CHANGES revision 167612
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5167612Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 6167612Ssimon 7167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 8167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 9167612Ssimon authentication-only ciphersuites. 10167612Ssimon [Bodo Moeller] 11167612Ssimon 12167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 13167612Ssimon Camellia256) share a single mask bit in the logic of 14167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 15167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 16167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 17167612Ssimon [Victor Duchovni] 18167612Ssimon 19167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 20167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 21167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 22167612Ssimon prevent the removal of trailing zero bits to get the proper DER 23167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 24167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 25167612Ssimon [Bodo Moeller] 26167612Ssimon 27167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 28167612Ssimon protocol version while receiving ClientHello even if the 29167612Ssimon ClientHello is fragmented. (The server can't insist on the 30167612Ssimon particular protocol version it has chosen before the ServerHello 31167612Ssimon message has informed the client about his choice.) 32167612Ssimon [Bodo Moeller] 33167612Ssimon 34167612Ssimon *) Add RFC 3779 support. 35167612Ssimon [Rob Austein for ARIN, Ben Laurie] 36167612Ssimon 37167612Ssimon *) Load error codes if they are not already present instead of using a 38167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 39167612Ssimon Improve header file function name parsing. 40167612Ssimon [Steve Henson] 41167612Ssimon 42167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 43167612Ssimon or CAPABILITY handshake as required by RFCs. 44167612Ssimon [Goetz Babin-Ebell] 45167612Ssimon 46162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 47162911Ssimon 48162911Ssimon *) Introduce limits to prevent malicious keys being able to 49162911Ssimon cause a denial of service. (CVE-2006-2940) 50162911Ssimon [Steve Henson, Bodo Moeller] 51162911Ssimon 52162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 53162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 54162911Ssimon 55162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 56162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 57162911Ssimon 58162911Ssimon *) Fix SSL client code which could crash if connecting to a 59162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 60162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 61162911Ssimon 62162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 63162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 64162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 65162911Ssimon the same strength classification in 0.9.7h) as we currently only 66162911Ssimon have a single AES bit in the ciphersuite description bitmap. 67162911Ssimon That change, however, also applied to ciphersuite strings such as 68162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 69162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 70162911Ssimon from SSL 3.0/TLS 1.0. 71162911Ssimon 72162911Ssimon So we change the selection algorithm again: Naming an explicit 73162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 74162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 75162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 76162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 77162911Ssimon 78162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 79162911Ssimon 128/256 bit distinction would be relevant, this works for now. 80162911Ssimon The proper fix will be to use different bits for AES128 and 81162911Ssimon AES256, which would have avoided the problems from the beginning; 82162911Ssimon however, bits are scarce, so we can only do this in a new release 83162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 84162911Ssimon definition to split the single 'unsigned long mask' bitmap into 85162911Ssimon multiple values to extend the available space. 86162911Ssimon 87162911Ssimon [Bodo Moeller] 88162911Ssimon 89162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 90162911Ssimon 91162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 92162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 93162911Ssimon 94162911Ssimon *) Add AES IGE and biIGE modes. 95162911Ssimon [Ben Laurie] 96162911Ssimon 97162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 98162911Ssimon possible instead of select(), since the latter has some 99162911Ssimon undesirable limitations. 100162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 101162911Ssimon 102162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 103162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 104162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 105162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 106162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 107162911Ssimon support, which is required for curve and point format negotiation 108162911Ssimon to avoid potential handshake problems. 109162911Ssimon [Bodo Moeller] 110162911Ssimon 111162911Ssimon *) Disable rogue ciphersuites: 112162911Ssimon 113162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 114162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 115162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 116162911Ssimon 117162911Ssimon The latter two were purportedly from 118162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 119162911Ssimon appear there. 120162911Ssimon 121167612Ssimon Also deactivate the remaining ciphersuites from 122162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 123162911Ssimon unofficial, and the ID has long expired. 124162911Ssimon [Bodo Moeller] 125162911Ssimon 126162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 127162911Ssimon dual-core machines) and other potential thread-safety issues. 128162911Ssimon [Bodo Moeller] 129162911Ssimon 130162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 131162911Ssimon versions), which is now available for royalty-free use 132162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 133162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 134162911Ssimon 135162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 136162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 137162911Ssimon is configured with 'enable-camellia'. 138162911Ssimon [NTT] 139162911Ssimon 140162911Ssimon *) Disable the padding bug check when compression is in use. The padding 141162911Ssimon bug check assumes the first packet is of even length, this is not 142162911Ssimon necessarily true if compresssion is enabled and can result in false 143162911Ssimon positives causing handshake failure. The actual bug test is ancient 144162911Ssimon code so it is hoped that implementations will either have fixed it by 145162911Ssimon now or any which still have the bug do not support compression. 146162911Ssimon [Steve Henson] 147162911Ssimon 148160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 149160814Ssimon 150160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 151160814Ssimon cipher suite and only match that one cipher suite if it is. 152160814Ssimon [Steve Henson] 153160814Ssimon 154160814Ssimon *) Link in manifests for VC++ if needed. 155160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 156160814Ssimon 157160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 158160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 159160814Ssimon TLS extensions, which are supported starting with the 0.9.9 160160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 161160814Ssimon [Douglas Stebila] 162160814Ssimon 163160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 164160814Ssimon opaque EVP_CIPHER_CTX handling. 165160814Ssimon [Steve Henson] 166160814Ssimon 167160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 168160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 169160814Ssimon to conform with the standards mentioned here: 170160814Ssimon http://www.zlib.net/DLL_FAQ.txt 171160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 172160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 173160814Ssimon of the headers and library. Gracefully handle case where zlib library 174160814Ssimon can't be loaded. 175160814Ssimon [Steve Henson] 176160814Ssimon 177160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 178160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 179160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 180160814Ssimon non standard OBJ_obj2txt() behaviour. 181160814Ssimon [Steve Henson] 182160814Ssimon 183160814Ssimon *) Add support for building of engines under engine/ as shared libraries 184160814Ssimon under VC++ build system. 185160814Ssimon [Steve Henson] 186160814Ssimon 187160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 188160814Ssimon Hopefully, we will not see any false combination of paths any more. 189160814Ssimon [Richard Levitte] 190160814Ssimon 191160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 192160814Ssimon 193160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 194160814Ssimon (part of SSL_OP_ALL). This option used to disable the 195160814Ssimon countermeasure against man-in-the-middle protocol-version 196160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 197160814Ssimon idea. (CVE-2005-2969) 198160814Ssimon 199160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 200160814Ssimon for Information Security, National Institute of Advanced Industrial 201160814Ssimon Science and Technology [AIST], Japan)] 202160814Ssimon 203160814Ssimon *) Add two function to clear and return the verify parameter flags. 204160814Ssimon [Steve Henson] 205160814Ssimon 206160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 207160814Ssimon runtime, thus removing the need for a lock. 208160814Ssimon [Nils Larsch] 209160814Ssimon 210160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 211160814Ssimon [Nick Mathewson and Ben Laurie] 212160814Ssimon 213160814Ssimon *) Add functions for well-known primes. 214160814Ssimon [Nick Mathewson] 215160814Ssimon 216160814Ssimon *) Extended Windows CE support. 217160814Ssimon [Satoshi Nakamura and Andy Polyakov] 218160814Ssimon 219160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 220160814Ssimon runtime, thus removing the need for a lock. 221160814Ssimon [Steve Henson] 222160814Ssimon 223160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 224160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 225160814Ssimon smime utility. 226160814Ssimon [Steve Henson] 227160814Ssimon 228160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 229160814Ssimon 230162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 231162911Ssimon OpenSSL 0.9.8.] 232162911Ssimon 233160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 234160814Ssimon [Richard Levitte] 235160814Ssimon 236160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 237160814Ssimon key into the same file any more. 238160814Ssimon [Richard Levitte] 239160814Ssimon 240160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 241160814Ssimon [Andy Polyakov] 242160814Ssimon 243160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 244160814Ssimon [Stefan <stf@udoma.org] 245160814Ssimon 246160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 247160814Ssimon libraries. Use DES_crypt(). 248160814Ssimon [Richard Levitte] 249160814Ssimon 250160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 251160814Ssimon involves renaming the source and generated shared-libs for 252160814Ssimon both. The engines will accept the corrected or legacy ids 253160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 254160814Ssimon this only applies when building 'shared'. 255160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 256160814Ssimon 257160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 258160814Ssimon PKCS12_create() to recognize a CSP name attribute and 259160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 260160814Ssimon [Steve Henson] 261160814Ssimon 262160814Ssimon *) Add new functionality to the bn blinding code: 263160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 264160814Ssimon a fixed number of uses (currently 32) 265160814Ssimon - add new function for parameter creation 266160814Ssimon - introduce flags to control the update behaviour of the 267160814Ssimon BN_BLINDING parameters 268160814Ssimon - hide BN_BLINDING structure 269160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 270160814Ssimon performance when a single RSA object is shared among several 271160814Ssimon threads. 272160814Ssimon [Nils Larsch] 273160814Ssimon 274160814Ssimon *) Add support for DTLS. 275160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 276160814Ssimon 277160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 278160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 279160814Ssimon [Walter Goulet] 280160814Ssimon 281160814Ssimon *) Remove buggy and incompletet DH cert support from 282160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 283160814Ssimon [Nils Larsch] 284160814Ssimon 285160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 286160814Ssimon the apps/openssl applications. 287160814Ssimon [Nils Larsch] 288160814Ssimon 289160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 290160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 291160814Ssimon DEBUG_SAFESTACK must also be set. 292160814Ssimon [Ben Laurie] 293160814Ssimon 294160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 295160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 296160814Ssimon 297160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 298160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 299160814Ssimon 300160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 301160814Ssimon is frequently required for interoperability, and there is no license 302160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 303160814Ssimon avoid this algorithm.) 304160814Ssimon 305160814Ssimon [Bodo Moeller] 306160814Ssimon 307160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 308160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 309160814Ssimon EGEE (Enabling Grids for E-science in Europe). 310160814Ssimon [Richard Levitte] 311160814Ssimon 312160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 313160814Ssimon as Intel P4, IA-64 and AMD64. 314160814Ssimon [Andy Polyakov] 315160814Ssimon 316160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 317160814Ssimon section number in a pod file instead of having to treat each file as 318160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 319160814Ssimon pod file: 320160814Ssimon 321160814Ssimon =for comment openssl_section:XXX 322160814Ssimon 323160814Ssimon The blank line is mandatory. 324160814Ssimon 325160814Ssimon [Steve Henson] 326160814Ssimon 327160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 328160814Ssimon to allow alternative format key and certificate files and passphrase 329160814Ssimon sources. 330160814Ssimon [Steve Henson] 331160814Ssimon 332160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 333160814Ssimon update associated structures and add various utility functions. 334160814Ssimon 335160814Ssimon Add new policy related verify parameters, include policy checking in 336160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 337160814Ssimon to support policy checking and print out. 338160814Ssimon [Steve Henson] 339160814Ssimon 340160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 341160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 342160814Ssimon as well as RNG (though RNG support is currently disabled). 343160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 344160814Ssimon 345160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 346160814Ssimon [Geoff Thorpe] 347160814Ssimon 348160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 349160814Ssimon [Andy Polyakov and a number of other people] 350160814Ssimon 351160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 352160814Ssimon implementation contributed by IBM. 353160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 354160814Ssimon 355160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 356160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 357160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 358160814Ssimon [Jelte Jansen, Geoff Thorpe] 359160814Ssimon 360160814Ssimon *) Functionality for creating the initial serial number file is now 361160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 362160814Ssimon 363160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 364160814Ssimon number file to 1, which is bound to cause problems. To avoid 365160814Ssimon the problems while respecting compatibility between different 0.9.7 366160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 367160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 368160814Ssimon we can fix the problem directly in the 'ca' utility.) 369160814Ssimon [Steve Henson] 370160814Ssimon 371160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 372160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 373160814Ssimon give fewer recursive includes, which could break lazy source code - so 374160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 375160814Ssimon developers should define this symbol when building and using openssl to 376160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 377160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 378160814Ssimon [Geoff Thorpe] 379160814Ssimon 380160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 381160814Ssimon [Steve Henson] 382160814Ssimon 383160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 384160814Ssimon This will generate a random key of the appropriate length based on the 385160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 386160814Ssimon routine to support keys of a specific form. This is used in the des and 387160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 388160814Ssimon code to use new functions and hence generate correct parity DES keys. 389160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 390160814Ssimon valid (weak or incorrect parity). 391160814Ssimon [Steve Henson] 392160814Ssimon 393160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 394160814Ssimon as looking them up. This is useful when the verified structure may contain 395160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 396160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 397160814Ssimon [Steve Henson] 398160814Ssimon 399160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 400160814Ssimon syntax: 401160814Ssimon 402160814Ssimon shortName = some long name, 1.2.3.4 403160814Ssimon [Steve Henson] 404160814Ssimon 405160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 406160814Ssimon limitation on the number of variables it can handle nor the depth of the 407160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 408160814Ssimon information can now expand as required, and rather than having a single 409160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 410160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 411160814Ssimon BN_CTX's "bundling". 412160814Ssimon [Geoff Thorpe] 413160814Ssimon 414160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 415160814Ssimon to allow all RSA operations to function using a single BN_CTX. 416160814Ssimon [Geoff Thorpe] 417160814Ssimon 418160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 419160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 420160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 421160814Ssimon [Steve Henson] 422160814Ssimon 423160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 424160814Ssimon remained unused and not that useful. A variety of other little bignum 425160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 426160814Ssimon below). 427160814Ssimon [Geoff Thorpe] 428160814Ssimon 429160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 430160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 431160814Ssimon [Richard Levitte] 432160814Ssimon 433160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 434160814Ssimon and this should never fail. So the return value from the use of 435160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 436160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 437160814Ssimon [Geoff Thorpe] 438160814Ssimon 439160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 440160814Ssimon initialised value as BN_new(). 441160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 442160814Ssimon 443160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 444160814Ssimon [Steve Henson] 445160814Ssimon 446160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 447160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 448160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 449160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 450160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 451160814Ssimon structures to try and expose faulty code further on. For now, openssl will 452160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 453160814Ssimon forms that it has tolerated in the past, but authors and packagers should 454160814Ssimon consider trying openssl and their own applications when compiled with 455160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 456160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 457160814Ssimon some point, these tighter rules will become openssl's default to improve 458160814Ssimon maintainability, though the assert()s and other overheads will remain only 459160814Ssimon in debugging configurations. See bn.h for more details. 460160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 461160814Ssimon 462160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 463160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 464160814Ssimon initialises it). The presence of this function only made it possible 465160814Ssimon to overwrite an existing structure (and cause memory leaks). 466160814Ssimon [Geoff Thorpe] 467160814Ssimon 468160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 469160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 470160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 471160814Ssimon to clean up those corresponding objects before destroying the hash table 472160814Ssimon (and losing the object pointers). So some over-zealous constifications in 473160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 474160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 475160814Ssimon prototyped to have "const" restrictions on the object pointers they are 476160814Ssimon given (and so aren't required to cast them away any more). 477160814Ssimon [Geoff Thorpe] 478160814Ssimon 479160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 480160814Ssimon (speed) prefers to use its own implementation. The two implementations 481160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 482160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 483160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 484160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 485160814Ssimon internally to the implementation so I've used that for now. 486160814Ssimon [Geoff Thorpe] 487160814Ssimon 488160814Ssimon *) Ensure that deprecated functions do not get compiled when 489160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 490160814Ssimon the self-tests were still using deprecated key-generation functions so 491160814Ssimon these have been updated also. 492160814Ssimon [Geoff Thorpe] 493160814Ssimon 494160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 495160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 496160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 497160814Ssimon digestedData type. Add additional code to correctly generate the 498160814Ssimon digestedData type and add support for this type in PKCS7 initialization 499160814Ssimon functions. 500160814Ssimon [Steve Henson] 501160814Ssimon 502160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 503160814Ssimon structure of type "other". 504160814Ssimon [Steve Henson] 505160814Ssimon 506160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 507160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 508160814Ssimon modulus operations are not performed. The (pre-generated) prime 509160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 510160814Ssimon re-generated on some platforms because of the "division by zero" 511160814Ssimon situation in the script. 512160814Ssimon [Ralf S. Engelschall] 513160814Ssimon 514160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 515160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 516160814Ssimon SHA-1 now is only used for "small" curves (where the 517160814Ssimon representation of a field element takes up to 24 bytes); for 518160814Ssimon larger curves, the field element resulting from ECDH is directly 519160814Ssimon used as premaster secret. 520160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 521160814Ssimon 522160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 523160814Ssimon curve secp160r1 to the tests. 524160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 525160814Ssimon 526160814Ssimon *) Add the possibility to load symbols globally with DSO. 527160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 528160814Ssimon 529160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 530160814Ssimon control of the error stack. 531160814Ssimon [Richard Levitte] 532160814Ssimon 533160814Ssimon *) Add support for STORE in ENGINE. 534160814Ssimon [Richard Levitte] 535160814Ssimon 536160814Ssimon *) Add the STORE type. The intention is to provide a common interface 537160814Ssimon to certificate and key stores, be they simple file-based stores, or 538160814Ssimon HSM-type store, or LDAP stores, or... 539160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 540160814Ssimon [Richard Levitte] 541160814Ssimon 542160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 543160814Ssimon pass a list of arguments to any function as well as provide a way 544160814Ssimon for a function to pass data back to the caller. 545160814Ssimon [Richard Levitte] 546160814Ssimon 547160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 548160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 549160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 550160814Ssimon a memory area. 551160814Ssimon [Richard Levitte] 552160814Ssimon 553160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 554160814Ssimon return an index to an element even if an exact match couldn't be 555160814Ssimon found. The index is guaranteed to point at the element where the 556160814Ssimon searched-for key would be inserted to preserve sorting order. 557160814Ssimon [Richard Levitte] 558160814Ssimon 559160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 560160814Ssimon takes an extra flags argument for optional functionality. Currently, 561160814Ssimon the following flags are defined: 562160814Ssimon 563160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 564160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 565160814Ssimon element where the comparing function returns a negative or zero 566160814Ssimon number. 567160814Ssimon 568160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 569160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 570160814Ssimon element where the comparing function returns zero. This is useful 571160814Ssimon if there are more than one element where the comparing function 572160814Ssimon returns zero. 573160814Ssimon [Richard Levitte] 574160814Ssimon 575160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 576160814Ssimon in such a way that the self-signed certificate becomes part of the 577160814Ssimon CA database and uses the same mechanisms for serial number generation 578160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 579160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 580160814Ssimon [Richard Levitte] 581160814Ssimon 582160814Ssimon *) Add functionality to check the public key of a certificate request 583160814Ssimon against a given private. This is useful to check that a certificate 584160814Ssimon request can be signed by that key (self-signing). 585160814Ssimon [Richard Levitte] 586160814Ssimon 587160814Ssimon *) Make it possible to have multiple active certificates with the same 588160814Ssimon subject in the CA index file. This is done only if the keyword 589160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 590160814Ssimon if 'CA_default') of the configuration file. The value is saved 591160814Ssimon with the database itself in a separate index attribute file, 592160814Ssimon named like the index file with '.attr' appended to the name. 593160814Ssimon [Richard Levitte] 594160814Ssimon 595160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 596160814Ssimon req and dirName. 597160814Ssimon [Steve Henson] 598160814Ssimon 599160814Ssimon *) Support for nameConstraints certificate extension. 600160814Ssimon [Steve Henson] 601160814Ssimon 602160814Ssimon *) Support for policyConstraints certificate extension. 603160814Ssimon [Steve Henson] 604160814Ssimon 605160814Ssimon *) Support for policyMappings certificate extension. 606160814Ssimon [Steve Henson] 607160814Ssimon 608160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 609160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 610160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 611160814Ssimon indirection. This lets alternative implementations fallback to the 612160814Ssimon default implementation more easily. 613160814Ssimon [Geoff Thorpe] 614160814Ssimon 615160814Ssimon *) Support for directoryName in GeneralName related extensions 616160814Ssimon in config files. 617160814Ssimon [Steve Henson] 618160814Ssimon 619160814Ssimon *) Make it possible to link applications using Makefile.shared. 620160814Ssimon Make that possible even when linking against static libraries! 621160814Ssimon [Richard Levitte] 622160814Ssimon 623160814Ssimon *) Support for single pass processing for S/MIME signing. This now 624160814Ssimon means that S/MIME signing can be done from a pipe, in addition 625160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 626160814Ssimon and the signed data does not need to be all held in memory. 627160814Ssimon 628160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 629160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 630160814Ssimon is done after the data is output (and digests calculated) in 631160814Ssimon SMIME_write_PKCS7(). 632160814Ssimon [Steve Henson] 633160814Ssimon 634160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 635160814Ssimon applications, at least on the platforms where it's known how 636160814Ssimon to do it. 637160814Ssimon [Richard Levitte] 638160814Ssimon 639160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 640160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 641160814Ssimon will now compute a table of multiples of the generator that 642160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 643160814Ssimon faster (notably in the case of a single point multiplication, 644160814Ssimon scalar * generator). 645160814Ssimon [Nils Larsch, Bodo Moeller] 646160814Ssimon 647160814Ssimon *) IPv6 support for certificate extensions. The various extensions 648160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 649160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 650160814Ssimon correctly. 651160814Ssimon [Steve Henson] 652160814Ssimon 653160814Ssimon *) Added an ENGINE that implements RSA by performing private key 654160814Ssimon exponentiations with the GMP library. The conversions to and from 655160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 656160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 657160814Ssimon However there are likely to be other architectures where GMP could 658160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 659160814Ssimon specified at Configure time and should be accompanied by the necessary 660160814Ssimon linker additions, eg; 661160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 662160814Ssimon [Geoff Thorpe] 663160814Ssimon 664160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 665160814Ssimon testing availability of engines with "-t" - the old behaviour is 666160814Ssimon produced by increasing the feature's verbosity with "-tt". 667160814Ssimon [Geoff Thorpe] 668160814Ssimon 669160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 670160814Ssimon could be freed. Solution: make sure initialization is performed early 671160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 672160814Ssimon via PR#459) 673160814Ssimon [Lutz Jaenicke] 674160814Ssimon 675160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 676160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 677160814Ssimon software implementations. For DSA and DH, parameter generation can 678160814Ssimon also be overriden by providing the appropriate method callbacks. 679160814Ssimon [Geoff Thorpe] 680160814Ssimon 681160814Ssimon *) Change the "progress" mechanism used in key-generation and 682160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 683160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 684160814Ssimon postfixes and the older functions are reimplemented as wrappers for 685160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 686160814Ssimon declarations of the old functions to help (graceful) attempts to 687160814Ssimon migrate to the new functions. Also, the new key-generation API 688160814Ssimon functions operate on a caller-supplied key-structure and return 689160814Ssimon success/failure rather than returning a key or NULL - this is to 690160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 691160814Ssimon 692160814Ssimon Example for using the new callback interface: 693160814Ssimon 694160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 695160814Ssimon void *my_arg = ...; 696160814Ssimon BN_GENCB my_cb; 697160814Ssimon 698160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 699160814Ssimon 700160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 701160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 702160814Ssimon * documentation of the function that calls the callback. 703160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 704160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 705160814Ssimon * to continue, or 0 to stop. 706160814Ssimon */ 707160814Ssimon 708160814Ssimon [Geoff Thorpe] 709160814Ssimon 710160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 711160814Ssimon available to TLS with the number defined in 712160814Ssimon draft-ietf-tls-compression-04.txt. 713160814Ssimon [Richard Levitte] 714160814Ssimon 715160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 716160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 717160814Ssimon 718160814Ssimon CertificatePair ::= SEQUENCE { 719160814Ssimon forward [0] Certificate OPTIONAL, 720160814Ssimon reverse [1] Certificate OPTIONAL, 721160814Ssimon -- at least one of the pair shall be present -- } 722160814Ssimon 723160814Ssimon Also implement the PEM functions to read and write certificate 724160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 725160814Ssimon 726160814Ssimon This needed to be defined, mostly for the sake of the LDAP 727160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 728160814Ssimon well. 729160814Ssimon [Richard Levitte] 730160814Ssimon 731160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 732160814Ssimon Makefile.shared, for Cygwin's sake. 733160814Ssimon [Richard Levitte] 734160814Ssimon 735160814Ssimon *) Extend the BIGNUM API by creating a function 736160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 737160814Ssimon and a macro that behave like 738160814Ssimon int BN_is_negative(const BIGNUM *a); 739160814Ssimon 740160814Ssimon to avoid the need to access 'a->neg' directly in applications. 741160814Ssimon [Nils Larsch] 742160814Ssimon 743160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 744160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 745160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 746160814Ssimon if applicable. 747160814Ssimon [Nils Larsch <nla@trustcenter.de>] 748160814Ssimon 749160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 750160814Ssimon [Bodo Moeller] 751160814Ssimon 752160814Ssimon *) Change the ENGINE framework to automatically load engines 753160814Ssimon dynamically from specific directories unless they could be 754160814Ssimon found to already be built in or loaded. Move all the 755160814Ssimon current engines except for the cryptodev one to a new 756160814Ssimon directory engines/. 757160814Ssimon The engines in engines/ are built as shared libraries if 758160814Ssimon the "shared" options was given to ./Configure or ./config. 759160814Ssimon Otherwise, they are inserted in libcrypto.a. 760160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 761160814Ssimon engines, but that can be overriden at configure time through 762160814Ssimon the usual use of --prefix and/or --openssldir, and at run 763160814Ssimon time with the environment variable OPENSSL_ENGINES. 764160814Ssimon [Geoff Thorpe and Richard Levitte] 765160814Ssimon 766160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 767160814Ssimon libraries. Addapt Makefile.org. 768160814Ssimon [Richard Levitte] 769160814Ssimon 770160814Ssimon *) Add version info to Win32 DLLs. 771160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 772160814Ssimon 773160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 774160814Ssimon can be added using this API to created arbitrary PKCS#12 775160814Ssimon files while avoiding the low level API. 776160814Ssimon 777160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 778160814Ssimon will then be omitted from the output file. The encryption 779160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 780160814Ssimon iteration count can be set to 0 to omit the mac. 781160814Ssimon 782160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 783160814Ssimon options work when creating a PKCS#12 file. New option -nomac 784160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 785160814Ssimon New code is modified to use the enhanced PKCS12_create() 786160814Ssimon instead of the low level API. 787160814Ssimon [Steve Henson] 788160814Ssimon 789160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 790160814Ssimon encoding. This can output sequences tags and octet strings in 791160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 792160814Ssimon encoding. This is experimental and needs additional code to 793160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 794160814Ssimon PKCS#7 code. 795160814Ssimon 796160814Ssimon Extend template encode functionality so that tagging is passed 797160814Ssimon down to the template encoder. 798160814Ssimon [Steve Henson] 799160814Ssimon 800160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 801160814Ssimon recognized instead of using RSA as a default. 802160814Ssimon [Bodo Moeller] 803160814Ssimon 804160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 805160814Ssimon As these are not official, they are not included in "ALL"; 806160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 807160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 808160814Ssimon 809160814Ssimon *) Add ECDH engine support. 810160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 811160814Ssimon 812160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 813160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 814160814Ssimon 815160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 816160814Ssimon without success (which indicates a broken PRNG). 817160814Ssimon [Bodo Moeller] 818160814Ssimon 819160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 820160814Ssimon is really the square of the return value. (Previously, 821160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 822160814Ssimon [Bodo Moeller] 823160814Ssimon 824160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 825160814Ssimon and WAP/WTLS; add OIDs that were still missing. 826160814Ssimon 827160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 828160814Ssimon (Sun Microsystems Laboratories)] 829160814Ssimon 830160814Ssimon *) Extend the EC library for elliptic curves over binary fields 831160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 832160814Ssimon New EC_METHOD: 833160814Ssimon 834160814Ssimon EC_GF2m_simple_method 835160814Ssimon 836160814Ssimon New API functions: 837160814Ssimon 838160814Ssimon EC_GROUP_new_curve_GF2m 839160814Ssimon EC_GROUP_set_curve_GF2m 840160814Ssimon EC_GROUP_get_curve_GF2m 841160814Ssimon EC_POINT_set_affine_coordinates_GF2m 842160814Ssimon EC_POINT_get_affine_coordinates_GF2m 843160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 844160814Ssimon 845160814Ssimon Point compression for binary fields is disabled by default for 846160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 847160814Ssimon enable it). 848160814Ssimon 849160814Ssimon As binary polynomials are represented as BIGNUMs, various members 850160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 851160814Ssimon between the implementations for prime fields and binary fields; 852160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 853160814Ssimon are essentially identical to their ..._GFp counterparts. 854160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 855160814Ssimon various internal method names.) 856160814Ssimon 857160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 858160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 859160814Ssimon 860160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 861160814Ssimon (Sun Microsystems Laboratories)] 862160814Ssimon 863160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 864160814Ssimon through methods ('mul', 'precompute_mult'). 865160814Ssimon 866160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 867160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 868160814Ssimon methods are undefined. 869160814Ssimon 870160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 871160814Ssimon (Sun Microsystems Laboratories)] 872160814Ssimon 873160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 874160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 875160814Ssimon length of the modulus. 876160814Ssimon 877160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 878160814Ssimon (Sun Microsystems Laboratories)] 879160814Ssimon 880160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 881160814Ssimon (These simply call ..._new and ..._copy). 882160814Ssimon 883160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 884160814Ssimon (Sun Microsystems Laboratories)] 885160814Ssimon 886160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 887160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 888160814Ssimon used) in the following functions [macros]: 889160814Ssimon 890160814Ssimon BN_GF2m_add 891160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 892160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 893160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 894160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 895160814Ssimon BN_GF2m_mod_inv 896160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 897160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 898160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 899160814Ssimon BN_GF2m_cmp [= BN_ucmp] 900160814Ssimon 901160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 902160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 903160814Ssimon 904160814Ssimon For some functions, an the irreducible polynomial defining a 905160814Ssimon field can be given as an 'unsigned int[]' with strictly 906160814Ssimon decreasing elements giving the indices of those bits that are set; 907160814Ssimon i.e., p[] represents the polynomial 908160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 909160814Ssimon where 910160814Ssimon p[0] > p[1] > ... > p[k] = 0. 911160814Ssimon This applies to the following functions: 912160814Ssimon 913160814Ssimon BN_GF2m_mod_arr 914160814Ssimon BN_GF2m_mod_mul_arr 915160814Ssimon BN_GF2m_mod_sqr_arr 916160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 917160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 918160814Ssimon BN_GF2m_mod_exp_arr 919160814Ssimon BN_GF2m_mod_sqrt_arr 920160814Ssimon BN_GF2m_mod_solve_quad_arr 921160814Ssimon BN_GF2m_poly2arr 922160814Ssimon BN_GF2m_arr2poly 923160814Ssimon 924160814Ssimon Conversion can be performed by the following functions: 925160814Ssimon 926160814Ssimon BN_GF2m_poly2arr 927160814Ssimon BN_GF2m_arr2poly 928160814Ssimon 929160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 930160814Ssimon 931160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 932160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 933160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 934160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 935160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 936160814Ssimon 937160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 938160814Ssimon (Sun Microsystems Laboratories)] 939160814Ssimon 940160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 941160814Ssimon functionality is disabled at compile-time. 942160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 943160814Ssimon 944160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 945160814Ssimon information is visible when viewing, e.g., a certificate: 946160814Ssimon 947160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 948160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 949160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 950160814Ssimon avoid the appearance of a printable string. 951160814Ssimon [Nils Larsch <nla@trustcenter.de>] 952160814Ssimon 953160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 954160814Ssimon functions 955160814Ssimon EC_GROUP_set_asn1_flag() 956160814Ssimon EC_GROUP_get_asn1_flag() 957160814Ssimon EC_GROUP_set_point_conversion_form() 958160814Ssimon EC_GROUP_get_point_conversion_form() 959160814Ssimon These control ASN1 encoding details: 960160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 961160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 962160814Ssimon - Points are encoded in uncompressed form by default; options for 963160814Ssimon asn1_for are as for point2oct, namely 964160814Ssimon POINT_CONVERSION_COMPRESSED 965160814Ssimon POINT_CONVERSION_UNCOMPRESSED 966160814Ssimon POINT_CONVERSION_HYBRID 967160814Ssimon 968160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 969160814Ssimon functions 970160814Ssimon EC_GROUP_set_seed() 971160814Ssimon EC_GROUP_get0_seed() 972160814Ssimon EC_GROUP_get_seed_len() 973160814Ssimon This is used only for ASN1 purposes (so far). 974160814Ssimon [Nils Larsch <nla@trustcenter.de>] 975160814Ssimon 976160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 977160814Ssimon of the appropriate field type OID. The new function 978160814Ssimon EC_METHOD_get_field_type() returns this value. 979160814Ssimon [Nils Larsch <nla@trustcenter.de>] 980160814Ssimon 981160814Ssimon *) Add functions 982160814Ssimon EC_POINT_point2bn() 983160814Ssimon EC_POINT_bn2point() 984160814Ssimon EC_POINT_point2hex() 985160814Ssimon EC_POINT_hex2point() 986160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 987160814Ssimon EC_POINT_oct2point(). 988160814Ssimon [Nils Larsch <nla@trustcenter.de>] 989160814Ssimon 990160814Ssimon *) Change internals of the EC library so that the functions 991160814Ssimon EC_GROUP_set_generator() 992160814Ssimon EC_GROUP_get_generator() 993160814Ssimon EC_GROUP_get_order() 994160814Ssimon EC_GROUP_get_cofactor() 995160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 996160814Ssimon to methods, which would lead to unnecessary code duplication when 997160814Ssimon adding different types of curves. 998160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 999160814Ssimon 1000160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 1001160814Ssimon arithmetic, and such that modified wNAFs are generated 1002160814Ssimon (which avoid length expansion in many cases). 1003160814Ssimon [Bodo Moeller] 1004160814Ssimon 1005160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 1006160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 1007160814Ssimon 1008160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 1009160814Ssimon on a EC_GROUP, its generator and order. This includes 1010160814Ssimon EC_GROUP_check_discriminant(). 1011160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1012160814Ssimon 1013160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 1014160814Ssimon 1015160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 1016160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 1017160814Ssimon 1018160814Ssimon ECDSA support is also included in various other files across the 1019160814Ssimon library. Most notably, 1020160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 1021160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 1022160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 1023160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 1024160814Ssimon them suitable for ECDSA where domain parameters must be 1025160814Ssimon extracted before the specific public key; 1026160814Ssimon - ECDSA engine support has been added. 