CHANGES revision 160814
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 6160814Ssimon 7160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 8160814Ssimon cipher suite and only match that one cipher suite if it is. 9160814Ssimon [Steve Henson] 10160814Ssimon 11160814Ssimon *) Link in manifests for VC++ if needed. 12160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 13160814Ssimon 14160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 15160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 16160814Ssimon TLS extensions, which are supported starting with the 0.9.9 17160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 18160814Ssimon [Douglas Stebila] 19160814Ssimon 20160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 21160814Ssimon opaque EVP_CIPHER_CTX handling. 22160814Ssimon [Steve Henson] 23160814Ssimon 24160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 25160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 26160814Ssimon to conform with the standards mentioned here: 27160814Ssimon http://www.zlib.net/DLL_FAQ.txt 28160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 29160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 30160814Ssimon of the headers and library. Gracefully handle case where zlib library 31160814Ssimon can't be loaded. 32160814Ssimon [Steve Henson] 33160814Ssimon 34160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 35160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 36160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 37160814Ssimon non standard OBJ_obj2txt() behaviour. 38160814Ssimon [Steve Henson] 39160814Ssimon 40160814Ssimon *) Add support for building of engines under engine/ as shared libraries 41160814Ssimon under VC++ build system. 42160814Ssimon [Steve Henson] 43160814Ssimon 44160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 45160814Ssimon Hopefully, we will not see any false combination of paths any more. 46160814Ssimon [Richard Levitte] 47160814Ssimon 48160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 49160814Ssimon 50160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 51160814Ssimon (part of SSL_OP_ALL). This option used to disable the 52160814Ssimon countermeasure against man-in-the-middle protocol-version 53160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 54160814Ssimon idea. (CVE-2005-2969) 55160814Ssimon 56160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 57160814Ssimon for Information Security, National Institute of Advanced Industrial 58160814Ssimon Science and Technology [AIST], Japan)] 59160814Ssimon 60160814Ssimon *) Add two function to clear and return the verify parameter flags. 61160814Ssimon [Steve Henson] 62160814Ssimon 63160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 64160814Ssimon runtime, thus removing the need for a lock. 65160814Ssimon [Nils Larsch] 66160814Ssimon 67160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 68160814Ssimon [Nick Mathewson and Ben Laurie] 69160814Ssimon 70160814Ssimon *) Add functions for well-known primes. 71160814Ssimon [Nick Mathewson] 72160814Ssimon 73160814Ssimon *) Extended Windows CE support. 74160814Ssimon [Satoshi Nakamura and Andy Polyakov] 75160814Ssimon 76160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 77160814Ssimon runtime, thus removing the need for a lock. 78160814Ssimon [Steve Henson] 79160814Ssimon 80160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 81160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 82160814Ssimon smime utility. 83160814Ssimon [Steve Henson] 84160814Ssimon 85160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 86160814Ssimon 87160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 88160814Ssimon [Richard Levitte] 89160814Ssimon 90160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 91160814Ssimon key into the same file any more. 92160814Ssimon [Richard Levitte] 93160814Ssimon 94160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 95160814Ssimon [Andy Polyakov] 96160814Ssimon 97160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 98160814Ssimon [Stefan <stf@udoma.org] 99160814Ssimon 100160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 101160814Ssimon libraries. Use DES_crypt(). 102160814Ssimon [Richard Levitte] 103160814Ssimon 104160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 105160814Ssimon involves renaming the source and generated shared-libs for 106160814Ssimon both. The engines will accept the corrected or legacy ids 107160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 108160814Ssimon this only applies when building 'shared'. 109160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 110160814Ssimon 111160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 112160814Ssimon PKCS12_create() to recognize a CSP name attribute and 113160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 114160814Ssimon [Steve Henson] 115160814Ssimon 116160814Ssimon *) Add new functionality to the bn blinding code: 117160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 118160814Ssimon a fixed number of uses (currently 32) 119160814Ssimon - add new function for parameter creation 120160814Ssimon - introduce flags to control the update behaviour of the 121160814Ssimon BN_BLINDING parameters 122160814Ssimon - hide BN_BLINDING structure 123160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 124160814Ssimon performance when a single RSA object is shared among several 125160814Ssimon threads. 126160814Ssimon [Nils Larsch] 127160814Ssimon 128160814Ssimon *) Add support for DTLS. 129160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 130160814Ssimon 131160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 132160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 133160814Ssimon [Walter Goulet] 134160814Ssimon 135160814Ssimon *) Remove buggy and incompletet DH cert support from 136160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 137160814Ssimon [Nils Larsch] 138160814Ssimon 139160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 140160814Ssimon the apps/openssl applications. 141160814Ssimon [Nils Larsch] 142160814Ssimon 143160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 144160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 145160814Ssimon DEBUG_SAFESTACK must also be set. 146160814Ssimon [Ben Laurie] 147160814Ssimon 148160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 149160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 150160814Ssimon 151160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 152160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 153160814Ssimon 154160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 155160814Ssimon is frequently required for interoperability, and there is no license 156160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 157160814Ssimon avoid this algorithm.) 158160814Ssimon 159160814Ssimon [Bodo Moeller] 160160814Ssimon 161160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 162160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 163160814Ssimon EGEE (Enabling Grids for E-science in Europe). 164160814Ssimon [Richard Levitte] 165160814Ssimon 166160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 167160814Ssimon as Intel P4, IA-64 and AMD64. 168160814Ssimon [Andy Polyakov] 169160814Ssimon 170160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 171160814Ssimon section number in a pod file instead of having to treat each file as 172160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 173160814Ssimon pod file: 174160814Ssimon 175160814Ssimon =for comment openssl_section:XXX 176160814Ssimon 177160814Ssimon The blank line is mandatory. 178160814Ssimon 179160814Ssimon [Steve Henson] 180160814Ssimon 181160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 182160814Ssimon to allow alternative format key and certificate files and passphrase 183160814Ssimon sources. 184160814Ssimon [Steve Henson] 185160814Ssimon 186160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 187160814Ssimon update associated structures and add various utility functions. 188160814Ssimon 189160814Ssimon Add new policy related verify parameters, include policy checking in 190160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 191160814Ssimon to support policy checking and print out. 192160814Ssimon [Steve Henson] 193160814Ssimon 194160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 195160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 196160814Ssimon as well as RNG (though RNG support is currently disabled). 197160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 198160814Ssimon 199160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 200160814Ssimon [Geoff Thorpe] 201160814Ssimon 202160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 203160814Ssimon [Andy Polyakov and a number of other people] 204160814Ssimon 205160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 206160814Ssimon implementation contributed by IBM. 207160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 208160814Ssimon 209160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 210160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 211160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 212160814Ssimon [Jelte Jansen, Geoff Thorpe] 213160814Ssimon 214160814Ssimon *) Functionality for creating the initial serial number file is now 215160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 216160814Ssimon 217160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 218160814Ssimon number file to 1, which is bound to cause problems. To avoid 219160814Ssimon the problems while respecting compatibility between different 0.9.7 220160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 221160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 222160814Ssimon we can fix the problem directly in the 'ca' utility.) 223160814Ssimon [Steve Henson] 224160814Ssimon 225160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 226160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 227160814Ssimon give fewer recursive includes, which could break lazy source code - so 228160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 229160814Ssimon developers should define this symbol when building and using openssl to 230160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 231160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 232160814Ssimon [Geoff Thorpe] 233160814Ssimon 234160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 235160814Ssimon [Steve Henson] 236160814Ssimon 237160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 238160814Ssimon This will generate a random key of the appropriate length based on the 239160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 240160814Ssimon routine to support keys of a specific form. This is used in the des and 241160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 242160814Ssimon code to use new functions and hence generate correct parity DES keys. 243160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 244160814Ssimon valid (weak or incorrect parity). 245160814Ssimon [Steve Henson] 246160814Ssimon 247160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 248160814Ssimon as looking them up. This is useful when the verified structure may contain 249160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 250160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 251160814Ssimon [Steve Henson] 252160814Ssimon 253160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 254160814Ssimon syntax: 255160814Ssimon 256160814Ssimon shortName = some long name, 1.2.3.4 257160814Ssimon [Steve Henson] 258160814Ssimon 259160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 260160814Ssimon limitation on the number of variables it can handle nor the depth of the 261160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 262160814Ssimon information can now expand as required, and rather than having a single 263160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 264160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 265160814Ssimon BN_CTX's "bundling". 266160814Ssimon [Geoff Thorpe] 267160814Ssimon 268160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 269160814Ssimon to allow all RSA operations to function using a single BN_CTX. 270160814Ssimon [Geoff Thorpe] 271160814Ssimon 272160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 273160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 274160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 275160814Ssimon [Steve Henson] 276160814Ssimon 277160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 278160814Ssimon remained unused and not that useful. A variety of other little bignum 279160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 280160814Ssimon below). 281160814Ssimon [Geoff Thorpe] 282160814Ssimon 283160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 284160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 285160814Ssimon [Richard Levitte] 286160814Ssimon 287160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 288160814Ssimon and this should never fail. So the return value from the use of 289160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 290160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 291160814Ssimon [Geoff Thorpe] 292160814Ssimon 293160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 294160814Ssimon initialised value as BN_new(). 295160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 296160814Ssimon 297160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 298160814Ssimon [Steve Henson] 299160814Ssimon 300160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 301160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 302160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 303160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 304160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 305160814Ssimon structures to try and expose faulty code further on. For now, openssl will 306160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 307160814Ssimon forms that it has tolerated in the past, but authors and packagers should 308160814Ssimon consider trying openssl and their own applications when compiled with 309160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 310160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 311160814Ssimon some point, these tighter rules will become openssl's default to improve 312160814Ssimon maintainability, though the assert()s and other overheads will remain only 313160814Ssimon in debugging configurations. See bn.h for more details. 314160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 315160814Ssimon 316160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 317160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 318160814Ssimon initialises it). The presence of this function only made it possible 319160814Ssimon to overwrite an existing structure (and cause memory leaks). 320160814Ssimon [Geoff Thorpe] 321160814Ssimon 322160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 323160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 324160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 325160814Ssimon to clean up those corresponding objects before destroying the hash table 326160814Ssimon (and losing the object pointers). So some over-zealous constifications in 327160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 328160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 329160814Ssimon prototyped to have "const" restrictions on the object pointers they are 330160814Ssimon given (and so aren't required to cast them away any more). 331160814Ssimon [Geoff Thorpe] 332160814Ssimon 333160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 334160814Ssimon (speed) prefers to use its own implementation. The two implementations 335160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 336160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 337160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 338160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 339160814Ssimon internally to the implementation so I've used that for now. 340160814Ssimon [Geoff Thorpe] 341160814Ssimon 342160814Ssimon *) Ensure that deprecated functions do not get compiled when 343160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 344160814Ssimon the self-tests were still using deprecated key-generation functions so 345160814Ssimon these have been updated also. 346160814Ssimon [Geoff Thorpe] 347160814Ssimon 348160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 349160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 350160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 351160814Ssimon digestedData type. Add additional code to correctly generate the 352160814Ssimon digestedData type and add support for this type in PKCS7 initialization 353160814Ssimon functions. 354160814Ssimon [Steve Henson] 355160814Ssimon 356160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 357160814Ssimon structure of type "other". 358160814Ssimon [Steve Henson] 359160814Ssimon 360160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 361160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 362160814Ssimon modulus operations are not performed. The (pre-generated) prime 363160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 364160814Ssimon re-generated on some platforms because of the "division by zero" 365160814Ssimon situation in the script. 366160814Ssimon [Ralf S. Engelschall] 367160814Ssimon 368160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 369160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 370160814Ssimon SHA-1 now is only used for "small" curves (where the 371160814Ssimon representation of a field element takes up to 24 bytes); for 372160814Ssimon larger curves, the field element resulting from ECDH is directly 373160814Ssimon used as premaster secret. 374160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 375160814Ssimon 376160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 377160814Ssimon curve secp160r1 to the tests. 378160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 379160814Ssimon 380160814Ssimon *) Add the possibility to load symbols globally with DSO. 381160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 382160814Ssimon 383160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 384160814Ssimon control of the error stack. 385160814Ssimon [Richard Levitte] 386160814Ssimon 387160814Ssimon *) Add support for STORE in ENGINE. 388160814Ssimon [Richard Levitte] 389160814Ssimon 390160814Ssimon *) Add the STORE type. The intention is to provide a common interface 391160814Ssimon to certificate and key stores, be they simple file-based stores, or 392160814Ssimon HSM-type store, or LDAP stores, or... 393160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 394160814Ssimon [Richard Levitte] 395160814Ssimon 396160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 397160814Ssimon pass a list of arguments to any function as well as provide a way 398160814Ssimon for a function to pass data back to the caller. 399160814Ssimon [Richard Levitte] 400160814Ssimon 401160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 402160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 403160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 404160814Ssimon a memory area. 405160814Ssimon [Richard Levitte] 406160814Ssimon 407160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 408160814Ssimon return an index to an element even if an exact match couldn't be 409160814Ssimon found. The index is guaranteed to point at the element where the 410160814Ssimon searched-for key would be inserted to preserve sorting order. 411160814Ssimon [Richard Levitte] 412160814Ssimon 413160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 414160814Ssimon takes an extra flags argument for optional functionality. Currently, 415160814Ssimon the following flags are defined: 416160814Ssimon 417160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 418160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 419160814Ssimon element where the comparing function returns a negative or zero 420160814Ssimon number. 421160814Ssimon 422160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 423160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 424160814Ssimon element where the comparing function returns zero. This is useful 425160814Ssimon if there are more than one element where the comparing function 426160814Ssimon returns zero. 427160814Ssimon [Richard Levitte] 428160814Ssimon 429160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 430160814Ssimon in such a way that the self-signed certificate becomes part of the 431160814Ssimon CA database and uses the same mechanisms for serial number generation 432160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 433160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 434160814Ssimon [Richard Levitte] 435160814Ssimon 436160814Ssimon *) Add functionality to check the public key of a certificate request 437160814Ssimon against a given private. This is useful to check that a certificate 438160814Ssimon request can be signed by that key (self-signing). 439160814Ssimon [Richard Levitte] 440160814Ssimon 441160814Ssimon *) Make it possible to have multiple active certificates with the same 442160814Ssimon subject in the CA index file. This is done only if the keyword 443160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 444160814Ssimon if 'CA_default') of the configuration file. The value is saved 445160814Ssimon with the database itself in a separate index attribute file, 446160814Ssimon named like the index file with '.attr' appended to the name. 447160814Ssimon [Richard Levitte] 448160814Ssimon 449160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 450160814Ssimon req and dirName. 451160814Ssimon [Steve Henson] 452160814Ssimon 453160814Ssimon *) Support for nameConstraints certificate extension. 454160814Ssimon [Steve Henson] 455160814Ssimon 456160814Ssimon *) Support for policyConstraints certificate extension. 457160814Ssimon [Steve Henson] 458160814Ssimon 459160814Ssimon *) Support for policyMappings certificate extension. 460160814Ssimon [Steve Henson] 461160814Ssimon 462160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 463160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 464160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 465160814Ssimon indirection. This lets alternative implementations fallback to the 466160814Ssimon default implementation more easily. 467160814Ssimon [Geoff Thorpe] 468160814Ssimon 469160814Ssimon *) Support for directoryName in GeneralName related extensions 470160814Ssimon in config files. 471160814Ssimon [Steve Henson] 472160814Ssimon 473160814Ssimon *) Make it possible to link applications using Makefile.shared. 474160814Ssimon Make that possible even when linking against static libraries! 475160814Ssimon [Richard Levitte] 476160814Ssimon 477160814Ssimon *) Support for single pass processing for S/MIME signing. This now 478160814Ssimon means that S/MIME signing can be done from a pipe, in addition 479160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 480160814Ssimon and the signed data does not need to be all held in memory. 481160814Ssimon 482160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 483160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 484160814Ssimon is done after the data is output (and digests calculated) in 485160814Ssimon SMIME_write_PKCS7(). 486160814Ssimon [Steve Henson] 487160814Ssimon 488160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 489160814Ssimon applications, at least on the platforms where it's known how 490160814Ssimon to do it. 491160814Ssimon [Richard Levitte] 492160814Ssimon 493160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 494160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 495160814Ssimon will now compute a table of multiples of the generator that 496160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 497160814Ssimon faster (notably in the case of a single point multiplication, 498160814Ssimon scalar * generator). 499160814Ssimon [Nils Larsch, Bodo Moeller] 500160814Ssimon 501160814Ssimon *) IPv6 support for certificate extensions. The various extensions 502160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 503160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 504160814Ssimon correctly. 505160814Ssimon [Steve Henson] 506160814Ssimon 507160814Ssimon *) Added an ENGINE that implements RSA by performing private key 508160814Ssimon exponentiations with the GMP library. The conversions to and from 509160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 510160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 511160814Ssimon However there are likely to be other architectures where GMP could 512160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 513160814Ssimon specified at Configure time and should be accompanied by the necessary 514160814Ssimon linker additions, eg; 515160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 516160814Ssimon [Geoff Thorpe] 517160814Ssimon 518160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 519160814Ssimon testing availability of engines with "-t" - the old behaviour is 520160814Ssimon produced by increasing the feature's verbosity with "-tt". 521160814Ssimon [Geoff Thorpe] 522160814Ssimon 523160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 524160814Ssimon could be freed. Solution: make sure initialization is performed early 525160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 526160814Ssimon via PR#459) 527160814Ssimon [Lutz Jaenicke] 528160814Ssimon 529160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 530160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 531160814Ssimon software implementations. For DSA and DH, parameter generation can 532160814Ssimon also be overriden by providing the appropriate method callbacks. 533160814Ssimon [Geoff Thorpe] 534160814Ssimon 535160814Ssimon *) Change the "progress" mechanism used in key-generation and 536160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 537160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 538160814Ssimon postfixes and the older functions are reimplemented as wrappers for 539160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 540160814Ssimon declarations of the old functions to help (graceful) attempts to 541160814Ssimon migrate to the new functions. Also, the new key-generation API 542160814Ssimon functions operate on a caller-supplied key-structure and return 543160814Ssimon success/failure rather than returning a key or NULL - this is to 544160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 545160814Ssimon 546160814Ssimon Example for using the new callback interface: 547160814Ssimon 548160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 549160814Ssimon void *my_arg = ...; 550160814Ssimon BN_GENCB my_cb; 551160814Ssimon 552160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 553160814Ssimon 554160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 555160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 556160814Ssimon * documentation of the function that calls the callback. 557160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 558160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 559160814Ssimon * to continue, or 0 to stop. 560160814Ssimon */ 561160814Ssimon 562160814Ssimon [Geoff Thorpe] 563160814Ssimon 564160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 565160814Ssimon available to TLS with the number defined in 566160814Ssimon draft-ietf-tls-compression-04.txt. 567160814Ssimon [Richard Levitte] 568160814Ssimon 569160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 570160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 571160814Ssimon 572160814Ssimon CertificatePair ::= SEQUENCE { 573160814Ssimon forward [0] Certificate OPTIONAL, 574160814Ssimon reverse [1] Certificate OPTIONAL, 575160814Ssimon -- at least one of the pair shall be present -- } 576160814Ssimon 577160814Ssimon Also implement the PEM functions to read and write certificate 578160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 579160814Ssimon 580160814Ssimon This needed to be defined, mostly for the sake of the LDAP 581160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 582160814Ssimon well. 583160814Ssimon [Richard Levitte] 584160814Ssimon 585160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 586160814Ssimon Makefile.shared, for Cygwin's sake. 587160814Ssimon [Richard Levitte] 588160814Ssimon 589160814Ssimon *) Extend the BIGNUM API by creating a function 590160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 591160814Ssimon and a macro that behave like 592160814Ssimon int BN_is_negative(const BIGNUM *a); 593160814Ssimon 594160814Ssimon to avoid the need to access 'a->neg' directly in applications. 595160814Ssimon [Nils Larsch] 596160814Ssimon 597160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 598160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 599160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 600160814Ssimon if applicable. 601160814Ssimon [Nils Larsch <nla@trustcenter.de>] 602160814Ssimon 603160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 604160814Ssimon [Bodo Moeller] 605160814Ssimon 606160814Ssimon *) Change the ENGINE framework to automatically load engines 607160814Ssimon dynamically from specific directories unless they could be 608160814Ssimon found to already be built in or loaded. Move all the 609160814Ssimon current engines except for the cryptodev one to a new 610160814Ssimon directory engines/. 611160814Ssimon The engines in engines/ are built as shared libraries if 612160814Ssimon the "shared" options was given to ./Configure or ./config. 613160814Ssimon Otherwise, they are inserted in libcrypto.a. 614160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 615160814Ssimon engines, but that can be overriden at configure time through 616160814Ssimon the usual use of --prefix and/or --openssldir, and at run 617160814Ssimon time with the environment variable OPENSSL_ENGINES. 618160814Ssimon [Geoff Thorpe and Richard Levitte] 619160814Ssimon 620160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 621160814Ssimon libraries. Addapt Makefile.org. 622160814Ssimon [Richard Levitte] 623160814Ssimon 624160814Ssimon *) Add version info to Win32 DLLs. 625160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 626160814Ssimon 627160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 628160814Ssimon can be added using this API to created arbitrary PKCS#12 629160814Ssimon files while avoiding the low level API. 630160814Ssimon 631160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 632160814Ssimon will then be omitted from the output file. The encryption 633160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 634160814Ssimon iteration count can be set to 0 to omit the mac. 635160814Ssimon 636160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 637160814Ssimon options work when creating a PKCS#12 file. New option -nomac 638160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 639160814Ssimon New code is modified to use the enhanced PKCS12_create() 640160814Ssimon instead of the low level API. 641160814Ssimon [Steve Henson] 642160814Ssimon 643160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 644160814Ssimon encoding. This can output sequences tags and octet strings in 645160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 646160814Ssimon encoding. This is experimental and needs additional code to 647160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 648160814Ssimon PKCS#7 code. 649160814Ssimon 650160814Ssimon Extend template encode functionality so that tagging is passed 651160814Ssimon down to the template encoder. 652160814Ssimon [Steve Henson] 653160814Ssimon 654160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 655160814Ssimon recognized instead of using RSA as a default. 656160814Ssimon [Bodo Moeller] 657160814Ssimon 658160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 659160814Ssimon As these are not official, they are not included in "ALL"; 660160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 661160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 662160814Ssimon 663160814Ssimon *) Add ECDH engine support. 664160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 665160814Ssimon 666160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 667160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 668160814Ssimon 669160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 670160814Ssimon without success (which indicates a broken PRNG). 671160814Ssimon [Bodo Moeller] 672160814Ssimon 673160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 674160814Ssimon is really the square of the return value. (Previously, 675160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 676160814Ssimon [Bodo Moeller] 677160814Ssimon 678160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 679160814Ssimon and WAP/WTLS; add OIDs that were still missing. 680160814Ssimon 681160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 682160814Ssimon (Sun Microsystems Laboratories)] 683160814Ssimon 684160814Ssimon *) Extend the EC library for elliptic curves over binary fields 685160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 686160814Ssimon New EC_METHOD: 687160814Ssimon 688160814Ssimon EC_GF2m_simple_method 689160814Ssimon 690160814Ssimon New API functions: 691160814Ssimon 692160814Ssimon EC_GROUP_new_curve_GF2m 693160814Ssimon EC_GROUP_set_curve_GF2m 694160814Ssimon EC_GROUP_get_curve_GF2m 695160814Ssimon EC_POINT_set_affine_coordinates_GF2m 696160814Ssimon EC_POINT_get_affine_coordinates_GF2m 697160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 698160814Ssimon 699160814Ssimon Point compression for binary fields is disabled by default for 700160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 701160814Ssimon enable it). 702160814Ssimon 703160814Ssimon As binary polynomials are represented as BIGNUMs, various members 704160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 705160814Ssimon between the implementations for prime fields and binary fields; 706160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 707160814Ssimon are essentially identical to their ..._GFp counterparts. 708160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 709160814Ssimon various internal method names.) 710160814Ssimon 711160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 712160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 713160814Ssimon 714160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 715160814Ssimon (Sun Microsystems Laboratories)] 716160814Ssimon 717160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 718160814Ssimon through methods ('mul', 'precompute_mult'). 719160814Ssimon 720160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 721160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 722160814Ssimon methods are undefined. 723160814Ssimon 724160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 725160814Ssimon (Sun Microsystems Laboratories)] 726160814Ssimon 727160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 728160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 729160814Ssimon length of the modulus. 730160814Ssimon 731160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 732160814Ssimon (Sun Microsystems Laboratories)] 733160814Ssimon 734160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 735160814Ssimon (These simply call ..._new and ..._copy). 736160814Ssimon 737160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 738160814Ssimon (Sun Microsystems Laboratories)] 739160814Ssimon 740160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 741160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 742160814Ssimon used) in the following functions [macros]: 743160814Ssimon 744160814Ssimon BN_GF2m_add 745160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 746160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 747160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 748160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 749160814Ssimon BN_GF2m_mod_inv 750160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 751160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 752160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 753160814Ssimon BN_GF2m_cmp [= BN_ucmp] 754160814Ssimon 755160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 756160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 757160814Ssimon 758160814Ssimon For some functions, an the irreducible polynomial defining a 759160814Ssimon field can be given as an 'unsigned int[]' with strictly 760160814Ssimon decreasing elements giving the indices of those bits that are set; 761160814Ssimon i.e., p[] represents the polynomial 762160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 763160814Ssimon where 764160814Ssimon p[0] > p[1] > ... > p[k] = 0. 765160814Ssimon This applies to the following functions: 766160814Ssimon 767160814Ssimon BN_GF2m_mod_arr 768160814Ssimon BN_GF2m_mod_mul_arr 769160814Ssimon BN_GF2m_mod_sqr_arr 770160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 771160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 772160814Ssimon BN_GF2m_mod_exp_arr 773160814Ssimon BN_GF2m_mod_sqrt_arr 774160814Ssimon BN_GF2m_mod_solve_quad_arr 775160814Ssimon BN_GF2m_poly2arr 776160814Ssimon BN_GF2m_arr2poly 777160814Ssimon 778160814Ssimon Conversion can be performed by the following functions: 779160814Ssimon 780160814Ssimon BN_GF2m_poly2arr 781160814Ssimon BN_GF2m_arr2poly 782160814Ssimon 783160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 784160814Ssimon 785160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 786160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 787160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 788160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 789160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 790160814Ssimon 791160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 792160814Ssimon (Sun Microsystems Laboratories)] 793160814Ssimon 794160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 795160814Ssimon functionality is disabled at compile-time. 796160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 797160814Ssimon 798160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 799160814Ssimon information is visible when viewing, e.g., a certificate: 800160814Ssimon 801160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 802160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 803160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 804160814Ssimon avoid the appearance of a printable string. 805160814Ssimon [Nils Larsch <nla@trustcenter.de>] 806160814Ssimon 807160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 808160814Ssimon functions 809160814Ssimon EC_GROUP_set_asn1_flag() 810160814Ssimon EC_GROUP_get_asn1_flag() 811160814Ssimon EC_GROUP_set_point_conversion_form() 812160814Ssimon EC_GROUP_get_point_conversion_form() 813160814Ssimon These control ASN1 encoding details: 814160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 815160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 816160814Ssimon - Points are encoded in uncompressed form by default; options for 817160814Ssimon asn1_for are as for point2oct, namely 818160814Ssimon POINT_CONVERSION_COMPRESSED 819160814Ssimon POINT_CONVERSION_UNCOMPRESSED 820160814Ssimon POINT_CONVERSION_HYBRID 821160814Ssimon 822160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 823160814Ssimon functions 824160814Ssimon EC_GROUP_set_seed() 825160814Ssimon EC_GROUP_get0_seed() 826160814Ssimon EC_GROUP_get_seed_len() 827160814Ssimon This is used only for ASN1 purposes (so far). 828160814Ssimon [Nils Larsch <nla@trustcenter.de>] 829160814Ssimon 830160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 831160814Ssimon of the appropriate field type OID. The new function 832160814Ssimon EC_METHOD_get_field_type() returns this value. 833160814Ssimon [Nils Larsch <nla@trustcenter.de>] 834160814Ssimon 835160814Ssimon *) Add functions 836160814Ssimon EC_POINT_point2bn() 837160814Ssimon EC_POINT_bn2point() 838160814Ssimon EC_POINT_point2hex() 839160814Ssimon EC_POINT_hex2point() 840160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 841160814Ssimon EC_POINT_oct2point(). 842160814Ssimon [Nils Larsch <nla@trustcenter.de>] 843160814Ssimon 844160814Ssimon *) Change internals of the EC library so that the functions 845160814Ssimon EC_GROUP_set_generator() 846160814Ssimon EC_GROUP_get_generator() 847160814Ssimon EC_GROUP_get_order() 848160814Ssimon EC_GROUP_get_cofactor() 849160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 850160814Ssimon to methods, which would lead to unnecessary code duplication when 851160814Ssimon adding different types of curves. 852160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 853160814Ssimon 854160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 855160814Ssimon arithmetic, and such that modified wNAFs are generated 856160814Ssimon (which avoid length expansion in many cases). 857160814Ssimon [Bodo Moeller] 858160814Ssimon 859160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 860160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 861160814Ssimon 862160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 863160814Ssimon on a EC_GROUP, its generator and order. This includes 864160814Ssimon EC_GROUP_check_discriminant(). 865160814Ssimon [Nils Larsch <nla@trustcenter.de>] 866160814Ssimon 867160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 868160814Ssimon 869160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 870160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 871160814Ssimon 872160814Ssimon ECDSA support is also included in various other files across the 873160814Ssimon library. Most notably, 874160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 875160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 876160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 877160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 878160814Ssimon them suitable for ECDSA where domain parameters must be 879160814Ssimon extracted before the specific public key; 880160814Ssimon - ECDSA engine support has been added. 