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 1123160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 1124160814Ssimon to dsa and dh code (which had race conditions before). 1125160814Ssimon [Steve Henson] 1126160814Ssimon 1127160814Ssimon *) Include the fixed error library code in the C error file definitions 1128160814Ssimon instead of fixing them up at runtime. This keeps the error code 1129160814Ssimon structures constant. 1130160814Ssimon [Steve Henson] 1131160814Ssimon 1132160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 1133160814Ssimon 1134160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 1135160814Ssimon OpenSSL 0.9.8.] 1136160814Ssimon 1137160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 1138160814Ssimon the 'length' field is signed on one version and unsigned on another 1139160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 1140160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 1141160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 1142160814Ssimon some needed definitions. 1143160814Ssimon [Steve Henson] 1144160814Ssimon 1145160814Ssimon *) Undo Cygwin change. 1146160814Ssimon [Ulf M�ller] 1147160814Ssimon 1148160814Ssimon *) Added support for proxy certificates according to RFC 3820. 1149160814Ssimon Because they may be a security thread to unaware applications, 1150160814Ssimon they must be explicitely allowed in run-time. See 1151160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 1152160814Ssimon [Richard Levitte] 1153160814Ssimon 1154160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 1155160814Ssimon 1156160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 1157160814Ssimon server and client random values. Previously 1158160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 1159160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 1160160814Ssimon 1161160814Ssimon This change has negligible security impact because: 1162160814Ssimon 1163160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 1164160814Ssimon data. 1165160814Ssimon 1166160814Ssimon 2. Server and client random values are sent in the clear in the initial 1167160814Ssimon handshake. 1168160814Ssimon 1169160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 1170160814Ssimon size for static RSA ciphersuites) as well as client server and random 1171160814Ssimon values. 1172160814Ssimon 1173160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 1174160814Ssimon to our attention. 1175160814Ssimon 1176160814Ssimon [Stephen Henson, reported by UK NISCC] 1177160814Ssimon 1178160814Ssimon *) Use Windows randomness collection on Cygwin. 1179160814Ssimon [Ulf M�ller] 1180160814Ssimon 1181160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 1182160814Ssimon prematurely by EGD/PRNGD. 1183160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 1184160814Ssimon 1185160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 1186160814Ssimon [Steve Henson] 1187160814Ssimon 1188160814Ssimon *) Back-port of selected performance improvements from development 1189160814Ssimon branch, as well as improved support for PowerPC platforms. 1190160814Ssimon [Andy Polyakov] 1191160814Ssimon 1192160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 1193160814Ssimon failure and freeing up memory if a failure occurs. 1194160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 1195160814Ssimon 1196160814Ssimon *) Add new -passin argument to dgst. 1197160814Ssimon [Steve Henson] 1198160814Ssimon 1199160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 1200160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 1201160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 1202160814Ssimon certificates. 1203160814Ssimon [Steve Henson] 1204160814Ssimon 1205160814Ssimon *) Make an explicit check during certificate validation to see that 1206160814Ssimon the CA setting in each certificate on the chain is correct. As a 1207160814Ssimon side effect always do the following basic checks on extensions, 1208160814Ssimon not just when there's an associated purpose to the check: 1209160814Ssimon 1210160814Ssimon - if there is an unhandled critical extension (unless the user 1211160814Ssimon has chosen to ignore this fault) 1212160814Ssimon - if the path length has been exceeded (if one is set at all) 1213160814Ssimon - that certain extensions fit the associated purpose (if one has 1214160814Ssimon been given) 1215160814Ssimon [Richard Levitte] 1216160814Ssimon 1217142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 1218142425Snectar 1219142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 1220142425Snectar environment. This would happen due to the reordering of the revoked 1221142425Snectar entries during signature checking and serial number lookup. Now the 1222142425Snectar encoding is cached and the serial number sort performed under a lock. 1223142425Snectar Add new STACK function sk_is_sorted(). 1224142425Snectar [Steve Henson] 1225142425Snectar 1226142425Snectar *) Add Delta CRL to the extension code. 1227142425Snectar [Steve Henson] 1228142425Snectar 1229142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 1230142425Snectar [David Holmes <d.holmes@f5.com>] 1231142425Snectar 1232142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 1233142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 1234142425Snectar This is done by creating a random 64 bit value for the initial serial 1235142425Snectar number when a serial number file is created or when a self signed 1236142425Snectar certificate is created using 'openssl req -x509'. The initial serial 1237142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 1238142425Snectar rather than being initialized to 1. 1239142425Snectar [Steve Henson] 1240142425Snectar 1241127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 1242127128Snectar 1243127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 1244160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 1245127128Snectar [Joe Orton, Steve Henson] 1246127128Snectar 1247127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 1248160814Ssimon (CVE-2004-0112) 1249127128Snectar [Joe Orton, Steve Henson] 1250127128Snectar 1251127128Snectar *) Make it possible to have multiple active certificates with the same 1252127128Snectar subject in the CA index file. This is done only if the keyword 1253127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 1254127128Snectar if 'CA_default') of the configuration file. The value is saved 1255127128Snectar with the database itself in a separate index attribute file, 1256127128Snectar named like the index file with '.attr' appended to the name. 1257127128Snectar [Richard Levitte] 1258127128Snectar 1259127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 1260127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 1261127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 1262127128Snectar extensions: since verify currently doesn't process CRL extensions this 1263127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 1264127128Snectar for these cases. 1265127128Snectar [Steve Henson] 1266127128Snectar 1267127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 1268127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 1269127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 1270127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 1271127128Snectar parsing them this should not create any compatibility issues. 1272127128Snectar [Steve Henson] 1273127128Snectar 1274127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 1275127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 1276127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 1277127128Snectar < 0.9.7. 1278127128Snectar [Steve Henson] 1279127128Snectar 1280127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 1281127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 1282127128Snectar 1283127128Snectar *) Use the correct content when signing type "other". 1284127128Snectar [Steve Henson] 1285127128Snectar 1286120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 1287120631Snectar 1288120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 1289120631Snectar 1290120631Snectar Stop out of bounds reads in the ASN1 code when presented with 1291160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 1292120631Snectar 1293160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 1294120631Snectar 1295120631Snectar If verify callback ignores invalid public key errors don't try to check 1296120631Snectar certificate signature with the NULL public key. 1297120631Snectar 1298120631Snectar [Steve Henson] 1299120631Snectar 1300120631Snectar *) New -ignore_err option in ocsp application to stop the server 1301120631Snectar exiting on the first error in a request. 1302120631Snectar [Steve Henson] 1303120631Snectar 1304120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 1305120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 1306120631Snectar specifications. 1307120631Snectar [Steve Henson] 1308120631Snectar 1309120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 1310120631Snectar extra data after the compression methods not only for TLS 1.0 1311120631Snectar but also for SSL 3.0 (as required by the specification). 1312120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 1313120631Snectar 1314120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 1315120631Snectar when it's 512 *bits* long, not 512 bytes. 1316120631Snectar [Richard Levitte] 1317120631Snectar 1318120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 1319120631Snectar blocks during encryption. 1320120631Snectar [Richard Levitte] 1321120631Snectar 1322120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 1323120631Snectar flushes were not handled properly if the BIO retried. On read 1324120631Snectar data was not being buffered properly and had various logic bugs. 1325120631Snectar This also affects blocking I/O when the data being decoded is a 1326120631Snectar certain size. 1327120631Snectar [Steve Henson] 1328120631Snectar 1329120631Snectar *) Various S/MIME bugfixes and compatibility changes: 1330120631Snectar output correct application/pkcs7 MIME type if 1331120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 1332120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 1333120631Snectar of files as .eml work). Correctly handle very long lines in MIME 1334120631Snectar parser. 1335120631Snectar [Steve Henson] 1336120631Snectar 1337120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 1338120631Snectar 1339120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 1340120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 1341120631Snectar a protocol version number mismatch like a decryption error 1342120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 1343120631Snectar [Bodo Moeller] 1344120631Snectar 1345120631Snectar *) Turn on RSA blinding by default in the default implementation 1346120631Snectar to avoid a timing attack. Applications that don't want it can call 1347120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 1348120631Snectar They would be ill-advised to do so in most cases. 1349120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 1350120631Snectar 1351120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 1352120631Snectar seeded (in this case, the secret RSA exponent is abused as 1353120631Snectar an unpredictable seed -- if it is not unpredictable, there 1354120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 1355120631Snectar by remembering the creator's thread ID in rsa->blinding and 1356120631Snectar having all other threads use local one-time blinding factors 1357120631Snectar (this requires more computation than sharing rsa->blinding, but 1358120631Snectar avoids excessive locking; and if an RSA object is not shared 1359120631Snectar between threads, blinding will still be very fast). 1360120631Snectar [Bodo Moeller] 1361120631Snectar 1362120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 1363120631Snectar ENGINE as defaults for all supported algorithms irrespective of 1364120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 1365120631Snectar should make sure they are passing it correctly. 1366120631Snectar [Geoff Thorpe] 1367120631Snectar 1368120631Snectar *) Target "mingw" now allows native Windows code to be generated in 1369120631Snectar the Cygwin environment as well as with the MinGW compiler. 1370120631Snectar [Ulf Moeller] 1371120631Snectar 1372111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 1373111147Snectar 1374111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 1375111147Snectar via timing by performing a MAC computation even if incorrrect 1376111147Snectar block cipher padding has been found. This is a countermeasure 1377111147Snectar against active attacks where the attacker has to distinguish 1378160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 1379111147Snectar 1380111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 1381111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 1382111147Snectar Martin Vuagnoux (EPFL, Ilion)] 1383111147Snectar 1384111147Snectar *) Make the no-err option work as intended. The intention with no-err 1385111147Snectar is not to have the whole error stack handling routines removed from 1386111147Snectar libcrypto, it's only intended to remove all the function name and 1387111147Snectar reason texts, thereby removing some of the footprint that may not 1388111147Snectar be interesting if those errors aren't displayed anyway. 1389111147Snectar 1390111147Snectar NOTE: it's still possible for any application or module to have it's 1391111147Snectar own set of error texts inserted. The routines are there, just not 1392111147Snectar used by default when no-err is given. 1393111147Snectar [Richard Levitte] 1394111147Snectar 1395111147Snectar *) Add support for FreeBSD on IA64. 1396111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 1397111147Snectar 1398111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 1399111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 1400111147Snectar the value returned by DES_cbc_cksum() was like the one from 1401111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 1402111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 1403111147Snectar 1404111147Snectar *) Allow an application to disable the automatic SSL chain building. 1405111147Snectar Before this a rather primitive chain build was always performed in 1406111147Snectar ssl3_output_cert_chain(): an application had no way to send the 1407111147Snectar correct chain if the automatic operation produced an incorrect result. 1408111147Snectar 1409111147Snectar Now the chain builder is disabled if either: 1410111147Snectar 1411111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 1412111147Snectar 1413111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 1414111147Snectar 1415111147Snectar The reasoning behind this is that an application would not want the 1416111147Snectar auto chain building to take place if extra chain certificates are 1417111147Snectar present and it might also want a means of sending no additional 1418111147Snectar certificates (for example the chain has two certificates and the 1419111147Snectar root is omitted). 1420111147Snectar [Steve Henson] 1421111147Snectar 1422111147Snectar *) Add the possibility to build without the ENGINE framework. 1423111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 1424111147Snectar 1425111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 1426111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 1427111147Snectar [Steve Henson] 1428111147Snectar 1429111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 1430111147Snectar could be freed. Solution: make sure initialization is performed early 1431111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 1432111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 1433111147Snectar [Lutz Jaenicke] 1434111147Snectar 1435111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 1436111147Snectar checked on reconnect on the client side, therefore session resumption 1437111147Snectar could still fail with a "ssl session id is different" error. This 1438111147Snectar behaviour is masked when SSL_OP_ALL is used due to 1439111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 1440111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 1441111147Snectar followup to PR #377. 1442111147Snectar [Lutz Jaenicke] 1443111147Snectar 1444111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 1445111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 1446111147Snectar [Andy Polyakov] 1447111147Snectar 1448111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 1449111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 1450111147Snectar the config script, much like the NetBSD support. 1451111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 1452111147Snectar 1453109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 1454109998Smarkm 1455120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 1456120631Snectar OpenSSL 0.9.7.] 1457120631Snectar 1458109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 1459109998Smarkm code (06) was taken as the first octet of the session ID and the last 1460109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 1461109998Smarkm caching could not have worked due to the session ID mismatch between 1462109998Smarkm client and server. 1463109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 1464109998Smarkm PR #377. 1465109998Smarkm [Lutz Jaenicke] 1466109998Smarkm 1467109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 1468109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 1469109998Smarkm removed entirely. 1470109998Smarkm [Richard Levitte] 1471109998Smarkm 1472109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 1473109998Smarkm seems that in spite of existing for more than a year, many application 1474109998Smarkm author have done nothing to provide the necessary callbacks, which 1475109998Smarkm means that this particular engine will not work properly anywhere. 1476109998Smarkm This is a very unfortunate situation which forces us, in the name 1477109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 1478109998Smarkm of libcrypto. 1479109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 1480109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 1481109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 1482109998Smarkm make such changes in the libcrypto locking code that changes will 1483109998Smarkm have to be made anyway). 1484109998Smarkm [Richard Levitte] 1485109998Smarkm 1486109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 1487109998Smarkm octets have been read, EOF or an error occurs. Without this change 1488109998Smarkm some truncated ASN1 structures will not produce an error. 1489109998Smarkm [Steve Henson] 1490109998Smarkm 1491109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 1492109998Smarkm Still give the possibility to force the use of Heimdal, but with 1493109998Smarkm warnings and a request that patches get sent to openssl-dev. 1494109998Smarkm [Richard Levitte] 1495109998Smarkm 1496109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 1497109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 1498109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 1499109998Smarkm 1500109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 1501109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 1502109998Smarkm edit numbers of the version. 1503109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 1504109998Smarkm 1505109998Smarkm *) Introduce safe string copy and catenation functions 1506109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 1507109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 1508109998Smarkm 1509109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 1510109998Smarkm [Ben Laurie (CHATS)] 1511109998Smarkm 1512109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 1513109998Smarkm resizing buffers containing secrets, and use where appropriate. 1514109998Smarkm [Ben Laurie (CHATS)] 1515109998Smarkm 1516109998Smarkm *) Avoid using fixed size buffers for configuration file location. 1517109998Smarkm [Ben Laurie (CHATS)] 1518109998Smarkm 1519109998Smarkm *) Avoid filename truncation for various CA files. 1520109998Smarkm [Ben Laurie (CHATS)] 1521109998Smarkm 1522109998Smarkm *) Use sizeof in preference to magic numbers. 1523109998Smarkm [Ben Laurie (CHATS)] 1524109998Smarkm 1525109998Smarkm *) Avoid filename truncation in cert requests. 1526109998Smarkm [Ben Laurie (CHATS)] 1527109998Smarkm 1528109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 1529109998Smarkm overflows. 1530109998Smarkm [Ben Laurie (CHATS)] 1531109998Smarkm 1532109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 1533109998Smarkm potentially lead to a spoofing attack). 1534109998Smarkm [Ben Laurie (CHATS)] 1535109998Smarkm 1536109998Smarkm *) Fix various buffers to be large enough for hex/decimal 1537109998Smarkm representations in a platform independent manner. 1538109998Smarkm [Ben Laurie (CHATS)] 1539109998Smarkm 1540109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 1541109998Smarkm resizing buffers containing secrets, and use where appropriate. 1542109998Smarkm [Ben Laurie (CHATS)] 1543109998Smarkm 1544109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 1545109998Smarkm indents. 1546109998Smarkm [Ben Laurie (CHATS)] 1547109998Smarkm 1548109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 1549109998Smarkm [Ben Laurie (CHATS)] 1550109998Smarkm 1551109998Smarkm *) buffer_gets() could terminate with the buffer only half 1552109998Smarkm full. Fixed. 1553109998Smarkm [Ben Laurie (CHATS)] 1554109998Smarkm 1555109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 1556109998Smarkm overflowing internal buffers by having large block sizes, etc. 1557109998Smarkm [Ben Laurie (CHATS)] 1558109998Smarkm 1559109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 1560109998Smarkm unconditionally). 1561109998Smarkm [Ben Laurie (CHATS)] 1562109998Smarkm 1563109998Smarkm *) Eliminate unused copy of key in RC4. 1564109998Smarkm [Ben Laurie (CHATS)] 1565109998Smarkm 1566109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 1567109998Smarkm [Ben Laurie (CHATS)] 1568109998Smarkm 1569109998Smarkm *) Fix off-by-one error in EGD path. 1570109998Smarkm [Ben Laurie (CHATS)] 1571109998Smarkm 1572109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 1573109998Smarkm [Ben Laurie (CHATS)] 1574109998Smarkm 1575109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 1576109998Smarkm CBCParameter. 1577109998Smarkm [Ben Laurie (CHATS)] 1578109998Smarkm 1579109998Smarkm *) Eliminate unused and dangerous function knumber(). 1580109998Smarkm [Ben Laurie (CHATS)] 1581109998Smarkm 1582109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 1583109998Smarkm [Ben Laurie (CHATS)] 1584109998Smarkm 1585109998Smarkm *) Protect against overlong session ID context length in an encoded 1586109998Smarkm session object. Since these are local, this does not appear to be 1587109998Smarkm exploitable. 1588109998Smarkm [Ben Laurie (CHATS)] 1589109998Smarkm 1590109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 1591109998Smarkm the 0.9.6 release series: 1592109998Smarkm 1593109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 1594109998Smarkm supply an oversized master key in Kerberos-enabled versions. 1595160814Ssimon (CVE-2002-0657) 1596109998Smarkm [Ben Laurie (CHATS)] 1597109998Smarkm 1598109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 1599109998Smarkm [Richard Levitte] 1600109998Smarkm 1601109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 1602109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 1603109998Smarkm 1604109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 1605109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 1606109998Smarkm 1607109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 1608109998Smarkm have been removed entirely. This was also the last step to make 1609109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 1610109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 1611109998Smarkm 1612109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 1613109998Smarkm to allow version independent disabling of normally unselected ciphers, 1614109998Smarkm which may be activated as a side-effect of selecting a single cipher. 1615109998Smarkm 1616109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 1617109998Smarkm out of "ALL" because they do not provide symmetric encryption. 1618109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 1619109998Smarkm [Lutz Jaenicke, Bodo Moeller] 1620109998Smarkm 1621109998Smarkm *) Add appropriate support for separate platform-dependent build 1622109998Smarkm directories. The recommended way to make a platform-dependent 1623109998Smarkm build directory is the following (tested on Linux), maybe with 1624109998Smarkm some local tweaks: 1625109998Smarkm 1626109998Smarkm # Place yourself outside of the OpenSSL source tree. In 1627109998Smarkm # this example, the environment variable OPENSSL_SOURCE 1628109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 1629109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 1630109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 1631111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 1632109998Smarkm mkdir -p `dirname $F` 1633109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 1634109998Smarkm done 1635109998Smarkm 1636109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 1637109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 1638109998Smarkm it probably means the source directory is very clean. 1639109998Smarkm [Richard Levitte] 1640109998Smarkm 1641109998Smarkm *) Make sure any ENGINE control commands make local copies of string 1642109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 1643109998Smarkm the caller may have overwritten (or deallocated) the original string 1644109998Smarkm data when a later ENGINE operation tries to use the stored values. 1645109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 1646109998Smarkm 1647109998Smarkm *) Improve diagnostics in file reading and command-line digests. 1648109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 1649109998Smarkm 1650109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 1651109998Smarkm error in AES-CFB decryption. 1652109998Smarkm [Richard Levitte] 1653109998Smarkm 1654109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 1655109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 1656109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 1657109998Smarkm BIOs and some applications. This has the side effect that 1658109998Smarkm applications must explicitly clean up cipher contexts with 1659109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 1660109998Smarkm [Steve Henson] 1661109998Smarkm 1662109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 1663109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 1664109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 1665109998Smarkm [Steve Henson] 1666109998Smarkm 1667109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 1668109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 1669109998Smarkm [Lutz Jaenicke] 1670109998Smarkm 1671109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 1672109998Smarkm form for "surname", serialNumber has no short form. 1673109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 1674109998Smarkm therefore remove "mail" short name for "internet 7". 1675109998Smarkm The OID for unique identifiers in X509 certificates is 1676109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 1677109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 1678109998Smarkm [Lutz Jaenicke] 1679109998Smarkm 1680109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 1681109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 1682109998Smarkm after all ctrl commands have been executed on it. If init=1 the 1683109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 1684109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 1685109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 1686109998Smarkm [Steve Henson] 1687109998Smarkm 1688109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 1689109998Smarkm argument is actually passed to the callback: In the 1690109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 1691109998Smarkm declaration has been changed from 1692109998Smarkm int (*cb)() 1693109998Smarkm into 1694109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 1695109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 1696109998Smarkm i=s->ctx->app_verify_callback(&ctx) 1697109998Smarkm has been changed into 1698109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 1699109998Smarkm 1700109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 1701109998Smarkm a dummy argument can be added to their callback functions. 1702109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 1703109998Smarkm 1704109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 1705109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 1706109998Smarkm 1707109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 1708109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 1709109998Smarkm This allows older applications to transparently support certain 1710109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 1711109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 1712109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 1713109998Smarkm always load it have also been added. 1714109998Smarkm [Steve Henson] 1715109998Smarkm 1716109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 1717109998Smarkm Adjust NIDs and EVP layer. 1718109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 1719109998Smarkm 1720109998Smarkm *) Config modules support in openssl utility. 1721109998Smarkm 1722109998Smarkm Most commands now load modules from the config file, 1723109998Smarkm though in a few (such as version) this isn't done 1724109998Smarkm because it couldn't be used for anything. 1725109998Smarkm 1726109998Smarkm In the case of ca and req the config file used is 1727109998Smarkm the same as the utility itself: that is the -config 1728109998Smarkm command line option can be used to specify an 1729109998Smarkm alternative file. 1730109998Smarkm [Steve Henson] 1731109998Smarkm 1732109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 1733109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 1734109998Smarkm [Steve Henson] 1735109998Smarkm 1736109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 1737109998Smarkm config section name. Add a new flag to tolerate a missing config file 1738109998Smarkm and move code to CONF_modules_load_file(). 1739109998Smarkm [Steve Henson] 1740109998Smarkm 1741109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 1742109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 1743109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 1744109998Smarkm to work with the new engine framework. 1745109998Smarkm [AEP Inc. and Richard Levitte] 1746109998Smarkm 1747109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 1748109998Smarkm Technologies. (Use engine 'sureware') 1749109998Smarkm The support was copied from 0.9.6c [engine] and adapted 1750109998Smarkm to work with the new engine framework. 1751109998Smarkm [Richard Levitte] 1752109998Smarkm 1753109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 1754109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 1755109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 1756109998Smarkm 1757109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 1758109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 1759109998Smarkm 1760109998Smarkm *) Add the configuration target debug-linux-ppro. 1761109998Smarkm Make 'openssl rsa' use the general key loading routines 1762109998Smarkm implemented in apps.c, and make those routines able to 1763109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 1764109998Smarkm FORMAT_IISSGC. 1765109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 1766109998Smarkm 1767109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 1768109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 1769109998Smarkm 1770109998Smarkm *) Add -keyform to rsautl, and document -engine. 1771109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 1772109998Smarkm 1773109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 1774109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 1775109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 1776109998Smarkm [Ben Laurie] 1777109998Smarkm 1778109998Smarkm *) Add new functions 1779109998Smarkm ERR_peek_last_error 1780109998Smarkm ERR_peek_last_error_line 1781109998Smarkm ERR_peek_last_error_line_data. 1782109998Smarkm These are similar to 1783109998Smarkm ERR_peek_error 1784109998Smarkm ERR_peek_error_line 1785109998Smarkm ERR_peek_error_line_data, 1786109998Smarkm but report on the latest error recorded rather than the first one 1787109998Smarkm still in the error queue. 1788109998Smarkm [Ben Laurie, Bodo Moeller] 1789109998Smarkm 1790109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 1791109998Smarkm like: 1792109998Smarkm default_algorithms = ALL 1793109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 1794109998Smarkm [Steve Henson] 1795109998Smarkm 1796109998Smarkm *) Prelminary ENGINE config module. 1797109998Smarkm [Steve Henson] 1798109998Smarkm 1799109998Smarkm *) New experimental application configuration code. 1800109998Smarkm [Steve Henson] 1801109998Smarkm 1802109998Smarkm *) Change the AES code to follow the same name structure as all other 1803109998Smarkm symmetric ciphers, and behave the same way. Move everything to 1804109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 1805109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 1806109998Smarkm 1807109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 1808109998Smarkm [Ben Laurie and Theo de Raadt] 1809109998Smarkm 1810109998Smarkm *) Add option to output public keys in req command. 1811109998Smarkm [Massimiliano Pala madwolf@openca.org] 1812109998Smarkm 1813109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 1814109998Smarkm (up to about 10% better than before for P-192 and P-224). 1815109998Smarkm [Bodo Moeller] 1816109998Smarkm 1817109998Smarkm *) New functions/macros 1818109998Smarkm 1819109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 1820109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 1821109998Smarkm SSL_set_msg_callback(ssl, cb) 1822109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 1823109998Smarkm 1824109998Smarkm to request calling a callback function 1825109998Smarkm 1826109998Smarkm void cb(int write_p, int version, int content_type, 1827109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 1828109998Smarkm 1829109998Smarkm whenever a protocol message has been completely received 1830109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 1831109998Smarkm protocol version according to which the SSL library interprets 1832109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 1833109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 1834109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 1835109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 1836109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 1837109998Smarkm SSL object, and 'arg' is the application-defined value set by 1838109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 1839109998Smarkm 1840109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 1841109998Smarkm to enable a callback that displays all protocol messages. 1842109998Smarkm [Bodo Moeller] 1843109998Smarkm 1844109998Smarkm *) Change the shared library support so shared libraries are built as 1845109998Smarkm soon as the corresponding static library is finished, and thereby get 1846109998Smarkm openssl and the test programs linked against the shared library. 1847109998Smarkm This still only happens when the keyword "shard" has been given to 1848109998Smarkm the configuration scripts. 1849109998Smarkm 1850109998Smarkm NOTE: shared library support is still an experimental thing, and 1851109998Smarkm backward binary compatibility is still not guaranteed. 1852109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 1853109998Smarkm 1854109998Smarkm *) Add support for Subject Information Access extension. 1855109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 1856109998Smarkm 1857109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 1858109998Smarkm additional bytes when new memory had to be allocated, not just 1859109998Smarkm when reusing an existing buffer. 1860109998Smarkm [Bodo Moeller] 1861109998Smarkm 1862109998Smarkm *) New command line and configuration option 'utf8' for the req command. 1863109998Smarkm This allows field values to be specified as UTF8 strings. 1864109998Smarkm [Steve Henson] 1865109998Smarkm 1866109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 1867109998Smarkm runs for the former and machine-readable output for the latter. 1868109998Smarkm [Ben Laurie] 1869109998Smarkm 1870109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 1871109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 1872109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 1873109998Smarkm has the same effect. 1874109998Smarkm [Massimiliano Pala madwolf@openca.org] 1875109998Smarkm 1876109998Smarkm *) Change all functions with names starting with des_ to be starting 1877109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 1878109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 1879109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 1880109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 1881109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 1882109998Smarkm exception. 1883109998Smarkm 1884109998Smarkm Since we provide two compatibility mappings, the user needs to 1885109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 1886109998Smarkm compatibility is desired. The default (i.e., when that macro 1887109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 1888109998Smarkm 1889109998Smarkm There are also macros that enable and disable the support of old 1890109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 1891109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 1892109998Smarkm are defined, the default will apply: to support the old des routines. 1893109998Smarkm 1894109998Smarkm In either case, one must include openssl/des.h to get the correct 1895109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 1896109998Smarkm won't work. 1897109998Smarkm 1898109998Smarkm NOTE: This is a major break of an old API into a new one. Software 1899109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 1900109998Smarkm time in the future, des_old.h and the libdes compatibility functions 1901109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 1902109998Smarkm default), and then completely removed. 1903109998Smarkm [Richard Levitte] 1904109998Smarkm 1905109998Smarkm *) Test for certificates which contain unsupported critical extensions. 1906109998Smarkm If such a certificate is found during a verify operation it is 1907109998Smarkm rejected by default: this behaviour can be overridden by either 1908109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 1909109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 1910109998Smarkm X509_supported_extension() has also been added which returns 1 if a 1911109998Smarkm particular extension is supported. 1912109998Smarkm [Steve Henson] 1913109998Smarkm 1914109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 1915109998Smarkm to retain compatibility with existing code. 1916109998Smarkm [Steve Henson] 1917109998Smarkm 1918109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 1919109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 1920109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 1921109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 1922109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 1923109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 1924109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 1925109998Smarkm requires the destination to be valid. 1926109998Smarkm 1927109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 1928109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 1929109998Smarkm [Steve Henson] 1930109998Smarkm 1931109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 1932109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 1933109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 1934109998Smarkm [Bodo Moeller] 1935109998Smarkm 1936109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 1937109998Smarkm [Massimo Santin via Richard Levitte] 1938109998Smarkm 1939109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 1940109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 1941109998Smarkm (initialisation, etc) from functionality dealing with implementations 1942109998Smarkm of specific crypto iterfaces. This change also introduces integrated 1943109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 1944109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 1945109998Smarkm implementations of their own. This is detailed in crypto/engine/README 1946109998Smarkm as it couldn't be adequately described here. However, there are a few 1947109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 1948109998Smarkm were changed in the original introduction of ENGINE code have now 1949109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 1950109998Smarkm deal more passive and at run-time, operations deal directly with 1951109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 1952109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 1953109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 1954109998Smarkm they were not being used by the framework as there is no concept of a 1955109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 1956109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 1957109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 1958109998Smarkm the new code. 1959109998Smarkm [Geoff Thorpe] 1960109998Smarkm 1961109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 1962109998Smarkm [Steve Henson] 1963109998Smarkm 1964109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 1965109998Smarkm and make sure the automatically generated functions ERR_load_* 1966109998Smarkm become part of libeay.num as well. 1967109998Smarkm [Richard Levitte] 1968109998Smarkm 1969109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 1970109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 1971109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 1972109998Smarkm false once a handshake has been completed. 