1027160814Ssimon [Nils Larsch <nla@trustcenter.de>] 1028160814Ssimon 1029160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 1030160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 1031160814Ssimon function 1032160814Ssimon EC_GROUP_new_by_curve_name(), 1033160814Ssimon and the list of available named curves can be obtained with 1034160814Ssimon EC_get_builtin_curves(). 1035160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 1036160814Ssimon accessed via 1037160814Ssimon EC_GROUP_set_curve_name() 1038160814Ssimon EC_GROUP_get_curve_name() 1039160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 1040160814Ssimon 1041160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1042160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 1043160814Ssimon required a small change in bn_mul_part_recursive() and the addition 1044160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 1045160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 1046160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 1047160814Ssimon differing sizes. 1048160814Ssimon [Richard Levitte] 1049160814Ssimon 1050167612Ssimon Changes between 0.9.7l and 0.9.7m [xx XXX xxxx] 1051162911Ssimon 1052167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 1053167612Ssimon sensitive data. 1054167612Ssimon [Benjamin Bennett <ben@psc.edu>] 1055167612Ssimon 1056167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 1057167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 1058167612Ssimon authentication-only ciphersuites. 1059167612Ssimon [Bodo Moeller] 1060167612Ssimon 1061167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 1062167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 1063167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 1064167612Ssimon [Victor Duchovni] 1065167612Ssimon 1066167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 1067167612Ssimon protocol version while receiving ClientHello even if the 1068167612Ssimon ClientHello is fragmented. (The server can't insist on the 1069167612Ssimon particular protocol version it has chosen before the ServerHello 1070167612Ssimon message has informed the client about his choice.) 1071167612Ssimon [Bodo Moeller] 1072167612Ssimon 1073167612Ssimon *) Load error codes if they are not already present instead of using a 1074167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 1075167612Ssimon [Steve Henson] 1076167612Ssimon 1077167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 1078167612Ssimon 1079167612Ssimon *) Introduce limits to prevent malicious keys being able to 1080167612Ssimon cause a denial of service. (CVE-2006-2940) 1081167612Ssimon [Steve Henson, Bodo Moeller] 1082167612Ssimon 1083167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 1084167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 1085167612Ssimon 1086167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 1087167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 1088167612Ssimon 1089167612Ssimon *) Fix SSL client code which could crash if connecting to a 1090167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 1091167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 1092167612Ssimon 1093162911Ssimon *) Change ciphersuite string processing so that an explicit 1094162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 1095162911Ssimon will no longer include "AES128-SHA"), and any other similar 1096162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 1097162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 1098162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 1099162911Ssimon changes from 0.9.8b and 0.9.8d. 1100162911Ssimon [Bodo Moeller] 1101162911Ssimon 1102162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 1103162911Ssimon 1104162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 1105162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 1106162911Ssimon 1107162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 1108162911Ssimon possible instead of select(), since the latter has some 1109162911Ssimon undesirable limitations. 1110162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 1111162911Ssimon 1112162911Ssimon *) Disable rogue ciphersuites: 1113162911Ssimon 1114162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 1115162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 1116162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 1117162911Ssimon 1118162911Ssimon The latter two were purportedly from 1119162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 1120162911Ssimon appear there. 1121162911Ssimon 1122162911Ssimon Also deactive the remaining ciphersuites from 1123162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 1124162911Ssimon unofficial, and the ID has long expired. 1125162911Ssimon [Bodo Moeller] 1126162911Ssimon 1127162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 1128162911Ssimon dual-core machines) and other potential thread-safety issues. 1129162911Ssimon [Bodo Moeller] 1130162911Ssimon 1131162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 1132162911Ssimon 1133162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 1134162911Ssimon module in FIPS mode. 1135162911Ssimon [Steve Henson] 1136162911Ssimon 1137162911Ssimon *) Fixes for VC++ 2005 build under Windows. 1138162911Ssimon [Steve Henson] 1139162911Ssimon 1140162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 1141162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 1142162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 1143162911Ssimon build to use fipscanister.o from the GNU make build. 1144162911Ssimon [Steve Henson] 1145162911Ssimon 1146160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 1147160814Ssimon 1148160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 1149160814Ssimon The value now differs depending on if you build for FIPS or not. 1150160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 1151160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 1152160814Ssimon the difference induced by this change. 1153160814Ssimon [Andy Polyakov] 1154160814Ssimon 1155160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 1156160814Ssimon 1157160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 1158160814Ssimon (part of SSL_OP_ALL). This option used to disable the 1159160814Ssimon countermeasure against man-in-the-middle protocol-version 1160160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 1161160814Ssimon idea. (CVE-2005-2969) 1162160814Ssimon 1163160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 1164160814Ssimon for Information Security, National Institute of Advanced Industrial 1165160814Ssimon Science and Technology [AIST], Japan)] 1166160814Ssimon 1167160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 1168160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 1169160814Ssimon [Steve Henson] 1170160814Ssimon 1171160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 1172160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 1173160814Ssimon the information leaked through timing could expose the secret key 1174160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 1175160814Ssimon biased k.) 1176160814Ssimon [Bodo Moeller] 1177160814Ssimon 1178160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 1179160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 1180160814Ssimon squares and multiplies and the memory access pattern are 1181160814Ssimon independent of the particular secret key. This will mitigate 1182160814Ssimon cache-timing and potential related attacks. 1183160814Ssimon 1184160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 1185160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 1186160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 1187160814Ssimon will use this BN flag for private exponents unless the flag 1188160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 1189160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 1190160814Ssimon 1191160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 1192160814Ssimon 1193160814Ssimon *) Change the client implementation for SSLv23_method() and 1194160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 1195160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 1196160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 1197160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 1198160814Ssimon [Bodo Moeller] 1199160814Ssimon 1200160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 1201160814Ssimon clients need. 1202160814Ssimon [Steve Henson] 1203160814Ssimon 1204160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 1205160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 1206160814Ssimon to dsa and dh code (which had race conditions before). 1207160814Ssimon [Steve Henson] 1208160814Ssimon 1209160814Ssimon *) Include the fixed error library code in the C error file definitions 1210160814Ssimon instead of fixing them up at runtime. This keeps the error code 1211160814Ssimon structures constant. 1212160814Ssimon [Steve Henson] 1213160814Ssimon 1214160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 1215160814Ssimon 1216160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 1217160814Ssimon OpenSSL 0.9.8.] 1218160814Ssimon 1219160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 1220160814Ssimon the 'length' field is signed on one version and unsigned on another 1221160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 1222160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 1223160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 1224160814Ssimon some needed definitions. 1225160814Ssimon [Steve Henson] 1226160814Ssimon 1227160814Ssimon *) Undo Cygwin change. 1228160814Ssimon [Ulf M�ller] 1229160814Ssimon 1230160814Ssimon *) Added support for proxy certificates according to RFC 3820. 1231160814Ssimon Because they may be a security thread to unaware applications, 1232160814Ssimon they must be explicitely allowed in run-time. See 1233160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 1234160814Ssimon [Richard Levitte] 1235160814Ssimon 1236160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 1237160814Ssimon 1238160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 1239160814Ssimon server and client random values. Previously 1240160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 1241160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 1242160814Ssimon 1243160814Ssimon This change has negligible security impact because: 1244160814Ssimon 1245160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 1246160814Ssimon data. 1247160814Ssimon 1248160814Ssimon 2. Server and client random values are sent in the clear in the initial 1249160814Ssimon handshake. 1250160814Ssimon 1251160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 1252160814Ssimon size for static RSA ciphersuites) as well as client server and random 1253160814Ssimon values. 1254160814Ssimon 1255160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 1256160814Ssimon to our attention. 1257160814Ssimon 1258160814Ssimon [Stephen Henson, reported by UK NISCC] 1259160814Ssimon 1260160814Ssimon *) Use Windows randomness collection on Cygwin. 1261160814Ssimon [Ulf M�ller] 1262160814Ssimon 1263160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 1264160814Ssimon prematurely by EGD/PRNGD. 1265160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 1266160814Ssimon 1267160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 1268160814Ssimon [Steve Henson] 1269160814Ssimon 1270160814Ssimon *) Back-port of selected performance improvements from development 1271160814Ssimon branch, as well as improved support for PowerPC platforms. 1272160814Ssimon [Andy Polyakov] 1273160814Ssimon 1274160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 1275160814Ssimon failure and freeing up memory if a failure occurs. 1276160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 1277160814Ssimon 1278160814Ssimon *) Add new -passin argument to dgst. 1279160814Ssimon [Steve Henson] 1280160814Ssimon 1281160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 1282160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 1283160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 1284160814Ssimon certificates. 1285160814Ssimon [Steve Henson] 1286160814Ssimon 1287160814Ssimon *) Make an explicit check during certificate validation to see that 1288160814Ssimon the CA setting in each certificate on the chain is correct. As a 1289160814Ssimon side effect always do the following basic checks on extensions, 1290160814Ssimon not just when there's an associated purpose to the check: 1291160814Ssimon 1292160814Ssimon - if there is an unhandled critical extension (unless the user 1293160814Ssimon has chosen to ignore this fault) 1294160814Ssimon - if the path length has been exceeded (if one is set at all) 1295160814Ssimon - that certain extensions fit the associated purpose (if one has 1296160814Ssimon been given) 1297160814Ssimon [Richard Levitte] 1298160814Ssimon 1299142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 1300142425Snectar 1301142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 1302142425Snectar environment. This would happen due to the reordering of the revoked 1303142425Snectar entries during signature checking and serial number lookup. Now the 1304142425Snectar encoding is cached and the serial number sort performed under a lock. 1305142425Snectar Add new STACK function sk_is_sorted(). 1306142425Snectar [Steve Henson] 1307142425Snectar 1308142425Snectar *) Add Delta CRL to the extension code. 1309142425Snectar [Steve Henson] 1310142425Snectar 1311142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 1312142425Snectar [David Holmes <d.holmes@f5.com>] 1313142425Snectar 1314142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 1315142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 1316142425Snectar This is done by creating a random 64 bit value for the initial serial 1317142425Snectar number when a serial number file is created or when a self signed 1318142425Snectar certificate is created using 'openssl req -x509'. The initial serial 1319142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 1320142425Snectar rather than being initialized to 1. 1321142425Snectar [Steve Henson] 1322142425Snectar 1323127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 1324127128Snectar 1325127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 1326160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 1327127128Snectar [Joe Orton, Steve Henson] 1328127128Snectar 1329127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 1330160814Ssimon (CVE-2004-0112) 1331127128Snectar [Joe Orton, Steve Henson] 1332127128Snectar 1333127128Snectar *) Make it possible to have multiple active certificates with the same 1334127128Snectar subject in the CA index file. This is done only if the keyword 1335127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 1336127128Snectar if 'CA_default') of the configuration file. The value is saved 1337127128Snectar with the database itself in a separate index attribute file, 1338127128Snectar named like the index file with '.attr' appended to the name. 1339127128Snectar [Richard Levitte] 1340127128Snectar 1341127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 1342127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 1343127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 1344127128Snectar extensions: since verify currently doesn't process CRL extensions this 1345127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 1346127128Snectar for these cases. 1347127128Snectar [Steve Henson] 1348127128Snectar 1349127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 1350127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 1351127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 1352127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 1353127128Snectar parsing them this should not create any compatibility issues. 1354127128Snectar [Steve Henson] 1355127128Snectar 1356127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 1357127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 1358127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 1359127128Snectar < 0.9.7. 1360127128Snectar [Steve Henson] 1361127128Snectar 1362127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 1363127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 1364127128Snectar 1365127128Snectar *) Use the correct content when signing type "other". 1366127128Snectar [Steve Henson] 1367127128Snectar 1368120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 1369120631Snectar 1370120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 1371120631Snectar 1372120631Snectar Stop out of bounds reads in the ASN1 code when presented with 1373160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 1374120631Snectar 1375160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 1376120631Snectar 1377120631Snectar If verify callback ignores invalid public key errors don't try to check 1378120631Snectar certificate signature with the NULL public key. 1379120631Snectar 1380120631Snectar [Steve Henson] 1381120631Snectar 1382120631Snectar *) New -ignore_err option in ocsp application to stop the server 1383120631Snectar exiting on the first error in a request. 1384120631Snectar [Steve Henson] 1385120631Snectar 1386120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 1387120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 1388120631Snectar specifications. 1389120631Snectar [Steve Henson] 1390120631Snectar 1391120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 1392120631Snectar extra data after the compression methods not only for TLS 1.0 1393120631Snectar but also for SSL 3.0 (as required by the specification). 1394120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 1395120631Snectar 1396120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 1397120631Snectar when it's 512 *bits* long, not 512 bytes. 1398120631Snectar [Richard Levitte] 1399120631Snectar 1400120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 1401120631Snectar blocks during encryption. 1402120631Snectar [Richard Levitte] 1403120631Snectar 1404120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 1405120631Snectar flushes were not handled properly if the BIO retried. On read 1406120631Snectar data was not being buffered properly and had various logic bugs. 1407120631Snectar This also affects blocking I/O when the data being decoded is a 1408120631Snectar certain size. 1409120631Snectar [Steve Henson] 1410120631Snectar 1411120631Snectar *) Various S/MIME bugfixes and compatibility changes: 1412120631Snectar output correct application/pkcs7 MIME type if 1413120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 1414120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 1415120631Snectar of files as .eml work). Correctly handle very long lines in MIME 1416120631Snectar parser. 1417120631Snectar [Steve Henson] 1418120631Snectar 1419120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 1420120631Snectar 1421120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 1422120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 1423120631Snectar a protocol version number mismatch like a decryption error 1424120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 1425120631Snectar [Bodo Moeller] 1426120631Snectar 1427120631Snectar *) Turn on RSA blinding by default in the default implementation 1428120631Snectar to avoid a timing attack. Applications that don't want it can call 1429120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 1430120631Snectar They would be ill-advised to do so in most cases. 1431120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 1432120631Snectar 1433120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 1434120631Snectar seeded (in this case, the secret RSA exponent is abused as 1435120631Snectar an unpredictable seed -- if it is not unpredictable, there 1436120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 1437120631Snectar by remembering the creator's thread ID in rsa->blinding and 1438120631Snectar having all other threads use local one-time blinding factors 1439120631Snectar (this requires more computation than sharing rsa->blinding, but 1440120631Snectar avoids excessive locking; and if an RSA object is not shared 1441120631Snectar between threads, blinding will still be very fast). 1442120631Snectar [Bodo Moeller] 1443120631Snectar 1444120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 1445120631Snectar ENGINE as defaults for all supported algorithms irrespective of 1446120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 1447120631Snectar should make sure they are passing it correctly. 1448120631Snectar [Geoff Thorpe] 1449120631Snectar 1450120631Snectar *) Target "mingw" now allows native Windows code to be generated in 1451120631Snectar the Cygwin environment as well as with the MinGW compiler. 1452120631Snectar [Ulf Moeller] 1453120631Snectar 1454111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 1455111147Snectar 1456111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 1457111147Snectar via timing by performing a MAC computation even if incorrrect 1458111147Snectar block cipher padding has been found. This is a countermeasure 1459111147Snectar against active attacks where the attacker has to distinguish 1460160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 1461111147Snectar 1462111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 1463111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 1464111147Snectar Martin Vuagnoux (EPFL, Ilion)] 1465111147Snectar 1466111147Snectar *) Make the no-err option work as intended. The intention with no-err 1467111147Snectar is not to have the whole error stack handling routines removed from 1468111147Snectar libcrypto, it's only intended to remove all the function name and 1469111147Snectar reason texts, thereby removing some of the footprint that may not 1470111147Snectar be interesting if those errors aren't displayed anyway. 1471111147Snectar 1472111147Snectar NOTE: it's still possible for any application or module to have it's 1473111147Snectar own set of error texts inserted. The routines are there, just not 1474111147Snectar used by default when no-err is given. 1475111147Snectar [Richard Levitte] 1476111147Snectar 1477111147Snectar *) Add support for FreeBSD on IA64. 1478111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 1479111147Snectar 1480111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 1481111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 1482111147Snectar the value returned by DES_cbc_cksum() was like the one from 1483111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 1484111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 1485111147Snectar 1486111147Snectar *) Allow an application to disable the automatic SSL chain building. 1487111147Snectar Before this a rather primitive chain build was always performed in 1488111147Snectar ssl3_output_cert_chain(): an application had no way to send the 1489111147Snectar correct chain if the automatic operation produced an incorrect result. 1490111147Snectar 1491111147Snectar Now the chain builder is disabled if either: 1492111147Snectar 1493111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 1494111147Snectar 1495111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 1496111147Snectar 1497111147Snectar The reasoning behind this is that an application would not want the 1498111147Snectar auto chain building to take place if extra chain certificates are 1499111147Snectar present and it might also want a means of sending no additional 1500111147Snectar certificates (for example the chain has two certificates and the 1501111147Snectar root is omitted). 1502111147Snectar [Steve Henson] 1503111147Snectar 1504111147Snectar *) Add the possibility to build without the ENGINE framework. 1505111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 1506111147Snectar 1507111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 1508111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 1509111147Snectar [Steve Henson] 1510111147Snectar 1511111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 1512111147Snectar could be freed. Solution: make sure initialization is performed early 1513111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 1514111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 1515111147Snectar [Lutz Jaenicke] 1516111147Snectar 1517111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 1518111147Snectar checked on reconnect on the client side, therefore session resumption 1519111147Snectar could still fail with a "ssl session id is different" error. This 1520111147Snectar behaviour is masked when SSL_OP_ALL is used due to 1521111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 1522111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 1523111147Snectar followup to PR #377. 1524111147Snectar [Lutz Jaenicke] 1525111147Snectar 1526111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 1527111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 1528111147Snectar [Andy Polyakov] 1529111147Snectar 1530111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 1531111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 1532111147Snectar the config script, much like the NetBSD support. 1533111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 1534111147Snectar 1535109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 1536109998Smarkm 1537120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 1538120631Snectar OpenSSL 0.9.7.] 1539120631Snectar 1540109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 1541109998Smarkm code (06) was taken as the first octet of the session ID and the last 1542109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 1543109998Smarkm caching could not have worked due to the session ID mismatch between 1544109998Smarkm client and server. 1545109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 1546109998Smarkm PR #377. 1547109998Smarkm [Lutz Jaenicke] 1548109998Smarkm 1549109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 1550109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 1551109998Smarkm removed entirely. 1552109998Smarkm [Richard Levitte] 1553109998Smarkm 1554109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 1555109998Smarkm seems that in spite of existing for more than a year, many application 1556109998Smarkm author have done nothing to provide the necessary callbacks, which 1557109998Smarkm means that this particular engine will not work properly anywhere. 1558109998Smarkm This is a very unfortunate situation which forces us, in the name 1559109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 1560109998Smarkm of libcrypto. 1561109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 1562109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 1563109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 1564109998Smarkm make such changes in the libcrypto locking code that changes will 1565109998Smarkm have to be made anyway). 1566109998Smarkm [Richard Levitte] 1567109998Smarkm 1568109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 1569109998Smarkm octets have been read, EOF or an error occurs. Without this change 1570109998Smarkm some truncated ASN1 structures will not produce an error. 1571109998Smarkm [Steve Henson] 1572109998Smarkm 1573109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 1574109998Smarkm Still give the possibility to force the use of Heimdal, but with 1575109998Smarkm warnings and a request that patches get sent to openssl-dev. 1576109998Smarkm [Richard Levitte] 1577109998Smarkm 1578109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 1579109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 1580109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 1581109998Smarkm 1582109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 1583109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 1584109998Smarkm edit numbers of the version. 1585109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 1586109998Smarkm 1587109998Smarkm *) Introduce safe string copy and catenation functions 1588109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 1589109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 1590109998Smarkm 1591109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 1592109998Smarkm [Ben Laurie (CHATS)] 1593109998Smarkm 1594109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 1595109998Smarkm resizing buffers containing secrets, and use where appropriate. 1596109998Smarkm [Ben Laurie (CHATS)] 1597109998Smarkm 1598109998Smarkm *) Avoid using fixed size buffers for configuration file location. 1599109998Smarkm [Ben Laurie (CHATS)] 1600109998Smarkm 1601109998Smarkm *) Avoid filename truncation for various CA files. 1602109998Smarkm [Ben Laurie (CHATS)] 1603109998Smarkm 1604109998Smarkm *) Use sizeof in preference to magic numbers. 1605109998Smarkm [Ben Laurie (CHATS)] 1606109998Smarkm 1607109998Smarkm *) Avoid filename truncation in cert requests. 1608109998Smarkm [Ben Laurie (CHATS)] 1609109998Smarkm 1610109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 1611109998Smarkm overflows. 1612109998Smarkm [Ben Laurie (CHATS)] 1613109998Smarkm 1614109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 1615109998Smarkm potentially lead to a spoofing attack). 1616109998Smarkm [Ben Laurie (CHATS)] 1617109998Smarkm 1618109998Smarkm *) Fix various buffers to be large enough for hex/decimal 1619109998Smarkm representations in a platform independent manner. 1620109998Smarkm [Ben Laurie (CHATS)] 1621109998Smarkm 1622109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 1623109998Smarkm resizing buffers containing secrets, and use where appropriate. 1624109998Smarkm [Ben Laurie (CHATS)] 1625109998Smarkm 1626109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 1627109998Smarkm indents. 1628109998Smarkm [Ben Laurie (CHATS)] 1629109998Smarkm 1630109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 1631109998Smarkm [Ben Laurie (CHATS)] 1632109998Smarkm 1633109998Smarkm *) buffer_gets() could terminate with the buffer only half 1634109998Smarkm full. Fixed. 1635109998Smarkm [Ben Laurie (CHATS)] 1636109998Smarkm 1637109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 1638109998Smarkm overflowing internal buffers by having large block sizes, etc. 1639109998Smarkm [Ben Laurie (CHATS)] 1640109998Smarkm 1641109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 1642109998Smarkm unconditionally). 1643109998Smarkm [Ben Laurie (CHATS)] 1644109998Smarkm 1645109998Smarkm *) Eliminate unused copy of key in RC4. 1646109998Smarkm [Ben Laurie (CHATS)] 1647109998Smarkm 1648109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 1649109998Smarkm [Ben Laurie (CHATS)] 1650109998Smarkm 1651109998Smarkm *) Fix off-by-one error in EGD path. 1652109998Smarkm [Ben Laurie (CHATS)] 1653109998Smarkm 1654109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 1655109998Smarkm [Ben Laurie (CHATS)] 1656109998Smarkm 1657109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 1658109998Smarkm CBCParameter. 1659109998Smarkm [Ben Laurie (CHATS)] 1660109998Smarkm 1661109998Smarkm *) Eliminate unused and dangerous function knumber(). 1662109998Smarkm [Ben Laurie (CHATS)] 1663109998Smarkm 1664109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 1665109998Smarkm [Ben Laurie (CHATS)] 1666109998Smarkm 1667109998Smarkm *) Protect against overlong session ID context length in an encoded 1668109998Smarkm session object. Since these are local, this does not appear to be 1669109998Smarkm exploitable. 1670109998Smarkm [Ben Laurie (CHATS)] 1671109998Smarkm 1672109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 1673109998Smarkm the 0.9.6 release series: 1674109998Smarkm 1675109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 1676109998Smarkm supply an oversized master key in Kerberos-enabled versions. 1677160814Ssimon (CVE-2002-0657) 1678109998Smarkm [Ben Laurie (CHATS)] 1679109998Smarkm 1680109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 1681109998Smarkm [Richard Levitte] 1682109998Smarkm 1683109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 1684109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 1685109998Smarkm 1686109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 1687109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 1688109998Smarkm 1689109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 1690109998Smarkm have been removed entirely. This was also the last step to make 1691109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 1692109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 1693109998Smarkm 1694109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 1695109998Smarkm to allow version independent disabling of normally unselected ciphers, 1696109998Smarkm which may be activated as a side-effect of selecting a single cipher. 1697109998Smarkm 1698109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 1699109998Smarkm out of "ALL" because they do not provide symmetric encryption. 1700109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 1701109998Smarkm [Lutz Jaenicke, Bodo Moeller] 1702109998Smarkm 1703109998Smarkm *) Add appropriate support for separate platform-dependent build 1704109998Smarkm directories. The recommended way to make a platform-dependent 1705109998Smarkm build directory is the following (tested on Linux), maybe with 1706109998Smarkm some local tweaks: 1707109998Smarkm 1708109998Smarkm # Place yourself outside of the OpenSSL source tree. In 1709109998Smarkm # this example, the environment variable OPENSSL_SOURCE 1710109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 1711109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 1712109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 1713111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 1714109998Smarkm mkdir -p `dirname $F` 1715109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 1716109998Smarkm done 1717109998Smarkm 1718109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 1719109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 1720109998Smarkm it probably means the source directory is very clean. 1721109998Smarkm [Richard Levitte] 1722109998Smarkm 1723109998Smarkm *) Make sure any ENGINE control commands make local copies of string 1724109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 1725109998Smarkm the caller may have overwritten (or deallocated) the original string 1726109998Smarkm data when a later ENGINE operation tries to use the stored values. 1727109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 1728109998Smarkm 1729109998Smarkm *) Improve diagnostics in file reading and command-line digests. 1730109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 1731109998Smarkm 1732109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 1733109998Smarkm error in AES-CFB decryption. 1734109998Smarkm [Richard Levitte] 1735109998Smarkm 1736109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 1737109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 1738109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 1739109998Smarkm BIOs and some applications. This has the side effect that 1740109998Smarkm applications must explicitly clean up cipher contexts with 1741109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 1742109998Smarkm [Steve Henson] 1743109998Smarkm 1744109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 1745109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 1746109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 1747109998Smarkm [Steve Henson] 1748109998Smarkm 1749109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 1750109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 1751109998Smarkm [Lutz Jaenicke] 1752109998Smarkm 1753109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 1754109998Smarkm form for "surname", serialNumber has no short form. 1755109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 1756109998Smarkm therefore remove "mail" short name for "internet 7". 1757109998Smarkm The OID for unique identifiers in X509 certificates is 1758109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 1759109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 1760109998Smarkm [Lutz Jaenicke] 1761109998Smarkm 1762109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 1763109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 1764109998Smarkm after all ctrl commands have been executed on it. If init=1 the 1765109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 1766109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 1767109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 1768109998Smarkm [Steve Henson] 1769109998Smarkm 1770109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 1771109998Smarkm argument is actually passed to the callback: In the 1772109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 1773109998Smarkm declaration has been changed from 1774109998Smarkm int (*cb)() 1775109998Smarkm into 1776109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 1777109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 1778109998Smarkm i=s->ctx->app_verify_callback(&ctx) 1779109998Smarkm has been changed into 1780109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 1781109998Smarkm 1782109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 1783109998Smarkm a dummy argument can be added to their callback functions. 1784109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 1785109998Smarkm 1786109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 1787109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 1788109998Smarkm 1789109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 1790109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 1791109998Smarkm This allows older applications to transparently support certain 1792109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 1793109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 1794109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 1795109998Smarkm always load it have also been added. 1796109998Smarkm [Steve Henson] 1797109998Smarkm 1798109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 1799109998Smarkm Adjust NIDs and EVP layer. 1800109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 1801109998Smarkm 1802109998Smarkm *) Config modules support in openssl utility. 1803109998Smarkm 1804109998Smarkm Most commands now load modules from the config file, 1805109998Smarkm though in a few (such as version) this isn't done 1806109998Smarkm because it couldn't be used for anything. 1807109998Smarkm 1808109998Smarkm In the case of ca and req the config file used is 1809109998Smarkm the same as the utility itself: that is the -config 1810109998Smarkm command line option can be used to specify an 1811109998Smarkm alternative file. 1812109998Smarkm [Steve Henson] 1813109998Smarkm 1814109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 1815109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 1816109998Smarkm [Steve Henson] 1817109998Smarkm 1818109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 1819109998Smarkm config section name. Add a new flag to tolerate a missing config file 1820109998Smarkm and move code to CONF_modules_load_file(). 1821109998Smarkm [Steve Henson] 1822109998Smarkm 1823109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 1824109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 1825109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 1826109998Smarkm to work with the new engine framework. 1827109998Smarkm [AEP Inc. and Richard Levitte] 1828109998Smarkm 1829109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 1830109998Smarkm Technologies. (Use engine 'sureware') 1831109998Smarkm The support was copied from 0.9.6c [engine] and adapted 1832109998Smarkm to work with the new engine framework. 1833109998Smarkm [Richard Levitte] 1834109998Smarkm 1835109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 1836109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 1837109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 1838109998Smarkm 1839109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 1840109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 1841109998Smarkm 1842109998Smarkm *) Add the configuration target debug-linux-ppro. 1843109998Smarkm Make 'openssl rsa' use the general key loading routines 1844109998Smarkm implemented in apps.c, and make those routines able to 1845109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 1846109998Smarkm FORMAT_IISSGC. 1847109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 1848109998Smarkm 1849109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 1850109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 1851109998Smarkm 1852109998Smarkm *) Add -keyform to rsautl, and document -engine. 1853109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 1854109998Smarkm 1855109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 1856109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 1857109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 1858109998Smarkm [Ben Laurie] 1859109998Smarkm 1860109998Smarkm *) Add new functions 1861109998Smarkm ERR_peek_last_error 1862109998Smarkm ERR_peek_last_error_line 1863109998Smarkm ERR_peek_last_error_line_data. 1864109998Smarkm These are similar to 1865109998Smarkm ERR_peek_error 1866109998Smarkm ERR_peek_error_line 1867109998Smarkm ERR_peek_error_line_data, 1868109998Smarkm but report on the latest error recorded rather than the first one 1869109998Smarkm still in the error queue. 1870109998Smarkm [Ben Laurie, Bodo Moeller] 1871109998Smarkm 1872109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 1873109998Smarkm like: 1874109998Smarkm default_algorithms = ALL 1875109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 1876109998Smarkm [Steve Henson] 1877109998Smarkm 1878109998Smarkm *) Prelminary ENGINE config module. 1879109998Smarkm [Steve Henson] 1880109998Smarkm 1881109998Smarkm *) New experimental application configuration code. 1882109998Smarkm [Steve Henson] 1883109998Smarkm 1884109998Smarkm *) Change the AES code to follow the same name structure as all other 1885109998Smarkm symmetric ciphers, and behave the same way. Move everything to 1886109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 1887109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 1888109998Smarkm 1889109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 1890109998Smarkm [Ben Laurie and Theo de Raadt] 1891109998Smarkm 1892109998Smarkm *) Add option to output public keys in req command. 1893109998Smarkm [Massimiliano Pala madwolf@openca.org] 1894109998Smarkm 1895109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 1896109998Smarkm (up to about 10% better than before for P-192 and P-224). 1897109998Smarkm [Bodo Moeller] 1898109998Smarkm 1899109998Smarkm *) New functions/macros 1900109998Smarkm 1901109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 1902109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 1903109998Smarkm SSL_set_msg_callback(ssl, cb) 1904109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 1905109998Smarkm 1906109998Smarkm to request calling a callback function 1907109998Smarkm 1908109998Smarkm void cb(int write_p, int version, int content_type, 1909109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 1910109998Smarkm 1911109998Smarkm whenever a protocol message has been completely received 1912109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 1913109998Smarkm protocol version according to which the SSL library interprets 1914109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 1915109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 1916109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 1917109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 1918109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 1919109998Smarkm SSL object, and 'arg' is the application-defined value set by 1920109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 1921109998Smarkm 1922109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 1923109998Smarkm to enable a callback that displays all protocol messages. 1924109998Smarkm [Bodo Moeller] 1925109998Smarkm 1926109998Smarkm *) Change the shared library support so shared libraries are built as 1927109998Smarkm soon as the corresponding static library is finished, and thereby get 1928109998Smarkm openssl and the test programs linked against the shared library. 1929109998Smarkm This still only happens when the keyword "shard" has been given to 1930109998Smarkm the configuration scripts. 1931109998Smarkm 1932109998Smarkm NOTE: shared library support is still an experimental thing, and 1933109998Smarkm backward binary compatibility is still not guaranteed. 1934109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 1935109998Smarkm 1936109998Smarkm *) Add support for Subject Information Access extension. 1937109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 1938109998Smarkm 1939109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 1940109998Smarkm additional bytes when new memory had to be allocated, not just 1941109998Smarkm when reusing an existing buffer. 1942109998Smarkm [Bodo Moeller] 1943109998Smarkm 1944109998Smarkm *) New command line and configuration option 'utf8' for the req command. 1945109998Smarkm This allows field values to be specified as UTF8 strings. 1946109998Smarkm [Steve Henson] 1947109998Smarkm 1948109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 1949109998Smarkm runs for the former and machine-readable output for the latter. 1950109998Smarkm [Ben Laurie] 1951109998Smarkm 1952109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 1953109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 1954109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 1955109998Smarkm has the same effect. 1956109998Smarkm [Massimiliano Pala madwolf@openca.org] 1957109998Smarkm 1958109998Smarkm *) Change all functions with names starting with des_ to be starting 1959109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 1960109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 1961109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 1962109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 1963109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 1964109998Smarkm exception. 1965109998Smarkm 1966109998Smarkm Since we provide two compatibility mappings, the user needs to 1967109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 1968109998Smarkm compatibility is desired. The default (i.e., when that macro 1969109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 1970109998Smarkm 1971109998Smarkm There are also macros that enable and disable the support of old 1972109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 1973109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 1974109998Smarkm are defined, the default will apply: to support the old des routines. 1975109998Smarkm 1976109998Smarkm In either case, one must include openssl/des.h to get the correct 1977109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 1978109998Smarkm won't work. 1979109998Smarkm 1980109998Smarkm NOTE: This is a major break of an old API into a new one. Software 1981109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 1982109998Smarkm time in the future, des_old.h and the libdes compatibility functions 1983109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 1984109998Smarkm default), and then completely removed. 1985109998Smarkm [Richard Levitte] 1986109998Smarkm 1987109998Smarkm *) Test for certificates which contain unsupported critical extensions. 1988109998Smarkm If such a certificate is found during a verify operation it is 1989109998Smarkm rejected by default: this behaviour can be overridden by either 1990109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 1991109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 1992109998Smarkm X509_supported_extension() has also been added which returns 1 if a 1993109998Smarkm particular extension is supported. 1994109998Smarkm [Steve Henson] 1995109998Smarkm 1996109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 1997109998Smarkm to retain compatibility with existing code. 1998109998Smarkm [Steve Henson] 1999109998Smarkm 2000109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 2001109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 2002109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 2003109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 2004109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 2005109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 2006109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 2007109998Smarkm requires the destination to be valid. 