881160814Ssimon [Nils Larsch <nla@trustcenter.de>] 882160814Ssimon 883160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 884160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 885160814Ssimon function 886160814Ssimon EC_GROUP_new_by_curve_name(), 887160814Ssimon and the list of available named curves can be obtained with 888160814Ssimon EC_get_builtin_curves(). 889160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 890160814Ssimon accessed via 891160814Ssimon EC_GROUP_set_curve_name() 892160814Ssimon EC_GROUP_get_curve_name() 893160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 894160814Ssimon 895160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 896160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 897160814Ssimon required a small change in bn_mul_part_recursive() and the addition 898160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 899160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 900160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 901160814Ssimon differing sizes. 902160814Ssimon [Richard Levitte] 903160814Ssimon 904160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 905160814Ssimon 906160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 907160814Ssimon The value now differs depending on if you build for FIPS or not. 908160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 909160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 910160814Ssimon the difference induced by this change. 911160814Ssimon [Andy Polyakov] 912160814Ssimon 913160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 914160814Ssimon 915160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 916160814Ssimon (part of SSL_OP_ALL). This option used to disable the 917160814Ssimon countermeasure against man-in-the-middle protocol-version 918160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 919160814Ssimon idea. (CVE-2005-2969) 920160814Ssimon 921160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 922160814Ssimon for Information Security, National Institute of Advanced Industrial 923160814Ssimon Science and Technology [AIST], Japan)] 924160814Ssimon 925160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 926160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 927160814Ssimon [Steve Henson] 928160814Ssimon 929160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 930160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 931160814Ssimon the information leaked through timing could expose the secret key 932160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 933160814Ssimon biased k.) 934160814Ssimon [Bodo Moeller] 935160814Ssimon 936160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 937160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 938160814Ssimon squares and multiplies and the memory access pattern are 939160814Ssimon independent of the particular secret key. This will mitigate 940160814Ssimon cache-timing and potential related attacks. 941160814Ssimon 942160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 943160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 944160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 945160814Ssimon will use this BN flag for private exponents unless the flag 946160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 947160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 948160814Ssimon 949160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 950160814Ssimon 951160814Ssimon *) Change the client implementation for SSLv23_method() and 952160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 953160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 954160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 955160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 956160814Ssimon [Bodo Moeller] 957160814Ssimon 958160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 959160814Ssimon clients need. 960160814Ssimon [Steve Henson] 961160814Ssimon 962160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 963160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 964160814Ssimon to dsa and dh code (which had race conditions before). 965160814Ssimon [Steve Henson] 966160814Ssimon 967160814Ssimon *) Include the fixed error library code in the C error file definitions 968160814Ssimon instead of fixing them up at runtime. This keeps the error code 969160814Ssimon structures constant. 970160814Ssimon [Steve Henson] 971160814Ssimon 972160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 973160814Ssimon 974160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 975160814Ssimon OpenSSL 0.9.8.] 976160814Ssimon 977160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 978160814Ssimon the 'length' field is signed on one version and unsigned on another 979160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 980160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 981160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 982160814Ssimon some needed definitions. 983160814Ssimon [Steve Henson] 984160814Ssimon 985160814Ssimon *) Undo Cygwin change. 986160814Ssimon [Ulf M�ller] 987160814Ssimon 988160814Ssimon *) Added support for proxy certificates according to RFC 3820. 989160814Ssimon Because they may be a security thread to unaware applications, 990160814Ssimon they must be explicitely allowed in run-time. See 991160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 992160814Ssimon [Richard Levitte] 993160814Ssimon 994160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 995160814Ssimon 996160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 997160814Ssimon server and client random values. Previously 998160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 999160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 1000160814Ssimon 1001160814Ssimon This change has negligible security impact because: 1002160814Ssimon 1003160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 1004160814Ssimon data. 1005160814Ssimon 1006160814Ssimon 2. Server and client random values are sent in the clear in the initial 1007160814Ssimon handshake. 1008160814Ssimon 1009160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 1010160814Ssimon size for static RSA ciphersuites) as well as client server and random 1011160814Ssimon values. 1012160814Ssimon 1013160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 1014160814Ssimon to our attention. 1015160814Ssimon 1016160814Ssimon [Stephen Henson, reported by UK NISCC] 1017160814Ssimon 1018160814Ssimon *) Use Windows randomness collection on Cygwin. 1019160814Ssimon [Ulf M�ller] 1020160814Ssimon 1021160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 1022160814Ssimon prematurely by EGD/PRNGD. 1023160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 1024160814Ssimon 1025160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 1026160814Ssimon [Steve Henson] 1027160814Ssimon 1028160814Ssimon *) Back-port of selected performance improvements from development 1029160814Ssimon branch, as well as improved support for PowerPC platforms. 1030160814Ssimon [Andy Polyakov] 1031160814Ssimon 1032160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 1033160814Ssimon failure and freeing up memory if a failure occurs. 1034160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 1035160814Ssimon 1036160814Ssimon *) Add new -passin argument to dgst. 1037160814Ssimon [Steve Henson] 1038160814Ssimon 1039160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 1040160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 1041160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 1042160814Ssimon certificates. 1043160814Ssimon [Steve Henson] 1044160814Ssimon 1045160814Ssimon *) Make an explicit check during certificate validation to see that 1046160814Ssimon the CA setting in each certificate on the chain is correct. As a 1047160814Ssimon side effect always do the following basic checks on extensions, 1048160814Ssimon not just when there's an associated purpose to the check: 1049160814Ssimon 1050160814Ssimon - if there is an unhandled critical extension (unless the user 1051160814Ssimon has chosen to ignore this fault) 1052160814Ssimon - if the path length has been exceeded (if one is set at all) 1053160814Ssimon - that certain extensions fit the associated purpose (if one has 1054160814Ssimon been given) 1055160814Ssimon [Richard Levitte] 1056160814Ssimon 1057142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 1058142425Snectar 1059142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 1060142425Snectar environment. This would happen due to the reordering of the revoked 1061142425Snectar entries during signature checking and serial number lookup. Now the 1062142425Snectar encoding is cached and the serial number sort performed under a lock. 1063142425Snectar Add new STACK function sk_is_sorted(). 1064142425Snectar [Steve Henson] 1065142425Snectar 1066142425Snectar *) Add Delta CRL to the extension code. 1067142425Snectar [Steve Henson] 1068142425Snectar 1069142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 1070142425Snectar [David Holmes <d.holmes@f5.com>] 1071142425Snectar 1072142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 1073142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 1074142425Snectar This is done by creating a random 64 bit value for the initial serial 1075142425Snectar number when a serial number file is created or when a self signed 1076142425Snectar certificate is created using 'openssl req -x509'. The initial serial 1077142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 1078142425Snectar rather than being initialized to 1. 1079142425Snectar [Steve Henson] 1080142425Snectar 1081127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 1082127128Snectar 1083127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 1084160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 1085127128Snectar [Joe Orton, Steve Henson] 1086127128Snectar 1087127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 1088160814Ssimon (CVE-2004-0112) 1089127128Snectar [Joe Orton, Steve Henson] 1090127128Snectar 1091127128Snectar *) Make it possible to have multiple active certificates with the same 1092127128Snectar subject in the CA index file. This is done only if the keyword 1093127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 1094127128Snectar if 'CA_default') of the configuration file. The value is saved 1095127128Snectar with the database itself in a separate index attribute file, 1096127128Snectar named like the index file with '.attr' appended to the name. 1097127128Snectar [Richard Levitte] 1098127128Snectar 1099127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 1100127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 1101127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 1102127128Snectar extensions: since verify currently doesn't process CRL extensions this 1103127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 1104127128Snectar for these cases. 1105127128Snectar [Steve Henson] 1106127128Snectar 1107127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 1108127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 1109127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 1110127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 1111127128Snectar parsing them this should not create any compatibility issues. 1112127128Snectar [Steve Henson] 1113127128Snectar 1114127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 1115127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 1116127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 1117127128Snectar < 0.9.7. 1118127128Snectar [Steve Henson] 1119127128Snectar 1120127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 1121127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 1122127128Snectar 1123127128Snectar *) Use the correct content when signing type "other". 1124127128Snectar [Steve Henson] 1125127128Snectar 1126120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 1127120631Snectar 1128120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 1129120631Snectar 1130120631Snectar Stop out of bounds reads in the ASN1 code when presented with 1131160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 1132120631Snectar 1133160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 1134120631Snectar 1135120631Snectar If verify callback ignores invalid public key errors don't try to check 1136120631Snectar certificate signature with the NULL public key. 1137120631Snectar 1138120631Snectar [Steve Henson] 1139120631Snectar 1140120631Snectar *) New -ignore_err option in ocsp application to stop the server 1141120631Snectar exiting on the first error in a request. 1142120631Snectar [Steve Henson] 1143120631Snectar 1144120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 1145120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 1146120631Snectar specifications. 1147120631Snectar [Steve Henson] 1148120631Snectar 1149120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 1150120631Snectar extra data after the compression methods not only for TLS 1.0 1151120631Snectar but also for SSL 3.0 (as required by the specification). 1152120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 1153120631Snectar 1154120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 1155120631Snectar when it's 512 *bits* long, not 512 bytes. 1156120631Snectar [Richard Levitte] 1157120631Snectar 1158120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 1159120631Snectar blocks during encryption. 1160120631Snectar [Richard Levitte] 1161120631Snectar 1162120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 1163120631Snectar flushes were not handled properly if the BIO retried. On read 1164120631Snectar data was not being buffered properly and had various logic bugs. 1165120631Snectar This also affects blocking I/O when the data being decoded is a 1166120631Snectar certain size. 1167120631Snectar [Steve Henson] 1168120631Snectar 1169120631Snectar *) Various S/MIME bugfixes and compatibility changes: 1170120631Snectar output correct application/pkcs7 MIME type if 1171120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 1172120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 1173120631Snectar of files as .eml work). Correctly handle very long lines in MIME 1174120631Snectar parser. 1175120631Snectar [Steve Henson] 1176120631Snectar 1177120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 1178120631Snectar 1179120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 1180120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 1181120631Snectar a protocol version number mismatch like a decryption error 1182120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 1183120631Snectar [Bodo Moeller] 1184120631Snectar 1185120631Snectar *) Turn on RSA blinding by default in the default implementation 1186120631Snectar to avoid a timing attack. Applications that don't want it can call 1187120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 1188120631Snectar They would be ill-advised to do so in most cases. 1189120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 1190120631Snectar 1191120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 1192120631Snectar seeded (in this case, the secret RSA exponent is abused as 1193120631Snectar an unpredictable seed -- if it is not unpredictable, there 1194120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 1195120631Snectar by remembering the creator's thread ID in rsa->blinding and 1196120631Snectar having all other threads use local one-time blinding factors 1197120631Snectar (this requires more computation than sharing rsa->blinding, but 1198120631Snectar avoids excessive locking; and if an RSA object is not shared 1199120631Snectar between threads, blinding will still be very fast). 1200120631Snectar [Bodo Moeller] 1201120631Snectar 1202120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 1203120631Snectar ENGINE as defaults for all supported algorithms irrespective of 1204120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 1205120631Snectar should make sure they are passing it correctly. 1206120631Snectar [Geoff Thorpe] 1207120631Snectar 1208120631Snectar *) Target "mingw" now allows native Windows code to be generated in 1209120631Snectar the Cygwin environment as well as with the MinGW compiler. 1210120631Snectar [Ulf Moeller] 1211120631Snectar 1212111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 1213111147Snectar 1214111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 1215111147Snectar via timing by performing a MAC computation even if incorrrect 1216111147Snectar block cipher padding has been found. This is a countermeasure 1217111147Snectar against active attacks where the attacker has to distinguish 1218160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 1219111147Snectar 1220111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 1221111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 1222111147Snectar Martin Vuagnoux (EPFL, Ilion)] 1223111147Snectar 1224111147Snectar *) Make the no-err option work as intended. The intention with no-err 1225111147Snectar is not to have the whole error stack handling routines removed from 1226111147Snectar libcrypto, it's only intended to remove all the function name and 1227111147Snectar reason texts, thereby removing some of the footprint that may not 1228111147Snectar be interesting if those errors aren't displayed anyway. 1229111147Snectar 1230111147Snectar NOTE: it's still possible for any application or module to have it's 1231111147Snectar own set of error texts inserted. The routines are there, just not 1232111147Snectar used by default when no-err is given. 1233111147Snectar [Richard Levitte] 1234111147Snectar 1235111147Snectar *) Add support for FreeBSD on IA64. 1236111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 1237111147Snectar 1238111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 1239111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 1240111147Snectar the value returned by DES_cbc_cksum() was like the one from 1241111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 1242111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 1243111147Snectar 1244111147Snectar *) Allow an application to disable the automatic SSL chain building. 1245111147Snectar Before this a rather primitive chain build was always performed in 1246111147Snectar ssl3_output_cert_chain(): an application had no way to send the 1247111147Snectar correct chain if the automatic operation produced an incorrect result. 1248111147Snectar 1249111147Snectar Now the chain builder is disabled if either: 1250111147Snectar 1251111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 1252111147Snectar 1253111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 1254111147Snectar 1255111147Snectar The reasoning behind this is that an application would not want the 1256111147Snectar auto chain building to take place if extra chain certificates are 1257111147Snectar present and it might also want a means of sending no additional 1258111147Snectar certificates (for example the chain has two certificates and the 1259111147Snectar root is omitted). 1260111147Snectar [Steve Henson] 1261111147Snectar 1262111147Snectar *) Add the possibility to build without the ENGINE framework. 1263111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 1264111147Snectar 1265111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 1266111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 1267111147Snectar [Steve Henson] 1268111147Snectar 1269111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 1270111147Snectar could be freed. Solution: make sure initialization is performed early 1271111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 1272111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 1273111147Snectar [Lutz Jaenicke] 1274111147Snectar 1275111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 1276111147Snectar checked on reconnect on the client side, therefore session resumption 1277111147Snectar could still fail with a "ssl session id is different" error. This 1278111147Snectar behaviour is masked when SSL_OP_ALL is used due to 1279111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 1280111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 1281111147Snectar followup to PR #377. 1282111147Snectar [Lutz Jaenicke] 1283111147Snectar 1284111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 1285111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 1286111147Snectar [Andy Polyakov] 1287111147Snectar 1288111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 1289111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 1290111147Snectar the config script, much like the NetBSD support. 1291111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 1292111147Snectar 1293109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 1294109998Smarkm 1295120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 1296120631Snectar OpenSSL 0.9.7.] 1297120631Snectar 1298109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 1299109998Smarkm code (06) was taken as the first octet of the session ID and the last 1300109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 1301109998Smarkm caching could not have worked due to the session ID mismatch between 1302109998Smarkm client and server. 1303109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 1304109998Smarkm PR #377. 1305109998Smarkm [Lutz Jaenicke] 1306109998Smarkm 1307109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 1308109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 1309109998Smarkm removed entirely. 1310109998Smarkm [Richard Levitte] 1311109998Smarkm 1312109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 1313109998Smarkm seems that in spite of existing for more than a year, many application 1314109998Smarkm author have done nothing to provide the necessary callbacks, which 1315109998Smarkm means that this particular engine will not work properly anywhere. 1316109998Smarkm This is a very unfortunate situation which forces us, in the name 1317109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 1318109998Smarkm of libcrypto. 1319109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 1320109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 1321109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 1322109998Smarkm make such changes in the libcrypto locking code that changes will 1323109998Smarkm have to be made anyway). 1324109998Smarkm [Richard Levitte] 1325109998Smarkm 1326109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 1327109998Smarkm octets have been read, EOF or an error occurs. Without this change 1328109998Smarkm some truncated ASN1 structures will not produce an error. 1329109998Smarkm [Steve Henson] 1330109998Smarkm 1331109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 1332109998Smarkm Still give the possibility to force the use of Heimdal, but with 1333109998Smarkm warnings and a request that patches get sent to openssl-dev. 1334109998Smarkm [Richard Levitte] 1335109998Smarkm 1336109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 1337109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 1338109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 1339109998Smarkm 1340109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 1341109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 1342109998Smarkm edit numbers of the version. 1343109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 1344109998Smarkm 1345109998Smarkm *) Introduce safe string copy and catenation functions 1346109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 1347109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 1348109998Smarkm 1349109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 1350109998Smarkm [Ben Laurie (CHATS)] 1351109998Smarkm 1352109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 1353109998Smarkm resizing buffers containing secrets, and use where appropriate. 1354109998Smarkm [Ben Laurie (CHATS)] 1355109998Smarkm 1356109998Smarkm *) Avoid using fixed size buffers for configuration file location. 1357109998Smarkm [Ben Laurie (CHATS)] 1358109998Smarkm 1359109998Smarkm *) Avoid filename truncation for various CA files. 1360109998Smarkm [Ben Laurie (CHATS)] 1361109998Smarkm 1362109998Smarkm *) Use sizeof in preference to magic numbers. 1363109998Smarkm [Ben Laurie (CHATS)] 1364109998Smarkm 1365109998Smarkm *) Avoid filename truncation in cert requests. 1366109998Smarkm [Ben Laurie (CHATS)] 1367109998Smarkm 1368109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 1369109998Smarkm overflows. 1370109998Smarkm [Ben Laurie (CHATS)] 1371109998Smarkm 1372109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 1373109998Smarkm potentially lead to a spoofing attack). 1374109998Smarkm [Ben Laurie (CHATS)] 1375109998Smarkm 1376109998Smarkm *) Fix various buffers to be large enough for hex/decimal 1377109998Smarkm representations in a platform independent manner. 1378109998Smarkm [Ben Laurie (CHATS)] 1379109998Smarkm 1380109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 1381109998Smarkm resizing buffers containing secrets, and use where appropriate. 1382109998Smarkm [Ben Laurie (CHATS)] 1383109998Smarkm 1384109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 1385109998Smarkm indents. 1386109998Smarkm [Ben Laurie (CHATS)] 1387109998Smarkm 1388109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 1389109998Smarkm [Ben Laurie (CHATS)] 1390109998Smarkm 1391109998Smarkm *) buffer_gets() could terminate with the buffer only half 1392109998Smarkm full. Fixed. 1393109998Smarkm [Ben Laurie (CHATS)] 1394109998Smarkm 1395109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 1396109998Smarkm overflowing internal buffers by having large block sizes, etc. 1397109998Smarkm [Ben Laurie (CHATS)] 1398109998Smarkm 1399109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 1400109998Smarkm unconditionally). 1401109998Smarkm [Ben Laurie (CHATS)] 1402109998Smarkm 1403109998Smarkm *) Eliminate unused copy of key in RC4. 1404109998Smarkm [Ben Laurie (CHATS)] 1405109998Smarkm 1406109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 1407109998Smarkm [Ben Laurie (CHATS)] 1408109998Smarkm 1409109998Smarkm *) Fix off-by-one error in EGD path. 1410109998Smarkm [Ben Laurie (CHATS)] 1411109998Smarkm 1412109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 1413109998Smarkm [Ben Laurie (CHATS)] 1414109998Smarkm 1415109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 1416109998Smarkm CBCParameter. 1417109998Smarkm [Ben Laurie (CHATS)] 1418109998Smarkm 1419109998Smarkm *) Eliminate unused and dangerous function knumber(). 1420109998Smarkm [Ben Laurie (CHATS)] 1421109998Smarkm 1422109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 1423109998Smarkm [Ben Laurie (CHATS)] 1424109998Smarkm 1425109998Smarkm *) Protect against overlong session ID context length in an encoded 1426109998Smarkm session object. Since these are local, this does not appear to be 1427109998Smarkm exploitable. 1428109998Smarkm [Ben Laurie (CHATS)] 1429109998Smarkm 1430109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 1431109998Smarkm the 0.9.6 release series: 1432109998Smarkm 1433109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 1434109998Smarkm supply an oversized master key in Kerberos-enabled versions. 1435160814Ssimon (CVE-2002-0657) 1436109998Smarkm [Ben Laurie (CHATS)] 1437109998Smarkm 1438109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 1439109998Smarkm [Richard Levitte] 1440109998Smarkm 1441109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 1442109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 1443109998Smarkm 1444109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 1445109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 1446109998Smarkm 1447109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 1448109998Smarkm have been removed entirely. This was also the last step to make 1449109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 1450109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 1451109998Smarkm 1452109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 1453109998Smarkm to allow version independent disabling of normally unselected ciphers, 1454109998Smarkm which may be activated as a side-effect of selecting a single cipher. 1455109998Smarkm 1456109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 1457109998Smarkm out of "ALL" because they do not provide symmetric encryption. 1458109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 1459109998Smarkm [Lutz Jaenicke, Bodo Moeller] 1460109998Smarkm 1461109998Smarkm *) Add appropriate support for separate platform-dependent build 1462109998Smarkm directories. The recommended way to make a platform-dependent 1463109998Smarkm build directory is the following (tested on Linux), maybe with 1464109998Smarkm some local tweaks: 1465109998Smarkm 1466109998Smarkm # Place yourself outside of the OpenSSL source tree. In 1467109998Smarkm # this example, the environment variable OPENSSL_SOURCE 1468109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 1469109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 1470109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 1471111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 1472109998Smarkm mkdir -p `dirname $F` 1473109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 1474109998Smarkm done 1475109998Smarkm 1476109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 1477109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 1478109998Smarkm it probably means the source directory is very clean. 1479109998Smarkm [Richard Levitte] 1480109998Smarkm 1481109998Smarkm *) Make sure any ENGINE control commands make local copies of string 1482109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 1483109998Smarkm the caller may have overwritten (or deallocated) the original string 1484109998Smarkm data when a later ENGINE operation tries to use the stored values. 1485109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 1486109998Smarkm 1487109998Smarkm *) Improve diagnostics in file reading and command-line digests. 1488109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 1489109998Smarkm 1490109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 1491109998Smarkm error in AES-CFB decryption. 1492109998Smarkm [Richard Levitte] 1493109998Smarkm 1494109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 1495109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 1496109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 1497109998Smarkm BIOs and some applications. This has the side effect that 1498109998Smarkm applications must explicitly clean up cipher contexts with 1499109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 1500109998Smarkm [Steve Henson] 1501109998Smarkm 1502109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 1503109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 1504109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 1505109998Smarkm [Steve Henson] 1506109998Smarkm 1507109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 1508109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 1509109998Smarkm [Lutz Jaenicke] 1510109998Smarkm 1511109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 1512109998Smarkm form for "surname", serialNumber has no short form. 1513109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 1514109998Smarkm therefore remove "mail" short name for "internet 7". 1515109998Smarkm The OID for unique identifiers in X509 certificates is 1516109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 1517109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 1518109998Smarkm [Lutz Jaenicke] 1519109998Smarkm 1520109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 1521109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 1522109998Smarkm after all ctrl commands have been executed on it. If init=1 the 1523109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 1524109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 1525109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 1526109998Smarkm [Steve Henson] 1527109998Smarkm 1528109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 1529109998Smarkm argument is actually passed to the callback: In the 1530109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 1531109998Smarkm declaration has been changed from 1532109998Smarkm int (*cb)() 1533109998Smarkm into 1534109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 1535109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 1536109998Smarkm i=s->ctx->app_verify_callback(&ctx) 1537109998Smarkm has been changed into 1538109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 1539109998Smarkm 1540109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 1541109998Smarkm a dummy argument can be added to their callback functions. 1542109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 1543109998Smarkm 1544109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 1545109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 1546109998Smarkm 1547109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 1548109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 1549109998Smarkm This allows older applications to transparently support certain 1550109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 1551109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 1552109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 1553109998Smarkm always load it have also been added. 1554109998Smarkm [Steve Henson] 1555109998Smarkm 1556109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 1557109998Smarkm Adjust NIDs and EVP layer. 1558109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 1559109998Smarkm 1560109998Smarkm *) Config modules support in openssl utility. 1561109998Smarkm 1562109998Smarkm Most commands now load modules from the config file, 1563109998Smarkm though in a few (such as version) this isn't done 1564109998Smarkm because it couldn't be used for anything. 1565109998Smarkm 1566109998Smarkm In the case of ca and req the config file used is 1567109998Smarkm the same as the utility itself: that is the -config 1568109998Smarkm command line option can be used to specify an 1569109998Smarkm alternative file. 1570109998Smarkm [Steve Henson] 1571109998Smarkm 1572109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 1573109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 1574109998Smarkm [Steve Henson] 1575109998Smarkm 1576109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 1577109998Smarkm config section name. Add a new flag to tolerate a missing config file 1578109998Smarkm and move code to CONF_modules_load_file(). 1579109998Smarkm [Steve Henson] 1580109998Smarkm 1581109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 1582109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 1583109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 1584109998Smarkm to work with the new engine framework. 1585109998Smarkm [AEP Inc. and Richard Levitte] 1586109998Smarkm 1587109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 1588109998Smarkm Technologies. (Use engine 'sureware') 1589109998Smarkm The support was copied from 0.9.6c [engine] and adapted 1590109998Smarkm to work with the new engine framework. 1591109998Smarkm [Richard Levitte] 1592109998Smarkm 1593109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 1594109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 1595109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 1596109998Smarkm 1597109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 1598109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 1599109998Smarkm 1600109998Smarkm *) Add the configuration target debug-linux-ppro. 1601109998Smarkm Make 'openssl rsa' use the general key loading routines 1602109998Smarkm implemented in apps.c, and make those routines able to 1603109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 1604109998Smarkm FORMAT_IISSGC. 1605109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 1606109998Smarkm 1607109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 1608109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 1609109998Smarkm 1610109998Smarkm *) Add -keyform to rsautl, and document -engine. 1611109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 1612109998Smarkm 1613109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 1614109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 1615109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 1616109998Smarkm [Ben Laurie] 1617109998Smarkm 1618109998Smarkm *) Add new functions 1619109998Smarkm ERR_peek_last_error 1620109998Smarkm ERR_peek_last_error_line 1621109998Smarkm ERR_peek_last_error_line_data. 1622109998Smarkm These are similar to 1623109998Smarkm ERR_peek_error 1624109998Smarkm ERR_peek_error_line 1625109998Smarkm ERR_peek_error_line_data, 1626109998Smarkm but report on the latest error recorded rather than the first one 1627109998Smarkm still in the error queue. 1628109998Smarkm [Ben Laurie, Bodo Moeller] 1629109998Smarkm 1630109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 1631109998Smarkm like: 1632109998Smarkm default_algorithms = ALL 1633109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 1634109998Smarkm [Steve Henson] 1635109998Smarkm 1636109998Smarkm *) Prelminary ENGINE config module. 1637109998Smarkm [Steve Henson] 1638109998Smarkm 1639109998Smarkm *) New experimental application configuration code. 1640109998Smarkm [Steve Henson] 1641109998Smarkm 1642109998Smarkm *) Change the AES code to follow the same name structure as all other 1643109998Smarkm symmetric ciphers, and behave the same way. Move everything to 1644109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 1645109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 1646109998Smarkm 1647109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 1648109998Smarkm [Ben Laurie and Theo de Raadt] 1649109998Smarkm 1650109998Smarkm *) Add option to output public keys in req command. 1651109998Smarkm [Massimiliano Pala madwolf@openca.org] 1652109998Smarkm 1653109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 1654109998Smarkm (up to about 10% better than before for P-192 and P-224). 1655109998Smarkm [Bodo Moeller] 1656109998Smarkm 1657109998Smarkm *) New functions/macros 1658109998Smarkm 1659109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 1660109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 1661109998Smarkm SSL_set_msg_callback(ssl, cb) 1662109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 1663109998Smarkm 1664109998Smarkm to request calling a callback function 1665109998Smarkm 1666109998Smarkm void cb(int write_p, int version, int content_type, 1667109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 1668109998Smarkm 1669109998Smarkm whenever a protocol message has been completely received 1670109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 1671109998Smarkm protocol version according to which the SSL library interprets 1672109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 1673109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 1674109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 1675109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 1676109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 1677109998Smarkm SSL object, and 'arg' is the application-defined value set by 1678109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 1679109998Smarkm 1680109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 1681109998Smarkm to enable a callback that displays all protocol messages. 1682109998Smarkm [Bodo Moeller] 1683109998Smarkm 1684109998Smarkm *) Change the shared library support so shared libraries are built as 1685109998Smarkm soon as the corresponding static library is finished, and thereby get 1686109998Smarkm openssl and the test programs linked against the shared library. 1687109998Smarkm This still only happens when the keyword "shard" has been given to 1688109998Smarkm the configuration scripts. 1689109998Smarkm 1690109998Smarkm NOTE: shared library support is still an experimental thing, and 1691109998Smarkm backward binary compatibility is still not guaranteed. 1692109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 1693109998Smarkm 1694109998Smarkm *) Add support for Subject Information Access extension. 1695109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 1696109998Smarkm 1697109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 1698109998Smarkm additional bytes when new memory had to be allocated, not just 1699109998Smarkm when reusing an existing buffer. 1700109998Smarkm [Bodo Moeller] 1701109998Smarkm 1702109998Smarkm *) New command line and configuration option 'utf8' for the req command. 1703109998Smarkm This allows field values to be specified as UTF8 strings. 1704109998Smarkm [Steve Henson] 1705109998Smarkm 1706109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 1707109998Smarkm runs for the former and machine-readable output for the latter. 1708109998Smarkm [Ben Laurie] 1709109998Smarkm 1710109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 1711109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 1712109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 1713109998Smarkm has the same effect. 1714109998Smarkm [Massimiliano Pala madwolf@openca.org] 1715109998Smarkm 1716109998Smarkm *) Change all functions with names starting with des_ to be starting 1717109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 1718109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 1719109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 1720109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 1721109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 1722109998Smarkm exception. 