1973109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 1974109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 1975109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 1976109998Smarkm client has followed the request.) 1977109998Smarkm [Bodo Moeller] 1978109998Smarkm 1979109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 1980109998Smarkm By default, clients may request session resumption even during 1981109998Smarkm renegotiation (if session ID contexts permit); with this option, 1982109998Smarkm session resumption is possible only in the first handshake. 1983109998Smarkm 1984109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 1985109998Smarkm more bits available for options that should not be part of 1986109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 1987109998Smarkm [Bodo Moeller] 1988109998Smarkm 1989109998Smarkm *) Add some demos for certificate and certificate request creation. 1990109998Smarkm [Steve Henson] 1991109998Smarkm 1992109998Smarkm *) Make maximum certificate chain size accepted from the peer application 1993109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 1994109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 1995109998Smarkm [Lutz Jaenicke] 1996109998Smarkm 1997109998Smarkm *) Add support for shared libraries for Unixware-7 1998109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 1999109998Smarkm [Lutz Jaenicke] 2000109998Smarkm 2001109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 2002109998Smarkm be done prior to destruction. Use this to unload error strings from 2003109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 2004109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 2005109998Smarkm [Geoff Thorpe] 2006109998Smarkm 2007109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 2008109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 2009109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 2010109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 2011109998Smarkm Also, add stub code to each that makes building them as self-contained 2012109998Smarkm shared-libraries easier (see README.ENGINE). 2013109998Smarkm [Geoff Thorpe] 2014109998Smarkm 2015109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 2016109998Smarkm implementations into applications that are completely implemented in 2017109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 2018109998Smarkm commands that can be used to configure what shared-library to load and 2019109998Smarkm to control aspects of the way it is handled. Also, made an update to 2020109998Smarkm the README.ENGINE file that brings its information up-to-date and 2021109998Smarkm provides some information and instructions on the "dynamic" ENGINE 2022109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 2023109998Smarkm [Geoff Thorpe] 2024109998Smarkm 2025109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 2026109998Smarkm "ERR_unload_strings" function. 2027109998Smarkm [Geoff Thorpe] 2028109998Smarkm 2029109998Smarkm *) Add a copy() function to EVP_MD. 2030109998Smarkm [Ben Laurie] 2031109998Smarkm 2032109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 2033109998Smarkm md_data void pointer. 2034109998Smarkm [Ben Laurie] 2035109998Smarkm 2036109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 2037109998Smarkm that the digest can only process a single chunk of data 2038109998Smarkm (typically because it is provided by a piece of 2039109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 2040109998Smarkm is only going to provide a single chunk of data, and hence the 2041109998Smarkm framework needn't accumulate the data for oneshot drivers. 2042109998Smarkm [Ben Laurie] 2043109998Smarkm 2044109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 2045109998Smarkm functions. This change also alters the storage and management of global 2046109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 2047109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 2048109998Smarkm index counters. The API functions that use this state have been changed 2049109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 2050109998Smarkm and counter, and there is now an API function to dynamically create new 2051109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 2052109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 2053109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 2054109998Smarkm such data would previously have always leaked in application code and 2055109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 2056109998Smarkm to it. Application code that doesn't use this new function will still 2057109998Smarkm leak as before, but their memory debugging output will announce it now 2058109998Smarkm rather than letting it slide. 2059109998Smarkm 2060109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 2061109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 2062109998Smarkm has a return value to indicate success or failure. 2063109998Smarkm [Geoff Thorpe] 2064109998Smarkm 2065109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 2066109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 2067109998Smarkm implementation. This change also adds two functions to "get" and "set" 2068109998Smarkm the implementation prior to it being automatically set the first time 2069109998Smarkm any other ERR function takes place. Ie. an application can call "get", 2070109998Smarkm pass the return value to a module it has just loaded, and that module 2071109998Smarkm can call its own "set" function using that value. This means the 2072109998Smarkm module's "ERR" operations will use (and modify) the error state in the 2073109998Smarkm application and not in its own statically linked copy of OpenSSL code. 2074109998Smarkm [Geoff Thorpe] 2075109998Smarkm 2076109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 2077109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 2078109998Smarkm the operation, and provides a more encapsulated way for external code 2079109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 2080109998Smarkm to use these functions rather than manually incrementing the counts. 2081109998Smarkm 2082109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 2083109998Smarkm [Geoff Thorpe] 2084109998Smarkm 2085109998Smarkm *) Add EVP test program. 2086109998Smarkm [Ben Laurie] 2087109998Smarkm 2088109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 2089109998Smarkm [Ben Laurie] 2090109998Smarkm 2091109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 2092109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 2093109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 2094109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 2095109998Smarkm directly. Modify 'ca' application to use new functions. 2096109998Smarkm [Steve Henson] 2097109998Smarkm 2098109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 2099109998Smarkm bug workarounds. Rollback attack detection is a security feature. 2100109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 2101109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 2102109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 2103109998Smarkm for their choice and can explicitly enable this option. 2104109998Smarkm [Bodo Moeller, Lutz Jaenicke] 2105109998Smarkm 2106109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 2107109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 2108109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 2109109998Smarkm Usage example: 2110109998Smarkm 2111109998Smarkm EVP_MD_CTX md; 2112109998Smarkm 2113109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 2114109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 2115109998Smarkm EVP_DigestUpdate(&md, in, len); 2116109998Smarkm EVP_DigestFinal(&md, out, NULL); 2117109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 2118109998Smarkm 2119109998Smarkm [Ben Laurie] 2120109998Smarkm 2121109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 2122109998Smarkm correcting its structure. This means that calls to DES functions 2123109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 2124109998Smarkm plain des_key_schedule (which was actually always a pointer 2125109998Smarkm anyway): E.g., 2126109998Smarkm 2127109998Smarkm des_key_schedule ks; 2128109998Smarkm 2129109998Smarkm des_set_key_checked(..., &ks); 2130109998Smarkm des_ncbc_encrypt(..., &ks, ...); 2131109998Smarkm 2132109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 2133109998Smarkm [Ben Laurie] 2134109998Smarkm 2135109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 2136109998Smarkm PEM causes large amounts of unused functions to be linked in due to 2137109998Smarkm poor organisation. For example pem_all.c contains every PEM function 2138109998Smarkm which has a knock on effect of linking in large amounts of (unused) 2139109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 2140109998Smarkm functions prevents this. 2141109998Smarkm [Steve Henson] 2142109998Smarkm 2143109998Smarkm *) Cleanup of EVP macros. 2144109998Smarkm [Ben Laurie] 2145109998Smarkm 2146109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 2147109998Smarkm correct _ecb suffix. 2148109998Smarkm [Ben Laurie] 2149109998Smarkm 2150109998Smarkm *) Add initial OCSP responder support to ocsp application. The 2151109998Smarkm revocation information is handled using the text based index 2152109998Smarkm use by the ca application. The responder can either handle 2153109998Smarkm requests generated internally, supplied in files (for example 2154109998Smarkm via a CGI script) or using an internal minimal server. 2155109998Smarkm [Steve Henson] 2156109998Smarkm 2157109998Smarkm *) Add configuration choices to get zlib compression for TLS. 2158109998Smarkm [Richard Levitte] 2159109998Smarkm 2160109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 2161109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 2162109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 2163109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 2164109998Smarkm 2165109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 2166109998Smarkm and authenticator structs; see crypto/krb5/. 2167109998Smarkm 2168109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 2169109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 2170109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 2171109998Smarkm via Richard Levitte] 2172109998Smarkm 2173109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 2174109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 2175109998Smarkm values for each of the key sizes rather than having just 2176109998Smarkm parameters (and 'speed' generating keys each time). 2177109998Smarkm [Geoff Thorpe] 2178109998Smarkm 2179109998Smarkm *) Speed up EVP routines. 2180109998Smarkm Before: 2181109998Smarkmencrypt 2182109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 2183109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 2184109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 2185109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 2186109998Smarkmdecrypt 2187109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 2188109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 2189109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 2190109998Smarkm After: 2191109998Smarkmencrypt 2192109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 2193109998Smarkmdecrypt 2194109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 2195109998Smarkm [Ben Laurie] 2196109998Smarkm 2197109998Smarkm *) Added the OS2-EMX target. 2198109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 2199109998Smarkm 2200109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 2201109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 2202109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 2203109998Smarkm structure: this means that the old CONF compatible routines can be 2204109998Smarkm retained (in particular wrt extensions) without having to duplicate the 2205109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 2206109998Smarkm [Steve Henson] 2207109998Smarkm 2208109998Smarkm *) Enhance the general user interface with mechanisms for inner control 2209109998Smarkm and with possibilities to have yes/no kind of prompts. 2210109998Smarkm [Richard Levitte] 2211109998Smarkm 2212109998Smarkm *) Change all calls to low level digest routines in the library and 2213109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 2214109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 2215109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 2216109998Smarkm 2217109998Smarkm *) Add the possibility to control engines through control names but with 2218109998Smarkm arbitrary arguments instead of just a string. 2219109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 2220109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 2221109998Smarkm versions of OpenSSL [engine]. 2222109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 2223109998Smarkm callback. 2224109998Smarkm [Richard Levitte] 2225109998Smarkm 2226109998Smarkm *) Enhance the general user interface with mechanisms to better support 2227109998Smarkm dialog box interfaces, application-defined prompts, the possibility 2228109998Smarkm to use defaults (for example default passwords from somewhere else) 2229109998Smarkm and interrupts/cancellations. 2230109998Smarkm [Richard Levitte] 2231109998Smarkm 2232109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 2233109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 2234109998Smarkm [Steve Henson] 2235109998Smarkm 2236109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 2237109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 2238109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 2239109998Smarkm 2240109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 2241109998Smarkm callback (pem_password_cb) as all other routines that need this 2242109998Smarkm kind of callback. 2243109998Smarkm [Richard Levitte] 2244109998Smarkm 2245109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 2246109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 2247109998Smarkm than this minimum value is recommended. 2248109998Smarkm [Lutz Jaenicke] 2249109998Smarkm 2250109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 2251109998Smarkm that are easily reachable. 2252109998Smarkm [Richard Levitte] 2253109998Smarkm 2254109998Smarkm *) Windows apparently can't transparently handle global 2255109998Smarkm variables defined in DLLs. Initialisations such as: 2256109998Smarkm 2257109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 2258109998Smarkm 2259109998Smarkm wont compile. This is used by the any applications that need to 2260109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 2261109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 2262109998Smarkm needed for static libraries under Win32. 2263109998Smarkm [Steve Henson] 2264109998Smarkm 2265109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 2266109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 2267109998Smarkm purpose functions and tidy up setting in other SSL functions. 2268109998Smarkm [Steve Henson] 2269109998Smarkm 2270109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 2271109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 2272109998Smarkm initialised. This allows various defaults to be set in the 2273109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 2274109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 2275109998Smarkm internally such as S/MIME. 2276109998Smarkm 2277109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 2278109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 2279109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 2280109998Smarkm 2281109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 2282109998Smarkm applications. 2283109998Smarkm [Steve Henson] 2284109998Smarkm 2285109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 2286109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 2287109998Smarkm its validity and signature checked, then if the certificate is found 2288109998Smarkm in the CRL the verify fails with a revoked error. 2289109998Smarkm 2290109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 2291109998Smarkm 2292109998Smarkm Command line options added to 'verify' application to support this. 2293109998Smarkm 2294109998Smarkm This needs some additional work, such as being able to handle multiple 2295109998Smarkm CRLs with different times, extension based lookup (rather than just 2296109998Smarkm by subject name) and ultimately more complete V2 CRL extension 2297109998Smarkm handling. 2298109998Smarkm [Steve Henson] 2299109998Smarkm 2300109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 2301109998Smarkm to replace things like des_read_password and friends (backward 2302109998Smarkm compatibility functions using this new API are provided). 2303109998Smarkm The purpose is to remove prompting functions from the DES code 2304109998Smarkm section as well as provide for prompting through dialog boxes in 2305109998Smarkm a window system and the like. 2306109998Smarkm [Richard Levitte] 2307109998Smarkm 2308109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 2309109998Smarkm per-structure level rather than having to store it globally. 2310109998Smarkm [Geoff] 2311109998Smarkm 2312109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 2313109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 2314109998Smarkm This causes the "original" ENGINE structure to act like a template, 2315109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 2316109998Smarkm operational state can be localised to each ENGINE structure, despite the 2317109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 2318109998Smarkm this case have no functional references and the return value is the single 2319109998Smarkm structural reference. This matches the single structural reference returned 2320109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 2321109998Smarkm ENGINE structure. 2322109998Smarkm [Geoff] 2323109998Smarkm 2324109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 2325109998Smarkm needs to match any other type at all we need to manually clear the 2326109998Smarkm tag cache. 2327109998Smarkm [Steve Henson] 2328109998Smarkm 2329109998Smarkm *) Changes to the "openssl engine" utility to include; 2330109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 2331109998Smarkm about an ENGINE's available control commands. 2332109998Smarkm - executing control commands from command line arguments using the 2333109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 2334109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 2335109998Smarkm the individual commands are colon-separated, for example; 2336109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 2337109998Smarkm [Geoff] 2338109998Smarkm 2339109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 2340109998Smarkm declare their own commands (numbers), names (strings), descriptions, 2341109998Smarkm and input types for run-time discovery by calling applications. A 2342109998Smarkm subset of these commands are implicitly classed as "executable" 2343109998Smarkm depending on their input type, and only these can be invoked through 2344109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 2345109998Smarkm can be based on user input, config files, etc). The distinction is 2346109998Smarkm that "executable" commands cannot return anything other than a boolean 2347109998Smarkm result and can only support numeric or string input, whereas some 2348109998Smarkm discoverable commands may only be for direct use through 2349109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 2350109998Smarkm pointers, or other custom uses. The "executable" commands are to 2351109998Smarkm support parameterisations of ENGINE behaviour that can be 2352109998Smarkm unambiguously defined by ENGINEs and used consistently across any 2353109998Smarkm OpenSSL-based application. Commands have been added to all the 2354109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 2355109998Smarkm control over shared-library paths without source code alterations. 2356109998Smarkm [Geoff] 2357109998Smarkm 2358109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 2359109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 2360109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 2361109998Smarkm this also allows the implementations to compile without using the 2362109998Smarkm internal engine_int.h header. 2363109998Smarkm [Geoff] 2364109998Smarkm 2365109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 2366109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 2367109998Smarkm should already have non-const pointers to it (ie. they should only 2368109998Smarkm modify their own ones). 2369109998Smarkm [Geoff] 2370109998Smarkm 2371109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 2372109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 2373109998Smarkm to C code. "nuron" string definitions were placed in variables 2374109998Smarkm rather than hard-coded - allowing parameterisation of these values 2375109998Smarkm later on via ctrl() commands. 2376109998Smarkm - Removed unused "#if 0"'d code. 2377109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 2378109998Smarkm structural references. 2379109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 2380109998Smarkm - Constified various get/set functions as appropriate and added 2381109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 2382109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 2383109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 2384109998Smarkm or function to NULL is a way of cancelling out a previously set 2385109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 2386109998Smarkm and doesn't justify the extra error symbols and code. 2387109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 2388109998Smarkm flags from engine_int.h to engine.h. 2389109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 2390109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 2391109998Smarkm [Geoff] 2392109998Smarkm 2393109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 2394109998Smarkm to the algorithm using long division. The binary algorithm can be 2395109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 2396109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 2397109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 2398109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 2399109998Smarkm appears to be advantageous for much longer moduli; here we use it 2400109998Smarkm for moduli up to 2048 bits. 2401109998Smarkm [Bodo Moeller] 2402109998Smarkm 2403109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 2404109998Smarkm could not support the combine flag in choice fields. 2405109998Smarkm [Steve Henson] 2406109998Smarkm 2407109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 2408109998Smarkm extensions from a certificate request to the certificate. 2409109998Smarkm [Steve Henson] 2410109998Smarkm 2411109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 2412109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 2413109998Smarkm file: this allows the display of the certificate about to be 2414109998Smarkm signed to be customised, to allow certain fields to be included 2415109998Smarkm or excluded and extension details. The old system didn't display 2416109998Smarkm multicharacter strings properly, omitted fields not in the policy 2417109998Smarkm and couldn't display additional details such as extensions. 2418109998Smarkm [Steve Henson] 2419109998Smarkm 2420109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 2421109998Smarkm of an arbitrary number of elliptic curve points 2422109998Smarkm \sum scalars[i]*points[i], 2423109998Smarkm optionally including the generator defined for the EC_GROUP: 2424109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 2425109998Smarkm 2426109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 2427109998Smarkm that the point list has just one item (besides the optional 2428109998Smarkm generator). 2429109998Smarkm [Bodo Moeller] 2430109998Smarkm 2431109998Smarkm *) First EC_METHODs for curves over GF(p): 2432109998Smarkm 2433109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 2434109998Smarkm operations and provides various method functions that can also 2435109998Smarkm operate with faster implementations of modular arithmetic. 2436109998Smarkm 2437109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 2438109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 2439109998Smarkm 2440109998Smarkm [Bodo Moeller; point addition and point doubling 2441109998Smarkm implementation directly derived from source code provided by 2442109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 2443109998Smarkm 2444109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 2445109998Smarkm crypto/ec/ec_lib.c): 2446109998Smarkm 2447109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 2448109998Smarkm based on EC_METHODs that are built into the library. 2449109998Smarkm 2450109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 2451109998Smarkm 2452109998Smarkm Most of the framework would be able to handle curves over arbitrary 2453109998Smarkm finite fields, but as there are no obvious types for fields other 2454109998Smarkm than GF(p), some functions are limited to that for now. 2455109998Smarkm [Bodo Moeller] 2456109998Smarkm 2457109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 2458109998Smarkm that the file contains a complete HTTP response. 2459109998Smarkm [Richard Levitte] 2460109998Smarkm 2461109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 2462109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 2463109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 2464109998Smarkm field while the former will cause them to run together if the field 2465109998Smarkm is 40 of more characters long. 2466109998Smarkm [Steve Henson] 2467109998Smarkm 2468109998Smarkm *) Constify the cipher and digest 'method' functions and structures 2469109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 2470109998Smarkm pointers. 2471109998Smarkm [Steve Henson] 2472109998Smarkm 2473109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 2474109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 2475109998Smarkm [Bodo Moeller] 2476109998Smarkm 2477109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 2478109998Smarkm internal software routines can never fail additional hardware versions 2479109998Smarkm might. 2480109998Smarkm [Steve Henson] 2481109998Smarkm 2482109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 2483109998Smarkm 2484109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 2485109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 2486109998Smarkm 2487109998Smarkm ASN1 error codes 2488109998Smarkm ERR_R_NESTED_ASN1_ERROR 2489109998Smarkm ... 2490109998Smarkm ERR_R_MISSING_ASN1_EOS 2491109998Smarkm were 4 .. 9, conflicting with 2492109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 2493109998Smarkm ... 2494109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 2495109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 2496109998Smarkm 2497109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 2498109998Smarkm [Bodo Moeller] 2499109998Smarkm 2500109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 2501109998Smarkm suffices. 2502109998Smarkm [Bodo Moeller] 2503109998Smarkm 2504109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 2505109998Smarkm sets the subject name for a new request or supersedes the 2506109998Smarkm subject name in a given request. Formats that can be parsed are 2507109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 2508109998Smarkm and 2509109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 2510109998Smarkm 2511109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 2512109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 2513109998Smarkm 2514109998Smarkm *) Introduce the possibility to access global variables through 2515109998Smarkm functions on platform were that's the best way to handle exporting 2516109998Smarkm global variables in shared libraries. To enable this functionality, 2517109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 2518109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 2519109998Smarkm is normally done by Configure or something similar). 2520109998Smarkm 2521109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 2522109998Smarkm in the source file (foo.c) like this: 2523109998Smarkm 2524109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 2525109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 2526109998Smarkm 2527109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 2528109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 2529109998Smarkm 2530109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 2531109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 2532109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 2533109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 2534109998Smarkm 2535109998Smarkm The #defines are very important, and therefore so is including the 2536109998Smarkm header file everywhere where the defined globals are used. 2537109998Smarkm 2538109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 2539109998Smarkm of ASN.1 items, but that structure is a bit different. 2540109998Smarkm 2541109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 2542109998Smarkm better and easier to understand logic to choose which symbols should 2543109998Smarkm go into the Windows .def files as well as a number of fixes and code 2544109998Smarkm cleanup (among others, algorithm keywords are now sorted 2545109998Smarkm lexicographically to avoid constant rewrites). 2546109998Smarkm [Richard Levitte] 2547109998Smarkm 2548109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 2549109998Smarkm result to 'rm' because if rm==num the value will be overwritten 2550109998Smarkm and produce the wrong result if 'num' is negative: this caused 2551109998Smarkm problems with BN_mod() and BN_nnmod(). 2552109998Smarkm [Steve Henson] 2553109998Smarkm 2554109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 2555109998Smarkm OCSP request and verifies the signer certificate. The signer 2556109998Smarkm certificate is just checked for a generic purpose and OCSP request 2557109998Smarkm trust settings. 2558109998Smarkm [Steve Henson] 2559109998Smarkm 2560109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 2561109998Smarkm responses. OCSP responses are prepared in real time and may only 2562109998Smarkm be a few seconds old. Simply checking that the current time lies 2563109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 2564109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 2565109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 2566109998Smarkm the current time. The age of the response can also optionally be 2567109998Smarkm checked. Two new options -validity_period and -status_age added to 2568109998Smarkm ocsp utility. 2569109998Smarkm [Steve Henson] 2570109998Smarkm 2571109998Smarkm *) If signature or public key algorithm is unrecognized print out its 2572109998Smarkm OID rather that just UNKNOWN. 2573109998Smarkm [Steve Henson] 2574109998Smarkm 2575109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 2576109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 2577109998Smarkm ID to be generated from the issuer certificate alone which can then be 2578109998Smarkm passed to OCSP_id_issuer_cmp(). 2579109998Smarkm [Steve Henson] 2580109998Smarkm 2581109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 2582109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 2583109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 2584109998Smarkm new macros which allow the underlying ASN1 function/structure to 2585109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 2586109998Smarkm references directly (such as &X509_it) but instead use the relevant 2587109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 2588109998Smarkm use of the new ASN1 code on platforms where exporting structures 2589109998Smarkm is problematical (for example in shared libraries) but exporting 2590109998Smarkm functions returning pointers to structures is not. 2591109998Smarkm [Steve Henson] 2592109998Smarkm 2593109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 2594109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 2595109998Smarkm The purpose of this is to allow applications to control, if they wish, 2596109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 2597109998Smarkm can be useful for session caching in multiple-server environments. A 2598109998Smarkm command-line switch for testing this (and any client code that wishes 2599109998Smarkm to use such a feature) has been added to "s_server". 2600109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 2601109998Smarkm 2602109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 2603109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 2604109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 2605109998Smarkm the growing number of special cases it was previously handling. 2606109998Smarkm [Richard Levitte] 2607109998Smarkm 2608109998Smarkm *) Make all configuration macros available for application by making 2609109998Smarkm sure they are available in opensslconf.h, by giving them names starting 2610109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 2611109998Smarkm sure e_os2.h will cover all platform-specific cases together with 2612109998Smarkm opensslconf.h. 2613109998Smarkm Additionally, it is now possible to define configuration/platform- 2614109998Smarkm specific names (called "system identities"). In the C code, these 2615109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 2616109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 2617109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 2618109998Smarkm what is available. 2619109998Smarkm [Richard Levitte] 2620109998Smarkm 2621109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 2622109998Smarkm number to use to be specified on the command line. Previously self 2623109998Smarkm signed certificates were hard coded with serial number 0 and the 2624109998Smarkm CA options of 'x509' had to use a serial number in a file which was 2625109998Smarkm auto incremented. 2626109998Smarkm [Steve Henson] 2627109998Smarkm 2628109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 2629109998Smarkm Currently CRL reason, invalidity date and hold instruction are 2630109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 2631109998Smarkm [Steve Henson] 2632109998Smarkm 2633109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 2634109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 2635109998Smarkm API, which was previously mandatory. This means that the data is 2636109998Smarkm not padded in any way and so the total length much be a multiple 2637109998Smarkm of the block size, otherwise an error occurs. 2638109998Smarkm [Steve Henson] 2639109998Smarkm 2640109998Smarkm *) Initial (incomplete) OCSP SSL support. 2641109998Smarkm [Steve Henson] 2642109998Smarkm 2643109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 2644109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 2645109998Smarkm option to ocsp utility. 2646109998Smarkm [Steve Henson] 2647109998Smarkm 2648109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 2649109998Smarkm reflects the various checks performed. Applications can decide 2650109998Smarkm whether to tolerate certain situations such as an absent nonce 2651109998Smarkm in a response when one was present in a request: the ocsp application 2652109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 2653109998Smarkm this is to allow responders to include a nonce in a response even if 2654109998Smarkm the request is nonce-less. 2655109998Smarkm [Steve Henson] 2656109998Smarkm 2657109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 2658109998Smarkm skipped when using openssl x509 multiple times on a single input file, 2659109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 2660109998Smarkm [Bodo Moeller] 2661109998Smarkm 2662109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 2663109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 2664109998Smarkm utility to correctly initialize revocation date of CRLs. 2665109998Smarkm [Steve Henson] 2666109998Smarkm 2667109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 2668109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 2669109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 2670109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 2671109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 2672109998Smarkm [Lutz Jaenicke] 2673109998Smarkm 2674109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 2675109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 2676109998Smarkm appear to exist. 2677109998Smarkm [Steve Henson] 2678109998Smarkm 2679109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 2680109998Smarkm additional certificates supplied. 2681109998Smarkm [Steve Henson] 2682109998Smarkm 2683109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 2684109998Smarkm OCSP client a number of certificate to only verify the response 2685109998Smarkm signature against. 2686109998Smarkm [Richard Levitte] 2687109998Smarkm 2688109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 2689109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 2690109998Smarkm AES OIDs. 2691109998Smarkm 2692109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 2693109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 2694109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 2695109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 2696109998Smarkm alias because they were not yet official; they could be 2697109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 2698109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 2699109998Smarkm alias is called "AES" and is part of "ALL".) 2700109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 2701109998Smarkm 2702109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 2703109998Smarkm request to response. 2704109998Smarkm [Steve Henson] 2705109998Smarkm 2706109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 2707109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 2708109998Smarkm extract information from a certificate request. OCSP_response_create() 2709109998Smarkm creates a response and optionally adds a basic response structure. 2710109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 2711109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 2712109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 2713109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 2714109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 2715109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 2716109998Smarkm (converts ASN1_TIME to GeneralizedTime). 2717109998Smarkm [Steve Henson] 2718109998Smarkm 2719109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 2720109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 2721109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 2722109998Smarkm contents: this is used in various key identifiers. 2723109998Smarkm [Steve Henson] 2724109998Smarkm 2725109998Smarkm *) Make sk_sort() tolerate a NULL argument. 2726109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 2727109998Smarkm 2728109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 2729109998Smarkm passed by the function are trusted implicitly. If any of them signed the 2730109998Smarkm response then it is assumed to be valid and is not verified. 2731109998Smarkm [Steve Henson] 2732109998Smarkm 2733109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 2734109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 2735109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 2736109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 2737109998Smarkm <support@securenetterm.com>] 2738109998Smarkm 2739109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 2740109998Smarkm routines: without these tracing memory leaks is very painful. 