2008109998Smarkm 2009109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 2010109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 2011109998Smarkm [Steve Henson] 2012109998Smarkm 2013109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 2014109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 2015109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 2016109998Smarkm [Bodo Moeller] 2017109998Smarkm 2018109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 2019109998Smarkm [Massimo Santin via Richard Levitte] 2020109998Smarkm 2021109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 2022109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 2023109998Smarkm (initialisation, etc) from functionality dealing with implementations 2024109998Smarkm of specific crypto iterfaces. This change also introduces integrated 2025109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 2026109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 2027109998Smarkm implementations of their own. This is detailed in crypto/engine/README 2028109998Smarkm as it couldn't be adequately described here. However, there are a few 2029109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 2030109998Smarkm were changed in the original introduction of ENGINE code have now 2031109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 2032109998Smarkm deal more passive and at run-time, operations deal directly with 2033109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 2034109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 2035109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 2036109998Smarkm they were not being used by the framework as there is no concept of a 2037109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 2038109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 2039109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 2040109998Smarkm the new code. 2041109998Smarkm [Geoff Thorpe] 2042109998Smarkm 2043109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 2044109998Smarkm [Steve Henson] 2045109998Smarkm 2046109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 2047109998Smarkm and make sure the automatically generated functions ERR_load_* 2048109998Smarkm become part of libeay.num as well. 2049109998Smarkm [Richard Levitte] 2050109998Smarkm 2051109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 2052109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 2053109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 2054109998Smarkm false once a handshake has been completed. 2055109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 2056109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 2057109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 2058109998Smarkm client has followed the request.) 2059109998Smarkm [Bodo Moeller] 2060109998Smarkm 2061109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 2062109998Smarkm By default, clients may request session resumption even during 2063109998Smarkm renegotiation (if session ID contexts permit); with this option, 2064109998Smarkm session resumption is possible only in the first handshake. 2065109998Smarkm 2066109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 2067109998Smarkm more bits available for options that should not be part of 2068109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 2069109998Smarkm [Bodo Moeller] 2070109998Smarkm 2071109998Smarkm *) Add some demos for certificate and certificate request creation. 2072109998Smarkm [Steve Henson] 2073109998Smarkm 2074109998Smarkm *) Make maximum certificate chain size accepted from the peer application 2075109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 2076109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 2077109998Smarkm [Lutz Jaenicke] 2078109998Smarkm 2079109998Smarkm *) Add support for shared libraries for Unixware-7 2080109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 2081109998Smarkm [Lutz Jaenicke] 2082109998Smarkm 2083109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 2084109998Smarkm be done prior to destruction. Use this to unload error strings from 2085109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 2086109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 2087109998Smarkm [Geoff Thorpe] 2088109998Smarkm 2089109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 2090109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 2091109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 2092109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 2093109998Smarkm Also, add stub code to each that makes building them as self-contained 2094109998Smarkm shared-libraries easier (see README.ENGINE). 2095109998Smarkm [Geoff Thorpe] 2096109998Smarkm 2097109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 2098109998Smarkm implementations into applications that are completely implemented in 2099109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 2100109998Smarkm commands that can be used to configure what shared-library to load and 2101109998Smarkm to control aspects of the way it is handled. Also, made an update to 2102109998Smarkm the README.ENGINE file that brings its information up-to-date and 2103109998Smarkm provides some information and instructions on the "dynamic" ENGINE 2104109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 2105109998Smarkm [Geoff Thorpe] 2106109998Smarkm 2107109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 2108109998Smarkm "ERR_unload_strings" function. 2109109998Smarkm [Geoff Thorpe] 2110109998Smarkm 2111109998Smarkm *) Add a copy() function to EVP_MD. 2112109998Smarkm [Ben Laurie] 2113109998Smarkm 2114109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 2115109998Smarkm md_data void pointer. 2116109998Smarkm [Ben Laurie] 2117109998Smarkm 2118109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 2119109998Smarkm that the digest can only process a single chunk of data 2120109998Smarkm (typically because it is provided by a piece of 2121109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 2122109998Smarkm is only going to provide a single chunk of data, and hence the 2123109998Smarkm framework needn't accumulate the data for oneshot drivers. 2124109998Smarkm [Ben Laurie] 2125109998Smarkm 2126109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 2127109998Smarkm functions. This change also alters the storage and management of global 2128109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 2129109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 2130109998Smarkm index counters. The API functions that use this state have been changed 2131109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 2132109998Smarkm and counter, and there is now an API function to dynamically create new 2133109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 2134109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 2135109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 2136109998Smarkm such data would previously have always leaked in application code and 2137109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 2138109998Smarkm to it. Application code that doesn't use this new function will still 2139109998Smarkm leak as before, but their memory debugging output will announce it now 2140109998Smarkm rather than letting it slide. 2141109998Smarkm 2142109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 2143109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 2144109998Smarkm has a return value to indicate success or failure. 2145109998Smarkm [Geoff Thorpe] 2146109998Smarkm 2147109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 2148109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 2149109998Smarkm implementation. This change also adds two functions to "get" and "set" 2150109998Smarkm the implementation prior to it being automatically set the first time 2151109998Smarkm any other ERR function takes place. Ie. an application can call "get", 2152109998Smarkm pass the return value to a module it has just loaded, and that module 2153109998Smarkm can call its own "set" function using that value. This means the 2154109998Smarkm module's "ERR" operations will use (and modify) the error state in the 2155109998Smarkm application and not in its own statically linked copy of OpenSSL code. 2156109998Smarkm [Geoff Thorpe] 2157109998Smarkm 2158109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 2159109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 2160109998Smarkm the operation, and provides a more encapsulated way for external code 2161109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 2162109998Smarkm to use these functions rather than manually incrementing the counts. 2163109998Smarkm 2164109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 2165109998Smarkm [Geoff Thorpe] 2166109998Smarkm 2167109998Smarkm *) Add EVP test program. 2168109998Smarkm [Ben Laurie] 2169109998Smarkm 2170109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 2171109998Smarkm [Ben Laurie] 2172109998Smarkm 2173109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 2174109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 2175109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 2176109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 2177109998Smarkm directly. Modify 'ca' application to use new functions. 2178109998Smarkm [Steve Henson] 2179109998Smarkm 2180109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 2181109998Smarkm bug workarounds. Rollback attack detection is a security feature. 2182109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 2183109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 2184109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 2185109998Smarkm for their choice and can explicitly enable this option. 2186109998Smarkm [Bodo Moeller, Lutz Jaenicke] 2187109998Smarkm 2188109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 2189109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 2190109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 2191109998Smarkm Usage example: 2192109998Smarkm 2193109998Smarkm EVP_MD_CTX md; 2194109998Smarkm 2195109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 2196109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 2197109998Smarkm EVP_DigestUpdate(&md, in, len); 2198109998Smarkm EVP_DigestFinal(&md, out, NULL); 2199109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 2200109998Smarkm 2201109998Smarkm [Ben Laurie] 2202109998Smarkm 2203109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 2204109998Smarkm correcting its structure. This means that calls to DES functions 2205109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 2206109998Smarkm plain des_key_schedule (which was actually always a pointer 2207109998Smarkm anyway): E.g., 2208109998Smarkm 2209109998Smarkm des_key_schedule ks; 2210109998Smarkm 2211109998Smarkm des_set_key_checked(..., &ks); 2212109998Smarkm des_ncbc_encrypt(..., &ks, ...); 2213109998Smarkm 2214109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 2215109998Smarkm [Ben Laurie] 2216109998Smarkm 2217109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 2218109998Smarkm PEM causes large amounts of unused functions to be linked in due to 2219109998Smarkm poor organisation. For example pem_all.c contains every PEM function 2220109998Smarkm which has a knock on effect of linking in large amounts of (unused) 2221109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 2222109998Smarkm functions prevents this. 2223109998Smarkm [Steve Henson] 2224109998Smarkm 2225109998Smarkm *) Cleanup of EVP macros. 2226109998Smarkm [Ben Laurie] 2227109998Smarkm 2228109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 2229109998Smarkm correct _ecb suffix. 2230109998Smarkm [Ben Laurie] 2231109998Smarkm 2232109998Smarkm *) Add initial OCSP responder support to ocsp application. The 2233109998Smarkm revocation information is handled using the text based index 2234109998Smarkm use by the ca application. The responder can either handle 2235109998Smarkm requests generated internally, supplied in files (for example 2236109998Smarkm via a CGI script) or using an internal minimal server. 2237109998Smarkm [Steve Henson] 2238109998Smarkm 2239109998Smarkm *) Add configuration choices to get zlib compression for TLS. 2240109998Smarkm [Richard Levitte] 2241109998Smarkm 2242109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 2243109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 2244109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 2245109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 2246109998Smarkm 2247109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 2248109998Smarkm and authenticator structs; see crypto/krb5/. 2249109998Smarkm 2250109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 2251109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 2252109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 2253109998Smarkm via Richard Levitte] 2254109998Smarkm 2255109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 2256109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 2257109998Smarkm values for each of the key sizes rather than having just 2258109998Smarkm parameters (and 'speed' generating keys each time). 2259109998Smarkm [Geoff Thorpe] 2260109998Smarkm 2261109998Smarkm *) Speed up EVP routines. 2262109998Smarkm Before: 2263109998Smarkmencrypt 2264109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 2265109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 2266109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 2267109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 2268109998Smarkmdecrypt 2269109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 2270109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 2271109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 2272109998Smarkm After: 2273109998Smarkmencrypt 2274109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 2275109998Smarkmdecrypt 2276109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 2277109998Smarkm [Ben Laurie] 2278109998Smarkm 2279109998Smarkm *) Added the OS2-EMX target. 2280109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 2281109998Smarkm 2282109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 2283109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 2284109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 2285109998Smarkm structure: this means that the old CONF compatible routines can be 2286109998Smarkm retained (in particular wrt extensions) without having to duplicate the 2287109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 2288109998Smarkm [Steve Henson] 2289109998Smarkm 2290109998Smarkm *) Enhance the general user interface with mechanisms for inner control 2291109998Smarkm and with possibilities to have yes/no kind of prompts. 2292109998Smarkm [Richard Levitte] 2293109998Smarkm 2294109998Smarkm *) Change all calls to low level digest routines in the library and 2295109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 2296109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 2297109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 2298109998Smarkm 2299109998Smarkm *) Add the possibility to control engines through control names but with 2300109998Smarkm arbitrary arguments instead of just a string. 2301109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 2302109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 2303109998Smarkm versions of OpenSSL [engine]. 2304109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 2305109998Smarkm callback. 2306109998Smarkm [Richard Levitte] 2307109998Smarkm 2308109998Smarkm *) Enhance the general user interface with mechanisms to better support 2309109998Smarkm dialog box interfaces, application-defined prompts, the possibility 2310109998Smarkm to use defaults (for example default passwords from somewhere else) 2311109998Smarkm and interrupts/cancellations. 2312109998Smarkm [Richard Levitte] 2313109998Smarkm 2314109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 2315109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 2316109998Smarkm [Steve Henson] 2317109998Smarkm 2318109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 2319109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 2320109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 2321109998Smarkm 2322109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 2323109998Smarkm callback (pem_password_cb) as all other routines that need this 2324109998Smarkm kind of callback. 2325109998Smarkm [Richard Levitte] 2326109998Smarkm 2327109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 2328109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 2329109998Smarkm than this minimum value is recommended. 2330109998Smarkm [Lutz Jaenicke] 2331109998Smarkm 2332109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 2333109998Smarkm that are easily reachable. 2334109998Smarkm [Richard Levitte] 2335109998Smarkm 2336109998Smarkm *) Windows apparently can't transparently handle global 2337109998Smarkm variables defined in DLLs. Initialisations such as: 2338109998Smarkm 2339109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 2340109998Smarkm 2341109998Smarkm wont compile. This is used by the any applications that need to 2342109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 2343109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 2344109998Smarkm needed for static libraries under Win32. 2345109998Smarkm [Steve Henson] 2346109998Smarkm 2347109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 2348109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 2349109998Smarkm purpose functions and tidy up setting in other SSL functions. 2350109998Smarkm [Steve Henson] 2351109998Smarkm 2352109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 2353109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 2354109998Smarkm initialised. This allows various defaults to be set in the 2355109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 2356109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 2357109998Smarkm internally such as S/MIME. 2358109998Smarkm 2359109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 2360109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 2361109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 2362109998Smarkm 2363109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 2364109998Smarkm applications. 2365109998Smarkm [Steve Henson] 2366109998Smarkm 2367109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 2368109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 2369109998Smarkm its validity and signature checked, then if the certificate is found 2370109998Smarkm in the CRL the verify fails with a revoked error. 2371109998Smarkm 2372109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 2373109998Smarkm 2374109998Smarkm Command line options added to 'verify' application to support this. 2375109998Smarkm 2376109998Smarkm This needs some additional work, such as being able to handle multiple 2377109998Smarkm CRLs with different times, extension based lookup (rather than just 2378109998Smarkm by subject name) and ultimately more complete V2 CRL extension 2379109998Smarkm handling. 2380109998Smarkm [Steve Henson] 2381109998Smarkm 2382109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 2383109998Smarkm to replace things like des_read_password and friends (backward 2384109998Smarkm compatibility functions using this new API are provided). 2385109998Smarkm The purpose is to remove prompting functions from the DES code 2386109998Smarkm section as well as provide for prompting through dialog boxes in 2387109998Smarkm a window system and the like. 2388109998Smarkm [Richard Levitte] 2389109998Smarkm 2390109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 2391109998Smarkm per-structure level rather than having to store it globally. 2392109998Smarkm [Geoff] 2393109998Smarkm 2394109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 2395109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 2396109998Smarkm This causes the "original" ENGINE structure to act like a template, 2397109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 2398109998Smarkm operational state can be localised to each ENGINE structure, despite the 2399109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 2400109998Smarkm this case have no functional references and the return value is the single 2401109998Smarkm structural reference. This matches the single structural reference returned 2402109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 2403109998Smarkm ENGINE structure. 2404109998Smarkm [Geoff] 2405109998Smarkm 2406109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 2407109998Smarkm needs to match any other type at all we need to manually clear the 2408109998Smarkm tag cache. 2409109998Smarkm [Steve Henson] 2410109998Smarkm 2411109998Smarkm *) Changes to the "openssl engine" utility to include; 2412109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 2413109998Smarkm about an ENGINE's available control commands. 2414109998Smarkm - executing control commands from command line arguments using the 2415109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 2416109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 2417109998Smarkm the individual commands are colon-separated, for example; 2418109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 2419109998Smarkm [Geoff] 2420109998Smarkm 2421109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 2422109998Smarkm declare their own commands (numbers), names (strings), descriptions, 2423109998Smarkm and input types for run-time discovery by calling applications. A 2424109998Smarkm subset of these commands are implicitly classed as "executable" 2425109998Smarkm depending on their input type, and only these can be invoked through 2426109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 2427109998Smarkm can be based on user input, config files, etc). The distinction is 2428109998Smarkm that "executable" commands cannot return anything other than a boolean 2429109998Smarkm result and can only support numeric or string input, whereas some 2430109998Smarkm discoverable commands may only be for direct use through 2431109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 2432109998Smarkm pointers, or other custom uses. The "executable" commands are to 2433109998Smarkm support parameterisations of ENGINE behaviour that can be 2434109998Smarkm unambiguously defined by ENGINEs and used consistently across any 2435109998Smarkm OpenSSL-based application. Commands have been added to all the 2436109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 2437109998Smarkm control over shared-library paths without source code alterations. 2438109998Smarkm [Geoff] 2439109998Smarkm 2440109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 2441109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 2442109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 2443109998Smarkm this also allows the implementations to compile without using the 2444109998Smarkm internal engine_int.h header. 2445109998Smarkm [Geoff] 2446109998Smarkm 2447109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 2448109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 2449109998Smarkm should already have non-const pointers to it (ie. they should only 2450109998Smarkm modify their own ones). 2451109998Smarkm [Geoff] 2452109998Smarkm 2453109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 2454109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 2455109998Smarkm to C code. "nuron" string definitions were placed in variables 2456109998Smarkm rather than hard-coded - allowing parameterisation of these values 2457109998Smarkm later on via ctrl() commands. 2458109998Smarkm - Removed unused "#if 0"'d code. 2459109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 2460109998Smarkm structural references. 2461109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 2462109998Smarkm - Constified various get/set functions as appropriate and added 2463109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 2464109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 2465109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 2466109998Smarkm or function to NULL is a way of cancelling out a previously set 2467109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 2468109998Smarkm and doesn't justify the extra error symbols and code. 2469109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 2470109998Smarkm flags from engine_int.h to engine.h. 2471109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 2472109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 2473109998Smarkm [Geoff] 2474109998Smarkm 2475109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 2476109998Smarkm to the algorithm using long division. The binary algorithm can be 2477109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 2478109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 2479109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 2480109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 2481109998Smarkm appears to be advantageous for much longer moduli; here we use it 2482109998Smarkm for moduli up to 2048 bits. 2483109998Smarkm [Bodo Moeller] 2484109998Smarkm 2485109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 2486109998Smarkm could not support the combine flag in choice fields. 2487109998Smarkm [Steve Henson] 2488109998Smarkm 2489109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 2490109998Smarkm extensions from a certificate request to the certificate. 2491109998Smarkm [Steve Henson] 2492109998Smarkm 2493109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 2494109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 2495109998Smarkm file: this allows the display of the certificate about to be 2496109998Smarkm signed to be customised, to allow certain fields to be included 2497109998Smarkm or excluded and extension details. The old system didn't display 2498109998Smarkm multicharacter strings properly, omitted fields not in the policy 2499109998Smarkm and couldn't display additional details such as extensions. 2500109998Smarkm [Steve Henson] 2501109998Smarkm 2502109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 2503109998Smarkm of an arbitrary number of elliptic curve points 2504109998Smarkm \sum scalars[i]*points[i], 2505109998Smarkm optionally including the generator defined for the EC_GROUP: 2506109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 2507109998Smarkm 2508109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 2509109998Smarkm that the point list has just one item (besides the optional 2510109998Smarkm generator). 2511109998Smarkm [Bodo Moeller] 2512109998Smarkm 2513109998Smarkm *) First EC_METHODs for curves over GF(p): 2514109998Smarkm 2515109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 2516109998Smarkm operations and provides various method functions that can also 2517109998Smarkm operate with faster implementations of modular arithmetic. 2518109998Smarkm 2519109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 2520109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 2521109998Smarkm 2522109998Smarkm [Bodo Moeller; point addition and point doubling 2523109998Smarkm implementation directly derived from source code provided by 2524109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 2525109998Smarkm 2526109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 2527109998Smarkm crypto/ec/ec_lib.c): 2528109998Smarkm 2529109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 2530109998Smarkm based on EC_METHODs that are built into the library. 2531109998Smarkm 2532109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 2533109998Smarkm 2534109998Smarkm Most of the framework would be able to handle curves over arbitrary 2535109998Smarkm finite fields, but as there are no obvious types for fields other 2536109998Smarkm than GF(p), some functions are limited to that for now. 2537109998Smarkm [Bodo Moeller] 2538109998Smarkm 2539109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 2540109998Smarkm that the file contains a complete HTTP response. 2541109998Smarkm [Richard Levitte] 2542109998Smarkm 2543109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 2544109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 2545109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 2546109998Smarkm field while the former will cause them to run together if the field 2547109998Smarkm is 40 of more characters long. 2548109998Smarkm [Steve Henson] 2549109998Smarkm 2550109998Smarkm *) Constify the cipher and digest 'method' functions and structures 2551109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 2552109998Smarkm pointers. 2553109998Smarkm [Steve Henson] 2554109998Smarkm 2555109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 2556109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 2557109998Smarkm [Bodo Moeller] 2558109998Smarkm 2559109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 2560109998Smarkm internal software routines can never fail additional hardware versions 2561109998Smarkm might. 2562109998Smarkm [Steve Henson] 2563109998Smarkm 2564109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 2565109998Smarkm 2566109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 2567109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 2568109998Smarkm 2569109998Smarkm ASN1 error codes 2570109998Smarkm ERR_R_NESTED_ASN1_ERROR 2571109998Smarkm ... 2572109998Smarkm ERR_R_MISSING_ASN1_EOS 2573109998Smarkm were 4 .. 9, conflicting with 2574109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 2575109998Smarkm ... 2576109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 2577109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 2578109998Smarkm 2579109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 2580109998Smarkm [Bodo Moeller] 2581109998Smarkm 2582109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 2583109998Smarkm suffices. 2584109998Smarkm [Bodo Moeller] 2585109998Smarkm 2586109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 2587109998Smarkm sets the subject name for a new request or supersedes the 2588109998Smarkm subject name in a given request. Formats that can be parsed are 2589109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 2590109998Smarkm and 2591109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 2592109998Smarkm 2593109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 2594109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 2595109998Smarkm 2596109998Smarkm *) Introduce the possibility to access global variables through 2597109998Smarkm functions on platform were that's the best way to handle exporting 2598109998Smarkm global variables in shared libraries. To enable this functionality, 2599109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 2600109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 2601109998Smarkm is normally done by Configure or something similar). 2602109998Smarkm 2603109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 2604109998Smarkm in the source file (foo.c) like this: 2605109998Smarkm 2606109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 2607109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 2608109998Smarkm 2609109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 2610109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 2611109998Smarkm 2612109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 2613109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 2614109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 2615109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 2616109998Smarkm 2617109998Smarkm The #defines are very important, and therefore so is including the 2618109998Smarkm header file everywhere where the defined globals are used. 2619109998Smarkm 2620109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 2621109998Smarkm of ASN.1 items, but that structure is a bit different. 2622109998Smarkm 2623109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 2624109998Smarkm better and easier to understand logic to choose which symbols should 2625109998Smarkm go into the Windows .def files as well as a number of fixes and code 2626109998Smarkm cleanup (among others, algorithm keywords are now sorted 2627109998Smarkm lexicographically to avoid constant rewrites). 2628109998Smarkm [Richard Levitte] 2629109998Smarkm 2630109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 2631109998Smarkm result to 'rm' because if rm==num the value will be overwritten 2632109998Smarkm and produce the wrong result if 'num' is negative: this caused 2633109998Smarkm problems with BN_mod() and BN_nnmod(). 2634109998Smarkm [Steve Henson] 2635109998Smarkm 2636109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 2637109998Smarkm OCSP request and verifies the signer certificate. The signer 2638109998Smarkm certificate is just checked for a generic purpose and OCSP request 2639109998Smarkm trust settings. 2640109998Smarkm [Steve Henson] 2641109998Smarkm 2642109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 2643109998Smarkm responses. OCSP responses are prepared in real time and may only 2644109998Smarkm be a few seconds old. Simply checking that the current time lies 2645109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 2646109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 2647109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 2648109998Smarkm the current time. The age of the response can also optionally be 2649109998Smarkm checked. Two new options -validity_period and -status_age added to 2650109998Smarkm ocsp utility. 2651109998Smarkm [Steve Henson] 2652109998Smarkm 2653109998Smarkm *) If signature or public key algorithm is unrecognized print out its 2654109998Smarkm OID rather that just UNKNOWN. 2655109998Smarkm [Steve Henson] 2656109998Smarkm 2657109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 2658109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 2659109998Smarkm ID to be generated from the issuer certificate alone which can then be 2660109998Smarkm passed to OCSP_id_issuer_cmp(). 2661109998Smarkm [Steve Henson] 2662109998Smarkm 2663109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 2664109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 2665109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 2666109998Smarkm new macros which allow the underlying ASN1 function/structure to 2667109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 2668109998Smarkm references directly (such as &X509_it) but instead use the relevant 2669109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 2670109998Smarkm use of the new ASN1 code on platforms where exporting structures 2671109998Smarkm is problematical (for example in shared libraries) but exporting 2672109998Smarkm functions returning pointers to structures is not. 2673109998Smarkm [Steve Henson] 2674109998Smarkm 2675109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 2676109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 2677109998Smarkm The purpose of this is to allow applications to control, if they wish, 2678109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 2679109998Smarkm can be useful for session caching in multiple-server environments. A 2680109998Smarkm command-line switch for testing this (and any client code that wishes 2681109998Smarkm to use such a feature) has been added to "s_server". 2682109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 2683109998Smarkm 2684109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 2685109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 2686109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 2687109998Smarkm the growing number of special cases it was previously handling. 2688109998Smarkm [Richard Levitte] 2689109998Smarkm 2690109998Smarkm *) Make all configuration macros available for application by making 2691109998Smarkm sure they are available in opensslconf.h, by giving them names starting 2692109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 2693109998Smarkm sure e_os2.h will cover all platform-specific cases together with 2694109998Smarkm opensslconf.h. 2695109998Smarkm Additionally, it is now possible to define configuration/platform- 2696109998Smarkm specific names (called "system identities"). In the C code, these 2697109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 2698109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 2699109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 2700109998Smarkm what is available. 2701109998Smarkm [Richard Levitte] 2702109998Smarkm 2703109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 2704109998Smarkm number to use to be specified on the command line. Previously self 2705109998Smarkm signed certificates were hard coded with serial number 0 and the 2706109998Smarkm CA options of 'x509' had to use a serial number in a file which was 2707109998Smarkm auto incremented. 2708109998Smarkm [Steve Henson] 2709109998Smarkm 2710109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 2711109998Smarkm Currently CRL reason, invalidity date and hold instruction are 2712109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 2713109998Smarkm [Steve Henson] 2714109998Smarkm 2715109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 2716109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 2717109998Smarkm API, which was previously mandatory. This means that the data is 2718109998Smarkm not padded in any way and so the total length much be a multiple 2719109998Smarkm of the block size, otherwise an error occurs. 2720109998Smarkm [Steve Henson] 2721109998Smarkm 2722109998Smarkm *) Initial (incomplete) OCSP SSL support. 2723109998Smarkm [Steve Henson] 2724109998Smarkm 2725109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 2726109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 2727109998Smarkm option to ocsp utility. 2728109998Smarkm [Steve Henson] 2729109998Smarkm 2730109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 2731109998Smarkm reflects the various checks performed. Applications can decide 2732109998Smarkm whether to tolerate certain situations such as an absent nonce 2733109998Smarkm in a response when one was present in a request: the ocsp application 2734109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 2735109998Smarkm this is to allow responders to include a nonce in a response even if 2736109998Smarkm the request is nonce-less. 2737109998Smarkm [Steve Henson] 2738109998Smarkm 2739109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 2740109998Smarkm skipped when using openssl x509 multiple times on a single input file, 2741109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 2742109998Smarkm [Bodo Moeller] 2743109998Smarkm 2744109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 2745109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 2746109998Smarkm utility to correctly initialize revocation date of CRLs. 2747109998Smarkm [Steve Henson] 2748109998Smarkm 2749109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 2750109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 2751109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 2752109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 2753109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 2754109998Smarkm [Lutz Jaenicke] 2755109998Smarkm 2756109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 2757109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 2758109998Smarkm appear to exist. 2759109998Smarkm [Steve Henson] 2760109998Smarkm 2761109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 2762109998Smarkm additional certificates supplied. 2763109998Smarkm [Steve Henson] 2764109998Smarkm 2765109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 2766109998Smarkm OCSP client a number of certificate to only verify the response 2767109998Smarkm signature against. 2768109998Smarkm [Richard Levitte] 2769109998Smarkm 2770109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 2771109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 2772109998Smarkm AES OIDs. 2773109998Smarkm 2774109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 2775109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 2776109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 2777109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 2778109998Smarkm alias because they were not yet official; they could be 2779109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 2780109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 2781109998Smarkm alias is called "AES" and is part of "ALL".) 2782109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 2783109998Smarkm 2784109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 2785109998Smarkm request to response. 2786109998Smarkm [Steve Henson] 2787109998Smarkm 2788109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 2789109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 2790109998Smarkm extract information from a certificate request. OCSP_response_create() 2791109998Smarkm creates a response and optionally adds a basic response structure. 2792109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 2793109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 2794109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 2795109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 2796109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 2797109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 2798109998Smarkm (converts ASN1_TIME to GeneralizedTime). 2799109998Smarkm [Steve Henson] 2800109998Smarkm 2801109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 2802109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 2803109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 2804109998Smarkm contents: this is used in various key identifiers. 2805109998Smarkm [Steve Henson] 2806109998Smarkm 2807109998Smarkm *) Make sk_sort() tolerate a NULL argument. 2808109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 2809109998Smarkm 2810109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 2811109998Smarkm passed by the function are trusted implicitly. If any of them signed the 2812109998Smarkm response then it is assumed to be valid and is not verified. 2813109998Smarkm [Steve Henson] 2814109998Smarkm 2815109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 2816109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 2817109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 2818109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 2819109998Smarkm <support@securenetterm.com>] 2820109998Smarkm 2821109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 2822109998Smarkm routines: without these tracing memory leaks is very painful. 2823109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 2824109998Smarkm [Steve Henson] 2825109998Smarkm 2826109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 2827109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 2828109998Smarkm effectively meant GeneralizedTime would never be used. Now it 2829109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 2830109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 2831109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 2832109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 2833109998Smarkm <support@securenetterm.com>] 2834109998Smarkm 2835109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 2836109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 2837109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 2838109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 2839109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 2840109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 2841109998Smarkm [Steve Henson] 2842109998Smarkm 2843109998Smarkm *) Add summary printout to ocsp utility. The various functions which 2844109998Smarkm convert status values to strings have been renamed to: 2845109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 2846109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 2847109998Smarkm to verify nonce values and to disable verification. OCSP response 2848109998Smarkm printout format cleaned up. 2849109998Smarkm [Steve Henson] 2850109998Smarkm 2851109998Smarkm *) Add additional OCSP certificate checks. These are those specified 2852109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 2853109998Smarkm certificate being checked must either be the OCSP signer certificate 2854109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 2855109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 2856109998Smarkm usage. This check is performed by attempting to match the OCSP 2857109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 2858109998Smarkm in the OCSP_CERTID structures of the response. 2859109998Smarkm [Steve Henson] 2860109998Smarkm 2861109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 2862109998Smarkm and related routines. This uses the standard OpenSSL certificate 2863109998Smarkm verify routines to perform initial checks (just CA validity) and 2864109998Smarkm to obtain the certificate chain. Then additional checks will be 2865109998Smarkm performed on the chain. Currently the root CA is checked to see 2866109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 2867109998Smarkm a root CA as a global signing root: that is any certificate that 2868109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 2869109998Smarkm [Steve Henson] 2870109998Smarkm 2871109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 2872109998Smarkm extensions from a separate configuration file. 2873109998Smarkm As when reading extensions from the main configuration file, 2874109998Smarkm the '-extensions ...' option may be used for specifying the 2875109998Smarkm section to use. 2876109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 2877109998Smarkm 2878109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 2879109998Smarkm read. The request can be sent to a responder and the output 2880109998Smarkm parsed, outputed or printed in text form. Not complete yet: 2881109998Smarkm still needs to check the OCSP response validity. 2882109998Smarkm [Steve Henson] 2883109998Smarkm 2884109998Smarkm *) New subcommands for 'openssl ca': 2885109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 2886109998Smarkm the given serial number (according to the index file). 2887109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 2888109998Smarkm in the index file. 2889109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 2890109998Smarkm 2891109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 2892109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 2893109998Smarkm so that the resulting key is not encrypted. 2894109998Smarkm [Damien Miller <djm@mindrot.org>] 2895109998Smarkm 2896109998Smarkm *) New configuration for the GNU Hurd. 2897109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 2898109998Smarkm 2899109998Smarkm *) Initial code to implement OCSP basic response verify. This 2900109998Smarkm is currently incomplete. Currently just finds the signer's 2901109998Smarkm certificate and verifies the signature on the response. 2902109998Smarkm [Steve Henson] 2903109998Smarkm 2904109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 2905109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 2906109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 2907109998Smarkm [Bodo Moeller] 2908109998Smarkm 2909109998Smarkm *) Allowing defining memory allocation callbacks that will be given 2910109998Smarkm file name and line number information in additional arguments 2911109998Smarkm (a const char* and an int). The basic functionality remains, as 2912109998Smarkm well as the original possibility to just replace malloc(), 2913109998Smarkm realloc() and free() by functions that do not know about these 2914109998Smarkm additional arguments. To register and find out the current 2915109998Smarkm settings for extended allocation functions, the following 2916109998Smarkm functions are provided: 2917109998Smarkm 2918109998Smarkm CRYPTO_set_mem_ex_functions 2919109998Smarkm CRYPTO_set_locked_mem_ex_functions 2920109998Smarkm CRYPTO_get_mem_ex_functions 2921109998Smarkm CRYPTO_get_locked_mem_ex_functions 2922109998Smarkm 2923109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 2924109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 2925109998Smarkm extended allocation function is enabled. 2926109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 2927109998Smarkm a conventional allocation function is enabled. 2928109998Smarkm [Richard Levitte, Bodo Moeller] 2929109998Smarkm 2930109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 2931109998Smarkm There should no longer be any prototype-casting required when using 2932109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 2933109998Smarkm the callback types and macros at the head of lhash.h for details 2934109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 2935109998Smarkm [Geoff Thorpe] 2936109998Smarkm 2937109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 2938109998Smarkm If /dev/[u]random devices are not available or do not return enough 2939109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 2940109998Smarkm be queried. 