1723109998Smarkm 1724109998Smarkm Since we provide two compatibility mappings, the user needs to 1725109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 1726109998Smarkm compatibility is desired. The default (i.e., when that macro 1727109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 1728109998Smarkm 1729109998Smarkm There are also macros that enable and disable the support of old 1730109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 1731109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 1732109998Smarkm are defined, the default will apply: to support the old des routines. 1733109998Smarkm 1734109998Smarkm In either case, one must include openssl/des.h to get the correct 1735109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 1736109998Smarkm won't work. 1737109998Smarkm 1738109998Smarkm NOTE: This is a major break of an old API into a new one. Software 1739109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 1740109998Smarkm time in the future, des_old.h and the libdes compatibility functions 1741109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 1742109998Smarkm default), and then completely removed. 1743109998Smarkm [Richard Levitte] 1744109998Smarkm 1745109998Smarkm *) Test for certificates which contain unsupported critical extensions. 1746109998Smarkm If such a certificate is found during a verify operation it is 1747109998Smarkm rejected by default: this behaviour can be overridden by either 1748109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 1749109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 1750109998Smarkm X509_supported_extension() has also been added which returns 1 if a 1751109998Smarkm particular extension is supported. 1752109998Smarkm [Steve Henson] 1753109998Smarkm 1754109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 1755109998Smarkm to retain compatibility with existing code. 1756109998Smarkm [Steve Henson] 1757109998Smarkm 1758109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 1759109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 1760109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 1761109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 1762109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 1763109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 1764109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 1765109998Smarkm requires the destination to be valid. 1766109998Smarkm 1767109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 1768109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 1769109998Smarkm [Steve Henson] 1770109998Smarkm 1771109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 1772109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 1773109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 1774109998Smarkm [Bodo Moeller] 1775109998Smarkm 1776109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 1777109998Smarkm [Massimo Santin via Richard Levitte] 1778109998Smarkm 1779109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 1780109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 1781109998Smarkm (initialisation, etc) from functionality dealing with implementations 1782109998Smarkm of specific crypto iterfaces. This change also introduces integrated 1783109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 1784109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 1785109998Smarkm implementations of their own. This is detailed in crypto/engine/README 1786109998Smarkm as it couldn't be adequately described here. However, there are a few 1787109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 1788109998Smarkm were changed in the original introduction of ENGINE code have now 1789109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 1790109998Smarkm deal more passive and at run-time, operations deal directly with 1791109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 1792109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 1793109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 1794109998Smarkm they were not being used by the framework as there is no concept of a 1795109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 1796109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 1797109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 1798109998Smarkm the new code. 1799109998Smarkm [Geoff Thorpe] 1800109998Smarkm 1801109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 1802109998Smarkm [Steve Henson] 1803109998Smarkm 1804109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 1805109998Smarkm and make sure the automatically generated functions ERR_load_* 1806109998Smarkm become part of libeay.num as well. 1807109998Smarkm [Richard Levitte] 1808109998Smarkm 1809109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 1810109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 1811109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 1812109998Smarkm false once a handshake has been completed. 1813109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 1814109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 1815109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 1816109998Smarkm client has followed the request.) 1817109998Smarkm [Bodo Moeller] 1818109998Smarkm 1819109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 1820109998Smarkm By default, clients may request session resumption even during 1821109998Smarkm renegotiation (if session ID contexts permit); with this option, 1822109998Smarkm session resumption is possible only in the first handshake. 1823109998Smarkm 1824109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 1825109998Smarkm more bits available for options that should not be part of 1826109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 1827109998Smarkm [Bodo Moeller] 1828109998Smarkm 1829109998Smarkm *) Add some demos for certificate and certificate request creation. 1830109998Smarkm [Steve Henson] 1831109998Smarkm 1832109998Smarkm *) Make maximum certificate chain size accepted from the peer application 1833109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 1834109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 1835109998Smarkm [Lutz Jaenicke] 1836109998Smarkm 1837109998Smarkm *) Add support for shared libraries for Unixware-7 1838109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 1839109998Smarkm [Lutz Jaenicke] 1840109998Smarkm 1841109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 1842109998Smarkm be done prior to destruction. Use this to unload error strings from 1843109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 1844109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 1845109998Smarkm [Geoff Thorpe] 1846109998Smarkm 1847109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 1848109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 1849109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 1850109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 1851109998Smarkm Also, add stub code to each that makes building them as self-contained 1852109998Smarkm shared-libraries easier (see README.ENGINE). 1853109998Smarkm [Geoff Thorpe] 1854109998Smarkm 1855109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 1856109998Smarkm implementations into applications that are completely implemented in 1857109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 1858109998Smarkm commands that can be used to configure what shared-library to load and 1859109998Smarkm to control aspects of the way it is handled. Also, made an update to 1860109998Smarkm the README.ENGINE file that brings its information up-to-date and 1861109998Smarkm provides some information and instructions on the "dynamic" ENGINE 1862109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 1863109998Smarkm [Geoff Thorpe] 1864109998Smarkm 1865109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 1866109998Smarkm "ERR_unload_strings" function. 1867109998Smarkm [Geoff Thorpe] 1868109998Smarkm 1869109998Smarkm *) Add a copy() function to EVP_MD. 1870109998Smarkm [Ben Laurie] 1871109998Smarkm 1872109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 1873109998Smarkm md_data void pointer. 1874109998Smarkm [Ben Laurie] 1875109998Smarkm 1876109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 1877109998Smarkm that the digest can only process a single chunk of data 1878109998Smarkm (typically because it is provided by a piece of 1879109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 1880109998Smarkm is only going to provide a single chunk of data, and hence the 1881109998Smarkm framework needn't accumulate the data for oneshot drivers. 1882109998Smarkm [Ben Laurie] 1883109998Smarkm 1884109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 1885109998Smarkm functions. This change also alters the storage and management of global 1886109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 1887109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 1888109998Smarkm index counters. The API functions that use this state have been changed 1889109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 1890109998Smarkm and counter, and there is now an API function to dynamically create new 1891109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 1892109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 1893109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 1894109998Smarkm such data would previously have always leaked in application code and 1895109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 1896109998Smarkm to it. Application code that doesn't use this new function will still 1897109998Smarkm leak as before, but their memory debugging output will announce it now 1898109998Smarkm rather than letting it slide. 1899109998Smarkm 1900109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 1901109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 1902109998Smarkm has a return value to indicate success or failure. 1903109998Smarkm [Geoff Thorpe] 1904109998Smarkm 1905109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 1906109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 1907109998Smarkm implementation. This change also adds two functions to "get" and "set" 1908109998Smarkm the implementation prior to it being automatically set the first time 1909109998Smarkm any other ERR function takes place. Ie. an application can call "get", 1910109998Smarkm pass the return value to a module it has just loaded, and that module 1911109998Smarkm can call its own "set" function using that value. This means the 1912109998Smarkm module's "ERR" operations will use (and modify) the error state in the 1913109998Smarkm application and not in its own statically linked copy of OpenSSL code. 1914109998Smarkm [Geoff Thorpe] 1915109998Smarkm 1916109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 1917109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 1918109998Smarkm the operation, and provides a more encapsulated way for external code 1919109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 1920109998Smarkm to use these functions rather than manually incrementing the counts. 1921109998Smarkm 1922109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 1923109998Smarkm [Geoff Thorpe] 1924109998Smarkm 1925109998Smarkm *) Add EVP test program. 1926109998Smarkm [Ben Laurie] 1927109998Smarkm 1928109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 1929109998Smarkm [Ben Laurie] 1930109998Smarkm 1931109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 1932109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 1933109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 1934109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 1935109998Smarkm directly. Modify 'ca' application to use new functions. 1936109998Smarkm [Steve Henson] 1937109998Smarkm 1938109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 1939109998Smarkm bug workarounds. Rollback attack detection is a security feature. 1940109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 1941109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 1942109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 1943109998Smarkm for their choice and can explicitly enable this option. 1944109998Smarkm [Bodo Moeller, Lutz Jaenicke] 1945109998Smarkm 1946109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 1947109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 1948109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 1949109998Smarkm Usage example: 1950109998Smarkm 1951109998Smarkm EVP_MD_CTX md; 1952109998Smarkm 1953109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 1954109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 1955109998Smarkm EVP_DigestUpdate(&md, in, len); 1956109998Smarkm EVP_DigestFinal(&md, out, NULL); 1957109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 1958109998Smarkm 1959109998Smarkm [Ben Laurie] 1960109998Smarkm 1961109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 1962109998Smarkm correcting its structure. This means that calls to DES functions 1963109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 1964109998Smarkm plain des_key_schedule (which was actually always a pointer 1965109998Smarkm anyway): E.g., 1966109998Smarkm 1967109998Smarkm des_key_schedule ks; 1968109998Smarkm 1969109998Smarkm des_set_key_checked(..., &ks); 1970109998Smarkm des_ncbc_encrypt(..., &ks, ...); 1971109998Smarkm 1972109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 1973109998Smarkm [Ben Laurie] 1974109998Smarkm 1975109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 1976109998Smarkm PEM causes large amounts of unused functions to be linked in due to 1977109998Smarkm poor organisation. For example pem_all.c contains every PEM function 1978109998Smarkm which has a knock on effect of linking in large amounts of (unused) 1979109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 1980109998Smarkm functions prevents this. 1981109998Smarkm [Steve Henson] 1982109998Smarkm 1983109998Smarkm *) Cleanup of EVP macros. 1984109998Smarkm [Ben Laurie] 1985109998Smarkm 1986109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 1987109998Smarkm correct _ecb suffix. 1988109998Smarkm [Ben Laurie] 1989109998Smarkm 1990109998Smarkm *) Add initial OCSP responder support to ocsp application. The 1991109998Smarkm revocation information is handled using the text based index 1992109998Smarkm use by the ca application. The responder can either handle 1993109998Smarkm requests generated internally, supplied in files (for example 1994109998Smarkm via a CGI script) or using an internal minimal server. 1995109998Smarkm [Steve Henson] 1996109998Smarkm 1997109998Smarkm *) Add configuration choices to get zlib compression for TLS. 1998109998Smarkm [Richard Levitte] 1999109998Smarkm 2000109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 2001109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 2002109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 2003109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 2004109998Smarkm 2005109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 2006109998Smarkm and authenticator structs; see crypto/krb5/. 2007109998Smarkm 2008109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 2009109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 2010109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 2011109998Smarkm via Richard Levitte] 2012109998Smarkm 2013109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 2014109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 2015109998Smarkm values for each of the key sizes rather than having just 2016109998Smarkm parameters (and 'speed' generating keys each time). 2017109998Smarkm [Geoff Thorpe] 2018109998Smarkm 2019109998Smarkm *) Speed up EVP routines. 2020109998Smarkm Before: 2021109998Smarkmencrypt 2022109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 2023109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 2024109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 2025109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 2026109998Smarkmdecrypt 2027109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 2028109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 2029109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 2030109998Smarkm After: 2031109998Smarkmencrypt 2032109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 2033109998Smarkmdecrypt 2034109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 2035109998Smarkm [Ben Laurie] 2036109998Smarkm 2037109998Smarkm *) Added the OS2-EMX target. 2038109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 2039109998Smarkm 2040109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 2041109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 2042109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 2043109998Smarkm structure: this means that the old CONF compatible routines can be 2044109998Smarkm retained (in particular wrt extensions) without having to duplicate the 2045109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 2046109998Smarkm [Steve Henson] 2047109998Smarkm 2048109998Smarkm *) Enhance the general user interface with mechanisms for inner control 2049109998Smarkm and with possibilities to have yes/no kind of prompts. 2050109998Smarkm [Richard Levitte] 2051109998Smarkm 2052109998Smarkm *) Change all calls to low level digest routines in the library and 2053109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 2054109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 2055109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 2056109998Smarkm 2057109998Smarkm *) Add the possibility to control engines through control names but with 2058109998Smarkm arbitrary arguments instead of just a string. 2059109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 2060109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 2061109998Smarkm versions of OpenSSL [engine]. 2062109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 2063109998Smarkm callback. 2064109998Smarkm [Richard Levitte] 2065109998Smarkm 2066109998Smarkm *) Enhance the general user interface with mechanisms to better support 2067109998Smarkm dialog box interfaces, application-defined prompts, the possibility 2068109998Smarkm to use defaults (for example default passwords from somewhere else) 2069109998Smarkm and interrupts/cancellations. 2070109998Smarkm [Richard Levitte] 2071109998Smarkm 2072109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 2073109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 2074109998Smarkm [Steve Henson] 2075109998Smarkm 2076109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 2077109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 2078109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 2079109998Smarkm 2080109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 2081109998Smarkm callback (pem_password_cb) as all other routines that need this 2082109998Smarkm kind of callback. 2083109998Smarkm [Richard Levitte] 2084109998Smarkm 2085109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 2086109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 2087109998Smarkm than this minimum value is recommended. 2088109998Smarkm [Lutz Jaenicke] 2089109998Smarkm 2090109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 2091109998Smarkm that are easily reachable. 2092109998Smarkm [Richard Levitte] 2093109998Smarkm 2094109998Smarkm *) Windows apparently can't transparently handle global 2095109998Smarkm variables defined in DLLs. Initialisations such as: 2096109998Smarkm 2097109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 2098109998Smarkm 2099109998Smarkm wont compile. This is used by the any applications that need to 2100109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 2101109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 2102109998Smarkm needed for static libraries under Win32. 2103109998Smarkm [Steve Henson] 2104109998Smarkm 2105109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 2106109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 2107109998Smarkm purpose functions and tidy up setting in other SSL functions. 2108109998Smarkm [Steve Henson] 2109109998Smarkm 2110109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 2111109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 2112109998Smarkm initialised. This allows various defaults to be set in the 2113109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 2114109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 2115109998Smarkm internally such as S/MIME. 2116109998Smarkm 2117109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 2118109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 2119109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 2120109998Smarkm 2121109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 2122109998Smarkm applications. 2123109998Smarkm [Steve Henson] 2124109998Smarkm 2125109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 2126109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 2127109998Smarkm its validity and signature checked, then if the certificate is found 2128109998Smarkm in the CRL the verify fails with a revoked error. 2129109998Smarkm 2130109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 2131109998Smarkm 2132109998Smarkm Command line options added to 'verify' application to support this. 2133109998Smarkm 2134109998Smarkm This needs some additional work, such as being able to handle multiple 2135109998Smarkm CRLs with different times, extension based lookup (rather than just 2136109998Smarkm by subject name) and ultimately more complete V2 CRL extension 2137109998Smarkm handling. 2138109998Smarkm [Steve Henson] 2139109998Smarkm 2140109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 2141109998Smarkm to replace things like des_read_password and friends (backward 2142109998Smarkm compatibility functions using this new API are provided). 2143109998Smarkm The purpose is to remove prompting functions from the DES code 2144109998Smarkm section as well as provide for prompting through dialog boxes in 2145109998Smarkm a window system and the like. 2146109998Smarkm [Richard Levitte] 2147109998Smarkm 2148109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 2149109998Smarkm per-structure level rather than having to store it globally. 2150109998Smarkm [Geoff] 2151109998Smarkm 2152109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 2153109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 2154109998Smarkm This causes the "original" ENGINE structure to act like a template, 2155109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 2156109998Smarkm operational state can be localised to each ENGINE structure, despite the 2157109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 2158109998Smarkm this case have no functional references and the return value is the single 2159109998Smarkm structural reference. This matches the single structural reference returned 2160109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 2161109998Smarkm ENGINE structure. 2162109998Smarkm [Geoff] 2163109998Smarkm 2164109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 2165109998Smarkm needs to match any other type at all we need to manually clear the 2166109998Smarkm tag cache. 2167109998Smarkm [Steve Henson] 2168109998Smarkm 2169109998Smarkm *) Changes to the "openssl engine" utility to include; 2170109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 2171109998Smarkm about an ENGINE's available control commands. 2172109998Smarkm - executing control commands from command line arguments using the 2173109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 2174109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 2175109998Smarkm the individual commands are colon-separated, for example; 2176109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 2177109998Smarkm [Geoff] 2178109998Smarkm 2179109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 2180109998Smarkm declare their own commands (numbers), names (strings), descriptions, 2181109998Smarkm and input types for run-time discovery by calling applications. A 2182109998Smarkm subset of these commands are implicitly classed as "executable" 2183109998Smarkm depending on their input type, and only these can be invoked through 2184109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 2185109998Smarkm can be based on user input, config files, etc). The distinction is 2186109998Smarkm that "executable" commands cannot return anything other than a boolean 2187109998Smarkm result and can only support numeric or string input, whereas some 2188109998Smarkm discoverable commands may only be for direct use through 2189109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 2190109998Smarkm pointers, or other custom uses. The "executable" commands are to 2191109998Smarkm support parameterisations of ENGINE behaviour that can be 2192109998Smarkm unambiguously defined by ENGINEs and used consistently across any 2193109998Smarkm OpenSSL-based application. Commands have been added to all the 2194109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 2195109998Smarkm control over shared-library paths without source code alterations. 2196109998Smarkm [Geoff] 2197109998Smarkm 2198109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 2199109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 2200109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 2201109998Smarkm this also allows the implementations to compile without using the 2202109998Smarkm internal engine_int.h header. 2203109998Smarkm [Geoff] 2204109998Smarkm 2205109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 2206109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 2207109998Smarkm should already have non-const pointers to it (ie. they should only 2208109998Smarkm modify their own ones). 2209109998Smarkm [Geoff] 2210109998Smarkm 2211109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 2212109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 2213109998Smarkm to C code. "nuron" string definitions were placed in variables 2214109998Smarkm rather than hard-coded - allowing parameterisation of these values 2215109998Smarkm later on via ctrl() commands. 2216109998Smarkm - Removed unused "#if 0"'d code. 2217109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 2218109998Smarkm structural references. 2219109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 2220109998Smarkm - Constified various get/set functions as appropriate and added 2221109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 2222109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 2223109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 2224109998Smarkm or function to NULL is a way of cancelling out a previously set 2225109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 2226109998Smarkm and doesn't justify the extra error symbols and code. 2227109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 2228109998Smarkm flags from engine_int.h to engine.h. 2229109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 2230109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 2231109998Smarkm [Geoff] 2232109998Smarkm 2233109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 2234109998Smarkm to the algorithm using long division. The binary algorithm can be 2235109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 2236109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 2237109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 2238109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 2239109998Smarkm appears to be advantageous for much longer moduli; here we use it 2240109998Smarkm for moduli up to 2048 bits. 2241109998Smarkm [Bodo Moeller] 2242109998Smarkm 2243109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 2244109998Smarkm could not support the combine flag in choice fields. 2245109998Smarkm [Steve Henson] 2246109998Smarkm 2247109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 2248109998Smarkm extensions from a certificate request to the certificate. 2249109998Smarkm [Steve Henson] 2250109998Smarkm 2251109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 2252109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 2253109998Smarkm file: this allows the display of the certificate about to be 2254109998Smarkm signed to be customised, to allow certain fields to be included 2255109998Smarkm or excluded and extension details. The old system didn't display 2256109998Smarkm multicharacter strings properly, omitted fields not in the policy 2257109998Smarkm and couldn't display additional details such as extensions. 2258109998Smarkm [Steve Henson] 2259109998Smarkm 2260109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 2261109998Smarkm of an arbitrary number of elliptic curve points 2262109998Smarkm \sum scalars[i]*points[i], 2263109998Smarkm optionally including the generator defined for the EC_GROUP: 2264109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 2265109998Smarkm 2266109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 2267109998Smarkm that the point list has just one item (besides the optional 2268109998Smarkm generator). 2269109998Smarkm [Bodo Moeller] 2270109998Smarkm 2271109998Smarkm *) First EC_METHODs for curves over GF(p): 2272109998Smarkm 2273109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 2274109998Smarkm operations and provides various method functions that can also 2275109998Smarkm operate with faster implementations of modular arithmetic. 2276109998Smarkm 2277109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 2278109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 2279109998Smarkm 2280109998Smarkm [Bodo Moeller; point addition and point doubling 2281109998Smarkm implementation directly derived from source code provided by 2282109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 2283109998Smarkm 2284109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 2285109998Smarkm crypto/ec/ec_lib.c): 2286109998Smarkm 2287109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 2288109998Smarkm based on EC_METHODs that are built into the library. 2289109998Smarkm 2290109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 2291109998Smarkm 2292109998Smarkm Most of the framework would be able to handle curves over arbitrary 2293109998Smarkm finite fields, but as there are no obvious types for fields other 2294109998Smarkm than GF(p), some functions are limited to that for now. 2295109998Smarkm [Bodo Moeller] 2296109998Smarkm 2297109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 2298109998Smarkm that the file contains a complete HTTP response. 2299109998Smarkm [Richard Levitte] 2300109998Smarkm 2301109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 2302109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 2303109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 2304109998Smarkm field while the former will cause them to run together if the field 2305109998Smarkm is 40 of more characters long. 2306109998Smarkm [Steve Henson] 2307109998Smarkm 2308109998Smarkm *) Constify the cipher and digest 'method' functions and structures 2309109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 2310109998Smarkm pointers. 2311109998Smarkm [Steve Henson] 2312109998Smarkm 2313109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 2314109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 2315109998Smarkm [Bodo Moeller] 2316109998Smarkm 2317109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 2318109998Smarkm internal software routines can never fail additional hardware versions 2319109998Smarkm might. 2320109998Smarkm [Steve Henson] 2321109998Smarkm 2322109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 2323109998Smarkm 2324109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 2325109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 2326109998Smarkm 2327109998Smarkm ASN1 error codes 2328109998Smarkm ERR_R_NESTED_ASN1_ERROR 2329109998Smarkm ... 2330109998Smarkm ERR_R_MISSING_ASN1_EOS 2331109998Smarkm were 4 .. 9, conflicting with 2332109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 2333109998Smarkm ... 2334109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 2335109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 2336109998Smarkm 2337109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 2338109998Smarkm [Bodo Moeller] 2339109998Smarkm 2340109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 2341109998Smarkm suffices. 2342109998Smarkm [Bodo Moeller] 2343109998Smarkm 2344109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 2345109998Smarkm sets the subject name for a new request or supersedes the 2346109998Smarkm subject name in a given request. Formats that can be parsed are 2347109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 2348109998Smarkm and 2349109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 2350109998Smarkm 2351109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 2352109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 2353109998Smarkm 2354109998Smarkm *) Introduce the possibility to access global variables through 2355109998Smarkm functions on platform were that's the best way to handle exporting 2356109998Smarkm global variables in shared libraries. To enable this functionality, 2357109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 2358109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 2359109998Smarkm is normally done by Configure or something similar). 2360109998Smarkm 2361109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 2362109998Smarkm in the source file (foo.c) like this: 2363109998Smarkm 2364109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 2365109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 2366109998Smarkm 2367109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 2368109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 2369109998Smarkm 2370109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 2371109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 2372109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 2373109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 2374109998Smarkm 2375109998Smarkm The #defines are very important, and therefore so is including the 2376109998Smarkm header file everywhere where the defined globals are used. 2377109998Smarkm 2378109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 2379109998Smarkm of ASN.1 items, but that structure is a bit different. 2380109998Smarkm 2381109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 2382109998Smarkm better and easier to understand logic to choose which symbols should 2383109998Smarkm go into the Windows .def files as well as a number of fixes and code 2384109998Smarkm cleanup (among others, algorithm keywords are now sorted 2385109998Smarkm lexicographically to avoid constant rewrites). 2386109998Smarkm [Richard Levitte] 2387109998Smarkm 2388109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 2389109998Smarkm result to 'rm' because if rm==num the value will be overwritten 2390109998Smarkm and produce the wrong result if 'num' is negative: this caused 2391109998Smarkm problems with BN_mod() and BN_nnmod(). 2392109998Smarkm [Steve Henson] 2393109998Smarkm 2394109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 2395109998Smarkm OCSP request and verifies the signer certificate. The signer 2396109998Smarkm certificate is just checked for a generic purpose and OCSP request 2397109998Smarkm trust settings. 2398109998Smarkm [Steve Henson] 2399109998Smarkm 2400109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 2401109998Smarkm responses. OCSP responses are prepared in real time and may only 2402109998Smarkm be a few seconds old. Simply checking that the current time lies 2403109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 2404109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 2405109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 2406109998Smarkm the current time. The age of the response can also optionally be 2407109998Smarkm checked. Two new options -validity_period and -status_age added to 2408109998Smarkm ocsp utility. 2409109998Smarkm [Steve Henson] 2410109998Smarkm 2411109998Smarkm *) If signature or public key algorithm is unrecognized print out its 2412109998Smarkm OID rather that just UNKNOWN. 2413109998Smarkm [Steve Henson] 2414109998Smarkm 2415109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 2416109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 2417109998Smarkm ID to be generated from the issuer certificate alone which can then be 2418109998Smarkm passed to OCSP_id_issuer_cmp(). 2419109998Smarkm [Steve Henson] 2420109998Smarkm 2421109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 2422109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 2423109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 2424109998Smarkm new macros which allow the underlying ASN1 function/structure to 2425109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 2426109998Smarkm references directly (such as &X509_it) but instead use the relevant 2427109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 2428109998Smarkm use of the new ASN1 code on platforms where exporting structures 2429109998Smarkm is problematical (for example in shared libraries) but exporting 2430109998Smarkm functions returning pointers to structures is not. 2431109998Smarkm [Steve Henson] 2432109998Smarkm 2433109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 2434109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 2435109998Smarkm The purpose of this is to allow applications to control, if they wish, 2436109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 2437109998Smarkm can be useful for session caching in multiple-server environments. A 2438109998Smarkm command-line switch for testing this (and any client code that wishes 2439109998Smarkm to use such a feature) has been added to "s_server". 2440109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 2441109998Smarkm 2442109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 2443109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 2444109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 2445109998Smarkm the growing number of special cases it was previously handling. 2446109998Smarkm [Richard Levitte] 2447109998Smarkm 2448109998Smarkm *) Make all configuration macros available for application by making 2449109998Smarkm sure they are available in opensslconf.h, by giving them names starting 2450109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 2451109998Smarkm sure e_os2.h will cover all platform-specific cases together with 2452109998Smarkm opensslconf.h. 2453109998Smarkm Additionally, it is now possible to define configuration/platform- 2454109998Smarkm specific names (called "system identities"). In the C code, these 2455109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 2456109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 2457109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 2458109998Smarkm what is available. 2459109998Smarkm [Richard Levitte] 2460109998Smarkm 2461109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 2462109998Smarkm number to use to be specified on the command line. Previously self 2463109998Smarkm signed certificates were hard coded with serial number 0 and the 2464109998Smarkm CA options of 'x509' had to use a serial number in a file which was 2465109998Smarkm auto incremented. 2466109998Smarkm [Steve Henson] 2467109998Smarkm 2468109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 2469109998Smarkm Currently CRL reason, invalidity date and hold instruction are 2470109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 2471109998Smarkm [Steve Henson] 2472109998Smarkm 2473109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 2474109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 2475109998Smarkm API, which was previously mandatory. This means that the data is 2476109998Smarkm not padded in any way and so the total length much be a multiple 2477109998Smarkm of the block size, otherwise an error occurs. 2478109998Smarkm [Steve Henson] 2479109998Smarkm 2480109998Smarkm *) Initial (incomplete) OCSP SSL support. 2481109998Smarkm [Steve Henson] 2482109998Smarkm 2483109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 2484109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 2485109998Smarkm option to ocsp utility. 2486109998Smarkm [Steve Henson] 2487109998Smarkm 2488109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 2489109998Smarkm reflects the various checks performed. Applications can decide 2490109998Smarkm whether to tolerate certain situations such as an absent nonce 2491109998Smarkm in a response when one was present in a request: the ocsp application 2492109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 2493109998Smarkm this is to allow responders to include a nonce in a response even if 2494109998Smarkm the request is nonce-less. 2495109998Smarkm [Steve Henson] 2496109998Smarkm 2497109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 2498109998Smarkm skipped when using openssl x509 multiple times on a single input file, 2499109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 2500109998Smarkm [Bodo Moeller] 2501109998Smarkm 2502109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 2503109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 2504109998Smarkm utility to correctly initialize revocation date of CRLs. 2505109998Smarkm [Steve Henson] 2506109998Smarkm 2507109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 2508109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 2509109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 2510109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 2511109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 2512109998Smarkm [Lutz Jaenicke] 2513109998Smarkm 2514109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 2515109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 2516109998Smarkm appear to exist. 2517109998Smarkm [Steve Henson] 2518109998Smarkm 2519109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 2520109998Smarkm additional certificates supplied. 2521109998Smarkm [Steve Henson] 2522109998Smarkm 2523109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 2524109998Smarkm OCSP client a number of certificate to only verify the response 2525109998Smarkm signature against. 