2741109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 2742109998Smarkm [Steve Henson] 2743109998Smarkm 2744109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 2745109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 2746109998Smarkm effectively meant GeneralizedTime would never be used. Now it 2747109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 2748109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 2749109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 2750109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 2751109998Smarkm <support@securenetterm.com>] 2752109998Smarkm 2753109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 2754109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 2755109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 2756109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 2757109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 2758109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 2759109998Smarkm [Steve Henson] 2760109998Smarkm 2761109998Smarkm *) Add summary printout to ocsp utility. The various functions which 2762109998Smarkm convert status values to strings have been renamed to: 2763109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 2764109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 2765109998Smarkm to verify nonce values and to disable verification. OCSP response 2766109998Smarkm printout format cleaned up. 2767109998Smarkm [Steve Henson] 2768109998Smarkm 2769109998Smarkm *) Add additional OCSP certificate checks. These are those specified 2770109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 2771109998Smarkm certificate being checked must either be the OCSP signer certificate 2772109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 2773109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 2774109998Smarkm usage. This check is performed by attempting to match the OCSP 2775109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 2776109998Smarkm in the OCSP_CERTID structures of the response. 2777109998Smarkm [Steve Henson] 2778109998Smarkm 2779109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 2780109998Smarkm and related routines. This uses the standard OpenSSL certificate 2781109998Smarkm verify routines to perform initial checks (just CA validity) and 2782109998Smarkm to obtain the certificate chain. Then additional checks will be 2783109998Smarkm performed on the chain. Currently the root CA is checked to see 2784109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 2785109998Smarkm a root CA as a global signing root: that is any certificate that 2786109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 2787109998Smarkm [Steve Henson] 2788109998Smarkm 2789109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 2790109998Smarkm extensions from a separate configuration file. 2791109998Smarkm As when reading extensions from the main configuration file, 2792109998Smarkm the '-extensions ...' option may be used for specifying the 2793109998Smarkm section to use. 2794109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 2795109998Smarkm 2796109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 2797109998Smarkm read. The request can be sent to a responder and the output 2798109998Smarkm parsed, outputed or printed in text form. Not complete yet: 2799109998Smarkm still needs to check the OCSP response validity. 2800109998Smarkm [Steve Henson] 2801109998Smarkm 2802109998Smarkm *) New subcommands for 'openssl ca': 2803109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 2804109998Smarkm the given serial number (according to the index file). 2805109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 2806109998Smarkm in the index file. 2807109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 2808109998Smarkm 2809109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 2810109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 2811109998Smarkm so that the resulting key is not encrypted. 2812109998Smarkm [Damien Miller <djm@mindrot.org>] 2813109998Smarkm 2814109998Smarkm *) New configuration for the GNU Hurd. 2815109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 2816109998Smarkm 2817109998Smarkm *) Initial code to implement OCSP basic response verify. This 2818109998Smarkm is currently incomplete. Currently just finds the signer's 2819109998Smarkm certificate and verifies the signature on the response. 2820109998Smarkm [Steve Henson] 2821109998Smarkm 2822109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 2823109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 2824109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 2825109998Smarkm [Bodo Moeller] 2826109998Smarkm 2827109998Smarkm *) Allowing defining memory allocation callbacks that will be given 2828109998Smarkm file name and line number information in additional arguments 2829109998Smarkm (a const char* and an int). The basic functionality remains, as 2830109998Smarkm well as the original possibility to just replace malloc(), 2831109998Smarkm realloc() and free() by functions that do not know about these 2832109998Smarkm additional arguments. To register and find out the current 2833109998Smarkm settings for extended allocation functions, the following 2834109998Smarkm functions are provided: 2835109998Smarkm 2836109998Smarkm CRYPTO_set_mem_ex_functions 2837109998Smarkm CRYPTO_set_locked_mem_ex_functions 2838109998Smarkm CRYPTO_get_mem_ex_functions 2839109998Smarkm CRYPTO_get_locked_mem_ex_functions 2840109998Smarkm 2841109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 2842109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 2843109998Smarkm extended allocation function is enabled. 2844109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 2845109998Smarkm a conventional allocation function is enabled. 2846109998Smarkm [Richard Levitte, Bodo Moeller] 2847109998Smarkm 2848109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 2849109998Smarkm There should no longer be any prototype-casting required when using 2850109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 2851109998Smarkm the callback types and macros at the head of lhash.h for details 2852109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 2853109998Smarkm [Geoff Thorpe] 2854109998Smarkm 2855109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 2856109998Smarkm If /dev/[u]random devices are not available or do not return enough 2857109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 2858109998Smarkm be queried. 2859109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 2860109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 2861109998Smarkm when enough entropy was collected without querying more sockets. 2862109998Smarkm [Lutz Jaenicke] 2863109998Smarkm 2864109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 2865109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 2866109998Smarkm of data has been collected. We spend at most 10 ms on each file 2867109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 2868109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 2869109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 2870109998Smarkm platforms the 10 ms delay will never occur. 2871109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 2872109998Smarkm For VMS, there's a currently-empty rand_vms.c. 2873109998Smarkm [Richard Levitte] 2874109998Smarkm 2875109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 2876109998Smarkm provide utility functions which an application needing 2877109998Smarkm to issue a request to an OCSP responder and analyse the 2878109998Smarkm response will typically need: as opposed to those which an 2879109998Smarkm OCSP responder itself would need which will be added later. 2880109998Smarkm 2881109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 2882109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 2883109998Smarkm response. OCSP_response_get1_basic() extracts basic response 2884109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 2885109998Smarkm information from an OCSP_CERTID structure (which will be created 2886109998Smarkm when the request structure is built). These are built from lower 2887109998Smarkm level functions which work on OCSP_SINGLERESP structures but 2888109998Smarkm wont normally be used unless the application wishes to examine 2889109998Smarkm extensions in the OCSP response for example. 2890109998Smarkm 2891109998Smarkm Replace nonce routines with a pair of functions. 2892109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 2893109998Smarkm generates a random value. OCSP_check_nonce() checks the 2894109998Smarkm validity of the nonce in an OCSP response. 2895109998Smarkm [Steve Henson] 2896109998Smarkm 2897109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 2898109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 2899109998Smarkm need to free up the newly created id. Change return type 2900109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 2901109998Smarkm This can then be used to add extensions to the request. 2902109998Smarkm Deleted OCSP_request_new(), since most of its functionality 2903109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 2904109998Smarkm clash) apart from the ability to set the request name which 2905109998Smarkm will be added elsewhere. 2906109998Smarkm [Steve Henson] 2907109998Smarkm 2908109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 2909109998Smarkm various functions. Extensions are now handled using the new 2910109998Smarkm OCSP extension code. New simple OCSP HTTP function which 2911109998Smarkm can be used to send requests and parse the response. 2912109998Smarkm [Steve Henson] 2913109998Smarkm 2914109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 2915109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 2916109998Smarkm uses the special reorder version of SET OF to sort the attributes 2917109998Smarkm and reorder them to match the encoded order. This resolves a long 2918109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 2919109998Smarkm it used to fail because the attribute order did not match the 2920109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 2921109998Smarkm it uses the received order. This is necessary to tolerate some broken 2922109998Smarkm software that does not order SET OF. This is handled by encoding 2923109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 2924109998Smarkm to produce the required SET OF. 2925109998Smarkm [Steve Henson] 2926109998Smarkm 2927109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 2928109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 2929109998Smarkm files to get correct declarations of the ASN.1 item variables. 2930109998Smarkm [Richard Levitte] 2931109998Smarkm 2932109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 2933109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 2934109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 2935109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 2936109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 2937109998Smarkm ASN1_ITEM and no wrapper functions. 2938109998Smarkm [Steve Henson] 2939109998Smarkm 2940109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 2941109998Smarkm replace the old function pointer based I/O routines. Change most of 2942109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 2943109998Smarkm [Steve Henson] 2944109998Smarkm 2945109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 2946109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 2947109998Smarkm it complain about algorithm deselection that isn't recognised. 2948109998Smarkm [Richard Levitte] 2949109998Smarkm 2950109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 2951109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 2952109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 2953109998Smarkm some old style ASN1 functions: this can be used to determine if old 2954109998Smarkm code will still work when these eventually go away. 2955109998Smarkm [Steve Henson] 2956109998Smarkm 2957109998Smarkm *) New extension functions for OCSP structures, these follow the 2958109998Smarkm same conventions as certificates and CRLs. 2959109998Smarkm [Steve Henson] 2960109998Smarkm 2961109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 2962109998Smarkm adds an extension. Its behaviour can be customised with various 2963109998Smarkm flags to append, replace or delete. Various wrappers added for 2964109998Smarkm certifcates and CRLs. 2965109998Smarkm [Steve Henson] 2966109998Smarkm 2967109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 2968109998Smarkm an extension cannot be parsed. Correct a typo in the 2969109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 2970109998Smarkm [Steve Henson] 2971109998Smarkm 2972109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 2973109998Smarkm entries for variables. 2974109998Smarkm [Steve Henson] 2975109998Smarkm 2976109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 2977109998Smarkm problems: As the program is single-threaded, all we have 2978109998Smarkm to do is register a locking callback using an array for 2979109998Smarkm storing which locks are currently held by the program. 2980109998Smarkm [Bodo Moeller] 2981109998Smarkm 2982109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 2983109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 2984109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 2985109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 2986109998Smarkm Unfortunately, the ex_data design is not at all suited 2987109998Smarkm for multi-threaded use, so it probably should be abolished. 2988109998Smarkm [Bodo Moeller] 2989109998Smarkm 2990109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 2991109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 2992109998Smarkm 2993109998Smarkm *) Move common extension printing code to new function 2994109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 2995109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 2996109998Smarkm [Steve Henson] 2997109998Smarkm 2998109998Smarkm *) New function X509_signature_print() to remove duplication in some 2999109998Smarkm print routines. 3000109998Smarkm [Steve Henson] 3001109998Smarkm 3002109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 3003109998Smarkm set (this was treated exactly the same as SET OF previously). This 3004109998Smarkm is used to reorder the STACK representing the structure to match the 3005109998Smarkm encoding. This will be used to get round a problem where a PKCS7 3006109998Smarkm structure which was signed could not be verified because the STACK 3007109998Smarkm order did not reflect the encoded order. 3008109998Smarkm [Steve Henson] 3009109998Smarkm 3010109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 3011109998Smarkm [Steve Henson] 3012109998Smarkm 3013109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 3014109998Smarkm for its ASN1 operations. The old style function pointers still exist 3015109998Smarkm for now but they will eventually go away. 3016109998Smarkm [Steve Henson] 3017109998Smarkm 3018109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 3019109998Smarkm completely replaces the old ASN1 functionality with a table driven 3020109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 3021109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 3022109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 3023109998Smarkm has also been converted to the new form. 3024109998Smarkm [Steve Henson] 3025109998Smarkm 3026109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 3027109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 3028109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 3029109998Smarkm for negative moduli. 3030109998Smarkm [Bodo Moeller] 3031109998Smarkm 3032109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 3033109998Smarkm of not touching the result's sign bit. 3034109998Smarkm [Bodo Moeller] 3035109998Smarkm 3036109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 3037109998Smarkm set. 3038109998Smarkm [Bodo Moeller] 3039109998Smarkm 3040109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 3041109998Smarkm macros to declare and implement thin (optionally static) functions 3042109998Smarkm that provide type-safety and avoid function pointer casting for the 3043109998Smarkm type-specific callbacks. 3044109998Smarkm [Geoff Thorpe] 3045109998Smarkm 3046109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 3047109998Smarkm RFC 2712. 3048109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 3049109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 3050109998Smarkm 3051109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 3052109998Smarkm in sections depending on the subject. 3053109998Smarkm [Richard Levitte] 3054109998Smarkm 3055109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 3056109998Smarkm Windows. 3057109998Smarkm [Richard Levitte] 3058109998Smarkm 3059109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 3060109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 3061109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 3062109998Smarkm be handled deterministically). 3063109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3064109998Smarkm 3065109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 3066109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 3067109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 3068109998Smarkm [Bodo Moeller] 3069109998Smarkm 3070109998Smarkm *) New function BN_kronecker. 3071109998Smarkm [Bodo Moeller] 3072109998Smarkm 3073109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 3074109998Smarkm positive unless both parameters are zero. 3075109998Smarkm Previously something reasonably close to an infinite loop was 3076109998Smarkm possible because numbers could be growing instead of shrinking 3077109998Smarkm in the implementation of Euclid's algorithm. 3078109998Smarkm [Bodo Moeller] 3079109998Smarkm 3080109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 3081109998Smarkm sign of the number in question. 3082109998Smarkm 3083109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 3084109998Smarkm 3085109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 3086109998Smarkm because its test if the absolute value of 'a' equals 'w'. 3087109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 3088109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 3089109998Smarkm BN_is_one(), and BN_is_word(). 3090109998Smarkm [Bodo Moeller] 3091109998Smarkm 3092109998Smarkm *) New function BN_swap. 3093109998Smarkm [Bodo Moeller] 3094109998Smarkm 3095109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 3096109998Smarkm the exponentiation functions are more likely to produce reasonable 3097109998Smarkm results on negative inputs. 3098109998Smarkm [Bodo Moeller] 3099109998Smarkm 3100109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 3101109998Smarkm Previously, it could be negative if one of the factors was negative; 3102109998Smarkm I don't think anyone really wanted that behaviour. 3103109998Smarkm [Bodo Moeller] 3104109998Smarkm 3105109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 3106109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 3107109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 3108109998Smarkm and add new functions: 3109109998Smarkm 3110109998Smarkm BN_nnmod 3111109998Smarkm BN_mod_sqr 3112109998Smarkm BN_mod_add 3113109998Smarkm BN_mod_add_quick 3114109998Smarkm BN_mod_sub 3115109998Smarkm BN_mod_sub_quick 3116109998Smarkm BN_mod_lshift1 3117109998Smarkm BN_mod_lshift1_quick 3118109998Smarkm BN_mod_lshift 3119109998Smarkm BN_mod_lshift_quick 3120109998Smarkm 3121109998Smarkm These functions always generate non-negative results. 3122109998Smarkm 3123109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 3124109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 3125109998Smarkm 3126109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 3127109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 3128109998Smarkm be reduced modulo m. 3129109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3130109998Smarkm 3131111147Snectar#if 0 3132111147Snectar The following entry accidentily appeared in the CHANGES file 3133111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 3134111147Snectar it do *not* apply to OpenSSL 0.9.7. 3135111147Snectar 3136109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3137109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 3138109998Smarkm required a small change in bn_mul_part_recursive() and the addition 3139109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 3140109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 3141109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 3142109998Smarkm differing sizes. 3143109998Smarkm [Richard Levitte] 3144111147Snectar#endif 3145109998Smarkm 3146109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 3147109998Smarkm unless the '-salt' option is used (which usually means that 3148109998Smarkm verification would just waste user's time since the resulting 3149109998Smarkm hash is going to be compared with some given password hash) 3150109998Smarkm or the new '-noverify' option is used. 3151109998Smarkm 3152109998Smarkm This is an incompatible change, but it does not affect 3153109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 3154109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 3155109998Smarkm cause any problems. 3156109998Smarkm [Bodo Moeller] 3157109998Smarkm 3158109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 3159109998Smarkm [Richard Levitte] 3160109998Smarkm 3161109998Smarkm *) Make DSO load along a path given through an environment variable 3162109998Smarkm (SHLIB_PATH) with shl_load(). 3163109998Smarkm [Richard Levitte] 3164109998Smarkm 3165109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 3166109998Smarkm Also constify the RSA code and most things related to it. In a 3167109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 3168109998Smarkm casts back to non-const were required (to be solved at a later 3169109998Smarkm time) 3170109998Smarkm [Richard Levitte] 3171109998Smarkm 3172109998Smarkm *) Make it so the openssl application has all engines loaded by default. 3173109998Smarkm [Richard Levitte] 3174109998Smarkm 3175109998Smarkm *) Constify the BIGNUM routines a little more. 3176109998Smarkm [Richard Levitte] 3177109998Smarkm 3178109998Smarkm *) Add the following functions: 3179109998Smarkm 3180109998Smarkm ENGINE_load_cswift() 3181109998Smarkm ENGINE_load_chil() 3182109998Smarkm ENGINE_load_atalla() 3183109998Smarkm ENGINE_load_nuron() 3184109998Smarkm ENGINE_load_builtin_engines() 3185109998Smarkm 3186109998Smarkm That way, an application can itself choose if external engines that 3187109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 3188109998Smarkm that applications won't have to be linked with libdl or other dso 3189109998Smarkm libraries unless it's really needed. 3190109998Smarkm 3191109998Smarkm Changed 'openssl engine' to load all engines on demand. 3192109998Smarkm Changed the engine header files to avoid the duplication of some 3193109998Smarkm declarations (they differed!). 3194109998Smarkm [Richard Levitte] 3195109998Smarkm 3196109998Smarkm *) 'openssl engine' can now list capabilities. 3197109998Smarkm [Richard Levitte] 3198109998Smarkm 3199109998Smarkm *) Better error reporting in 'openssl engine'. 3200109998Smarkm [Richard Levitte] 3201109998Smarkm 3202109998Smarkm *) Never call load_dh_param(NULL) in s_server. 3203109998Smarkm [Bodo Moeller] 3204109998Smarkm 3205109998Smarkm *) Add engine application. It can currently list engines by name and 3206109998Smarkm identity, and test if they are actually available. 3207109998Smarkm [Richard Levitte] 3208109998Smarkm 3209109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 3210109998Smarkm sure the installed documentation is also owned by root.root. 3211109998Smarkm [Damien Miller <djm@mindrot.org>] 3212109998Smarkm 3213109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 3214109998Smarkm keys (public as well as private) handled by engines. 3215109998Smarkm [Richard Levitte] 3216109998Smarkm 3217109998Smarkm *) Add OCSP code that comes from CertCo. 3218109998Smarkm [Richard Levitte] 3219109998Smarkm 3220109998Smarkm *) Add VMS support for the Rijndael code. 3221109998Smarkm [Richard Levitte] 3222109998Smarkm 3223109998Smarkm *) Added untested support for Nuron crypto accelerator. 3224109998Smarkm [Ben Laurie] 3225109998Smarkm 3226109998Smarkm *) Add support for external cryptographic devices. This code was 3227109998Smarkm previously distributed separately as the "engine" branch. 3228109998Smarkm [Geoff Thorpe, Richard Levitte] 3229109998Smarkm 3230109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 3231109998Smarkm have far greater control over how a "name" is turned into a filename 3232109998Smarkm depending on the operating environment and any oddities about the 3233109998Smarkm different shared library filenames on each system. 3234109998Smarkm [Geoff Thorpe] 3235109998Smarkm 3236109998Smarkm *) Support threads on FreeBSD-elf in Configure. 3237109998Smarkm [Richard Levitte] 3238109998Smarkm 3239109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 3240109998Smarkm warnings about corrupt line number information when assembling 3241109998Smarkm with debugging information. This is caused by the overlapping 3242109998Smarkm of two sections. 3243109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 3244109998Smarkm 3245109998Smarkm *) NCONF changes. 3246109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 3247109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 3248109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 3249109998Smarkm binary backward compatibility. 3250109998Smarkm Make it possible for methods to load from something other than a BIO, 3251109998Smarkm by providing a function pointer that is given a name instead of a BIO. 3252109998Smarkm For example, this could be used to load configuration data from an 3253109998Smarkm LDAP server. 3254109998Smarkm [Richard Levitte] 3255109998Smarkm 3256109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 3257109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 3258109998Smarkm with non blocking I/O was not possible because no retry code was 3259109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 3260109998Smarkm this case. 3261109998Smarkm [Steve Henson] 3262109998Smarkm 3263109998Smarkm *) Added the beginnings of Rijndael support. 3264109998Smarkm [Ben Laurie] 3265109998Smarkm 3266109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 3267109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 3268109998Smarkm to allow certificate printing to more controllable, additional 3269109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 3270109998Smarkm set. 3271109998Smarkm [Steve Henson] 3272109998Smarkm 3273109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 3274109998Smarkm [Richard Levitte] 3275109998Smarkm 3276142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 3277142425Snectar 3278142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3279160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3280142425Snectar [Joe Orton, Steve Henson] 3281142425Snectar 3282142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 3283142425Snectar 3284142425Snectar *) Fix additional bug revealed by the NISCC test suite: 3285142425Snectar 3286142425Snectar Stop bug triggering large recursion when presented with 3287160814Ssimon certain ASN.1 tags (CVE-2003-0851) 3288142425Snectar [Steve Henson] 3289142425Snectar 3290120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 3291120631Snectar 3292120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3293120631Snectar 3294120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3295160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3296120631Snectar 3297120631Snectar If verify callback ignores invalid public key errors don't try to check 3298120631Snectar certificate signature with the NULL public key. 3299120631Snectar 3300120631Snectar [Steve Henson] 3301120631Snectar 3302120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3303120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3304120631Snectar specifications. 3305120631Snectar [Steve Henson] 3306120631Snectar 3307120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3308120631Snectar extra data after the compression methods not only for TLS 1.0 3309120631Snectar but also for SSL 3.0 (as required by the specification). 3310120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3311120631Snectar 3312120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3313120631Snectar when it's 512 *bits* long, not 512 bytes. 3314120631Snectar [Richard Levitte] 3315120631Snectar 3316120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 3317120631Snectar 3318120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3319120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3320120631Snectar a protocol version number mismatch like a decryption error 3321120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3322120631Snectar [Bodo Moeller] 3323120631Snectar 3324120631Snectar *) Turn on RSA blinding by default in the default implementation 3325120631Snectar to avoid a timing attack. Applications that don't want it can call 3326120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3327120631Snectar They would be ill-advised to do so in most cases. 3328120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3329120631Snectar 3330120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 3331120631Snectar seeded (in this case, the secret RSA exponent is abused as 3332120631Snectar an unpredictable seed -- if it is not unpredictable, there 3333120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 3334120631Snectar by remembering the creator's thread ID in rsa->blinding and 3335120631Snectar having all other threads use local one-time blinding factors 3336120631Snectar (this requires more computation than sharing rsa->blinding, but 3337120631Snectar avoids excessive locking; and if an RSA object is not shared 3338120631Snectar between threads, blinding will still be very fast). 3339120631Snectar [Bodo Moeller] 3340120631Snectar 3341111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 3342111147Snectar 3343111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 3344111147Snectar via timing by performing a MAC computation even if incorrrect 3345111147Snectar block cipher padding has been found. This is a countermeasure 3346111147Snectar against active attacks where the attacker has to distinguish 3347160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 3348111147Snectar 3349111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 3350111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 3351111147Snectar Martin Vuagnoux (EPFL, Ilion)] 3352111147Snectar 3353109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 3354109998Smarkm 3355109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 3356109998Smarkm memory from it's contents. This is done with a counter that will 3357109998Smarkm place alternating values in each byte. This can be used to solve 3358109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 3359109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 3360109998Smarkm be read through on certain media, for example a swap space on disk. 3361109998Smarkm [Geoff Thorpe] 3362109998Smarkm 3363109998Smarkm *) Bugfix: client side session caching did not work with external caching, 3364109998Smarkm because the session->cipher setting was not restored when reloading 3365109998Smarkm from the external cache. This problem was masked, when 3366109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 3367109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 3368109998Smarkm [Lutz Jaenicke] 3369109998Smarkm 3370109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 3371109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 3372109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 3373109998Smarkm 3374109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 3375109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 3376109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 3377109998Smarkm EVP_cleanup(). 3378109998Smarkm [Richard Levitte] 3379109998Smarkm 3380109998Smarkm *) Change the default configuration reader to deal with last line not 3381109998Smarkm being properly terminated. 3382109998Smarkm [Richard Levitte] 3383109998Smarkm 3384109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 3385109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 3386109998Smarkm emailAddress where the value has the type ia5String. 3387109998Smarkm [stefank@valicert.com via Richard Levitte] 3388109998Smarkm 3389109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 3390109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 3391109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 3392109998Smarkm the bitwise-OR of the two for use by the majority of applications 3393109998Smarkm wanting this behaviour, and update the docs. The documented 3394109998Smarkm behaviour and actual behaviour were inconsistent and had been 3395109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 3396109998Smarkm change. 3397109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 3398109998Smarkm 3399109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 3400109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 3401109998Smarkm [Bodo Moeller] 3402109998Smarkm 3403109998Smarkm *) Fix initialization code race conditions in 3404109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 3405109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 3406109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 3407109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 3408109998Smarkm ssl2_get_cipher_by_char(), 3409109998Smarkm ssl3_get_cipher_by_char(). 3410109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 3411109998Smarkm 3412109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 3413109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 3414109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 3415109998Smarkm (see [openssl.org #212]). 3416109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 3417109998Smarkm 3418109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 3419109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 3420109998Smarkm [Steve Henson] 3421109998Smarkm 3422101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 3423101618Snectar 3424101618Snectar *) [In 0.9.6g-engine release:] 3425101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 3426101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 3427101618Snectar 3428101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 3429101613Snectar 3430101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 3431101613Snectar and get fix the header length calculation. 3432101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 3433101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 3434101613Snectar Steve Henson] 3435101613Snectar 3436101613Snectar *) Use proper error handling instead of 'assertions' in buffer 3437101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 3438101613Snectar assertions could call abort()). 3439101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 3440101613Snectar 3441100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 3442100936Snectar 3443109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 3444109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 3445109998Smarkm negative or the content length exceeds the length of the 3446109998Smarkm supplied buffer. 3447109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 3448109998Smarkm 3449100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 3450100936Snectar for the cipher strength set and where therefore not handled correctly 3451100936Snectar by the selection routines (PR #130). 3452100936Snectar [Lutz Jaenicke] 3453100936Snectar 3454100936Snectar *) Fix EVP_dsa_sha macro. 3455100936Snectar [Nils Larsch] 3456100936Snectar 3457100936Snectar *) New option 3458100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 3459100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 3460100936Snectar that was added in OpenSSL 0.9.6d. 3461100936Snectar 3462100936Snectar As the countermeasure turned out to be incompatible with some 3463100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 3464100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 3465100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 3466100936Snectar 's_server'), so the new option is automatically set in many 3467100936Snectar applications. 3468100936Snectar [Bodo Moeller] 3469100936Snectar 3470100936Snectar *) Changes in security patch: 3471100936Snectar 3472100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 3473100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 3474100936Snectar Air Force Materiel Command, USAF, under agreement number 3475100936Snectar F30602-01-2-0537. 3476100936Snectar 3477100936Snectar *) Add various sanity checks to asn1_get_length() to reject 3478100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 3479100936Snectar negative or the content length exceeds the length of the 3480160814Ssimon supplied buffer. (CVE-2002-0659) 3481100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 3482100936Snectar 3483100936Snectar *) Assertions for various potential buffer overflows, not known to 3484100936Snectar happen in practice. 3485100936Snectar [Ben Laurie (CHATS)] 3486100936Snectar 3487100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 3488160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 3489100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 3490100936Snectar 3491100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 3492160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 3493100936Snectar [Ben Laurie (CHATS)] 3494100936Snectar 3495100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 3496160814Ssimon supply an oversized client master key. (CVE-2002-0656) 3497100936Snectar [Ben Laurie (CHATS)] 3498100936Snectar 3499100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 3500100928Snectar 3501100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 3502100928Snectar encoded as NULL) with id-dsa-with-sha1. 3503100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 3504100928Snectar 3505100928Snectar *) Check various X509_...() return values in apps/req.c. 3506100928Snectar [Nils Larsch <nla@trustcenter.de>] 3507100928Snectar 3508100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 3509100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 3510100928Snectar was just at the end of a processed block. The bug was discovered when 3511100928Snectar processing data through a buffering memory BIO handing the data to a 3512100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 3513100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 3514100928Snectar [Lutz Jaenicke] 3515100928Snectar 3516100928Snectar *) Implement a countermeasure against a vulnerability recently found 3517100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 3518100928Snectar before application data chunks to avoid the use of known IVs 3519100928Snectar with data potentially chosen by the attacker. 3520100928Snectar [Bodo Moeller] 3521100928Snectar 3522100928Snectar *) Fix length checks in ssl3_get_client_hello(). 3523100928Snectar [Bodo Moeller] 3524100928Snectar 3525100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 3526100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 3527100928Snectar ssl3_read_bytes() found application data while handshake 3528100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 3529100928Snectar merely automatically cleared during the initial handshake. 3530100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 3531100928Snectar 3532100928Snectar *) Fix object definitions for Private and Enterprise: they were not 3533100928Snectar recognized in their shortname (=lowercase) representation. Extend 3534100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 3535100928Snectar of silently ignoring the problem (Svenning Sorensen 3536100928Snectar <sss@sss.dnsalias.net>). 3537100928Snectar [Lutz Jaenicke] 3538100928Snectar 3539100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 3540100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 3541100928Snectar code did not properly initialise the 'add' and 'rem' values to 3542100928Snectar BN_generate_prime().) 3543100928Snectar 3544100928Snectar In the new general case, we do not insist that 'generator' is 3545100928Snectar actually a primitive root: This requirement is rather pointless; 3546100928Snectar a generator of the order-q subgroup is just as good, if not 3547100928Snectar better. 3548100928Snectar [Bodo Moeller] 3549100928Snectar 3550100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 3551100928Snectar Tom Wu <tom@arcot.com>. 3552100928Snectar [Lutz Jaenicke] 3553100928Snectar 3554100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 3555100928Snectar returning non-zero before the data has been completely received 3556100928Snectar when using non-blocking I/O. 3557100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 3558100928Snectar 3559100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 3560100928Snectar [Ben Laurie, Lutz Jaenicke] 3561100928Snectar 3562100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 3563100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 3564100928Snectar [Lutz Jaenicke] 3565100928Snectar 3566100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 3567100928Snectar configuration for the versions before that. 3568100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 3569100928Snectar 3570100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 3571100928Snectar check whether we deal with a copy of a session and do not delete from 3572100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 3573100928Snectar <izhar@checkpoint.com>. 3574100928Snectar [Lutz Jaenicke] 3575100928Snectar 3576100928Snectar *) Do not store session data into the internal session cache, if it 3577100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 3578100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 3579100928Snectar [Lutz Jaenicke] 3580100928Snectar 3581100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 3582100928Snectar value is 0. 