2941109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 2942109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 2943109998Smarkm when enough entropy was collected without querying more sockets. 2944109998Smarkm [Lutz Jaenicke] 2945109998Smarkm 2946109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 2947109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 2948109998Smarkm of data has been collected. We spend at most 10 ms on each file 2949109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 2950109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 2951109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 2952109998Smarkm platforms the 10 ms delay will never occur. 2953109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 2954109998Smarkm For VMS, there's a currently-empty rand_vms.c. 2955109998Smarkm [Richard Levitte] 2956109998Smarkm 2957109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 2958109998Smarkm provide utility functions which an application needing 2959109998Smarkm to issue a request to an OCSP responder and analyse the 2960109998Smarkm response will typically need: as opposed to those which an 2961109998Smarkm OCSP responder itself would need which will be added later. 2962109998Smarkm 2963109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 2964109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 2965109998Smarkm response. OCSP_response_get1_basic() extracts basic response 2966109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 2967109998Smarkm information from an OCSP_CERTID structure (which will be created 2968109998Smarkm when the request structure is built). These are built from lower 2969109998Smarkm level functions which work on OCSP_SINGLERESP structures but 2970109998Smarkm wont normally be used unless the application wishes to examine 2971109998Smarkm extensions in the OCSP response for example. 2972109998Smarkm 2973109998Smarkm Replace nonce routines with a pair of functions. 2974109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 2975109998Smarkm generates a random value. OCSP_check_nonce() checks the 2976109998Smarkm validity of the nonce in an OCSP response. 2977109998Smarkm [Steve Henson] 2978109998Smarkm 2979109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 2980109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 2981109998Smarkm need to free up the newly created id. Change return type 2982109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 2983109998Smarkm This can then be used to add extensions to the request. 2984109998Smarkm Deleted OCSP_request_new(), since most of its functionality 2985109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 2986109998Smarkm clash) apart from the ability to set the request name which 2987109998Smarkm will be added elsewhere. 2988109998Smarkm [Steve Henson] 2989109998Smarkm 2990109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 2991109998Smarkm various functions. Extensions are now handled using the new 2992109998Smarkm OCSP extension code. New simple OCSP HTTP function which 2993109998Smarkm can be used to send requests and parse the response. 2994109998Smarkm [Steve Henson] 2995109998Smarkm 2996109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 2997109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 2998109998Smarkm uses the special reorder version of SET OF to sort the attributes 2999109998Smarkm and reorder them to match the encoded order. This resolves a long 3000109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 3001109998Smarkm it used to fail because the attribute order did not match the 3002109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 3003109998Smarkm it uses the received order. This is necessary to tolerate some broken 3004109998Smarkm software that does not order SET OF. This is handled by encoding 3005109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 3006109998Smarkm to produce the required SET OF. 3007109998Smarkm [Steve Henson] 3008109998Smarkm 3009109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 3010109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 3011109998Smarkm files to get correct declarations of the ASN.1 item variables. 3012109998Smarkm [Richard Levitte] 3013109998Smarkm 3014109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 3015109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 3016109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 3017109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 3018109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 3019109998Smarkm ASN1_ITEM and no wrapper functions. 3020109998Smarkm [Steve Henson] 3021109998Smarkm 3022109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 3023109998Smarkm replace the old function pointer based I/O routines. Change most of 3024109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 3025109998Smarkm [Steve Henson] 3026109998Smarkm 3027109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 3028109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 3029109998Smarkm it complain about algorithm deselection that isn't recognised. 3030109998Smarkm [Richard Levitte] 3031109998Smarkm 3032109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 3033109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 3034109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 3035109998Smarkm some old style ASN1 functions: this can be used to determine if old 3036109998Smarkm code will still work when these eventually go away. 3037109998Smarkm [Steve Henson] 3038109998Smarkm 3039109998Smarkm *) New extension functions for OCSP structures, these follow the 3040109998Smarkm same conventions as certificates and CRLs. 3041109998Smarkm [Steve Henson] 3042109998Smarkm 3043109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 3044109998Smarkm adds an extension. Its behaviour can be customised with various 3045109998Smarkm flags to append, replace or delete. Various wrappers added for 3046109998Smarkm certifcates and CRLs. 3047109998Smarkm [Steve Henson] 3048109998Smarkm 3049109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 3050109998Smarkm an extension cannot be parsed. Correct a typo in the 3051109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 3052109998Smarkm [Steve Henson] 3053109998Smarkm 3054109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 3055109998Smarkm entries for variables. 3056109998Smarkm [Steve Henson] 3057109998Smarkm 3058109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 3059109998Smarkm problems: As the program is single-threaded, all we have 3060109998Smarkm to do is register a locking callback using an array for 3061109998Smarkm storing which locks are currently held by the program. 3062109998Smarkm [Bodo Moeller] 3063109998Smarkm 3064109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 3065109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 3066109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 3067109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 3068109998Smarkm Unfortunately, the ex_data design is not at all suited 3069109998Smarkm for multi-threaded use, so it probably should be abolished. 3070109998Smarkm [Bodo Moeller] 3071109998Smarkm 3072109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 3073109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 3074109998Smarkm 3075109998Smarkm *) Move common extension printing code to new function 3076109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 3077109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 3078109998Smarkm [Steve Henson] 3079109998Smarkm 3080109998Smarkm *) New function X509_signature_print() to remove duplication in some 3081109998Smarkm print routines. 3082109998Smarkm [Steve Henson] 3083109998Smarkm 3084109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 3085109998Smarkm set (this was treated exactly the same as SET OF previously). This 3086109998Smarkm is used to reorder the STACK representing the structure to match the 3087109998Smarkm encoding. This will be used to get round a problem where a PKCS7 3088109998Smarkm structure which was signed could not be verified because the STACK 3089109998Smarkm order did not reflect the encoded order. 3090109998Smarkm [Steve Henson] 3091109998Smarkm 3092109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 3093109998Smarkm [Steve Henson] 3094109998Smarkm 3095109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 3096109998Smarkm for its ASN1 operations. The old style function pointers still exist 3097109998Smarkm for now but they will eventually go away. 3098109998Smarkm [Steve Henson] 3099109998Smarkm 3100109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 3101109998Smarkm completely replaces the old ASN1 functionality with a table driven 3102109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 3103109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 3104109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 3105109998Smarkm has also been converted to the new form. 3106109998Smarkm [Steve Henson] 3107109998Smarkm 3108109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 3109109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 3110109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 3111109998Smarkm for negative moduli. 3112109998Smarkm [Bodo Moeller] 3113109998Smarkm 3114109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 3115109998Smarkm of not touching the result's sign bit. 3116109998Smarkm [Bodo Moeller] 3117109998Smarkm 3118109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 3119109998Smarkm set. 3120109998Smarkm [Bodo Moeller] 3121109998Smarkm 3122109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 3123109998Smarkm macros to declare and implement thin (optionally static) functions 3124109998Smarkm that provide type-safety and avoid function pointer casting for the 3125109998Smarkm type-specific callbacks. 3126109998Smarkm [Geoff Thorpe] 3127109998Smarkm 3128109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 3129109998Smarkm RFC 2712. 3130109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 3131109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 3132109998Smarkm 3133109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 3134109998Smarkm in sections depending on the subject. 3135109998Smarkm [Richard Levitte] 3136109998Smarkm 3137109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 3138109998Smarkm Windows. 3139109998Smarkm [Richard Levitte] 3140109998Smarkm 3141109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 3142109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 3143109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 3144109998Smarkm be handled deterministically). 3145109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3146109998Smarkm 3147109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 3148109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 3149109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 3150109998Smarkm [Bodo Moeller] 3151109998Smarkm 3152109998Smarkm *) New function BN_kronecker. 3153109998Smarkm [Bodo Moeller] 3154109998Smarkm 3155109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 3156109998Smarkm positive unless both parameters are zero. 3157109998Smarkm Previously something reasonably close to an infinite loop was 3158109998Smarkm possible because numbers could be growing instead of shrinking 3159109998Smarkm in the implementation of Euclid's algorithm. 3160109998Smarkm [Bodo Moeller] 3161109998Smarkm 3162109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 3163109998Smarkm sign of the number in question. 3164109998Smarkm 3165109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 3166109998Smarkm 3167109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 3168109998Smarkm because its test if the absolute value of 'a' equals 'w'. 3169109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 3170109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 3171109998Smarkm BN_is_one(), and BN_is_word(). 3172109998Smarkm [Bodo Moeller] 3173109998Smarkm 3174109998Smarkm *) New function BN_swap. 3175109998Smarkm [Bodo Moeller] 3176109998Smarkm 3177109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 3178109998Smarkm the exponentiation functions are more likely to produce reasonable 3179109998Smarkm results on negative inputs. 3180109998Smarkm [Bodo Moeller] 3181109998Smarkm 3182109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 3183109998Smarkm Previously, it could be negative if one of the factors was negative; 3184109998Smarkm I don't think anyone really wanted that behaviour. 3185109998Smarkm [Bodo Moeller] 3186109998Smarkm 3187109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 3188109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 3189109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 3190109998Smarkm and add new functions: 3191109998Smarkm 3192109998Smarkm BN_nnmod 3193109998Smarkm BN_mod_sqr 3194109998Smarkm BN_mod_add 3195109998Smarkm BN_mod_add_quick 3196109998Smarkm BN_mod_sub 3197109998Smarkm BN_mod_sub_quick 3198109998Smarkm BN_mod_lshift1 3199109998Smarkm BN_mod_lshift1_quick 3200109998Smarkm BN_mod_lshift 3201109998Smarkm BN_mod_lshift_quick 3202109998Smarkm 3203109998Smarkm These functions always generate non-negative results. 3204109998Smarkm 3205109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 3206109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 3207109998Smarkm 3208109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 3209109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 3210109998Smarkm be reduced modulo m. 3211109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 3212109998Smarkm 3213111147Snectar#if 0 3214111147Snectar The following entry accidentily appeared in the CHANGES file 3215111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 3216111147Snectar it do *not* apply to OpenSSL 0.9.7. 3217111147Snectar 3218109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3219109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 3220109998Smarkm required a small change in bn_mul_part_recursive() and the addition 3221109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 3222109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 3223109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 3224109998Smarkm differing sizes. 3225109998Smarkm [Richard Levitte] 3226111147Snectar#endif 3227109998Smarkm 3228109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 3229109998Smarkm unless the '-salt' option is used (which usually means that 3230109998Smarkm verification would just waste user's time since the resulting 3231109998Smarkm hash is going to be compared with some given password hash) 3232109998Smarkm or the new '-noverify' option is used. 3233109998Smarkm 3234109998Smarkm This is an incompatible change, but it does not affect 3235109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 3236109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 3237109998Smarkm cause any problems. 3238109998Smarkm [Bodo Moeller] 3239109998Smarkm 3240109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 3241109998Smarkm [Richard Levitte] 3242109998Smarkm 3243109998Smarkm *) Make DSO load along a path given through an environment variable 3244109998Smarkm (SHLIB_PATH) with shl_load(). 3245109998Smarkm [Richard Levitte] 3246109998Smarkm 3247109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 3248109998Smarkm Also constify the RSA code and most things related to it. In a 3249109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 3250109998Smarkm casts back to non-const were required (to be solved at a later 3251109998Smarkm time) 3252109998Smarkm [Richard Levitte] 3253109998Smarkm 3254109998Smarkm *) Make it so the openssl application has all engines loaded by default. 3255109998Smarkm [Richard Levitte] 3256109998Smarkm 3257109998Smarkm *) Constify the BIGNUM routines a little more. 3258109998Smarkm [Richard Levitte] 3259109998Smarkm 3260109998Smarkm *) Add the following functions: 3261109998Smarkm 3262109998Smarkm ENGINE_load_cswift() 3263109998Smarkm ENGINE_load_chil() 3264109998Smarkm ENGINE_load_atalla() 3265109998Smarkm ENGINE_load_nuron() 3266109998Smarkm ENGINE_load_builtin_engines() 3267109998Smarkm 3268109998Smarkm That way, an application can itself choose if external engines that 3269109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 3270109998Smarkm that applications won't have to be linked with libdl or other dso 3271109998Smarkm libraries unless it's really needed. 3272109998Smarkm 3273109998Smarkm Changed 'openssl engine' to load all engines on demand. 3274109998Smarkm Changed the engine header files to avoid the duplication of some 3275109998Smarkm declarations (they differed!). 3276109998Smarkm [Richard Levitte] 3277109998Smarkm 3278109998Smarkm *) 'openssl engine' can now list capabilities. 3279109998Smarkm [Richard Levitte] 3280109998Smarkm 3281109998Smarkm *) Better error reporting in 'openssl engine'. 3282109998Smarkm [Richard Levitte] 3283109998Smarkm 3284109998Smarkm *) Never call load_dh_param(NULL) in s_server. 3285109998Smarkm [Bodo Moeller] 3286109998Smarkm 3287109998Smarkm *) Add engine application. It can currently list engines by name and 3288109998Smarkm identity, and test if they are actually available. 3289109998Smarkm [Richard Levitte] 3290109998Smarkm 3291109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 3292109998Smarkm sure the installed documentation is also owned by root.root. 3293109998Smarkm [Damien Miller <djm@mindrot.org>] 3294109998Smarkm 3295109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 3296109998Smarkm keys (public as well as private) handled by engines. 3297109998Smarkm [Richard Levitte] 3298109998Smarkm 3299109998Smarkm *) Add OCSP code that comes from CertCo. 3300109998Smarkm [Richard Levitte] 3301109998Smarkm 3302109998Smarkm *) Add VMS support for the Rijndael code. 3303109998Smarkm [Richard Levitte] 3304109998Smarkm 3305109998Smarkm *) Added untested support for Nuron crypto accelerator. 3306109998Smarkm [Ben Laurie] 3307109998Smarkm 3308109998Smarkm *) Add support for external cryptographic devices. This code was 3309109998Smarkm previously distributed separately as the "engine" branch. 3310109998Smarkm [Geoff Thorpe, Richard Levitte] 3311109998Smarkm 3312109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 3313109998Smarkm have far greater control over how a "name" is turned into a filename 3314109998Smarkm depending on the operating environment and any oddities about the 3315109998Smarkm different shared library filenames on each system. 3316109998Smarkm [Geoff Thorpe] 3317109998Smarkm 3318109998Smarkm *) Support threads on FreeBSD-elf in Configure. 3319109998Smarkm [Richard Levitte] 3320109998Smarkm 3321109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 3322109998Smarkm warnings about corrupt line number information when assembling 3323109998Smarkm with debugging information. This is caused by the overlapping 3324109998Smarkm of two sections. 3325109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 3326109998Smarkm 3327109998Smarkm *) NCONF changes. 3328109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 3329109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 3330109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 3331109998Smarkm binary backward compatibility. 3332109998Smarkm Make it possible for methods to load from something other than a BIO, 3333109998Smarkm by providing a function pointer that is given a name instead of a BIO. 3334109998Smarkm For example, this could be used to load configuration data from an 3335109998Smarkm LDAP server. 3336109998Smarkm [Richard Levitte] 3337109998Smarkm 3338109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 3339109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 3340109998Smarkm with non blocking I/O was not possible because no retry code was 3341109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 3342109998Smarkm this case. 3343109998Smarkm [Steve Henson] 3344109998Smarkm 3345109998Smarkm *) Added the beginnings of Rijndael support. 3346109998Smarkm [Ben Laurie] 3347109998Smarkm 3348109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 3349109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 3350109998Smarkm to allow certificate printing to more controllable, additional 3351109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 3352109998Smarkm set. 3353109998Smarkm [Steve Henson] 3354109998Smarkm 3355109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 3356109998Smarkm [Richard Levitte] 3357109998Smarkm 3358142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 3359142425Snectar 3360142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3361160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3362142425Snectar [Joe Orton, Steve Henson] 3363142425Snectar 3364142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 3365142425Snectar 3366142425Snectar *) Fix additional bug revealed by the NISCC test suite: 3367142425Snectar 3368142425Snectar Stop bug triggering large recursion when presented with 3369160814Ssimon certain ASN.1 tags (CVE-2003-0851) 3370142425Snectar [Steve Henson] 3371142425Snectar 3372120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 3373120631Snectar 3374120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3375120631Snectar 3376120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3377160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3378120631Snectar 3379120631Snectar If verify callback ignores invalid public key errors don't try to check 3380120631Snectar certificate signature with the NULL public key. 3381120631Snectar 3382120631Snectar [Steve Henson] 3383120631Snectar 3384120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3385120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3386120631Snectar specifications. 3387120631Snectar [Steve Henson] 3388120631Snectar 3389120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3390120631Snectar extra data after the compression methods not only for TLS 1.0 3391120631Snectar but also for SSL 3.0 (as required by the specification). 3392120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3393120631Snectar 3394120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3395120631Snectar when it's 512 *bits* long, not 512 bytes. 3396120631Snectar [Richard Levitte] 3397120631Snectar 3398120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 3399120631Snectar 3400120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3401120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3402120631Snectar a protocol version number mismatch like a decryption error 3403120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3404120631Snectar [Bodo Moeller] 3405120631Snectar 3406120631Snectar *) Turn on RSA blinding by default in the default implementation 3407120631Snectar to avoid a timing attack. Applications that don't want it can call 3408120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3409120631Snectar They would be ill-advised to do so in most cases. 3410120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3411120631Snectar 3412120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 3413120631Snectar seeded (in this case, the secret RSA exponent is abused as 3414120631Snectar an unpredictable seed -- if it is not unpredictable, there 3415120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 3416120631Snectar by remembering the creator's thread ID in rsa->blinding and 3417120631Snectar having all other threads use local one-time blinding factors 3418120631Snectar (this requires more computation than sharing rsa->blinding, but 3419120631Snectar avoids excessive locking; and if an RSA object is not shared 3420120631Snectar between threads, blinding will still be very fast). 3421120631Snectar [Bodo Moeller] 3422120631Snectar 3423111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 3424111147Snectar 3425111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 3426111147Snectar via timing by performing a MAC computation even if incorrrect 3427111147Snectar block cipher padding has been found. This is a countermeasure 3428111147Snectar against active attacks where the attacker has to distinguish 3429160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 3430111147Snectar 3431111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 3432111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 3433111147Snectar Martin Vuagnoux (EPFL, Ilion)] 3434111147Snectar 3435109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 3436109998Smarkm 3437109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 3438109998Smarkm memory from it's contents. This is done with a counter that will 3439109998Smarkm place alternating values in each byte. This can be used to solve 3440109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 3441109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 3442109998Smarkm be read through on certain media, for example a swap space on disk. 3443109998Smarkm [Geoff Thorpe] 3444109998Smarkm 3445109998Smarkm *) Bugfix: client side session caching did not work with external caching, 3446109998Smarkm because the session->cipher setting was not restored when reloading 3447109998Smarkm from the external cache. This problem was masked, when 3448109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 3449109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 3450109998Smarkm [Lutz Jaenicke] 3451109998Smarkm 3452109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 3453109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 3454109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 3455109998Smarkm 3456109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 3457109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 3458109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 3459109998Smarkm EVP_cleanup(). 3460109998Smarkm [Richard Levitte] 3461109998Smarkm 3462109998Smarkm *) Change the default configuration reader to deal with last line not 3463109998Smarkm being properly terminated. 3464109998Smarkm [Richard Levitte] 3465109998Smarkm 3466109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 3467109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 3468109998Smarkm emailAddress where the value has the type ia5String. 3469109998Smarkm [stefank@valicert.com via Richard Levitte] 3470109998Smarkm 3471109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 3472109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 3473109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 3474109998Smarkm the bitwise-OR of the two for use by the majority of applications 3475109998Smarkm wanting this behaviour, and update the docs. The documented 3476109998Smarkm behaviour and actual behaviour were inconsistent and had been 3477109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 3478109998Smarkm change. 3479109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 3480109998Smarkm 3481109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 3482109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 3483109998Smarkm [Bodo Moeller] 3484109998Smarkm 3485109998Smarkm *) Fix initialization code race conditions in 3486109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 3487109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 3488109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 3489109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 3490109998Smarkm ssl2_get_cipher_by_char(), 3491109998Smarkm ssl3_get_cipher_by_char(). 3492109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 3493109998Smarkm 3494109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 3495109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 3496109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 3497109998Smarkm (see [openssl.org #212]). 3498109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 3499109998Smarkm 3500109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 3501109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 3502109998Smarkm [Steve Henson] 3503109998Smarkm 3504101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 3505101618Snectar 3506101618Snectar *) [In 0.9.6g-engine release:] 3507101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 3508101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 3509101618Snectar 3510101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 3511101613Snectar 3512101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 3513101613Snectar and get fix the header length calculation. 3514101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 3515101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 3516101613Snectar Steve Henson] 3517101613Snectar 3518101613Snectar *) Use proper error handling instead of 'assertions' in buffer 3519101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 3520101613Snectar assertions could call abort()). 3521101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 3522101613Snectar 3523100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 3524100936Snectar 3525109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 3526109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 3527109998Smarkm negative or the content length exceeds the length of the 3528109998Smarkm supplied buffer. 3529109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 3530109998Smarkm 3531100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 3532100936Snectar for the cipher strength set and where therefore not handled correctly 3533100936Snectar by the selection routines (PR #130). 3534100936Snectar [Lutz Jaenicke] 3535100936Snectar 3536100936Snectar *) Fix EVP_dsa_sha macro. 3537100936Snectar [Nils Larsch] 3538100936Snectar 3539100936Snectar *) New option 3540100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 3541100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 3542100936Snectar that was added in OpenSSL 0.9.6d. 3543100936Snectar 3544100936Snectar As the countermeasure turned out to be incompatible with some 3545100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 3546100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 3547100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 3548100936Snectar 's_server'), so the new option is automatically set in many 3549100936Snectar applications. 3550100936Snectar [Bodo Moeller] 3551100936Snectar 3552100936Snectar *) Changes in security patch: 3553100936Snectar 3554100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 3555100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 3556100936Snectar Air Force Materiel Command, USAF, under agreement number 3557100936Snectar F30602-01-2-0537. 3558100936Snectar 3559100936Snectar *) Add various sanity checks to asn1_get_length() to reject 3560100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 3561100936Snectar negative or the content length exceeds the length of the 3562160814Ssimon supplied buffer. (CVE-2002-0659) 3563100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 3564100936Snectar 3565100936Snectar *) Assertions for various potential buffer overflows, not known to 3566100936Snectar happen in practice. 3567100936Snectar [Ben Laurie (CHATS)] 3568100936Snectar 3569100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 3570160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 3571100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 3572100936Snectar 3573100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 3574160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 3575100936Snectar [Ben Laurie (CHATS)] 3576100936Snectar 3577100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 3578160814Ssimon supply an oversized client master key. (CVE-2002-0656) 3579100936Snectar [Ben Laurie (CHATS)] 3580100936Snectar 3581100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 3582100928Snectar 3583100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 3584100928Snectar encoded as NULL) with id-dsa-with-sha1. 3585100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 3586100928Snectar 3587100928Snectar *) Check various X509_...() return values in apps/req.c. 3588100928Snectar [Nils Larsch <nla@trustcenter.de>] 3589100928Snectar 3590100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 3591100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 3592100928Snectar was just at the end of a processed block. The bug was discovered when 3593100928Snectar processing data through a buffering memory BIO handing the data to a 3594100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 3595100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 3596100928Snectar [Lutz Jaenicke] 3597100928Snectar 3598100928Snectar *) Implement a countermeasure against a vulnerability recently found 3599100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 3600100928Snectar before application data chunks to avoid the use of known IVs 3601100928Snectar with data potentially chosen by the attacker. 3602100928Snectar [Bodo Moeller] 3603100928Snectar 3604100928Snectar *) Fix length checks in ssl3_get_client_hello(). 3605100928Snectar [Bodo Moeller] 3606100928Snectar 3607100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 3608100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 3609100928Snectar ssl3_read_bytes() found application data while handshake 3610100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 3611100928Snectar merely automatically cleared during the initial handshake. 3612100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 3613100928Snectar 3614100928Snectar *) Fix object definitions for Private and Enterprise: they were not 3615100928Snectar recognized in their shortname (=lowercase) representation. Extend 3616100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 3617100928Snectar of silently ignoring the problem (Svenning Sorensen 3618100928Snectar <sss@sss.dnsalias.net>). 3619100928Snectar [Lutz Jaenicke] 3620100928Snectar 3621100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 3622100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 3623100928Snectar code did not properly initialise the 'add' and 'rem' values to 3624100928Snectar BN_generate_prime().) 3625100928Snectar 3626100928Snectar In the new general case, we do not insist that 'generator' is 3627100928Snectar actually a primitive root: This requirement is rather pointless; 3628100928Snectar a generator of the order-q subgroup is just as good, if not 3629100928Snectar better. 3630100928Snectar [Bodo Moeller] 3631100928Snectar 3632100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 3633100928Snectar Tom Wu <tom@arcot.com>. 3634100928Snectar [Lutz Jaenicke] 3635100928Snectar 3636100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 3637100928Snectar returning non-zero before the data has been completely received 3638100928Snectar when using non-blocking I/O. 3639100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 3640100928Snectar 3641100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 3642100928Snectar [Ben Laurie, Lutz Jaenicke] 3643100928Snectar 3644100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 3645100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 3646100928Snectar [Lutz Jaenicke] 3647100928Snectar 3648100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 3649100928Snectar configuration for the versions before that. 3650100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 3651100928Snectar 3652100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 3653100928Snectar check whether we deal with a copy of a session and do not delete from 3654100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 3655100928Snectar <izhar@checkpoint.com>. 3656100928Snectar [Lutz Jaenicke] 3657100928Snectar 3658100928Snectar *) Do not store session data into the internal session cache, if it 3659100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 3660100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 3661100928Snectar [Lutz Jaenicke] 3662100928Snectar 3663100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 3664100928Snectar value is 0. 3665100928Snectar [Richard Levitte] 3666100928Snectar 3667109998Smarkm *) [In 0.9.6d-engine release:] 3668109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 3669100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 3670100928Snectar 3671100928Snectar *) Add the configuration target linux-s390x. 3672100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 3673100928Snectar 3674100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 3675100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 3676100928Snectar variable as an indication that a ClientHello message has been 3677100928Snectar received. As the flag value will be lost between multiple 3678100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 3679100928Snectar function may not be aware that a handshake has actually taken 3680100928Snectar place, thus preventing a new session from being added to the 3681100928Snectar session cache. 3682100928Snectar 3683100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 3684100928Snectar using a local variable. 3685100928Snectar [Lutz Jaenicke, Bodo Moeller] 3686100928Snectar 3687100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 3688100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 3689100928Snectar [Geoff Thorpe, Bodo Moeller] 3690100928Snectar 3691100928Snectar *) New 'shared_ldflag' column in Configure platform table. 3692100928Snectar [Richard Levitte] 3693100928Snectar 3694100928Snectar *) Fix EVP_CIPHER_mode macro. 3695100928Snectar ["Dan S. Camper" <dan@bti.net>] 3696100928Snectar 3697100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 3698100928Snectar type, we must throw them away by setting rr->length to 0. 3699100928Snectar [D P Chang <dpc@qualys.com>] 3700100928Snectar 370189837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 370289837Skris 370389837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 370489837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 370589837Skris worked incorrectly for those cases where range = 10..._2 and 370689837Skris 3*range is two bits longer than range.) 370789837Skris [Bodo Moeller] 370889837Skris 370989837Skris *) Only add signing time to PKCS7 structures if it is not already 371089837Skris present. 371189837Skris [Steve Henson] 371289837Skris 371389837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 371489837Skris OBJ_ld_ce should be OBJ_id_ce. 371589837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 371689837Skris incorrect (cf. RFC 3039). 371789837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 371889837Skris 371989837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 372089837Skris returns early because it has nothing to do. 372189837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 372289837Skris 372389837Skris *) [In 0.9.6c-engine release:] 372489837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 372589837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 372689837Skris 372789837Skris *) [In 0.9.6c-engine release:] 372889837Skris Add support for Cryptographic Appliance's keyserver technology. 372989837Skris (Use engine 'keyclient') 373089837Skris [Cryptographic Appliances and Geoff Thorpe] 373189837Skris 373289837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 373389837Skris is called via tools/c89.sh because arguments have to be 373489837Skris rearranged (all '-L' options must appear before the first object 373589837Skris modules). 373689837Skris [Richard Shapiro <rshapiro@abinitio.com>] 373789837Skris 373889837Skris *) [In 0.9.6c-engine release:] 373989837Skris Add support for Broadcom crypto accelerator cards, backported 374089837Skris from 0.9.7. 374189837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 374289837Skris 374389837Skris *) [In 0.9.6c-engine release:] 374489837Skris Add support for SureWare crypto accelerator cards from 374589837Skris Baltimore Technologies. (Use engine 'sureware') 374689837Skris [Baltimore Technologies and Mark Cox] 374789837Skris 374889837Skris *) [In 0.9.6c-engine release:] 374989837Skris Add support for crypto accelerator cards from Accelerated 375089837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 375189837Skris [AEP Inc. and Mark Cox] 375289837Skris 375389837Skris *) Add a configuration entry for gcc on UnixWare. 375489837Skris [Gary Benson <gbenson@redhat.com>] 375589837Skris 375689837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 375789837Skris messages are stored in a single piece (fixed-length part and 375889837Skris variable-length part combined) and fix various bugs found on the way. 375989837Skris [Bodo Moeller] 376089837Skris 376189837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 376289837Skris instead. BIO_gethostbyname() does not know what timeouts are 376389837Skris appropriate, so entries would stay in cache even when they have 376489837Skris become invalid. 376589837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 376689837Skris 376789837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 376889837Skris faced with a pathologically small ClientHello fragment that does 376989837Skris not contain client_version: Instead of aborting with an error, 377089837Skris simply choose the highest available protocol version (i.e., 377189837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 377289837Skris messages are never sent like this, but this change gives us 377389837Skris strictly correct behaviour at least for TLS. 377489837Skris [Bodo Moeller] 377589837Skris 377689837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 377789837Skris never resets s->method to s->ctx->method when called from within 377889837Skris one of the SSL handshake functions. 377989837Skris [Bodo Moeller; problem pointed out by Niko Baric] 378089837Skris 378189837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 378289837Skris (sent using the client's version number) if client_version is 378389837Skris smaller than the protocol version in use. Also change 378489837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 378589837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 378689837Skris the client will at least see that alert. 378789837Skris [Bodo Moeller] 378889837Skris 378989837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 379089837Skris correctly. 379189837Skris [Bodo Moeller] 379289837Skris 379389837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 379489837Skris client receives HelloRequest while in a handshake. 379589837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 379689837Skris 379789837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 379889837Skris should end in 'break', not 'goto end' which circuments various 379989837Skris cleanups done in state SSL_ST_OK. But session related stuff 380089837Skris must be disabled for SSL_ST_OK in the case that we just sent a 380189837Skris HelloRequest. 380289837Skris 380389837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 380489837Skris before just sending a HelloRequest. 380589837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 380689837Skris 380789837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 380889837Skris reveal whether illegal block cipher padding was found or a MAC 380989837Skris verification error occured. (Neither SSLerr() codes nor alerts 381089837Skris are directly visible to potential attackers, but the information 381189837Skris may leak via logfiles.) 381289837Skris 381389837Skris Similar changes are not required for the SSL 2.0 implementation 381489837Skris because the number of padding bytes is sent in clear for SSL 2.0, 381589837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 381689837Skris failed to verify that the purported number of padding bytes is in 381789837Skris the legal range. 381889837Skris [Bodo Moeller] 381989837Skris 382089837Skris *) Add OpenUNIX-8 support including shared libraries 382189837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 382289837Skris [Lutz Jaenicke] 382389837Skris 382489837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 382589837Skris 'wristwatch attack' using huge encoding parameters (cf. 382689837Skris James H. Manger's CRYPTO 2001 paper). Note that the 382789837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 382889837Skris encoding parameters and hence was not vulnerable. 382989837Skris [Bodo Moeller] 383089837Skris 383189837Skris *) BN_sqr() bug fix. 383289837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 383389837Skris 383489837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 383589837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 383689837Skris followed by modular reduction. 383789837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 383889837Skris 383989837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 384089837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 384189837Skris [Bodo Moeller] 384289837Skris 384389837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 384489837Skris This function was broken, as the check for a new client hello message 384589837Skris to handle SGC did not allow these large messages. 384689837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 384789837Skris [Lutz Jaenicke] 384889837Skris 384989837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 385089837Skris [Lutz Jaenicke] 385189837Skris 385289837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 385389837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 385489837Skris [Lutz Jaenicke] 385589837Skris 385689837Skris *) Rework the configuration and shared library support for Tru64 Unix. 385789837Skris The configuration part makes use of modern compiler features and 385889837Skris still retains old compiler behavior for those that run older versions 385989837Skris of the OS. The shared library support part includes a variant that 386089837Skris uses the RPATH feature, and is available through the special 386189837Skris configuration target "alpha-cc-rpath", which will never be selected 386289837Skris automatically. 386389837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 386489837Skris 386589837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 386689837Skris with the same message size as in ssl3_get_certificate_request(). 386789837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 386889837Skris messages might inadvertently be reject as too long. 386989837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 387089837Skris 387189837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 387289837Skris [Andy Polyakov] 387389837Skris 387489837Skris *) Modified SSL library such that the verify_callback that has been set 387589837Skris specificly for an SSL object with SSL_set_verify() is actually being 387689837Skris used. Before the change, a verify_callback set with this function was 387789837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 387889837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 387989837Skris to allow the necessary settings. 388089837Skris [Lutz Jaenicke] 388189837Skris 388289837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 388389837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 388489837Skris done automatically (in contradiction to the requirements of the C 388589837Skris standard). This made problems when used from OpenSSH. 388689837Skris [Lutz Jaenicke] 388789837Skris 388889837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 388989837Skris dh->length and always used 389089837Skris 389189837Skris BN_rand_range(priv_key, dh->p). 389289837Skris 389389837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 389489837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 389589837Skris dh->length (recommended exponent length) is much smaller than the 389689837Skris length of dh->p. We could use BN_rand_range() if the order of 389789837Skris the subgroup was stored in the DH structure, but we only have 389889837Skris dh->length. 389989837Skris 390089837Skris So switch back to 390189837Skris 390289837Skris BN_rand(priv_key, l, ...) 390389837Skris 390489837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 390589837Skris otherwise. 390689837Skris [Bodo Moeller] 390789837Skris 390889837Skris *) In 390989837Skris 391089837Skris RSA_eay_public_encrypt 391189837Skris RSA_eay_private_decrypt 391289837Skris RSA_eay_private_encrypt (signing) 391389837Skris RSA_eay_public_decrypt (signature verification) 391489837Skris 391589837Skris (default implementations for RSA_public_encrypt, 391689837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 391789837Skris always reject numbers >= n. 391889837Skris [Bodo Moeller] 391989837Skris 392089837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 392189837Skris to synchronize access to 'locking_thread'. This is necessary on 392289837Skris systems where access to 'locking_thread' (an 'unsigned long' 392389837Skris variable) is not atomic. 