2526109998Smarkm [Richard Levitte] 2527109998Smarkm 2528109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 2529109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 2530109998Smarkm AES OIDs. 2531109998Smarkm 2532109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 2533109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 2534109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 2535109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 2536109998Smarkm alias because they were not yet official; they could be 2537109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 2538109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 2539109998Smarkm alias is called "AES" and is part of "ALL".) 2540109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 2541109998Smarkm 2542109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 2543109998Smarkm request to response. 2544109998Smarkm [Steve Henson] 2545109998Smarkm 2546109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 2547109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 2548109998Smarkm extract information from a certificate request. OCSP_response_create() 2549109998Smarkm creates a response and optionally adds a basic response structure. 2550109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 2551109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 2552109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 2553109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 2554109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 2555109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 2556109998Smarkm (converts ASN1_TIME to GeneralizedTime). 2557109998Smarkm [Steve Henson] 2558109998Smarkm 2559109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 2560109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 2561109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 2562109998Smarkm contents: this is used in various key identifiers. 2563109998Smarkm [Steve Henson] 2564109998Smarkm 2565109998Smarkm *) Make sk_sort() tolerate a NULL argument. 2566109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 2567109998Smarkm 2568109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 2569109998Smarkm passed by the function are trusted implicitly. If any of them signed the 2570109998Smarkm response then it is assumed to be valid and is not verified. 2571109998Smarkm [Steve Henson] 2572109998Smarkm 2573109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 2574109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 2575109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 2576109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 2577109998Smarkm <support@securenetterm.com>] 2578109998Smarkm 2579109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 2580109998Smarkm routines: without these tracing memory leaks is very painful. 2581109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 2582109998Smarkm [Steve Henson] 2583109998Smarkm 2584109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 2585109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 2586109998Smarkm effectively meant GeneralizedTime would never be used. Now it 2587109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 2588109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 2589109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 2590109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 2591109998Smarkm <support@securenetterm.com>] 2592109998Smarkm 2593109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 2594109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 2595109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 2596109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 2597109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 2598109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 2599109998Smarkm [Steve Henson] 2600109998Smarkm 2601109998Smarkm *) Add summary printout to ocsp utility. The various functions which 2602109998Smarkm convert status values to strings have been renamed to: 2603109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 2604109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 2605109998Smarkm to verify nonce values and to disable verification. OCSP response 2606109998Smarkm printout format cleaned up. 2607109998Smarkm [Steve Henson] 2608109998Smarkm 2609109998Smarkm *) Add additional OCSP certificate checks. These are those specified 2610109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 2611109998Smarkm certificate being checked must either be the OCSP signer certificate 2612109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 2613109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 2614109998Smarkm usage. This check is performed by attempting to match the OCSP 2615109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 2616109998Smarkm in the OCSP_CERTID structures of the response. 2617109998Smarkm [Steve Henson] 2618109998Smarkm 2619109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 2620109998Smarkm and related routines. This uses the standard OpenSSL certificate 2621109998Smarkm verify routines to perform initial checks (just CA validity) and 2622109998Smarkm to obtain the certificate chain. Then additional checks will be 2623109998Smarkm performed on the chain. Currently the root CA is checked to see 2624109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 2625109998Smarkm a root CA as a global signing root: that is any certificate that 2626109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 2627109998Smarkm [Steve Henson] 2628109998Smarkm 2629109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 2630109998Smarkm extensions from a separate configuration file. 2631109998Smarkm As when reading extensions from the main configuration file, 2632109998Smarkm the '-extensions ...' option may be used for specifying the 2633109998Smarkm section to use. 2634109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 2635109998Smarkm 2636109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 2637109998Smarkm read. The request can be sent to a responder and the output 2638109998Smarkm parsed, outputed or printed in text form. Not complete yet: 2639109998Smarkm still needs to check the OCSP response validity. 2640109998Smarkm [Steve Henson] 2641109998Smarkm 2642109998Smarkm *) New subcommands for 'openssl ca': 2643109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 2644109998Smarkm the given serial number (according to the index file). 2645109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 2646109998Smarkm in the index file. 2647109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 2648109998Smarkm 2649109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 2650109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 2651109998Smarkm so that the resulting key is not encrypted. 2652109998Smarkm [Damien Miller <djm@mindrot.org>] 2653109998Smarkm 2654109998Smarkm *) New configuration for the GNU Hurd. 2655109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 2656109998Smarkm 2657109998Smarkm *) Initial code to implement OCSP basic response verify. This 2658109998Smarkm is currently incomplete. Currently just finds the signer's 2659109998Smarkm certificate and verifies the signature on the response. 2660109998Smarkm [Steve Henson] 2661109998Smarkm 2662109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 2663109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 2664109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 2665109998Smarkm [Bodo Moeller] 2666109998Smarkm 2667109998Smarkm *) Allowing defining memory allocation callbacks that will be given 2668109998Smarkm file name and line number information in additional arguments 2669109998Smarkm (a const char* and an int). The basic functionality remains, as 2670109998Smarkm well as the original possibility to just replace malloc(), 2671109998Smarkm realloc() and free() by functions that do not know about these 2672109998Smarkm additional arguments. To register and find out the current 2673109998Smarkm settings for extended allocation functions, the following 2674109998Smarkm functions are provided: 2675109998Smarkm 2676109998Smarkm CRYPTO_set_mem_ex_functions 2677109998Smarkm CRYPTO_set_locked_mem_ex_functions 2678109998Smarkm CRYPTO_get_mem_ex_functions 2679109998Smarkm CRYPTO_get_locked_mem_ex_functions 2680109998Smarkm 2681109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 2682109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 2683109998Smarkm extended allocation function is enabled. 2684109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 2685109998Smarkm a conventional allocation function is enabled. 2686109998Smarkm [Richard Levitte, Bodo Moeller] 2687109998Smarkm 2688109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 2689109998Smarkm There should no longer be any prototype-casting required when using 2690109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 2691109998Smarkm the callback types and macros at the head of lhash.h for details 2692109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 2693109998Smarkm [Geoff Thorpe] 2694109998Smarkm 2695109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 2696109998Smarkm If /dev/[u]random devices are not available or do not return enough 2697109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 2698109998Smarkm be queried. 2699109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 2700109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 2701109998Smarkm when enough entropy was collected without querying more sockets. 2702109998Smarkm [Lutz Jaenicke] 2703109998Smarkm 2704109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 2705109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 2706109998Smarkm of data has been collected. We spend at most 10 ms on each file 2707109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 2708109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 2709109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 2710109998Smarkm platforms the 10 ms delay will never occur. 2711109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 2712109998Smarkm For VMS, there's a currently-empty rand_vms.c. 2713109998Smarkm [Richard Levitte] 2714109998Smarkm 2715109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 2716109998Smarkm provide utility functions which an application needing 2717109998Smarkm to issue a request to an OCSP responder and analyse the 2718109998Smarkm response will typically need: as opposed to those which an 2719109998Smarkm OCSP responder itself would need which will be added later. 2720109998Smarkm 2721109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 2722109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 2723109998Smarkm response. OCSP_response_get1_basic() extracts basic response 2724109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 2725109998Smarkm information from an OCSP_CERTID structure (which will be created 2726109998Smarkm when the request structure is built). These are built from lower 2727109998Smarkm level functions which work on OCSP_SINGLERESP structures but 2728109998Smarkm wont normally be used unless the application wishes to examine 2729109998Smarkm extensions in the OCSP response for example. 2730109998Smarkm 2731109998Smarkm Replace nonce routines with a pair of functions. 2732109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 2733109998Smarkm generates a random value. OCSP_check_nonce() checks the 2734109998Smarkm validity of the nonce in an OCSP response. 2735109998Smarkm [Steve Henson] 2736109998Smarkm 2737109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 2738109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 2739109998Smarkm need to free up the newly created id. Change return type 2740109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 2741109998Smarkm This can then be used to add extensions to the request. 2742109998Smarkm Deleted OCSP_request_new(), since most of its functionality 2743109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 2744109998Smarkm clash) apart from the ability to set the request name which 2745109998Smarkm will be added elsewhere. 2746109998Smarkm [Steve Henson] 2747109998Smarkm 2748109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 2749109998Smarkm various functions. Extensions are now handled using the new 2750109998Smarkm OCSP extension code. New simple OCSP HTTP function which 2751109998Smarkm can be used to send requests and parse the response. 2752109998Smarkm [Steve Henson] 2753109998Smarkm 2754109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 2755109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 2756109998Smarkm uses the special reorder version of SET OF to sort the attributes 2757109998Smarkm and reorder them to match the encoded order. This resolves a long 2758109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 2759109998Smarkm it used to fail because the attribute order did not match the 2760109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 2761109998Smarkm it uses the received order. This is necessary to tolerate some broken 2762109998Smarkm software that does not order SET OF. This is handled by encoding 2763109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 2764109998Smarkm to produce the required SET OF. 2765109998Smarkm [Steve Henson] 2766109998Smarkm 2767109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 2768109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 2769109998Smarkm files to get correct declarations of the ASN.1 item variables. 2770109998Smarkm [Richard Levitte] 2771109998Smarkm 2772109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 2773109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 2774109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 2775109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 2776109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 2777109998Smarkm ASN1_ITEM and no wrapper functions. 2778109998Smarkm [Steve Henson] 2779109998Smarkm 2780109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 2781109998Smarkm replace the old function pointer based I/O routines. Change most of 2782109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 2783109998Smarkm [Steve Henson] 2784109998Smarkm 2785109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 2786109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 2787109998Smarkm it complain about algorithm deselection that isn't recognised. 2788109998Smarkm [Richard Levitte] 2789109998Smarkm 2790109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 2791109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 2792109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 2793109998Smarkm some old style ASN1 functions: this can be used to determine if old 2794109998Smarkm code will still work when these eventually go away. 2795109998Smarkm [Steve Henson] 2796109998Smarkm 2797109998Smarkm *) New extension functions for OCSP structures, these follow the 2798109998Smarkm same conventions as certificates and CRLs. 2799109998Smarkm [Steve Henson] 2800109998Smarkm 2801109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 2802109998Smarkm adds an extension. Its behaviour can be customised with various 2803109998Smarkm flags to append, replace or delete. Various wrappers added for 2804109998Smarkm certifcates and CRLs. 2805109998Smarkm [Steve Henson] 2806109998Smarkm 2807109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 2808109998Smarkm an extension cannot be parsed. Correct a typo in the 2809109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 2810109998Smarkm [Steve Henson] 2811109998Smarkm 2812109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 2813109998Smarkm entries for variables. 2814109998Smarkm [Steve Henson] 2815109998Smarkm 2816109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 2817109998Smarkm problems: As the program is single-threaded, all we have 2818109998Smarkm to do is register a locking callback using an array for 2819109998Smarkm storing which locks are currently held by the program. 2820109998Smarkm [Bodo Moeller] 2821109998Smarkm 2822109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 2823109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 2824109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 2825109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 2826109998Smarkm Unfortunately, the ex_data design is not at all suited 2827109998Smarkm for multi-threaded use, so it probably should be abolished. 2828109998Smarkm [Bodo Moeller] 2829109998Smarkm 2830109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 2831109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 2832109998Smarkm 2833109998Smarkm *) Move common extension printing code to new function 2834109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 2835109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 2836109998Smarkm [Steve Henson] 2837109998Smarkm 2838109998Smarkm *) New function X509_signature_print() to remove duplication in some 2839109998Smarkm print routines. 2840109998Smarkm [Steve Henson] 2841109998Smarkm 2842109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 2843109998Smarkm set (this was treated exactly the same as SET OF previously). This 2844109998Smarkm is used to reorder the STACK representing the structure to match the 2845109998Smarkm encoding. This will be used to get round a problem where a PKCS7 2846109998Smarkm structure which was signed could not be verified because the STACK 2847109998Smarkm order did not reflect the encoded order. 2848109998Smarkm [Steve Henson] 2849109998Smarkm 2850109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 2851109998Smarkm [Steve Henson] 2852109998Smarkm 2853109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 2854109998Smarkm for its ASN1 operations. The old style function pointers still exist 2855109998Smarkm for now but they will eventually go away. 2856109998Smarkm [Steve Henson] 2857109998Smarkm 2858109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 2859109998Smarkm completely replaces the old ASN1 functionality with a table driven 2860109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 2861109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 2862109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 2863109998Smarkm has also been converted to the new form. 2864109998Smarkm [Steve Henson] 2865109998Smarkm 2866109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 2867109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 2868109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 2869109998Smarkm for negative moduli. 2870109998Smarkm [Bodo Moeller] 2871109998Smarkm 2872109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 2873109998Smarkm of not touching the result's sign bit. 2874109998Smarkm [Bodo Moeller] 2875109998Smarkm 2876109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 2877109998Smarkm set. 2878109998Smarkm [Bodo Moeller] 2879109998Smarkm 2880109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 2881109998Smarkm macros to declare and implement thin (optionally static) functions 2882109998Smarkm that provide type-safety and avoid function pointer casting for the 2883109998Smarkm type-specific callbacks. 2884109998Smarkm [Geoff Thorpe] 2885109998Smarkm 2886109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 2887109998Smarkm RFC 2712. 2888109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 2889109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 2890109998Smarkm 2891109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 2892109998Smarkm in sections depending on the subject. 2893109998Smarkm [Richard Levitte] 2894109998Smarkm 2895109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 2896109998Smarkm Windows. 2897109998Smarkm [Richard Levitte] 2898109998Smarkm 2899109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 2900109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 2901109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 2902109998Smarkm be handled deterministically). 2903109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 2904109998Smarkm 2905109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 2906109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 2907109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 2908109998Smarkm [Bodo Moeller] 2909109998Smarkm 2910109998Smarkm *) New function BN_kronecker. 2911109998Smarkm [Bodo Moeller] 2912109998Smarkm 2913109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 2914109998Smarkm positive unless both parameters are zero. 2915109998Smarkm Previously something reasonably close to an infinite loop was 2916109998Smarkm possible because numbers could be growing instead of shrinking 2917109998Smarkm in the implementation of Euclid's algorithm. 2918109998Smarkm [Bodo Moeller] 2919109998Smarkm 2920109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 2921109998Smarkm sign of the number in question. 2922109998Smarkm 2923109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 2924109998Smarkm 2925109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 2926109998Smarkm because its test if the absolute value of 'a' equals 'w'. 2927109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 2928109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 2929109998Smarkm BN_is_one(), and BN_is_word(). 2930109998Smarkm [Bodo Moeller] 2931109998Smarkm 2932109998Smarkm *) New function BN_swap. 2933109998Smarkm [Bodo Moeller] 2934109998Smarkm 2935109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 2936109998Smarkm the exponentiation functions are more likely to produce reasonable 2937109998Smarkm results on negative inputs. 2938109998Smarkm [Bodo Moeller] 2939109998Smarkm 2940109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 2941109998Smarkm Previously, it could be negative if one of the factors was negative; 2942109998Smarkm I don't think anyone really wanted that behaviour. 2943109998Smarkm [Bodo Moeller] 2944109998Smarkm 2945109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 2946109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 2947109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 2948109998Smarkm and add new functions: 2949109998Smarkm 2950109998Smarkm BN_nnmod 2951109998Smarkm BN_mod_sqr 2952109998Smarkm BN_mod_add 2953109998Smarkm BN_mod_add_quick 2954109998Smarkm BN_mod_sub 2955109998Smarkm BN_mod_sub_quick 2956109998Smarkm BN_mod_lshift1 2957109998Smarkm BN_mod_lshift1_quick 2958109998Smarkm BN_mod_lshift 2959109998Smarkm BN_mod_lshift_quick 2960109998Smarkm 2961109998Smarkm These functions always generate non-negative results. 2962109998Smarkm 2963109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 2964109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 2965109998Smarkm 2966109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 2967109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 2968109998Smarkm be reduced modulo m. 2969109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 2970109998Smarkm 2971111147Snectar#if 0 2972111147Snectar The following entry accidentily appeared in the CHANGES file 2973111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 2974111147Snectar it do *not* apply to OpenSSL 0.9.7. 2975111147Snectar 2976109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 2977109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 2978109998Smarkm required a small change in bn_mul_part_recursive() and the addition 2979109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 2980109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 2981109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 2982109998Smarkm differing sizes. 2983109998Smarkm [Richard Levitte] 2984111147Snectar#endif 2985109998Smarkm 2986109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 2987109998Smarkm unless the '-salt' option is used (which usually means that 2988109998Smarkm verification would just waste user's time since the resulting 2989109998Smarkm hash is going to be compared with some given password hash) 2990109998Smarkm or the new '-noverify' option is used. 2991109998Smarkm 2992109998Smarkm This is an incompatible change, but it does not affect 2993109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 2994109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 2995109998Smarkm cause any problems. 2996109998Smarkm [Bodo Moeller] 2997109998Smarkm 2998109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 2999109998Smarkm [Richard Levitte] 3000109998Smarkm 3001109998Smarkm *) Make DSO load along a path given through an environment variable 3002109998Smarkm (SHLIB_PATH) with shl_load(). 3003109998Smarkm [Richard Levitte] 3004109998Smarkm 3005109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 3006109998Smarkm Also constify the RSA code and most things related to it. In a 3007109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 3008109998Smarkm casts back to non-const were required (to be solved at a later 3009109998Smarkm time) 3010109998Smarkm [Richard Levitte] 3011109998Smarkm 3012109998Smarkm *) Make it so the openssl application has all engines loaded by default. 3013109998Smarkm [Richard Levitte] 3014109998Smarkm 3015109998Smarkm *) Constify the BIGNUM routines a little more. 3016109998Smarkm [Richard Levitte] 3017109998Smarkm 3018109998Smarkm *) Add the following functions: 3019109998Smarkm 3020109998Smarkm ENGINE_load_cswift() 3021109998Smarkm ENGINE_load_chil() 3022109998Smarkm ENGINE_load_atalla() 3023109998Smarkm ENGINE_load_nuron() 3024109998Smarkm ENGINE_load_builtin_engines() 3025109998Smarkm 3026109998Smarkm That way, an application can itself choose if external engines that 3027109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 3028109998Smarkm that applications won't have to be linked with libdl or other dso 3029109998Smarkm libraries unless it's really needed. 3030109998Smarkm 3031109998Smarkm Changed 'openssl engine' to load all engines on demand. 3032109998Smarkm Changed the engine header files to avoid the duplication of some 3033109998Smarkm declarations (they differed!). 3034109998Smarkm [Richard Levitte] 3035109998Smarkm 3036109998Smarkm *) 'openssl engine' can now list capabilities. 3037109998Smarkm [Richard Levitte] 3038109998Smarkm 3039109998Smarkm *) Better error reporting in 'openssl engine'. 3040109998Smarkm [Richard Levitte] 3041109998Smarkm 3042109998Smarkm *) Never call load_dh_param(NULL) in s_server. 3043109998Smarkm [Bodo Moeller] 3044109998Smarkm 3045109998Smarkm *) Add engine application. It can currently list engines by name and 3046109998Smarkm identity, and test if they are actually available. 3047109998Smarkm [Richard Levitte] 3048109998Smarkm 3049109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 3050109998Smarkm sure the installed documentation is also owned by root.root. 3051109998Smarkm [Damien Miller <djm@mindrot.org>] 3052109998Smarkm 3053109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 3054109998Smarkm keys (public as well as private) handled by engines. 3055109998Smarkm [Richard Levitte] 3056109998Smarkm 3057109998Smarkm *) Add OCSP code that comes from CertCo. 3058109998Smarkm [Richard Levitte] 3059109998Smarkm 3060109998Smarkm *) Add VMS support for the Rijndael code. 3061109998Smarkm [Richard Levitte] 3062109998Smarkm 3063109998Smarkm *) Added untested support for Nuron crypto accelerator. 3064109998Smarkm [Ben Laurie] 3065109998Smarkm 3066109998Smarkm *) Add support for external cryptographic devices. This code was 3067109998Smarkm previously distributed separately as the "engine" branch. 3068109998Smarkm [Geoff Thorpe, Richard Levitte] 3069109998Smarkm 3070109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 3071109998Smarkm have far greater control over how a "name" is turned into a filename 3072109998Smarkm depending on the operating environment and any oddities about the 3073109998Smarkm different shared library filenames on each system. 3074109998Smarkm [Geoff Thorpe] 3075109998Smarkm 3076109998Smarkm *) Support threads on FreeBSD-elf in Configure. 3077109998Smarkm [Richard Levitte] 3078109998Smarkm 3079109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 3080109998Smarkm warnings about corrupt line number information when assembling 3081109998Smarkm with debugging information. This is caused by the overlapping 3082109998Smarkm of two sections. 3083109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 3084109998Smarkm 3085109998Smarkm *) NCONF changes. 3086109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 3087109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 3088109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 3089109998Smarkm binary backward compatibility. 3090109998Smarkm Make it possible for methods to load from something other than a BIO, 3091109998Smarkm by providing a function pointer that is given a name instead of a BIO. 3092109998Smarkm For example, this could be used to load configuration data from an 3093109998Smarkm LDAP server. 3094109998Smarkm [Richard Levitte] 3095109998Smarkm 3096109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 3097109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 3098109998Smarkm with non blocking I/O was not possible because no retry code was 3099109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 3100109998Smarkm this case. 3101109998Smarkm [Steve Henson] 3102109998Smarkm 3103109998Smarkm *) Added the beginnings of Rijndael support. 3104109998Smarkm [Ben Laurie] 3105109998Smarkm 3106109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 3107109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 3108109998Smarkm to allow certificate printing to more controllable, additional 3109109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 3110109998Smarkm set. 3111109998Smarkm [Steve Henson] 3112109998Smarkm 3113109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 3114109998Smarkm [Richard Levitte] 3115109998Smarkm 3116142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 3117142425Snectar 3118142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3119160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3120142425Snectar [Joe Orton, Steve Henson] 3121142425Snectar 3122142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 3123142425Snectar 3124142425Snectar *) Fix additional bug revealed by the NISCC test suite: 3125142425Snectar 3126142425Snectar Stop bug triggering large recursion when presented with 3127160814Ssimon certain ASN.1 tags (CVE-2003-0851) 3128142425Snectar [Steve Henson] 3129142425Snectar 3130120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 3131120631Snectar 3132120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3133120631Snectar 3134120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3135160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3136120631Snectar 3137120631Snectar If verify callback ignores invalid public key errors don't try to check 3138120631Snectar certificate signature with the NULL public key. 3139120631Snectar 3140120631Snectar [Steve Henson] 3141120631Snectar 3142120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3143120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3144120631Snectar specifications. 3145120631Snectar [Steve Henson] 3146120631Snectar 3147120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3148120631Snectar extra data after the compression methods not only for TLS 1.0 3149120631Snectar but also for SSL 3.0 (as required by the specification). 3150120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3151120631Snectar 3152120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3153120631Snectar when it's 512 *bits* long, not 512 bytes. 3154120631Snectar [Richard Levitte] 3155120631Snectar 3156120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 3157120631Snectar 3158120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3159120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3160120631Snectar a protocol version number mismatch like a decryption error 3161120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3162120631Snectar [Bodo Moeller] 3163120631Snectar 3164120631Snectar *) Turn on RSA blinding by default in the default implementation 3165120631Snectar to avoid a timing attack. Applications that don't want it can call 3166120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3167120631Snectar They would be ill-advised to do so in most cases. 3168120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3169120631Snectar 3170120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 3171120631Snectar seeded (in this case, the secret RSA exponent is abused as 3172120631Snectar an unpredictable seed -- if it is not unpredictable, there 3173120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 3174120631Snectar by remembering the creator's thread ID in rsa->blinding and 3175120631Snectar having all other threads use local one-time blinding factors 3176120631Snectar (this requires more computation than sharing rsa->blinding, but 3177120631Snectar avoids excessive locking; and if an RSA object is not shared 3178120631Snectar between threads, blinding will still be very fast). 3179120631Snectar [Bodo Moeller] 3180120631Snectar 3181111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 3182111147Snectar 3183111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 3184111147Snectar via timing by performing a MAC computation even if incorrrect 3185111147Snectar block cipher padding has been found. This is a countermeasure 3186111147Snectar against active attacks where the attacker has to distinguish 3187160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 3188111147Snectar 3189111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 3190111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 3191111147Snectar Martin Vuagnoux (EPFL, Ilion)] 3192111147Snectar 3193109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 3194109998Smarkm 3195109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 3196109998Smarkm memory from it's contents. This is done with a counter that will 3197109998Smarkm place alternating values in each byte. This can be used to solve 3198109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 3199109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 3200109998Smarkm be read through on certain media, for example a swap space on disk. 3201109998Smarkm [Geoff Thorpe] 3202109998Smarkm 3203109998Smarkm *) Bugfix: client side session caching did not work with external caching, 3204109998Smarkm because the session->cipher setting was not restored when reloading 3205109998Smarkm from the external cache. This problem was masked, when 3206109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 3207109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 3208109998Smarkm [Lutz Jaenicke] 3209109998Smarkm 3210109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 3211109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 3212109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 3213109998Smarkm 3214109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 3215109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 3216109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 3217109998Smarkm EVP_cleanup(). 3218109998Smarkm [Richard Levitte] 3219109998Smarkm 3220109998Smarkm *) Change the default configuration reader to deal with last line not 3221109998Smarkm being properly terminated. 3222109998Smarkm [Richard Levitte] 3223109998Smarkm 3224109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 3225109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 3226109998Smarkm emailAddress where the value has the type ia5String. 3227109998Smarkm [stefank@valicert.com via Richard Levitte] 3228109998Smarkm 3229109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 3230109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 3231109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 3232109998Smarkm the bitwise-OR of the two for use by the majority of applications 3233109998Smarkm wanting this behaviour, and update the docs. The documented 3234109998Smarkm behaviour and actual behaviour were inconsistent and had been 3235109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 3236109998Smarkm change. 3237109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 3238109998Smarkm 3239109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 3240109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 3241109998Smarkm [Bodo Moeller] 3242109998Smarkm 3243109998Smarkm *) Fix initialization code race conditions in 3244109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 3245109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 3246109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 3247109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 3248109998Smarkm ssl2_get_cipher_by_char(), 3249109998Smarkm ssl3_get_cipher_by_char(). 3250109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 3251109998Smarkm 3252109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 3253109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 3254109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 3255109998Smarkm (see [openssl.org #212]). 3256109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 3257109998Smarkm 3258109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 3259109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 3260109998Smarkm [Steve Henson] 3261109998Smarkm 3262101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 3263101618Snectar 3264101618Snectar *) [In 0.9.6g-engine release:] 3265101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 3266101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 3267101618Snectar 3268101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 3269101613Snectar 3270101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 3271101613Snectar and get fix the header length calculation. 3272101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 3273101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 3274101613Snectar Steve Henson] 3275101613Snectar 3276101613Snectar *) Use proper error handling instead of 'assertions' in buffer 3277101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 3278101613Snectar assertions could call abort()). 3279101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 3280101613Snectar 3281100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 3282100936Snectar 3283109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 3284109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 3285109998Smarkm negative or the content length exceeds the length of the 3286109998Smarkm supplied buffer. 3287109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 3288109998Smarkm 3289100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 3290100936Snectar for the cipher strength set and where therefore not handled correctly 3291100936Snectar by the selection routines (PR #130). 3292100936Snectar [Lutz Jaenicke] 3293100936Snectar 3294100936Snectar *) Fix EVP_dsa_sha macro. 3295100936Snectar [Nils Larsch] 3296100936Snectar 3297100936Snectar *) New option 3298100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 3299100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 3300100936Snectar that was added in OpenSSL 0.9.6d. 3301100936Snectar 3302100936Snectar As the countermeasure turned out to be incompatible with some 3303100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 3304100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 3305100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 3306100936Snectar 's_server'), so the new option is automatically set in many 3307100936Snectar applications. 3308100936Snectar [Bodo Moeller] 3309100936Snectar 3310100936Snectar *) Changes in security patch: 3311100936Snectar 3312100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 3313100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 3314100936Snectar Air Force Materiel Command, USAF, under agreement number 3315100936Snectar F30602-01-2-0537. 3316100936Snectar 3317100936Snectar *) Add various sanity checks to asn1_get_length() to reject 3318100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 3319100936Snectar negative or the content length exceeds the length of the 3320160814Ssimon supplied buffer. (CVE-2002-0659) 3321100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 3322100936Snectar 3323100936Snectar *) Assertions for various potential buffer overflows, not known to 3324100936Snectar happen in practice. 3325100936Snectar [Ben Laurie (CHATS)] 3326100936Snectar 3327100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 3328160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 3329100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 3330100936Snectar 3331100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 3332160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 3333100936Snectar [Ben Laurie (CHATS)] 3334100936Snectar 3335100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 3336160814Ssimon supply an oversized client master key. (CVE-2002-0656) 3337100936Snectar [Ben Laurie (CHATS)] 3338100936Snectar 3339100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 3340100928Snectar 3341100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 3342100928Snectar encoded as NULL) with id-dsa-with-sha1. 3343100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 3344100928Snectar 3345100928Snectar *) Check various X509_...() return values in apps/req.c. 3346100928Snectar [Nils Larsch <nla@trustcenter.de>] 3347100928Snectar 3348100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 3349100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 3350100928Snectar was just at the end of a processed block. The bug was discovered when 3351100928Snectar processing data through a buffering memory BIO handing the data to a 3352100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 3353100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 3354100928Snectar [Lutz Jaenicke] 3355100928Snectar 3356100928Snectar *) Implement a countermeasure against a vulnerability recently found 3357100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 3358100928Snectar before application data chunks to avoid the use of known IVs 3359100928Snectar with data potentially chosen by the attacker. 