3583100928Snectar [Richard Levitte] 3584100928Snectar 3585109998Smarkm *) [In 0.9.6d-engine release:] 3586109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 3587100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 3588100928Snectar 3589100928Snectar *) Add the configuration target linux-s390x. 3590100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 3591100928Snectar 3592100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 3593100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 3594100928Snectar variable as an indication that a ClientHello message has been 3595100928Snectar received. As the flag value will be lost between multiple 3596100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 3597100928Snectar function may not be aware that a handshake has actually taken 3598100928Snectar place, thus preventing a new session from being added to the 3599100928Snectar session cache. 3600100928Snectar 3601100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 3602100928Snectar using a local variable. 3603100928Snectar [Lutz Jaenicke, Bodo Moeller] 3604100928Snectar 3605100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 3606100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 3607100928Snectar [Geoff Thorpe, Bodo Moeller] 3608100928Snectar 3609100928Snectar *) New 'shared_ldflag' column in Configure platform table. 3610100928Snectar [Richard Levitte] 3611100928Snectar 3612100928Snectar *) Fix EVP_CIPHER_mode macro. 3613100928Snectar ["Dan S. Camper" <dan@bti.net>] 3614100928Snectar 3615100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 3616100928Snectar type, we must throw them away by setting rr->length to 0. 3617100928Snectar [D P Chang <dpc@qualys.com>] 3618100928Snectar 361989837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 362089837Skris 362189837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 362289837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 362389837Skris worked incorrectly for those cases where range = 10..._2 and 362489837Skris 3*range is two bits longer than range.) 362589837Skris [Bodo Moeller] 362689837Skris 362789837Skris *) Only add signing time to PKCS7 structures if it is not already 362889837Skris present. 362989837Skris [Steve Henson] 363089837Skris 363189837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 363289837Skris OBJ_ld_ce should be OBJ_id_ce. 363389837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 363489837Skris incorrect (cf. RFC 3039). 363589837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 363689837Skris 363789837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 363889837Skris returns early because it has nothing to do. 363989837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 364089837Skris 364189837Skris *) [In 0.9.6c-engine release:] 364289837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 364389837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 364489837Skris 364589837Skris *) [In 0.9.6c-engine release:] 364689837Skris Add support for Cryptographic Appliance's keyserver technology. 364789837Skris (Use engine 'keyclient') 364889837Skris [Cryptographic Appliances and Geoff Thorpe] 364989837Skris 365089837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 365189837Skris is called via tools/c89.sh because arguments have to be 365289837Skris rearranged (all '-L' options must appear before the first object 365389837Skris modules). 365489837Skris [Richard Shapiro <rshapiro@abinitio.com>] 365589837Skris 365689837Skris *) [In 0.9.6c-engine release:] 365789837Skris Add support for Broadcom crypto accelerator cards, backported 365889837Skris from 0.9.7. 365989837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 366089837Skris 366189837Skris *) [In 0.9.6c-engine release:] 366289837Skris Add support for SureWare crypto accelerator cards from 366389837Skris Baltimore Technologies. (Use engine 'sureware') 366489837Skris [Baltimore Technologies and Mark Cox] 366589837Skris 366689837Skris *) [In 0.9.6c-engine release:] 366789837Skris Add support for crypto accelerator cards from Accelerated 366889837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 366989837Skris [AEP Inc. and Mark Cox] 367089837Skris 367189837Skris *) Add a configuration entry for gcc on UnixWare. 367289837Skris [Gary Benson <gbenson@redhat.com>] 367389837Skris 367489837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 367589837Skris messages are stored in a single piece (fixed-length part and 367689837Skris variable-length part combined) and fix various bugs found on the way. 367789837Skris [Bodo Moeller] 367889837Skris 367989837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 368089837Skris instead. BIO_gethostbyname() does not know what timeouts are 368189837Skris appropriate, so entries would stay in cache even when they have 368289837Skris become invalid. 368389837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 368489837Skris 368589837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 368689837Skris faced with a pathologically small ClientHello fragment that does 368789837Skris not contain client_version: Instead of aborting with an error, 368889837Skris simply choose the highest available protocol version (i.e., 368989837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 369089837Skris messages are never sent like this, but this change gives us 369189837Skris strictly correct behaviour at least for TLS. 369289837Skris [Bodo Moeller] 369389837Skris 369489837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 369589837Skris never resets s->method to s->ctx->method when called from within 369689837Skris one of the SSL handshake functions. 369789837Skris [Bodo Moeller; problem pointed out by Niko Baric] 369889837Skris 369989837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 370089837Skris (sent using the client's version number) if client_version is 370189837Skris smaller than the protocol version in use. Also change 370289837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 370389837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 370489837Skris the client will at least see that alert. 370589837Skris [Bodo Moeller] 370689837Skris 370789837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 370889837Skris correctly. 370989837Skris [Bodo Moeller] 371089837Skris 371189837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 371289837Skris client receives HelloRequest while in a handshake. 371389837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 371489837Skris 371589837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 371689837Skris should end in 'break', not 'goto end' which circuments various 371789837Skris cleanups done in state SSL_ST_OK. But session related stuff 371889837Skris must be disabled for SSL_ST_OK in the case that we just sent a 371989837Skris HelloRequest. 372089837Skris 372189837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 372289837Skris before just sending a HelloRequest. 372389837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 372489837Skris 372589837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 372689837Skris reveal whether illegal block cipher padding was found or a MAC 372789837Skris verification error occured. (Neither SSLerr() codes nor alerts 372889837Skris are directly visible to potential attackers, but the information 372989837Skris may leak via logfiles.) 373089837Skris 373189837Skris Similar changes are not required for the SSL 2.0 implementation 373289837Skris because the number of padding bytes is sent in clear for SSL 2.0, 373389837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 373489837Skris failed to verify that the purported number of padding bytes is in 373589837Skris the legal range. 373689837Skris [Bodo Moeller] 373789837Skris 373889837Skris *) Add OpenUNIX-8 support including shared libraries 373989837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 374089837Skris [Lutz Jaenicke] 374189837Skris 374289837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 374389837Skris 'wristwatch attack' using huge encoding parameters (cf. 374489837Skris James H. Manger's CRYPTO 2001 paper). Note that the 374589837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 374689837Skris encoding parameters and hence was not vulnerable. 374789837Skris [Bodo Moeller] 374889837Skris 374989837Skris *) BN_sqr() bug fix. 375089837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 375189837Skris 375289837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 375389837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 375489837Skris followed by modular reduction. 375589837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 375689837Skris 375789837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 375889837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 375989837Skris [Bodo Moeller] 376089837Skris 376189837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 376289837Skris This function was broken, as the check for a new client hello message 376389837Skris to handle SGC did not allow these large messages. 376489837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 376589837Skris [Lutz Jaenicke] 376689837Skris 376789837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 376889837Skris [Lutz Jaenicke] 376989837Skris 377089837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 377189837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 377289837Skris [Lutz Jaenicke] 377389837Skris 377489837Skris *) Rework the configuration and shared library support for Tru64 Unix. 377589837Skris The configuration part makes use of modern compiler features and 377689837Skris still retains old compiler behavior for those that run older versions 377789837Skris of the OS. The shared library support part includes a variant that 377889837Skris uses the RPATH feature, and is available through the special 377989837Skris configuration target "alpha-cc-rpath", which will never be selected 378089837Skris automatically. 378189837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 378289837Skris 378389837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 378489837Skris with the same message size as in ssl3_get_certificate_request(). 378589837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 378689837Skris messages might inadvertently be reject as too long. 378789837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 378889837Skris 378989837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 379089837Skris [Andy Polyakov] 379189837Skris 379289837Skris *) Modified SSL library such that the verify_callback that has been set 379389837Skris specificly for an SSL object with SSL_set_verify() is actually being 379489837Skris used. Before the change, a verify_callback set with this function was 379589837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 379689837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 379789837Skris to allow the necessary settings. 379889837Skris [Lutz Jaenicke] 379989837Skris 380089837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 380189837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 380289837Skris done automatically (in contradiction to the requirements of the C 380389837Skris standard). This made problems when used from OpenSSH. 380489837Skris [Lutz Jaenicke] 380589837Skris 380689837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 380789837Skris dh->length and always used 380889837Skris 380989837Skris BN_rand_range(priv_key, dh->p). 381089837Skris 381189837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 381289837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 381389837Skris dh->length (recommended exponent length) is much smaller than the 381489837Skris length of dh->p. We could use BN_rand_range() if the order of 381589837Skris the subgroup was stored in the DH structure, but we only have 381689837Skris dh->length. 381789837Skris 381889837Skris So switch back to 381989837Skris 382089837Skris BN_rand(priv_key, l, ...) 382189837Skris 382289837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 382389837Skris otherwise. 382489837Skris [Bodo Moeller] 382589837Skris 382689837Skris *) In 382789837Skris 382889837Skris RSA_eay_public_encrypt 382989837Skris RSA_eay_private_decrypt 383089837Skris RSA_eay_private_encrypt (signing) 383189837Skris RSA_eay_public_decrypt (signature verification) 383289837Skris 383389837Skris (default implementations for RSA_public_encrypt, 383489837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 383589837Skris always reject numbers >= n. 383689837Skris [Bodo Moeller] 383789837Skris 383889837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 383989837Skris to synchronize access to 'locking_thread'. This is necessary on 384089837Skris systems where access to 'locking_thread' (an 'unsigned long' 384189837Skris variable) is not atomic. 384289837Skris [Bodo Moeller] 384389837Skris 384489837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 384589837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 384689837Skris a race condition if 0 is a valid thread ID. 384789837Skris [Travis Vitek <vitek@roguewave.com>] 384889837Skris 384989837Skris *) Add support for shared libraries under Irix. 385089837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 385189837Skris 385289837Skris *) Add configuration option to build on Linux on both big-endian and 385389837Skris little-endian MIPS. 385489837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 385589837Skris 385689837Skris *) Add the possibility to create shared libraries on HP-UX. 385789837Skris [Richard Levitte] 385889837Skris 385979998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 386079998Skris 386179998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 386279998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 386379998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 386479998Skris PRNG state recovery was possible based on the output of 386579998Skris one PRNG request appropriately sized to gain knowledge on 386679998Skris 'md' followed by enough consecutive 1-byte PRNG requests 386779998Skris to traverse all of 'state'. 386879998Skris 386979998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 387079998Skris during PRNG output generation, hash all of the previous 387179998Skris 'md_local' value, not just the half used for PRNG output. 387279998Skris 387379998Skris 2. Make the number of bytes from 'state' included into the hash 387479998Skris independent from the number of PRNG bytes requested. 387579998Skris 387679998Skris The first measure alone would be sufficient to avoid 387779998Skris Markku-Juhani's attack. (Actually it had never occurred 387879998Skris to me that the half of 'md_local' used for chaining was the 387979998Skris half from which PRNG output bytes were taken -- I had always 388079998Skris assumed that the secret half would be used.) The second 388179998Skris measure makes sure that additional data from 'state' is never 388279998Skris mixed into 'md_local' in small portions; this heuristically 388379998Skris further strengthens the PRNG. 388479998Skris [Bodo Moeller] 388579998Skris 388679998Skris *) Fix crypto/bn/asm/mips3.s. 388779998Skris [Andy Polyakov] 388879998Skris 388979998Skris *) When only the key is given to "enc", the IV is undefined. Print out 389079998Skris an error message in this case. 389179998Skris [Lutz Jaenicke] 389279998Skris 389379998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 389479998Skris [Steve Henson] 389579998Skris 389679998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 389779998Skris positive and less than q. 389879998Skris [Bodo Moeller] 389979998Skris 390079998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 390179998Skris used: it isn't thread safe and the add_lock_callback should handle 390279998Skris that itself. 390379998Skris [Paul Rose <Paul.Rose@bridge.com>] 390479998Skris 390579998Skris *) Verify that incoming data obeys the block size in 390679998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 390779998Skris [Bodo Moeller] 390879998Skris 390979998Skris *) Fix OAEP check. 391079998Skris [Ulf M�ller, Bodo M�ller] 391179998Skris 391279998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 391389837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 391479998Skris when fixing the server behaviour for backwards-compatible 'client 391579998Skris hello' messages. (Note that the attack is impractical against 391679998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 391779998Skris means that the probability of guessing a valid ciphertext is 391879998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 391979998Skris paper.) 392079998Skris 392179998Skris Before 0.9.5, the countermeasure (hide the error by generating a 392279998Skris random 'decryption result') did not work properly because 392379998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 392479998Skris detect the supposedly ignored error. 392579998Skris 392679998Skris Both problems are now fixed. 392779998Skris [Bodo Moeller] 392879998Skris 392979998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 393079998Skris (previously it was 1024). 393179998Skris [Bodo Moeller] 393279998Skris 393379998Skris *) Fix for compatibility mode trust settings: ignore trust settings 393479998Skris unless some valid trust or reject settings are present. 393579998Skris [Steve Henson] 393679998Skris 393779998Skris *) Fix for blowfish EVP: its a variable length cipher. 393879998Skris [Steve Henson] 393979998Skris 394079998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 394179998Skris parameters in DSA public key structures and return an error in the 394279998Skris DSA routines if parameters are absent. 394379998Skris [Steve Henson] 394479998Skris 394579998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 394679998Skris in the current directory if neither $RANDFILE nor $HOME was set. 394779998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 394879998Skris caused some confusion to Windows users who haven't defined $HOME. 394979998Skris Thus RAND_file_name() is changed again: e_os.h can define a 395079998Skris DEFAULT_HOME, which will be used if $HOME is not set. 395179998Skris For Windows, we use "C:"; on other platforms, we still require 395279998Skris environment variables. 395379998Skris 395479998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 395579998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 395679998Skris having multiple threads call RAND_poll() concurrently. 395779998Skris [Bodo Moeller] 395879998Skris 395979998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 396079998Skris combination of a flag and a thread ID variable. 396179998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 396279998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 396389837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 396479998Skris that they do not hold after the first thread unsets add_do_not_lock). 396579998Skris [Bodo Moeller] 396679998Skris 396779998Skris *) Change bctest again: '-x' expressions are not available in all 396879998Skris versions of 'test'. 396979998Skris [Bodo Moeller] 397079998Skris 397176866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 397272613Skris 397376866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 397476866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 397576866Skris 397676866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 397776866Skris the default extension for executables, if any. Also, make the perl 397876866Skris scripts that use symlink() to test if it really exists and use "cp" 397976866Skris if it doesn't. All this made OpenSSL compilable and installable in 398076866Skris CygWin. 398176866Skris [Richard Levitte] 398276866Skris 398376866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 398476866Skris If SEQUENCE is length is indefinite just set c->slen to the total 398576866Skris amount of data available. 398676866Skris [Steve Henson, reported by shige@FreeBSD.org] 398776866Skris [This change does not apply to 0.9.7.] 398876866Skris 398976866Skris *) Change bctest to avoid here-documents inside command substitution 399076866Skris (workaround for FreeBSD /bin/sh bug). 399176866Skris For compatibility with Ultrix, avoid shell functions (introduced 399276866Skris in the bctest version that searches along $PATH). 399376866Skris [Bodo Moeller] 399476866Skris 399576866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 399676866Skris with des_encrypt() defined on some operating systems, like Solaris 399776866Skris and UnixWare. 399876866Skris [Richard Levitte] 399976866Skris 400076866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 400176866Skris On the Importance of Eliminating Errors in Cryptographic 400276866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 400376866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 400476866Skris [Ulf Moeller] 400576866Skris 400676866Skris *) MIPS assembler BIGNUM division bug fix. 400776866Skris [Andy Polyakov] 400876866Skris 400976866Skris *) Disabled incorrect Alpha assembler code. 401076866Skris [Richard Levitte] 401176866Skris 401276866Skris *) Fix PKCS#7 decode routines so they correctly update the length 401376866Skris after reading an EOC for the EXPLICIT tag. 401476866Skris [Steve Henson] 401576866Skris [This change does not apply to 0.9.7.] 401676866Skris 401776866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 401876866Skris if a 3DES key was generated with a 0 initial byte. Include 401976866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 402076866Skris (but broken) behaviour. 402176866Skris [Steve Henson] 402276866Skris 402376866Skris *) Enhance bctest to search for a working bc along $PATH and print 402476866Skris it when found. 402576866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 402676866Skris 402776866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 402876866Skris don't write to the wrong index in ERR_set_error_data. 402976866Skris [Bodo Moeller] 403076866Skris 403176866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 403276866Skris did not exist. 403376866Skris [Bodo Moeller] 403476866Skris 403576866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 403676866Skris [Jeremy Cooper <jeremy@baymoo.org>] 403776866Skris 403876866Skris *) Make it possible to reuse SSLv2 sessions. 403976866Skris [Richard Levitte] 404076866Skris 404176866Skris *) In copy_email() check for >= 0 as a return value for 404276866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 404376866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 404476866Skris 404576866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 404676866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 404776866Skris PKCS7_verify() fails with non detached data. 404876866Skris [Steve Henson] 404976866Skris 405076866Skris *) Don't use getenv in library functions when run as setuid/setgid. 405176866Skris New function OPENSSL_issetugid(). 405276866Skris [Ulf Moeller] 405376866Skris 405476866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 405576866Skris due to incorrect handling of multi-threading: 405676866Skris 405776866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 405876866Skris 405976866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 406076866Skris 406176866Skris 3. Count how many times MemCheck_off() has been called so that 406276866Skris nested use can be treated correctly. This also avoids 406376866Skris inband-signalling in the previous code (which relied on the 406476866Skris assumption that thread ID 0 is impossible). 406576866Skris [Bodo Moeller] 406676866Skris 406776866Skris *) Add "-rand" option also to s_client and s_server. 406876866Skris [Lutz Jaenicke] 406976866Skris 407076866Skris *) Fix CPU detection on Irix 6.x. 407176866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 407276866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 407376866Skris 407476866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 407576866Skris was empty. 407676866Skris [Steve Henson] 407776866Skris [This change does not apply to 0.9.7.] 407876866Skris 407976866Skris *) Use the cached encoding of an X509_NAME structure rather than 408076866Skris copying it. This is apparently the reason for the libsafe "errors" 408176866Skris but the code is actually correct. 408276866Skris [Steve Henson] 408376866Skris 408472613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 408572613Skris Bleichenbacher's DSA attack. 408676866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 408776866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 408876866Skris and leaves the highest bit random. 408976866Skris [Ulf Moeller, Bodo Moeller] 409072613Skris 409172613Skris *) In the NCONF_...-based implementations for CONF_... queries 409272613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 409372613Skris a temporary CONF structure with the data component set to NULL 409472613Skris (which gives segmentation faults in lh_retrieve). 409572613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 409672613Skris CONF_get_number (which may use environment variables) and directly 409772613Skris return NULL from CONF_get_section. 409872613Skris [Bodo Moeller] 409972613Skris 410072613Skris *) Fix potential buffer overrun for EBCDIC. 410172613Skris [Ulf Moeller] 410272613Skris 410372613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 410472613Skris keyUsage if basicConstraints absent for a CA. 410572613Skris [Steve Henson] 410672613Skris 410772613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 410872613Skris is more generally accepted (no spaces before the semicolon), since 410972613Skris some programs can't parse those values properly otherwise. Also make 411072613Skris sure BIO's that break lines after each write do not create invalid 411172613Skris headers. 411272613Skris [Richard Levitte] 411372613Skris 411472613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 411572613Skris macros previously used would not encode an empty SEQUENCE OF 411672613Skris and break the signature. 411772613Skris [Steve Henson] 411876866Skris [This change does not apply to 0.9.7.] 411972613Skris 412072613Skris *) Zero the premaster secret after deriving the master secret in 412172613Skris DH ciphersuites. 412272613Skris [Steve Henson] 412372613Skris 412472613Skris *) Add some EVP_add_digest_alias registrations (as found in 412572613Skris OpenSSL_add_all_digests()) to SSL_library_init() 412672613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 412772613Skris compatibility with peers using X.509 certificates 412872613Skris with unconventional AlgorithmIdentifier OIDs. 412972613Skris [Bodo Moeller] 413072613Skris 413172613Skris *) Fix for Irix with NO_ASM. 413272613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 413372613Skris 413472613Skris *) ./config script fixes. 413572613Skris [Ulf Moeller, Richard Levitte] 413672613Skris 413772613Skris *) Fix 'openssl passwd -1'. 413872613Skris [Bodo Moeller] 413972613Skris 414072613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 414172613Skris terminated strings whose length is passed in the passlen 414272613Skris parameter, for example from PEM callbacks. This was done 414372613Skris by adding an extra length parameter to asc2uni(). 414472613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 414572613Skris 414672613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 414772613Skris call failed, free the DSA structure. 414872613Skris [Bodo Moeller] 414972613Skris 415072613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 415172613Skris These are present in some PKCS#12 files. 415272613Skris [Steve Henson] 415372613Skris 415472613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 415572613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 415672613Skris when writing a 32767 byte record. 415772613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 415872613Skris 415972613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 416072613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 416172613Skris 416272613Skris (RSA objects have a reference count access to which is protected 416372613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 416472613Skris so they are meant to be shared between threads.) 416572613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 416672613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 416772613Skris 416872613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 416972613Skris [Bodo Moeller] 417072613Skris 417176866Skris *) Use better test patterns in bntest. 417276866Skris [Ulf M�ller] 417376866Skris 417472613Skris *) rand_win.c fix for Borland C. 417572613Skris [Ulf M�ller] 417672613Skris 417772613Skris *) BN_rshift bugfix for n == 0. 417872613Skris [Bodo Moeller] 417972613Skris 418076866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 418176866Skris so that 'make test' does not abort just because 'bc' is broken. 418276866Skris [Bodo Moeller] 418376866Skris 418472613Skris *) Store verify_result within SSL_SESSION also for client side to 418572613Skris avoid potential security hole. (Re-used sessions on the client side 418672613Skris always resulted in verify_result==X509_V_OK, not using the original 418772613Skris result of the server certificate verification.) 418872613Skris [Lutz Jaenicke] 418972613Skris 419072613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 419172613Skris SSL3_RT_APPLICATION_DATA, return 0. 419272613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 419372613Skris [Bodo Moeller] 419472613Skris 419572613Skris *) Fix SSL_peek: 419672613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 419772613Skris releases, have been re-implemented by renaming the previous 419872613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 419972613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 420072613Skris to them. The new ssl[23]_{read,peek} functions are calls to 420172613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 420272613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 420372613Skris does the actual work for ssl3_read_internal. 420472613Skris [Bodo Moeller] 420572613Skris 420676866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 420776866Skris the method-specific "init()" handler. Also clean up ex_data after 420876866Skris calling the method-specific "finish()" handler. Previously, this was 420976866Skris happening the other way round. 421076866Skris [Geoff Thorpe] 421176866Skris 421272613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 421372613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 421472613Skris [Bodo Moeller] 421572613Skris 421676866Skris *) Make sure that shared libraries get the internal name engine with 421776866Skris the full version number and not just 0. This should mark the 421876866Skris shared libraries as not backward compatible. Of course, this should 421976866Skris be changed again when we can guarantee backward binary compatibility. 422076866Skris [Richard Levitte] 422176866Skris 422272613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 422372613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 422472613Skris 422576866Skris *) Rework the system to generate shared libraries: 422676866Skris 422776866Skris - Make note of the expected extension for the shared libraries and 422876866Skris if there is a need for symbolic links from for example libcrypto.so.0 422976866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 423076866Skris that. 423176866Skris 423276866Skris - Make as few rebuilds of the shared libraries as possible. 423376866Skris 423476866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 423576866Skris 423676866Skris - When installing, install the shared libraries separately from the 423776866Skris static ones. 423876866Skris [Richard Levitte] 423976866Skris 424072613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 424172613Skris 424272613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 424372613Skris and not in SSL_clear because the latter is also used by the 424472613Skris accept/connect functions; previously, the settings made by 424572613Skris SSL_set_read_ahead would be lost during the handshake. 424672613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 424772613Skris 424872613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 424972613Skris Previously, it would create entries for disableed algorithms no 425072613Skris matter what. 425172613Skris [Richard Levitte] 425272613Skris 425372613Skris *) Added several new manual pages for SSL_* function. 425472613Skris [Lutz Jaenicke] 425572613Skris 425668651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 425768651Skris 425868651Skris *) In ssl23_get_client_hello, generate an error message when faced 425968651Skris with an initial SSL 3.0/TLS record that is too small to contain the 426068651Skris first two bytes of the ClientHello message, i.e. client_version. 426168651Skris (Note that this is a pathologic case that probably has never happened 426268651Skris in real life.) The previous approach was to use the version number 426368651Skris from the record header as a substitute; but our protocol choice 426468651Skris should not depend on that one because it is not authenticated 426568651Skris by the Finished messages. 426668651Skris [Bodo Moeller] 426768651Skris 426868651Skris *) More robust randomness gathering functions for Windows. 426968651Skris [Jeffrey Altman <jaltman@columbia.edu>] 427068651Skris 427168651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 427268651Skris not set then we don't setup the error code for issuer check errors 427368651Skris to avoid possibly overwriting other errors which the callback does 427468651Skris handle. If an application does set the flag then we assume it knows 427568651Skris what it is doing and can handle the new informational codes 427668651Skris appropriately. 427768651Skris [Steve Henson] 427868651Skris 427968651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 428068651Skris a general "ANY" type, as such it should be able to decode anything 428168651Skris including tagged types. However it didn't check the class so it would 428268651Skris wrongly interpret tagged types in the same way as their universal 428368651Skris counterpart and unknown types were just rejected. Changed so that the 428468651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 428568651Skris that is the encoding is stored intact. There is also a new type 428668651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 428768651Skris case we have no idea what the actual type is so we just lump them all 428868651Skris together. 428968651Skris [Steve Henson] 429068651Skris 429168651Skris *) On VMS, stdout may very well lead to a file that is written to 429268651Skris in a record-oriented fashion. That means that every write() will 429368651Skris write a separate record, which will be read separately by the 429468651Skris programs trying to read from it. This can be very confusing. 429568651Skris 429668651Skris The solution is to put a BIO filter in the way that will buffer 429768651Skris text until a linefeed is reached, and then write everything a 429868651Skris line at a time, so every record written will be an actual line, 429968651Skris not chunks of lines and not (usually doesn't happen, but I've 430068651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 430168651Skris the answer. 430268651Skris 430368651Skris Currently, it's a VMS-only method, because that's where it has 430468651Skris been tested well enough. 430568651Skris [Richard Levitte] 430668651Skris 430768651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 430868651Skris it can return incorrect results. 430968651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 431068651Skris but it was in 0.9.6-beta[12].) 431168651Skris [Bodo Moeller] 431268651Skris 431368651Skris *) Disable the check for content being present when verifying detached 431468651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 431568651Skris include zero length content when signing messages. 431668651Skris [Steve Henson] 431768651Skris 431868651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 431968651Skris BIO_ctrl (for BIO pairs). 432068651Skris [Bodo M�ller] 432168651Skris 432268651Skris *) Add DSO method for VMS. 432368651Skris [Richard Levitte] 432468651Skris 432568651Skris *) Bug fix: Montgomery multiplication could produce results with the 432668651Skris wrong sign. 432768651Skris [Ulf M�ller] 432868651Skris 432968651Skris *) Add RPM specification openssl.spec and modify it to build three 433068651Skris packages. The default package contains applications, application 433168651Skris documentation and run-time libraries. The devel package contains 433268651Skris include files, static libraries and function documentation. The 433368651Skris doc package contains the contents of the doc directory. The original 433468651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 433568651Skris [Richard Levitte] 433668651Skris 433768651Skris *) Add a large number of documentation files for many SSL routines. 433868651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 433968651Skris 434068651Skris *) Add a configuration entry for Sony News 4. 434168651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 434268651Skris 434368651Skris *) Don't set the two most significant bits to one when generating a 434468651Skris random number < q in the DSA library. 434568651Skris [Ulf M�ller] 434668651Skris 434768651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 434868651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 434968651Skris the underlying transport is blocking) if a handshake took place. 435068651Skris (The default behaviour is needed by applications such as s_client 435168651Skris and s_server that use select() to determine when to use SSL_read; 435268651Skris but for applications that know in advance when to expect data, it 435368651Skris just makes things more complicated.) 435468651Skris [Bodo Moeller] 435568651Skris 435668651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 435768651Skris from EGD. 435868651Skris [Ben Laurie] 435968651Skris 436068651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 436168651Skris work better on such systems. 436268651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 436368651Skris 436468651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 436568651Skris Update PKCS12_parse() so it copies the friendlyName and the 436668651Skris keyid to the certificates aux info. 436768651Skris [Steve Henson] 436868651Skris 436968651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 437068651Skris if there was more than one signature. 437168651Skris [Sven Uszpelkat <su@celocom.de>] 437268651Skris 437368651Skris *) Major change in util/mkdef.pl to include extra information 437468651Skris about each symbol, as well as presentig variables as well 437568651Skris as functions. This change means that there's n more need 437668651Skris to rebuild the .num files when some algorithms are excluded. 437768651Skris [Richard Levitte] 437868651Skris 437968651Skris *) Allow the verify time to be set by an application, 438068651Skris rather than always using the current time. 438168651Skris [Steve Henson] 438268651Skris 438368651Skris *) Phase 2 verify code reorganisation. The certificate 438468651Skris verify code now looks up an issuer certificate by a 438568651Skris number of criteria: subject name, authority key id 438668651Skris and key usage. It also verifies self signed certificates 438768651Skris by the same criteria. The main comparison function is 438868651Skris X509_check_issued() which performs these checks. 438968651Skris 439068651Skris Lot of changes were necessary in order to support this 439168651Skris without completely rewriting the lookup code. 439268651Skris 439368651Skris Authority and subject key identifier are now cached. 439468651Skris 439568651Skris The LHASH 'certs' is X509_STORE has now been replaced 439668651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 439768651Skris LHASH can't store or retrieve multiple objects with 439868651Skris the same hash value. 439968651Skris 440068651Skris As a result various functions (which were all internal 440168651Skris use only) have changed to handle the new X509_STORE 440268651Skris structure. This will break anything that messed round 440368651Skris with X509_STORE internally. 440468651Skris 440568651Skris The functions X509_STORE_add_cert() now checks for an 440668651Skris exact match, rather than just subject name. 440768651Skris 440868651Skris The X509_STORE API doesn't directly support the retrieval 440968651Skris of multiple certificates matching a given criteria, however 441068651Skris this can be worked round by performing a lookup first 441168651Skris (which will fill the cache with candidate certificates) 441268651Skris and then examining the cache for matches. This is probably 441368651Skris the best we can do without throwing out X509_LOOKUP 441468651Skris entirely (maybe later...). 441568651Skris 441668651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 441768651Skris 441868651Skris All certificate lookup operations now go via a get_issuer() 441968651Skris callback. Although this currently uses an X509_STORE it 442068651Skris can be replaced by custom lookups. This is a simple way 442168651Skris to bypass the X509_STORE hackery necessary to make this 442268651Skris work and makes it possible to use more efficient techniques 442368651Skris in future. A very simple version which uses a simple 442468651Skris STACK for its trusted certificate store is also provided 442568651Skris using X509_STORE_CTX_trusted_stack(). 442668651Skris 442768651Skris The verify_cb() and verify() callbacks now have equivalents 442868651Skris in the X509_STORE_CTX structure. 442968651Skris 443068651Skris X509_STORE_CTX also has a 'flags' field which can be used 443168651Skris to customise the verify behaviour. 443268651Skris [Steve Henson] 443368651Skris 443468651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 443568651Skris excludes S/MIME capabilities. 443668651Skris [Steve Henson] 443768651Skris 443868651Skris *) When a certificate request is read in keep a copy of the 443968651Skris original encoding of the signed data and use it when outputing 444068651Skris again. Signatures then use the original encoding rather than 444168651Skris a decoded, encoded version which may cause problems if the 444268651Skris request is improperly encoded. 