392489837Skris [Bodo Moeller] 392589837Skris 392689837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 392789837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 392889837Skris a race condition if 0 is a valid thread ID. 392989837Skris [Travis Vitek <vitek@roguewave.com>] 393089837Skris 393189837Skris *) Add support for shared libraries under Irix. 393289837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 393389837Skris 393489837Skris *) Add configuration option to build on Linux on both big-endian and 393589837Skris little-endian MIPS. 393689837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 393789837Skris 393889837Skris *) Add the possibility to create shared libraries on HP-UX. 393989837Skris [Richard Levitte] 394089837Skris 394179998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 394279998Skris 394379998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 394479998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 394579998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 394679998Skris PRNG state recovery was possible based on the output of 394779998Skris one PRNG request appropriately sized to gain knowledge on 394879998Skris 'md' followed by enough consecutive 1-byte PRNG requests 394979998Skris to traverse all of 'state'. 395079998Skris 395179998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 395279998Skris during PRNG output generation, hash all of the previous 395379998Skris 'md_local' value, not just the half used for PRNG output. 395479998Skris 395579998Skris 2. Make the number of bytes from 'state' included into the hash 395679998Skris independent from the number of PRNG bytes requested. 395779998Skris 395879998Skris The first measure alone would be sufficient to avoid 395979998Skris Markku-Juhani's attack. (Actually it had never occurred 396079998Skris to me that the half of 'md_local' used for chaining was the 396179998Skris half from which PRNG output bytes were taken -- I had always 396279998Skris assumed that the secret half would be used.) The second 396379998Skris measure makes sure that additional data from 'state' is never 396479998Skris mixed into 'md_local' in small portions; this heuristically 396579998Skris further strengthens the PRNG. 396679998Skris [Bodo Moeller] 396779998Skris 396879998Skris *) Fix crypto/bn/asm/mips3.s. 396979998Skris [Andy Polyakov] 397079998Skris 397179998Skris *) When only the key is given to "enc", the IV is undefined. Print out 397279998Skris an error message in this case. 397379998Skris [Lutz Jaenicke] 397479998Skris 397579998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 397679998Skris [Steve Henson] 397779998Skris 397879998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 397979998Skris positive and less than q. 398079998Skris [Bodo Moeller] 398179998Skris 398279998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 398379998Skris used: it isn't thread safe and the add_lock_callback should handle 398479998Skris that itself. 398579998Skris [Paul Rose <Paul.Rose@bridge.com>] 398679998Skris 398779998Skris *) Verify that incoming data obeys the block size in 398879998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 398979998Skris [Bodo Moeller] 399079998Skris 399179998Skris *) Fix OAEP check. 399279998Skris [Ulf M�ller, Bodo M�ller] 399379998Skris 399479998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 399589837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 399679998Skris when fixing the server behaviour for backwards-compatible 'client 399779998Skris hello' messages. (Note that the attack is impractical against 399879998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 399979998Skris means that the probability of guessing a valid ciphertext is 400079998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 400179998Skris paper.) 400279998Skris 400379998Skris Before 0.9.5, the countermeasure (hide the error by generating a 400479998Skris random 'decryption result') did not work properly because 400579998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 400679998Skris detect the supposedly ignored error. 400779998Skris 400879998Skris Both problems are now fixed. 400979998Skris [Bodo Moeller] 401079998Skris 401179998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 401279998Skris (previously it was 1024). 401379998Skris [Bodo Moeller] 401479998Skris 401579998Skris *) Fix for compatibility mode trust settings: ignore trust settings 401679998Skris unless some valid trust or reject settings are present. 401779998Skris [Steve Henson] 401879998Skris 401979998Skris *) Fix for blowfish EVP: its a variable length cipher. 402079998Skris [Steve Henson] 402179998Skris 402279998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 402379998Skris parameters in DSA public key structures and return an error in the 402479998Skris DSA routines if parameters are absent. 402579998Skris [Steve Henson] 402679998Skris 402779998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 402879998Skris in the current directory if neither $RANDFILE nor $HOME was set. 402979998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 403079998Skris caused some confusion to Windows users who haven't defined $HOME. 403179998Skris Thus RAND_file_name() is changed again: e_os.h can define a 403279998Skris DEFAULT_HOME, which will be used if $HOME is not set. 403379998Skris For Windows, we use "C:"; on other platforms, we still require 403479998Skris environment variables. 403579998Skris 403679998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 403779998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 403879998Skris having multiple threads call RAND_poll() concurrently. 403979998Skris [Bodo Moeller] 404079998Skris 404179998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 404279998Skris combination of a flag and a thread ID variable. 404379998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 404479998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 404589837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 404679998Skris that they do not hold after the first thread unsets add_do_not_lock). 404779998Skris [Bodo Moeller] 404879998Skris 404979998Skris *) Change bctest again: '-x' expressions are not available in all 405079998Skris versions of 'test'. 405179998Skris [Bodo Moeller] 405279998Skris 405376866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 405472613Skris 405576866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 405676866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 405776866Skris 405876866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 405976866Skris the default extension for executables, if any. Also, make the perl 406076866Skris scripts that use symlink() to test if it really exists and use "cp" 406176866Skris if it doesn't. All this made OpenSSL compilable and installable in 406276866Skris CygWin. 406376866Skris [Richard Levitte] 406476866Skris 406576866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 406676866Skris If SEQUENCE is length is indefinite just set c->slen to the total 406776866Skris amount of data available. 406876866Skris [Steve Henson, reported by shige@FreeBSD.org] 406976866Skris [This change does not apply to 0.9.7.] 407076866Skris 407176866Skris *) Change bctest to avoid here-documents inside command substitution 407276866Skris (workaround for FreeBSD /bin/sh bug). 407376866Skris For compatibility with Ultrix, avoid shell functions (introduced 407476866Skris in the bctest version that searches along $PATH). 407576866Skris [Bodo Moeller] 407676866Skris 407776866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 407876866Skris with des_encrypt() defined on some operating systems, like Solaris 407976866Skris and UnixWare. 408076866Skris [Richard Levitte] 408176866Skris 408276866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 408376866Skris On the Importance of Eliminating Errors in Cryptographic 408476866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 408576866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 408676866Skris [Ulf Moeller] 408776866Skris 408876866Skris *) MIPS assembler BIGNUM division bug fix. 408976866Skris [Andy Polyakov] 409076866Skris 409176866Skris *) Disabled incorrect Alpha assembler code. 409276866Skris [Richard Levitte] 409376866Skris 409476866Skris *) Fix PKCS#7 decode routines so they correctly update the length 409576866Skris after reading an EOC for the EXPLICIT tag. 409676866Skris [Steve Henson] 409776866Skris [This change does not apply to 0.9.7.] 409876866Skris 409976866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 410076866Skris if a 3DES key was generated with a 0 initial byte. Include 410176866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 410276866Skris (but broken) behaviour. 410376866Skris [Steve Henson] 410476866Skris 410576866Skris *) Enhance bctest to search for a working bc along $PATH and print 410676866Skris it when found. 410776866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 410876866Skris 410976866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 411076866Skris don't write to the wrong index in ERR_set_error_data. 411176866Skris [Bodo Moeller] 411276866Skris 411376866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 411476866Skris did not exist. 411576866Skris [Bodo Moeller] 411676866Skris 411776866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 411876866Skris [Jeremy Cooper <jeremy@baymoo.org>] 411976866Skris 412076866Skris *) Make it possible to reuse SSLv2 sessions. 412176866Skris [Richard Levitte] 412276866Skris 412376866Skris *) In copy_email() check for >= 0 as a return value for 412476866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 412576866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 412676866Skris 412776866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 412876866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 412976866Skris PKCS7_verify() fails with non detached data. 413076866Skris [Steve Henson] 413176866Skris 413276866Skris *) Don't use getenv in library functions when run as setuid/setgid. 413376866Skris New function OPENSSL_issetugid(). 413476866Skris [Ulf Moeller] 413576866Skris 413676866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 413776866Skris due to incorrect handling of multi-threading: 413876866Skris 413976866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 414076866Skris 414176866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 414276866Skris 414376866Skris 3. Count how many times MemCheck_off() has been called so that 414476866Skris nested use can be treated correctly. This also avoids 414576866Skris inband-signalling in the previous code (which relied on the 414676866Skris assumption that thread ID 0 is impossible). 414776866Skris [Bodo Moeller] 414876866Skris 414976866Skris *) Add "-rand" option also to s_client and s_server. 415076866Skris [Lutz Jaenicke] 415176866Skris 415276866Skris *) Fix CPU detection on Irix 6.x. 415376866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 415476866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 415576866Skris 415676866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 415776866Skris was empty. 415876866Skris [Steve Henson] 415976866Skris [This change does not apply to 0.9.7.] 416076866Skris 416176866Skris *) Use the cached encoding of an X509_NAME structure rather than 416276866Skris copying it. This is apparently the reason for the libsafe "errors" 416376866Skris but the code is actually correct. 416476866Skris [Steve Henson] 416576866Skris 416672613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 416772613Skris Bleichenbacher's DSA attack. 416876866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 416976866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 417076866Skris and leaves the highest bit random. 417176866Skris [Ulf Moeller, Bodo Moeller] 417272613Skris 417372613Skris *) In the NCONF_...-based implementations for CONF_... queries 417472613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 417572613Skris a temporary CONF structure with the data component set to NULL 417672613Skris (which gives segmentation faults in lh_retrieve). 417772613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 417872613Skris CONF_get_number (which may use environment variables) and directly 417972613Skris return NULL from CONF_get_section. 418072613Skris [Bodo Moeller] 418172613Skris 418272613Skris *) Fix potential buffer overrun for EBCDIC. 418372613Skris [Ulf Moeller] 418472613Skris 418572613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 418672613Skris keyUsage if basicConstraints absent for a CA. 418772613Skris [Steve Henson] 418872613Skris 418972613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 419072613Skris is more generally accepted (no spaces before the semicolon), since 419172613Skris some programs can't parse those values properly otherwise. Also make 419272613Skris sure BIO's that break lines after each write do not create invalid 419372613Skris headers. 419472613Skris [Richard Levitte] 419572613Skris 419672613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 419772613Skris macros previously used would not encode an empty SEQUENCE OF 419872613Skris and break the signature. 419972613Skris [Steve Henson] 420076866Skris [This change does not apply to 0.9.7.] 420172613Skris 420272613Skris *) Zero the premaster secret after deriving the master secret in 420372613Skris DH ciphersuites. 420472613Skris [Steve Henson] 420572613Skris 420672613Skris *) Add some EVP_add_digest_alias registrations (as found in 420772613Skris OpenSSL_add_all_digests()) to SSL_library_init() 420872613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 420972613Skris compatibility with peers using X.509 certificates 421072613Skris with unconventional AlgorithmIdentifier OIDs. 421172613Skris [Bodo Moeller] 421272613Skris 421372613Skris *) Fix for Irix with NO_ASM. 421472613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 421572613Skris 421672613Skris *) ./config script fixes. 421772613Skris [Ulf Moeller, Richard Levitte] 421872613Skris 421972613Skris *) Fix 'openssl passwd -1'. 422072613Skris [Bodo Moeller] 422172613Skris 422272613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 422372613Skris terminated strings whose length is passed in the passlen 422472613Skris parameter, for example from PEM callbacks. This was done 422572613Skris by adding an extra length parameter to asc2uni(). 422672613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 422772613Skris 422872613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 422972613Skris call failed, free the DSA structure. 423072613Skris [Bodo Moeller] 423172613Skris 423272613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 423372613Skris These are present in some PKCS#12 files. 423472613Skris [Steve Henson] 423572613Skris 423672613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 423772613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 423872613Skris when writing a 32767 byte record. 423972613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 424072613Skris 424172613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 424272613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 424372613Skris 424472613Skris (RSA objects have a reference count access to which is protected 424572613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 424672613Skris so they are meant to be shared between threads.) 424772613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 424872613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 424972613Skris 425072613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 425172613Skris [Bodo Moeller] 425272613Skris 425376866Skris *) Use better test patterns in bntest. 425476866Skris [Ulf M�ller] 425576866Skris 425672613Skris *) rand_win.c fix for Borland C. 425772613Skris [Ulf M�ller] 425872613Skris 425972613Skris *) BN_rshift bugfix for n == 0. 426072613Skris [Bodo Moeller] 426172613Skris 426276866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 426376866Skris so that 'make test' does not abort just because 'bc' is broken. 426476866Skris [Bodo Moeller] 426576866Skris 426672613Skris *) Store verify_result within SSL_SESSION also for client side to 426772613Skris avoid potential security hole. (Re-used sessions on the client side 426872613Skris always resulted in verify_result==X509_V_OK, not using the original 426972613Skris result of the server certificate verification.) 427072613Skris [Lutz Jaenicke] 427172613Skris 427272613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 427372613Skris SSL3_RT_APPLICATION_DATA, return 0. 427472613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 427572613Skris [Bodo Moeller] 427672613Skris 427772613Skris *) Fix SSL_peek: 427872613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 427972613Skris releases, have been re-implemented by renaming the previous 428072613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 428172613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 428272613Skris to them. The new ssl[23]_{read,peek} functions are calls to 428372613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 428472613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 428572613Skris does the actual work for ssl3_read_internal. 428672613Skris [Bodo Moeller] 428772613Skris 428876866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 428976866Skris the method-specific "init()" handler. Also clean up ex_data after 429076866Skris calling the method-specific "finish()" handler. Previously, this was 429176866Skris happening the other way round. 429276866Skris [Geoff Thorpe] 429376866Skris 429472613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 429572613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 429672613Skris [Bodo Moeller] 429772613Skris 429876866Skris *) Make sure that shared libraries get the internal name engine with 429976866Skris the full version number and not just 0. This should mark the 430076866Skris shared libraries as not backward compatible. Of course, this should 430176866Skris be changed again when we can guarantee backward binary compatibility. 430276866Skris [Richard Levitte] 430376866Skris 430472613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 430572613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 430672613Skris 430776866Skris *) Rework the system to generate shared libraries: 430876866Skris 430976866Skris - Make note of the expected extension for the shared libraries and 431076866Skris if there is a need for symbolic links from for example libcrypto.so.0 431176866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 431276866Skris that. 431376866Skris 431476866Skris - Make as few rebuilds of the shared libraries as possible. 431576866Skris 431676866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 431776866Skris 431876866Skris - When installing, install the shared libraries separately from the 431976866Skris static ones. 432076866Skris [Richard Levitte] 432176866Skris 432272613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 432372613Skris 432472613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 432572613Skris and not in SSL_clear because the latter is also used by the 432672613Skris accept/connect functions; previously, the settings made by 432772613Skris SSL_set_read_ahead would be lost during the handshake. 432872613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 432972613Skris 433072613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 433172613Skris Previously, it would create entries for disableed algorithms no 433272613Skris matter what. 433372613Skris [Richard Levitte] 433472613Skris 433572613Skris *) Added several new manual pages for SSL_* function. 433672613Skris [Lutz Jaenicke] 433772613Skris 433868651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 433968651Skris 434068651Skris *) In ssl23_get_client_hello, generate an error message when faced 434168651Skris with an initial SSL 3.0/TLS record that is too small to contain the 434268651Skris first two bytes of the ClientHello message, i.e. client_version. 434368651Skris (Note that this is a pathologic case that probably has never happened 434468651Skris in real life.) The previous approach was to use the version number 434568651Skris from the record header as a substitute; but our protocol choice 434668651Skris should not depend on that one because it is not authenticated 434768651Skris by the Finished messages. 434868651Skris [Bodo Moeller] 434968651Skris 435068651Skris *) More robust randomness gathering functions for Windows. 435168651Skris [Jeffrey Altman <jaltman@columbia.edu>] 435268651Skris 435368651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 435468651Skris not set then we don't setup the error code for issuer check errors 435568651Skris to avoid possibly overwriting other errors which the callback does 435668651Skris handle. If an application does set the flag then we assume it knows 435768651Skris what it is doing and can handle the new informational codes 435868651Skris appropriately. 435968651Skris [Steve Henson] 436068651Skris 436168651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 436268651Skris a general "ANY" type, as such it should be able to decode anything 436368651Skris including tagged types. However it didn't check the class so it would 436468651Skris wrongly interpret tagged types in the same way as their universal 436568651Skris counterpart and unknown types were just rejected. Changed so that the 436668651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 436768651Skris that is the encoding is stored intact. There is also a new type 436868651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 436968651Skris case we have no idea what the actual type is so we just lump them all 437068651Skris together. 437168651Skris [Steve Henson] 437268651Skris 437368651Skris *) On VMS, stdout may very well lead to a file that is written to 437468651Skris in a record-oriented fashion. That means that every write() will 437568651Skris write a separate record, which will be read separately by the 437668651Skris programs trying to read from it. This can be very confusing. 437768651Skris 437868651Skris The solution is to put a BIO filter in the way that will buffer 437968651Skris text until a linefeed is reached, and then write everything a 438068651Skris line at a time, so every record written will be an actual line, 438168651Skris not chunks of lines and not (usually doesn't happen, but I've 438268651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 438368651Skris the answer. 438468651Skris 438568651Skris Currently, it's a VMS-only method, because that's where it has 438668651Skris been tested well enough. 438768651Skris [Richard Levitte] 438868651Skris 438968651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 439068651Skris it can return incorrect results. 439168651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 439268651Skris but it was in 0.9.6-beta[12].) 439368651Skris [Bodo Moeller] 439468651Skris 439568651Skris *) Disable the check for content being present when verifying detached 439668651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 439768651Skris include zero length content when signing messages. 439868651Skris [Steve Henson] 439968651Skris 440068651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 440168651Skris BIO_ctrl (for BIO pairs). 440268651Skris [Bodo M�ller] 440368651Skris 440468651Skris *) Add DSO method for VMS. 440568651Skris [Richard Levitte] 440668651Skris 440768651Skris *) Bug fix: Montgomery multiplication could produce results with the 440868651Skris wrong sign. 440968651Skris [Ulf M�ller] 441068651Skris 441168651Skris *) Add RPM specification openssl.spec and modify it to build three 441268651Skris packages. The default package contains applications, application 441368651Skris documentation and run-time libraries. The devel package contains 441468651Skris include files, static libraries and function documentation. The 441568651Skris doc package contains the contents of the doc directory. The original 441668651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 441768651Skris [Richard Levitte] 441868651Skris 441968651Skris *) Add a large number of documentation files for many SSL routines. 442068651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 442168651Skris 442268651Skris *) Add a configuration entry for Sony News 4. 442368651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 442468651Skris 442568651Skris *) Don't set the two most significant bits to one when generating a 442668651Skris random number < q in the DSA library. 442768651Skris [Ulf M�ller] 442868651Skris 442968651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 443068651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 443168651Skris the underlying transport is blocking) if a handshake took place. 443268651Skris (The default behaviour is needed by applications such as s_client 443368651Skris and s_server that use select() to determine when to use SSL_read; 443468651Skris but for applications that know in advance when to expect data, it 443568651Skris just makes things more complicated.) 443668651Skris [Bodo Moeller] 443768651Skris 443868651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 443968651Skris from EGD. 444068651Skris [Ben Laurie] 444168651Skris 444268651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 444368651Skris work better on such systems. 444468651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 444568651Skris 444668651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 444768651Skris Update PKCS12_parse() so it copies the friendlyName and the 444868651Skris keyid to the certificates aux info. 444968651Skris [Steve Henson] 445068651Skris 445168651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 445268651Skris if there was more than one signature. 445368651Skris [Sven Uszpelkat <su@celocom.de>] 445468651Skris 445568651Skris *) Major change in util/mkdef.pl to include extra information 445668651Skris about each symbol, as well as presentig variables as well 445768651Skris as functions. This change means that there's n more need 445868651Skris to rebuild the .num files when some algorithms are excluded. 445968651Skris [Richard Levitte] 446068651Skris 446168651Skris *) Allow the verify time to be set by an application, 446268651Skris rather than always using the current time. 446368651Skris [Steve Henson] 446468651Skris 446568651Skris *) Phase 2 verify code reorganisation. The certificate 446668651Skris verify code now looks up an issuer certificate by a 446768651Skris number of criteria: subject name, authority key id 446868651Skris and key usage. It also verifies self signed certificates 446968651Skris by the same criteria. The main comparison function is 447068651Skris X509_check_issued() which performs these checks. 447168651Skris 447268651Skris Lot of changes were necessary in order to support this 447368651Skris without completely rewriting the lookup code. 447468651Skris 447568651Skris Authority and subject key identifier are now cached. 447668651Skris 447768651Skris The LHASH 'certs' is X509_STORE has now been replaced 447868651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 447968651Skris LHASH can't store or retrieve multiple objects with 448068651Skris the same hash value. 448168651Skris 448268651Skris As a result various functions (which were all internal 448368651Skris use only) have changed to handle the new X509_STORE 448468651Skris structure. This will break anything that messed round 448568651Skris with X509_STORE internally. 448668651Skris 448768651Skris The functions X509_STORE_add_cert() now checks for an 448868651Skris exact match, rather than just subject name. 448968651Skris 449068651Skris The X509_STORE API doesn't directly support the retrieval 449168651Skris of multiple certificates matching a given criteria, however 449268651Skris this can be worked round by performing a lookup first 449368651Skris (which will fill the cache with candidate certificates) 449468651Skris and then examining the cache for matches. This is probably 449568651Skris the best we can do without throwing out X509_LOOKUP 449668651Skris entirely (maybe later...). 449768651Skris 449868651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 449968651Skris 450068651Skris All certificate lookup operations now go via a get_issuer() 450168651Skris callback. Although this currently uses an X509_STORE it 450268651Skris can be replaced by custom lookups. This is a simple way 450368651Skris to bypass the X509_STORE hackery necessary to make this 450468651Skris work and makes it possible to use more efficient techniques 450568651Skris in future. A very simple version which uses a simple 450668651Skris STACK for its trusted certificate store is also provided 450768651Skris using X509_STORE_CTX_trusted_stack(). 450868651Skris 450968651Skris The verify_cb() and verify() callbacks now have equivalents 451068651Skris in the X509_STORE_CTX structure. 451168651Skris 451268651Skris X509_STORE_CTX also has a 'flags' field which can be used 451368651Skris to customise the verify behaviour. 451468651Skris [Steve Henson] 451568651Skris 451668651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 451768651Skris excludes S/MIME capabilities. 451868651Skris [Steve Henson] 451968651Skris 452068651Skris *) When a certificate request is read in keep a copy of the 452168651Skris original encoding of the signed data and use it when outputing 452268651Skris again. Signatures then use the original encoding rather than 452368651Skris a decoded, encoded version which may cause problems if the 452468651Skris request is improperly encoded. 452568651Skris [Steve Henson] 452668651Skris 452768651Skris *) For consistency with other BIO_puts implementations, call 452868651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 452968651Skris BIO_write(b, ...). 453068651Skris 453168651Skris In BIO_puts, increment b->num_write as in BIO_write. 453268651Skris [Peter.Sylvester@EdelWeb.fr] 453368651Skris 453468651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 453568651Skris BN_zero, we may not return a BIGNUM with an array consisting of 453668651Skris words set to zero.) 453768651Skris [Bodo Moeller] 453868651Skris 453968651Skris *) Avoid calling abort() from within the library when problems are 454068651Skris detected, except if preprocessor symbols have been defined 454168651Skris (such as REF_CHECK, BN_DEBUG etc.). 454268651Skris [Bodo Moeller] 454368651Skris 454468651Skris *) New openssl application 'rsautl'. This utility can be 454568651Skris used for low level RSA operations. DER public key 454668651Skris BIO/fp routines also added. 454768651Skris [Steve Henson] 454868651Skris 454968651Skris *) New Configure entry and patches for compiling on QNX 4. 455068651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 455168651Skris 455268651Skris *) A demo state-machine implementation was sponsored by 455368651Skris Nuron (http://www.nuron.com/) and is now available in 455468651Skris demos/state_machine. 455568651Skris [Ben Laurie] 455668651Skris 455768651Skris *) New options added to the 'dgst' utility for signature 455868651Skris generation and verification. 455968651Skris [Steve Henson] 456068651Skris 456168651Skris *) Unrecognized PKCS#7 content types are now handled via a 456268651Skris catch all ASN1_TYPE structure. This allows unsupported 456368651Skris types to be stored as a "blob" and an application can 456468651Skris encode and decode it manually. 456568651Skris [Steve Henson] 456668651Skris 456768651Skris *) Fix various signed/unsigned issues to make a_strex.c 456868651Skris compile under VC++. 456968651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 457068651Skris 457168651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 457268651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 457368651Skris if passed a NULL BN and its argument was negative. 457468651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 457568651Skris 457668651Skris *) Modification to PKCS#7 encoding routines to output definite 457768651Skris length encoding. Since currently the whole structures are in 457868651Skris memory there's not real point in using indefinite length 457968651Skris constructed encoding. However if OpenSSL is compiled with 458068651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 458168651Skris [Steve Henson] 458268651Skris 458368651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 458468651Skris [Richard Levitte] 458568651Skris 458668651Skris *) Added more prefixes to parse for in the the strings written 458768651Skris through a logging bio, to cover all the levels that are available 458868651Skris through syslog. The prefixes are now: 458968651Skris 459068651Skris PANIC, EMERG, EMR => LOG_EMERG 459168651Skris ALERT, ALR => LOG_ALERT 459268651Skris CRIT, CRI => LOG_CRIT 459368651Skris ERROR, ERR => LOG_ERR 459468651Skris WARNING, WARN, WAR => LOG_WARNING 459568651Skris NOTICE, NOTE, NOT => LOG_NOTICE 459668651Skris INFO, INF => LOG_INFO 459768651Skris DEBUG, DBG => LOG_DEBUG 459868651Skris 459968651Skris and as before, if none of those prefixes are present at the 460068651Skris beginning of the string, LOG_ERR is chosen. 460168651Skris 460268651Skris On Win32, the LOG_* levels are mapped according to this: 460368651Skris 460468651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 460568651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 460668651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 460768651Skris 460868651Skris [Richard Levitte] 460968651Skris 461068651Skris *) Made it possible to reconfigure with just the configuration 461168651Skris argument "reconf" or "reconfigure". The command line arguments 461268651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 461368651Skris and are retrieved from there when reconfiguring. 461468651Skris [Richard Levitte] 461568651Skris 461668651Skris *) MD4 implemented. 461768651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 461868651Skris 461968651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 462068651Skris [Richard Levitte] 462168651Skris 462268651Skris *) The obj_dat.pl script was messing up the sorting of object 462368651Skris names. The reason was that it compared the quoted version 462468651Skris of strings as a result "OCSP" > "OCSP Signing" because 462568651Skris " > SPACE. Changed script to store unquoted versions of 462668651Skris names and add quotes on output. It was also omitting some 462768651Skris names from the lookup table if they were given a default 462868651Skris value (that is if SN is missing it is given the same 462968651Skris value as LN and vice versa), these are now added on the 463068651Skris grounds that if an object has a name we should be able to 463168651Skris look it up. Finally added warning output when duplicate 463268651Skris short or long names are found. 463368651Skris [Steve Henson] 463468651Skris 463568651Skris *) Changes needed for Tandem NSK. 463668651Skris [Scott Uroff <scott@xypro.com>] 463768651Skris 463868651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 463968651Skris RSA_padding_check_SSLv23(), special padding was never detected 464068651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 464168651Skris version rollback attacks was not effective. 464268651Skris 464368651Skris In s23_clnt.c, don't use special rollback-attack detection padding 464468651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 464568651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 464668651Skris SSL 2.0 is the only protocol enabled in the server. 464768651Skris [Bodo Moeller] 464868651Skris 464968651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 465068651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 465168651Skris BIO_dump_indent() are added. 465268651Skris [Richard Levitte] 465368651Skris 465468651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 465568651Skris these print out strings and name structures based on various 465668651Skris flags including RFC2253 support and proper handling of 465768651Skris multibyte characters. Added options to the 'x509' utility 465868651Skris to allow the various flags to be set. 465968651Skris [Steve Henson] 466068651Skris 466168651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 466268651Skris Also change the functions X509_cmp_current_time() and 466368651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 466468651Skris this will enable certificates using GeneralizedTime in validity 466568651Skris dates to be checked. 466668651Skris [Steve Henson] 466768651Skris 466868651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 466968651Skris negative public key encodings) on by default, 467068651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 467168651Skris [Steve Henson] 467268651Skris 467368651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 467468651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 467568651Skris the encoding can be trivially obtained from the structure. 467668651Skris [Steve Henson] 467768651Skris 467868651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 467968651Skris not read locks (CRYPTO_r_[un]lock). 468068651Skris [Bodo Moeller] 468168651Skris 468268651Skris *) A first attempt at creating official support for shared 468368651Skris libraries through configuration. I've kept it so the 468468651Skris default is static libraries only, and the OpenSSL programs 468568651Skris are always statically linked for now, but there are 468668651Skris preparations for dynamic linking in place. 468789837Skris This has been tested on Linux and Tru64. 468868651Skris [Richard Levitte] 468968651Skris 469068651Skris *) Randomness polling function for Win9x, as described in: 469168651Skris Peter Gutmann, Software Generation of Practically Strong 469268651Skris Random Numbers. 469368651Skris [Ulf M�ller] 469468651Skris 469568651Skris *) Fix so PRNG is seeded in req if using an already existing 469668651Skris DSA key. 469768651Skris [Steve Henson] 469868651Skris 469968651Skris *) New options to smime application. -inform and -outform 470068651Skris allow alternative formats for the S/MIME message including 470168651Skris PEM and DER. The -content option allows the content to be 470268651Skris specified separately. This should allow things like Netscape 470368651Skris form signing output easier to verify. 470468651Skris [Steve Henson] 470568651Skris 470668651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 470768651Skris [Steve Henson] 470868651Skris 470968651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 471068651Skris STRING types. These convert content octets to and from the 471168651Skris underlying type. The actual tag and length octets are 471268651Skris already assumed to have been read in and checked. These 471368651Skris are needed because all other string types have virtually 471468651Skris identical handling apart from the tag. By having versions 471568651Skris of the ASN1 functions that just operate on content octets 471668651Skris IMPLICIT tagging can be handled properly. It also allows 471768651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 471868651Skris and ASN1_INTEGER are identical apart from the tag. 471968651Skris [Steve Henson] 472068651Skris 472168651Skris *) Change the handling of OID objects as follows: 472268651Skris 472368651Skris - New object identifiers are inserted in objects.txt, following 472468651Skris the syntax given in objects.README. 472568651Skris - objects.pl is used to process obj_mac.num and create a new 472668651Skris obj_mac.h. 472768651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 472868651Skris obj_mac.h. 472968651Skris 473068651Skris This is currently kind of a hack, and the perl code in objects.pl 473168651Skris isn't very elegant, but it works as I intended. The simplest way 473268651Skris to check that it worked correctly is to look in obj_dat.h and 473368651Skris check the array nid_objs and make sure the objects haven't moved 473468651Skris around (this is important!). Additions are OK, as well as 473568651Skris consistent name changes. 473668651Skris [Richard Levitte] 473768651Skris 473868651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 473968651Skris [Bodo Moeller] 474068651Skris 474168651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 474268651Skris The given file adds to whatever has already been seeded into the 474368651Skris random pool through the RANDFILE configuration file option or 474468651Skris environment variable, or the default random state file. 474568651Skris [Richard Levitte] 474668651Skris 474768651Skris *) mkstack.pl now sorts each macro group into lexical order. 474868651Skris Previously the output order depended on the order the files 474968651Skris appeared in the directory, resulting in needless rewriting 475068651Skris of safestack.h . 475168651Skris [Steve Henson] 475268651Skris 475368651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 475468651Skris work arounds for the VC++ problem that it treats func() as 475568651Skris func(void). Also stripped out the parts of mkdef.pl that 475668651Skris added extra typesafe functions: these no longer exist. 475768651Skris [Steve Henson] 475868651Skris 475968651Skris *) Reorganisation of the stack code. The macros are now all 476068651Skris collected in safestack.h . Each macro is defined in terms of 476168651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 476268651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 476368651Skris this has the advantage of retaining type safety without the 476468651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 476568651Skris then the non typesafe macros are used instead. Also modified the 476668651Skris mkstack.pl script to handle the new form. Needs testing to see 476768651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 476868651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 476968651Skris and PKCS12_STACK_OF. 477068651Skris [Steve Henson] 477168651Skris 477268651Skris *) When some versions of IIS use the 'NET' form of private key the 477368651Skris key derivation algorithm is different. Normally MD5(password) is 477468651Skris used as a 128 bit RC4 key. In the modified case 477568651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 477668651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 477768651Skris as the old Netscape_RSA functions except they have an additional 477868651Skris 'sgckey' parameter which uses the modified algorithm. Also added 477968651Skris an -sgckey command line option to the rsa utility. Thanks to 478068651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 478168651Skris algorithm to openssl-dev. 478268651Skris [Steve Henson] 478368651Skris 478468651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 478568651Skris invalid expansion on some systems (SCO 5.0.5 for example). 478668651Skris Corrected to 'c.kname'. 478768651Skris [Phillip Porch <root@theporch.com>] 478868651Skris 478968651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 479068651Skris a STACK of email addresses from a certificate or request, these look 479168651Skris in the subject name and the subject alternative name extensions and 479268651Skris omit any duplicate addresses. 479368651Skris [Steve Henson] 479468651Skris 479568651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 479668651Skris This makes DSA verification about 2 % faster. 479768651Skris [Bodo Moeller] 479868651Skris 479968651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 480068651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 480168651Skris plus overhead for 1024 bit moduli). 480268651Skris This makes exponentiations about 0.5 % faster for 1024 bit 480368651Skris exponents (as measured by "openssl speed rsa2048"). 480468651Skris [Bodo Moeller] 480568651Skris 480668651Skris *) Rename memory handling macros to avoid conflicts with other 480768651Skris software: 480868651Skris Malloc => OPENSSL_malloc 480968651Skris Malloc_locked => OPENSSL_malloc_locked 481068651Skris Realloc => OPENSSL_realloc 481168651Skris Free => OPENSSL_free 481268651Skris [Richard Levitte] 481368651Skris 481468651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 481568651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 481668651Skris [Bodo Moeller] 481768651Skris 481868651Skris *) CygWin32 support. 481968651Skris [John Jarvie <jjarvie@newsguy.com>] 482068651Skris 482168651Skris *) The type-safe stack code has been rejigged. It is now only compiled 482268651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 482368651Skris by default all type-specific stack functions are "#define"d back to 482468651Skris standard stack functions. This results in more streamlined output 482568651Skris but retains the type-safety checking possibilities of the original 482668651Skris approach. 482768651Skris [Geoff Thorpe] 482868651Skris 482968651Skris *) The STACK code has been cleaned up, and certain type declarations 483068651Skris that didn't make a lot of sense have been brought in line. This has 483168651Skris also involved a cleanup of sorts in safestack.h to more correctly 483268651Skris map type-safe stack functions onto their plain stack counterparts. 483368651Skris This work has also resulted in a variety of "const"ifications of 483468651Skris lots of the code, especially "_cmp" operations which should normally 483568651Skris be prototyped with "const" parameters anyway. 483668651Skris [Geoff Thorpe] 483768651Skris 483868651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 483968651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 484068651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 484168651Skris where all of 'md' is used each time the PRNG is used, but 'state' 484268651Skris is used only indexed by a cyclic counter. As entropy may not be 484368651Skris well distributed from the beginning, 'md' is important as a 484468651Skris chaining variable. However, the output function chains only half 484568651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 484668651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 484768651Skris in all of 'state' being rewritten, with the new values depending 484868651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 484968651Skris [Bodo Moeller] 485068651Skris 485168651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 485268651Skris the handshake is continued after ssl_verify_cert_chain(); 485368651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 485468651Skris can lead to 'unexplainable' connection aborts later. 485568651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 485668651Skris 485768651Skris *) Major EVP API cipher revision. 485868651Skris Add hooks for extra EVP features. This allows various cipher 485968651Skris parameters to be set in the EVP interface. Support added for variable 486068651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 486168651Skris setting of RC2 and RC5 parameters. 486268651Skris 486368651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 486468651Skris ciphers. 486568651Skris 486668651Skris Remove lots of duplicated code from the EVP library. For example *every* 486768651Skris cipher init() function handles the 'iv' in the same way according to the 486868651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 486968651Skris for CFB and OFB modes they zero ctx->num. 487068651Skris 487168651Skris New functionality allows removal of S/MIME code RC2 hack. 487268651Skris 487368651Skris Most of the routines have the same form and so can be declared in terms 487468651Skris of macros. 487568651Skris 487668651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 487768651Skris all individual ciphers. If the cipher wants to handle IVs or keys 487868651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 487968651Skris flags. 488068651Skris 488168651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 488268651Skris value: although software versions of the algorithms cannot fail 488368651Skris any installed hardware versions can. 488468651Skris [Steve Henson] 488568651Skris 488668651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 488768651Skris this option is set, tolerate broken clients that send the negotiated 488868651Skris protocol version number instead of the requested protocol version 488968651Skris number. 489068651Skris [Bodo Moeller] 489168651Skris 489268651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 489368651Skris i.e. non-zero for export ciphersuites, zero otherwise. 489468651Skris Previous versions had this flag inverted, inconsistent with 489568651Skris rsa_tmp_cb (..._TMP_RSA_CB). 489668651Skris [Bodo Moeller; problem reported by Amit Chopra] 489768651Skris 489868651Skris *) Add missing DSA library text string. Work around for some IIS 489968651Skris key files with invalid SEQUENCE encoding. 490068651Skris [Steve Henson] 490168651Skris 490268651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 490368651Skris and so on that are implemented in OpenSSL. 490468651Skris [Richard Levitte] 490568651Skris 490668651Skris *) Enhance c_rehash script. Old version would mishandle certificates 490768651Skris with the same subject name hash and wouldn't handle CRLs at all. 490868651Skris Added -fingerprint option to crl utility, to support new c_rehash 490968651Skris features. 491068651Skris [Steve Henson] 491168651Skris 491268651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 491368651Skris [Ulf M�ller] 491468651Skris 491568651Skris *) Fix for SSL server purpose checking. Server checking was 491668651Skris rejecting certificates which had extended key usage present 491768651Skris but no ssl client purpose. 491868651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 491968651Skris 492068651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 492168651Skris is a little unclear about how a blank password is handled. 