3360100928Snectar [Bodo Moeller] 3361100928Snectar 3362100928Snectar *) Fix length checks in ssl3_get_client_hello(). 3363100928Snectar [Bodo Moeller] 3364100928Snectar 3365100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 3366100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 3367100928Snectar ssl3_read_bytes() found application data while handshake 3368100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 3369100928Snectar merely automatically cleared during the initial handshake. 3370100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 3371100928Snectar 3372100928Snectar *) Fix object definitions for Private and Enterprise: they were not 3373100928Snectar recognized in their shortname (=lowercase) representation. Extend 3374100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 3375100928Snectar of silently ignoring the problem (Svenning Sorensen 3376100928Snectar <sss@sss.dnsalias.net>). 3377100928Snectar [Lutz Jaenicke] 3378100928Snectar 3379100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 3380100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 3381100928Snectar code did not properly initialise the 'add' and 'rem' values to 3382100928Snectar BN_generate_prime().) 3383100928Snectar 3384100928Snectar In the new general case, we do not insist that 'generator' is 3385100928Snectar actually a primitive root: This requirement is rather pointless; 3386100928Snectar a generator of the order-q subgroup is just as good, if not 3387100928Snectar better. 3388100928Snectar [Bodo Moeller] 3389100928Snectar 3390100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 3391100928Snectar Tom Wu <tom@arcot.com>. 3392100928Snectar [Lutz Jaenicke] 3393100928Snectar 3394100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 3395100928Snectar returning non-zero before the data has been completely received 3396100928Snectar when using non-blocking I/O. 3397100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 3398100928Snectar 3399100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 3400100928Snectar [Ben Laurie, Lutz Jaenicke] 3401100928Snectar 3402100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 3403100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 3404100928Snectar [Lutz Jaenicke] 3405100928Snectar 3406100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 3407100928Snectar configuration for the versions before that. 3408100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 3409100928Snectar 3410100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 3411100928Snectar check whether we deal with a copy of a session and do not delete from 3412100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 3413100928Snectar <izhar@checkpoint.com>. 3414100928Snectar [Lutz Jaenicke] 3415100928Snectar 3416100928Snectar *) Do not store session data into the internal session cache, if it 3417100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 3418100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 3419100928Snectar [Lutz Jaenicke] 3420100928Snectar 3421100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 3422100928Snectar value is 0. 3423100928Snectar [Richard Levitte] 3424100928Snectar 3425109998Smarkm *) [In 0.9.6d-engine release:] 3426109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 3427100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 3428100928Snectar 3429100928Snectar *) Add the configuration target linux-s390x. 3430100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 3431100928Snectar 3432100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 3433100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 3434100928Snectar variable as an indication that a ClientHello message has been 3435100928Snectar received. As the flag value will be lost between multiple 3436100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 3437100928Snectar function may not be aware that a handshake has actually taken 3438100928Snectar place, thus preventing a new session from being added to the 3439100928Snectar session cache. 3440100928Snectar 3441100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 3442100928Snectar using a local variable. 3443100928Snectar [Lutz Jaenicke, Bodo Moeller] 3444100928Snectar 3445100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 3446100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 3447100928Snectar [Geoff Thorpe, Bodo Moeller] 3448100928Snectar 3449100928Snectar *) New 'shared_ldflag' column in Configure platform table. 3450100928Snectar [Richard Levitte] 3451100928Snectar 3452100928Snectar *) Fix EVP_CIPHER_mode macro. 3453100928Snectar ["Dan S. Camper" <dan@bti.net>] 3454100928Snectar 3455100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 3456100928Snectar type, we must throw them away by setting rr->length to 0. 3457100928Snectar [D P Chang <dpc@qualys.com>] 3458100928Snectar 345989837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 346089837Skris 346189837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 346289837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 346389837Skris worked incorrectly for those cases where range = 10..._2 and 346489837Skris 3*range is two bits longer than range.) 346589837Skris [Bodo Moeller] 346689837Skris 346789837Skris *) Only add signing time to PKCS7 structures if it is not already 346889837Skris present. 346989837Skris [Steve Henson] 347089837Skris 347189837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 347289837Skris OBJ_ld_ce should be OBJ_id_ce. 347389837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 347489837Skris incorrect (cf. RFC 3039). 347589837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 347689837Skris 347789837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 347889837Skris returns early because it has nothing to do. 347989837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 348089837Skris 348189837Skris *) [In 0.9.6c-engine release:] 348289837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 348389837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 348489837Skris 348589837Skris *) [In 0.9.6c-engine release:] 348689837Skris Add support for Cryptographic Appliance's keyserver technology. 348789837Skris (Use engine 'keyclient') 348889837Skris [Cryptographic Appliances and Geoff Thorpe] 348989837Skris 349089837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 349189837Skris is called via tools/c89.sh because arguments have to be 349289837Skris rearranged (all '-L' options must appear before the first object 349389837Skris modules). 349489837Skris [Richard Shapiro <rshapiro@abinitio.com>] 349589837Skris 349689837Skris *) [In 0.9.6c-engine release:] 349789837Skris Add support for Broadcom crypto accelerator cards, backported 349889837Skris from 0.9.7. 349989837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 350089837Skris 350189837Skris *) [In 0.9.6c-engine release:] 350289837Skris Add support for SureWare crypto accelerator cards from 350389837Skris Baltimore Technologies. (Use engine 'sureware') 350489837Skris [Baltimore Technologies and Mark Cox] 350589837Skris 350689837Skris *) [In 0.9.6c-engine release:] 350789837Skris Add support for crypto accelerator cards from Accelerated 350889837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 350989837Skris [AEP Inc. and Mark Cox] 351089837Skris 351189837Skris *) Add a configuration entry for gcc on UnixWare. 351289837Skris [Gary Benson <gbenson@redhat.com>] 351389837Skris 351489837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 351589837Skris messages are stored in a single piece (fixed-length part and 351689837Skris variable-length part combined) and fix various bugs found on the way. 351789837Skris [Bodo Moeller] 351889837Skris 351989837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 352089837Skris instead. BIO_gethostbyname() does not know what timeouts are 352189837Skris appropriate, so entries would stay in cache even when they have 352289837Skris become invalid. 352389837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 352489837Skris 352589837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 352689837Skris faced with a pathologically small ClientHello fragment that does 352789837Skris not contain client_version: Instead of aborting with an error, 352889837Skris simply choose the highest available protocol version (i.e., 352989837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 353089837Skris messages are never sent like this, but this change gives us 353189837Skris strictly correct behaviour at least for TLS. 353289837Skris [Bodo Moeller] 353389837Skris 353489837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 353589837Skris never resets s->method to s->ctx->method when called from within 353689837Skris one of the SSL handshake functions. 353789837Skris [Bodo Moeller; problem pointed out by Niko Baric] 353889837Skris 353989837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 354089837Skris (sent using the client's version number) if client_version is 354189837Skris smaller than the protocol version in use. Also change 354289837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 354389837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 354489837Skris the client will at least see that alert. 354589837Skris [Bodo Moeller] 354689837Skris 354789837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 354889837Skris correctly. 354989837Skris [Bodo Moeller] 355089837Skris 355189837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 355289837Skris client receives HelloRequest while in a handshake. 355389837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 355489837Skris 355589837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 355689837Skris should end in 'break', not 'goto end' which circuments various 355789837Skris cleanups done in state SSL_ST_OK. But session related stuff 355889837Skris must be disabled for SSL_ST_OK in the case that we just sent a 355989837Skris HelloRequest. 356089837Skris 356189837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 356289837Skris before just sending a HelloRequest. 356389837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 356489837Skris 356589837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 356689837Skris reveal whether illegal block cipher padding was found or a MAC 356789837Skris verification error occured. (Neither SSLerr() codes nor alerts 356889837Skris are directly visible to potential attackers, but the information 356989837Skris may leak via logfiles.) 357089837Skris 357189837Skris Similar changes are not required for the SSL 2.0 implementation 357289837Skris because the number of padding bytes is sent in clear for SSL 2.0, 357389837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 357489837Skris failed to verify that the purported number of padding bytes is in 357589837Skris the legal range. 357689837Skris [Bodo Moeller] 357789837Skris 357889837Skris *) Add OpenUNIX-8 support including shared libraries 357989837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 358089837Skris [Lutz Jaenicke] 358189837Skris 358289837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 358389837Skris 'wristwatch attack' using huge encoding parameters (cf. 358489837Skris James H. Manger's CRYPTO 2001 paper). Note that the 358589837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 358689837Skris encoding parameters and hence was not vulnerable. 358789837Skris [Bodo Moeller] 358889837Skris 358989837Skris *) BN_sqr() bug fix. 359089837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 359189837Skris 359289837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 359389837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 359489837Skris followed by modular reduction. 359589837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 359689837Skris 359789837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 359889837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 359989837Skris [Bodo Moeller] 360089837Skris 360189837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 360289837Skris This function was broken, as the check for a new client hello message 360389837Skris to handle SGC did not allow these large messages. 360489837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 360589837Skris [Lutz Jaenicke] 360689837Skris 360789837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 360889837Skris [Lutz Jaenicke] 360989837Skris 361089837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 361189837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 361289837Skris [Lutz Jaenicke] 361389837Skris 361489837Skris *) Rework the configuration and shared library support for Tru64 Unix. 361589837Skris The configuration part makes use of modern compiler features and 361689837Skris still retains old compiler behavior for those that run older versions 361789837Skris of the OS. The shared library support part includes a variant that 361889837Skris uses the RPATH feature, and is available through the special 361989837Skris configuration target "alpha-cc-rpath", which will never be selected 362089837Skris automatically. 362189837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 362289837Skris 362389837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 362489837Skris with the same message size as in ssl3_get_certificate_request(). 362589837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 362689837Skris messages might inadvertently be reject as too long. 362789837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 362889837Skris 362989837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 363089837Skris [Andy Polyakov] 363189837Skris 363289837Skris *) Modified SSL library such that the verify_callback that has been set 363389837Skris specificly for an SSL object with SSL_set_verify() is actually being 363489837Skris used. Before the change, a verify_callback set with this function was 363589837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 363689837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 363789837Skris to allow the necessary settings. 363889837Skris [Lutz Jaenicke] 363989837Skris 364089837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 364189837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 364289837Skris done automatically (in contradiction to the requirements of the C 364389837Skris standard). This made problems when used from OpenSSH. 364489837Skris [Lutz Jaenicke] 364589837Skris 364689837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 364789837Skris dh->length and always used 364889837Skris 364989837Skris BN_rand_range(priv_key, dh->p). 365089837Skris 365189837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 365289837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 365389837Skris dh->length (recommended exponent length) is much smaller than the 365489837Skris length of dh->p. We could use BN_rand_range() if the order of 365589837Skris the subgroup was stored in the DH structure, but we only have 365689837Skris dh->length. 365789837Skris 365889837Skris So switch back to 365989837Skris 366089837Skris BN_rand(priv_key, l, ...) 366189837Skris 366289837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 366389837Skris otherwise. 366489837Skris [Bodo Moeller] 366589837Skris 366689837Skris *) In 366789837Skris 366889837Skris RSA_eay_public_encrypt 366989837Skris RSA_eay_private_decrypt 367089837Skris RSA_eay_private_encrypt (signing) 367189837Skris RSA_eay_public_decrypt (signature verification) 367289837Skris 367389837Skris (default implementations for RSA_public_encrypt, 367489837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 367589837Skris always reject numbers >= n. 367689837Skris [Bodo Moeller] 367789837Skris 367889837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 367989837Skris to synchronize access to 'locking_thread'. This is necessary on 368089837Skris systems where access to 'locking_thread' (an 'unsigned long' 368189837Skris variable) is not atomic. 368289837Skris [Bodo Moeller] 368389837Skris 368489837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 368589837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 368689837Skris a race condition if 0 is a valid thread ID. 368789837Skris [Travis Vitek <vitek@roguewave.com>] 368889837Skris 368989837Skris *) Add support for shared libraries under Irix. 369089837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 369189837Skris 369289837Skris *) Add configuration option to build on Linux on both big-endian and 369389837Skris little-endian MIPS. 369489837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 369589837Skris 369689837Skris *) Add the possibility to create shared libraries on HP-UX. 369789837Skris [Richard Levitte] 369889837Skris 369979998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 370079998Skris 370179998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 370279998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 370379998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 370479998Skris PRNG state recovery was possible based on the output of 370579998Skris one PRNG request appropriately sized to gain knowledge on 370679998Skris 'md' followed by enough consecutive 1-byte PRNG requests 370779998Skris to traverse all of 'state'. 370879998Skris 370979998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 371079998Skris during PRNG output generation, hash all of the previous 371179998Skris 'md_local' value, not just the half used for PRNG output. 371279998Skris 371379998Skris 2. Make the number of bytes from 'state' included into the hash 371479998Skris independent from the number of PRNG bytes requested. 371579998Skris 371679998Skris The first measure alone would be sufficient to avoid 371779998Skris Markku-Juhani's attack. (Actually it had never occurred 371879998Skris to me that the half of 'md_local' used for chaining was the 371979998Skris half from which PRNG output bytes were taken -- I had always 372079998Skris assumed that the secret half would be used.) The second 372179998Skris measure makes sure that additional data from 'state' is never 372279998Skris mixed into 'md_local' in small portions; this heuristically 372379998Skris further strengthens the PRNG. 372479998Skris [Bodo Moeller] 372579998Skris 372679998Skris *) Fix crypto/bn/asm/mips3.s. 372779998Skris [Andy Polyakov] 372879998Skris 372979998Skris *) When only the key is given to "enc", the IV is undefined. Print out 373079998Skris an error message in this case. 373179998Skris [Lutz Jaenicke] 373279998Skris 373379998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 373479998Skris [Steve Henson] 373579998Skris 373679998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 373779998Skris positive and less than q. 373879998Skris [Bodo Moeller] 373979998Skris 374079998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 374179998Skris used: it isn't thread safe and the add_lock_callback should handle 374279998Skris that itself. 374379998Skris [Paul Rose <Paul.Rose@bridge.com>] 374479998Skris 374579998Skris *) Verify that incoming data obeys the block size in 374679998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 374779998Skris [Bodo Moeller] 374879998Skris 374979998Skris *) Fix OAEP check. 375079998Skris [Ulf M�ller, Bodo M�ller] 375179998Skris 375279998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 375389837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 375479998Skris when fixing the server behaviour for backwards-compatible 'client 375579998Skris hello' messages. (Note that the attack is impractical against 375679998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 375779998Skris means that the probability of guessing a valid ciphertext is 375879998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 375979998Skris paper.) 376079998Skris 376179998Skris Before 0.9.5, the countermeasure (hide the error by generating a 376279998Skris random 'decryption result') did not work properly because 376379998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 376479998Skris detect the supposedly ignored error. 376579998Skris 376679998Skris Both problems are now fixed. 376779998Skris [Bodo Moeller] 376879998Skris 376979998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 377079998Skris (previously it was 1024). 377179998Skris [Bodo Moeller] 377279998Skris 377379998Skris *) Fix for compatibility mode trust settings: ignore trust settings 377479998Skris unless some valid trust or reject settings are present. 377579998Skris [Steve Henson] 377679998Skris 377779998Skris *) Fix for blowfish EVP: its a variable length cipher. 377879998Skris [Steve Henson] 377979998Skris 378079998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 378179998Skris parameters in DSA public key structures and return an error in the 378279998Skris DSA routines if parameters are absent. 378379998Skris [Steve Henson] 378479998Skris 378579998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 378679998Skris in the current directory if neither $RANDFILE nor $HOME was set. 378779998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 378879998Skris caused some confusion to Windows users who haven't defined $HOME. 378979998Skris Thus RAND_file_name() is changed again: e_os.h can define a 379079998Skris DEFAULT_HOME, which will be used if $HOME is not set. 379179998Skris For Windows, we use "C:"; on other platforms, we still require 379279998Skris environment variables. 379379998Skris 379479998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 379579998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 379679998Skris having multiple threads call RAND_poll() concurrently. 379779998Skris [Bodo Moeller] 379879998Skris 379979998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 380079998Skris combination of a flag and a thread ID variable. 380179998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 380279998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 380389837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 380479998Skris that they do not hold after the first thread unsets add_do_not_lock). 380579998Skris [Bodo Moeller] 380679998Skris 380779998Skris *) Change bctest again: '-x' expressions are not available in all 380879998Skris versions of 'test'. 380979998Skris [Bodo Moeller] 381079998Skris 381176866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 381272613Skris 381376866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 381476866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 381576866Skris 381676866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 381776866Skris the default extension for executables, if any. Also, make the perl 381876866Skris scripts that use symlink() to test if it really exists and use "cp" 381976866Skris if it doesn't. All this made OpenSSL compilable and installable in 382076866Skris CygWin. 382176866Skris [Richard Levitte] 382276866Skris 382376866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 382476866Skris If SEQUENCE is length is indefinite just set c->slen to the total 382576866Skris amount of data available. 382676866Skris [Steve Henson, reported by shige@FreeBSD.org] 382776866Skris [This change does not apply to 0.9.7.] 382876866Skris 382976866Skris *) Change bctest to avoid here-documents inside command substitution 383076866Skris (workaround for FreeBSD /bin/sh bug). 383176866Skris For compatibility with Ultrix, avoid shell functions (introduced 383276866Skris in the bctest version that searches along $PATH). 383376866Skris [Bodo Moeller] 383476866Skris 383576866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 383676866Skris with des_encrypt() defined on some operating systems, like Solaris 383776866Skris and UnixWare. 383876866Skris [Richard Levitte] 383976866Skris 384076866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 384176866Skris On the Importance of Eliminating Errors in Cryptographic 384276866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 384376866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 384476866Skris [Ulf Moeller] 384576866Skris 384676866Skris *) MIPS assembler BIGNUM division bug fix. 384776866Skris [Andy Polyakov] 384876866Skris 384976866Skris *) Disabled incorrect Alpha assembler code. 385076866Skris [Richard Levitte] 385176866Skris 385276866Skris *) Fix PKCS#7 decode routines so they correctly update the length 385376866Skris after reading an EOC for the EXPLICIT tag. 385476866Skris [Steve Henson] 385576866Skris [This change does not apply to 0.9.7.] 385676866Skris 385776866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 385876866Skris if a 3DES key was generated with a 0 initial byte. Include 385976866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 386076866Skris (but broken) behaviour. 386176866Skris [Steve Henson] 386276866Skris 386376866Skris *) Enhance bctest to search for a working bc along $PATH and print 386476866Skris it when found. 386576866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 386676866Skris 386776866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 386876866Skris don't write to the wrong index in ERR_set_error_data. 386976866Skris [Bodo Moeller] 387076866Skris 387176866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 387276866Skris did not exist. 387376866Skris [Bodo Moeller] 387476866Skris 387576866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 387676866Skris [Jeremy Cooper <jeremy@baymoo.org>] 387776866Skris 387876866Skris *) Make it possible to reuse SSLv2 sessions. 387976866Skris [Richard Levitte] 388076866Skris 388176866Skris *) In copy_email() check for >= 0 as a return value for 388276866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 388376866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 388476866Skris 388576866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 388676866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 388776866Skris PKCS7_verify() fails with non detached data. 388876866Skris [Steve Henson] 388976866Skris 389076866Skris *) Don't use getenv in library functions when run as setuid/setgid. 389176866Skris New function OPENSSL_issetugid(). 389276866Skris [Ulf Moeller] 389376866Skris 389476866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 389576866Skris due to incorrect handling of multi-threading: 389676866Skris 389776866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 389876866Skris 389976866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 390076866Skris 390176866Skris 3. Count how many times MemCheck_off() has been called so that 390276866Skris nested use can be treated correctly. This also avoids 390376866Skris inband-signalling in the previous code (which relied on the 390476866Skris assumption that thread ID 0 is impossible). 390576866Skris [Bodo Moeller] 390676866Skris 390776866Skris *) Add "-rand" option also to s_client and s_server. 390876866Skris [Lutz Jaenicke] 390976866Skris 391076866Skris *) Fix CPU detection on Irix 6.x. 391176866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 391276866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 391376866Skris 391476866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 391576866Skris was empty. 391676866Skris [Steve Henson] 391776866Skris [This change does not apply to 0.9.7.] 391876866Skris 391976866Skris *) Use the cached encoding of an X509_NAME structure rather than 392076866Skris copying it. This is apparently the reason for the libsafe "errors" 392176866Skris but the code is actually correct. 392276866Skris [Steve Henson] 392376866Skris 392472613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 392572613Skris Bleichenbacher's DSA attack. 392676866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 392776866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 392876866Skris and leaves the highest bit random. 392976866Skris [Ulf Moeller, Bodo Moeller] 393072613Skris 393172613Skris *) In the NCONF_...-based implementations for CONF_... queries 393272613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 393372613Skris a temporary CONF structure with the data component set to NULL 393472613Skris (which gives segmentation faults in lh_retrieve). 393572613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 393672613Skris CONF_get_number (which may use environment variables) and directly 393772613Skris return NULL from CONF_get_section. 393872613Skris [Bodo Moeller] 393972613Skris 394072613Skris *) Fix potential buffer overrun for EBCDIC. 394172613Skris [Ulf Moeller] 394272613Skris 394372613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 394472613Skris keyUsage if basicConstraints absent for a CA. 394572613Skris [Steve Henson] 394672613Skris 394772613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 394872613Skris is more generally accepted (no spaces before the semicolon), since 394972613Skris some programs can't parse those values properly otherwise. Also make 395072613Skris sure BIO's that break lines after each write do not create invalid 395172613Skris headers. 395272613Skris [Richard Levitte] 395372613Skris 395472613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 395572613Skris macros previously used would not encode an empty SEQUENCE OF 395672613Skris and break the signature. 395772613Skris [Steve Henson] 395876866Skris [This change does not apply to 0.9.7.] 395972613Skris 396072613Skris *) Zero the premaster secret after deriving the master secret in 396172613Skris DH ciphersuites. 396272613Skris [Steve Henson] 396372613Skris 396472613Skris *) Add some EVP_add_digest_alias registrations (as found in 396572613Skris OpenSSL_add_all_digests()) to SSL_library_init() 396672613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 396772613Skris compatibility with peers using X.509 certificates 396872613Skris with unconventional AlgorithmIdentifier OIDs. 396972613Skris [Bodo Moeller] 397072613Skris 397172613Skris *) Fix for Irix with NO_ASM. 397272613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 397372613Skris 397472613Skris *) ./config script fixes. 397572613Skris [Ulf Moeller, Richard Levitte] 397672613Skris 397772613Skris *) Fix 'openssl passwd -1'. 397872613Skris [Bodo Moeller] 397972613Skris 398072613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 398172613Skris terminated strings whose length is passed in the passlen 398272613Skris parameter, for example from PEM callbacks. This was done 398372613Skris by adding an extra length parameter to asc2uni(). 398472613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 398572613Skris 398672613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 398772613Skris call failed, free the DSA structure. 398872613Skris [Bodo Moeller] 398972613Skris 399072613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 399172613Skris These are present in some PKCS#12 files. 399272613Skris [Steve Henson] 399372613Skris 399472613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 399572613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 399672613Skris when writing a 32767 byte record. 399772613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 399872613Skris 399972613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 400072613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 400172613Skris 400272613Skris (RSA objects have a reference count access to which is protected 400372613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 400472613Skris so they are meant to be shared between threads.) 400572613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 400672613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 400772613Skris 400872613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 400972613Skris [Bodo Moeller] 401072613Skris 401176866Skris *) Use better test patterns in bntest. 401276866Skris [Ulf M�ller] 401376866Skris 401472613Skris *) rand_win.c fix for Borland C. 401572613Skris [Ulf M�ller] 401672613Skris 401772613Skris *) BN_rshift bugfix for n == 0. 401872613Skris [Bodo Moeller] 401972613Skris 402076866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 402176866Skris so that 'make test' does not abort just because 'bc' is broken. 402276866Skris [Bodo Moeller] 402376866Skris 402472613Skris *) Store verify_result within SSL_SESSION also for client side to 402572613Skris avoid potential security hole. (Re-used sessions on the client side 402672613Skris always resulted in verify_result==X509_V_OK, not using the original 402772613Skris result of the server certificate verification.) 402872613Skris [Lutz Jaenicke] 402972613Skris 403072613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 403172613Skris SSL3_RT_APPLICATION_DATA, return 0. 403272613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 403372613Skris [Bodo Moeller] 403472613Skris 403572613Skris *) Fix SSL_peek: 403672613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 403772613Skris releases, have been re-implemented by renaming the previous 403872613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 403972613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 404072613Skris to them. The new ssl[23]_{read,peek} functions are calls to 404172613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 404272613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 404372613Skris does the actual work for ssl3_read_internal. 404472613Skris [Bodo Moeller] 404572613Skris 404676866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 404776866Skris the method-specific "init()" handler. Also clean up ex_data after 404876866Skris calling the method-specific "finish()" handler. Previously, this was 404976866Skris happening the other way round. 405076866Skris [Geoff Thorpe] 405176866Skris 405272613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 405372613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 405472613Skris [Bodo Moeller] 405572613Skris 405676866Skris *) Make sure that shared libraries get the internal name engine with 405776866Skris the full version number and not just 0. This should mark the 405876866Skris shared libraries as not backward compatible. Of course, this should 405976866Skris be changed again when we can guarantee backward binary compatibility. 406076866Skris [Richard Levitte] 406176866Skris 406272613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 406372613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 406472613Skris 406576866Skris *) Rework the system to generate shared libraries: 406676866Skris 406776866Skris - Make note of the expected extension for the shared libraries and 406876866Skris if there is a need for symbolic links from for example libcrypto.so.0 406976866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 407076866Skris that. 407176866Skris 407276866Skris - Make as few rebuilds of the shared libraries as possible. 407376866Skris 407476866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 407576866Skris 407676866Skris - When installing, install the shared libraries separately from the 407776866Skris static ones. 407876866Skris [Richard Levitte] 407976866Skris 408072613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 408172613Skris 408272613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 408372613Skris and not in SSL_clear because the latter is also used by the 408472613Skris accept/connect functions; previously, the settings made by 408572613Skris SSL_set_read_ahead would be lost during the handshake. 408672613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 408772613Skris 408872613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 408972613Skris Previously, it would create entries for disableed algorithms no 409072613Skris matter what. 409172613Skris [Richard Levitte] 409272613Skris 409372613Skris *) Added several new manual pages for SSL_* function. 409472613Skris [Lutz Jaenicke] 409572613Skris 409668651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 409768651Skris 409868651Skris *) In ssl23_get_client_hello, generate an error message when faced 409968651Skris with an initial SSL 3.0/TLS record that is too small to contain the 410068651Skris first two bytes of the ClientHello message, i.e. client_version. 410168651Skris (Note that this is a pathologic case that probably has never happened 410268651Skris in real life.) The previous approach was to use the version number 410368651Skris from the record header as a substitute; but our protocol choice 410468651Skris should not depend on that one because it is not authenticated 410568651Skris by the Finished messages. 410668651Skris [Bodo Moeller] 410768651Skris 410868651Skris *) More robust randomness gathering functions for Windows. 410968651Skris [Jeffrey Altman <jaltman@columbia.edu>] 411068651Skris 411168651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 411268651Skris not set then we don't setup the error code for issuer check errors 411368651Skris to avoid possibly overwriting other errors which the callback does 411468651Skris handle. If an application does set the flag then we assume it knows 411568651Skris what it is doing and can handle the new informational codes 411668651Skris appropriately. 411768651Skris [Steve Henson] 411868651Skris 411968651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 412068651Skris a general "ANY" type, as such it should be able to decode anything 412168651Skris including tagged types. However it didn't check the class so it would 412268651Skris wrongly interpret tagged types in the same way as their universal 412368651Skris counterpart and unknown types were just rejected. Changed so that the 412468651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 412568651Skris that is the encoding is stored intact. There is also a new type 412668651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 412768651Skris case we have no idea what the actual type is so we just lump them all 412868651Skris together. 412968651Skris [Steve Henson] 413068651Skris 413168651Skris *) On VMS, stdout may very well lead to a file that is written to 413268651Skris in a record-oriented fashion. That means that every write() will 413368651Skris write a separate record, which will be read separately by the 413468651Skris programs trying to read from it. This can be very confusing. 413568651Skris 413668651Skris The solution is to put a BIO filter in the way that will buffer 413768651Skris text until a linefeed is reached, and then write everything a 413868651Skris line at a time, so every record written will be an actual line, 413968651Skris not chunks of lines and not (usually doesn't happen, but I've 414068651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 414168651Skris the answer. 414268651Skris 414368651Skris Currently, it's a VMS-only method, because that's where it has 414468651Skris been tested well enough. 414568651Skris [Richard Levitte] 414668651Skris 414768651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 414868651Skris it can return incorrect results. 414968651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 415068651Skris but it was in 0.9.6-beta[12].) 415168651Skris [Bodo Moeller] 415268651Skris 415368651Skris *) Disable the check for content being present when verifying detached 415468651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 415568651Skris include zero length content when signing messages. 415668651Skris [Steve Henson] 415768651Skris 415868651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 415968651Skris BIO_ctrl (for BIO pairs). 416068651Skris [Bodo M�ller] 416168651Skris 416268651Skris *) Add DSO method for VMS. 416368651Skris [Richard Levitte] 416468651Skris 416568651Skris *) Bug fix: Montgomery multiplication could produce results with the 416668651Skris wrong sign. 416768651Skris [Ulf M�ller] 416868651Skris 416968651Skris *) Add RPM specification openssl.spec and modify it to build three 417068651Skris packages. The default package contains applications, application 417168651Skris documentation and run-time libraries. The devel package contains 417268651Skris include files, static libraries and function documentation. The 417368651Skris doc package contains the contents of the doc directory. The original 417468651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 417568651Skris [Richard Levitte] 417668651Skris 417768651Skris *) Add a large number of documentation files for many SSL routines. 417868651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 417968651Skris 418068651Skris *) Add a configuration entry for Sony News 4. 418168651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 418268651Skris 418368651Skris *) Don't set the two most significant bits to one when generating a 418468651Skris random number < q in the DSA library. 418568651Skris [Ulf M�ller] 418668651Skris 418768651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 418868651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 418968651Skris the underlying transport is blocking) if a handshake took place. 419068651Skris (The default behaviour is needed by applications such as s_client 419168651Skris and s_server that use select() to determine when to use SSL_read; 419268651Skris but for applications that know in advance when to expect data, it 419368651Skris just makes things more complicated.) 419468651Skris [Bodo Moeller] 419568651Skris 419668651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 419768651Skris from EGD. 419868651Skris [Ben Laurie] 419968651Skris 420068651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 420168651Skris work better on such systems. 420268651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 420368651Skris 420468651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 420568651Skris Update PKCS12_parse() so it copies the friendlyName and the 420668651Skris keyid to the certificates aux info. 420768651Skris [Steve Henson] 420868651Skris 420968651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 421068651Skris if there was more than one signature. 421168651Skris [Sven Uszpelkat <su@celocom.de>] 421268651Skris 421368651Skris *) Major change in util/mkdef.pl to include extra information 421468651Skris about each symbol, as well as presentig variables as well 421568651Skris as functions. This change means that there's n more need 421668651Skris to rebuild the .num files when some algorithms are excluded. 