444368651Skris [Steve Henson] 444468651Skris 444568651Skris *) For consistency with other BIO_puts implementations, call 444668651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 444768651Skris BIO_write(b, ...). 444868651Skris 444968651Skris In BIO_puts, increment b->num_write as in BIO_write. 445068651Skris [Peter.Sylvester@EdelWeb.fr] 445168651Skris 445268651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 445368651Skris BN_zero, we may not return a BIGNUM with an array consisting of 445468651Skris words set to zero.) 445568651Skris [Bodo Moeller] 445668651Skris 445768651Skris *) Avoid calling abort() from within the library when problems are 445868651Skris detected, except if preprocessor symbols have been defined 445968651Skris (such as REF_CHECK, BN_DEBUG etc.). 446068651Skris [Bodo Moeller] 446168651Skris 446268651Skris *) New openssl application 'rsautl'. This utility can be 446368651Skris used for low level RSA operations. DER public key 446468651Skris BIO/fp routines also added. 446568651Skris [Steve Henson] 446668651Skris 446768651Skris *) New Configure entry and patches for compiling on QNX 4. 446868651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 446968651Skris 447068651Skris *) A demo state-machine implementation was sponsored by 447168651Skris Nuron (http://www.nuron.com/) and is now available in 447268651Skris demos/state_machine. 447368651Skris [Ben Laurie] 447468651Skris 447568651Skris *) New options added to the 'dgst' utility for signature 447668651Skris generation and verification. 447768651Skris [Steve Henson] 447868651Skris 447968651Skris *) Unrecognized PKCS#7 content types are now handled via a 448068651Skris catch all ASN1_TYPE structure. This allows unsupported 448168651Skris types to be stored as a "blob" and an application can 448268651Skris encode and decode it manually. 448368651Skris [Steve Henson] 448468651Skris 448568651Skris *) Fix various signed/unsigned issues to make a_strex.c 448668651Skris compile under VC++. 448768651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 448868651Skris 448968651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 449068651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 449168651Skris if passed a NULL BN and its argument was negative. 449268651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 449368651Skris 449468651Skris *) Modification to PKCS#7 encoding routines to output definite 449568651Skris length encoding. Since currently the whole structures are in 449668651Skris memory there's not real point in using indefinite length 449768651Skris constructed encoding. However if OpenSSL is compiled with 449868651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 449968651Skris [Steve Henson] 450068651Skris 450168651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 450268651Skris [Richard Levitte] 450368651Skris 450468651Skris *) Added more prefixes to parse for in the the strings written 450568651Skris through a logging bio, to cover all the levels that are available 450668651Skris through syslog. The prefixes are now: 450768651Skris 450868651Skris PANIC, EMERG, EMR => LOG_EMERG 450968651Skris ALERT, ALR => LOG_ALERT 451068651Skris CRIT, CRI => LOG_CRIT 451168651Skris ERROR, ERR => LOG_ERR 451268651Skris WARNING, WARN, WAR => LOG_WARNING 451368651Skris NOTICE, NOTE, NOT => LOG_NOTICE 451468651Skris INFO, INF => LOG_INFO 451568651Skris DEBUG, DBG => LOG_DEBUG 451668651Skris 451768651Skris and as before, if none of those prefixes are present at the 451868651Skris beginning of the string, LOG_ERR is chosen. 451968651Skris 452068651Skris On Win32, the LOG_* levels are mapped according to this: 452168651Skris 452268651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 452368651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 452468651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 452568651Skris 452668651Skris [Richard Levitte] 452768651Skris 452868651Skris *) Made it possible to reconfigure with just the configuration 452968651Skris argument "reconf" or "reconfigure". The command line arguments 453068651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 453168651Skris and are retrieved from there when reconfiguring. 453268651Skris [Richard Levitte] 453368651Skris 453468651Skris *) MD4 implemented. 453568651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 453668651Skris 453768651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 453868651Skris [Richard Levitte] 453968651Skris 454068651Skris *) The obj_dat.pl script was messing up the sorting of object 454168651Skris names. The reason was that it compared the quoted version 454268651Skris of strings as a result "OCSP" > "OCSP Signing" because 454368651Skris " > SPACE. Changed script to store unquoted versions of 454468651Skris names and add quotes on output. It was also omitting some 454568651Skris names from the lookup table if they were given a default 454668651Skris value (that is if SN is missing it is given the same 454768651Skris value as LN and vice versa), these are now added on the 454868651Skris grounds that if an object has a name we should be able to 454968651Skris look it up. Finally added warning output when duplicate 455068651Skris short or long names are found. 455168651Skris [Steve Henson] 455268651Skris 455368651Skris *) Changes needed for Tandem NSK. 455468651Skris [Scott Uroff <scott@xypro.com>] 455568651Skris 455668651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 455768651Skris RSA_padding_check_SSLv23(), special padding was never detected 455868651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 455968651Skris version rollback attacks was not effective. 456068651Skris 456168651Skris In s23_clnt.c, don't use special rollback-attack detection padding 456268651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 456368651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 456468651Skris SSL 2.0 is the only protocol enabled in the server. 456568651Skris [Bodo Moeller] 456668651Skris 456768651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 456868651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 456968651Skris BIO_dump_indent() are added. 457068651Skris [Richard Levitte] 457168651Skris 457268651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 457368651Skris these print out strings and name structures based on various 457468651Skris flags including RFC2253 support and proper handling of 457568651Skris multibyte characters. Added options to the 'x509' utility 457668651Skris to allow the various flags to be set. 457768651Skris [Steve Henson] 457868651Skris 457968651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 458068651Skris Also change the functions X509_cmp_current_time() and 458168651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 458268651Skris this will enable certificates using GeneralizedTime in validity 458368651Skris dates to be checked. 458468651Skris [Steve Henson] 458568651Skris 458668651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 458768651Skris negative public key encodings) on by default, 458868651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 458968651Skris [Steve Henson] 459068651Skris 459168651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 459268651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 459368651Skris the encoding can be trivially obtained from the structure. 459468651Skris [Steve Henson] 459568651Skris 459668651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 459768651Skris not read locks (CRYPTO_r_[un]lock). 459868651Skris [Bodo Moeller] 459968651Skris 460068651Skris *) A first attempt at creating official support for shared 460168651Skris libraries through configuration. I've kept it so the 460268651Skris default is static libraries only, and the OpenSSL programs 460368651Skris are always statically linked for now, but there are 460468651Skris preparations for dynamic linking in place. 460589837Skris This has been tested on Linux and Tru64. 460668651Skris [Richard Levitte] 460768651Skris 460868651Skris *) Randomness polling function for Win9x, as described in: 460968651Skris Peter Gutmann, Software Generation of Practically Strong 461068651Skris Random Numbers. 461168651Skris [Ulf M�ller] 461268651Skris 461368651Skris *) Fix so PRNG is seeded in req if using an already existing 461468651Skris DSA key. 461568651Skris [Steve Henson] 461668651Skris 461768651Skris *) New options to smime application. -inform and -outform 461868651Skris allow alternative formats for the S/MIME message including 461968651Skris PEM and DER. The -content option allows the content to be 462068651Skris specified separately. This should allow things like Netscape 462168651Skris form signing output easier to verify. 462268651Skris [Steve Henson] 462368651Skris 462468651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 462568651Skris [Steve Henson] 462668651Skris 462768651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 462868651Skris STRING types. These convert content octets to and from the 462968651Skris underlying type. The actual tag and length octets are 463068651Skris already assumed to have been read in and checked. These 463168651Skris are needed because all other string types have virtually 463268651Skris identical handling apart from the tag. By having versions 463368651Skris of the ASN1 functions that just operate on content octets 463468651Skris IMPLICIT tagging can be handled properly. It also allows 463568651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 463668651Skris and ASN1_INTEGER are identical apart from the tag. 463768651Skris [Steve Henson] 463868651Skris 463968651Skris *) Change the handling of OID objects as follows: 464068651Skris 464168651Skris - New object identifiers are inserted in objects.txt, following 464268651Skris the syntax given in objects.README. 464368651Skris - objects.pl is used to process obj_mac.num and create a new 464468651Skris obj_mac.h. 464568651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 464668651Skris obj_mac.h. 464768651Skris 464868651Skris This is currently kind of a hack, and the perl code in objects.pl 464968651Skris isn't very elegant, but it works as I intended. The simplest way 465068651Skris to check that it worked correctly is to look in obj_dat.h and 465168651Skris check the array nid_objs and make sure the objects haven't moved 465268651Skris around (this is important!). Additions are OK, as well as 465368651Skris consistent name changes. 465468651Skris [Richard Levitte] 465568651Skris 465668651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 465768651Skris [Bodo Moeller] 465868651Skris 465968651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 466068651Skris The given file adds to whatever has already been seeded into the 466168651Skris random pool through the RANDFILE configuration file option or 466268651Skris environment variable, or the default random state file. 466368651Skris [Richard Levitte] 466468651Skris 466568651Skris *) mkstack.pl now sorts each macro group into lexical order. 466668651Skris Previously the output order depended on the order the files 466768651Skris appeared in the directory, resulting in needless rewriting 466868651Skris of safestack.h . 466968651Skris [Steve Henson] 467068651Skris 467168651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 467268651Skris work arounds for the VC++ problem that it treats func() as 467368651Skris func(void). Also stripped out the parts of mkdef.pl that 467468651Skris added extra typesafe functions: these no longer exist. 467568651Skris [Steve Henson] 467668651Skris 467768651Skris *) Reorganisation of the stack code. The macros are now all 467868651Skris collected in safestack.h . Each macro is defined in terms of 467968651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 468068651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 468168651Skris this has the advantage of retaining type safety without the 468268651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 468368651Skris then the non typesafe macros are used instead. Also modified the 468468651Skris mkstack.pl script to handle the new form. Needs testing to see 468568651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 468668651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 468768651Skris and PKCS12_STACK_OF. 468868651Skris [Steve Henson] 468968651Skris 469068651Skris *) When some versions of IIS use the 'NET' form of private key the 469168651Skris key derivation algorithm is different. Normally MD5(password) is 469268651Skris used as a 128 bit RC4 key. In the modified case 469368651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 469468651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 469568651Skris as the old Netscape_RSA functions except they have an additional 469668651Skris 'sgckey' parameter which uses the modified algorithm. Also added 469768651Skris an -sgckey command line option to the rsa utility. Thanks to 469868651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 469968651Skris algorithm to openssl-dev. 470068651Skris [Steve Henson] 470168651Skris 470268651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 470368651Skris invalid expansion on some systems (SCO 5.0.5 for example). 470468651Skris Corrected to 'c.kname'. 470568651Skris [Phillip Porch <root@theporch.com>] 470668651Skris 470768651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 470868651Skris a STACK of email addresses from a certificate or request, these look 470968651Skris in the subject name and the subject alternative name extensions and 471068651Skris omit any duplicate addresses. 471168651Skris [Steve Henson] 471268651Skris 471368651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 471468651Skris This makes DSA verification about 2 % faster. 471568651Skris [Bodo Moeller] 471668651Skris 471768651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 471868651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 471968651Skris plus overhead for 1024 bit moduli). 472068651Skris This makes exponentiations about 0.5 % faster for 1024 bit 472168651Skris exponents (as measured by "openssl speed rsa2048"). 472268651Skris [Bodo Moeller] 472368651Skris 472468651Skris *) Rename memory handling macros to avoid conflicts with other 472568651Skris software: 472668651Skris Malloc => OPENSSL_malloc 472768651Skris Malloc_locked => OPENSSL_malloc_locked 472868651Skris Realloc => OPENSSL_realloc 472968651Skris Free => OPENSSL_free 473068651Skris [Richard Levitte] 473168651Skris 473268651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 473368651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 473468651Skris [Bodo Moeller] 473568651Skris 473668651Skris *) CygWin32 support. 473768651Skris [John Jarvie <jjarvie@newsguy.com>] 473868651Skris 473968651Skris *) The type-safe stack code has been rejigged. It is now only compiled 474068651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 474168651Skris by default all type-specific stack functions are "#define"d back to 474268651Skris standard stack functions. This results in more streamlined output 474368651Skris but retains the type-safety checking possibilities of the original 474468651Skris approach. 474568651Skris [Geoff Thorpe] 474668651Skris 474768651Skris *) The STACK code has been cleaned up, and certain type declarations 474868651Skris that didn't make a lot of sense have been brought in line. This has 474968651Skris also involved a cleanup of sorts in safestack.h to more correctly 475068651Skris map type-safe stack functions onto their plain stack counterparts. 475168651Skris This work has also resulted in a variety of "const"ifications of 475268651Skris lots of the code, especially "_cmp" operations which should normally 475368651Skris be prototyped with "const" parameters anyway. 475468651Skris [Geoff Thorpe] 475568651Skris 475668651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 475768651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 475868651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 475968651Skris where all of 'md' is used each time the PRNG is used, but 'state' 476068651Skris is used only indexed by a cyclic counter. As entropy may not be 476168651Skris well distributed from the beginning, 'md' is important as a 476268651Skris chaining variable. However, the output function chains only half 476368651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 476468651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 476568651Skris in all of 'state' being rewritten, with the new values depending 476668651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 476768651Skris [Bodo Moeller] 476868651Skris 476968651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 477068651Skris the handshake is continued after ssl_verify_cert_chain(); 477168651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 477268651Skris can lead to 'unexplainable' connection aborts later. 477368651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 477468651Skris 477568651Skris *) Major EVP API cipher revision. 477668651Skris Add hooks for extra EVP features. This allows various cipher 477768651Skris parameters to be set in the EVP interface. Support added for variable 477868651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 477968651Skris setting of RC2 and RC5 parameters. 478068651Skris 478168651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 478268651Skris ciphers. 478368651Skris 478468651Skris Remove lots of duplicated code from the EVP library. For example *every* 478568651Skris cipher init() function handles the 'iv' in the same way according to the 478668651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 478768651Skris for CFB and OFB modes they zero ctx->num. 478868651Skris 478968651Skris New functionality allows removal of S/MIME code RC2 hack. 479068651Skris 479168651Skris Most of the routines have the same form and so can be declared in terms 479268651Skris of macros. 479368651Skris 479468651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 479568651Skris all individual ciphers. If the cipher wants to handle IVs or keys 479668651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 479768651Skris flags. 479868651Skris 479968651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 480068651Skris value: although software versions of the algorithms cannot fail 480168651Skris any installed hardware versions can. 480268651Skris [Steve Henson] 480368651Skris 480468651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 480568651Skris this option is set, tolerate broken clients that send the negotiated 480668651Skris protocol version number instead of the requested protocol version 480768651Skris number. 480868651Skris [Bodo Moeller] 480968651Skris 481068651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 481168651Skris i.e. non-zero for export ciphersuites, zero otherwise. 481268651Skris Previous versions had this flag inverted, inconsistent with 481368651Skris rsa_tmp_cb (..._TMP_RSA_CB). 481468651Skris [Bodo Moeller; problem reported by Amit Chopra] 481568651Skris 481668651Skris *) Add missing DSA library text string. Work around for some IIS 481768651Skris key files with invalid SEQUENCE encoding. 481868651Skris [Steve Henson] 481968651Skris 482068651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 482168651Skris and so on that are implemented in OpenSSL. 482268651Skris [Richard Levitte] 482368651Skris 482468651Skris *) Enhance c_rehash script. Old version would mishandle certificates 482568651Skris with the same subject name hash and wouldn't handle CRLs at all. 482668651Skris Added -fingerprint option to crl utility, to support new c_rehash 482768651Skris features. 482868651Skris [Steve Henson] 482968651Skris 483068651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 483168651Skris [Ulf M�ller] 483268651Skris 483368651Skris *) Fix for SSL server purpose checking. Server checking was 483468651Skris rejecting certificates which had extended key usage present 483568651Skris but no ssl client purpose. 483668651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 483768651Skris 483868651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 483968651Skris is a little unclear about how a blank password is handled. 484068651Skris Since the password in encoded as a BMPString with terminating 484168651Skris double NULL a zero length password would end up as just the 484268651Skris double NULL. However no password at all is different and is 484368651Skris handled differently in the PKCS#12 key generation code. NS 484468651Skris treats a blank password as zero length. MSIE treats it as no 484568651Skris password on export: but it will try both on import. We now do 484668651Skris the same: PKCS12_parse() tries zero length and no password if 484768651Skris the password is set to "" or NULL (NULL is now a valid password: 484868651Skris it wasn't before) as does the pkcs12 application. 484968651Skris [Steve Henson] 485068651Skris 485168651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 485268651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 485368651Skris be obtained from the error queue. 485468651Skris [Bodo Moeller] 485568651Skris 485668651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 485768651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 485868651Skris accordingly to avoid race conditions (this is necessary because 485968651Skris thread_hash is no longer constant once set). 486068651Skris [Bodo Moeller] 486168651Skris 486268651Skris *) Bugfix for linux-elf makefile.one. 486368651Skris [Ulf M�ller] 486468651Skris 486568651Skris *) RSA_get_default_method() will now cause a default 486668651Skris RSA_METHOD to be chosen if one doesn't exist already. 486768651Skris Previously this was only set during a call to RSA_new() 486868651Skris or RSA_new_method(NULL) meaning it was possible for 486968651Skris RSA_get_default_method() to return NULL. 487068651Skris [Geoff Thorpe] 487168651Skris 487268651Skris *) Added native name translation to the existing DSO code 487368651Skris that will convert (if the flag to do so is set) filenames 487468651Skris that are sufficiently small and have no path information 487568651Skris into a canonical native form. Eg. "blah" converted to 487668651Skris "libblah.so" or "blah.dll" etc. 487768651Skris [Geoff Thorpe] 487868651Skris 487968651Skris *) New function ERR_error_string_n(e, buf, len) which is like 488068651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 488168651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 488268651Skris may not be NULL. 488368651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 488468651Skris 488568651Skris *) CONF library reworked to become more general. A new CONF 488668651Skris configuration file reader "class" is implemented as well as a 488768651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 488868651Skris old CONF_* functions are still there, but are reimplemented to 488968651Skris work in terms of the new functions. Also, a set of functions 489068651Skris to handle the internal storage of the configuration data is 489168651Skris provided to make it easier to write new configuration file 489268651Skris reader "classes" (I can definitely see something reading a 489368651Skris configuration file in XML format, for example), called _CONF_*, 489468651Skris or "the configuration storage API"... 489568651Skris 489668651Skris The new configuration file reading functions are: 489768651Skris 489868651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 489968651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 490068651Skris 490168651Skris NCONF_default, NCONF_WIN32 490268651Skris 490368651Skris NCONF_dump_fp, NCONF_dump_bio 490468651Skris 490568651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 490668651Skris NCONF_new creates a new CONF object. This works in the same way 490768651Skris as other interfaces in OpenSSL, like the BIO interface. 490868651Skris NCONF_dump_* dump the internal storage of the configuration file, 490968651Skris which is useful for debugging. All other functions take the same 491068651Skris arguments as the old CONF_* functions wth the exception of the 491168651Skris first that must be a `CONF *' instead of a `LHASH *'. 491268651Skris 491368651Skris To make it easer to use the new classes with the old CONF_* functions, 491468651Skris the function CONF_set_default_method is provided. 491568651Skris [Richard Levitte] 491668651Skris 491768651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 491868651Skris mentioned in the documentation but had not been implemented. 491968651Skris (This option is not yet really useful because even the additional 492068651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 492168651Skris [Bodo Moeller] 492268651Skris 492368651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 492468651Skris OpenSSL-based applications) load shared libraries and bind to 492568651Skris them in a portable way. 492668651Skris [Geoff Thorpe, with contributions from Richard Levitte] 492768651Skris 492859191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 492959191Skris 493059191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 493159191Skris 493259191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 493359191Skris (the default implementation of RAND_status). 493459191Skris 493559191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 493659191Skris to '-clrext' (= clear extensions), as intended and documented. 493759191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 493859191Skris <attili@amaxo.com>] 493959191Skris 494059191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 494159191Skris was larger than the MD block size. 494259191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 494359191Skris 494459191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 494559191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 494659191Skris using the passed key: if the passed key was a private key the result 494759191Skris of X509_print(), for example, would be to print out all the private key 494859191Skris components. 494959191Skris [Steve Henson] 495059191Skris 495159191Skris *) des_quad_cksum() byte order bug fix. 495259191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 495359191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 495459191Skris 495559191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 495659191Skris discouraged. 495759191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 495859191Skris 495959191Skris *) For easily testing in shell scripts whether some command 496059191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 496159191Skris returns with exit code 0 iff no command of the given name is available. 496259191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 496359191Skris the output goes to stdout and nothing is printed to stderr. 496459191Skris Additional arguments are always ignored. 496559191Skris 496659191Skris Since for each cipher there is a command of the same name, 496759191Skris the 'no-cipher' compilation switches can be tested this way. 496859191Skris 496959191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 497059191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 497159191Skris [Bodo Moeller] 497259191Skris 497359191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 497459191Skris [Bodo Moeller] 497559191Skris 497659191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 497759191Skris is set; it will be thrown away anyway because each handshake creates 497859191Skris its own key. 497959191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 498059191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 498159191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 498259191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 498359191Skris [Bodo Moeller] 498459191Skris 498559191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 498659191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 498759191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 498859191Skris does not suppress any output. 498959191Skris [Richard Levitte] 499059191Skris 499159191Skris *) Add compatibility options to the purpose and trust code. The 499259191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 499359191Skris accepts a certificate or CA, this was the previous behaviour, 499459191Skris with all the associated security issues. 499559191Skris 499659191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 499759191Skris automatically trust self signed roots in certificate store. A 499859191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 499959191Skris a purpose has no associated trust setting and it should instead 500059191Skris use the value in the default purpose. 500159191Skris [Steve Henson] 500259191Skris 500359191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 500459191Skris and fix a memory leak. 500559191Skris [Steve Henson] 500659191Skris 500759191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 500859191Skris reason strings from the previous version of the .c file, as 500959191Skris the default to have only downcase letters (and digits) in 501059191Skris automatically generated reasons codes is not always appropriate. 501159191Skris [Bodo Moeller] 501259191Skris 501359191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 501459191Skris using strerror. Previously, ERR_reason_error_string() returned 501559191Skris library names as reason strings for SYSerr; but SYSerr is a special 501659191Skris case where small numbers are errno values, not library numbers. 501759191Skris [Bodo Moeller] 501859191Skris 501959191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 502059191Skris converts DSA parameters into DH parameters. (When creating parameters, 502159191Skris DSA_generate_parameters is used.) 502259191Skris [Bodo Moeller] 502359191Skris 502459191Skris *) Include 'length' (recommended exponent length) in C code generated 502559191Skris by 'openssl dhparam -C'. 502659191Skris [Bodo Moeller] 502759191Skris 502859191Skris *) The second argument to set_label in perlasm was already being used 502959191Skris so couldn't be used as a "file scope" flag. Moved to third argument 503059191Skris which was free. 503159191Skris [Steve Henson] 503259191Skris 503359191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 503459191Skris instead of RAND_bytes for encryption IVs and salts. 503559191Skris [Bodo Moeller] 503659191Skris 503759191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 503859191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 503959191Skris RAND_set_rand_method would be impossible. 504059191Skris [Bodo Moeller] 504159191Skris 504259191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 504359191Skris number generation fails. 504459191Skris [Bodo Moeller] 504559191Skris 504659191Skris *) New 'rand' application for creating pseudo-random output. 504759191Skris [Bodo Moeller] 504859191Skris 504959191Skris *) Added configuration support for Linux/IA64 505059191Skris [Rolf Haberrecker <rolf@suse.de>] 505159191Skris 505259191Skris *) Assembler module support for Mingw32. 505359191Skris [Ulf M�ller] 505459191Skris 505559191Skris *) Shared library support for HPUX (in shlib/). 505659191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 505759191Skris 505859191Skris *) Shared library support for Solaris gcc. 505959191Skris [Lutz Behnke <behnke@trustcenter.de>] 506059191Skris 506159191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 506259191Skris 506359191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 506459191Skris were added manually and by SMIME_crlf_copy(). 506559191Skris [Steve Henson] 506659191Skris 506759191Skris *) In bntest.c don't call BN_rand with zero bits argument. 506859191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 506959191Skris 507059191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 507159191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 507259191Skris [Ulf M�ller] 507359191Skris 507459191Skris *) Add an optional second argument to the set_label() in the perl 507559191Skris assembly language builder. If this argument exists and is set 507659191Skris to 1 it signals that the assembler should use a symbol whose 507759191Skris scope is the entire file, not just the current function. This 507859191Skris is needed with MASM which uses the format label:: for this scope. 507959191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 508059191Skris 508159191Skris *) Change the ASN1 types so they are typedefs by default. Before 508259191Skris almost all types were #define'd to ASN1_STRING which was causing 508359191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 508459191Skris for example. 508559191Skris [Steve Henson] 508659191Skris 508759191Skris *) Change names of new functions to the new get1/get0 naming 508859191Skris convention: After 'get1', the caller owns a reference count 508959191Skris and has to call ..._free; 'get0' returns a pointer to some 509059191Skris data structure without incrementing reference counters. 509159191Skris (Some of the existing 'get' functions increment a reference 509259191Skris counter, some don't.) 509359191Skris Similarly, 'set1' and 'add1' functions increase reference 509459191Skris counters or duplicate objects. 509559191Skris [Steve Henson] 509659191Skris 509759191Skris *) Allow for the possibility of temp RSA key generation failure: 509859191Skris the code used to assume it always worked and crashed on failure. 509959191Skris [Steve Henson] 510059191Skris 510159191Skris *) Fix potential buffer overrun problem in BIO_printf(). 510259191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 510359191Skris pointed out by David Sacerdote <das33@cornell.edu>] 510459191Skris 510559191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 510659191Skris RAND_egd() and RAND_status(). In the command line application, 510759191Skris the EGD socket can be specified like a seed file using RANDFILE 510859191Skris or -rand. 510959191Skris [Ulf M�ller] 511059191Skris 511159191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 511259191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 511359191Skris [Steve Henson] 511459191Skris 511559191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 511659191Skris list to exclude them. This means that no special compilation option 511759191Skris is needed to use anonymous DH: it just needs to be included in the 511859191Skris cipher list. 511959191Skris [Steve Henson] 512059191Skris 512159191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 512259191Skris EVP_MD_type. The old functionality is available in a new macro called 512359191Skris EVP_MD_md(). Change code that uses it and update docs. 512459191Skris [Steve Henson] 512559191Skris 512659191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 512759191Skris where the 'void *' argument is replaced by a function pointer argument. 512859191Skris Previously 'void *' was abused to point to functions, which works on 512959191Skris many platforms, but is not correct. As these functions are usually 513059191Skris called by macros defined in OpenSSL header files, most source code 513159191Skris should work without changes. 513259191Skris [Richard Levitte] 513359191Skris 513459191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 513559191Skris sections with information on -D... compiler switches used for 513659191Skris compiling the library so that applications can see them. To enable 513759191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 513859191Skris must be defined. E.g., 513959191Skris #define OPENSSL_ALGORITHM_DEFINES 514059191Skris #include <openssl/opensslconf.h> 514159191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 514259191Skris [Richard Levitte, Ulf and Bodo M�ller] 514359191Skris 514459191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 514559191Skris record layer. 514659191Skris [Bodo Moeller] 514759191Skris 514859191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 514959191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 515059191Skris the required ASN1 format: arbitrary types determined by an OID. 515159191Skris [Steve Henson] 515259191Skris 515359191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 515459191Skris argument to 'req'. This is not because the function is newer or 515559191Skris better than others it just uses the work 'NEW' in the certificate 515659191Skris request header lines. Some software needs this. 515759191Skris [Steve Henson] 515859191Skris 515959191Skris *) Reorganise password command line arguments: now passwords can be 516059191Skris obtained from various sources. Delete the PEM_cb function and make 516159191Skris it the default behaviour: i.e. if the callback is NULL and the 516259191Skris usrdata argument is not NULL interpret it as a null terminated pass 516359191Skris phrase. If usrdata and the callback are NULL then the pass phrase 516459191Skris is prompted for as usual. 516559191Skris [Steve Henson] 516659191Skris 516759191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 516859191Skris the support is automatically enabled. The resulting binaries will 516959191Skris autodetect the card and use it if present. 517059191Skris [Ben Laurie and Compaq Inc.] 517159191Skris 517259191Skris *) Work around for Netscape hang bug. This sends certificate request 517359191Skris and server done in one record. Since this is perfectly legal in the 517459191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 517559191Skris the bugs/SSLv3 entry for more info. 517659191Skris [Steve Henson] 517759191Skris 517859191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 517959191Skris [Andy Polyakov] 518059191Skris 518159191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 518259191Skris of seed file. 518359191Skris [Steve Henson] 518459191Skris 518559191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 518659191Skris [Bodo Moeller] 518759191Skris 518859191Skris *) Add command line password options to the remaining applications. 518959191Skris [Steve Henson] 519059191Skris 519159191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 519259191Skris bits. 519359191Skris [Ulf M�ller] 519459191Skris 519559191Skris *) More tests in bntest.c, and changed test_bn output. 519659191Skris [Ulf M�ller] 519759191Skris 519859191Skris *) ./config recognizes MacOS X now. 519959191Skris [Andy Polyakov] 520059191Skris 520159191Skris *) Bug fix for BN_div() when the first words of num and divsor are 520259191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 520359191Skris [Ulf M�ller] 520459191Skris 520559191Skris *) Add support for various broken PKCS#8 formats, and command line 520659191Skris options to produce them. 520759191Skris [Steve Henson] 520859191Skris 520959191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 521059191Skris get temporary BIGNUMs from a BN_CTX. 521159191Skris [Ulf M�ller] 521259191Skris 521359191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 521459191Skris for p == 0. 521559191Skris [Ulf M�ller] 521659191Skris 521759191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 521859191Skris include a #define from the old name to the new. The original intent 521959191Skris was that statically linked binaries could for example just call 522059191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 522159191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 522259191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 522359191Skris one would link with the other. They are now in separate source files. 522459191Skris [Steve Henson] 522559191Skris 522659191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 522759191Skris [Steve Henson] 522859191Skris 522959191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 523059191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 523159191Skris loop, our standard modexp algorithms are faster). 523259191Skris [Bodo Moeller] 523359191Skris 523459191Skris *) Support for the EBCDIC character set completed. 523559191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 523659191Skris 523759191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 523859191Skris use void * instead of char * in lhash. 523959191Skris [Ulf M�ller] 524059191Skris 524159191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 524259191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 524359191Skris this the server could overwrite ephemeral keys that the client 524459191Skris has already seen). 524559191Skris [Bodo Moeller] 524659191Skris 524759191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 524859191Skris using 50 iterations of the Rabin-Miller test. 524959191Skris 525059191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 525159191Skris iterations of the Rabin-Miller test as required by the appendix 525259191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 525359191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 525459191Skris generation becomes much faster. 525559191Skris 525659191Skris This implies a change for the callback functions in DSA_is_prime 525759191Skris and DSA_generate_parameters: The callback function is called once 525859191Skris for each positive witness in the Rabin-Miller test, not just 525959191Skris occasionally in the inner loop; and the parameters to the 526059191Skris callback function now provide an iteration count for the outer 526159191Skris loop rather than for the current invocation of the inner loop. 526259191Skris DSA_generate_parameters additionally can call the callback 526359191Skris function with an 'iteration count' of -1, meaning that a 526459191Skris candidate has passed the trial division test (when q is generated 526559191Skris from an application-provided seed, trial division is skipped). 526659191Skris [Bodo Moeller] 526759191Skris 526859191Skris *) New function BN_is_prime_fasttest that optionally does trial 526959191Skris division before starting the Rabin-Miller test and has 527059191Skris an additional BN_CTX * argument (whereas BN_is_prime always 527159191Skris has to allocate at least one BN_CTX). 