492268651Skris Since the password in encoded as a BMPString with terminating 492368651Skris double NULL a zero length password would end up as just the 492468651Skris double NULL. However no password at all is different and is 492568651Skris handled differently in the PKCS#12 key generation code. NS 492668651Skris treats a blank password as zero length. MSIE treats it as no 492768651Skris password on export: but it will try both on import. We now do 492868651Skris the same: PKCS12_parse() tries zero length and no password if 492968651Skris the password is set to "" or NULL (NULL is now a valid password: 493068651Skris it wasn't before) as does the pkcs12 application. 493168651Skris [Steve Henson] 493268651Skris 493368651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 493468651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 493568651Skris be obtained from the error queue. 493668651Skris [Bodo Moeller] 493768651Skris 493868651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 493968651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 494068651Skris accordingly to avoid race conditions (this is necessary because 494168651Skris thread_hash is no longer constant once set). 494268651Skris [Bodo Moeller] 494368651Skris 494468651Skris *) Bugfix for linux-elf makefile.one. 494568651Skris [Ulf M�ller] 494668651Skris 494768651Skris *) RSA_get_default_method() will now cause a default 494868651Skris RSA_METHOD to be chosen if one doesn't exist already. 494968651Skris Previously this was only set during a call to RSA_new() 495068651Skris or RSA_new_method(NULL) meaning it was possible for 495168651Skris RSA_get_default_method() to return NULL. 495268651Skris [Geoff Thorpe] 495368651Skris 495468651Skris *) Added native name translation to the existing DSO code 495568651Skris that will convert (if the flag to do so is set) filenames 495668651Skris that are sufficiently small and have no path information 495768651Skris into a canonical native form. Eg. "blah" converted to 495868651Skris "libblah.so" or "blah.dll" etc. 495968651Skris [Geoff Thorpe] 496068651Skris 496168651Skris *) New function ERR_error_string_n(e, buf, len) which is like 496268651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 496368651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 496468651Skris may not be NULL. 496568651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 496668651Skris 496768651Skris *) CONF library reworked to become more general. A new CONF 496868651Skris configuration file reader "class" is implemented as well as a 496968651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 497068651Skris old CONF_* functions are still there, but are reimplemented to 497168651Skris work in terms of the new functions. Also, a set of functions 497268651Skris to handle the internal storage of the configuration data is 497368651Skris provided to make it easier to write new configuration file 497468651Skris reader "classes" (I can definitely see something reading a 497568651Skris configuration file in XML format, for example), called _CONF_*, 497668651Skris or "the configuration storage API"... 497768651Skris 497868651Skris The new configuration file reading functions are: 497968651Skris 498068651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 498168651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 498268651Skris 498368651Skris NCONF_default, NCONF_WIN32 498468651Skris 498568651Skris NCONF_dump_fp, NCONF_dump_bio 498668651Skris 498768651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 498868651Skris NCONF_new creates a new CONF object. This works in the same way 498968651Skris as other interfaces in OpenSSL, like the BIO interface. 499068651Skris NCONF_dump_* dump the internal storage of the configuration file, 499168651Skris which is useful for debugging. All other functions take the same 499268651Skris arguments as the old CONF_* functions wth the exception of the 499368651Skris first that must be a `CONF *' instead of a `LHASH *'. 499468651Skris 499568651Skris To make it easer to use the new classes with the old CONF_* functions, 499668651Skris the function CONF_set_default_method is provided. 499768651Skris [Richard Levitte] 499868651Skris 499968651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 500068651Skris mentioned in the documentation but had not been implemented. 500168651Skris (This option is not yet really useful because even the additional 500268651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 500368651Skris [Bodo Moeller] 500468651Skris 500568651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 500668651Skris OpenSSL-based applications) load shared libraries and bind to 500768651Skris them in a portable way. 500868651Skris [Geoff Thorpe, with contributions from Richard Levitte] 500968651Skris 501059191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 501159191Skris 501259191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 501359191Skris 501459191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 501559191Skris (the default implementation of RAND_status). 501659191Skris 501759191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 501859191Skris to '-clrext' (= clear extensions), as intended and documented. 501959191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 502059191Skris <attili@amaxo.com>] 502159191Skris 502259191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 502359191Skris was larger than the MD block size. 502459191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 502559191Skris 502659191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 502759191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 502859191Skris using the passed key: if the passed key was a private key the result 502959191Skris of X509_print(), for example, would be to print out all the private key 503059191Skris components. 503159191Skris [Steve Henson] 503259191Skris 503359191Skris *) des_quad_cksum() byte order bug fix. 503459191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 503559191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 503659191Skris 503759191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 503859191Skris discouraged. 503959191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 504059191Skris 504159191Skris *) For easily testing in shell scripts whether some command 504259191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 504359191Skris returns with exit code 0 iff no command of the given name is available. 504459191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 504559191Skris the output goes to stdout and nothing is printed to stderr. 504659191Skris Additional arguments are always ignored. 504759191Skris 504859191Skris Since for each cipher there is a command of the same name, 504959191Skris the 'no-cipher' compilation switches can be tested this way. 505059191Skris 505159191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 505259191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 505359191Skris [Bodo Moeller] 505459191Skris 505559191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 505659191Skris [Bodo Moeller] 505759191Skris 505859191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 505959191Skris is set; it will be thrown away anyway because each handshake creates 506059191Skris its own key. 506159191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 506259191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 506359191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 506459191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 506559191Skris [Bodo Moeller] 506659191Skris 506759191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 506859191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 506959191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 507059191Skris does not suppress any output. 507159191Skris [Richard Levitte] 507259191Skris 507359191Skris *) Add compatibility options to the purpose and trust code. The 507459191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 507559191Skris accepts a certificate or CA, this was the previous behaviour, 507659191Skris with all the associated security issues. 507759191Skris 507859191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 507959191Skris automatically trust self signed roots in certificate store. A 508059191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 508159191Skris a purpose has no associated trust setting and it should instead 508259191Skris use the value in the default purpose. 508359191Skris [Steve Henson] 508459191Skris 508559191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 508659191Skris and fix a memory leak. 508759191Skris [Steve Henson] 508859191Skris 508959191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 509059191Skris reason strings from the previous version of the .c file, as 509159191Skris the default to have only downcase letters (and digits) in 509259191Skris automatically generated reasons codes is not always appropriate. 509359191Skris [Bodo Moeller] 509459191Skris 509559191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 509659191Skris using strerror. Previously, ERR_reason_error_string() returned 509759191Skris library names as reason strings for SYSerr; but SYSerr is a special 509859191Skris case where small numbers are errno values, not library numbers. 509959191Skris [Bodo Moeller] 510059191Skris 510159191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 510259191Skris converts DSA parameters into DH parameters. (When creating parameters, 510359191Skris DSA_generate_parameters is used.) 510459191Skris [Bodo Moeller] 510559191Skris 510659191Skris *) Include 'length' (recommended exponent length) in C code generated 510759191Skris by 'openssl dhparam -C'. 510859191Skris [Bodo Moeller] 510959191Skris 511059191Skris *) The second argument to set_label in perlasm was already being used 511159191Skris so couldn't be used as a "file scope" flag. Moved to third argument 511259191Skris which was free. 511359191Skris [Steve Henson] 511459191Skris 511559191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 511659191Skris instead of RAND_bytes for encryption IVs and salts. 511759191Skris [Bodo Moeller] 511859191Skris 511959191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 512059191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 512159191Skris RAND_set_rand_method would be impossible. 512259191Skris [Bodo Moeller] 512359191Skris 512459191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 512559191Skris number generation fails. 512659191Skris [Bodo Moeller] 512759191Skris 512859191Skris *) New 'rand' application for creating pseudo-random output. 512959191Skris [Bodo Moeller] 513059191Skris 513159191Skris *) Added configuration support for Linux/IA64 513259191Skris [Rolf Haberrecker <rolf@suse.de>] 513359191Skris 513459191Skris *) Assembler module support for Mingw32. 513559191Skris [Ulf M�ller] 513659191Skris 513759191Skris *) Shared library support for HPUX (in shlib/). 513859191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 513959191Skris 514059191Skris *) Shared library support for Solaris gcc. 514159191Skris [Lutz Behnke <behnke@trustcenter.de>] 514259191Skris 514359191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 514459191Skris 514559191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 514659191Skris were added manually and by SMIME_crlf_copy(). 514759191Skris [Steve Henson] 514859191Skris 514959191Skris *) In bntest.c don't call BN_rand with zero bits argument. 515059191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 515159191Skris 515259191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 515359191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 515459191Skris [Ulf M�ller] 515559191Skris 515659191Skris *) Add an optional second argument to the set_label() in the perl 515759191Skris assembly language builder. If this argument exists and is set 515859191Skris to 1 it signals that the assembler should use a symbol whose 515959191Skris scope is the entire file, not just the current function. This 516059191Skris is needed with MASM which uses the format label:: for this scope. 516159191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 516259191Skris 516359191Skris *) Change the ASN1 types so they are typedefs by default. Before 516459191Skris almost all types were #define'd to ASN1_STRING which was causing 516559191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 516659191Skris for example. 516759191Skris [Steve Henson] 516859191Skris 516959191Skris *) Change names of new functions to the new get1/get0 naming 517059191Skris convention: After 'get1', the caller owns a reference count 517159191Skris and has to call ..._free; 'get0' returns a pointer to some 517259191Skris data structure without incrementing reference counters. 517359191Skris (Some of the existing 'get' functions increment a reference 517459191Skris counter, some don't.) 517559191Skris Similarly, 'set1' and 'add1' functions increase reference 517659191Skris counters or duplicate objects. 517759191Skris [Steve Henson] 517859191Skris 517959191Skris *) Allow for the possibility of temp RSA key generation failure: 518059191Skris the code used to assume it always worked and crashed on failure. 518159191Skris [Steve Henson] 518259191Skris 518359191Skris *) Fix potential buffer overrun problem in BIO_printf(). 518459191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 518559191Skris pointed out by David Sacerdote <das33@cornell.edu>] 518659191Skris 518759191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 518859191Skris RAND_egd() and RAND_status(). In the command line application, 518959191Skris the EGD socket can be specified like a seed file using RANDFILE 519059191Skris or -rand. 519159191Skris [Ulf M�ller] 519259191Skris 519359191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 519459191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 519559191Skris [Steve Henson] 519659191Skris 519759191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 519859191Skris list to exclude them. This means that no special compilation option 519959191Skris is needed to use anonymous DH: it just needs to be included in the 520059191Skris cipher list. 520159191Skris [Steve Henson] 520259191Skris 520359191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 520459191Skris EVP_MD_type. The old functionality is available in a new macro called 520559191Skris EVP_MD_md(). Change code that uses it and update docs. 520659191Skris [Steve Henson] 520759191Skris 520859191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 520959191Skris where the 'void *' argument is replaced by a function pointer argument. 521059191Skris Previously 'void *' was abused to point to functions, which works on 521159191Skris many platforms, but is not correct. As these functions are usually 521259191Skris called by macros defined in OpenSSL header files, most source code 521359191Skris should work without changes. 521459191Skris [Richard Levitte] 521559191Skris 521659191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 521759191Skris sections with information on -D... compiler switches used for 521859191Skris compiling the library so that applications can see them. To enable 521959191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 522059191Skris must be defined. E.g., 522159191Skris #define OPENSSL_ALGORITHM_DEFINES 522259191Skris #include <openssl/opensslconf.h> 522359191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 522459191Skris [Richard Levitte, Ulf and Bodo M�ller] 522559191Skris 522659191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 522759191Skris record layer. 522859191Skris [Bodo Moeller] 522959191Skris 523059191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 523159191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 523259191Skris the required ASN1 format: arbitrary types determined by an OID. 523359191Skris [Steve Henson] 523459191Skris 523559191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 523659191Skris argument to 'req'. This is not because the function is newer or 523759191Skris better than others it just uses the work 'NEW' in the certificate 523859191Skris request header lines. Some software needs this. 523959191Skris [Steve Henson] 524059191Skris 524159191Skris *) Reorganise password command line arguments: now passwords can be 524259191Skris obtained from various sources. Delete the PEM_cb function and make 524359191Skris it the default behaviour: i.e. if the callback is NULL and the 524459191Skris usrdata argument is not NULL interpret it as a null terminated pass 524559191Skris phrase. If usrdata and the callback are NULL then the pass phrase 524659191Skris is prompted for as usual. 524759191Skris [Steve Henson] 524859191Skris 524959191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 525059191Skris the support is automatically enabled. The resulting binaries will 525159191Skris autodetect the card and use it if present. 525259191Skris [Ben Laurie and Compaq Inc.] 525359191Skris 525459191Skris *) Work around for Netscape hang bug. This sends certificate request 525559191Skris and server done in one record. Since this is perfectly legal in the 525659191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 525759191Skris the bugs/SSLv3 entry for more info. 525859191Skris [Steve Henson] 525959191Skris 526059191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 526159191Skris [Andy Polyakov] 526259191Skris 526359191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 526459191Skris of seed file. 526559191Skris [Steve Henson] 526659191Skris 526759191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 526859191Skris [Bodo Moeller] 526959191Skris 527059191Skris *) Add command line password options to the remaining applications. 527159191Skris [Steve Henson] 527259191Skris 527359191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 527459191Skris bits. 527559191Skris [Ulf M�ller] 527659191Skris 527759191Skris *) More tests in bntest.c, and changed test_bn output. 527859191Skris [Ulf M�ller] 527959191Skris 528059191Skris *) ./config recognizes MacOS X now. 528159191Skris [Andy Polyakov] 528259191Skris 528359191Skris *) Bug fix for BN_div() when the first words of num and divsor are 528459191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 528559191Skris [Ulf M�ller] 528659191Skris 528759191Skris *) Add support for various broken PKCS#8 formats, and command line 528859191Skris options to produce them. 528959191Skris [Steve Henson] 529059191Skris 529159191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 529259191Skris get temporary BIGNUMs from a BN_CTX. 529359191Skris [Ulf M�ller] 529459191Skris 529559191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 529659191Skris for p == 0. 529759191Skris [Ulf M�ller] 529859191Skris 529959191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 530059191Skris include a #define from the old name to the new. The original intent 530159191Skris was that statically linked binaries could for example just call 530259191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 530359191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 530459191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 530559191Skris one would link with the other. They are now in separate source files. 530659191Skris [Steve Henson] 530759191Skris 530859191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 530959191Skris [Steve Henson] 531059191Skris 531159191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 531259191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 531359191Skris loop, our standard modexp algorithms are faster). 531459191Skris [Bodo Moeller] 531559191Skris 531659191Skris *) Support for the EBCDIC character set completed. 531759191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 531859191Skris 531959191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 532059191Skris use void * instead of char * in lhash. 532159191Skris [Ulf M�ller] 532259191Skris 532359191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 532459191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 532559191Skris this the server could overwrite ephemeral keys that the client 532659191Skris has already seen). 532759191Skris [Bodo Moeller] 532859191Skris 532959191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 533059191Skris using 50 iterations of the Rabin-Miller test. 533159191Skris 533259191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 533359191Skris iterations of the Rabin-Miller test as required by the appendix 533459191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 533559191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 533659191Skris generation becomes much faster. 533759191Skris 533859191Skris This implies a change for the callback functions in DSA_is_prime 533959191Skris and DSA_generate_parameters: The callback function is called once 534059191Skris for each positive witness in the Rabin-Miller test, not just 534159191Skris occasionally in the inner loop; and the parameters to the 534259191Skris callback function now provide an iteration count for the outer 534359191Skris loop rather than for the current invocation of the inner loop. 534459191Skris DSA_generate_parameters additionally can call the callback 534559191Skris function with an 'iteration count' of -1, meaning that a 534659191Skris candidate has passed the trial division test (when q is generated 534759191Skris from an application-provided seed, trial division is skipped). 534859191Skris [Bodo Moeller] 534959191Skris 535059191Skris *) New function BN_is_prime_fasttest that optionally does trial 535159191Skris division before starting the Rabin-Miller test and has 535259191Skris an additional BN_CTX * argument (whereas BN_is_prime always 535359191Skris has to allocate at least one BN_CTX). 535459191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 535559191Skris trial division stage. 535659191Skris [Bodo Moeller] 535759191Skris 535859191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 535959191Skris as ASN1_TIME. 536059191Skris [Steve Henson] 536159191Skris 536259191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 536359191Skris [Steve Henson] 536459191Skris 536559191Skris *) New function BN_pseudo_rand(). 536659191Skris [Ulf M�ller] 536759191Skris 536859191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 536959191Skris bignum version of BN_from_montgomery() with the working code from 537059191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 537159191Skris the comments. 537259191Skris [Ulf M�ller] 537359191Skris 537459191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 537559191Skris made it impossible to use the same SSL_SESSION data structure in 537659191Skris SSL2 clients in multiple threads. 537759191Skris [Bodo Moeller] 537859191Skris 537959191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 538059191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 538159191Skris to seed the PRNG (previously an explicit byte count was required). 538259191Skris [Ulf M�ller, Bodo M�ller] 538359191Skris 538459191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 538559191Skris used (char *) instead of (void *) and had casts all over the place. 538659191Skris [Steve Henson] 538759191Skris 538859191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 538959191Skris [Ulf M�ller] 539059191Skris 539159191Skris *) Retain source code compatibility for BN_prime_checks macro: 539259191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 539359191Skris BN_prime_checks_for_size to determine the appropriate number of 539459191Skris Rabin-Miller iterations. 539559191Skris [Ulf M�ller] 539659191Skris 539759191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 539859191Skris DH_CHECK_P_NOT_SAFE_PRIME. 539959191Skris (Check if this is true? OpenPGP calls them "strong".) 540059191Skris [Ulf M�ller] 540159191Skris 540259191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 540359191Skris "dhparam". The old programs are retained for now but will handle DH keys 540459191Skris (instead of parameters) in future. 540559191Skris [Steve Henson] 540659191Skris 540759191Skris *) Make the ciphers, s_server and s_client programs check the return values 540859191Skris when a new cipher list is set. 540959191Skris [Steve Henson] 541059191Skris 541159191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 541259191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 541359191Skris wrong. 541459191Skris 541559191Skris The syntax for the cipher sorting has been extended to support sorting by 541659191Skris cipher-strength (using the strength_bits hard coded in the tables). 541759191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 541859191Skris 541959191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 542059191Skris string with an "undefined" symbol (neither command nor alphanumeric 542159191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 542259191Skris an error is flagged. 542359191Skris 542459191Skris Due to the strength-sorting extension, the code of the 542559191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 542659191Skris the readability was also increased :-) 542759191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 542859191Skris 542959191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 543059191Skris for the first serial number and places 2 in the serial number file. This 543159191Skris avoids problems when the root CA is created with serial number zero and 543259191Skris the first user certificate has the same issuer name and serial number 543359191Skris as the root CA. 543459191Skris [Steve Henson] 543559191Skris 543659191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 543759191Skris the new code. Add documentation for this stuff. 543859191Skris [Steve Henson] 543959191Skris 544059191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 544159191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 544259191Skris structures and behave in an analagous way to the X509v3 functions: 544359191Skris they shouldn't be called directly but wrapper functions should be used 544459191Skris instead. 544559191Skris 544659191Skris So we also now have some wrapper functions that call the X509at functions 544759191Skris when passed certificate requests. (TO DO: similar things can be done with 544859191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 544959191Skris things. Some of these need some d2i or i2d and print functionality 545059191Skris because they handle more complex structures.) 545159191Skris [Steve Henson] 545259191Skris 545359191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 545459191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 545559191Skris NO_RSA in ssl/s2*.c. 545659191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 545759191Skris 545859191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 545959191Skris has a return value which indicates the quality of the random data 546059191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 546159191Skris error queue. New function RAND_pseudo_bytes() generates output that is 546259191Skris guaranteed to be unique but not unpredictable. RAND_add is like 546359191Skris RAND_seed, but takes an extra argument for an entropy estimate 546459191Skris (RAND_seed always assumes full entropy). 546559191Skris [Ulf M�ller] 546659191Skris 546759191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 546859191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 546959191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 547059191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 547159191Skris false-positive rate of at most 2^-80 for random input. 547259191Skris [Bodo Moeller] 547359191Skris 547459191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 547559191Skris [Bodo Moeller] 547659191Skris 547759191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 547859191Skris in the 0.9.5 release), this returns the chain 547959191Skris from an X509_CTX structure with a dup of the stack and all 548059191Skris the X509 reference counts upped: so the stack will exist 548159191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 548259191Skris to use this. 548359191Skris 548459191Skris Also make SSL_SESSION_print() print out the verify return 548559191Skris code. 548659191Skris [Steve Henson] 548759191Skris 548859191Skris *) Add manpage for the pkcs12 command. Also change the default 548959191Skris behaviour so MAC iteration counts are used unless the new 549059191Skris -nomaciter option is used. This improves file security and 549159191Skris only older versions of MSIE (4.0 for example) need it. 549259191Skris [Steve Henson] 549359191Skris 549459191Skris *) Honor the no-xxx Configure options when creating .DEF files. 549559191Skris [Ulf M�ller] 549659191Skris 549759191Skris *) Add PKCS#10 attributes to field table: challengePassword, 549859191Skris unstructuredName and unstructuredAddress. These are taken from 549959191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 550059191Skris international characters are used. 550159191Skris 550259191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 550359191Skris based on strings. Remove the 'loc' parameter when adding 550459191Skris attributes because these will be a SET OF encoding which is sorted 550559191Skris in ASN1 order. 550659191Skris [Steve Henson] 550759191Skris 550859191Skris *) Initial changes to the 'req' utility to allow request generation 550959191Skris automation. This will allow an application to just generate a template 551059191Skris file containing all the field values and have req construct the 551159191Skris request. 551259191Skris 551359191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 551459191Skris used all over the place including certificate requests and PKCS#7 551559191Skris structures. They are currently handled manually where necessary with 551659191Skris some primitive wrappers for PKCS#7. The new functions behave in a 551759191Skris manner analogous to the X509 extension functions: they allow 551859191Skris attributes to be looked up by NID and added. 551959191Skris 552059191Skris Later something similar to the X509V3 code would be desirable to 552159191Skris automatically handle the encoding, decoding and printing of the 552259191Skris more complex types. The string types like challengePassword can 552359191Skris be handled by the string table functions. 552459191Skris 552559191Skris Also modified the multi byte string table handling. Now there is 552659191Skris a 'global mask' which masks out certain types. The table itself 552759191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 552859191Skris is useful when for example there is only one permissible type 552959191Skris (as in countryName) and using the mask might result in no valid 553059191Skris types at all. 553159191Skris [Steve Henson] 553259191Skris 553359191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 553459191Skris SSL_get_peer_finished to allow applications to obtain the latest 553559191Skris Finished messages sent to the peer or expected from the peer, 553659191Skris respectively. (SSL_get_peer_finished is usually the Finished message 553759191Skris actually received from the peer, otherwise the protocol will be aborted.) 553859191Skris 553959191Skris As the Finished message are message digests of the complete handshake 554059191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 554159191Skris be used for external authentication procedures when the authentication 554259191Skris provided by SSL/TLS is not desired or is not enough. 554359191Skris [Bodo Moeller] 554459191Skris 554559191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 554659191Skris the host supports BWX extension and if Compaq C is present on the 554759191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 554859191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 554959191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 555059191Skris SHA1. 555159191Skris [Andy Polyakov] 555259191Skris 555359191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 555459191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 555559191Skris weak crypto and after checking the certificate is SGC a second one 555659191Skris with strong crypto. MS SGC stops the first handshake after receiving 555759191Skris the server certificate message and sends a second client hello. Since 555859191Skris a server will typically do all the time consuming operations before 555959191Skris expecting any further messages from the client (server key exchange 556059191Skris is the most expensive) there is little difference between the two. 556159191Skris 556259191Skris To get OpenSSL to support MS SGC we have to permit a second client 556359191Skris hello message after we have sent server done. In addition we have to 556459191Skris reset the MAC if we do get this second client hello. 556559191Skris [Steve Henson] 556659191Skris 556759191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 556859191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 556959191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 557059191Skris format DER encoded private key. Newer code should use PKCS#8 format which 557159191Skris has the key type encoded in the ASN1 structure. Added DER private key 557259191Skris support to pkcs8 application. 557359191Skris [Steve Henson] 557459191Skris 557559191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 557659191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 557759191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 557859191Skris is set, we interpret this as a request to violate the specification 557959191Skris (the worst that can happen is a handshake failure, and 'correct' 558059191Skris behaviour would result in a handshake failure anyway). 558159191Skris [Bodo Moeller] 558259191Skris 558359191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 558459191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 558559191Skris concurrently obtain them from an external cache). 558659191Skris The internal cache can handle only one SSL_SESSION with a given ID, 558759191Skris so if there's a conflict, we now throw out the old one to achieve 558859191Skris consistency. 558959191Skris [Bodo Moeller] 559059191Skris 559159191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 559259191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 559359191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 559459191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 559559191Skris example. 559659191Skris [Steve Henson] 559759191Skris 559859191Skris *) Simplify the trust setting structure and code. Now we just have 559959191Skris two sequences of OIDs for trusted and rejected settings. These will 560059191Skris typically have values the same as the extended key usage extension 560159191Skris and any application specific purposes. 560259191Skris 560359191Skris The trust checking code now has a default behaviour: it will just 560459191Skris check for an object with the same NID as the passed id. Functions can 560559191Skris be provided to override either the default behaviour or the behaviour 560659191Skris for a given id. SSL client, server and email already have functions 560759191Skris in place for compatibility: they check the NID and also return "trusted" 560859191Skris if the certificate is self signed. 560959191Skris [Steve Henson] 561059191Skris 561159191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 561259191Skris traditional format into an EVP_PKEY structure. 561359191Skris [Steve Henson] 561459191Skris 561559191Skris *) Add a password callback function PEM_cb() which either prompts for 561659191Skris a password if usr_data is NULL or otherwise assumes it is a null 561759191Skris terminated password. Allow passwords to be passed on command line 561859191Skris environment or config files in a few more utilities. 561959191Skris [Steve Henson] 562059191Skris 562159191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 562259191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 562359191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 562459191Skris Update documentation. 562559191Skris [Steve Henson] 562659191Skris 562759191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 562859191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 562959191Skris and produce an error if it couldn't. For compatibility we also have 563059191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 563159191Skris don't allocate anything because they don't need to. 563259191Skris [Steve Henson] 563359191Skris 563459191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 563559191Skris for details. 563659191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 563759191Skris 563859191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 563959191Skris possibly others as well. The purpose is to make an interface that 564059191Skris provide hooks so anyone can build a separate set of allocation and 564159191Skris deallocation routines to be used by OpenSSL, for example memory 564259191Skris pool implementations, or something else, which was previously hard 564359191Skris since Malloc(), Realloc() and Free() were defined as macros having 564459191Skris the values malloc, realloc and free, respectively (except for Win32 564559191Skris compilations). The same is provided for memory debugging code. 564659191Skris OpenSSL already comes with functionality to find memory leaks, but 564759191Skris this gives people a chance to debug other memory problems. 564859191Skris 564959191Skris With these changes, a new set of functions and macros have appeared: 565059191Skris 565168651Skris CRYPTO_set_mem_debug_functions() [F] 565268651Skris CRYPTO_get_mem_debug_functions() [F] 565368651Skris CRYPTO_dbg_set_options() [F] 565468651Skris CRYPTO_dbg_get_options() [F] 565568651Skris CRYPTO_malloc_debug_init() [M] 565659191Skris 565759191Skris The memory debug functions are NULL by default, unless the library 565859191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 565959191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 566059191Skris gives the standard debugging functions that come with OpenSSL) or 566159191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 566259191Skris provided by the library user) must be used. When the standard 566359191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 566459191Skris request additional information: 566559191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 566659191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 566759191Skris 566859191Skris Also, things like CRYPTO_set_mem_functions will always give the 566959191Skris expected result (the new set of functions is used for allocation 567059191Skris and deallocation) at all times, regardless of platform and compiler 567159191Skris options. 567259191Skris 567359191Skris To finish it up, some functions that were never use in any other 567459191Skris way than through macros have a new API and new semantic: 567559191Skris 567659191Skris CRYPTO_dbg_malloc() 567759191Skris CRYPTO_dbg_realloc() 567859191Skris CRYPTO_dbg_free() 567959191Skris 568059191Skris All macros of value have retained their old syntax. 568159191Skris [Richard Levitte and Bodo Moeller] 568259191Skris 568359191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 568459191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 568559191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 568659191Skris algorithm. 568759191Skris [Steve Henson] 568859191Skris 568959191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 569059191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 569159191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 569259191Skris 569359191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 569459191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 569559191Skris functionality to handle multipart/signed properly) and a utility 569659191Skris called 'smime' to call all this stuff. This is based on code I 569759191Skris originally wrote for Celo who have kindly allowed it to be 569859191Skris included in OpenSSL. 569959191Skris [Steve Henson] 570059191Skris 570159191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 570259191Skris des_set_key (aka des_key_sched). Global variable des_check_key 570359191Skris decides which of these is called by des_set_key; this way 570459191Skris des_check_key behaves as it always did, but applications and 570559191Skris the library itself, which was buggy for des_check_key == 1, 570659191Skris have a cleaner way to pick the version they need. 570759191Skris [Bodo Moeller] 570859191Skris 570959191Skris *) New function PKCS12_newpass() which changes the password of a 571059191Skris PKCS12 structure. 571159191Skris [Steve Henson] 571259191Skris 571359191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 571459191Skris dynamic mix. In both cases the ids can be used as an index into the 571559191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 571659191Skris functions so they accept a list of the field values and the 571759191Skris application doesn't need to directly manipulate the X509_TRUST 571859191Skris structure. 571959191Skris [Steve Henson] 572059191Skris 572159191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 572259191Skris need initialising. 572359191Skris [Steve Henson] 572459191Skris 572559191Skris *) Modify the way the V3 extension code looks up extensions. This now 572659191Skris works in a similar way to the object code: we have some "standard" 572759191Skris extensions in a static table which is searched with OBJ_bsearch() 572859191Skris and the application can add dynamic ones if needed. The file 572959191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 573059191Skris updated whenever a new extension is added to the core code and kept 573159191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 573259191Skris this. New extensions are not added too often so this file can readily 573359191Skris be maintained manually. 573459191Skris 573559191Skris There are two big advantages in doing things this way. The extensions 573659191Skris can be looked up immediately and no longer need to be "added" using 573759191Skris X509V3_add_standard_extensions(): this function now does nothing. 573859191Skris [Side note: I get *lots* of email saying the extension code doesn't 573959191Skris work because people forget to call this function] 574059191Skris Also no dynamic allocation is done unless new extensions are added: 574159191Skris so if we don't add custom extensions there is no need to call 574259191Skris X509V3_EXT_cleanup(). 574359191Skris [Steve Henson] 574459191Skris 574559191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 574659191Skris magic header, so unsalted files fail gracefully instead of just decrypting 574759191Skris to garbage. This is because not salting is a big security hole, so people 574859191Skris should be discouraged from doing it. 574959191Skris [Ben Laurie] 575059191Skris 575159191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 575259191Skris digest to be passed on the command line but it only used this 575359191Skris parameter when signing a certificate. Modified so all relevant 575459191Skris operations are affected by the digest parameter including the 575559191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 575659191Skris DSA key was used because it didn't fix the digest. 575759191Skris [Steve Henson] 575859191Skris 575959191Skris *) Initial certificate chain verify code. Currently tests the untrusted 576059191Skris certificates for consistency with the verify purpose (which is set 576159191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 576259191Skris 576359191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 576459191Skris this is because it will reject chains with invalid extensions whereas 576559191Skris every previous version of OpenSSL and SSLeay made no checks at all. 576659191Skris 576759191Skris Trust code: checks the root CA for the relevant trust settings. Trust 576859191Skris settings have an initial value consistent with the verify purpose: e.g. 576959191Skris if the verify purpose is for SSL client use it expects the CA to be 577059191Skris trusted for SSL client use. However the default value can be changed to 577159191Skris permit custom trust settings: one example of this would be to only trust 577259191Skris certificates from a specific "secure" set of CAs. 577359191Skris 577459191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 577559191Skris which should be used for version portability: especially since the 577659191Skris verify structure is likely to change more often now. 577759191Skris 577859191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 577959191Skris to set them. If not set then assume SSL clients will verify SSL servers 578059191Skris and vice versa. 578159191Skris 578259191Skris Two new options to the verify program: -untrusted allows a set of 578359191Skris untrusted certificates to be passed in and -purpose which sets the 578459191Skris intended purpose of the certificate. If a purpose is set then the 578559191Skris new chain verify code is used to check extension consistency. 578659191Skris [Steve Henson] 578759191Skris 578859191Skris *) Support for the authority information access extension. 578959191Skris [Steve Henson] 579059191Skris 579159191Skris *) Modify RSA and DSA PEM read routines to transparently handle 579259191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 579359191Skris public keys in a format compatible with certificate 579459191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 579559191Skris functions called *_PublicKey_* which used various odd formats so 579659191Skris these are retained for compatibility: however the DSA variants were 579759191Skris never in a public release so they have been deleted. Changed dsa/rsa 579859191Skris utilities to handle the new format: note no releases ever handled public 579959191Skris keys so we should be OK. 580059191Skris 580159191Skris The primary motivation for this change is to avoid the same fiasco 580259191Skris that dogs private keys: there are several incompatible private key 580359191Skris formats some of which are standard and some OpenSSL specific and 580459191Skris require various evil hacks to allow partial transparent handling and 580559191Skris even then it doesn't work with DER formats. Given the option anything 580659191Skris other than PKCS#8 should be dumped: but the other formats have to 580759191Skris stay in the name of compatibility. 580859191Skris 580959191Skris With public keys and the benefit of hindsight one standard format 581059191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 581159191Skris it clearly returns an error if you try to read the wrong kind of key. 581259191Skris 581359191Skris Added a -pubkey option to the 'x509' utility to output the public key. 581459191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 581559191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 581659191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 581759191Skris that do the same as the EVP_PKEY_assign_*() except they up the 581859191Skris reference count of the added key (they don't "swallow" the 581959191Skris supplied key). 582059191Skris [Steve Henson] 582159191Skris 582259191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 582359191Skris CRLs would fail if the file contained no certificates or no CRLs: 582459191Skris added a new function to read in both types and return the number 582559191Skris read: this means that if none are read it will be an error. The 582659191Skris DER versions of the certificate and CRL reader would always fail 582759191Skris because it isn't possible to mix certificates and CRLs in DER format 582859191Skris without choking one or the other routine. Changed this to just read 582959191Skris a certificate: this is the best we can do. Also modified the code 583059191Skris in apps/verify.c to take notice of return codes: it was previously 583159191Skris attempting to read in certificates from NULL pointers and ignoring 583259191Skris any errors: this is one reason why the cert and CRL reader seemed 583359191Skris to work. It doesn't check return codes from the default certificate 583459191Skris routines: these may well fail if the certificates aren't installed. 583559191Skris [Steve Henson] 583659191Skris 583759191Skris *) Code to support otherName option in GeneralName. 583859191Skris [Steve Henson] 583959191Skris 584059191Skris *) First update to verify code. Change the verify utility 584159191Skris so it warns if it is passed a self signed certificate: 584259191Skris for consistency with the normal behaviour. X509_verify 584359191Skris has been modified to it will now verify a self signed 584459191Skris certificate if *exactly* the same certificate appears 584559191Skris in the store: it was previously impossible to trust a 584659191Skris single self signed certificate. This means that: 584759191Skris openssl verify ss.pem 584859191Skris now gives a warning about a self signed certificate but 584959191Skris openssl verify -CAfile ss.pem ss.pem 585059191Skris is OK. 585159191Skris [Steve Henson] 585259191Skris 585359191Skris *) For servers, store verify_result in SSL_SESSION data structure 585459191Skris (and add it to external session representation). 