421768651Skris [Richard Levitte] 421868651Skris 421968651Skris *) Allow the verify time to be set by an application, 422068651Skris rather than always using the current time. 422168651Skris [Steve Henson] 422268651Skris 422368651Skris *) Phase 2 verify code reorganisation. The certificate 422468651Skris verify code now looks up an issuer certificate by a 422568651Skris number of criteria: subject name, authority key id 422668651Skris and key usage. It also verifies self signed certificates 422768651Skris by the same criteria. The main comparison function is 422868651Skris X509_check_issued() which performs these checks. 422968651Skris 423068651Skris Lot of changes were necessary in order to support this 423168651Skris without completely rewriting the lookup code. 423268651Skris 423368651Skris Authority and subject key identifier are now cached. 423468651Skris 423568651Skris The LHASH 'certs' is X509_STORE has now been replaced 423668651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 423768651Skris LHASH can't store or retrieve multiple objects with 423868651Skris the same hash value. 423968651Skris 424068651Skris As a result various functions (which were all internal 424168651Skris use only) have changed to handle the new X509_STORE 424268651Skris structure. This will break anything that messed round 424368651Skris with X509_STORE internally. 424468651Skris 424568651Skris The functions X509_STORE_add_cert() now checks for an 424668651Skris exact match, rather than just subject name. 424768651Skris 424868651Skris The X509_STORE API doesn't directly support the retrieval 424968651Skris of multiple certificates matching a given criteria, however 425068651Skris this can be worked round by performing a lookup first 425168651Skris (which will fill the cache with candidate certificates) 425268651Skris and then examining the cache for matches. This is probably 425368651Skris the best we can do without throwing out X509_LOOKUP 425468651Skris entirely (maybe later...). 425568651Skris 425668651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 425768651Skris 425868651Skris All certificate lookup operations now go via a get_issuer() 425968651Skris callback. Although this currently uses an X509_STORE it 426068651Skris can be replaced by custom lookups. This is a simple way 426168651Skris to bypass the X509_STORE hackery necessary to make this 426268651Skris work and makes it possible to use more efficient techniques 426368651Skris in future. A very simple version which uses a simple 426468651Skris STACK for its trusted certificate store is also provided 426568651Skris using X509_STORE_CTX_trusted_stack(). 426668651Skris 426768651Skris The verify_cb() and verify() callbacks now have equivalents 426868651Skris in the X509_STORE_CTX structure. 426968651Skris 427068651Skris X509_STORE_CTX also has a 'flags' field which can be used 427168651Skris to customise the verify behaviour. 427268651Skris [Steve Henson] 427368651Skris 427468651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 427568651Skris excludes S/MIME capabilities. 427668651Skris [Steve Henson] 427768651Skris 427868651Skris *) When a certificate request is read in keep a copy of the 427968651Skris original encoding of the signed data and use it when outputing 428068651Skris again. Signatures then use the original encoding rather than 428168651Skris a decoded, encoded version which may cause problems if the 428268651Skris request is improperly encoded. 428368651Skris [Steve Henson] 428468651Skris 428568651Skris *) For consistency with other BIO_puts implementations, call 428668651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 428768651Skris BIO_write(b, ...). 428868651Skris 428968651Skris In BIO_puts, increment b->num_write as in BIO_write. 429068651Skris [Peter.Sylvester@EdelWeb.fr] 429168651Skris 429268651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 429368651Skris BN_zero, we may not return a BIGNUM with an array consisting of 429468651Skris words set to zero.) 429568651Skris [Bodo Moeller] 429668651Skris 429768651Skris *) Avoid calling abort() from within the library when problems are 429868651Skris detected, except if preprocessor symbols have been defined 429968651Skris (such as REF_CHECK, BN_DEBUG etc.). 430068651Skris [Bodo Moeller] 430168651Skris 430268651Skris *) New openssl application 'rsautl'. This utility can be 430368651Skris used for low level RSA operations. DER public key 430468651Skris BIO/fp routines also added. 430568651Skris [Steve Henson] 430668651Skris 430768651Skris *) New Configure entry and patches for compiling on QNX 4. 430868651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 430968651Skris 431068651Skris *) A demo state-machine implementation was sponsored by 431168651Skris Nuron (http://www.nuron.com/) and is now available in 431268651Skris demos/state_machine. 431368651Skris [Ben Laurie] 431468651Skris 431568651Skris *) New options added to the 'dgst' utility for signature 431668651Skris generation and verification. 431768651Skris [Steve Henson] 431868651Skris 431968651Skris *) Unrecognized PKCS#7 content types are now handled via a 432068651Skris catch all ASN1_TYPE structure. This allows unsupported 432168651Skris types to be stored as a "blob" and an application can 432268651Skris encode and decode it manually. 432368651Skris [Steve Henson] 432468651Skris 432568651Skris *) Fix various signed/unsigned issues to make a_strex.c 432668651Skris compile under VC++. 432768651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 432868651Skris 432968651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 433068651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 433168651Skris if passed a NULL BN and its argument was negative. 433268651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 433368651Skris 433468651Skris *) Modification to PKCS#7 encoding routines to output definite 433568651Skris length encoding. Since currently the whole structures are in 433668651Skris memory there's not real point in using indefinite length 433768651Skris constructed encoding. However if OpenSSL is compiled with 433868651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 433968651Skris [Steve Henson] 434068651Skris 434168651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 434268651Skris [Richard Levitte] 434368651Skris 434468651Skris *) Added more prefixes to parse for in the the strings written 434568651Skris through a logging bio, to cover all the levels that are available 434668651Skris through syslog. The prefixes are now: 434768651Skris 434868651Skris PANIC, EMERG, EMR => LOG_EMERG 434968651Skris ALERT, ALR => LOG_ALERT 435068651Skris CRIT, CRI => LOG_CRIT 435168651Skris ERROR, ERR => LOG_ERR 435268651Skris WARNING, WARN, WAR => LOG_WARNING 435368651Skris NOTICE, NOTE, NOT => LOG_NOTICE 435468651Skris INFO, INF => LOG_INFO 435568651Skris DEBUG, DBG => LOG_DEBUG 435668651Skris 435768651Skris and as before, if none of those prefixes are present at the 435868651Skris beginning of the string, LOG_ERR is chosen. 435968651Skris 436068651Skris On Win32, the LOG_* levels are mapped according to this: 436168651Skris 436268651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 436368651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 436468651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 436568651Skris 436668651Skris [Richard Levitte] 436768651Skris 436868651Skris *) Made it possible to reconfigure with just the configuration 436968651Skris argument "reconf" or "reconfigure". The command line arguments 437068651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 437168651Skris and are retrieved from there when reconfiguring. 437268651Skris [Richard Levitte] 437368651Skris 437468651Skris *) MD4 implemented. 437568651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 437668651Skris 437768651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 437868651Skris [Richard Levitte] 437968651Skris 438068651Skris *) The obj_dat.pl script was messing up the sorting of object 438168651Skris names. The reason was that it compared the quoted version 438268651Skris of strings as a result "OCSP" > "OCSP Signing" because 438368651Skris " > SPACE. Changed script to store unquoted versions of 438468651Skris names and add quotes on output. It was also omitting some 438568651Skris names from the lookup table if they were given a default 438668651Skris value (that is if SN is missing it is given the same 438768651Skris value as LN and vice versa), these are now added on the 438868651Skris grounds that if an object has a name we should be able to 438968651Skris look it up. Finally added warning output when duplicate 439068651Skris short or long names are found. 439168651Skris [Steve Henson] 439268651Skris 439368651Skris *) Changes needed for Tandem NSK. 439468651Skris [Scott Uroff <scott@xypro.com>] 439568651Skris 439668651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 439768651Skris RSA_padding_check_SSLv23(), special padding was never detected 439868651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 439968651Skris version rollback attacks was not effective. 440068651Skris 440168651Skris In s23_clnt.c, don't use special rollback-attack detection padding 440268651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 440368651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 440468651Skris SSL 2.0 is the only protocol enabled in the server. 440568651Skris [Bodo Moeller] 440668651Skris 440768651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 440868651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 440968651Skris BIO_dump_indent() are added. 441068651Skris [Richard Levitte] 441168651Skris 441268651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 441368651Skris these print out strings and name structures based on various 441468651Skris flags including RFC2253 support and proper handling of 441568651Skris multibyte characters. Added options to the 'x509' utility 441668651Skris to allow the various flags to be set. 441768651Skris [Steve Henson] 441868651Skris 441968651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 442068651Skris Also change the functions X509_cmp_current_time() and 442168651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 442268651Skris this will enable certificates using GeneralizedTime in validity 442368651Skris dates to be checked. 442468651Skris [Steve Henson] 442568651Skris 442668651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 442768651Skris negative public key encodings) on by default, 442868651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 442968651Skris [Steve Henson] 443068651Skris 443168651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 443268651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 443368651Skris the encoding can be trivially obtained from the structure. 443468651Skris [Steve Henson] 443568651Skris 443668651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 443768651Skris not read locks (CRYPTO_r_[un]lock). 443868651Skris [Bodo Moeller] 443968651Skris 444068651Skris *) A first attempt at creating official support for shared 444168651Skris libraries through configuration. I've kept it so the 444268651Skris default is static libraries only, and the OpenSSL programs 444368651Skris are always statically linked for now, but there are 444468651Skris preparations for dynamic linking in place. 444589837Skris This has been tested on Linux and Tru64. 444668651Skris [Richard Levitte] 444768651Skris 444868651Skris *) Randomness polling function for Win9x, as described in: 444968651Skris Peter Gutmann, Software Generation of Practically Strong 445068651Skris Random Numbers. 445168651Skris [Ulf M�ller] 445268651Skris 445368651Skris *) Fix so PRNG is seeded in req if using an already existing 445468651Skris DSA key. 445568651Skris [Steve Henson] 445668651Skris 445768651Skris *) New options to smime application. -inform and -outform 445868651Skris allow alternative formats for the S/MIME message including 445968651Skris PEM and DER. The -content option allows the content to be 446068651Skris specified separately. This should allow things like Netscape 446168651Skris form signing output easier to verify. 446268651Skris [Steve Henson] 446368651Skris 446468651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 446568651Skris [Steve Henson] 446668651Skris 446768651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 446868651Skris STRING types. These convert content octets to and from the 446968651Skris underlying type. The actual tag and length octets are 447068651Skris already assumed to have been read in and checked. These 447168651Skris are needed because all other string types have virtually 447268651Skris identical handling apart from the tag. By having versions 447368651Skris of the ASN1 functions that just operate on content octets 447468651Skris IMPLICIT tagging can be handled properly. It also allows 447568651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 447668651Skris and ASN1_INTEGER are identical apart from the tag. 447768651Skris [Steve Henson] 447868651Skris 447968651Skris *) Change the handling of OID objects as follows: 448068651Skris 448168651Skris - New object identifiers are inserted in objects.txt, following 448268651Skris the syntax given in objects.README. 448368651Skris - objects.pl is used to process obj_mac.num and create a new 448468651Skris obj_mac.h. 448568651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 448668651Skris obj_mac.h. 448768651Skris 448868651Skris This is currently kind of a hack, and the perl code in objects.pl 448968651Skris isn't very elegant, but it works as I intended. The simplest way 449068651Skris to check that it worked correctly is to look in obj_dat.h and 449168651Skris check the array nid_objs and make sure the objects haven't moved 449268651Skris around (this is important!). Additions are OK, as well as 449368651Skris consistent name changes. 449468651Skris [Richard Levitte] 449568651Skris 449668651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 449768651Skris [Bodo Moeller] 449868651Skris 449968651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 450068651Skris The given file adds to whatever has already been seeded into the 450168651Skris random pool through the RANDFILE configuration file option or 450268651Skris environment variable, or the default random state file. 450368651Skris [Richard Levitte] 450468651Skris 450568651Skris *) mkstack.pl now sorts each macro group into lexical order. 450668651Skris Previously the output order depended on the order the files 450768651Skris appeared in the directory, resulting in needless rewriting 450868651Skris of safestack.h . 450968651Skris [Steve Henson] 451068651Skris 451168651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 451268651Skris work arounds for the VC++ problem that it treats func() as 451368651Skris func(void). Also stripped out the parts of mkdef.pl that 451468651Skris added extra typesafe functions: these no longer exist. 451568651Skris [Steve Henson] 451668651Skris 451768651Skris *) Reorganisation of the stack code. The macros are now all 451868651Skris collected in safestack.h . Each macro is defined in terms of 451968651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 452068651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 452168651Skris this has the advantage of retaining type safety without the 452268651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 452368651Skris then the non typesafe macros are used instead. Also modified the 452468651Skris mkstack.pl script to handle the new form. Needs testing to see 452568651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 452668651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 452768651Skris and PKCS12_STACK_OF. 452868651Skris [Steve Henson] 452968651Skris 453068651Skris *) When some versions of IIS use the 'NET' form of private key the 453168651Skris key derivation algorithm is different. Normally MD5(password) is 453268651Skris used as a 128 bit RC4 key. In the modified case 453368651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 453468651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 453568651Skris as the old Netscape_RSA functions except they have an additional 453668651Skris 'sgckey' parameter which uses the modified algorithm. Also added 453768651Skris an -sgckey command line option to the rsa utility. Thanks to 453868651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 453968651Skris algorithm to openssl-dev. 454068651Skris [Steve Henson] 454168651Skris 454268651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 454368651Skris invalid expansion on some systems (SCO 5.0.5 for example). 454468651Skris Corrected to 'c.kname'. 454568651Skris [Phillip Porch <root@theporch.com>] 454668651Skris 454768651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 454868651Skris a STACK of email addresses from a certificate or request, these look 454968651Skris in the subject name and the subject alternative name extensions and 455068651Skris omit any duplicate addresses. 455168651Skris [Steve Henson] 455268651Skris 455368651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 455468651Skris This makes DSA verification about 2 % faster. 455568651Skris [Bodo Moeller] 455668651Skris 455768651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 455868651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 455968651Skris plus overhead for 1024 bit moduli). 456068651Skris This makes exponentiations about 0.5 % faster for 1024 bit 456168651Skris exponents (as measured by "openssl speed rsa2048"). 456268651Skris [Bodo Moeller] 456368651Skris 456468651Skris *) Rename memory handling macros to avoid conflicts with other 456568651Skris software: 456668651Skris Malloc => OPENSSL_malloc 456768651Skris Malloc_locked => OPENSSL_malloc_locked 456868651Skris Realloc => OPENSSL_realloc 456968651Skris Free => OPENSSL_free 457068651Skris [Richard Levitte] 457168651Skris 457268651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 457368651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 457468651Skris [Bodo Moeller] 457568651Skris 457668651Skris *) CygWin32 support. 457768651Skris [John Jarvie <jjarvie@newsguy.com>] 457868651Skris 457968651Skris *) The type-safe stack code has been rejigged. It is now only compiled 458068651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 458168651Skris by default all type-specific stack functions are "#define"d back to 458268651Skris standard stack functions. This results in more streamlined output 458368651Skris but retains the type-safety checking possibilities of the original 458468651Skris approach. 458568651Skris [Geoff Thorpe] 458668651Skris 458768651Skris *) The STACK code has been cleaned up, and certain type declarations 458868651Skris that didn't make a lot of sense have been brought in line. This has 458968651Skris also involved a cleanup of sorts in safestack.h to more correctly 459068651Skris map type-safe stack functions onto their plain stack counterparts. 459168651Skris This work has also resulted in a variety of "const"ifications of 459268651Skris lots of the code, especially "_cmp" operations which should normally 459368651Skris be prototyped with "const" parameters anyway. 459468651Skris [Geoff Thorpe] 459568651Skris 459668651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 459768651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 459868651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 459968651Skris where all of 'md' is used each time the PRNG is used, but 'state' 460068651Skris is used only indexed by a cyclic counter. As entropy may not be 460168651Skris well distributed from the beginning, 'md' is important as a 460268651Skris chaining variable. However, the output function chains only half 460368651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 460468651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 460568651Skris in all of 'state' being rewritten, with the new values depending 460668651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 460768651Skris [Bodo Moeller] 460868651Skris 460968651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 461068651Skris the handshake is continued after ssl_verify_cert_chain(); 461168651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 461268651Skris can lead to 'unexplainable' connection aborts later. 461368651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 461468651Skris 461568651Skris *) Major EVP API cipher revision. 461668651Skris Add hooks for extra EVP features. This allows various cipher 461768651Skris parameters to be set in the EVP interface. Support added for variable 461868651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 461968651Skris setting of RC2 and RC5 parameters. 462068651Skris 462168651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 462268651Skris ciphers. 462368651Skris 462468651Skris Remove lots of duplicated code from the EVP library. For example *every* 462568651Skris cipher init() function handles the 'iv' in the same way according to the 462668651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 462768651Skris for CFB and OFB modes they zero ctx->num. 462868651Skris 462968651Skris New functionality allows removal of S/MIME code RC2 hack. 463068651Skris 463168651Skris Most of the routines have the same form and so can be declared in terms 463268651Skris of macros. 463368651Skris 463468651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 463568651Skris all individual ciphers. If the cipher wants to handle IVs or keys 463668651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 463768651Skris flags. 463868651Skris 463968651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 464068651Skris value: although software versions of the algorithms cannot fail 464168651Skris any installed hardware versions can. 464268651Skris [Steve Henson] 464368651Skris 464468651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 464568651Skris this option is set, tolerate broken clients that send the negotiated 464668651Skris protocol version number instead of the requested protocol version 464768651Skris number. 464868651Skris [Bodo Moeller] 464968651Skris 465068651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 465168651Skris i.e. non-zero for export ciphersuites, zero otherwise. 465268651Skris Previous versions had this flag inverted, inconsistent with 465368651Skris rsa_tmp_cb (..._TMP_RSA_CB). 465468651Skris [Bodo Moeller; problem reported by Amit Chopra] 465568651Skris 465668651Skris *) Add missing DSA library text string. Work around for some IIS 465768651Skris key files with invalid SEQUENCE encoding. 465868651Skris [Steve Henson] 465968651Skris 466068651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 466168651Skris and so on that are implemented in OpenSSL. 466268651Skris [Richard Levitte] 466368651Skris 466468651Skris *) Enhance c_rehash script. Old version would mishandle certificates 466568651Skris with the same subject name hash and wouldn't handle CRLs at all. 466668651Skris Added -fingerprint option to crl utility, to support new c_rehash 466768651Skris features. 466868651Skris [Steve Henson] 466968651Skris 467068651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 467168651Skris [Ulf M�ller] 467268651Skris 467368651Skris *) Fix for SSL server purpose checking. Server checking was 467468651Skris rejecting certificates which had extended key usage present 467568651Skris but no ssl client purpose. 467668651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 467768651Skris 467868651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 467968651Skris is a little unclear about how a blank password is handled. 468068651Skris Since the password in encoded as a BMPString with terminating 468168651Skris double NULL a zero length password would end up as just the 468268651Skris double NULL. However no password at all is different and is 468368651Skris handled differently in the PKCS#12 key generation code. NS 468468651Skris treats a blank password as zero length. MSIE treats it as no 468568651Skris password on export: but it will try both on import. We now do 468668651Skris the same: PKCS12_parse() tries zero length and no password if 468768651Skris the password is set to "" or NULL (NULL is now a valid password: 468868651Skris it wasn't before) as does the pkcs12 application. 468968651Skris [Steve Henson] 469068651Skris 469168651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 469268651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 469368651Skris be obtained from the error queue. 469468651Skris [Bodo Moeller] 469568651Skris 469668651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 469768651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 469868651Skris accordingly to avoid race conditions (this is necessary because 469968651Skris thread_hash is no longer constant once set). 470068651Skris [Bodo Moeller] 470168651Skris 470268651Skris *) Bugfix for linux-elf makefile.one. 470368651Skris [Ulf M�ller] 470468651Skris 470568651Skris *) RSA_get_default_method() will now cause a default 470668651Skris RSA_METHOD to be chosen if one doesn't exist already. 470768651Skris Previously this was only set during a call to RSA_new() 470868651Skris or RSA_new_method(NULL) meaning it was possible for 470968651Skris RSA_get_default_method() to return NULL. 471068651Skris [Geoff Thorpe] 471168651Skris 471268651Skris *) Added native name translation to the existing DSO code 471368651Skris that will convert (if the flag to do so is set) filenames 471468651Skris that are sufficiently small and have no path information 471568651Skris into a canonical native form. Eg. "blah" converted to 471668651Skris "libblah.so" or "blah.dll" etc. 471768651Skris [Geoff Thorpe] 471868651Skris 471968651Skris *) New function ERR_error_string_n(e, buf, len) which is like 472068651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 472168651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 472268651Skris may not be NULL. 472368651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 472468651Skris 472568651Skris *) CONF library reworked to become more general. A new CONF 472668651Skris configuration file reader "class" is implemented as well as a 472768651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 472868651Skris old CONF_* functions are still there, but are reimplemented to 472968651Skris work in terms of the new functions. Also, a set of functions 473068651Skris to handle the internal storage of the configuration data is 473168651Skris provided to make it easier to write new configuration file 473268651Skris reader "classes" (I can definitely see something reading a 473368651Skris configuration file in XML format, for example), called _CONF_*, 473468651Skris or "the configuration storage API"... 473568651Skris 473668651Skris The new configuration file reading functions are: 473768651Skris 473868651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 473968651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 474068651Skris 474168651Skris NCONF_default, NCONF_WIN32 474268651Skris 474368651Skris NCONF_dump_fp, NCONF_dump_bio 474468651Skris 474568651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 474668651Skris NCONF_new creates a new CONF object. This works in the same way 474768651Skris as other interfaces in OpenSSL, like the BIO interface. 474868651Skris NCONF_dump_* dump the internal storage of the configuration file, 474968651Skris which is useful for debugging. All other functions take the same 475068651Skris arguments as the old CONF_* functions wth the exception of the 475168651Skris first that must be a `CONF *' instead of a `LHASH *'. 475268651Skris 475368651Skris To make it easer to use the new classes with the old CONF_* functions, 475468651Skris the function CONF_set_default_method is provided. 475568651Skris [Richard Levitte] 475668651Skris 475768651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 475868651Skris mentioned in the documentation but had not been implemented. 475968651Skris (This option is not yet really useful because even the additional 476068651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 476168651Skris [Bodo Moeller] 476268651Skris 476368651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 476468651Skris OpenSSL-based applications) load shared libraries and bind to 476568651Skris them in a portable way. 476668651Skris [Geoff Thorpe, with contributions from Richard Levitte] 476768651Skris 476859191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 476959191Skris 477059191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 477159191Skris 477259191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 477359191Skris (the default implementation of RAND_status). 477459191Skris 477559191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 477659191Skris to '-clrext' (= clear extensions), as intended and documented. 477759191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 477859191Skris <attili@amaxo.com>] 477959191Skris 478059191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 478159191Skris was larger than the MD block size. 478259191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 478359191Skris 478459191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 478559191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 478659191Skris using the passed key: if the passed key was a private key the result 478759191Skris of X509_print(), for example, would be to print out all the private key 478859191Skris components. 478959191Skris [Steve Henson] 479059191Skris 479159191Skris *) des_quad_cksum() byte order bug fix. 479259191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 479359191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 479459191Skris 479559191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 479659191Skris discouraged. 479759191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 479859191Skris 479959191Skris *) For easily testing in shell scripts whether some command 480059191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 480159191Skris returns with exit code 0 iff no command of the given name is available. 480259191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 480359191Skris the output goes to stdout and nothing is printed to stderr. 480459191Skris Additional arguments are always ignored. 480559191Skris 480659191Skris Since for each cipher there is a command of the same name, 480759191Skris the 'no-cipher' compilation switches can be tested this way. 480859191Skris 480959191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 481059191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 481159191Skris [Bodo Moeller] 481259191Skris 481359191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 481459191Skris [Bodo Moeller] 481559191Skris 481659191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 481759191Skris is set; it will be thrown away anyway because each handshake creates 481859191Skris its own key. 481959191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 482059191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 482159191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 482259191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 482359191Skris [Bodo Moeller] 482459191Skris 482559191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 482659191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 482759191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 482859191Skris does not suppress any output. 482959191Skris [Richard Levitte] 483059191Skris 483159191Skris *) Add compatibility options to the purpose and trust code. The 483259191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 483359191Skris accepts a certificate or CA, this was the previous behaviour, 483459191Skris with all the associated security issues. 483559191Skris 483659191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 483759191Skris automatically trust self signed roots in certificate store. A 483859191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 483959191Skris a purpose has no associated trust setting and it should instead 484059191Skris use the value in the default purpose. 484159191Skris [Steve Henson] 484259191Skris 484359191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 484459191Skris and fix a memory leak. 484559191Skris [Steve Henson] 484659191Skris 484759191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 484859191Skris reason strings from the previous version of the .c file, as 484959191Skris the default to have only downcase letters (and digits) in 485059191Skris automatically generated reasons codes is not always appropriate. 485159191Skris [Bodo Moeller] 485259191Skris 485359191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 485459191Skris using strerror. Previously, ERR_reason_error_string() returned 485559191Skris library names as reason strings for SYSerr; but SYSerr is a special 485659191Skris case where small numbers are errno values, not library numbers. 485759191Skris [Bodo Moeller] 485859191Skris 485959191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 486059191Skris converts DSA parameters into DH parameters. (When creating parameters, 486159191Skris DSA_generate_parameters is used.) 486259191Skris [Bodo Moeller] 486359191Skris 486459191Skris *) Include 'length' (recommended exponent length) in C code generated 486559191Skris by 'openssl dhparam -C'. 486659191Skris [Bodo Moeller] 486759191Skris 486859191Skris *) The second argument to set_label in perlasm was already being used 486959191Skris so couldn't be used as a "file scope" flag. Moved to third argument 487059191Skris which was free. 487159191Skris [Steve Henson] 487259191Skris 487359191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 487459191Skris instead of RAND_bytes for encryption IVs and salts. 487559191Skris [Bodo Moeller] 487659191Skris 487759191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 487859191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 487959191Skris RAND_set_rand_method would be impossible. 488059191Skris [Bodo Moeller] 488159191Skris 488259191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 488359191Skris number generation fails. 488459191Skris [Bodo Moeller] 488559191Skris 488659191Skris *) New 'rand' application for creating pseudo-random output. 488759191Skris [Bodo Moeller] 488859191Skris 488959191Skris *) Added configuration support for Linux/IA64 489059191Skris [Rolf Haberrecker <rolf@suse.de>] 489159191Skris 489259191Skris *) Assembler module support for Mingw32. 489359191Skris [Ulf M�ller] 489459191Skris 489559191Skris *) Shared library support for HPUX (in shlib/). 489659191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 489759191Skris 489859191Skris *) Shared library support for Solaris gcc. 489959191Skris [Lutz Behnke <behnke@trustcenter.de>] 490059191Skris 490159191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 490259191Skris 490359191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 490459191Skris were added manually and by SMIME_crlf_copy(). 490559191Skris [Steve Henson] 490659191Skris 490759191Skris *) In bntest.c don't call BN_rand with zero bits argument. 490859191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 490959191Skris 491059191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 491159191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 491259191Skris [Ulf M�ller] 491359191Skris 491459191Skris *) Add an optional second argument to the set_label() in the perl 491559191Skris assembly language builder. If this argument exists and is set 491659191Skris to 1 it signals that the assembler should use a symbol whose 491759191Skris scope is the entire file, not just the current function. This 491859191Skris is needed with MASM which uses the format label:: for this scope. 491959191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 492059191Skris 492159191Skris *) Change the ASN1 types so they are typedefs by default. Before 492259191Skris almost all types were #define'd to ASN1_STRING which was causing 492359191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 492459191Skris for example. 492559191Skris [Steve Henson] 492659191Skris 492759191Skris *) Change names of new functions to the new get1/get0 naming 492859191Skris convention: After 'get1', the caller owns a reference count 492959191Skris and has to call ..._free; 'get0' returns a pointer to some 493059191Skris data structure without incrementing reference counters. 493159191Skris (Some of the existing 'get' functions increment a reference 493259191Skris counter, some don't.) 493359191Skris Similarly, 'set1' and 'add1' functions increase reference 493459191Skris counters or duplicate objects. 493559191Skris [Steve Henson] 493659191Skris 493759191Skris *) Allow for the possibility of temp RSA key generation failure: 493859191Skris the code used to assume it always worked and crashed on failure. 493959191Skris [Steve Henson] 494059191Skris 494159191Skris *) Fix potential buffer overrun problem in BIO_printf(). 494259191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 494359191Skris pointed out by David Sacerdote <das33@cornell.edu>] 494459191Skris 494559191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 494659191Skris RAND_egd() and RAND_status(). In the command line application, 494759191Skris the EGD socket can be specified like a seed file using RANDFILE 494859191Skris or -rand. 494959191Skris [Ulf M�ller] 495059191Skris 495159191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 495259191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 495359191Skris [Steve Henson] 495459191Skris 495559191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 495659191Skris list to exclude them. This means that no special compilation option 495759191Skris is needed to use anonymous DH: it just needs to be included in the 495859191Skris cipher list. 495959191Skris [Steve Henson] 496059191Skris 496159191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 496259191Skris EVP_MD_type. The old functionality is available in a new macro called 496359191Skris EVP_MD_md(). Change code that uses it and update docs. 496459191Skris [Steve Henson] 496559191Skris 496659191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 496759191Skris where the 'void *' argument is replaced by a function pointer argument. 496859191Skris Previously 'void *' was abused to point to functions, which works on 496959191Skris many platforms, but is not correct. As these functions are usually 497059191Skris called by macros defined in OpenSSL header files, most source code 497159191Skris should work without changes. 497259191Skris [Richard Levitte] 497359191Skris 497459191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 497559191Skris sections with information on -D... compiler switches used for 497659191Skris compiling the library so that applications can see them. To enable 497759191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 497859191Skris must be defined. E.g., 497959191Skris #define OPENSSL_ALGORITHM_DEFINES 498059191Skris #include <openssl/opensslconf.h> 498159191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 498259191Skris [Richard Levitte, Ulf and Bodo M�ller] 498359191Skris 498459191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 498559191Skris record layer. 498659191Skris [Bodo Moeller] 498759191Skris 498859191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 498959191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 499059191Skris the required ASN1 format: arbitrary types determined by an OID. 499159191Skris [Steve Henson] 499259191Skris 499359191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 499459191Skris argument to 'req'. This is not because the function is newer or 499559191Skris better than others it just uses the work 'NEW' in the certificate 499659191Skris request header lines. Some software needs this. 499759191Skris [Steve Henson] 499859191Skris 499959191Skris *) Reorganise password command line arguments: now passwords can be 500059191Skris obtained from various sources. Delete the PEM_cb function and make 500159191Skris it the default behaviour: i.e. if the callback is NULL and the 500259191Skris usrdata argument is not NULL interpret it as a null terminated pass 500359191Skris phrase. If usrdata and the callback are NULL then the pass phrase 500459191Skris is prompted for as usual. 500559191Skris [Steve Henson] 500659191Skris 500759191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 500859191Skris the support is automatically enabled. The resulting binaries will 500959191Skris autodetect the card and use it if present. 501059191Skris [Ben Laurie and Compaq Inc.] 501159191Skris 501259191Skris *) Work around for Netscape hang bug. This sends certificate request 501359191Skris and server done in one record. Since this is perfectly legal in the 501459191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 501559191Skris the bugs/SSLv3 entry for more info. 501659191Skris [Steve Henson] 501759191Skris 501859191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 501959191Skris [Andy Polyakov] 502059191Skris 502159191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 502259191Skris of seed file. 502359191Skris [Steve Henson] 502459191Skris 502559191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 502659191Skris [Bodo Moeller] 502759191Skris 502859191Skris *) Add command line password options to the remaining applications. 502959191Skris [Steve Henson] 503059191Skris 503159191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 503259191Skris bits. 503359191Skris [Ulf M�ller] 503459191Skris 503559191Skris *) More tests in bntest.c, and changed test_bn output. 