527259191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 527359191Skris trial division stage. 527459191Skris [Bodo Moeller] 527559191Skris 527659191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 527759191Skris as ASN1_TIME. 527859191Skris [Steve Henson] 527959191Skris 528059191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 528159191Skris [Steve Henson] 528259191Skris 528359191Skris *) New function BN_pseudo_rand(). 528459191Skris [Ulf M�ller] 528559191Skris 528659191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 528759191Skris bignum version of BN_from_montgomery() with the working code from 528859191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 528959191Skris the comments. 529059191Skris [Ulf M�ller] 529159191Skris 529259191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 529359191Skris made it impossible to use the same SSL_SESSION data structure in 529459191Skris SSL2 clients in multiple threads. 529559191Skris [Bodo Moeller] 529659191Skris 529759191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 529859191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 529959191Skris to seed the PRNG (previously an explicit byte count was required). 530059191Skris [Ulf M�ller, Bodo M�ller] 530159191Skris 530259191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 530359191Skris used (char *) instead of (void *) and had casts all over the place. 530459191Skris [Steve Henson] 530559191Skris 530659191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 530759191Skris [Ulf M�ller] 530859191Skris 530959191Skris *) Retain source code compatibility for BN_prime_checks macro: 531059191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 531159191Skris BN_prime_checks_for_size to determine the appropriate number of 531259191Skris Rabin-Miller iterations. 531359191Skris [Ulf M�ller] 531459191Skris 531559191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 531659191Skris DH_CHECK_P_NOT_SAFE_PRIME. 531759191Skris (Check if this is true? OpenPGP calls them "strong".) 531859191Skris [Ulf M�ller] 531959191Skris 532059191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 532159191Skris "dhparam". The old programs are retained for now but will handle DH keys 532259191Skris (instead of parameters) in future. 532359191Skris [Steve Henson] 532459191Skris 532559191Skris *) Make the ciphers, s_server and s_client programs check the return values 532659191Skris when a new cipher list is set. 532759191Skris [Steve Henson] 532859191Skris 532959191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 533059191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 533159191Skris wrong. 533259191Skris 533359191Skris The syntax for the cipher sorting has been extended to support sorting by 533459191Skris cipher-strength (using the strength_bits hard coded in the tables). 533559191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 533659191Skris 533759191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 533859191Skris string with an "undefined" symbol (neither command nor alphanumeric 533959191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 534059191Skris an error is flagged. 534159191Skris 534259191Skris Due to the strength-sorting extension, the code of the 534359191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 534459191Skris the readability was also increased :-) 534559191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 534659191Skris 534759191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 534859191Skris for the first serial number and places 2 in the serial number file. This 534959191Skris avoids problems when the root CA is created with serial number zero and 535059191Skris the first user certificate has the same issuer name and serial number 535159191Skris as the root CA. 535259191Skris [Steve Henson] 535359191Skris 535459191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 535559191Skris the new code. Add documentation for this stuff. 535659191Skris [Steve Henson] 535759191Skris 535859191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 535959191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 536059191Skris structures and behave in an analagous way to the X509v3 functions: 536159191Skris they shouldn't be called directly but wrapper functions should be used 536259191Skris instead. 536359191Skris 536459191Skris So we also now have some wrapper functions that call the X509at functions 536559191Skris when passed certificate requests. (TO DO: similar things can be done with 536659191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 536759191Skris things. Some of these need some d2i or i2d and print functionality 536859191Skris because they handle more complex structures.) 536959191Skris [Steve Henson] 537059191Skris 537159191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 537259191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 537359191Skris NO_RSA in ssl/s2*.c. 537459191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 537559191Skris 537659191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 537759191Skris has a return value which indicates the quality of the random data 537859191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 537959191Skris error queue. New function RAND_pseudo_bytes() generates output that is 538059191Skris guaranteed to be unique but not unpredictable. RAND_add is like 538159191Skris RAND_seed, but takes an extra argument for an entropy estimate 538259191Skris (RAND_seed always assumes full entropy). 538359191Skris [Ulf M�ller] 538459191Skris 538559191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 538659191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 538759191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 538859191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 538959191Skris false-positive rate of at most 2^-80 for random input. 539059191Skris [Bodo Moeller] 539159191Skris 539259191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 539359191Skris [Bodo Moeller] 539459191Skris 539559191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 539659191Skris in the 0.9.5 release), this returns the chain 539759191Skris from an X509_CTX structure with a dup of the stack and all 539859191Skris the X509 reference counts upped: so the stack will exist 539959191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 540059191Skris to use this. 540159191Skris 540259191Skris Also make SSL_SESSION_print() print out the verify return 540359191Skris code. 540459191Skris [Steve Henson] 540559191Skris 540659191Skris *) Add manpage for the pkcs12 command. Also change the default 540759191Skris behaviour so MAC iteration counts are used unless the new 540859191Skris -nomaciter option is used. This improves file security and 540959191Skris only older versions of MSIE (4.0 for example) need it. 541059191Skris [Steve Henson] 541159191Skris 541259191Skris *) Honor the no-xxx Configure options when creating .DEF files. 541359191Skris [Ulf M�ller] 541459191Skris 541559191Skris *) Add PKCS#10 attributes to field table: challengePassword, 541659191Skris unstructuredName and unstructuredAddress. These are taken from 541759191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 541859191Skris international characters are used. 541959191Skris 542059191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 542159191Skris based on strings. Remove the 'loc' parameter when adding 542259191Skris attributes because these will be a SET OF encoding which is sorted 542359191Skris in ASN1 order. 542459191Skris [Steve Henson] 542559191Skris 542659191Skris *) Initial changes to the 'req' utility to allow request generation 542759191Skris automation. This will allow an application to just generate a template 542859191Skris file containing all the field values and have req construct the 542959191Skris request. 543059191Skris 543159191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 543259191Skris used all over the place including certificate requests and PKCS#7 543359191Skris structures. They are currently handled manually where necessary with 543459191Skris some primitive wrappers for PKCS#7. The new functions behave in a 543559191Skris manner analogous to the X509 extension functions: they allow 543659191Skris attributes to be looked up by NID and added. 543759191Skris 543859191Skris Later something similar to the X509V3 code would be desirable to 543959191Skris automatically handle the encoding, decoding and printing of the 544059191Skris more complex types. The string types like challengePassword can 544159191Skris be handled by the string table functions. 544259191Skris 544359191Skris Also modified the multi byte string table handling. Now there is 544459191Skris a 'global mask' which masks out certain types. The table itself 544559191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 544659191Skris is useful when for example there is only one permissible type 544759191Skris (as in countryName) and using the mask might result in no valid 544859191Skris types at all. 544959191Skris [Steve Henson] 545059191Skris 545159191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 545259191Skris SSL_get_peer_finished to allow applications to obtain the latest 545359191Skris Finished messages sent to the peer or expected from the peer, 545459191Skris respectively. (SSL_get_peer_finished is usually the Finished message 545559191Skris actually received from the peer, otherwise the protocol will be aborted.) 545659191Skris 545759191Skris As the Finished message are message digests of the complete handshake 545859191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 545959191Skris be used for external authentication procedures when the authentication 546059191Skris provided by SSL/TLS is not desired or is not enough. 546159191Skris [Bodo Moeller] 546259191Skris 546359191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 546459191Skris the host supports BWX extension and if Compaq C is present on the 546559191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 546659191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 546759191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 546859191Skris SHA1. 546959191Skris [Andy Polyakov] 547059191Skris 547159191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 547259191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 547359191Skris weak crypto and after checking the certificate is SGC a second one 547459191Skris with strong crypto. MS SGC stops the first handshake after receiving 547559191Skris the server certificate message and sends a second client hello. Since 547659191Skris a server will typically do all the time consuming operations before 547759191Skris expecting any further messages from the client (server key exchange 547859191Skris is the most expensive) there is little difference between the two. 547959191Skris 548059191Skris To get OpenSSL to support MS SGC we have to permit a second client 548159191Skris hello message after we have sent server done. In addition we have to 548259191Skris reset the MAC if we do get this second client hello. 548359191Skris [Steve Henson] 548459191Skris 548559191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 548659191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 548759191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 548859191Skris format DER encoded private key. Newer code should use PKCS#8 format which 548959191Skris has the key type encoded in the ASN1 structure. Added DER private key 549059191Skris support to pkcs8 application. 549159191Skris [Steve Henson] 549259191Skris 549359191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 549459191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 549559191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 549659191Skris is set, we interpret this as a request to violate the specification 549759191Skris (the worst that can happen is a handshake failure, and 'correct' 549859191Skris behaviour would result in a handshake failure anyway). 549959191Skris [Bodo Moeller] 550059191Skris 550159191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 550259191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 550359191Skris concurrently obtain them from an external cache). 550459191Skris The internal cache can handle only one SSL_SESSION with a given ID, 550559191Skris so if there's a conflict, we now throw out the old one to achieve 550659191Skris consistency. 550759191Skris [Bodo Moeller] 550859191Skris 550959191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 551059191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 551159191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 551259191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 551359191Skris example. 551459191Skris [Steve Henson] 551559191Skris 551659191Skris *) Simplify the trust setting structure and code. Now we just have 551759191Skris two sequences of OIDs for trusted and rejected settings. These will 551859191Skris typically have values the same as the extended key usage extension 551959191Skris and any application specific purposes. 552059191Skris 552159191Skris The trust checking code now has a default behaviour: it will just 552259191Skris check for an object with the same NID as the passed id. Functions can 552359191Skris be provided to override either the default behaviour or the behaviour 552459191Skris for a given id. SSL client, server and email already have functions 552559191Skris in place for compatibility: they check the NID and also return "trusted" 552659191Skris if the certificate is self signed. 552759191Skris [Steve Henson] 552859191Skris 552959191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 553059191Skris traditional format into an EVP_PKEY structure. 553159191Skris [Steve Henson] 553259191Skris 553359191Skris *) Add a password callback function PEM_cb() which either prompts for 553459191Skris a password if usr_data is NULL or otherwise assumes it is a null 553559191Skris terminated password. Allow passwords to be passed on command line 553659191Skris environment or config files in a few more utilities. 553759191Skris [Steve Henson] 553859191Skris 553959191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 554059191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 554159191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 554259191Skris Update documentation. 554359191Skris [Steve Henson] 554459191Skris 554559191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 554659191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 554759191Skris and produce an error if it couldn't. For compatibility we also have 554859191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 554959191Skris don't allocate anything because they don't need to. 555059191Skris [Steve Henson] 555159191Skris 555259191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 555359191Skris for details. 555459191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 555559191Skris 555659191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 555759191Skris possibly others as well. The purpose is to make an interface that 555859191Skris provide hooks so anyone can build a separate set of allocation and 555959191Skris deallocation routines to be used by OpenSSL, for example memory 556059191Skris pool implementations, or something else, which was previously hard 556159191Skris since Malloc(), Realloc() and Free() were defined as macros having 556259191Skris the values malloc, realloc and free, respectively (except for Win32 556359191Skris compilations). The same is provided for memory debugging code. 556459191Skris OpenSSL already comes with functionality to find memory leaks, but 556559191Skris this gives people a chance to debug other memory problems. 556659191Skris 556759191Skris With these changes, a new set of functions and macros have appeared: 556859191Skris 556968651Skris CRYPTO_set_mem_debug_functions() [F] 557068651Skris CRYPTO_get_mem_debug_functions() [F] 557168651Skris CRYPTO_dbg_set_options() [F] 557268651Skris CRYPTO_dbg_get_options() [F] 557368651Skris CRYPTO_malloc_debug_init() [M] 557459191Skris 557559191Skris The memory debug functions are NULL by default, unless the library 557659191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 557759191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 557859191Skris gives the standard debugging functions that come with OpenSSL) or 557959191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 558059191Skris provided by the library user) must be used. When the standard 558159191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 558259191Skris request additional information: 558359191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 558459191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 558559191Skris 558659191Skris Also, things like CRYPTO_set_mem_functions will always give the 558759191Skris expected result (the new set of functions is used for allocation 558859191Skris and deallocation) at all times, regardless of platform and compiler 558959191Skris options. 559059191Skris 559159191Skris To finish it up, some functions that were never use in any other 559259191Skris way than through macros have a new API and new semantic: 559359191Skris 559459191Skris CRYPTO_dbg_malloc() 559559191Skris CRYPTO_dbg_realloc() 559659191Skris CRYPTO_dbg_free() 559759191Skris 559859191Skris All macros of value have retained their old syntax. 559959191Skris [Richard Levitte and Bodo Moeller] 560059191Skris 560159191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 560259191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 560359191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 560459191Skris algorithm. 560559191Skris [Steve Henson] 560659191Skris 560759191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 560859191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 560959191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 561059191Skris 561159191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 561259191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 561359191Skris functionality to handle multipart/signed properly) and a utility 561459191Skris called 'smime' to call all this stuff. This is based on code I 561559191Skris originally wrote for Celo who have kindly allowed it to be 561659191Skris included in OpenSSL. 561759191Skris [Steve Henson] 561859191Skris 561959191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 562059191Skris des_set_key (aka des_key_sched). Global variable des_check_key 562159191Skris decides which of these is called by des_set_key; this way 562259191Skris des_check_key behaves as it always did, but applications and 562359191Skris the library itself, which was buggy for des_check_key == 1, 562459191Skris have a cleaner way to pick the version they need. 562559191Skris [Bodo Moeller] 562659191Skris 562759191Skris *) New function PKCS12_newpass() which changes the password of a 562859191Skris PKCS12 structure. 562959191Skris [Steve Henson] 563059191Skris 563159191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 563259191Skris dynamic mix. In both cases the ids can be used as an index into the 563359191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 563459191Skris functions so they accept a list of the field values and the 563559191Skris application doesn't need to directly manipulate the X509_TRUST 563659191Skris structure. 563759191Skris [Steve Henson] 563859191Skris 563959191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 564059191Skris need initialising. 564159191Skris [Steve Henson] 564259191Skris 564359191Skris *) Modify the way the V3 extension code looks up extensions. This now 564459191Skris works in a similar way to the object code: we have some "standard" 564559191Skris extensions in a static table which is searched with OBJ_bsearch() 564659191Skris and the application can add dynamic ones if needed. The file 564759191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 564859191Skris updated whenever a new extension is added to the core code and kept 564959191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 565059191Skris this. New extensions are not added too often so this file can readily 565159191Skris be maintained manually. 565259191Skris 565359191Skris There are two big advantages in doing things this way. The extensions 565459191Skris can be looked up immediately and no longer need to be "added" using 565559191Skris X509V3_add_standard_extensions(): this function now does nothing. 565659191Skris [Side note: I get *lots* of email saying the extension code doesn't 565759191Skris work because people forget to call this function] 565859191Skris Also no dynamic allocation is done unless new extensions are added: 565959191Skris so if we don't add custom extensions there is no need to call 566059191Skris X509V3_EXT_cleanup(). 566159191Skris [Steve Henson] 566259191Skris 566359191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 566459191Skris magic header, so unsalted files fail gracefully instead of just decrypting 566559191Skris to garbage. This is because not salting is a big security hole, so people 566659191Skris should be discouraged from doing it. 566759191Skris [Ben Laurie] 566859191Skris 566959191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 567059191Skris digest to be passed on the command line but it only used this 567159191Skris parameter when signing a certificate. Modified so all relevant 567259191Skris operations are affected by the digest parameter including the 567359191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 567459191Skris DSA key was used because it didn't fix the digest. 567559191Skris [Steve Henson] 567659191Skris 567759191Skris *) Initial certificate chain verify code. Currently tests the untrusted 567859191Skris certificates for consistency with the verify purpose (which is set 567959191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 568059191Skris 568159191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 568259191Skris this is because it will reject chains with invalid extensions whereas 568359191Skris every previous version of OpenSSL and SSLeay made no checks at all. 568459191Skris 568559191Skris Trust code: checks the root CA for the relevant trust settings. Trust 568659191Skris settings have an initial value consistent with the verify purpose: e.g. 568759191Skris if the verify purpose is for SSL client use it expects the CA to be 568859191Skris trusted for SSL client use. However the default value can be changed to 568959191Skris permit custom trust settings: one example of this would be to only trust 569059191Skris certificates from a specific "secure" set of CAs. 569159191Skris 569259191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 569359191Skris which should be used for version portability: especially since the 569459191Skris verify structure is likely to change more often now. 569559191Skris 569659191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 569759191Skris to set them. If not set then assume SSL clients will verify SSL servers 569859191Skris and vice versa. 569959191Skris 570059191Skris Two new options to the verify program: -untrusted allows a set of 570159191Skris untrusted certificates to be passed in and -purpose which sets the 570259191Skris intended purpose of the certificate. If a purpose is set then the 570359191Skris new chain verify code is used to check extension consistency. 570459191Skris [Steve Henson] 570559191Skris 570659191Skris *) Support for the authority information access extension. 570759191Skris [Steve Henson] 570859191Skris 570959191Skris *) Modify RSA and DSA PEM read routines to transparently handle 571059191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 571159191Skris public keys in a format compatible with certificate 571259191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 571359191Skris functions called *_PublicKey_* which used various odd formats so 571459191Skris these are retained for compatibility: however the DSA variants were 571559191Skris never in a public release so they have been deleted. Changed dsa/rsa 571659191Skris utilities to handle the new format: note no releases ever handled public 571759191Skris keys so we should be OK. 571859191Skris 571959191Skris The primary motivation for this change is to avoid the same fiasco 572059191Skris that dogs private keys: there are several incompatible private key 572159191Skris formats some of which are standard and some OpenSSL specific and 572259191Skris require various evil hacks to allow partial transparent handling and 572359191Skris even then it doesn't work with DER formats. Given the option anything 572459191Skris other than PKCS#8 should be dumped: but the other formats have to 572559191Skris stay in the name of compatibility. 572659191Skris 572759191Skris With public keys and the benefit of hindsight one standard format 572859191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 572959191Skris it clearly returns an error if you try to read the wrong kind of key. 573059191Skris 573159191Skris Added a -pubkey option to the 'x509' utility to output the public key. 573259191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 573359191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 573459191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 573559191Skris that do the same as the EVP_PKEY_assign_*() except they up the 573659191Skris reference count of the added key (they don't "swallow" the 573759191Skris supplied key). 573859191Skris [Steve Henson] 573959191Skris 574059191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 574159191Skris CRLs would fail if the file contained no certificates or no CRLs: 574259191Skris added a new function to read in both types and return the number 574359191Skris read: this means that if none are read it will be an error. The 574459191Skris DER versions of the certificate and CRL reader would always fail 574559191Skris because it isn't possible to mix certificates and CRLs in DER format 574659191Skris without choking one or the other routine. Changed this to just read 574759191Skris a certificate: this is the best we can do. Also modified the code 574859191Skris in apps/verify.c to take notice of return codes: it was previously 574959191Skris attempting to read in certificates from NULL pointers and ignoring 575059191Skris any errors: this is one reason why the cert and CRL reader seemed 575159191Skris to work. It doesn't check return codes from the default certificate 575259191Skris routines: these may well fail if the certificates aren't installed. 575359191Skris [Steve Henson] 575459191Skris 575559191Skris *) Code to support otherName option in GeneralName. 575659191Skris [Steve Henson] 575759191Skris 575859191Skris *) First update to verify code. Change the verify utility 575959191Skris so it warns if it is passed a self signed certificate: 576059191Skris for consistency with the normal behaviour. X509_verify 576159191Skris has been modified to it will now verify a self signed 576259191Skris certificate if *exactly* the same certificate appears 576359191Skris in the store: it was previously impossible to trust a 576459191Skris single self signed certificate. This means that: 576559191Skris openssl verify ss.pem 576659191Skris now gives a warning about a self signed certificate but 576759191Skris openssl verify -CAfile ss.pem ss.pem 576859191Skris is OK. 576959191Skris [Steve Henson] 577059191Skris 577159191Skris *) For servers, store verify_result in SSL_SESSION data structure 577259191Skris (and add it to external session representation). 577359191Skris This is needed when client certificate verifications fails, 577459191Skris but an application-provided verification callback (set by 577559191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 577659191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 577759191Skris but returns 1): When the session is reused, we have to set 577859191Skris ssl->verify_result to the appropriate error code to avoid 577959191Skris security holes. 578059191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 578159191Skris 578259191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 578359191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 578459191Skris didn't contain any existing data because it was being created. 578559191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 578659191Skris 578759191Skris *) Add a salt to the key derivation routines in enc.c. This 578859191Skris forms the first 8 bytes of the encrypted file. Also add a 578959191Skris -S option to allow a salt to be input on the command line. 579059191Skris [Steve Henson] 579159191Skris 579259191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 579359191Skris to compare two certificates. We do this by working out the SHA1 579459191Skris hash and comparing that. X509_cmp() will be needed by the trust 579559191Skris code. 579659191Skris [Steve Henson] 579759191Skris 579859191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 579959191Skris the reference count in the SSL_SESSION returned. 580059191Skris [Geoff Thorpe <geoff@eu.c2.net>] 580159191Skris 580259191Skris *) Fix for 'req': it was adding a null to request attributes. 580359191Skris Also change the X509_LOOKUP and X509_INFO code to handle 580459191Skris certificate auxiliary information. 580559191Skris [Steve Henson] 580659191Skris 580759191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 580859191Skris the 'enc' command. 580959191Skris [Steve Henson] 581059191Skris 581159191Skris *) Add the possibility to add extra information to the memory leak 581259191Skris detecting output, to form tracebacks, showing from where each 581359191Skris allocation was originated: CRYPTO_push_info("constant string") adds 581459191Skris the string plus current file name and line number to a per-thread 581559191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 581659191Skris is like calling CYRPTO_pop_info() until the stack is empty. 581759191Skris Also updated memory leak detection code to be multi-thread-safe. 581859191Skris [Richard Levitte] 581959191Skris 582059191Skris *) Add options -text and -noout to pkcs7 utility and delete the 582159191Skris encryption options which never did anything. Update docs. 582259191Skris [Steve Henson] 582359191Skris 582459191Skris *) Add options to some of the utilities to allow the pass phrase 582559191Skris to be included on either the command line (not recommended on 582659191Skris OSes like Unix) or read from the environment. Update the 582759191Skris manpages and fix a few bugs. 582859191Skris [Steve Henson] 582959191Skris 583059191Skris *) Add a few manpages for some of the openssl commands. 583159191Skris [Steve Henson] 583259191Skris 583359191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 583459191Skris leaking and not finding already revoked certificates. 583559191Skris [Steve Henson] 583659191Skris 583759191Skris *) Extensive changes to support certificate auxiliary information. 583859191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 583959191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 584059191Skris can still read in a certificate file in the usual way but it 584159191Skris will also read in any additional "auxiliary information". By 584259191Skris doing things this way a fair degree of compatibility can be 584359191Skris retained: existing certificates can have this information added 584459191Skris using the new 'x509' options. 584559191Skris 584659191Skris Current auxiliary information includes an "alias" and some trust 584759191Skris settings. The trust settings will ultimately be used in enhanced 584859191Skris certificate chain verification routines: currently a certificate 584959191Skris can only be trusted if it is self signed and then it is trusted 585059191Skris for all purposes. 585159191Skris [Steve Henson] 585259191Skris 585359191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 585459191Skris The problem was that one of the replacement routines had not been working 585559191Skris since SSLeay releases. For now the offending routine has been replaced 585659191Skris with non-optimised assembler. Even so, this now gives around 95% 585759191Skris performance improvement for 1024 bit RSA signs. 585859191Skris [Mark Cox] 585959191Skris 586059191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 586159191Skris handling. Most clients have the effective key size in bits equal to 586259191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 586359191Skris A few however don't do this and instead use the size of the decrypted key 586459191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 586559191Skris the effective key length. In this case the effective key length can still 586659191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 586759191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 586859191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 586959191Skris the key length and effective key length are equal. 587059191Skris [Steve Henson] 587159191Skris 587259191Skris *) Add a bunch of functions that should simplify the creation of 587359191Skris X509_NAME structures. Now you should be able to do: 587459191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 587559191Skris and have it automatically work out the correct field type and fill in 587659191Skris the structures. The more adventurous can try: 587759191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 587859191Skris and it will (hopefully) work out the correct multibyte encoding. 587959191Skris [Steve Henson] 588059191Skris 588159191Skris *) Change the 'req' utility to use the new field handling and multibyte 588259191Skris copy routines. Before the DN field creation was handled in an ad hoc 588359191Skris way in req, ca, and x509 which was rather broken and didn't support 588459191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 588559191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 588659191Skris using the dirstring_type option. See the new comment in the default 588759191Skris openssl.cnf for more info. 588859191Skris [Steve Henson] 588959191Skris 589059191Skris *) Make crypto/rand/md_rand.c more robust: 589159191Skris - Assure unique random numbers after fork(). 589259191Skris - Make sure that concurrent threads access the global counter and 589359191Skris md serializably so that we never lose entropy in them 589459191Skris or use exactly the same state in multiple threads. 589559191Skris Access to the large state is not always serializable because 589659191Skris the additional locking could be a performance killer, and 589759191Skris md should be large enough anyway. 589859191Skris [Bodo Moeller] 589959191Skris 590059191Skris *) New file apps/app_rand.c with commonly needed functionality 590159191Skris for handling the random seed file. 590259191Skris 590359191Skris Use the random seed file in some applications that previously did not: 590459191Skris ca, 590559191Skris dsaparam -genkey (which also ignored its '-rand' option), 590659191Skris s_client, 590759191Skris s_server, 590859191Skris x509 (when signing). 590959191Skris Except on systems with /dev/urandom, it is crucial to have a random 591059191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 591159191Skris for RSA signatures we could do without one. 591259191Skris 591359191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 591459191Skris of each file listed in the '-rand' option. The function as previously 591559191Skris found in genrsa is now in app_rand.c and is used by all programs 591659191Skris that support '-rand'. 591759191Skris [Bodo Moeller] 591859191Skris 591959191Skris *) In RAND_write_file, use mode 0600 for creating files; 592059191Skris don't just chmod when it may be too late. 592159191Skris [Bodo Moeller] 592259191Skris 592359191Skris *) Report an error from X509_STORE_load_locations 592459191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 592559191Skris [Bill Perry] 592659191Skris 592759191Skris *) New function ASN1_mbstring_copy() this copies a string in either 592859191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 592959191Skris into an ASN1_STRING type. A mask of permissible types is passed 593059191Skris and it chooses the "minimal" type to use or an error if not type 593159191Skris is suitable. 593259191Skris [Steve Henson] 593359191Skris 593459191Skris *) Add function equivalents to the various macros in asn1.h. The old 593559191Skris macros are retained with an M_ prefix. Code inside the library can 593659191Skris use the M_ macros. External code (including the openssl utility) 593759191Skris should *NOT* in order to be "shared library friendly". 593859191Skris [Steve Henson] 593959191Skris 594059191Skris *) Add various functions that can check a certificate's extensions 594159191Skris to see if it usable for various purposes such as SSL client, 594259191Skris server or S/MIME and CAs of these types. This is currently 594359191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 594459191Skris verification. Also added a -purpose flag to x509 utility to 594559191Skris print out all the purposes. 594659191Skris [Steve Henson] 594759191Skris 594859191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 594959191Skris functions. 595059191Skris [Steve Henson] 595159191Skris 595259191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 595359191Skris for, obtain and decode and extension and obtain its critical flag. 595459191Skris This allows all the necessary extension code to be handled in a 595559191Skris single function call. 595659191Skris [Steve Henson] 595759191Skris 595859191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 595959191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 596059191Skris [Andy Polyakov] 596159191Skris 596259191Skris *) New -noout option to asn1parse. This causes no output to be produced 596359191Skris its main use is when combined with -strparse and -out to extract data 596459191Skris from a file (which may not be in ASN.1 format). 596559191Skris [Steve Henson] 596659191Skris 596759191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 596859191Skris when producing the local key id. 596959191Skris [Richard Levitte <levitte@stacken.kth.se>] 597059191Skris 597159191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 597259191Skris stated explicitly. If it is not stated then it tries the first server 597359191Skris certificate file. The previous behaviour hard coded the filename 597459191Skris "server.pem". 597559191Skris [Steve Henson] 597659191Skris 597759191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 597859191Skris a public key to be input or output. For example: 597959191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 598059191Skris Also added necessary DSA public key functions to handle this. 598159191Skris [Steve Henson] 598259191Skris 598359191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 598459191Skris in the message. This was handled by allowing 598559191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 598659191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 598759191Skris 598859191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 598959191Skris to the end of the strings whereas this didn't. This would cause problems 599059191Skris if strings read with d2i_ASN1_bytes() were later modified. 599159191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 599259191Skris 599359191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 599459191Skris data and it contains EOF it will end up returning an error. This is 599559191Skris caused by input 46 bytes long. The cause is due to the way base64 599659191Skris BIOs find the start of base64 encoded data. They do this by trying a 599759191Skris trial decode on each line until they find one that works. When they 599859191Skris do a flag is set and it starts again knowing it can pass all the 599959191Skris data directly through the decoder. Unfortunately it doesn't reset 600059191Skris the context it uses. This means that if EOF is reached an attempt 600159191Skris is made to pass two EOFs through the context and this causes the 600259191Skris resulting error. This can also cause other problems as well. As is 600359191Skris usual with these problems it takes *ages* to find and the fix is 600459191Skris trivial: move one line. 600559191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 600659191Skris 600759191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 600859191Skris old code wouldn't work because it needed to select() on sockets and the 600959191Skris tty (for keypresses and to see if data could be written). Win32 only 601059191Skris supports select() on sockets so we select() with a 1s timeout on the 601159191Skris sockets and then see if any characters are waiting to be read, if none 601259191Skris are present then we retry, we also assume we can always write data to 601359191Skris the tty. This isn't nice because the code then blocks until we've 601459191Skris received a complete line of data and it is effectively polling the 601559191Skris keyboard at 1s intervals: however it's quite a bit better than not 601659191Skris working at all :-) A dedicated Windows application might handle this 601759191Skris with an event loop for example. 601859191Skris [Steve Henson] 601959191Skris 602059191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 602159191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 602259191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 602359191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 602459191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 602559191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 602659191Skris This necessitated the support of an extra signature type NID_md5_sha1 602759191Skris for SSL signatures and modifications to the SSL library to use it instead 602859191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 602959191Skris [Steve Henson] 603059191Skris 603159191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 603259191Skris will lookup a CRL issuers certificate and verify the signature in a 603359191Skris similar way to the verify program. Tidy up the crl program so it 603459191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 603559191Skris less strict. It will now permit CRL extensions even if it is not 603659191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 603759191Skris [Steve Henson] 603859191Skris 603959191Skris *) Initialize all non-automatic variables each time one of the openssl 604059191Skris sub-programs is started (this is necessary as they may be started 604159191Skris multiple times from the "OpenSSL>" prompt). 