585559191Skris This is needed when client certificate verifications fails, 585659191Skris but an application-provided verification callback (set by 585759191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 585859191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 585959191Skris but returns 1): When the session is reused, we have to set 586059191Skris ssl->verify_result to the appropriate error code to avoid 586159191Skris security holes. 586259191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 586359191Skris 586459191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 586559191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 586659191Skris didn't contain any existing data because it was being created. 586759191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 586859191Skris 586959191Skris *) Add a salt to the key derivation routines in enc.c. This 587059191Skris forms the first 8 bytes of the encrypted file. Also add a 587159191Skris -S option to allow a salt to be input on the command line. 587259191Skris [Steve Henson] 587359191Skris 587459191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 587559191Skris to compare two certificates. We do this by working out the SHA1 587659191Skris hash and comparing that. X509_cmp() will be needed by the trust 587759191Skris code. 587859191Skris [Steve Henson] 587959191Skris 588059191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 588159191Skris the reference count in the SSL_SESSION returned. 588259191Skris [Geoff Thorpe <geoff@eu.c2.net>] 588359191Skris 588459191Skris *) Fix for 'req': it was adding a null to request attributes. 588559191Skris Also change the X509_LOOKUP and X509_INFO code to handle 588659191Skris certificate auxiliary information. 588759191Skris [Steve Henson] 588859191Skris 588959191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 589059191Skris the 'enc' command. 589159191Skris [Steve Henson] 589259191Skris 589359191Skris *) Add the possibility to add extra information to the memory leak 589459191Skris detecting output, to form tracebacks, showing from where each 589559191Skris allocation was originated: CRYPTO_push_info("constant string") adds 589659191Skris the string plus current file name and line number to a per-thread 589759191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 589859191Skris is like calling CYRPTO_pop_info() until the stack is empty. 589959191Skris Also updated memory leak detection code to be multi-thread-safe. 590059191Skris [Richard Levitte] 590159191Skris 590259191Skris *) Add options -text and -noout to pkcs7 utility and delete the 590359191Skris encryption options which never did anything. Update docs. 590459191Skris [Steve Henson] 590559191Skris 590659191Skris *) Add options to some of the utilities to allow the pass phrase 590759191Skris to be included on either the command line (not recommended on 590859191Skris OSes like Unix) or read from the environment. Update the 590959191Skris manpages and fix a few bugs. 591059191Skris [Steve Henson] 591159191Skris 591259191Skris *) Add a few manpages for some of the openssl commands. 591359191Skris [Steve Henson] 591459191Skris 591559191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 591659191Skris leaking and not finding already revoked certificates. 591759191Skris [Steve Henson] 591859191Skris 591959191Skris *) Extensive changes to support certificate auxiliary information. 592059191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 592159191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 592259191Skris can still read in a certificate file in the usual way but it 592359191Skris will also read in any additional "auxiliary information". By 592459191Skris doing things this way a fair degree of compatibility can be 592559191Skris retained: existing certificates can have this information added 592659191Skris using the new 'x509' options. 592759191Skris 592859191Skris Current auxiliary information includes an "alias" and some trust 592959191Skris settings. The trust settings will ultimately be used in enhanced 593059191Skris certificate chain verification routines: currently a certificate 593159191Skris can only be trusted if it is self signed and then it is trusted 593259191Skris for all purposes. 593359191Skris [Steve Henson] 593459191Skris 593559191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 593659191Skris The problem was that one of the replacement routines had not been working 593759191Skris since SSLeay releases. For now the offending routine has been replaced 593859191Skris with non-optimised assembler. Even so, this now gives around 95% 593959191Skris performance improvement for 1024 bit RSA signs. 594059191Skris [Mark Cox] 594159191Skris 594259191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 594359191Skris handling. Most clients have the effective key size in bits equal to 594459191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 594559191Skris A few however don't do this and instead use the size of the decrypted key 594659191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 594759191Skris the effective key length. In this case the effective key length can still 594859191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 594959191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 595059191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 595159191Skris the key length and effective key length are equal. 595259191Skris [Steve Henson] 595359191Skris 595459191Skris *) Add a bunch of functions that should simplify the creation of 595559191Skris X509_NAME structures. Now you should be able to do: 595659191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 595759191Skris and have it automatically work out the correct field type and fill in 595859191Skris the structures. The more adventurous can try: 595959191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 596059191Skris and it will (hopefully) work out the correct multibyte encoding. 596159191Skris [Steve Henson] 596259191Skris 596359191Skris *) Change the 'req' utility to use the new field handling and multibyte 596459191Skris copy routines. Before the DN field creation was handled in an ad hoc 596559191Skris way in req, ca, and x509 which was rather broken and didn't support 596659191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 596759191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 596859191Skris using the dirstring_type option. See the new comment in the default 596959191Skris openssl.cnf for more info. 597059191Skris [Steve Henson] 597159191Skris 597259191Skris *) Make crypto/rand/md_rand.c more robust: 597359191Skris - Assure unique random numbers after fork(). 597459191Skris - Make sure that concurrent threads access the global counter and 597559191Skris md serializably so that we never lose entropy in them 597659191Skris or use exactly the same state in multiple threads. 597759191Skris Access to the large state is not always serializable because 597859191Skris the additional locking could be a performance killer, and 597959191Skris md should be large enough anyway. 598059191Skris [Bodo Moeller] 598159191Skris 598259191Skris *) New file apps/app_rand.c with commonly needed functionality 598359191Skris for handling the random seed file. 598459191Skris 598559191Skris Use the random seed file in some applications that previously did not: 598659191Skris ca, 598759191Skris dsaparam -genkey (which also ignored its '-rand' option), 598859191Skris s_client, 598959191Skris s_server, 599059191Skris x509 (when signing). 599159191Skris Except on systems with /dev/urandom, it is crucial to have a random 599259191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 599359191Skris for RSA signatures we could do without one. 599459191Skris 599559191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 599659191Skris of each file listed in the '-rand' option. The function as previously 599759191Skris found in genrsa is now in app_rand.c and is used by all programs 599859191Skris that support '-rand'. 599959191Skris [Bodo Moeller] 600059191Skris 600159191Skris *) In RAND_write_file, use mode 0600 for creating files; 600259191Skris don't just chmod when it may be too late. 600359191Skris [Bodo Moeller] 600459191Skris 600559191Skris *) Report an error from X509_STORE_load_locations 600659191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 600759191Skris [Bill Perry] 600859191Skris 600959191Skris *) New function ASN1_mbstring_copy() this copies a string in either 601059191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 601159191Skris into an ASN1_STRING type. A mask of permissible types is passed 601259191Skris and it chooses the "minimal" type to use or an error if not type 601359191Skris is suitable. 601459191Skris [Steve Henson] 601559191Skris 601659191Skris *) Add function equivalents to the various macros in asn1.h. The old 601759191Skris macros are retained with an M_ prefix. Code inside the library can 601859191Skris use the M_ macros. External code (including the openssl utility) 601959191Skris should *NOT* in order to be "shared library friendly". 602059191Skris [Steve Henson] 602159191Skris 602259191Skris *) Add various functions that can check a certificate's extensions 602359191Skris to see if it usable for various purposes such as SSL client, 602459191Skris server or S/MIME and CAs of these types. This is currently 602559191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 602659191Skris verification. Also added a -purpose flag to x509 utility to 602759191Skris print out all the purposes. 602859191Skris [Steve Henson] 602959191Skris 603059191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 603159191Skris functions. 603259191Skris [Steve Henson] 603359191Skris 603459191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 603559191Skris for, obtain and decode and extension and obtain its critical flag. 603659191Skris This allows all the necessary extension code to be handled in a 603759191Skris single function call. 603859191Skris [Steve Henson] 603959191Skris 604059191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 604159191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 604259191Skris [Andy Polyakov] 604359191Skris 604459191Skris *) New -noout option to asn1parse. This causes no output to be produced 604559191Skris its main use is when combined with -strparse and -out to extract data 604659191Skris from a file (which may not be in ASN.1 format). 604759191Skris [Steve Henson] 604859191Skris 604959191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 605059191Skris when producing the local key id. 605159191Skris [Richard Levitte <levitte@stacken.kth.se>] 605259191Skris 605359191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 605459191Skris stated explicitly. If it is not stated then it tries the first server 605559191Skris certificate file. The previous behaviour hard coded the filename 605659191Skris "server.pem". 605759191Skris [Steve Henson] 605859191Skris 605959191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 606059191Skris a public key to be input or output. For example: 606159191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 606259191Skris Also added necessary DSA public key functions to handle this. 606359191Skris [Steve Henson] 606459191Skris 606559191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 606659191Skris in the message. This was handled by allowing 606759191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 606859191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 606959191Skris 607059191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 607159191Skris to the end of the strings whereas this didn't. This would cause problems 607259191Skris if strings read with d2i_ASN1_bytes() were later modified. 607359191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 607459191Skris 607559191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 607659191Skris data and it contains EOF it will end up returning an error. This is 607759191Skris caused by input 46 bytes long. The cause is due to the way base64 607859191Skris BIOs find the start of base64 encoded data. They do this by trying a 607959191Skris trial decode on each line until they find one that works. When they 608059191Skris do a flag is set and it starts again knowing it can pass all the 608159191Skris data directly through the decoder. Unfortunately it doesn't reset 608259191Skris the context it uses. This means that if EOF is reached an attempt 608359191Skris is made to pass two EOFs through the context and this causes the 608459191Skris resulting error. This can also cause other problems as well. As is 608559191Skris usual with these problems it takes *ages* to find and the fix is 608659191Skris trivial: move one line. 608759191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 608859191Skris 608959191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 609059191Skris old code wouldn't work because it needed to select() on sockets and the 609159191Skris tty (for keypresses and to see if data could be written). Win32 only 609259191Skris supports select() on sockets so we select() with a 1s timeout on the 609359191Skris sockets and then see if any characters are waiting to be read, if none 609459191Skris are present then we retry, we also assume we can always write data to 609559191Skris the tty. This isn't nice because the code then blocks until we've 609659191Skris received a complete line of data and it is effectively polling the 609759191Skris keyboard at 1s intervals: however it's quite a bit better than not 609859191Skris working at all :-) A dedicated Windows application might handle this 609959191Skris with an event loop for example. 610059191Skris [Steve Henson] 610159191Skris 610259191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 610359191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 610459191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 610559191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 610659191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 610759191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 610859191Skris This necessitated the support of an extra signature type NID_md5_sha1 610959191Skris for SSL signatures and modifications to the SSL library to use it instead 611059191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 611159191Skris [Steve Henson] 611259191Skris 611359191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 611459191Skris will lookup a CRL issuers certificate and verify the signature in a 611559191Skris similar way to the verify program. Tidy up the crl program so it 611659191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 611759191Skris less strict. It will now permit CRL extensions even if it is not 611859191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 611959191Skris [Steve Henson] 612059191Skris 612159191Skris *) Initialize all non-automatic variables each time one of the openssl 612259191Skris sub-programs is started (this is necessary as they may be started 612359191Skris multiple times from the "OpenSSL>" prompt). 612459191Skris [Lennart Bang, Bodo Moeller] 612559191Skris 612659191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 612759191Skris removing all other RSA functionality (this is what NO_RSA does). This 612859191Skris is so (for example) those in the US can disable those operations covered 612959191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 613059191Skris key generation. 613159191Skris [Steve Henson] 613259191Skris 613359191Skris *) Non-copying interface to BIO pairs. 613459191Skris (still largely untested) 613559191Skris [Bodo Moeller] 613659191Skris 613759191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 613859191Skris ASCII string. This was handled independently in various places before. 613959191Skris [Steve Henson] 614059191Skris 614159191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 614259191Skris UTF8 strings a character at a time. 614359191Skris [Steve Henson] 614459191Skris 614559191Skris *) Use client_version from client hello to select the protocol 614659191Skris (s23_srvr.c) and for RSA client key exchange verification 614759191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 614859191Skris [Bodo Moeller] 614959191Skris 615059191Skris *) Add various utility functions to handle SPKACs, these were previously 615159191Skris handled by poking round in the structure internals. Added new function 615259191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 615359191Skris print, verify and generate SPKACs. Based on an original idea from 615459191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 615559191Skris [Steve Henson] 615659191Skris 615759191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 615859191Skris [Andy Polyakov] 615959191Skris 616059191Skris *) Allow the config file extension section to be overwritten on the 616159191Skris command line. Based on an original idea from Massimiliano Pala 616259191Skris <madwolf@comune.modena.it>. The new option is called -extensions 616359191Skris and can be applied to ca, req and x509. Also -reqexts to override 616459191Skris the request extensions in req and -crlexts to override the crl extensions 616559191Skris in ca. 616659191Skris [Steve Henson] 616759191Skris 616859191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 616959191Skris the same field multiple times by preceding it by "XXXX." for example: 617059191Skris 1.OU="Unit name 1" 617159191Skris 2.OU="Unit name 2" 617259191Skris this is the same syntax as used in the req config file. 617359191Skris [Steve Henson] 617459191Skris 617559191Skris *) Allow certificate extensions to be added to certificate requests. These 617659191Skris are specified in a 'req_extensions' option of the req section of the 617759191Skris config file. They can be printed out with the -text option to req but 617859191Skris are otherwise ignored at present. 617959191Skris [Steve Henson] 618059191Skris 618159191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 618259191Skris data read consists of only the final block it would not decrypted because 618359191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 618459191Skris A misplaced 'break' also meant the decrypted final block might not be 618559191Skris copied until the next read. 618659191Skris [Steve Henson] 618759191Skris 618859191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 618959191Skris a few extra parameters to the DH structure: these will be useful if 619059191Skris for example we want the value of 'q' or implement X9.42 DH. 619159191Skris [Steve Henson] 619259191Skris 619359191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 619459191Skris provides hooks that allow the default DSA functions or functions on a 619559191Skris "per key" basis to be replaced. This allows hardware acceleration and 619659191Skris hardware key storage to be handled without major modification to the 619759191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 619859191Skris associated functions. 619959191Skris [Steve Henson] 620059191Skris 620159191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 620259191Skris as "read only": it can't be written to and the buffer it points to will 620359191Skris not be freed. Reading from a read only BIO is much more efficient than 620459191Skris a normal memory BIO. This was added because there are several times when 620559191Skris an area of memory needs to be read from a BIO. The previous method was 620659191Skris to create a memory BIO and write the data to it, this results in two 620759191Skris copies of the data and an O(n^2) reading algorithm. There is a new 620859191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 620959191Skris an area of memory. Also modified the PKCS#7 routines to use read only 621059191Skris memory BIOs. 621159191Skris [Steve Henson] 621259191Skris 621359191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 621459191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 621559191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 621659191Skris but a retry condition occured while trying to read the rest. 621759191Skris [Bodo Moeller] 621859191Skris 621959191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 622059191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 622159191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 622259191Skris the encrypted data type: this is a more sensible place to put it and it 622359191Skris allows the PKCS#12 code to be tidied up that duplicated this 622459191Skris functionality. 622559191Skris [Steve Henson] 622659191Skris 622759191Skris *) Changed obj_dat.pl script so it takes its input and output files on 622859191Skris the command line. This should avoid shell escape redirection problems 622959191Skris under Win32. 623059191Skris [Steve Henson] 623159191Skris 623259191Skris *) Initial support for certificate extension requests, these are included 623359191Skris in things like Xenroll certificate requests. Included functions to allow 623459191Skris extensions to be obtained and added. 623559191Skris [Steve Henson] 623659191Skris 623759191Skris *) -crlf option to s_client and s_server for sending newlines as 623859191Skris CRLF (as required by many protocols). 623959191Skris [Bodo Moeller] 624059191Skris 624155714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 624255714Skris 624355714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 624455714Skris [Ralf S. Engelschall] 624555714Skris 624655714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 624755714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 624855714Skris 624955714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 625055714Skris program. 625155714Skris [Steve Henson] 625255714Skris 625355714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 625455714Skris DH parameters/keys (q is lost during that conversion, but the resulting 625555714Skris DH parameters contain its length). 625655714Skris 625755714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 625855714Skris much faster than DH_generate_parameters (which creates parameters 625955714Skris where p = 2*q + 1), and also the smaller q makes DH computations 626055714Skris much more efficient (160-bit exponentiation instead of 1024-bit 626155714Skris exponentiation); so this provides a convenient way to support DHE 626255714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 626355714Skris utter importance to use 626455714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 626555714Skris or 626655714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 626755714Skris when such DH parameters are used, because otherwise small subgroup 626855714Skris attacks may become possible! 626955714Skris [Bodo Moeller] 627055714Skris 627155714Skris *) Avoid memory leak in i2d_DHparams. 627255714Skris [Bodo Moeller] 627355714Skris 627455714Skris *) Allow the -k option to be used more than once in the enc program: 627555714Skris this allows the same encrypted message to be read by multiple recipients. 627655714Skris [Steve Henson] 627755714Skris 627855714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 627955714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 628055714Skris it will always use the numerical form of the OID, even if it has a short 628155714Skris or long name. 628255714Skris [Steve Henson] 628355714Skris 628455714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 628555714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 628655714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 628755714Skris no private key components need be present and it might store extra data 628859191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 628959191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 629059191Skris private key operations. 629155714Skris [Steve Henson] 629255714Skris 629355714Skris *) Added support for SPARC Linux. 629455714Skris [Andy Polyakov] 629555714Skris 629655714Skris *) pem_password_cb function type incompatibly changed from 629755714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 629855714Skris to 629955714Skris ....(char *buf, int size, int rwflag, void *userdata); 630055714Skris so that applications can pass data to their callbacks: 630155714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 630255714Skris additional void * argument, which is just handed through whenever 630355714Skris the password callback is called. 630459191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 630555714Skris 630655714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 630755714Skris 630855714Skris Compatibility note: As many C implementations push function arguments 630955714Skris onto the stack in reverse order, the new library version is likely to 631055714Skris interoperate with programs that have been compiled with the old 631155714Skris pem_password_cb definition (PEM_whatever takes some data that 631255714Skris happens to be on the stack as its last argument, and the callback 631355714Skris just ignores this garbage); but there is no guarantee whatsoever that 631455714Skris this will work. 631555714Skris 631655714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 631755714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 631855714Skris problems not only on Windows, but also on some Unix platforms. 631955714Skris To avoid problematic command lines, these definitions are now in an 632055714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 632155714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 632255714Skris [Bodo Moeller] 632355714Skris 632455714Skris *) MIPS III/IV assembler module is reimplemented. 632555714Skris [Andy Polyakov] 632655714Skris 632755714Skris *) More DES library cleanups: remove references to srand/rand and 632855714Skris delete an unused file. 632955714Skris [Ulf M�ller] 633055714Skris 633155714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 633255714Skris since not many people have MASM (ml) and it can be hard to obtain. 633355714Skris This is currently experimental but it seems to work OK and pass all 633455714Skris the tests. Check out INSTALL.W32 for info. 633555714Skris [Steve Henson] 633655714Skris 633755714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 633855714Skris without temporary keys kept an extra copy of the server key, 633955714Skris and connections with temporary keys did not free everything in case 634055714Skris of an error. 634155714Skris [Bodo Moeller] 634255714Skris 634355714Skris *) New function RSA_check_key and new openssl rsa option -check 634455714Skris for verifying the consistency of RSA keys. 634555714Skris [Ulf Moeller, Bodo Moeller] 634655714Skris 634755714Skris *) Various changes to make Win32 compile work: 634855714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 634955714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 635055714Skris comparison" warnings. 635155714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 635255714Skris [Steve Henson] 635355714Skris 635455714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 635555714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 635655714Skris derived keys are printed to stderr. 635755714Skris [Steve Henson] 635855714Skris 635955714Skris *) Copy the flags in ASN1_STRING_dup(). 636055714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 636155714Skris 636255714Skris *) The x509 application mishandled signing requests containing DSA 636355714Skris keys when the signing key was also DSA and the parameters didn't match. 636455714Skris 636555714Skris It was supposed to omit the parameters when they matched the signing key: 636655714Skris the verifying software was then supposed to automatically use the CA's 636755714Skris parameters if they were absent from the end user certificate. 636855714Skris 636955714Skris Omitting parameters is no longer recommended. The test was also 637055714Skris the wrong way round! This was probably due to unusual behaviour in 637155714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 637255714Skris This meant that parameters were omitted when they *didn't* match and 637355714Skris the certificate was useless. Certificates signed with 'ca' didn't have 637455714Skris this bug. 637555714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 637655714Skris 637755714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 637855714Skris The interface is as follows: 637955714Skris Applications can use 638055714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 638155714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 638255714Skris "off" is now the default. 638355714Skris The library internally uses 638455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 638555714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 638655714Skris to disable memory-checking temporarily. 638755714Skris 638855714Skris Some inconsistent states that previously were possible (and were 638955714Skris even the default) are now avoided. 639055714Skris 639155714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 639255714Skris with each memory chunk allocated; this is occasionally more helpful 639355714Skris than just having a counter. 639455714Skris 639555714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 639655714Skris 639755714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 639855714Skris extensions. 639955714Skris [Bodo Moeller] 640055714Skris 640155714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 640255714Skris which largely parallels "options", but is for changing API behaviour, 640355714Skris whereas "options" are about protocol behaviour. 640455714Skris Initial "mode" flags are: 640555714Skris 640655714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 640755714Skris a single record has been written. 640855714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 640955714Skris retries use the same buffer location. 641055714Skris (But all of the contents must be 641155714Skris copied!) 641255714Skris [Bodo Moeller] 641355714Skris 641479998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 641555714Skris worked. 641655714Skris 641755714Skris *) Fix problems with no-hmac etc. 641855714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 641955714Skris 642055714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 642155714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 642255714Skris to mess around with the internals of an RSA structure. 642355714Skris [Steve Henson] 642455714Skris 642555714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 642655714Skris Also really enable memory leak checks in openssl.c and in some 642755714Skris test programs. 642855714Skris [Chad C. Mulligan, Bodo Moeller] 642955714Skris 643055714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 643155714Skris up the length of negative integers. This has now been simplified to just 643255714Skris store the length when it is first determined and use it later, rather 643355714Skris than trying to keep track of where data is copied and updating it to 643455714Skris point to the end. 643555714Skris [Steve Henson, reported by Brien Wheeler 643655714Skris <bwheeler@authentica-security.com>] 643755714Skris 643855714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 643955714Skris of a PKCS#7 signature but with the signing certificate passed to the 644055714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 644155714Skris certificate is present in the PKCS#7 structure. This isn't always the 644255714Skris case: certificates can be omitted from a PKCS#7 structure and be 644355714Skris distributed by "out of band" means (such as a certificate database). 644455714Skris [Steve Henson] 644555714Skris 644655714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 644755714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 644855714Skris necessary function names. 644955714Skris [Steve Henson] 645055714Skris 645155714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 645255714Skris options set by Configure in the top level Makefile, and Configure 645355714Skris was not even able to write more than one option correctly. 645455714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 645555714Skris [Bodo Moeller] 645655714Skris 645755714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 645855714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 645955714Skris for example allow memory BIOs to contain config info. 646055714Skris [Steve Henson] 646155714Skris 646255714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 646355714Skris Whoever hopes to achieve shared-library compatibility across versions 646455714Skris must use this, not the compile-time macro. 646555714Skris (Exercise 0.9.4: Which is the minimum library version required by 646655714Skris such programs?) 646755714Skris Note: All this applies only to multi-threaded programs, others don't 646855714Skris need locks. 646955714Skris [Bodo Moeller] 647055714Skris 647155714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 647255714Skris through a BIO pair triggered the default case, i.e. 647355714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 647455714Skris [Bodo Moeller] 647555714Skris 647655714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 647755714Skris can use the SSL library even if none of the specific BIOs is 647855714Skris appropriate. 647955714Skris [Bodo Moeller] 648055714Skris 648155714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 648255714Skris for the encoded length. 648355714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 648455714Skris 648555714Skris *) Add initial documentation of the X509V3 functions. 648655714Skris [Steve Henson] 648755714Skris 648855714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 648955714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 649055714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 649155714Skris secure PKCS#8 private key format with a high iteration count. 649255714Skris [Steve Henson] 649355714Skris 649455714Skris *) Fix determination of Perl interpreter: A perl or perl5 649555714Skris _directory_ in $PATH was also accepted as the interpreter. 649655714Skris [Ralf S. Engelschall] 649755714Skris 649855714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 649955714Skris wrong with it but it was very old and did things like calling 650055714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 650155714Skris unusual formatting. 650255714Skris [Steve Henson] 650355714Skris 650455714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 650555714Skris to use the new extension code. 650655714Skris [Steve Henson] 650755714Skris 650855714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 650955714Skris with macros. This should make it easier to change their form, add extra 651055714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 651155714Skris constant. 651255714Skris [Steve Henson] 651355714Skris 651455714Skris *) Add to configuration table a new entry that can specify an alternative 651555714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 651655714Skris according to Mark Crispin <MRC@Panda.COM>. 651755714Skris [Bodo Moeller] 651855714Skris 651955714Skris#if 0 652055714Skris *) DES CBC did not update the IV. Weird. 652155714Skris [Ben Laurie] 652255714Skris#else 652355714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 652455714Skris Changing the behaviour of the former might break existing programs -- 652555714Skris where IV updating is needed, des_ncbc_encrypt can be used. 652655714Skris#endif 652755714Skris 652855714Skris *) When bntest is run from "make test" it drives bc to check its 652955714Skris calculations, as well as internally checking them. If an internal check 653055714Skris fails, it needs to cause bc to give a non-zero result or make test carries 653155714Skris on without noticing the failure. Fixed. 653255714Skris [Ben Laurie] 653355714Skris 653455714Skris *) DES library cleanups. 653555714Skris [Ulf M�ller] 653655714Skris 653755714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 653855714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 653955714Skris ciphers. NOTE: although the key derivation function has been verified 654055714Skris against some published test vectors it has not been extensively tested 654155714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 654255714Skris of v2.0. 654355714Skris [Steve Henson] 654455714Skris 654555714Skris *) Instead of "mkdir -p", which is not fully portable, use new 654655714Skris Perl script "util/mkdir-p.pl". 654755714Skris [Bodo Moeller] 654855714Skris 654955714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 655055714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 655155714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 655255714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 655355714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 655455714Skris underlying key generation function so it must do its own ASN1 parsing. 655555714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 655655714Skris 'parameter' argument instead of literal salt and iteration count values 655755714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 655855714Skris [Steve Henson] 655955714Skris 656055714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 656155714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 656255714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 656355714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 656455714Skris value was just used as a "magic string" and not used directly its 656555714Skris value doesn't matter. 656655714Skris [Steve Henson] 656755714Skris 656855714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 656955714Skris support mutable. 657055714Skris [Ben Laurie] 657155714Skris 657255714Skris *) "linux-sparc64" configuration (ultrapenguin). 657355714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 657455714Skris "linux-sparc" configuration. 657555714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 657655714Skris 657755714Skris *) config now generates no-xxx options for missing ciphers. 657855714Skris [Ulf M�ller] 657955714Skris 658055714Skris *) Support the EBCDIC character set (work in progress). 658155714Skris File ebcdic.c not yet included because it has a different license. 658255714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 658355714Skris 658455714Skris *) Support BS2000/OSD-POSIX. 658555714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 658655714Skris 658755714Skris *) Make callbacks for key generation use void * instead of char *. 658855714Skris [Ben Laurie] 658955714Skris 659055714Skris *) Make S/MIME samples compile (not yet tested). 659155714Skris [Ben Laurie] 659255714Skris 659355714Skris *) Additional typesafe stacks. 659455714Skris [Ben Laurie] 659555714Skris 659655714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 659755714Skris [Bodo Moeller] 659855714Skris 659955714Skris 660055714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 660155714Skris 660255714Skris *) New configuration variant "sco5-gcc". 660355714Skris 660455714Skris *) Updated some demos. 660555714Skris [Sean O Riordain, Wade Scholine] 660655714Skris 660755714Skris *) Add missing BIO_free at exit of pkcs12 application. 660855714Skris [Wu Zhigang] 660955714Skris 661055714Skris *) Fix memory leak in conf.c. 661155714Skris [Steve Henson] 661255714Skris 661355714Skris *) Updates for Win32 to assembler version of MD5. 661455714Skris [Steve Henson] 661555714Skris 661655714Skris *) Set #! path to perl in apps/der_chop to where we found it 661755714Skris instead of using a fixed path. 661855714Skris [Bodo Moeller] 661955714Skris 662055714Skris *) SHA library changes for irix64-mips4-cc. 662155714Skris [Andy Polyakov] 662255714Skris 662355714Skris *) Improvements for VMS support. 662455714Skris [Richard Levitte] 662555714Skris 662655714Skris 662755714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 662855714Skris 662955714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 663055714Skris This also avoids the problems with SC4.2 and unpatched SC5. 663155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 663255714Skris 663355714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 663455714Skris These are required because of the typesafe stack would otherwise break 663555714Skris existing code. If old code used a structure member which used to be STACK 663655714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 663755714Skris sk_num or sk_value it would produce an error because the num, data members 663855714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 663955714Skris replaces the old method of assigning a value to sk_value 664055714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 664155714Skris that does this will no longer work (and should use sk_set instead) but 664255714Skris this could be regarded as a "questionable" behaviour anyway. 664355714Skris [Steve Henson] 664455714Skris 664555714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 664655714Skris correctly handle encrypted S/MIME data. 664755714Skris [Steve Henson] 664855714Skris 664955714Skris *) Change type of various DES function arguments from des_cblock 665055714Skris (which means, in function argument declarations, pointer to char) 665155714Skris to des_cblock * (meaning pointer to array with 8 char elements), 665255714Skris which allows the compiler to do more typechecking; it was like 665355714Skris that back in SSLeay, but with lots of ugly casts. 665455714Skris 665555714Skris Introduce new type const_des_cblock. 665655714Skris [Bodo Moeller] 665755714Skris 665855714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 665955714Skris problems: find RecipientInfo structure that matches recipient certificate 666055714Skris and initialise the ASN1 structures properly based on passed cipher. 666155714Skris [Steve Henson] 666255714Skris 666355714Skris *) Belatedly make the BN tests actually check the results. 666455714Skris [Ben Laurie] 666555714Skris 666655714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 666755714Skris to and from BNs: it was completely broken. New compilation option 666855714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 666955714Skris key elements as negative integers. 667055714Skris [Steve Henson] 667155714Skris 667255714Skris *) Reorganize and speed up MD5. 667355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 667455714Skris 667555714Skris *) VMS support. 667655714Skris [Richard Levitte <richard@levitte.org>] 667755714Skris 667855714Skris *) New option -out to asn1parse to allow the parsed structure to be 667955714Skris output to a file. This is most useful when combined with the -strparse 668055714Skris option to examine the output of things like OCTET STRINGS. 668155714Skris [Steve Henson] 668255714Skris 668355714Skris *) Make SSL library a little more fool-proof by not requiring any longer 668455714Skris that SSL_set_{accept,connect}_state be called before 668555714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 668655714Skris in many applications because usually everything *appeared* to work as 668755714Skris intended anyway -- now it really works as intended). 668855714Skris [Bodo Moeller] 668955714Skris 669055714Skris *) Move openssl.cnf out of lib/. 669155714Skris [Ulf M�ller] 669255714Skris 669355714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 669455714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 669555714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 669655714Skris [Ralf S. Engelschall] 669755714Skris 669855714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 669955714Skris handle PKCS#7 enveloped data properly. 670055714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 670155714Skris 670255714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 670355714Skris copying pointers. The cert_st handling is changed by this in 670455714Skris various ways (and thus what used to be known as ctx->default_cert 670555714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 670655714Skris any longer when s->cert does not give us what we need). 670755714Skris ssl_cert_instantiate becomes obsolete by this change. 670855714Skris As soon as we've got the new code right (possibly it already is?), 670955714Skris we have solved a couple of bugs of the earlier code where s->cert 671055714Skris was used as if it could not have been shared with other SSL structures. 671155714Skris 671255714Skris Note that using the SSL API in certain dirty ways now will result 671355714Skris in different behaviour than observed with earlier library versions: 671455714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 671555714Skris does not influence s as it used to. 671655714Skris 671755714Skris In order to clean up things more thoroughly, inside SSL_SESSION 671855714Skris we don't use CERT any longer, but a new structure SESS_CERT 671955714Skris that holds per-session data (if available); currently, this is 672055714Skris the peer's certificate chain and, for clients, the server's certificate 672155714Skris and temporary key. CERT holds only those values that can have 672255714Skris meaningful defaults in an SSL_CTX. 672355714Skris [Bodo Moeller] 672455714Skris 672555714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 672655714Skris from the internal representation. Various PKCS#7 fixes: remove some 672755714Skris evil casts and set the enc_dig_alg field properly based on the signing 672855714Skris key type. 672955714Skris [Steve Henson] 673055714Skris 673155714Skris *) Allow PKCS#12 password to be set from the command line or the 673255714Skris environment. Let 'ca' get its config file name from the environment 673355714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 673455714Skris and 'x509'). 673555714Skris [Steve Henson] 673655714Skris 673755714Skris *) Allow certificate policies extension to use an IA5STRING for the 673855714Skris organization field. This is contrary to the PKIX definition but 673955714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 674055714Skris extension option. 674155714Skris [Steve Henson] 674255714Skris 674355714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 674455714Skris without disallowing inline assembler and the like for non-pedantic builds. 674555714Skris [Ben Laurie] 674655714Skris 674755714Skris *) Support Borland C++ builder. 674855714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 674955714Skris 675055714Skris *) Support Mingw32. 675155714Skris [Ulf M�ller] 675255714Skris 675355714Skris *) SHA-1 cleanups and performance enhancements. 675455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 675555714Skris 675655714Skris *) Sparc v8plus assembler for the bignum library. 675755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 675855714Skris 675955714Skris *) Accept any -xxx and +xxx compiler options in Configure. 676055714Skris [Ulf M�ller] 676155714Skris 676255714Skris *) Update HPUX configuration. 676355714Skris [Anonymous] 676455714Skris 676555714Skris *) Add missing sk_<type>_unshift() function to safestack.h 676655714Skris [Ralf S. Engelschall] 676755714Skris 676855714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 676955714Skris "extra_cert"s in addition to the certificate. (This makes sense 677055714Skris only for "PEM" format files, as chains as a whole are not 677155714Skris DER-encoded.) 677255714Skris [Bodo Moeller] 677355714Skris 677455714Skris *) Support verify_depth from the SSL API. 677555714Skris x509_vfy.c had what can be considered an off-by-one-error: 677655714Skris Its depth (which was not part of the external interface) 677755714Skris was actually counting the number of certificates in a chain; 677855714Skris now it really counts the depth. 677955714Skris [Bodo Moeller] 678055714Skris 678155714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 678255714Skris instead of X509err, which often resulted in confusing error 678355714Skris messages since the error codes are not globally unique 678455714Skris (e.g. an alleged error in ssl3_accept when a certificate 678555714Skris didn't match the private key). 678655714Skris 678755714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 678855714Skris value (so that you don't need SSL_set_session_id_context for each 678955714Skris connection using the SSL_CTX). 679055714Skris [Bodo Moeller] 679155714Skris 679255714Skris *) OAEP decoding bug fix. 679355714Skris [Ulf M�ller] 679455714Skris 679555714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 679655714Skris David Harris. 679755714Skris [Bodo Moeller] 679855714Skris 679955714Skris *) New Configure options "threads" and "no-threads". For systems 680055714Skris where the proper compiler options are known (currently Solaris 680155714Skris and Linux), "threads" is the default. 680255714Skris [Bodo Moeller] 680355714Skris 680455714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 680555714Skris [Bodo Moeller] 680655714Skris 680755714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 680855714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 680955714Skris such as /usr/local/bin. 681055714Skris [Bodo Moeller] 681155714Skris 681255714Skris *) "make linux-shared" to build shared libraries. 