503659191Skris [Ulf M�ller] 503759191Skris 503859191Skris *) ./config recognizes MacOS X now. 503959191Skris [Andy Polyakov] 504059191Skris 504159191Skris *) Bug fix for BN_div() when the first words of num and divsor are 504259191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 504359191Skris [Ulf M�ller] 504459191Skris 504559191Skris *) Add support for various broken PKCS#8 formats, and command line 504659191Skris options to produce them. 504759191Skris [Steve Henson] 504859191Skris 504959191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 505059191Skris get temporary BIGNUMs from a BN_CTX. 505159191Skris [Ulf M�ller] 505259191Skris 505359191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 505459191Skris for p == 0. 505559191Skris [Ulf M�ller] 505659191Skris 505759191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 505859191Skris include a #define from the old name to the new. The original intent 505959191Skris was that statically linked binaries could for example just call 506059191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 506159191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 506259191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 506359191Skris one would link with the other. They are now in separate source files. 506459191Skris [Steve Henson] 506559191Skris 506659191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 506759191Skris [Steve Henson] 506859191Skris 506959191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 507059191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 507159191Skris loop, our standard modexp algorithms are faster). 507259191Skris [Bodo Moeller] 507359191Skris 507459191Skris *) Support for the EBCDIC character set completed. 507559191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 507659191Skris 507759191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 507859191Skris use void * instead of char * in lhash. 507959191Skris [Ulf M�ller] 508059191Skris 508159191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 508259191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 508359191Skris this the server could overwrite ephemeral keys that the client 508459191Skris has already seen). 508559191Skris [Bodo Moeller] 508659191Skris 508759191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 508859191Skris using 50 iterations of the Rabin-Miller test. 508959191Skris 509059191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 509159191Skris iterations of the Rabin-Miller test as required by the appendix 509259191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 509359191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 509459191Skris generation becomes much faster. 509559191Skris 509659191Skris This implies a change for the callback functions in DSA_is_prime 509759191Skris and DSA_generate_parameters: The callback function is called once 509859191Skris for each positive witness in the Rabin-Miller test, not just 509959191Skris occasionally in the inner loop; and the parameters to the 510059191Skris callback function now provide an iteration count for the outer 510159191Skris loop rather than for the current invocation of the inner loop. 510259191Skris DSA_generate_parameters additionally can call the callback 510359191Skris function with an 'iteration count' of -1, meaning that a 510459191Skris candidate has passed the trial division test (when q is generated 510559191Skris from an application-provided seed, trial division is skipped). 510659191Skris [Bodo Moeller] 510759191Skris 510859191Skris *) New function BN_is_prime_fasttest that optionally does trial 510959191Skris division before starting the Rabin-Miller test and has 511059191Skris an additional BN_CTX * argument (whereas BN_is_prime always 511159191Skris has to allocate at least one BN_CTX). 511259191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 511359191Skris trial division stage. 511459191Skris [Bodo Moeller] 511559191Skris 511659191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 511759191Skris as ASN1_TIME. 511859191Skris [Steve Henson] 511959191Skris 512059191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 512159191Skris [Steve Henson] 512259191Skris 512359191Skris *) New function BN_pseudo_rand(). 512459191Skris [Ulf M�ller] 512559191Skris 512659191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 512759191Skris bignum version of BN_from_montgomery() with the working code from 512859191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 512959191Skris the comments. 513059191Skris [Ulf M�ller] 513159191Skris 513259191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 513359191Skris made it impossible to use the same SSL_SESSION data structure in 513459191Skris SSL2 clients in multiple threads. 513559191Skris [Bodo Moeller] 513659191Skris 513759191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 513859191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 513959191Skris to seed the PRNG (previously an explicit byte count was required). 514059191Skris [Ulf M�ller, Bodo M�ller] 514159191Skris 514259191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 514359191Skris used (char *) instead of (void *) and had casts all over the place. 514459191Skris [Steve Henson] 514559191Skris 514659191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 514759191Skris [Ulf M�ller] 514859191Skris 514959191Skris *) Retain source code compatibility for BN_prime_checks macro: 515059191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 515159191Skris BN_prime_checks_for_size to determine the appropriate number of 515259191Skris Rabin-Miller iterations. 515359191Skris [Ulf M�ller] 515459191Skris 515559191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 515659191Skris DH_CHECK_P_NOT_SAFE_PRIME. 515759191Skris (Check if this is true? OpenPGP calls them "strong".) 515859191Skris [Ulf M�ller] 515959191Skris 516059191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 516159191Skris "dhparam". The old programs are retained for now but will handle DH keys 516259191Skris (instead of parameters) in future. 516359191Skris [Steve Henson] 516459191Skris 516559191Skris *) Make the ciphers, s_server and s_client programs check the return values 516659191Skris when a new cipher list is set. 516759191Skris [Steve Henson] 516859191Skris 516959191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 517059191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 517159191Skris wrong. 517259191Skris 517359191Skris The syntax for the cipher sorting has been extended to support sorting by 517459191Skris cipher-strength (using the strength_bits hard coded in the tables). 517559191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 517659191Skris 517759191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 517859191Skris string with an "undefined" symbol (neither command nor alphanumeric 517959191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 518059191Skris an error is flagged. 518159191Skris 518259191Skris Due to the strength-sorting extension, the code of the 518359191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 518459191Skris the readability was also increased :-) 518559191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 518659191Skris 518759191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 518859191Skris for the first serial number and places 2 in the serial number file. This 518959191Skris avoids problems when the root CA is created with serial number zero and 519059191Skris the first user certificate has the same issuer name and serial number 519159191Skris as the root CA. 519259191Skris [Steve Henson] 519359191Skris 519459191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 519559191Skris the new code. Add documentation for this stuff. 519659191Skris [Steve Henson] 519759191Skris 519859191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 519959191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 520059191Skris structures and behave in an analagous way to the X509v3 functions: 520159191Skris they shouldn't be called directly but wrapper functions should be used 520259191Skris instead. 520359191Skris 520459191Skris So we also now have some wrapper functions that call the X509at functions 520559191Skris when passed certificate requests. (TO DO: similar things can be done with 520659191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 520759191Skris things. Some of these need some d2i or i2d and print functionality 520859191Skris because they handle more complex structures.) 520959191Skris [Steve Henson] 521059191Skris 521159191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 521259191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 521359191Skris NO_RSA in ssl/s2*.c. 521459191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 521559191Skris 521659191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 521759191Skris has a return value which indicates the quality of the random data 521859191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 521959191Skris error queue. New function RAND_pseudo_bytes() generates output that is 522059191Skris guaranteed to be unique but not unpredictable. RAND_add is like 522159191Skris RAND_seed, but takes an extra argument for an entropy estimate 522259191Skris (RAND_seed always assumes full entropy). 522359191Skris [Ulf M�ller] 522459191Skris 522559191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 522659191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 522759191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 522859191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 522959191Skris false-positive rate of at most 2^-80 for random input. 523059191Skris [Bodo Moeller] 523159191Skris 523259191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 523359191Skris [Bodo Moeller] 523459191Skris 523559191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 523659191Skris in the 0.9.5 release), this returns the chain 523759191Skris from an X509_CTX structure with a dup of the stack and all 523859191Skris the X509 reference counts upped: so the stack will exist 523959191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 524059191Skris to use this. 524159191Skris 524259191Skris Also make SSL_SESSION_print() print out the verify return 524359191Skris code. 524459191Skris [Steve Henson] 524559191Skris 524659191Skris *) Add manpage for the pkcs12 command. Also change the default 524759191Skris behaviour so MAC iteration counts are used unless the new 524859191Skris -nomaciter option is used. This improves file security and 524959191Skris only older versions of MSIE (4.0 for example) need it. 525059191Skris [Steve Henson] 525159191Skris 525259191Skris *) Honor the no-xxx Configure options when creating .DEF files. 525359191Skris [Ulf M�ller] 525459191Skris 525559191Skris *) Add PKCS#10 attributes to field table: challengePassword, 525659191Skris unstructuredName and unstructuredAddress. These are taken from 525759191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 525859191Skris international characters are used. 525959191Skris 526059191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 526159191Skris based on strings. Remove the 'loc' parameter when adding 526259191Skris attributes because these will be a SET OF encoding which is sorted 526359191Skris in ASN1 order. 526459191Skris [Steve Henson] 526559191Skris 526659191Skris *) Initial changes to the 'req' utility to allow request generation 526759191Skris automation. This will allow an application to just generate a template 526859191Skris file containing all the field values and have req construct the 526959191Skris request. 527059191Skris 527159191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 527259191Skris used all over the place including certificate requests and PKCS#7 527359191Skris structures. They are currently handled manually where necessary with 527459191Skris some primitive wrappers for PKCS#7. The new functions behave in a 527559191Skris manner analogous to the X509 extension functions: they allow 527659191Skris attributes to be looked up by NID and added. 527759191Skris 527859191Skris Later something similar to the X509V3 code would be desirable to 527959191Skris automatically handle the encoding, decoding and printing of the 528059191Skris more complex types. The string types like challengePassword can 528159191Skris be handled by the string table functions. 528259191Skris 528359191Skris Also modified the multi byte string table handling. Now there is 528459191Skris a 'global mask' which masks out certain types. The table itself 528559191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 528659191Skris is useful when for example there is only one permissible type 528759191Skris (as in countryName) and using the mask might result in no valid 528859191Skris types at all. 528959191Skris [Steve Henson] 529059191Skris 529159191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 529259191Skris SSL_get_peer_finished to allow applications to obtain the latest 529359191Skris Finished messages sent to the peer or expected from the peer, 529459191Skris respectively. (SSL_get_peer_finished is usually the Finished message 529559191Skris actually received from the peer, otherwise the protocol will be aborted.) 529659191Skris 529759191Skris As the Finished message are message digests of the complete handshake 529859191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 529959191Skris be used for external authentication procedures when the authentication 530059191Skris provided by SSL/TLS is not desired or is not enough. 530159191Skris [Bodo Moeller] 530259191Skris 530359191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 530459191Skris the host supports BWX extension and if Compaq C is present on the 530559191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 530659191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 530759191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 530859191Skris SHA1. 530959191Skris [Andy Polyakov] 531059191Skris 531159191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 531259191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 531359191Skris weak crypto and after checking the certificate is SGC a second one 531459191Skris with strong crypto. MS SGC stops the first handshake after receiving 531559191Skris the server certificate message and sends a second client hello. Since 531659191Skris a server will typically do all the time consuming operations before 531759191Skris expecting any further messages from the client (server key exchange 531859191Skris is the most expensive) there is little difference between the two. 531959191Skris 532059191Skris To get OpenSSL to support MS SGC we have to permit a second client 532159191Skris hello message after we have sent server done. In addition we have to 532259191Skris reset the MAC if we do get this second client hello. 532359191Skris [Steve Henson] 532459191Skris 532559191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 532659191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 532759191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 532859191Skris format DER encoded private key. Newer code should use PKCS#8 format which 532959191Skris has the key type encoded in the ASN1 structure. Added DER private key 533059191Skris support to pkcs8 application. 533159191Skris [Steve Henson] 533259191Skris 533359191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 533459191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 533559191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 533659191Skris is set, we interpret this as a request to violate the specification 533759191Skris (the worst that can happen is a handshake failure, and 'correct' 533859191Skris behaviour would result in a handshake failure anyway). 533959191Skris [Bodo Moeller] 534059191Skris 534159191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 534259191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 534359191Skris concurrently obtain them from an external cache). 534459191Skris The internal cache can handle only one SSL_SESSION with a given ID, 534559191Skris so if there's a conflict, we now throw out the old one to achieve 534659191Skris consistency. 534759191Skris [Bodo Moeller] 534859191Skris 534959191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 535059191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 535159191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 535259191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 535359191Skris example. 535459191Skris [Steve Henson] 535559191Skris 535659191Skris *) Simplify the trust setting structure and code. Now we just have 535759191Skris two sequences of OIDs for trusted and rejected settings. These will 535859191Skris typically have values the same as the extended key usage extension 535959191Skris and any application specific purposes. 536059191Skris 536159191Skris The trust checking code now has a default behaviour: it will just 536259191Skris check for an object with the same NID as the passed id. Functions can 536359191Skris be provided to override either the default behaviour or the behaviour 536459191Skris for a given id. SSL client, server and email already have functions 536559191Skris in place for compatibility: they check the NID and also return "trusted" 536659191Skris if the certificate is self signed. 536759191Skris [Steve Henson] 536859191Skris 536959191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 537059191Skris traditional format into an EVP_PKEY structure. 537159191Skris [Steve Henson] 537259191Skris 537359191Skris *) Add a password callback function PEM_cb() which either prompts for 537459191Skris a password if usr_data is NULL or otherwise assumes it is a null 537559191Skris terminated password. Allow passwords to be passed on command line 537659191Skris environment or config files in a few more utilities. 537759191Skris [Steve Henson] 537859191Skris 537959191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 538059191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 538159191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 538259191Skris Update documentation. 538359191Skris [Steve Henson] 538459191Skris 538559191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 538659191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 538759191Skris and produce an error if it couldn't. For compatibility we also have 538859191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 538959191Skris don't allocate anything because they don't need to. 539059191Skris [Steve Henson] 539159191Skris 539259191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 539359191Skris for details. 539459191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 539559191Skris 539659191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 539759191Skris possibly others as well. The purpose is to make an interface that 539859191Skris provide hooks so anyone can build a separate set of allocation and 539959191Skris deallocation routines to be used by OpenSSL, for example memory 540059191Skris pool implementations, or something else, which was previously hard 540159191Skris since Malloc(), Realloc() and Free() were defined as macros having 540259191Skris the values malloc, realloc and free, respectively (except for Win32 540359191Skris compilations). The same is provided for memory debugging code. 540459191Skris OpenSSL already comes with functionality to find memory leaks, but 540559191Skris this gives people a chance to debug other memory problems. 540659191Skris 540759191Skris With these changes, a new set of functions and macros have appeared: 540859191Skris 540968651Skris CRYPTO_set_mem_debug_functions() [F] 541068651Skris CRYPTO_get_mem_debug_functions() [F] 541168651Skris CRYPTO_dbg_set_options() [F] 541268651Skris CRYPTO_dbg_get_options() [F] 541368651Skris CRYPTO_malloc_debug_init() [M] 541459191Skris 541559191Skris The memory debug functions are NULL by default, unless the library 541659191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 541759191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 541859191Skris gives the standard debugging functions that come with OpenSSL) or 541959191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 542059191Skris provided by the library user) must be used. When the standard 542159191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 542259191Skris request additional information: 542359191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 542459191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 542559191Skris 542659191Skris Also, things like CRYPTO_set_mem_functions will always give the 542759191Skris expected result (the new set of functions is used for allocation 542859191Skris and deallocation) at all times, regardless of platform and compiler 542959191Skris options. 543059191Skris 543159191Skris To finish it up, some functions that were never use in any other 543259191Skris way than through macros have a new API and new semantic: 543359191Skris 543459191Skris CRYPTO_dbg_malloc() 543559191Skris CRYPTO_dbg_realloc() 543659191Skris CRYPTO_dbg_free() 543759191Skris 543859191Skris All macros of value have retained their old syntax. 543959191Skris [Richard Levitte and Bodo Moeller] 544059191Skris 544159191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 544259191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 544359191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 544459191Skris algorithm. 544559191Skris [Steve Henson] 544659191Skris 544759191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 544859191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 544959191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 545059191Skris 545159191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 545259191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 545359191Skris functionality to handle multipart/signed properly) and a utility 545459191Skris called 'smime' to call all this stuff. This is based on code I 545559191Skris originally wrote for Celo who have kindly allowed it to be 545659191Skris included in OpenSSL. 545759191Skris [Steve Henson] 545859191Skris 545959191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 546059191Skris des_set_key (aka des_key_sched). Global variable des_check_key 546159191Skris decides which of these is called by des_set_key; this way 546259191Skris des_check_key behaves as it always did, but applications and 546359191Skris the library itself, which was buggy for des_check_key == 1, 546459191Skris have a cleaner way to pick the version they need. 546559191Skris [Bodo Moeller] 546659191Skris 546759191Skris *) New function PKCS12_newpass() which changes the password of a 546859191Skris PKCS12 structure. 546959191Skris [Steve Henson] 547059191Skris 547159191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 547259191Skris dynamic mix. In both cases the ids can be used as an index into the 547359191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 547459191Skris functions so they accept a list of the field values and the 547559191Skris application doesn't need to directly manipulate the X509_TRUST 547659191Skris structure. 547759191Skris [Steve Henson] 547859191Skris 547959191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 548059191Skris need initialising. 548159191Skris [Steve Henson] 548259191Skris 548359191Skris *) Modify the way the V3 extension code looks up extensions. This now 548459191Skris works in a similar way to the object code: we have some "standard" 548559191Skris extensions in a static table which is searched with OBJ_bsearch() 548659191Skris and the application can add dynamic ones if needed. The file 548759191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 548859191Skris updated whenever a new extension is added to the core code and kept 548959191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 549059191Skris this. New extensions are not added too often so this file can readily 549159191Skris be maintained manually. 549259191Skris 549359191Skris There are two big advantages in doing things this way. The extensions 549459191Skris can be looked up immediately and no longer need to be "added" using 549559191Skris X509V3_add_standard_extensions(): this function now does nothing. 549659191Skris [Side note: I get *lots* of email saying the extension code doesn't 549759191Skris work because people forget to call this function] 549859191Skris Also no dynamic allocation is done unless new extensions are added: 549959191Skris so if we don't add custom extensions there is no need to call 550059191Skris X509V3_EXT_cleanup(). 550159191Skris [Steve Henson] 550259191Skris 550359191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 550459191Skris magic header, so unsalted files fail gracefully instead of just decrypting 550559191Skris to garbage. This is because not salting is a big security hole, so people 550659191Skris should be discouraged from doing it. 550759191Skris [Ben Laurie] 550859191Skris 550959191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 551059191Skris digest to be passed on the command line but it only used this 551159191Skris parameter when signing a certificate. Modified so all relevant 551259191Skris operations are affected by the digest parameter including the 551359191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 551459191Skris DSA key was used because it didn't fix the digest. 551559191Skris [Steve Henson] 551659191Skris 551759191Skris *) Initial certificate chain verify code. Currently tests the untrusted 551859191Skris certificates for consistency with the verify purpose (which is set 551959191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 552059191Skris 552159191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 552259191Skris this is because it will reject chains with invalid extensions whereas 552359191Skris every previous version of OpenSSL and SSLeay made no checks at all. 552459191Skris 552559191Skris Trust code: checks the root CA for the relevant trust settings. Trust 552659191Skris settings have an initial value consistent with the verify purpose: e.g. 552759191Skris if the verify purpose is for SSL client use it expects the CA to be 552859191Skris trusted for SSL client use. However the default value can be changed to 552959191Skris permit custom trust settings: one example of this would be to only trust 553059191Skris certificates from a specific "secure" set of CAs. 553159191Skris 553259191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 553359191Skris which should be used for version portability: especially since the 553459191Skris verify structure is likely to change more often now. 553559191Skris 553659191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 553759191Skris to set them. If not set then assume SSL clients will verify SSL servers 553859191Skris and vice versa. 553959191Skris 554059191Skris Two new options to the verify program: -untrusted allows a set of 554159191Skris untrusted certificates to be passed in and -purpose which sets the 554259191Skris intended purpose of the certificate. If a purpose is set then the 554359191Skris new chain verify code is used to check extension consistency. 554459191Skris [Steve Henson] 554559191Skris 554659191Skris *) Support for the authority information access extension. 554759191Skris [Steve Henson] 554859191Skris 554959191Skris *) Modify RSA and DSA PEM read routines to transparently handle 555059191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 555159191Skris public keys in a format compatible with certificate 555259191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 555359191Skris functions called *_PublicKey_* which used various odd formats so 555459191Skris these are retained for compatibility: however the DSA variants were 555559191Skris never in a public release so they have been deleted. Changed dsa/rsa 555659191Skris utilities to handle the new format: note no releases ever handled public 555759191Skris keys so we should be OK. 555859191Skris 555959191Skris The primary motivation for this change is to avoid the same fiasco 556059191Skris that dogs private keys: there are several incompatible private key 556159191Skris formats some of which are standard and some OpenSSL specific and 556259191Skris require various evil hacks to allow partial transparent handling and 556359191Skris even then it doesn't work with DER formats. Given the option anything 556459191Skris other than PKCS#8 should be dumped: but the other formats have to 556559191Skris stay in the name of compatibility. 556659191Skris 556759191Skris With public keys and the benefit of hindsight one standard format 556859191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 556959191Skris it clearly returns an error if you try to read the wrong kind of key. 557059191Skris 557159191Skris Added a -pubkey option to the 'x509' utility to output the public key. 557259191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 557359191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 557459191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 557559191Skris that do the same as the EVP_PKEY_assign_*() except they up the 557659191Skris reference count of the added key (they don't "swallow" the 557759191Skris supplied key). 557859191Skris [Steve Henson] 557959191Skris 558059191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 558159191Skris CRLs would fail if the file contained no certificates or no CRLs: 558259191Skris added a new function to read in both types and return the number 558359191Skris read: this means that if none are read it will be an error. The 558459191Skris DER versions of the certificate and CRL reader would always fail 558559191Skris because it isn't possible to mix certificates and CRLs in DER format 558659191Skris without choking one or the other routine. Changed this to just read 558759191Skris a certificate: this is the best we can do. Also modified the code 558859191Skris in apps/verify.c to take notice of return codes: it was previously 558959191Skris attempting to read in certificates from NULL pointers and ignoring 559059191Skris any errors: this is one reason why the cert and CRL reader seemed 559159191Skris to work. It doesn't check return codes from the default certificate 559259191Skris routines: these may well fail if the certificates aren't installed. 559359191Skris [Steve Henson] 559459191Skris 559559191Skris *) Code to support otherName option in GeneralName. 559659191Skris [Steve Henson] 559759191Skris 559859191Skris *) First update to verify code. Change the verify utility 559959191Skris so it warns if it is passed a self signed certificate: 560059191Skris for consistency with the normal behaviour. X509_verify 560159191Skris has been modified to it will now verify a self signed 560259191Skris certificate if *exactly* the same certificate appears 560359191Skris in the store: it was previously impossible to trust a 560459191Skris single self signed certificate. This means that: 560559191Skris openssl verify ss.pem 560659191Skris now gives a warning about a self signed certificate but 560759191Skris openssl verify -CAfile ss.pem ss.pem 560859191Skris is OK. 560959191Skris [Steve Henson] 561059191Skris 561159191Skris *) For servers, store verify_result in SSL_SESSION data structure 561259191Skris (and add it to external session representation). 561359191Skris This is needed when client certificate verifications fails, 561459191Skris but an application-provided verification callback (set by 561559191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 561659191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 561759191Skris but returns 1): When the session is reused, we have to set 561859191Skris ssl->verify_result to the appropriate error code to avoid 561959191Skris security holes. 562059191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 562159191Skris 562259191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 562359191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 562459191Skris didn't contain any existing data because it was being created. 562559191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 562659191Skris 562759191Skris *) Add a salt to the key derivation routines in enc.c. This 562859191Skris forms the first 8 bytes of the encrypted file. Also add a 562959191Skris -S option to allow a salt to be input on the command line. 563059191Skris [Steve Henson] 563159191Skris 563259191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 563359191Skris to compare two certificates. We do this by working out the SHA1 563459191Skris hash and comparing that. X509_cmp() will be needed by the trust 563559191Skris code. 563659191Skris [Steve Henson] 563759191Skris 563859191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 563959191Skris the reference count in the SSL_SESSION returned. 564059191Skris [Geoff Thorpe <geoff@eu.c2.net>] 564159191Skris 564259191Skris *) Fix for 'req': it was adding a null to request attributes. 564359191Skris Also change the X509_LOOKUP and X509_INFO code to handle 564459191Skris certificate auxiliary information. 564559191Skris [Steve Henson] 564659191Skris 564759191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 564859191Skris the 'enc' command. 564959191Skris [Steve Henson] 565059191Skris 565159191Skris *) Add the possibility to add extra information to the memory leak 565259191Skris detecting output, to form tracebacks, showing from where each 565359191Skris allocation was originated: CRYPTO_push_info("constant string") adds 565459191Skris the string plus current file name and line number to a per-thread 565559191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 565659191Skris is like calling CYRPTO_pop_info() until the stack is empty. 565759191Skris Also updated memory leak detection code to be multi-thread-safe. 565859191Skris [Richard Levitte] 565959191Skris 566059191Skris *) Add options -text and -noout to pkcs7 utility and delete the 566159191Skris encryption options which never did anything. Update docs. 566259191Skris [Steve Henson] 566359191Skris 566459191Skris *) Add options to some of the utilities to allow the pass phrase 566559191Skris to be included on either the command line (not recommended on 566659191Skris OSes like Unix) or read from the environment. Update the 566759191Skris manpages and fix a few bugs. 566859191Skris [Steve Henson] 566959191Skris 567059191Skris *) Add a few manpages for some of the openssl commands. 567159191Skris [Steve Henson] 567259191Skris 567359191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 567459191Skris leaking and not finding already revoked certificates. 567559191Skris [Steve Henson] 567659191Skris 567759191Skris *) Extensive changes to support certificate auxiliary information. 567859191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 567959191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 568059191Skris can still read in a certificate file in the usual way but it 568159191Skris will also read in any additional "auxiliary information". By 568259191Skris doing things this way a fair degree of compatibility can be 568359191Skris retained: existing certificates can have this information added 568459191Skris using the new 'x509' options. 568559191Skris 568659191Skris Current auxiliary information includes an "alias" and some trust 568759191Skris settings. The trust settings will ultimately be used in enhanced 568859191Skris certificate chain verification routines: currently a certificate 568959191Skris can only be trusted if it is self signed and then it is trusted 569059191Skris for all purposes. 569159191Skris [Steve Henson] 569259191Skris 569359191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 569459191Skris The problem was that one of the replacement routines had not been working 569559191Skris since SSLeay releases. For now the offending routine has been replaced 569659191Skris with non-optimised assembler. Even so, this now gives around 95% 569759191Skris performance improvement for 1024 bit RSA signs. 569859191Skris [Mark Cox] 569959191Skris 570059191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 570159191Skris handling. Most clients have the effective key size in bits equal to 570259191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 570359191Skris A few however don't do this and instead use the size of the decrypted key 570459191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 570559191Skris the effective key length. In this case the effective key length can still 570659191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 570759191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 570859191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 570959191Skris the key length and effective key length are equal. 571059191Skris [Steve Henson] 571159191Skris 571259191Skris *) Add a bunch of functions that should simplify the creation of 571359191Skris X509_NAME structures. Now you should be able to do: 571459191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 571559191Skris and have it automatically work out the correct field type and fill in 571659191Skris the structures. The more adventurous can try: 571759191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 571859191Skris and it will (hopefully) work out the correct multibyte encoding. 571959191Skris [Steve Henson] 572059191Skris 572159191Skris *) Change the 'req' utility to use the new field handling and multibyte 572259191Skris copy routines. Before the DN field creation was handled in an ad hoc 572359191Skris way in req, ca, and x509 which was rather broken and didn't support 572459191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 572559191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 572659191Skris using the dirstring_type option. See the new comment in the default 572759191Skris openssl.cnf for more info. 572859191Skris [Steve Henson] 572959191Skris 573059191Skris *) Make crypto/rand/md_rand.c more robust: 573159191Skris - Assure unique random numbers after fork(). 573259191Skris - Make sure that concurrent threads access the global counter and 573359191Skris md serializably so that we never lose entropy in them 573459191Skris or use exactly the same state in multiple threads. 573559191Skris Access to the large state is not always serializable because 573659191Skris the additional locking could be a performance killer, and 573759191Skris md should be large enough anyway. 573859191Skris [Bodo Moeller] 573959191Skris 574059191Skris *) New file apps/app_rand.c with commonly needed functionality 574159191Skris for handling the random seed file. 574259191Skris 574359191Skris Use the random seed file in some applications that previously did not: 574459191Skris ca, 574559191Skris dsaparam -genkey (which also ignored its '-rand' option), 574659191Skris s_client, 574759191Skris s_server, 574859191Skris x509 (when signing). 574959191Skris Except on systems with /dev/urandom, it is crucial to have a random 575059191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 575159191Skris for RSA signatures we could do without one. 575259191Skris 575359191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 575459191Skris of each file listed in the '-rand' option. The function as previously 575559191Skris found in genrsa is now in app_rand.c and is used by all programs 575659191Skris that support '-rand'. 575759191Skris [Bodo Moeller] 575859191Skris 575959191Skris *) In RAND_write_file, use mode 0600 for creating files; 576059191Skris don't just chmod when it may be too late. 576159191Skris [Bodo Moeller] 576259191Skris 576359191Skris *) Report an error from X509_STORE_load_locations 576459191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 576559191Skris [Bill Perry] 576659191Skris 576759191Skris *) New function ASN1_mbstring_copy() this copies a string in either 576859191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 576959191Skris into an ASN1_STRING type. A mask of permissible types is passed 577059191Skris and it chooses the "minimal" type to use or an error if not type 577159191Skris is suitable. 577259191Skris [Steve Henson] 577359191Skris 577459191Skris *) Add function equivalents to the various macros in asn1.h. The old 577559191Skris macros are retained with an M_ prefix. Code inside the library can 577659191Skris use the M_ macros. External code (including the openssl utility) 577759191Skris should *NOT* in order to be "shared library friendly". 577859191Skris [Steve Henson] 577959191Skris 578059191Skris *) Add various functions that can check a certificate's extensions 578159191Skris to see if it usable for various purposes such as SSL client, 578259191Skris server or S/MIME and CAs of these types. This is currently 578359191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 578459191Skris verification. Also added a -purpose flag to x509 utility to 578559191Skris print out all the purposes. 578659191Skris [Steve Henson] 578759191Skris 578859191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 578959191Skris functions. 579059191Skris [Steve Henson] 579159191Skris 579259191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 579359191Skris for, obtain and decode and extension and obtain its critical flag. 579459191Skris This allows all the necessary extension code to be handled in a 579559191Skris single function call. 579659191Skris [Steve Henson] 579759191Skris 579859191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 579959191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 580059191Skris [Andy Polyakov] 580159191Skris 580259191Skris *) New -noout option to asn1parse. This causes no output to be produced 580359191Skris its main use is when combined with -strparse and -out to extract data 580459191Skris from a file (which may not be in ASN.1 format). 580559191Skris [Steve Henson] 580659191Skris 580759191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 580859191Skris when producing the local key id. 580959191Skris [Richard Levitte <levitte@stacken.kth.se>] 581059191Skris 581159191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 581259191Skris stated explicitly. If it is not stated then it tries the first server 581359191Skris certificate file. The previous behaviour hard coded the filename 581459191Skris "server.pem". 