604259191Skris [Lennart Bang, Bodo Moeller] 604359191Skris 604459191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 604559191Skris removing all other RSA functionality (this is what NO_RSA does). This 604659191Skris is so (for example) those in the US can disable those operations covered 604759191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 604859191Skris key generation. 604959191Skris [Steve Henson] 605059191Skris 605159191Skris *) Non-copying interface to BIO pairs. 605259191Skris (still largely untested) 605359191Skris [Bodo Moeller] 605459191Skris 605559191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 605659191Skris ASCII string. This was handled independently in various places before. 605759191Skris [Steve Henson] 605859191Skris 605959191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 606059191Skris UTF8 strings a character at a time. 606159191Skris [Steve Henson] 606259191Skris 606359191Skris *) Use client_version from client hello to select the protocol 606459191Skris (s23_srvr.c) and for RSA client key exchange verification 606559191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 606659191Skris [Bodo Moeller] 606759191Skris 606859191Skris *) Add various utility functions to handle SPKACs, these were previously 606959191Skris handled by poking round in the structure internals. Added new function 607059191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 607159191Skris print, verify and generate SPKACs. Based on an original idea from 607259191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 607359191Skris [Steve Henson] 607459191Skris 607559191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 607659191Skris [Andy Polyakov] 607759191Skris 607859191Skris *) Allow the config file extension section to be overwritten on the 607959191Skris command line. Based on an original idea from Massimiliano Pala 608059191Skris <madwolf@comune.modena.it>. The new option is called -extensions 608159191Skris and can be applied to ca, req and x509. Also -reqexts to override 608259191Skris the request extensions in req and -crlexts to override the crl extensions 608359191Skris in ca. 608459191Skris [Steve Henson] 608559191Skris 608659191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 608759191Skris the same field multiple times by preceding it by "XXXX." for example: 608859191Skris 1.OU="Unit name 1" 608959191Skris 2.OU="Unit name 2" 609059191Skris this is the same syntax as used in the req config file. 609159191Skris [Steve Henson] 609259191Skris 609359191Skris *) Allow certificate extensions to be added to certificate requests. These 609459191Skris are specified in a 'req_extensions' option of the req section of the 609559191Skris config file. They can be printed out with the -text option to req but 609659191Skris are otherwise ignored at present. 609759191Skris [Steve Henson] 609859191Skris 609959191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 610059191Skris data read consists of only the final block it would not decrypted because 610159191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 610259191Skris A misplaced 'break' also meant the decrypted final block might not be 610359191Skris copied until the next read. 610459191Skris [Steve Henson] 610559191Skris 610659191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 610759191Skris a few extra parameters to the DH structure: these will be useful if 610859191Skris for example we want the value of 'q' or implement X9.42 DH. 610959191Skris [Steve Henson] 611059191Skris 611159191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 611259191Skris provides hooks that allow the default DSA functions or functions on a 611359191Skris "per key" basis to be replaced. This allows hardware acceleration and 611459191Skris hardware key storage to be handled without major modification to the 611559191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 611659191Skris associated functions. 611759191Skris [Steve Henson] 611859191Skris 611959191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 612059191Skris as "read only": it can't be written to and the buffer it points to will 612159191Skris not be freed. Reading from a read only BIO is much more efficient than 612259191Skris a normal memory BIO. This was added because there are several times when 612359191Skris an area of memory needs to be read from a BIO. The previous method was 612459191Skris to create a memory BIO and write the data to it, this results in two 612559191Skris copies of the data and an O(n^2) reading algorithm. There is a new 612659191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 612759191Skris an area of memory. Also modified the PKCS#7 routines to use read only 612859191Skris memory BIOs. 612959191Skris [Steve Henson] 613059191Skris 613159191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 613259191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 613359191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 613459191Skris but a retry condition occured while trying to read the rest. 613559191Skris [Bodo Moeller] 613659191Skris 613759191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 613859191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 613959191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 614059191Skris the encrypted data type: this is a more sensible place to put it and it 614159191Skris allows the PKCS#12 code to be tidied up that duplicated this 614259191Skris functionality. 614359191Skris [Steve Henson] 614459191Skris 614559191Skris *) Changed obj_dat.pl script so it takes its input and output files on 614659191Skris the command line. This should avoid shell escape redirection problems 614759191Skris under Win32. 614859191Skris [Steve Henson] 614959191Skris 615059191Skris *) Initial support for certificate extension requests, these are included 615159191Skris in things like Xenroll certificate requests. Included functions to allow 615259191Skris extensions to be obtained and added. 615359191Skris [Steve Henson] 615459191Skris 615559191Skris *) -crlf option to s_client and s_server for sending newlines as 615659191Skris CRLF (as required by many protocols). 615759191Skris [Bodo Moeller] 615859191Skris 615955714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 616055714Skris 616155714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 616255714Skris [Ralf S. Engelschall] 616355714Skris 616455714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 616555714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 616655714Skris 616755714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 616855714Skris program. 616955714Skris [Steve Henson] 617055714Skris 617155714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 617255714Skris DH parameters/keys (q is lost during that conversion, but the resulting 617355714Skris DH parameters contain its length). 617455714Skris 617555714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 617655714Skris much faster than DH_generate_parameters (which creates parameters 617755714Skris where p = 2*q + 1), and also the smaller q makes DH computations 617855714Skris much more efficient (160-bit exponentiation instead of 1024-bit 617955714Skris exponentiation); so this provides a convenient way to support DHE 618055714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 618155714Skris utter importance to use 618255714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 618355714Skris or 618455714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 618555714Skris when such DH parameters are used, because otherwise small subgroup 618655714Skris attacks may become possible! 618755714Skris [Bodo Moeller] 618855714Skris 618955714Skris *) Avoid memory leak in i2d_DHparams. 619055714Skris [Bodo Moeller] 619155714Skris 619255714Skris *) Allow the -k option to be used more than once in the enc program: 619355714Skris this allows the same encrypted message to be read by multiple recipients. 619455714Skris [Steve Henson] 619555714Skris 619655714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 619755714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 619855714Skris it will always use the numerical form of the OID, even if it has a short 619955714Skris or long name. 620055714Skris [Steve Henson] 620155714Skris 620255714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 620355714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 620455714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 620555714Skris no private key components need be present and it might store extra data 620659191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 620759191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 620859191Skris private key operations. 620955714Skris [Steve Henson] 621055714Skris 621155714Skris *) Added support for SPARC Linux. 621255714Skris [Andy Polyakov] 621355714Skris 621455714Skris *) pem_password_cb function type incompatibly changed from 621555714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 621655714Skris to 621755714Skris ....(char *buf, int size, int rwflag, void *userdata); 621855714Skris so that applications can pass data to their callbacks: 621955714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 622055714Skris additional void * argument, which is just handed through whenever 622155714Skris the password callback is called. 622259191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 622355714Skris 622455714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 622555714Skris 622655714Skris Compatibility note: As many C implementations push function arguments 622755714Skris onto the stack in reverse order, the new library version is likely to 622855714Skris interoperate with programs that have been compiled with the old 622955714Skris pem_password_cb definition (PEM_whatever takes some data that 623055714Skris happens to be on the stack as its last argument, and the callback 623155714Skris just ignores this garbage); but there is no guarantee whatsoever that 623255714Skris this will work. 623355714Skris 623455714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 623555714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 623655714Skris problems not only on Windows, but also on some Unix platforms. 623755714Skris To avoid problematic command lines, these definitions are now in an 623855714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 623955714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 624055714Skris [Bodo Moeller] 624155714Skris 624255714Skris *) MIPS III/IV assembler module is reimplemented. 624355714Skris [Andy Polyakov] 624455714Skris 624555714Skris *) More DES library cleanups: remove references to srand/rand and 624655714Skris delete an unused file. 624755714Skris [Ulf M�ller] 624855714Skris 624955714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 625055714Skris since not many people have MASM (ml) and it can be hard to obtain. 625155714Skris This is currently experimental but it seems to work OK and pass all 625255714Skris the tests. Check out INSTALL.W32 for info. 625355714Skris [Steve Henson] 625455714Skris 625555714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 625655714Skris without temporary keys kept an extra copy of the server key, 625755714Skris and connections with temporary keys did not free everything in case 625855714Skris of an error. 625955714Skris [Bodo Moeller] 626055714Skris 626155714Skris *) New function RSA_check_key and new openssl rsa option -check 626255714Skris for verifying the consistency of RSA keys. 626355714Skris [Ulf Moeller, Bodo Moeller] 626455714Skris 626555714Skris *) Various changes to make Win32 compile work: 626655714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 626755714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 626855714Skris comparison" warnings. 626955714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 627055714Skris [Steve Henson] 627155714Skris 627255714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 627355714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 627455714Skris derived keys are printed to stderr. 627555714Skris [Steve Henson] 627655714Skris 627755714Skris *) Copy the flags in ASN1_STRING_dup(). 627855714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 627955714Skris 628055714Skris *) The x509 application mishandled signing requests containing DSA 628155714Skris keys when the signing key was also DSA and the parameters didn't match. 628255714Skris 628355714Skris It was supposed to omit the parameters when they matched the signing key: 628455714Skris the verifying software was then supposed to automatically use the CA's 628555714Skris parameters if they were absent from the end user certificate. 628655714Skris 628755714Skris Omitting parameters is no longer recommended. The test was also 628855714Skris the wrong way round! This was probably due to unusual behaviour in 628955714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 629055714Skris This meant that parameters were omitted when they *didn't* match and 629155714Skris the certificate was useless. Certificates signed with 'ca' didn't have 629255714Skris this bug. 629355714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 629455714Skris 629555714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 629655714Skris The interface is as follows: 629755714Skris Applications can use 629855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 629955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 630055714Skris "off" is now the default. 630155714Skris The library internally uses 630255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 630355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 630455714Skris to disable memory-checking temporarily. 630555714Skris 630655714Skris Some inconsistent states that previously were possible (and were 630755714Skris even the default) are now avoided. 630855714Skris 630955714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 631055714Skris with each memory chunk allocated; this is occasionally more helpful 631155714Skris than just having a counter. 631255714Skris 631355714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 631455714Skris 631555714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 631655714Skris extensions. 631755714Skris [Bodo Moeller] 631855714Skris 631955714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 632055714Skris which largely parallels "options", but is for changing API behaviour, 632155714Skris whereas "options" are about protocol behaviour. 632255714Skris Initial "mode" flags are: 632355714Skris 632455714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 632555714Skris a single record has been written. 632655714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 632755714Skris retries use the same buffer location. 632855714Skris (But all of the contents must be 632955714Skris copied!) 633055714Skris [Bodo Moeller] 633155714Skris 633279998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 633355714Skris worked. 633455714Skris 633555714Skris *) Fix problems with no-hmac etc. 633655714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 633755714Skris 633855714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 633955714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 634055714Skris to mess around with the internals of an RSA structure. 634155714Skris [Steve Henson] 634255714Skris 634355714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 634455714Skris Also really enable memory leak checks in openssl.c and in some 634555714Skris test programs. 634655714Skris [Chad C. Mulligan, Bodo Moeller] 634755714Skris 634855714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 634955714Skris up the length of negative integers. This has now been simplified to just 635055714Skris store the length when it is first determined and use it later, rather 635155714Skris than trying to keep track of where data is copied and updating it to 635255714Skris point to the end. 635355714Skris [Steve Henson, reported by Brien Wheeler 635455714Skris <bwheeler@authentica-security.com>] 635555714Skris 635655714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 635755714Skris of a PKCS#7 signature but with the signing certificate passed to the 635855714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 635955714Skris certificate is present in the PKCS#7 structure. This isn't always the 636055714Skris case: certificates can be omitted from a PKCS#7 structure and be 636155714Skris distributed by "out of band" means (such as a certificate database). 636255714Skris [Steve Henson] 636355714Skris 636455714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 636555714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 636655714Skris necessary function names. 636755714Skris [Steve Henson] 636855714Skris 636955714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 637055714Skris options set by Configure in the top level Makefile, and Configure 637155714Skris was not even able to write more than one option correctly. 637255714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 637355714Skris [Bodo Moeller] 637455714Skris 637555714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 637655714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 637755714Skris for example allow memory BIOs to contain config info. 637855714Skris [Steve Henson] 637955714Skris 638055714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 638155714Skris Whoever hopes to achieve shared-library compatibility across versions 638255714Skris must use this, not the compile-time macro. 638355714Skris (Exercise 0.9.4: Which is the minimum library version required by 638455714Skris such programs?) 638555714Skris Note: All this applies only to multi-threaded programs, others don't 638655714Skris need locks. 638755714Skris [Bodo Moeller] 638855714Skris 638955714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 639055714Skris through a BIO pair triggered the default case, i.e. 639155714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 639255714Skris [Bodo Moeller] 639355714Skris 639455714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 639555714Skris can use the SSL library even if none of the specific BIOs is 639655714Skris appropriate. 639755714Skris [Bodo Moeller] 639855714Skris 639955714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 640055714Skris for the encoded length. 640155714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 640255714Skris 640355714Skris *) Add initial documentation of the X509V3 functions. 640455714Skris [Steve Henson] 640555714Skris 640655714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 640755714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 640855714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 640955714Skris secure PKCS#8 private key format with a high iteration count. 641055714Skris [Steve Henson] 641155714Skris 641255714Skris *) Fix determination of Perl interpreter: A perl or perl5 641355714Skris _directory_ in $PATH was also accepted as the interpreter. 641455714Skris [Ralf S. Engelschall] 641555714Skris 641655714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 641755714Skris wrong with it but it was very old and did things like calling 641855714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 641955714Skris unusual formatting. 642055714Skris [Steve Henson] 642155714Skris 642255714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 642355714Skris to use the new extension code. 642455714Skris [Steve Henson] 642555714Skris 642655714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 642755714Skris with macros. This should make it easier to change their form, add extra 642855714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 642955714Skris constant. 643055714Skris [Steve Henson] 643155714Skris 643255714Skris *) Add to configuration table a new entry that can specify an alternative 643355714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 643455714Skris according to Mark Crispin <MRC@Panda.COM>. 643555714Skris [Bodo Moeller] 643655714Skris 643755714Skris#if 0 643855714Skris *) DES CBC did not update the IV. Weird. 643955714Skris [Ben Laurie] 644055714Skris#else 644155714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 644255714Skris Changing the behaviour of the former might break existing programs -- 644355714Skris where IV updating is needed, des_ncbc_encrypt can be used. 644455714Skris#endif 644555714Skris 644655714Skris *) When bntest is run from "make test" it drives bc to check its 644755714Skris calculations, as well as internally checking them. If an internal check 644855714Skris fails, it needs to cause bc to give a non-zero result or make test carries 644955714Skris on without noticing the failure. Fixed. 645055714Skris [Ben Laurie] 645155714Skris 645255714Skris *) DES library cleanups. 645355714Skris [Ulf M�ller] 645455714Skris 645555714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 645655714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 645755714Skris ciphers. NOTE: although the key derivation function has been verified 645855714Skris against some published test vectors it has not been extensively tested 645955714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 646055714Skris of v2.0. 646155714Skris [Steve Henson] 646255714Skris 646355714Skris *) Instead of "mkdir -p", which is not fully portable, use new 646455714Skris Perl script "util/mkdir-p.pl". 646555714Skris [Bodo Moeller] 646655714Skris 646755714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 646855714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 646955714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 647055714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 647155714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 647255714Skris underlying key generation function so it must do its own ASN1 parsing. 647355714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 647455714Skris 'parameter' argument instead of literal salt and iteration count values 647555714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 647655714Skris [Steve Henson] 647755714Skris 647855714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 647955714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 648055714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 648155714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 648255714Skris value was just used as a "magic string" and not used directly its 648355714Skris value doesn't matter. 648455714Skris [Steve Henson] 648555714Skris 648655714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 648755714Skris support mutable. 648855714Skris [Ben Laurie] 648955714Skris 649055714Skris *) "linux-sparc64" configuration (ultrapenguin). 649155714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 649255714Skris "linux-sparc" configuration. 649355714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 649455714Skris 649555714Skris *) config now generates no-xxx options for missing ciphers. 649655714Skris [Ulf M�ller] 649755714Skris 649855714Skris *) Support the EBCDIC character set (work in progress). 649955714Skris File ebcdic.c not yet included because it has a different license. 650055714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 650155714Skris 650255714Skris *) Support BS2000/OSD-POSIX. 650355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 650455714Skris 650555714Skris *) Make callbacks for key generation use void * instead of char *. 650655714Skris [Ben Laurie] 650755714Skris 650855714Skris *) Make S/MIME samples compile (not yet tested). 650955714Skris [Ben Laurie] 651055714Skris 651155714Skris *) Additional typesafe stacks. 651255714Skris [Ben Laurie] 651355714Skris 651455714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 651555714Skris [Bodo Moeller] 651655714Skris 651755714Skris 651855714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 651955714Skris 652055714Skris *) New configuration variant "sco5-gcc". 652155714Skris 652255714Skris *) Updated some demos. 652355714Skris [Sean O Riordain, Wade Scholine] 652455714Skris 652555714Skris *) Add missing BIO_free at exit of pkcs12 application. 652655714Skris [Wu Zhigang] 652755714Skris 652855714Skris *) Fix memory leak in conf.c. 652955714Skris [Steve Henson] 653055714Skris 653155714Skris *) Updates for Win32 to assembler version of MD5. 653255714Skris [Steve Henson] 653355714Skris 653455714Skris *) Set #! path to perl in apps/der_chop to where we found it 653555714Skris instead of using a fixed path. 653655714Skris [Bodo Moeller] 653755714Skris 653855714Skris *) SHA library changes for irix64-mips4-cc. 653955714Skris [Andy Polyakov] 654055714Skris 654155714Skris *) Improvements for VMS support. 654255714Skris [Richard Levitte] 654355714Skris 654455714Skris 654555714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 654655714Skris 654755714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 654855714Skris This also avoids the problems with SC4.2 and unpatched SC5. 654955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 655055714Skris 655155714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 655255714Skris These are required because of the typesafe stack would otherwise break 655355714Skris existing code. If old code used a structure member which used to be STACK 655455714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 655555714Skris sk_num or sk_value it would produce an error because the num, data members 655655714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 655755714Skris replaces the old method of assigning a value to sk_value 655855714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 655955714Skris that does this will no longer work (and should use sk_set instead) but 656055714Skris this could be regarded as a "questionable" behaviour anyway. 656155714Skris [Steve Henson] 656255714Skris 656355714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 656455714Skris correctly handle encrypted S/MIME data. 656555714Skris [Steve Henson] 656655714Skris 656755714Skris *) Change type of various DES function arguments from des_cblock 656855714Skris (which means, in function argument declarations, pointer to char) 656955714Skris to des_cblock * (meaning pointer to array with 8 char elements), 657055714Skris which allows the compiler to do more typechecking; it was like 657155714Skris that back in SSLeay, but with lots of ugly casts. 657255714Skris 657355714Skris Introduce new type const_des_cblock. 657455714Skris [Bodo Moeller] 657555714Skris 657655714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 657755714Skris problems: find RecipientInfo structure that matches recipient certificate 657855714Skris and initialise the ASN1 structures properly based on passed cipher. 657955714Skris [Steve Henson] 658055714Skris 658155714Skris *) Belatedly make the BN tests actually check the results. 658255714Skris [Ben Laurie] 658355714Skris 658455714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 658555714Skris to and from BNs: it was completely broken. New compilation option 658655714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 658755714Skris key elements as negative integers. 658855714Skris [Steve Henson] 658955714Skris 659055714Skris *) Reorganize and speed up MD5. 659155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 659255714Skris 659355714Skris *) VMS support. 659455714Skris [Richard Levitte <richard@levitte.org>] 659555714Skris 659655714Skris *) New option -out to asn1parse to allow the parsed structure to be 659755714Skris output to a file. This is most useful when combined with the -strparse 659855714Skris option to examine the output of things like OCTET STRINGS. 659955714Skris [Steve Henson] 660055714Skris 660155714Skris *) Make SSL library a little more fool-proof by not requiring any longer 660255714Skris that SSL_set_{accept,connect}_state be called before 660355714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 660455714Skris in many applications because usually everything *appeared* to work as 660555714Skris intended anyway -- now it really works as intended). 660655714Skris [Bodo Moeller] 660755714Skris 660855714Skris *) Move openssl.cnf out of lib/. 660955714Skris [Ulf M�ller] 661055714Skris 661155714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 661255714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 661355714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 661455714Skris [Ralf S. Engelschall] 661555714Skris 661655714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 661755714Skris handle PKCS#7 enveloped data properly. 661855714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 661955714Skris 662055714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 662155714Skris copying pointers. The cert_st handling is changed by this in 662255714Skris various ways (and thus what used to be known as ctx->default_cert 662355714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 662455714Skris any longer when s->cert does not give us what we need). 662555714Skris ssl_cert_instantiate becomes obsolete by this change. 662655714Skris As soon as we've got the new code right (possibly it already is?), 662755714Skris we have solved a couple of bugs of the earlier code where s->cert 662855714Skris was used as if it could not have been shared with other SSL structures. 662955714Skris 663055714Skris Note that using the SSL API in certain dirty ways now will result 663155714Skris in different behaviour than observed with earlier library versions: 663255714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 663355714Skris does not influence s as it used to. 663455714Skris 663555714Skris In order to clean up things more thoroughly, inside SSL_SESSION 663655714Skris we don't use CERT any longer, but a new structure SESS_CERT 663755714Skris that holds per-session data (if available); currently, this is 663855714Skris the peer's certificate chain and, for clients, the server's certificate 663955714Skris and temporary key. CERT holds only those values that can have 664055714Skris meaningful defaults in an SSL_CTX. 664155714Skris [Bodo Moeller] 664255714Skris 664355714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 664455714Skris from the internal representation. Various PKCS#7 fixes: remove some 664555714Skris evil casts and set the enc_dig_alg field properly based on the signing 664655714Skris key type. 664755714Skris [Steve Henson] 664855714Skris 664955714Skris *) Allow PKCS#12 password to be set from the command line or the 665055714Skris environment. Let 'ca' get its config file name from the environment 665155714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 665255714Skris and 'x509'). 665355714Skris [Steve Henson] 665455714Skris 665555714Skris *) Allow certificate policies extension to use an IA5STRING for the 665655714Skris organization field. This is contrary to the PKIX definition but 665755714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 665855714Skris extension option. 665955714Skris [Steve Henson] 666055714Skris 666155714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 666255714Skris without disallowing inline assembler and the like for non-pedantic builds. 666355714Skris [Ben Laurie] 666455714Skris 666555714Skris *) Support Borland C++ builder. 666655714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 666755714Skris 666855714Skris *) Support Mingw32. 666955714Skris [Ulf M�ller] 667055714Skris 667155714Skris *) SHA-1 cleanups and performance enhancements. 667255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 667355714Skris 667455714Skris *) Sparc v8plus assembler for the bignum library. 667555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 667655714Skris 667755714Skris *) Accept any -xxx and +xxx compiler options in Configure. 667855714Skris [Ulf M�ller] 667955714Skris 668055714Skris *) Update HPUX configuration. 668155714Skris [Anonymous] 668255714Skris 668355714Skris *) Add missing sk_<type>_unshift() function to safestack.h 668455714Skris [Ralf S. Engelschall] 668555714Skris 668655714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 668755714Skris "extra_cert"s in addition to the certificate. (This makes sense 668855714Skris only for "PEM" format files, as chains as a whole are not 668955714Skris DER-encoded.) 669055714Skris [Bodo Moeller] 669155714Skris 669255714Skris *) Support verify_depth from the SSL API. 669355714Skris x509_vfy.c had what can be considered an off-by-one-error: 669455714Skris Its depth (which was not part of the external interface) 669555714Skris was actually counting the number of certificates in a chain; 669655714Skris now it really counts the depth. 669755714Skris [Bodo Moeller] 669855714Skris 669955714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 670055714Skris instead of X509err, which often resulted in confusing error 670155714Skris messages since the error codes are not globally unique 670255714Skris (e.g. an alleged error in ssl3_accept when a certificate 670355714Skris didn't match the private key). 670455714Skris 670555714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 670655714Skris value (so that you don't need SSL_set_session_id_context for each 670755714Skris connection using the SSL_CTX). 670855714Skris [Bodo Moeller] 670955714Skris 671055714Skris *) OAEP decoding bug fix. 671155714Skris [Ulf M�ller] 671255714Skris 671355714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 671455714Skris David Harris. 671555714Skris [Bodo Moeller] 671655714Skris 671755714Skris *) New Configure options "threads" and "no-threads". For systems 671855714Skris where the proper compiler options are known (currently Solaris 671955714Skris and Linux), "threads" is the default. 672055714Skris [Bodo Moeller] 672155714Skris 672255714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 672355714Skris [Bodo Moeller] 672455714Skris 672555714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 672655714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 672755714Skris such as /usr/local/bin. 672855714Skris [Bodo Moeller] 672955714Skris 673055714Skris *) "make linux-shared" to build shared libraries. 673155714Skris [Niels Poppe <niels@netbox.org>] 673255714Skris 673355714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 673455714Skris [Ulf M�ller] 673555714Skris 673655714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 673755714Skris extension adding in x509 utility. 673855714Skris [Steve Henson] 673955714Skris 674055714Skris *) Remove NOPROTO sections and error code comments. 674155714Skris [Ulf M�ller] 674255714Skris 674355714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 674455714Skris prototypes. 674555714Skris [Steve Henson] 674655714Skris 674755714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 674855714Skris [Ulf M�ller] 674955714Skris 675055714Skris *) Complete rewrite of the error code script(s). It is all now handled 675155714Skris by one script at the top level which handles error code gathering, 675255714Skris header rewriting and C source file generation. It should be much better 675355714Skris than the old method: it now uses a modified version of Ulf's parser to 675455714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 675555714Skris aren't needed for error creation any more) and do a better job of 675655714Skris translating function codes into names. The old 'ASN1 error code imbedded 675755714Skris in a comment' is no longer necessary and it doesn't use .err files which 675855714Skris have now been deleted. Also the error code call doesn't have to appear all 675955714Skris on one line (which resulted in some large lines...). 676055714Skris [Steve Henson] 676155714Skris 676255714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 676355714Skris [Bodo Moeller] 676455714Skris 676555714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 676655714Skris 0 (which usually indicates a closed connection), but continue reading. 676755714Skris [Bodo Moeller] 676855714Skris 676955714Skris *) Fix some race conditions. 677055714Skris [Bodo Moeller] 677155714Skris 677255714Skris *) Add support for CRL distribution points extension. Add Certificate 677355714Skris Policies and CRL distribution points documentation. 677455714Skris [Steve Henson] 677555714Skris 677655714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 677755714Skris [Ulf M�ller] 677855714Skris 677955714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 678055714Skris 8 of keying material. Merlin has also confirmed interop with this fix 678155714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 678255714Skris [Merlin Hughes <merlin@baltimore.ie>] 678355714Skris 678455714Skris *) Fix lots of warnings. 678555714Skris [Richard Levitte <levitte@stacken.kth.se>] 678655714Skris 678755714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 678855714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 678955714Skris [Richard Levitte <levitte@stacken.kth.se>] 679055714Skris 679155714Skris *) Fix problems with sizeof(long) == 8. 679255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 679355714Skris 679455714Skris *) Change functions to ANSI C. 679555714Skris [Ulf M�ller] 679655714Skris 679755714Skris *) Fix typos in error codes. 679855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 679955714Skris 680055714Skris *) Remove defunct assembler files from Configure. 680155714Skris [Ulf M�ller] 680255714Skris 680355714Skris *) SPARC v8 assembler BIGNUM implementation. 680455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 680555714Skris 680655714Skris *) Support for Certificate Policies extension: both print and set. 680755714Skris Various additions to support the r2i method this uses. 680855714Skris [Steve Henson] 680955714Skris 681055714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 681155714Skris return a const string when you are expecting an allocated buffer. 681255714Skris [Ben Laurie] 681355714Skris 681455714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 681555714Skris types DirectoryString and DisplayText. 681655714Skris [Steve Henson] 681755714Skris 681855714Skris *) Add code to allow r2i extensions to access the configuration database, 681955714Skris add an LHASH database driver and add several ctx helper functions. 682055714Skris [Steve Henson] 682155714Skris 682255714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 682355714Skris fail when they extended the size of a BIGNUM. 682455714Skris [Steve Henson] 682555714Skris 682655714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 682755714Skris support typesafe stack. 682855714Skris [Steve Henson] 682955714Skris 683055714Skris *) Fix typo in SSL_[gs]et_options(). 683155714Skris [Nils Frostberg <nils@medcom.se>] 683255714Skris 683355714Skris *) Delete various functions and files that belonged to the (now obsolete) 683455714Skris old X509V3 handling code. 683555714Skris [Steve Henson] 683655714Skris 683755714Skris *) New Configure option "rsaref". 683855714Skris [Ulf M�ller] 683955714Skris 684055714Skris *) Don't auto-generate pem.h. 684155714Skris [Bodo Moeller] 684255714Skris 684355714Skris *) Introduce type-safe ASN.1 SETs. 684455714Skris [Ben Laurie] 684555714Skris 684655714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 684755714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 684855714Skris 684955714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 685055714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 685155714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 685255714Skris few STACKed things have been converted already. Feel free to convert more. 685355714Skris In the fullness of time, I'll do away with the STACK type altogether. 685455714Skris [Ben Laurie] 685555714Skris 685655714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 685755714Skris specified in <certfile> by updating the entry in the index.txt file. 685855714Skris This way one no longer has to edit the index.txt file manually for 685955714Skris revoking a certificate. The -revoke option does the gory details now. 686055714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 686155714Skris 686255714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 686355714Skris `-text' option at all and this way the `-noout -text' combination was 686455714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 686555714Skris [Ralf S. Engelschall] 686655714Skris 686755714Skris *) Make sure a corresponding plain text error message exists for the 686855714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 686955714Skris verify callback function determined that a certificate was revoked. 687055714Skris [Ralf S. Engelschall] 687155714Skris 687255714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 687355714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 687455714Skris all available cipers including rc5, which was forgotten until now. 687555714Skris In order to let the testing shell script know which algorithms 687655714Skris are available, a new (up to now undocumented) command 687755714Skris "openssl list-cipher-commands" is used. 687855714Skris [Bodo Moeller] 687955714Skris 688055714Skris *) Bugfix: s_client occasionally would sleep in select() when 688155714Skris it should have checked SSL_pending() first. 688255714Skris [Bodo Moeller] 688355714Skris 688455714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 688555714Skris the raw DSA values prior to ASN.1 encoding. 