681355714Skris [Niels Poppe <niels@netbox.org>] 681455714Skris 681555714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 681655714Skris [Ulf M�ller] 681755714Skris 681855714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 681955714Skris extension adding in x509 utility. 682055714Skris [Steve Henson] 682155714Skris 682255714Skris *) Remove NOPROTO sections and error code comments. 682355714Skris [Ulf M�ller] 682455714Skris 682555714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 682655714Skris prototypes. 682755714Skris [Steve Henson] 682855714Skris 682955714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 683055714Skris [Ulf M�ller] 683155714Skris 683255714Skris *) Complete rewrite of the error code script(s). It is all now handled 683355714Skris by one script at the top level which handles error code gathering, 683455714Skris header rewriting and C source file generation. It should be much better 683555714Skris than the old method: it now uses a modified version of Ulf's parser to 683655714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 683755714Skris aren't needed for error creation any more) and do a better job of 683855714Skris translating function codes into names. The old 'ASN1 error code imbedded 683955714Skris in a comment' is no longer necessary and it doesn't use .err files which 684055714Skris have now been deleted. Also the error code call doesn't have to appear all 684155714Skris on one line (which resulted in some large lines...). 684255714Skris [Steve Henson] 684355714Skris 684455714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 684555714Skris [Bodo Moeller] 684655714Skris 684755714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 684855714Skris 0 (which usually indicates a closed connection), but continue reading. 684955714Skris [Bodo Moeller] 685055714Skris 685155714Skris *) Fix some race conditions. 685255714Skris [Bodo Moeller] 685355714Skris 685455714Skris *) Add support for CRL distribution points extension. Add Certificate 685555714Skris Policies and CRL distribution points documentation. 685655714Skris [Steve Henson] 685755714Skris 685855714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 685955714Skris [Ulf M�ller] 686055714Skris 686155714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 686255714Skris 8 of keying material. Merlin has also confirmed interop with this fix 686355714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 686455714Skris [Merlin Hughes <merlin@baltimore.ie>] 686555714Skris 686655714Skris *) Fix lots of warnings. 686755714Skris [Richard Levitte <levitte@stacken.kth.se>] 686855714Skris 686955714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 687055714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 687155714Skris [Richard Levitte <levitte@stacken.kth.se>] 687255714Skris 687355714Skris *) Fix problems with sizeof(long) == 8. 687455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 687555714Skris 687655714Skris *) Change functions to ANSI C. 687755714Skris [Ulf M�ller] 687855714Skris 687955714Skris *) Fix typos in error codes. 688055714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 688155714Skris 688255714Skris *) Remove defunct assembler files from Configure. 688355714Skris [Ulf M�ller] 688455714Skris 688555714Skris *) SPARC v8 assembler BIGNUM implementation. 688655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 688755714Skris 688855714Skris *) Support for Certificate Policies extension: both print and set. 688955714Skris Various additions to support the r2i method this uses. 689055714Skris [Steve Henson] 689155714Skris 689255714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 689355714Skris return a const string when you are expecting an allocated buffer. 689455714Skris [Ben Laurie] 689555714Skris 689655714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 689755714Skris types DirectoryString and DisplayText. 689855714Skris [Steve Henson] 689955714Skris 690055714Skris *) Add code to allow r2i extensions to access the configuration database, 690155714Skris add an LHASH database driver and add several ctx helper functions. 690255714Skris [Steve Henson] 690355714Skris 690455714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 690555714Skris fail when they extended the size of a BIGNUM. 690655714Skris [Steve Henson] 690755714Skris 690855714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 690955714Skris support typesafe stack. 691055714Skris [Steve Henson] 691155714Skris 691255714Skris *) Fix typo in SSL_[gs]et_options(). 691355714Skris [Nils Frostberg <nils@medcom.se>] 691455714Skris 691555714Skris *) Delete various functions and files that belonged to the (now obsolete) 691655714Skris old X509V3 handling code. 691755714Skris [Steve Henson] 691855714Skris 691955714Skris *) New Configure option "rsaref". 692055714Skris [Ulf M�ller] 692155714Skris 692255714Skris *) Don't auto-generate pem.h. 692355714Skris [Bodo Moeller] 692455714Skris 692555714Skris *) Introduce type-safe ASN.1 SETs. 692655714Skris [Ben Laurie] 692755714Skris 692855714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 692955714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 693055714Skris 693155714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 693255714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 693355714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 693455714Skris few STACKed things have been converted already. Feel free to convert more. 693555714Skris In the fullness of time, I'll do away with the STACK type altogether. 693655714Skris [Ben Laurie] 693755714Skris 693855714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 693955714Skris specified in <certfile> by updating the entry in the index.txt file. 694055714Skris This way one no longer has to edit the index.txt file manually for 694155714Skris revoking a certificate. The -revoke option does the gory details now. 694255714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 694355714Skris 694455714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 694555714Skris `-text' option at all and this way the `-noout -text' combination was 694655714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 694755714Skris [Ralf S. Engelschall] 694855714Skris 694955714Skris *) Make sure a corresponding plain text error message exists for the 695055714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 695155714Skris verify callback function determined that a certificate was revoked. 695255714Skris [Ralf S. Engelschall] 695355714Skris 695455714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 695555714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 695655714Skris all available cipers including rc5, which was forgotten until now. 695755714Skris In order to let the testing shell script know which algorithms 695855714Skris are available, a new (up to now undocumented) command 695955714Skris "openssl list-cipher-commands" is used. 696055714Skris [Bodo Moeller] 696155714Skris 696255714Skris *) Bugfix: s_client occasionally would sleep in select() when 696355714Skris it should have checked SSL_pending() first. 696455714Skris [Bodo Moeller] 696555714Skris 696655714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 696755714Skris the raw DSA values prior to ASN.1 encoding. 696855714Skris [Ulf M�ller] 696955714Skris 697055714Skris *) Tweaks to Configure 697155714Skris [Niels Poppe <niels@netbox.org>] 697255714Skris 697355714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 697455714Skris yet... 697555714Skris [Steve Henson] 697655714Skris 697755714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 697855714Skris [Ulf M�ller] 697955714Skris 698055714Skris *) New config option to avoid instructions that are illegal on the 80386. 698155714Skris The default code is faster, but requires at least a 486. 698255714Skris [Ulf M�ller] 698355714Skris 698455714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 698555714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 698655714Skris same as SSL2_VERSION anyway. 698755714Skris [Bodo Moeller] 698855714Skris 698955714Skris *) New "-showcerts" option for s_client. 699055714Skris [Bodo Moeller] 699155714Skris 699255714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 699355714Skris application. Various cleanups and fixes. 699455714Skris [Steve Henson] 699555714Skris 699655714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 699755714Skris modify error routines to work internally. Add error codes and PBE init 699855714Skris to library startup routines. 699955714Skris [Steve Henson] 700055714Skris 700155714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 700255714Skris packing functions to asn1 and evp. Changed function names and error 700355714Skris codes along the way. 700455714Skris [Steve Henson] 700555714Skris 700655714Skris *) PKCS12 integration: and so it begins... First of several patches to 700755714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 700855714Skris objects to objects.h 700955714Skris [Steve Henson] 701055714Skris 701155714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 701255714Skris and display support for Thawte strong extranet extension. 701355714Skris [Steve Henson] 701455714Skris 701555714Skris *) Add LinuxPPC support. 701655714Skris [Jeff Dubrule <igor@pobox.org>] 701755714Skris 701855714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 701955714Skris bn_div_words in alpha.s. 702055714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 702155714Skris 702255714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 702355714Skris OAEP isn't supported when OpenSSL is built with RSAref. 702455714Skris [Ulf Moeller <ulf@fitug.de>] 702555714Skris 702655714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 702755714Skris so they no longer are missing under -DNOPROTO. 702855714Skris [Soren S. Jorvang <soren@t.dk>] 702955714Skris 703055714Skris 703155714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 703255714Skris 703355714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 703455714Skris doesn't work when the session is reused. Coming soon! 703555714Skris [Ben Laurie] 703655714Skris 703755714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 703855714Skris context thus bypassing client cert protection! All software that uses 703955714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 704055714Skris allow session reuse! A fuller solution is in the works. 704155714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 704255714Skris 704355714Skris *) Some more source tree cleanups (removed obsolete files 704455714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 704555714Skris permission on "config" script to be executable) and a fix for the INSTALL 704655714Skris document. 704755714Skris [Ulf Moeller <ulf@fitug.de>] 704855714Skris 704955714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 705055714Skris Malloc, Free. 705155714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 705255714Skris 705355714Skris *) Make rsa_oaep_test return non-zero on error. 705455714Skris [Ulf Moeller <ulf@fitug.de>] 705555714Skris 705655714Skris *) Add support for native Solaris shared libraries. Configure 705755714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 705855714Skris if someone would make that last step automatic. 705955714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 706055714Skris 706155714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 706255714Skris [Ben Laurie] 706355714Skris 706455714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 706555714Skris except NULL ciphers". This means the default cipher list will no longer 706655714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 706755714Skris the string "DEFAULT:eNULL". 706855714Skris [Steve Henson] 706955714Skris 707055714Skris *) Fix to RSA private encryption routines: if p < q then it would 707155714Skris occasionally produce an invalid result. This will only happen with 707255714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 707355714Skris [Steve Henson] 707455714Skris 707555714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 707655714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 707755714Skris because this way one can also use an interpreter named `perl5' (which is 707855714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 707955714Skris installed as `perl'). 708055714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 708155714Skris 708255714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 708355714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 708455714Skris 708555714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 708655714Skris advapi32.lib to Win32 build and change the pem test comparision 708755714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 708855714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 708955714Skris and crypto/des/ede_cbcm_enc.c. 709055714Skris [Steve Henson] 709155714Skris 709255714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 709355714Skris [Ben Laurie] 709455714Skris 709555714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 709655714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 709755714Skris is horrible: I feel ill.... 709855714Skris [Steve Henson] 709955714Skris 710055714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 710155714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 710255714Skris sections: 10 functions were absent from non ANSI section and not exported 710355714Skris from Windows DLLs. Fixed up libeay.num for new functions. 710455714Skris [Steve Henson] 710555714Skris 710655714Skris *) Make `openssl version' output lines consistent. 710755714Skris [Ralf S. Engelschall] 710855714Skris 710955714Skris *) Fix Win32 symbol export lists for BIO functions: Added 711055714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 711155714Skris to ms/libeay{16,32}.def. 711255714Skris [Ralf S. Engelschall] 711355714Skris 711455714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 711555714Skris fine under Unix and passes some trivial tests I've now added. But the 711655714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 711755714Skris added to make sure no one expects that this stuff really works in the 711855714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 711955714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 712055714Skris openssl_bio.xs. 712155714Skris [Ralf S. Engelschall] 712255714Skris 712355714Skris *) Fix the generation of two part addresses in perl. 712455714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 712555714Skris 712655714Skris *) Add config entry for Linux on MIPS. 712755714Skris [John Tobey <jtobey@channel1.com>] 712855714Skris 712955714Skris *) Make links whenever Configure is run, unless we are on Windoze. 713055714Skris [Ben Laurie] 713155714Skris 713255714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 713355714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 713455714Skris in CRLs. 713555714Skris [Steve Henson] 713655714Skris 713755714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 713855714Skris other platforms details on the command line without having to patch the 713955714Skris Configure script everytime: One now can use ``perl Configure 714055714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 714155714Skris to them (seperated by colons). This is treated as there would be a static 714255714Skris pre-configured entry in Configure's %table under key <id> with value 714355714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 714455714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 714555714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 714655714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 714755714Skris [Ralf S. Engelschall] 714855714Skris 714955714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 715055714Skris [Ben Laurie] 715155714Skris 715255714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 715355714Skris on the `perl Configure ...' command line. This way one can compile 715455714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 715555714Skris for linking it into DSOs. 715655714Skris [Ralf S. Engelschall] 715755714Skris 715855714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 715955714Skris Fixed. 716055714Skris [Ben Laurie] 716155714Skris 716255714Skris *) Cleaned up the LICENSE document: The official contact for any license 716355714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 716455714Skris And add a paragraph about the dual-license situation to make sure people 716555714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 716655714Skris to the OpenSSL toolkit. 716755714Skris [Ralf S. Engelschall] 716855714Skris 716955714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 717055714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 717155714Skris Additonally cleaned up the `make links' target: Remove unnecessary 717255714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 717355714Skris to speed processing and no longer clutter the display with confusing 717455714Skris stuff. Instead only the actually done links are displayed. 717555714Skris [Ralf S. Engelschall] 717655714Skris 717755714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 717855714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 717955714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 718055714Skris encryption. 718155714Skris [Ben Laurie] 718255714Skris 718355714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 718455714Skris signed attributes when verifying signatures (this would break them), 718555714Skris the detached data encoding was wrong and public keys obtained using 718655714Skris X509_get_pubkey() weren't freed. 718755714Skris [Steve Henson] 718855714Skris 718955714Skris *) Add text documentation for the BUFFER functions. Also added a work around 719055714Skris to a Win95 console bug. This was triggered by the password read stuff: the 719155714Skris last character typed gets carried over to the next fread(). If you were 719255714Skris generating a new cert request using 'req' for example then the last 719355714Skris character of the passphrase would be CR which would then enter the first 719455714Skris field as blank. 719555714Skris [Steve Henson] 719655714Skris 719755714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 719855714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 719955714Skris button and can be used by applications based on OpenSSL to show the 720055714Skris relationship to the OpenSSL project. 720155714Skris [Ralf S. Engelschall] 720255714Skris 720355714Skris *) Remove confusing variables in function signatures in files 720455714Skris ssl/ssl_lib.c and ssl/ssl.h. 720555714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 720655714Skris 720755714Skris *) Don't install bss_file.c under PREFIX/include/ 720855714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 720955714Skris 721055714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 721155714Skris functions that return function pointers and has support for NT specific 721255714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 721355714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 721455714Skris unsigned to signed types: this was killing the Win32 compile. 721555714Skris [Steve Henson] 721655714Skris 721755714Skris *) Add new certificate file to stack functions, 721855714Skris SSL_add_dir_cert_subjects_to_stack() and 721955714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 722055714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 722155714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 722255714Skris This means that Apache-SSL and similar packages don't have to mess around 722355714Skris to add as many CAs as they want to the preferred list. 722455714Skris [Ben Laurie] 722555714Skris 722655714Skris *) Experiment with doxygen documentation. Currently only partially applied to 722755714Skris ssl/ssl_lib.c. 722855714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 722955714Skris openssl.doxy as the configuration file. 723055714Skris [Ben Laurie] 723155714Skris 723255714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 723355714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 723455714Skris 723555714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 723655714Skris compiled in by default: it has problems with large keys. 723755714Skris [Steve Henson] 723855714Skris 723955714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 724055714Skris DH private keys and/or callback functions which directly correspond to 724155714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 724255714Skris is needed for applications which have to configure certificates on a 724355714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 724455714Skris (e.g. s_server). 724555714Skris For the RSA certificate situation is makes no difference, but 724655714Skris for the DSA certificate situation this fixes the "no shared cipher" 724755714Skris problem where the OpenSSL cipher selection procedure failed because the 724855714Skris temporary keys were not overtaken from the context and the API provided 724955714Skris no way to reconfigure them. 725055714Skris The new functions now let applications reconfigure the stuff and they 725155714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 725255714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 725355714Skris non-public-API function ssl_cert_instantiate() is used as a helper 725455714Skris function and also to reduce code redundancy inside ssl_rsa.c. 725555714Skris [Ralf S. Engelschall] 725655714Skris 725755714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 725855714Skris area because they are useful for the DSA situation and should be 725955714Skris recognized by the users. 726055714Skris [Ralf S. Engelschall] 726155714Skris 726255714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 726355714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 726455714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 726555714Skris already masked variable. 726655714Skris [Richard Levitte <levitte@stacken.kth.se>] 726755714Skris 726855714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 726955714Skris [Richard Levitte <levitte@stacken.kth.se>] 727055714Skris 727155714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 727255714Skris from `int' to `unsigned int' because it's a length and initialized by 727355714Skris EVP_DigestFinal() which expects an `unsigned int *'. 727455714Skris [Richard Levitte <levitte@stacken.kth.se>] 727555714Skris 727655714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 727755714Skris script. Instead use the usual Shell->Perl transition trick. 727855714Skris [Ralf S. Engelschall] 727955714Skris 728055714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 728155714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 728255714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 728355714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 728455714Skris currently the public key is printed (a decision which was already done by 728555714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 728655714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 728755714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 728855714Skris now, too. 728955714Skris [Ralf S. Engelschall] 729055714Skris 729155714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 729255714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 729355714Skris [Arne Ansper <arne@ats.cyber.ee>] 729455714Skris 729555714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 729655714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 729755714Skris config file. 729855714Skris [Steve Henson] 729955714Skris 730055714Skris *) Add cool BIO that does syslog (or event log on NT). 730155714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 730255714Skris 730355714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 730455714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 730555714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 730655714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 730755714Skris [Ben Laurie] 730855714Skris 730955714Skris *) Add preliminary config info for new extension code. 731055714Skris [Steve Henson] 731155714Skris 731255714Skris *) Make RSA_NO_PADDING really use no padding. 731355714Skris [Ulf Moeller <ulf@fitug.de>] 731455714Skris 731555714Skris *) Generate errors when private/public key check is done. 731655714Skris [Ben Laurie] 731755714Skris 731855714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 731955714Skris for some CRL extensions and new objects added. 732055714Skris [Steve Henson] 732155714Skris 732255714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 732355714Skris key usage extension and fuller support for authority key id. 732455714Skris [Steve Henson] 732555714Skris 732655714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 732755714Skris padding method for RSA, which is recommended for new applications in PKCS 732855714Skris #1 v2.0 (RFC 2437, October 1998). 732955714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 733055714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 733155714Skris against Bleichbacher's attack on RSA. 733255714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 733355714Skris Ben Laurie] 733455714Skris 733555714Skris *) Updates to the new SSL compression code 733655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 733755714Skris 733855714Skris *) Fix so that the version number in the master secret, when passed 733955714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 734055714Skris (because the server will not accept higher), that the version number 734155714Skris is 0x03,0x01, not 0x03,0x00 734255714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 734355714Skris 734455714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 734555714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 734655714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 734755714Skris [Steve Henson] 734855714Skris 734955714Skris *) Support for RAW extensions where an arbitrary extension can be 735055714Skris created by including its DER encoding. See apps/openssl.cnf for 735155714Skris an example. 735255714Skris [Steve Henson] 735355714Skris 735455714Skris *) Make sure latest Perl versions don't interpret some generated C array 735555714Skris code as Perl array code in the crypto/err/err_genc.pl script. 735655714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 735755714Skris 735855714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 735955714Skris not many people have the assembler. Various Win32 compilation fixes and 736055714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 736155714Skris build instructions. 736255714Skris [Steve Henson] 736355714Skris 736455714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 736555714Skris file under Win32 and also build pem.h from pem.org. New script 736655714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 736755714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 736855714Skris [Steve Henson] 736955714Skris 737055714Skris *) Major rework of DES function declarations, in the pursuit of correctness 737155714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 737255714Skris too. You may find this causes warnings in your code. Zapping your evil 737355714Skris casts will probably fix them. Mostly. 737455714Skris [Ben Laurie] 737555714Skris 737655714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 737755714Skris obj_dat.pl. It considered a zero in an object definition to mean 737855714Skris "end of object": none of the objects in objects.h have any zeros 737955714Skris so it wasn't spotted. 738055714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 738155714Skris 738255714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 738355714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 738455714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 738555714Skris vectors if you have them. 738655714Skris [Ben Laurie] 738755714Skris 738855714Skris *) Correct calculation of key length for export ciphers (too much space was 738955714Skris allocated for null ciphers). This has not been tested! 739055714Skris [Ben Laurie] 739155714Skris 739255714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 739355714Skris message is now correct (it understands "crypto" and "ssl" on its 739455714Skris command line). There is also now an "update" option. This will update 739555714Skris the util/ssleay.num and util/libeay.num files with any new functions. 739655714Skris If you do a: 739755714Skris perl util/mkdef.pl crypto ssl update 739855714Skris it will update them. 739955714Skris [Steve Henson] 740055714Skris 740155714Skris *) Overhauled the Perl interface (perl/*): 740255714Skris - ported BN stuff to OpenSSL's different BN library 740355714Skris - made the perl/ source tree CVS-aware 740455714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 740555714Skris their history because I've copied them in the repository) 740655714Skris - removed obsolete files (the test scripts will be replaced 740755714Skris by better Test::Harness variants in the future) 740855714Skris [Ralf S. Engelschall] 740955714Skris 741055714Skris *) First cut for a very conservative source tree cleanup: 741155714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 741255714Skris where we collect the old documents and readme texts. 741355714Skris 2. remove the first part of files where I'm already sure that we no 741455714Skris longer need them because of three reasons: either they are just temporary 741555714Skris files which were left by Eric or they are preserved original files where 741655714Skris I've verified that the diff is also available in the CVS via "cvs diff 741755714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 741855714Skris the crypto/md/ stuff). 741955714Skris [Ralf S. Engelschall] 742055714Skris 742155714Skris *) More extension code. Incomplete support for subject and issuer alt 742255714Skris name, issuer and authority key id. Change the i2v function parameters 742355714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 742455714Skris what that's for :-) Fix to ASN1 macro which messed up 742555714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 742655714Skris [Steve Henson] 742755714Skris 742855714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 742955714Skris INTEGER code. 743055714Skris [Steve Henson] 743155714Skris 743255714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 743355714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 743455714Skris 743555714Skris *) Make sure `make rehash' target really finds the `openssl' program. 743655714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 743755714Skris 743855714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 743955714Skris like to hear about it if this slows down other processors. 744055714Skris [Ben Laurie] 744155714Skris 744255714Skris *) Add CygWin32 platform information to Configure script. 744355714Skris [Alan Batie <batie@aahz.jf.intel.com>] 744455714Skris 744555714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 744655714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 744755714Skris 744855714Skris *) New program nseq to manipulate netscape certificate sequences 744955714Skris [Steve Henson] 745055714Skris 745155714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 745255714Skris few typos. 745355714Skris [Steve Henson] 745455714Skris 745555714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 745655714Skris but the BN code had some problems that would cause failures when 745755714Skris doing certificate verification and some other functions. 745855714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 745955714Skris 746055714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 746155714Skris [Steve Henson] 746255714Skris 746355714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 746455714Skris [Steve Henson] 746555714Skris 746655714Skris *) Add several PKIX and private extended key usage OIDs. 746755714Skris [Steve Henson] 746855714Skris 746955714Skris *) Modify the 'ca' program to handle the new extension code. Modify 747055714Skris openssl.cnf for new extension format, add comments. 747155714Skris [Steve Henson] 747255714Skris 747355714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 747455714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 747555714Skris CA extensions. 747655714Skris [Steve Henson] 747755714Skris 747855714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 747955714Skris error code, add initial support to X509_print() and x509 application. 748055714Skris [Steve Henson] 748155714Skris 748255714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 748355714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 748455714Skris stuff is currently isolated and isn't even compiled yet. 748555714Skris [Steve Henson] 748655714Skris 748755714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 748855714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 748955714Skris Removed the versions check from X509 routines when loading extensions: 749055714Skris this allows certain broken certificates that don't set the version 749155714Skris properly to be processed. 749255714Skris [Steve Henson] 749355714Skris 749455714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 749555714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 749655714Skris can still be regenerated with "make depend". 749755714Skris [Ben Laurie] 749855714Skris 749955714Skris *) Spelling mistake in C version of CAST-128. 750055714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 750155714Skris 750255714Skris *) Changes to the error generation code. The perl script err-code.pl 750355714Skris now reads in the old error codes and retains the old numbers, only 750455714Skris adding new ones if necessary. It also only changes the .err files if new 750555714Skris codes are added. The makefiles have been modified to only insert errors 750655714Skris when needed (to avoid needlessly modifying header files). This is done 750755714Skris by only inserting errors if the .err file is newer than the auto generated 750855714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 750955714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 751055714Skris or delete all the .err files. 751155714Skris [Steve Henson] 751255714Skris 751355714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 751455714Skris been fixed, but is untested. The assembler versions are also fixed, but 751555714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 751655714Skris to regenerate it if needed. 751755714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 751855714Skris Hagino <itojun@kame.net>] 751955714Skris 752055714Skris *) File was opened incorrectly in randfile.c. 752155714Skris [Ulf M�ller <ulf@fitug.de>] 752255714Skris 752355714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 752455714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 752555714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 752655714Skris al: it's just almost always a UTCTime. Note this patch adds new error 752755714Skris codes so do a "make errors" if there are problems. 752855714Skris [Steve Henson] 752955714Skris 753055714Skris *) Correct Linux 1 recognition in config. 753155714Skris [Ulf M�ller <ulf@fitug.de>] 753255714Skris 753355714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 753455714Skris [Anonymous <nobody@replay.com>] 753555714Skris 753655714Skris *) Generate an error if given an empty string as a cert directory. Also 753755714Skris generate an error if handed NULL (previously returned 0 to indicate an 753855714Skris error, but didn't set one). 753955714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 754055714Skris 754155714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 754255714Skris [Ben Laurie] 754355714Skris 754455714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 754555714Skris parameters. This was causing a warning which killed off the Win32 compile. 754655714Skris [Steve Henson] 754755714Skris 754855714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 754955714Skris [Neil Costigan <neil.costigan@celocom.com>] 755055714Skris 755155714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 755255714Skris based on a text string, looking up short and long names and finally 755355714Skris "dot" format. The "dot" format stuff didn't work. Added new function 755455714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 755555714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 755655714Skris OID is not part of the table. 755755714Skris [Steve Henson] 755855714Skris 755955714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 756055714Skris X509_LOOKUP_by_alias(). 756155714Skris [Ben Laurie] 756255714Skris 756355714Skris *) Sort openssl functions by name. 756455714Skris [Ben Laurie] 756555714Skris 756655714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 756755714Skris encryption from sample DSA keys (in case anyone is interested the password 756855714Skris was "1234"). 756955714Skris [Steve Henson] 757055714Skris 757155714Skris *) Make _all_ *_free functions accept a NULL pointer. 757255714Skris [Frans Heymans <fheymans@isaserver.be>] 757355714Skris 757455714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 757555714Skris NULL pointers. 757655714Skris [Anonymous <nobody@replay.com>] 757755714Skris 757855714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 757955714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 758055714Skris 758155714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 758255714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 758355714Skris 758455714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 758555714Skris [Anonymous <nobody@replay.com>] 758655714Skris 758755714Skris *) Add prototype for temp key callback functions 758855714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 758955714Skris [Ben Laurie] 759055714Skris 759155714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 759255714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 759355714Skris [Steve Henson] 759455714Skris 759555714Skris *) X509_name_add_entry() freed the wrong thing after an error. 759655714Skris [Arne Ansper <arne@ats.cyber.ee>] 759755714Skris 759855714Skris *) rsa_eay.c would attempt to free a NULL context. 759955714Skris [Arne Ansper <arne@ats.cyber.ee>] 760055714Skris 760155714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 760255714Skris [Arne Ansper <arne@ats.cyber.ee>] 760355714Skris 760455714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 760555714Skris [Arne Ansper <arne@ats.cyber.ee>] 760655714Skris 760755714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 760855714Skris in X509_STORE_new(), but document the fact that this variable is still 760955714Skris unused in the certificate verification process. 761055714Skris [Ralf S. Engelschall] 761155714Skris 761255714Skris *) Fix the various library and apps files to free up pkeys obtained from 761355714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 761455714Skris [Steve Henson] 761555714Skris 761655714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 761755714Skris demos/maurice/example2.c work, amongst others, probably. 761855714Skris [Steve Henson and Ben Laurie] 761955714Skris 762055714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 762155714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 762255714Skris are no longer created. This way we have a single and consistent command 762355714Skris line interface `openssl <command>', similar to `cvs <command>'. 762455714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 762555714Skris 762655714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 762755714Skris BIT STRING wrapper always have zero unused bits. 762855714Skris [Steve Henson] 762955714Skris 763055714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 763155714Skris [Steve Henson] 763255714Skris 763355714Skris *) Make the top-level INSTALL documentation easier to understand. 763455714Skris [Paul Sutton] 763555714Skris 763655714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 763755714Skris make (including if user presses ^C) [Paul Sutton] 763855714Skris 763955714Skris *) Make Montgomery context stuff explicit in RSA data structure. 764055714Skris [Ben Laurie] 764155714Skris 764255714Skris *) Fix build order of pem and err to allow for generated pem.h. 764355714Skris [Ben Laurie] 764455714Skris 764555714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 764655714Skris [Ben Laurie] 764755714Skris 764855714Skris *) Enhanced the err-ins.pl script so it makes the error library number 764955714Skris global and can add a library name. This is needed for external ASN1 and 765055714Skris other error libraries. 765155714Skris [Steve Henson] 765255714Skris 765355714Skris *) Fixed sk_insert which never worked properly. 765455714Skris [Steve Henson] 765555714Skris 765655714Skris *) Fix ASN1 macros so they can handle indefinite length construted 765755714Skris EXPLICIT tags. Some non standard certificates use these: they can now 765855714Skris be read in. 765955714Skris [Steve Henson] 766055714Skris 766155714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 766255714Skris into a single doc/ssleay.txt bundle. This way the information is still 766355714Skris preserved but no longer messes up this directory. Now it's new room for 766455714Skris the new set of documenation files. 766555714Skris [Ralf S. Engelschall] 766655714Skris 766755714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 766855714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 766955714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 767055714Skris number of arguments. 767155714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 767255714Skris 767355714Skris *) Fix test data to work with the above. 767455714Skris [Ben Laurie] 767555714Skris 767655714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 767755714Skris was already fixed by Eric for 0.9.1 it seems. 767855714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 767955714Skris 768055714Skris *) Autodetect FreeBSD3. 768155714Skris [Ben Laurie] 768255714Skris 768355714Skris *) Fix various bugs in Configure. This affects the following platforms: 768455714Skris nextstep 768555714Skris ncr-scde 768655714Skris unixware-2.0 768755714Skris unixware-2.0-pentium 768855714Skris sco5-cc. 768955714Skris [Ben Laurie] 769055714Skris 769155714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 769255714Skris before they are needed. 769355714Skris [Ben Laurie] 769455714Skris 769555714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 769655714Skris [Ben Laurie] 769755714Skris 769855714Skris 769955714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 770055714Skris 770155714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 770255714Skris changed SSLeay to OpenSSL in version strings. 770355714Skris [Ralf S. Engelschall] 770455714Skris 770555714Skris *) Some fixups to the top-level documents. 770655714Skris [Paul Sutton] 770755714Skris 770855714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 770955714Skris because the symlink to include/ was missing. 771055714Skris [Ralf S. Engelschall] 771155714Skris 771255714Skris *) Incorporated the popular no-RSA/DSA-only patches 771355714Skris which allow to compile a RSA-free SSLeay. 771455714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 771555714Skris 771655714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 771755714Skris when "ssleay" is still not found. 771855714Skris [Ralf S. Engelschall] 771955714Skris 772055714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 772155714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 772255714Skris 772355714Skris *) Updated the README file. 772455714Skris [Ralf S. Engelschall] 772555714Skris 772655714Skris *) Added various .cvsignore files in the CVS repository subdirs 772755714Skris to make a "cvs update" really silent. 772855714Skris [Ralf S. Engelschall] 772955714Skris 773055714Skris *) Recompiled the error-definition header files and added 773155714Skris missing symbols to the Win32 linker tables. 773255714Skris [Ralf S. Engelschall] 773355714Skris 773455714Skris *) Cleaned up the top-level documents; 773555714Skris o new files: CHANGES and LICENSE 773655714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 773755714Skris o merged COPYRIGHT into LICENSE 773855714Skris o removed obsolete TODO file 773955714Skris o renamed MICROSOFT to INSTALL.W32 774055714Skris [Ralf S. Engelschall] 774155714Skris 774255714Skris *) Removed dummy files from the 0.9.1b source tree: 774355714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 774455714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 774555714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 774655714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 774755714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 774855714Skris [Ralf S. Engelschall] 774955714Skris 775055714Skris *) Added various platform portability fixes. 775155714Skris [Mark J. Cox] 775255714Skris 775355714Skris *) The Genesis of the OpenSSL rpject: 775455714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 775555714Skris Young and Tim J. Hudson created while they were working for C2Net until 775655714Skris summer 1998. 775755714Skris [The OpenSSL Project] 775855714Skris 775955714Skris 776055714Skris Changes between 0.9.0b and 0.9.1b [not released] 776155714Skris 776255714Skris *) Updated a few CA certificates under certs/ 776355714Skris [Eric A. Young] 776455714Skris 776555714Skris *) Changed some BIGNUM api stuff. 776655714Skris [Eric A. Young] 776755714Skris 776855714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 776955714Skris DGUX x86, Linux Alpha, etc. 777055714Skris [Eric A. Young] 777155714Skris 777255714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 777355714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 777455714Skris available). 777555714Skris [Eric A. Young] 777655714Skris 777755714Skris *) Add -strparse option to asn1pars program which parses nested 777855714Skris binary structures 777955714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 778055714Skris 778155714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 778255714Skris [Eric A. Young] 778355714Skris 778455714Skris *) DSA fix for "ca" program. 778555714Skris [Eric A. Young] 778655714Skris 778755714Skris *) Added "-genkey" option to "dsaparam" program. 778855714Skris [Eric A. Young] 778955714Skris 779055714Skris *) Added RIPE MD160 (rmd160) message digest. 779155714Skris [Eric A. Young] 779255714Skris 779355714Skris *) Added -a (all) option to "ssleay version" command. 779455714Skris [Eric A. Young] 779555714Skris 779655714Skris *) Added PLATFORM define which is the id given to Configure. 779755714Skris [Eric A. Young] 779855714Skris 779955714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 780055714Skris [Eric A. Young] 780155714Skris 780255714Skris *) Extended the ASN.1 parser routines. 780355714Skris [Eric A. Young] 780455714Skris 780555714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 780655714Skris [Eric A. Young] 780755714Skris 780855714Skris *) Added a BN_CTX to the BN library. 780955714Skris [Eric A. Young] 781055714Skris 781155714Skris *) Fixed the weak key values in DES library 781255714Skris [Eric A. Young] 781355714Skris 781455714Skris *) Changed API in EVP library for cipher aliases. 781555714Skris [Eric A. Young] 781655714Skris 781755714Skris *) Added support for RC2/64bit cipher. 781855714Skris [Eric A. Young] 781955714Skris 782055714Skris *) Converted the lhash library to the crypto/mem.c functions. 782155714Skris [Eric A. Young] 782255714Skris 782355714Skris *) Added more recognized ASN.1 object ids. 782455714Skris [Eric A. Young] 782555714Skris 782655714Skris *) Added more RSA padding checks for SSL/TLS. 782755714Skris [Eric A. Young] 782855714Skris 782955714Skris *) Added BIO proxy/filter functionality. 783055714Skris [Eric A. Young] 783155714Skris 783255714Skris *) Added extra_certs to SSL_CTX which can be used 783355714Skris send extra CA certificates to the client in the CA cert chain sending 783455714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 783555714Skris [Eric A. Young] 783655714Skris 783755714Skris *) Now Fortezza is denied in the authentication phase because 783855714Skris this is key exchange mechanism is not supported by SSLeay at all. 783955714Skris [Eric A. Young] 784055714Skris 784155714Skris *) Additional PKCS1 checks. 784255714Skris [Eric A. Young] 784355714Skris 784455714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 784555714Skris [Eric A. Young] 784655714Skris 784755714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 784855714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 784955714Skris [Eric A. Young] 785055714Skris 785155714Skris *) Fixed a few memory leaks. 785255714Skris [Eric A. Young] 785355714Skris 785455714Skris *) Fixed various code and comment typos. 785555714Skris [Eric A. Young] 785655714Skris 785755714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 785855714Skris bytes sent in the client random. 785955714Skris [Edward Bishop <ebishop@spyglass.com>] 786055714Skris 7861