688655714Skris [Ulf M�ller] 688755714Skris 688855714Skris *) Tweaks to Configure 688955714Skris [Niels Poppe <niels@netbox.org>] 689055714Skris 689155714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 689255714Skris yet... 689355714Skris [Steve Henson] 689455714Skris 689555714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 689655714Skris [Ulf M�ller] 689755714Skris 689855714Skris *) New config option to avoid instructions that are illegal on the 80386. 689955714Skris The default code is faster, but requires at least a 486. 690055714Skris [Ulf M�ller] 690155714Skris 690255714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 690355714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 690455714Skris same as SSL2_VERSION anyway. 690555714Skris [Bodo Moeller] 690655714Skris 690755714Skris *) New "-showcerts" option for s_client. 690855714Skris [Bodo Moeller] 690955714Skris 691055714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 691155714Skris application. Various cleanups and fixes. 691255714Skris [Steve Henson] 691355714Skris 691455714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 691555714Skris modify error routines to work internally. Add error codes and PBE init 691655714Skris to library startup routines. 691755714Skris [Steve Henson] 691855714Skris 691955714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 692055714Skris packing functions to asn1 and evp. Changed function names and error 692155714Skris codes along the way. 692255714Skris [Steve Henson] 692355714Skris 692455714Skris *) PKCS12 integration: and so it begins... First of several patches to 692555714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 692655714Skris objects to objects.h 692755714Skris [Steve Henson] 692855714Skris 692955714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 693055714Skris and display support for Thawte strong extranet extension. 693155714Skris [Steve Henson] 693255714Skris 693355714Skris *) Add LinuxPPC support. 693455714Skris [Jeff Dubrule <igor@pobox.org>] 693555714Skris 693655714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 693755714Skris bn_div_words in alpha.s. 693855714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 693955714Skris 694055714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 694155714Skris OAEP isn't supported when OpenSSL is built with RSAref. 694255714Skris [Ulf Moeller <ulf@fitug.de>] 694355714Skris 694455714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 694555714Skris so they no longer are missing under -DNOPROTO. 694655714Skris [Soren S. Jorvang <soren@t.dk>] 694755714Skris 694855714Skris 694955714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 695055714Skris 695155714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 695255714Skris doesn't work when the session is reused. Coming soon! 695355714Skris [Ben Laurie] 695455714Skris 695555714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 695655714Skris context thus bypassing client cert protection! All software that uses 695755714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 695855714Skris allow session reuse! A fuller solution is in the works. 695955714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 696055714Skris 696155714Skris *) Some more source tree cleanups (removed obsolete files 696255714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 696355714Skris permission on "config" script to be executable) and a fix for the INSTALL 696455714Skris document. 696555714Skris [Ulf Moeller <ulf@fitug.de>] 696655714Skris 696755714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 696855714Skris Malloc, Free. 696955714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 697055714Skris 697155714Skris *) Make rsa_oaep_test return non-zero on error. 697255714Skris [Ulf Moeller <ulf@fitug.de>] 697355714Skris 697455714Skris *) Add support for native Solaris shared libraries. Configure 697555714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 697655714Skris if someone would make that last step automatic. 697755714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 697855714Skris 697955714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 698055714Skris [Ben Laurie] 698155714Skris 698255714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 698355714Skris except NULL ciphers". This means the default cipher list will no longer 698455714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 698555714Skris the string "DEFAULT:eNULL". 698655714Skris [Steve Henson] 698755714Skris 698855714Skris *) Fix to RSA private encryption routines: if p < q then it would 698955714Skris occasionally produce an invalid result. This will only happen with 699055714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 699155714Skris [Steve Henson] 699255714Skris 699355714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 699455714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 699555714Skris because this way one can also use an interpreter named `perl5' (which is 699655714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 699755714Skris installed as `perl'). 699855714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 699955714Skris 700055714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 700155714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 700255714Skris 700355714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 700455714Skris advapi32.lib to Win32 build and change the pem test comparision 700555714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 700655714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 700755714Skris and crypto/des/ede_cbcm_enc.c. 700855714Skris [Steve Henson] 700955714Skris 701055714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 701155714Skris [Ben Laurie] 701255714Skris 701355714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 701455714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 701555714Skris is horrible: I feel ill.... 701655714Skris [Steve Henson] 701755714Skris 701855714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 701955714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 702055714Skris sections: 10 functions were absent from non ANSI section and not exported 702155714Skris from Windows DLLs. Fixed up libeay.num for new functions. 702255714Skris [Steve Henson] 702355714Skris 702455714Skris *) Make `openssl version' output lines consistent. 702555714Skris [Ralf S. Engelschall] 702655714Skris 702755714Skris *) Fix Win32 symbol export lists for BIO functions: Added 702855714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 702955714Skris to ms/libeay{16,32}.def. 703055714Skris [Ralf S. Engelschall] 703155714Skris 703255714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 703355714Skris fine under Unix and passes some trivial tests I've now added. But the 703455714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 703555714Skris added to make sure no one expects that this stuff really works in the 703655714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 703755714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 703855714Skris openssl_bio.xs. 703955714Skris [Ralf S. Engelschall] 704055714Skris 704155714Skris *) Fix the generation of two part addresses in perl. 704255714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 704355714Skris 704455714Skris *) Add config entry for Linux on MIPS. 704555714Skris [John Tobey <jtobey@channel1.com>] 704655714Skris 704755714Skris *) Make links whenever Configure is run, unless we are on Windoze. 704855714Skris [Ben Laurie] 704955714Skris 705055714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 705155714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 705255714Skris in CRLs. 705355714Skris [Steve Henson] 705455714Skris 705555714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 705655714Skris other platforms details on the command line without having to patch the 705755714Skris Configure script everytime: One now can use ``perl Configure 705855714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 705955714Skris to them (seperated by colons). This is treated as there would be a static 706055714Skris pre-configured entry in Configure's %table under key <id> with value 706155714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 706255714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 706355714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 706455714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 706555714Skris [Ralf S. Engelschall] 706655714Skris 706755714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 706855714Skris [Ben Laurie] 706955714Skris 707055714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 707155714Skris on the `perl Configure ...' command line. This way one can compile 707255714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 707355714Skris for linking it into DSOs. 707455714Skris [Ralf S. Engelschall] 707555714Skris 707655714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 707755714Skris Fixed. 707855714Skris [Ben Laurie] 707955714Skris 708055714Skris *) Cleaned up the LICENSE document: The official contact for any license 708155714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 708255714Skris And add a paragraph about the dual-license situation to make sure people 708355714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 708455714Skris to the OpenSSL toolkit. 708555714Skris [Ralf S. Engelschall] 708655714Skris 708755714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 708855714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 708955714Skris Additonally cleaned up the `make links' target: Remove unnecessary 709055714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 709155714Skris to speed processing and no longer clutter the display with confusing 709255714Skris stuff. Instead only the actually done links are displayed. 709355714Skris [Ralf S. Engelschall] 709455714Skris 709555714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 709655714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 709755714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 709855714Skris encryption. 709955714Skris [Ben Laurie] 710055714Skris 710155714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 710255714Skris signed attributes when verifying signatures (this would break them), 710355714Skris the detached data encoding was wrong and public keys obtained using 710455714Skris X509_get_pubkey() weren't freed. 710555714Skris [Steve Henson] 710655714Skris 710755714Skris *) Add text documentation for the BUFFER functions. Also added a work around 710855714Skris to a Win95 console bug. This was triggered by the password read stuff: the 710955714Skris last character typed gets carried over to the next fread(). If you were 711055714Skris generating a new cert request using 'req' for example then the last 711155714Skris character of the passphrase would be CR which would then enter the first 711255714Skris field as blank. 711355714Skris [Steve Henson] 711455714Skris 711555714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 711655714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 711755714Skris button and can be used by applications based on OpenSSL to show the 711855714Skris relationship to the OpenSSL project. 711955714Skris [Ralf S. Engelschall] 712055714Skris 712155714Skris *) Remove confusing variables in function signatures in files 712255714Skris ssl/ssl_lib.c and ssl/ssl.h. 712355714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 712455714Skris 712555714Skris *) Don't install bss_file.c under PREFIX/include/ 712655714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 712755714Skris 712855714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 712955714Skris functions that return function pointers and has support for NT specific 713055714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 713155714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 713255714Skris unsigned to signed types: this was killing the Win32 compile. 713355714Skris [Steve Henson] 713455714Skris 713555714Skris *) Add new certificate file to stack functions, 713655714Skris SSL_add_dir_cert_subjects_to_stack() and 713755714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 713855714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 713955714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 714055714Skris This means that Apache-SSL and similar packages don't have to mess around 714155714Skris to add as many CAs as they want to the preferred list. 714255714Skris [Ben Laurie] 714355714Skris 714455714Skris *) Experiment with doxygen documentation. Currently only partially applied to 714555714Skris ssl/ssl_lib.c. 714655714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 714755714Skris openssl.doxy as the configuration file. 714855714Skris [Ben Laurie] 714955714Skris 715055714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 715155714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 715255714Skris 715355714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 715455714Skris compiled in by default: it has problems with large keys. 715555714Skris [Steve Henson] 715655714Skris 715755714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 715855714Skris DH private keys and/or callback functions which directly correspond to 715955714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 716055714Skris is needed for applications which have to configure certificates on a 716155714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 716255714Skris (e.g. s_server). 716355714Skris For the RSA certificate situation is makes no difference, but 716455714Skris for the DSA certificate situation this fixes the "no shared cipher" 716555714Skris problem where the OpenSSL cipher selection procedure failed because the 716655714Skris temporary keys were not overtaken from the context and the API provided 716755714Skris no way to reconfigure them. 716855714Skris The new functions now let applications reconfigure the stuff and they 716955714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 717055714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 717155714Skris non-public-API function ssl_cert_instantiate() is used as a helper 717255714Skris function and also to reduce code redundancy inside ssl_rsa.c. 717355714Skris [Ralf S. Engelschall] 717455714Skris 717555714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 717655714Skris area because they are useful for the DSA situation and should be 717755714Skris recognized by the users. 717855714Skris [Ralf S. Engelschall] 717955714Skris 718055714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 718155714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 718255714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 718355714Skris already masked variable. 718455714Skris [Richard Levitte <levitte@stacken.kth.se>] 718555714Skris 718655714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 718755714Skris [Richard Levitte <levitte@stacken.kth.se>] 718855714Skris 718955714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 719055714Skris from `int' to `unsigned int' because it's a length and initialized by 719155714Skris EVP_DigestFinal() which expects an `unsigned int *'. 719255714Skris [Richard Levitte <levitte@stacken.kth.se>] 719355714Skris 719455714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 719555714Skris script. Instead use the usual Shell->Perl transition trick. 719655714Skris [Ralf S. Engelschall] 719755714Skris 719855714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 719955714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 720055714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 720155714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 720255714Skris currently the public key is printed (a decision which was already done by 720355714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 720455714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 720555714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 720655714Skris now, too. 720755714Skris [Ralf S. Engelschall] 720855714Skris 720955714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 721055714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 721155714Skris [Arne Ansper <arne@ats.cyber.ee>] 721255714Skris 721355714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 721455714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 721555714Skris config file. 721655714Skris [Steve Henson] 721755714Skris 721855714Skris *) Add cool BIO that does syslog (or event log on NT). 721955714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 722055714Skris 722155714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 722255714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 722355714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 722455714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 722555714Skris [Ben Laurie] 722655714Skris 722755714Skris *) Add preliminary config info for new extension code. 722855714Skris [Steve Henson] 722955714Skris 723055714Skris *) Make RSA_NO_PADDING really use no padding. 723155714Skris [Ulf Moeller <ulf@fitug.de>] 723255714Skris 723355714Skris *) Generate errors when private/public key check is done. 723455714Skris [Ben Laurie] 723555714Skris 723655714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 723755714Skris for some CRL extensions and new objects added. 723855714Skris [Steve Henson] 723955714Skris 724055714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 724155714Skris key usage extension and fuller support for authority key id. 724255714Skris [Steve Henson] 724355714Skris 724455714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 724555714Skris padding method for RSA, which is recommended for new applications in PKCS 724655714Skris #1 v2.0 (RFC 2437, October 1998). 724755714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 724855714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 724955714Skris against Bleichbacher's attack on RSA. 725055714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 725155714Skris Ben Laurie] 725255714Skris 725355714Skris *) Updates to the new SSL compression code 725455714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 725555714Skris 725655714Skris *) Fix so that the version number in the master secret, when passed 725755714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 725855714Skris (because the server will not accept higher), that the version number 725955714Skris is 0x03,0x01, not 0x03,0x00 726055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 726155714Skris 726255714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 726355714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 726455714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 726555714Skris [Steve Henson] 726655714Skris 726755714Skris *) Support for RAW extensions where an arbitrary extension can be 726855714Skris created by including its DER encoding. See apps/openssl.cnf for 726955714Skris an example. 727055714Skris [Steve Henson] 727155714Skris 727255714Skris *) Make sure latest Perl versions don't interpret some generated C array 727355714Skris code as Perl array code in the crypto/err/err_genc.pl script. 727455714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 727555714Skris 727655714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 727755714Skris not many people have the assembler. Various Win32 compilation fixes and 727855714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 727955714Skris build instructions. 728055714Skris [Steve Henson] 728155714Skris 728255714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 728355714Skris file under Win32 and also build pem.h from pem.org. New script 728455714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 728555714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 728655714Skris [Steve Henson] 728755714Skris 728855714Skris *) Major rework of DES function declarations, in the pursuit of correctness 728955714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 729055714Skris too. You may find this causes warnings in your code. Zapping your evil 729155714Skris casts will probably fix them. Mostly. 729255714Skris [Ben Laurie] 729355714Skris 729455714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 729555714Skris obj_dat.pl. It considered a zero in an object definition to mean 729655714Skris "end of object": none of the objects in objects.h have any zeros 729755714Skris so it wasn't spotted. 729855714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 729955714Skris 730055714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 730155714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 730255714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 730355714Skris vectors if you have them. 730455714Skris [Ben Laurie] 730555714Skris 730655714Skris *) Correct calculation of key length for export ciphers (too much space was 730755714Skris allocated for null ciphers). This has not been tested! 730855714Skris [Ben Laurie] 730955714Skris 731055714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 731155714Skris message is now correct (it understands "crypto" and "ssl" on its 731255714Skris command line). There is also now an "update" option. This will update 731355714Skris the util/ssleay.num and util/libeay.num files with any new functions. 731455714Skris If you do a: 731555714Skris perl util/mkdef.pl crypto ssl update 731655714Skris it will update them. 731755714Skris [Steve Henson] 731855714Skris 731955714Skris *) Overhauled the Perl interface (perl/*): 732055714Skris - ported BN stuff to OpenSSL's different BN library 732155714Skris - made the perl/ source tree CVS-aware 732255714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 732355714Skris their history because I've copied them in the repository) 732455714Skris - removed obsolete files (the test scripts will be replaced 732555714Skris by better Test::Harness variants in the future) 732655714Skris [Ralf S. Engelschall] 732755714Skris 732855714Skris *) First cut for a very conservative source tree cleanup: 732955714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 733055714Skris where we collect the old documents and readme texts. 733155714Skris 2. remove the first part of files where I'm already sure that we no 733255714Skris longer need them because of three reasons: either they are just temporary 733355714Skris files which were left by Eric or they are preserved original files where 733455714Skris I've verified that the diff is also available in the CVS via "cvs diff 733555714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 733655714Skris the crypto/md/ stuff). 733755714Skris [Ralf S. Engelschall] 733855714Skris 733955714Skris *) More extension code. Incomplete support for subject and issuer alt 734055714Skris name, issuer and authority key id. Change the i2v function parameters 734155714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 734255714Skris what that's for :-) Fix to ASN1 macro which messed up 734355714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 734455714Skris [Steve Henson] 734555714Skris 734655714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 734755714Skris INTEGER code. 734855714Skris [Steve Henson] 734955714Skris 735055714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 735155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 735255714Skris 735355714Skris *) Make sure `make rehash' target really finds the `openssl' program. 735455714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 735555714Skris 735655714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 735755714Skris like to hear about it if this slows down other processors. 735855714Skris [Ben Laurie] 735955714Skris 736055714Skris *) Add CygWin32 platform information to Configure script. 736155714Skris [Alan Batie <batie@aahz.jf.intel.com>] 736255714Skris 736355714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 736455714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 736555714Skris 736655714Skris *) New program nseq to manipulate netscape certificate sequences 736755714Skris [Steve Henson] 736855714Skris 736955714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 737055714Skris few typos. 737155714Skris [Steve Henson] 737255714Skris 737355714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 737455714Skris but the BN code had some problems that would cause failures when 737555714Skris doing certificate verification and some other functions. 737655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 737755714Skris 737855714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 737955714Skris [Steve Henson] 738055714Skris 738155714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 738255714Skris [Steve Henson] 738355714Skris 738455714Skris *) Add several PKIX and private extended key usage OIDs. 738555714Skris [Steve Henson] 738655714Skris 738755714Skris *) Modify the 'ca' program to handle the new extension code. Modify 738855714Skris openssl.cnf for new extension format, add comments. 738955714Skris [Steve Henson] 739055714Skris 739155714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 739255714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 739355714Skris CA extensions. 739455714Skris [Steve Henson] 739555714Skris 739655714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 739755714Skris error code, add initial support to X509_print() and x509 application. 739855714Skris [Steve Henson] 739955714Skris 740055714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 740155714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 740255714Skris stuff is currently isolated and isn't even compiled yet. 740355714Skris [Steve Henson] 740455714Skris 740555714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 740655714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 740755714Skris Removed the versions check from X509 routines when loading extensions: 740855714Skris this allows certain broken certificates that don't set the version 740955714Skris properly to be processed. 741055714Skris [Steve Henson] 741155714Skris 741255714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 741355714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 741455714Skris can still be regenerated with "make depend". 741555714Skris [Ben Laurie] 741655714Skris 741755714Skris *) Spelling mistake in C version of CAST-128. 741855714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 741955714Skris 742055714Skris *) Changes to the error generation code. The perl script err-code.pl 742155714Skris now reads in the old error codes and retains the old numbers, only 742255714Skris adding new ones if necessary. It also only changes the .err files if new 742355714Skris codes are added. The makefiles have been modified to only insert errors 742455714Skris when needed (to avoid needlessly modifying header files). This is done 742555714Skris by only inserting errors if the .err file is newer than the auto generated 742655714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 742755714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 742855714Skris or delete all the .err files. 742955714Skris [Steve Henson] 743055714Skris 743155714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 743255714Skris been fixed, but is untested. The assembler versions are also fixed, but 743355714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 743455714Skris to regenerate it if needed. 743555714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 743655714Skris Hagino <itojun@kame.net>] 743755714Skris 743855714Skris *) File was opened incorrectly in randfile.c. 743955714Skris [Ulf M�ller <ulf@fitug.de>] 744055714Skris 744155714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 744255714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 744355714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 744455714Skris al: it's just almost always a UTCTime. Note this patch adds new error 744555714Skris codes so do a "make errors" if there are problems. 744655714Skris [Steve Henson] 744755714Skris 744855714Skris *) Correct Linux 1 recognition in config. 744955714Skris [Ulf M�ller <ulf@fitug.de>] 745055714Skris 745155714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 745255714Skris [Anonymous <nobody@replay.com>] 745355714Skris 745455714Skris *) Generate an error if given an empty string as a cert directory. Also 745555714Skris generate an error if handed NULL (previously returned 0 to indicate an 745655714Skris error, but didn't set one). 745755714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 745855714Skris 745955714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 746055714Skris [Ben Laurie] 746155714Skris 746255714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 746355714Skris parameters. This was causing a warning which killed off the Win32 compile. 746455714Skris [Steve Henson] 746555714Skris 746655714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 746755714Skris [Neil Costigan <neil.costigan@celocom.com>] 746855714Skris 746955714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 747055714Skris based on a text string, looking up short and long names and finally 747155714Skris "dot" format. The "dot" format stuff didn't work. Added new function 747255714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 747355714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 747455714Skris OID is not part of the table. 747555714Skris [Steve Henson] 747655714Skris 747755714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 747855714Skris X509_LOOKUP_by_alias(). 747955714Skris [Ben Laurie] 748055714Skris 748155714Skris *) Sort openssl functions by name. 748255714Skris [Ben Laurie] 748355714Skris 748455714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 748555714Skris encryption from sample DSA keys (in case anyone is interested the password 748655714Skris was "1234"). 748755714Skris [Steve Henson] 748855714Skris 748955714Skris *) Make _all_ *_free functions accept a NULL pointer. 749055714Skris [Frans Heymans <fheymans@isaserver.be>] 749155714Skris 749255714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 749355714Skris NULL pointers. 749455714Skris [Anonymous <nobody@replay.com>] 749555714Skris 749655714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 749755714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 749855714Skris 749955714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 750055714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 750155714Skris 750255714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 750355714Skris [Anonymous <nobody@replay.com>] 750455714Skris 750555714Skris *) Add prototype for temp key callback functions 750655714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 750755714Skris [Ben Laurie] 750855714Skris 750955714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 751055714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 751155714Skris [Steve Henson] 751255714Skris 751355714Skris *) X509_name_add_entry() freed the wrong thing after an error. 751455714Skris [Arne Ansper <arne@ats.cyber.ee>] 751555714Skris 751655714Skris *) rsa_eay.c would attempt to free a NULL context. 751755714Skris [Arne Ansper <arne@ats.cyber.ee>] 751855714Skris 751955714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 752055714Skris [Arne Ansper <arne@ats.cyber.ee>] 752155714Skris 752255714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 752355714Skris [Arne Ansper <arne@ats.cyber.ee>] 752455714Skris 752555714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 752655714Skris in X509_STORE_new(), but document the fact that this variable is still 752755714Skris unused in the certificate verification process. 752855714Skris [Ralf S. Engelschall] 752955714Skris 753055714Skris *) Fix the various library and apps files to free up pkeys obtained from 753155714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 753255714Skris [Steve Henson] 753355714Skris 753455714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 753555714Skris demos/maurice/example2.c work, amongst others, probably. 753655714Skris [Steve Henson and Ben Laurie] 753755714Skris 753855714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 753955714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 754055714Skris are no longer created. This way we have a single and consistent command 754155714Skris line interface `openssl <command>', similar to `cvs <command>'. 754255714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 754355714Skris 754455714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 754555714Skris BIT STRING wrapper always have zero unused bits. 754655714Skris [Steve Henson] 754755714Skris 754855714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 754955714Skris [Steve Henson] 755055714Skris 755155714Skris *) Make the top-level INSTALL documentation easier to understand. 755255714Skris [Paul Sutton] 755355714Skris 755455714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 755555714Skris make (including if user presses ^C) [Paul Sutton] 755655714Skris 755755714Skris *) Make Montgomery context stuff explicit in RSA data structure. 755855714Skris [Ben Laurie] 755955714Skris 756055714Skris *) Fix build order of pem and err to allow for generated pem.h. 756155714Skris [Ben Laurie] 756255714Skris 756355714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 756455714Skris [Ben Laurie] 756555714Skris 756655714Skris *) Enhanced the err-ins.pl script so it makes the error library number 756755714Skris global and can add a library name. This is needed for external ASN1 and 756855714Skris other error libraries. 756955714Skris [Steve Henson] 757055714Skris 757155714Skris *) Fixed sk_insert which never worked properly. 757255714Skris [Steve Henson] 757355714Skris 757455714Skris *) Fix ASN1 macros so they can handle indefinite length construted 757555714Skris EXPLICIT tags. Some non standard certificates use these: they can now 757655714Skris be read in. 757755714Skris [Steve Henson] 757855714Skris 757955714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 758055714Skris into a single doc/ssleay.txt bundle. This way the information is still 758155714Skris preserved but no longer messes up this directory. Now it's new room for 758255714Skris the new set of documenation files. 758355714Skris [Ralf S. Engelschall] 758455714Skris 758555714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 758655714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 758755714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 758855714Skris number of arguments. 758955714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 759055714Skris 759155714Skris *) Fix test data to work with the above. 759255714Skris [Ben Laurie] 759355714Skris 759455714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 759555714Skris was already fixed by Eric for 0.9.1 it seems. 759655714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 759755714Skris 759855714Skris *) Autodetect FreeBSD3. 759955714Skris [Ben Laurie] 760055714Skris 760155714Skris *) Fix various bugs in Configure. This affects the following platforms: 760255714Skris nextstep 760355714Skris ncr-scde 760455714Skris unixware-2.0 760555714Skris unixware-2.0-pentium 760655714Skris sco5-cc. 760755714Skris [Ben Laurie] 760855714Skris 760955714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 761055714Skris before they are needed. 761155714Skris [Ben Laurie] 761255714Skris 761355714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 761455714Skris [Ben Laurie] 761555714Skris 761655714Skris 761755714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 761855714Skris 761955714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 762055714Skris changed SSLeay to OpenSSL in version strings. 762155714Skris [Ralf S. Engelschall] 762255714Skris 762355714Skris *) Some fixups to the top-level documents. 762455714Skris [Paul Sutton] 762555714Skris 762655714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 762755714Skris because the symlink to include/ was missing. 762855714Skris [Ralf S. Engelschall] 762955714Skris 763055714Skris *) Incorporated the popular no-RSA/DSA-only patches 763155714Skris which allow to compile a RSA-free SSLeay. 763255714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 763355714Skris 763455714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 763555714Skris when "ssleay" is still not found. 763655714Skris [Ralf S. Engelschall] 763755714Skris 763855714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 763955714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 764055714Skris 764155714Skris *) Updated the README file. 764255714Skris [Ralf S. Engelschall] 764355714Skris 764455714Skris *) Added various .cvsignore files in the CVS repository subdirs 764555714Skris to make a "cvs update" really silent. 764655714Skris [Ralf S. Engelschall] 764755714Skris 764855714Skris *) Recompiled the error-definition header files and added 764955714Skris missing symbols to the Win32 linker tables. 765055714Skris [Ralf S. Engelschall] 765155714Skris 765255714Skris *) Cleaned up the top-level documents; 765355714Skris o new files: CHANGES and LICENSE 765455714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 765555714Skris o merged COPYRIGHT into LICENSE 765655714Skris o removed obsolete TODO file 765755714Skris o renamed MICROSOFT to INSTALL.W32 765855714Skris [Ralf S. Engelschall] 765955714Skris 766055714Skris *) Removed dummy files from the 0.9.1b source tree: 766155714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 766255714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 766355714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 766455714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 766555714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 766655714Skris [Ralf S. Engelschall] 766755714Skris 766855714Skris *) Added various platform portability fixes. 766955714Skris [Mark J. Cox] 767055714Skris 767155714Skris *) The Genesis of the OpenSSL rpject: 767255714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 767355714Skris Young and Tim J. Hudson created while they were working for C2Net until 767455714Skris summer 1998. 767555714Skris [The OpenSSL Project] 767655714Skris 767755714Skris 767855714Skris Changes between 0.9.0b and 0.9.1b [not released] 767955714Skris 768055714Skris *) Updated a few CA certificates under certs/ 768155714Skris [Eric A. Young] 768255714Skris 768355714Skris *) Changed some BIGNUM api stuff. 768455714Skris [Eric A. Young] 768555714Skris 768655714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 768755714Skris DGUX x86, Linux Alpha, etc. 768855714Skris [Eric A. Young] 768955714Skris 769055714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 769155714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 769255714Skris available). 769355714Skris [Eric A. Young] 769455714Skris 769555714Skris *) Add -strparse option to asn1pars program which parses nested 769655714Skris binary structures 769755714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 769855714Skris 769955714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 770055714Skris [Eric A. Young] 770155714Skris 770255714Skris *) DSA fix for "ca" program. 770355714Skris [Eric A. Young] 770455714Skris 770555714Skris *) Added "-genkey" option to "dsaparam" program. 770655714Skris [Eric A. Young] 770755714Skris 770855714Skris *) Added RIPE MD160 (rmd160) message digest. 770955714Skris [Eric A. Young] 771055714Skris 771155714Skris *) Added -a (all) option to "ssleay version" command. 771255714Skris [Eric A. Young] 771355714Skris 771455714Skris *) Added PLATFORM define which is the id given to Configure. 771555714Skris [Eric A. Young] 771655714Skris 771755714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 771855714Skris [Eric A. Young] 771955714Skris 772055714Skris *) Extended the ASN.1 parser routines. 772155714Skris [Eric A. Young] 772255714Skris 772355714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 772455714Skris [Eric A. Young] 772555714Skris 772655714Skris *) Added a BN_CTX to the BN library. 772755714Skris [Eric A. Young] 772855714Skris 772955714Skris *) Fixed the weak key values in DES library 773055714Skris [Eric A. Young] 773155714Skris 773255714Skris *) Changed API in EVP library for cipher aliases. 773355714Skris [Eric A. Young] 773455714Skris 773555714Skris *) Added support for RC2/64bit cipher. 773655714Skris [Eric A. Young] 773755714Skris 773855714Skris *) Converted the lhash library to the crypto/mem.c functions. 773955714Skris [Eric A. Young] 774055714Skris 774155714Skris *) Added more recognized ASN.1 object ids. 774255714Skris [Eric A. Young] 774355714Skris 774455714Skris *) Added more RSA padding checks for SSL/TLS. 774555714Skris [Eric A. Young] 774655714Skris 774755714Skris *) Added BIO proxy/filter functionality. 774855714Skris [Eric A. Young] 774955714Skris 775055714Skris *) Added extra_certs to SSL_CTX which can be used 775155714Skris send extra CA certificates to the client in the CA cert chain sending 775255714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 775355714Skris [Eric A. Young] 775455714Skris 775555714Skris *) Now Fortezza is denied in the authentication phase because 775655714Skris this is key exchange mechanism is not supported by SSLeay at all. 775755714Skris [Eric A. Young] 775855714Skris 775955714Skris *) Additional PKCS1 checks. 776055714Skris [Eric A. Young] 776155714Skris 776255714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 776355714Skris [Eric A. Young] 776455714Skris 776555714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 776655714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 776755714Skris [Eric A. Young] 776855714Skris 776955714Skris *) Fixed a few memory leaks. 777055714Skris [Eric A. Young] 777155714Skris 777255714Skris *) Fixed various code and comment typos. 777355714Skris [Eric A. Young] 777455714Skris 777555714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 777655714Skris bytes sent in the client random. 777755714Skris [Edward Bishop <ebishop@spyglass.com>] 777855714Skris 7779