581559191Skris [Steve Henson] 581659191Skris 581759191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 581859191Skris a public key to be input or output. For example: 581959191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 582059191Skris Also added necessary DSA public key functions to handle this. 582159191Skris [Steve Henson] 582259191Skris 582359191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 582459191Skris in the message. This was handled by allowing 582559191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 582659191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 582759191Skris 582859191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 582959191Skris to the end of the strings whereas this didn't. This would cause problems 583059191Skris if strings read with d2i_ASN1_bytes() were later modified. 583159191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 583259191Skris 583359191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 583459191Skris data and it contains EOF it will end up returning an error. This is 583559191Skris caused by input 46 bytes long. The cause is due to the way base64 583659191Skris BIOs find the start of base64 encoded data. They do this by trying a 583759191Skris trial decode on each line until they find one that works. When they 583859191Skris do a flag is set and it starts again knowing it can pass all the 583959191Skris data directly through the decoder. Unfortunately it doesn't reset 584059191Skris the context it uses. This means that if EOF is reached an attempt 584159191Skris is made to pass two EOFs through the context and this causes the 584259191Skris resulting error. This can also cause other problems as well. As is 584359191Skris usual with these problems it takes *ages* to find and the fix is 584459191Skris trivial: move one line. 584559191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 584659191Skris 584759191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 584859191Skris old code wouldn't work because it needed to select() on sockets and the 584959191Skris tty (for keypresses and to see if data could be written). Win32 only 585059191Skris supports select() on sockets so we select() with a 1s timeout on the 585159191Skris sockets and then see if any characters are waiting to be read, if none 585259191Skris are present then we retry, we also assume we can always write data to 585359191Skris the tty. This isn't nice because the code then blocks until we've 585459191Skris received a complete line of data and it is effectively polling the 585559191Skris keyboard at 1s intervals: however it's quite a bit better than not 585659191Skris working at all :-) A dedicated Windows application might handle this 585759191Skris with an event loop for example. 585859191Skris [Steve Henson] 585959191Skris 586059191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 586159191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 586259191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 586359191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 586459191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 586559191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 586659191Skris This necessitated the support of an extra signature type NID_md5_sha1 586759191Skris for SSL signatures and modifications to the SSL library to use it instead 586859191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 586959191Skris [Steve Henson] 587059191Skris 587159191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 587259191Skris will lookup a CRL issuers certificate and verify the signature in a 587359191Skris similar way to the verify program. Tidy up the crl program so it 587459191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 587559191Skris less strict. It will now permit CRL extensions even if it is not 587659191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 587759191Skris [Steve Henson] 587859191Skris 587959191Skris *) Initialize all non-automatic variables each time one of the openssl 588059191Skris sub-programs is started (this is necessary as they may be started 588159191Skris multiple times from the "OpenSSL>" prompt). 588259191Skris [Lennart Bang, Bodo Moeller] 588359191Skris 588459191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 588559191Skris removing all other RSA functionality (this is what NO_RSA does). This 588659191Skris is so (for example) those in the US can disable those operations covered 588759191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 588859191Skris key generation. 588959191Skris [Steve Henson] 589059191Skris 589159191Skris *) Non-copying interface to BIO pairs. 589259191Skris (still largely untested) 589359191Skris [Bodo Moeller] 589459191Skris 589559191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 589659191Skris ASCII string. This was handled independently in various places before. 589759191Skris [Steve Henson] 589859191Skris 589959191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 590059191Skris UTF8 strings a character at a time. 590159191Skris [Steve Henson] 590259191Skris 590359191Skris *) Use client_version from client hello to select the protocol 590459191Skris (s23_srvr.c) and for RSA client key exchange verification 590559191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 590659191Skris [Bodo Moeller] 590759191Skris 590859191Skris *) Add various utility functions to handle SPKACs, these were previously 590959191Skris handled by poking round in the structure internals. Added new function 591059191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 591159191Skris print, verify and generate SPKACs. Based on an original idea from 591259191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 591359191Skris [Steve Henson] 591459191Skris 591559191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 591659191Skris [Andy Polyakov] 591759191Skris 591859191Skris *) Allow the config file extension section to be overwritten on the 591959191Skris command line. Based on an original idea from Massimiliano Pala 592059191Skris <madwolf@comune.modena.it>. The new option is called -extensions 592159191Skris and can be applied to ca, req and x509. Also -reqexts to override 592259191Skris the request extensions in req and -crlexts to override the crl extensions 592359191Skris in ca. 592459191Skris [Steve Henson] 592559191Skris 592659191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 592759191Skris the same field multiple times by preceding it by "XXXX." for example: 592859191Skris 1.OU="Unit name 1" 592959191Skris 2.OU="Unit name 2" 593059191Skris this is the same syntax as used in the req config file. 593159191Skris [Steve Henson] 593259191Skris 593359191Skris *) Allow certificate extensions to be added to certificate requests. These 593459191Skris are specified in a 'req_extensions' option of the req section of the 593559191Skris config file. They can be printed out with the -text option to req but 593659191Skris are otherwise ignored at present. 593759191Skris [Steve Henson] 593859191Skris 593959191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 594059191Skris data read consists of only the final block it would not decrypted because 594159191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 594259191Skris A misplaced 'break' also meant the decrypted final block might not be 594359191Skris copied until the next read. 594459191Skris [Steve Henson] 594559191Skris 594659191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 594759191Skris a few extra parameters to the DH structure: these will be useful if 594859191Skris for example we want the value of 'q' or implement X9.42 DH. 594959191Skris [Steve Henson] 595059191Skris 595159191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 595259191Skris provides hooks that allow the default DSA functions or functions on a 595359191Skris "per key" basis to be replaced. This allows hardware acceleration and 595459191Skris hardware key storage to be handled without major modification to the 595559191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 595659191Skris associated functions. 595759191Skris [Steve Henson] 595859191Skris 595959191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 596059191Skris as "read only": it can't be written to and the buffer it points to will 596159191Skris not be freed. Reading from a read only BIO is much more efficient than 596259191Skris a normal memory BIO. This was added because there are several times when 596359191Skris an area of memory needs to be read from a BIO. The previous method was 596459191Skris to create a memory BIO and write the data to it, this results in two 596559191Skris copies of the data and an O(n^2) reading algorithm. There is a new 596659191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 596759191Skris an area of memory. Also modified the PKCS#7 routines to use read only 596859191Skris memory BIOs. 596959191Skris [Steve Henson] 597059191Skris 597159191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 597259191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 597359191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 597459191Skris but a retry condition occured while trying to read the rest. 597559191Skris [Bodo Moeller] 597659191Skris 597759191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 597859191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 597959191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 598059191Skris the encrypted data type: this is a more sensible place to put it and it 598159191Skris allows the PKCS#12 code to be tidied up that duplicated this 598259191Skris functionality. 598359191Skris [Steve Henson] 598459191Skris 598559191Skris *) Changed obj_dat.pl script so it takes its input and output files on 598659191Skris the command line. This should avoid shell escape redirection problems 598759191Skris under Win32. 598859191Skris [Steve Henson] 598959191Skris 599059191Skris *) Initial support for certificate extension requests, these are included 599159191Skris in things like Xenroll certificate requests. Included functions to allow 599259191Skris extensions to be obtained and added. 599359191Skris [Steve Henson] 599459191Skris 599559191Skris *) -crlf option to s_client and s_server for sending newlines as 599659191Skris CRLF (as required by many protocols). 599759191Skris [Bodo Moeller] 599859191Skris 599955714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 600055714Skris 600155714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 600255714Skris [Ralf S. Engelschall] 600355714Skris 600455714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 600555714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 600655714Skris 600755714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 600855714Skris program. 600955714Skris [Steve Henson] 601055714Skris 601155714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 601255714Skris DH parameters/keys (q is lost during that conversion, but the resulting 601355714Skris DH parameters contain its length). 601455714Skris 601555714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 601655714Skris much faster than DH_generate_parameters (which creates parameters 601755714Skris where p = 2*q + 1), and also the smaller q makes DH computations 601855714Skris much more efficient (160-bit exponentiation instead of 1024-bit 601955714Skris exponentiation); so this provides a convenient way to support DHE 602055714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 602155714Skris utter importance to use 602255714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 602355714Skris or 602455714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 602555714Skris when such DH parameters are used, because otherwise small subgroup 602655714Skris attacks may become possible! 602755714Skris [Bodo Moeller] 602855714Skris 602955714Skris *) Avoid memory leak in i2d_DHparams. 603055714Skris [Bodo Moeller] 603155714Skris 603255714Skris *) Allow the -k option to be used more than once in the enc program: 603355714Skris this allows the same encrypted message to be read by multiple recipients. 603455714Skris [Steve Henson] 603555714Skris 603655714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 603755714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 603855714Skris it will always use the numerical form of the OID, even if it has a short 603955714Skris or long name. 604055714Skris [Steve Henson] 604155714Skris 604255714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 604355714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 604455714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 604555714Skris no private key components need be present and it might store extra data 604659191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 604759191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 604859191Skris private key operations. 604955714Skris [Steve Henson] 605055714Skris 605155714Skris *) Added support for SPARC Linux. 605255714Skris [Andy Polyakov] 605355714Skris 605455714Skris *) pem_password_cb function type incompatibly changed from 605555714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 605655714Skris to 605755714Skris ....(char *buf, int size, int rwflag, void *userdata); 605855714Skris so that applications can pass data to their callbacks: 605955714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 606055714Skris additional void * argument, which is just handed through whenever 606155714Skris the password callback is called. 606259191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 606355714Skris 606455714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 606555714Skris 606655714Skris Compatibility note: As many C implementations push function arguments 606755714Skris onto the stack in reverse order, the new library version is likely to 606855714Skris interoperate with programs that have been compiled with the old 606955714Skris pem_password_cb definition (PEM_whatever takes some data that 607055714Skris happens to be on the stack as its last argument, and the callback 607155714Skris just ignores this garbage); but there is no guarantee whatsoever that 607255714Skris this will work. 607355714Skris 607455714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 607555714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 607655714Skris problems not only on Windows, but also on some Unix platforms. 607755714Skris To avoid problematic command lines, these definitions are now in an 607855714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 607955714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 608055714Skris [Bodo Moeller] 608155714Skris 608255714Skris *) MIPS III/IV assembler module is reimplemented. 608355714Skris [Andy Polyakov] 608455714Skris 608555714Skris *) More DES library cleanups: remove references to srand/rand and 608655714Skris delete an unused file. 608755714Skris [Ulf M�ller] 608855714Skris 608955714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 609055714Skris since not many people have MASM (ml) and it can be hard to obtain. 609155714Skris This is currently experimental but it seems to work OK and pass all 609255714Skris the tests. Check out INSTALL.W32 for info. 609355714Skris [Steve Henson] 609455714Skris 609555714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 609655714Skris without temporary keys kept an extra copy of the server key, 609755714Skris and connections with temporary keys did not free everything in case 609855714Skris of an error. 609955714Skris [Bodo Moeller] 610055714Skris 610155714Skris *) New function RSA_check_key and new openssl rsa option -check 610255714Skris for verifying the consistency of RSA keys. 610355714Skris [Ulf Moeller, Bodo Moeller] 610455714Skris 610555714Skris *) Various changes to make Win32 compile work: 610655714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 610755714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 610855714Skris comparison" warnings. 610955714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 611055714Skris [Steve Henson] 611155714Skris 611255714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 611355714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 611455714Skris derived keys are printed to stderr. 611555714Skris [Steve Henson] 611655714Skris 611755714Skris *) Copy the flags in ASN1_STRING_dup(). 611855714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 611955714Skris 612055714Skris *) The x509 application mishandled signing requests containing DSA 612155714Skris keys when the signing key was also DSA and the parameters didn't match. 612255714Skris 612355714Skris It was supposed to omit the parameters when they matched the signing key: 612455714Skris the verifying software was then supposed to automatically use the CA's 612555714Skris parameters if they were absent from the end user certificate. 612655714Skris 612755714Skris Omitting parameters is no longer recommended. The test was also 612855714Skris the wrong way round! This was probably due to unusual behaviour in 612955714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 613055714Skris This meant that parameters were omitted when they *didn't* match and 613155714Skris the certificate was useless. Certificates signed with 'ca' didn't have 613255714Skris this bug. 613355714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 613455714Skris 613555714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 613655714Skris The interface is as follows: 613755714Skris Applications can use 613855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 613955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 614055714Skris "off" is now the default. 614155714Skris The library internally uses 614255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 614355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 614455714Skris to disable memory-checking temporarily. 614555714Skris 614655714Skris Some inconsistent states that previously were possible (and were 614755714Skris even the default) are now avoided. 614855714Skris 614955714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 615055714Skris with each memory chunk allocated; this is occasionally more helpful 615155714Skris than just having a counter. 615255714Skris 615355714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 615455714Skris 615555714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 615655714Skris extensions. 615755714Skris [Bodo Moeller] 615855714Skris 615955714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 616055714Skris which largely parallels "options", but is for changing API behaviour, 616155714Skris whereas "options" are about protocol behaviour. 616255714Skris Initial "mode" flags are: 616355714Skris 616455714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 616555714Skris a single record has been written. 616655714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 616755714Skris retries use the same buffer location. 616855714Skris (But all of the contents must be 616955714Skris copied!) 617055714Skris [Bodo Moeller] 617155714Skris 617279998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 617355714Skris worked. 617455714Skris 617555714Skris *) Fix problems with no-hmac etc. 617655714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 617755714Skris 617855714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 617955714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 618055714Skris to mess around with the internals of an RSA structure. 618155714Skris [Steve Henson] 618255714Skris 618355714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 618455714Skris Also really enable memory leak checks in openssl.c and in some 618555714Skris test programs. 618655714Skris [Chad C. Mulligan, Bodo Moeller] 618755714Skris 618855714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 618955714Skris up the length of negative integers. This has now been simplified to just 619055714Skris store the length when it is first determined and use it later, rather 619155714Skris than trying to keep track of where data is copied and updating it to 619255714Skris point to the end. 619355714Skris [Steve Henson, reported by Brien Wheeler 619455714Skris <bwheeler@authentica-security.com>] 619555714Skris 619655714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 619755714Skris of a PKCS#7 signature but with the signing certificate passed to the 619855714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 619955714Skris certificate is present in the PKCS#7 structure. This isn't always the 620055714Skris case: certificates can be omitted from a PKCS#7 structure and be 620155714Skris distributed by "out of band" means (such as a certificate database). 620255714Skris [Steve Henson] 620355714Skris 620455714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 620555714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 620655714Skris necessary function names. 620755714Skris [Steve Henson] 620855714Skris 620955714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 621055714Skris options set by Configure in the top level Makefile, and Configure 621155714Skris was not even able to write more than one option correctly. 621255714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 621355714Skris [Bodo Moeller] 621455714Skris 621555714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 621655714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 621755714Skris for example allow memory BIOs to contain config info. 621855714Skris [Steve Henson] 621955714Skris 622055714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 622155714Skris Whoever hopes to achieve shared-library compatibility across versions 622255714Skris must use this, not the compile-time macro. 622355714Skris (Exercise 0.9.4: Which is the minimum library version required by 622455714Skris such programs?) 622555714Skris Note: All this applies only to multi-threaded programs, others don't 622655714Skris need locks. 622755714Skris [Bodo Moeller] 622855714Skris 622955714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 623055714Skris through a BIO pair triggered the default case, i.e. 623155714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 623255714Skris [Bodo Moeller] 623355714Skris 623455714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 623555714Skris can use the SSL library even if none of the specific BIOs is 623655714Skris appropriate. 623755714Skris [Bodo Moeller] 623855714Skris 623955714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 624055714Skris for the encoded length. 624155714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 624255714Skris 624355714Skris *) Add initial documentation of the X509V3 functions. 624455714Skris [Steve Henson] 624555714Skris 624655714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 624755714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 624855714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 624955714Skris secure PKCS#8 private key format with a high iteration count. 625055714Skris [Steve Henson] 625155714Skris 625255714Skris *) Fix determination of Perl interpreter: A perl or perl5 625355714Skris _directory_ in $PATH was also accepted as the interpreter. 625455714Skris [Ralf S. Engelschall] 625555714Skris 625655714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 625755714Skris wrong with it but it was very old and did things like calling 625855714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 625955714Skris unusual formatting. 626055714Skris [Steve Henson] 626155714Skris 626255714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 626355714Skris to use the new extension code. 626455714Skris [Steve Henson] 626555714Skris 626655714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 626755714Skris with macros. This should make it easier to change their form, add extra 626855714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 626955714Skris constant. 627055714Skris [Steve Henson] 627155714Skris 627255714Skris *) Add to configuration table a new entry that can specify an alternative 627355714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 627455714Skris according to Mark Crispin <MRC@Panda.COM>. 627555714Skris [Bodo Moeller] 627655714Skris 627755714Skris#if 0 627855714Skris *) DES CBC did not update the IV. Weird. 627955714Skris [Ben Laurie] 628055714Skris#else 628155714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 628255714Skris Changing the behaviour of the former might break existing programs -- 628355714Skris where IV updating is needed, des_ncbc_encrypt can be used. 628455714Skris#endif 628555714Skris 628655714Skris *) When bntest is run from "make test" it drives bc to check its 628755714Skris calculations, as well as internally checking them. If an internal check 628855714Skris fails, it needs to cause bc to give a non-zero result or make test carries 628955714Skris on without noticing the failure. Fixed. 629055714Skris [Ben Laurie] 629155714Skris 629255714Skris *) DES library cleanups. 629355714Skris [Ulf M�ller] 629455714Skris 629555714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 629655714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 629755714Skris ciphers. NOTE: although the key derivation function has been verified 629855714Skris against some published test vectors it has not been extensively tested 629955714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 630055714Skris of v2.0. 630155714Skris [Steve Henson] 630255714Skris 630355714Skris *) Instead of "mkdir -p", which is not fully portable, use new 630455714Skris Perl script "util/mkdir-p.pl". 630555714Skris [Bodo Moeller] 630655714Skris 630755714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 630855714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 630955714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 631055714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 631155714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 631255714Skris underlying key generation function so it must do its own ASN1 parsing. 631355714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 631455714Skris 'parameter' argument instead of literal salt and iteration count values 631555714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 631655714Skris [Steve Henson] 631755714Skris 631855714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 631955714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 632055714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 632155714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 632255714Skris value was just used as a "magic string" and not used directly its 632355714Skris value doesn't matter. 632455714Skris [Steve Henson] 632555714Skris 632655714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 632755714Skris support mutable. 632855714Skris [Ben Laurie] 632955714Skris 633055714Skris *) "linux-sparc64" configuration (ultrapenguin). 633155714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 633255714Skris "linux-sparc" configuration. 633355714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 633455714Skris 633555714Skris *) config now generates no-xxx options for missing ciphers. 633655714Skris [Ulf M�ller] 633755714Skris 633855714Skris *) Support the EBCDIC character set (work in progress). 633955714Skris File ebcdic.c not yet included because it has a different license. 634055714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 634155714Skris 634255714Skris *) Support BS2000/OSD-POSIX. 634355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 634455714Skris 634555714Skris *) Make callbacks for key generation use void * instead of char *. 634655714Skris [Ben Laurie] 634755714Skris 634855714Skris *) Make S/MIME samples compile (not yet tested). 634955714Skris [Ben Laurie] 635055714Skris 635155714Skris *) Additional typesafe stacks. 635255714Skris [Ben Laurie] 635355714Skris 635455714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 635555714Skris [Bodo Moeller] 635655714Skris 635755714Skris 635855714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 635955714Skris 636055714Skris *) New configuration variant "sco5-gcc". 636155714Skris 636255714Skris *) Updated some demos. 636355714Skris [Sean O Riordain, Wade Scholine] 636455714Skris 636555714Skris *) Add missing BIO_free at exit of pkcs12 application. 636655714Skris [Wu Zhigang] 636755714Skris 636855714Skris *) Fix memory leak in conf.c. 636955714Skris [Steve Henson] 637055714Skris 637155714Skris *) Updates for Win32 to assembler version of MD5. 637255714Skris [Steve Henson] 637355714Skris 637455714Skris *) Set #! path to perl in apps/der_chop to where we found it 637555714Skris instead of using a fixed path. 637655714Skris [Bodo Moeller] 637755714Skris 637855714Skris *) SHA library changes for irix64-mips4-cc. 637955714Skris [Andy Polyakov] 638055714Skris 638155714Skris *) Improvements for VMS support. 638255714Skris [Richard Levitte] 638355714Skris 638455714Skris 638555714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 638655714Skris 638755714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 638855714Skris This also avoids the problems with SC4.2 and unpatched SC5. 638955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 639055714Skris 639155714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 639255714Skris These are required because of the typesafe stack would otherwise break 639355714Skris existing code. If old code used a structure member which used to be STACK 639455714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 639555714Skris sk_num or sk_value it would produce an error because the num, data members 639655714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 639755714Skris replaces the old method of assigning a value to sk_value 639855714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 639955714Skris that does this will no longer work (and should use sk_set instead) but 640055714Skris this could be regarded as a "questionable" behaviour anyway. 640155714Skris [Steve Henson] 640255714Skris 640355714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 640455714Skris correctly handle encrypted S/MIME data. 640555714Skris [Steve Henson] 640655714Skris 640755714Skris *) Change type of various DES function arguments from des_cblock 640855714Skris (which means, in function argument declarations, pointer to char) 640955714Skris to des_cblock * (meaning pointer to array with 8 char elements), 641055714Skris which allows the compiler to do more typechecking; it was like 641155714Skris that back in SSLeay, but with lots of ugly casts. 641255714Skris 641355714Skris Introduce new type const_des_cblock. 641455714Skris [Bodo Moeller] 641555714Skris 641655714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 641755714Skris problems: find RecipientInfo structure that matches recipient certificate 641855714Skris and initialise the ASN1 structures properly based on passed cipher. 641955714Skris [Steve Henson] 642055714Skris 642155714Skris *) Belatedly make the BN tests actually check the results. 642255714Skris [Ben Laurie] 642355714Skris 642455714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 642555714Skris to and from BNs: it was completely broken. New compilation option 642655714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 642755714Skris key elements as negative integers. 642855714Skris [Steve Henson] 642955714Skris 643055714Skris *) Reorganize and speed up MD5. 643155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 643255714Skris 643355714Skris *) VMS support. 643455714Skris [Richard Levitte <richard@levitte.org>] 643555714Skris 643655714Skris *) New option -out to asn1parse to allow the parsed structure to be 643755714Skris output to a file. This is most useful when combined with the -strparse 643855714Skris option to examine the output of things like OCTET STRINGS. 643955714Skris [Steve Henson] 644055714Skris 644155714Skris *) Make SSL library a little more fool-proof by not requiring any longer 644255714Skris that SSL_set_{accept,connect}_state be called before 644355714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 644455714Skris in many applications because usually everything *appeared* to work as 644555714Skris intended anyway -- now it really works as intended). 644655714Skris [Bodo Moeller] 644755714Skris 644855714Skris *) Move openssl.cnf out of lib/. 644955714Skris [Ulf M�ller] 645055714Skris 645155714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 645255714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 645355714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 645455714Skris [Ralf S. Engelschall] 645555714Skris 645655714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 645755714Skris handle PKCS#7 enveloped data properly. 645855714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 645955714Skris 646055714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 646155714Skris copying pointers. The cert_st handling is changed by this in 646255714Skris various ways (and thus what used to be known as ctx->default_cert 646355714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 646455714Skris any longer when s->cert does not give us what we need). 646555714Skris ssl_cert_instantiate becomes obsolete by this change. 646655714Skris As soon as we've got the new code right (possibly it already is?), 646755714Skris we have solved a couple of bugs of the earlier code where s->cert 646855714Skris was used as if it could not have been shared with other SSL structures. 646955714Skris 647055714Skris Note that using the SSL API in certain dirty ways now will result 647155714Skris in different behaviour than observed with earlier library versions: 647255714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 647355714Skris does not influence s as it used to. 647455714Skris 647555714Skris In order to clean up things more thoroughly, inside SSL_SESSION 647655714Skris we don't use CERT any longer, but a new structure SESS_CERT 647755714Skris that holds per-session data (if available); currently, this is 647855714Skris the peer's certificate chain and, for clients, the server's certificate 647955714Skris and temporary key. CERT holds only those values that can have 648055714Skris meaningful defaults in an SSL_CTX. 648155714Skris [Bodo Moeller] 648255714Skris 648355714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 648455714Skris from the internal representation. Various PKCS#7 fixes: remove some 648555714Skris evil casts and set the enc_dig_alg field properly based on the signing 648655714Skris key type. 648755714Skris [Steve Henson] 648855714Skris 648955714Skris *) Allow PKCS#12 password to be set from the command line or the 649055714Skris environment. Let 'ca' get its config file name from the environment 649155714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 649255714Skris and 'x509'). 649355714Skris [Steve Henson] 649455714Skris 649555714Skris *) Allow certificate policies extension to use an IA5STRING for the 649655714Skris organization field. This is contrary to the PKIX definition but 649755714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 649855714Skris extension option. 649955714Skris [Steve Henson] 650055714Skris 650155714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 650255714Skris without disallowing inline assembler and the like for non-pedantic builds. 650355714Skris [Ben Laurie] 650455714Skris 650555714Skris *) Support Borland C++ builder. 650655714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 650755714Skris 650855714Skris *) Support Mingw32. 650955714Skris [Ulf M�ller] 651055714Skris 651155714Skris *) SHA-1 cleanups and performance enhancements. 651255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 651355714Skris 651455714Skris *) Sparc v8plus assembler for the bignum library. 651555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 651655714Skris 651755714Skris *) Accept any -xxx and +xxx compiler options in Configure. 651855714Skris [Ulf M�ller] 651955714Skris 652055714Skris *) Update HPUX configuration. 652155714Skris [Anonymous] 652255714Skris 652355714Skris *) Add missing sk_<type>_unshift() function to safestack.h 652455714Skris [Ralf S. Engelschall] 652555714Skris 652655714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 652755714Skris "extra_cert"s in addition to the certificate. (This makes sense 652855714Skris only for "PEM" format files, as chains as a whole are not 652955714Skris DER-encoded.) 653055714Skris [Bodo Moeller] 653155714Skris 653255714Skris *) Support verify_depth from the SSL API. 653355714Skris x509_vfy.c had what can be considered an off-by-one-error: 653455714Skris Its depth (which was not part of the external interface) 653555714Skris was actually counting the number of certificates in a chain; 653655714Skris now it really counts the depth. 653755714Skris [Bodo Moeller] 653855714Skris 653955714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 654055714Skris instead of X509err, which often resulted in confusing error 654155714Skris messages since the error codes are not globally unique 654255714Skris (e.g. an alleged error in ssl3_accept when a certificate 654355714Skris didn't match the private key). 654455714Skris 654555714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 654655714Skris value (so that you don't need SSL_set_session_id_context for each 654755714Skris connection using the SSL_CTX). 654855714Skris [Bodo Moeller] 654955714Skris 655055714Skris *) OAEP decoding bug fix. 655155714Skris [Ulf M�ller] 655255714Skris 655355714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 655455714Skris David Harris. 655555714Skris [Bodo Moeller] 655655714Skris 655755714Skris *) New Configure options "threads" and "no-threads". For systems 655855714Skris where the proper compiler options are known (currently Solaris 655955714Skris and Linux), "threads" is the default. 656055714Skris [Bodo Moeller] 656155714Skris 656255714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 656355714Skris [Bodo Moeller] 656455714Skris 656555714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 656655714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 656755714Skris such as /usr/local/bin. 656855714Skris [Bodo Moeller] 656955714Skris 657055714Skris *) "make linux-shared" to build shared libraries. 657155714Skris [Niels Poppe <niels@netbox.org>] 657255714Skris 657355714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 657455714Skris [Ulf M�ller] 657555714Skris 657655714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 657755714Skris extension adding in x509 utility. 657855714Skris [Steve Henson] 657955714Skris 658055714Skris *) Remove NOPROTO sections and error code comments. 658155714Skris [Ulf M�ller] 658255714Skris 658355714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 658455714Skris prototypes. 658555714Skris [Steve Henson] 658655714Skris 658755714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 658855714Skris [Ulf M�ller] 658955714Skris 659055714Skris *) Complete rewrite of the error code script(s). It is all now handled 659155714Skris by one script at the top level which handles error code gathering, 659255714Skris header rewriting and C source file generation. It should be much better 659355714Skris than the old method: it now uses a modified version of Ulf's parser to 659455714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 659555714Skris aren't needed for error creation any more) and do a better job of 659655714Skris translating function codes into names. The old 'ASN1 error code imbedded 659755714Skris in a comment' is no longer necessary and it doesn't use .err files which 659855714Skris have now been deleted. Also the error code call doesn't have to appear all 659955714Skris on one line (which resulted in some large lines...). 660055714Skris [Steve Henson] 660155714Skris 660255714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 660355714Skris [Bodo Moeller] 660455714Skris 660555714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 660655714Skris 0 (which usually indicates a closed connection), but continue reading. 660755714Skris [Bodo Moeller] 660855714Skris 660955714Skris *) Fix some race conditions. 661055714Skris [Bodo Moeller] 661155714Skris 661255714Skris *) Add support for CRL distribution points extension. Add Certificate 661355714Skris Policies and CRL distribution points documentation. 661455714Skris [Steve Henson] 661555714Skris 661655714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 661755714Skris [Ulf M�ller] 661855714Skris 661955714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 662055714Skris 8 of keying material. Merlin has also confirmed interop with this fix 662155714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 662255714Skris [Merlin Hughes <merlin@baltimore.ie>] 662355714Skris 662455714Skris *) Fix lots of warnings. 662555714Skris [Richard Levitte <levitte@stacken.kth.se>] 662655714Skris 662755714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 662855714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 662955714Skris [Richard Levitte <levitte@stacken.kth.se>] 663055714Skris 663155714Skris *) Fix problems with sizeof(long) == 8. 663255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 663355714Skris 663455714Skris *) Change functions to ANSI C. 663555714Skris [Ulf M�ller] 663655714Skris 663755714Skris *) Fix typos in error codes. 663855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 663955714Skris 664055714Skris *) Remove defunct assembler files from Configure. 664155714Skris [Ulf M�ller] 664255714Skris 664355714Skris *) SPARC v8 assembler BIGNUM implementation. 664455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 664555714Skris 664655714Skris *) Support for Certificate Policies extension: both print and set. 664755714Skris Various additions to support the r2i method this uses. 664855714Skris [Steve Henson] 664955714Skris 665055714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 665155714Skris return a const string when you are expecting an allocated buffer. 665255714Skris [Ben Laurie] 665355714Skris 665455714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 665555714Skris types DirectoryString and DisplayText. 665655714Skris [Steve Henson] 665755714Skris 665855714Skris *) Add code to allow r2i extensions to access the configuration database, 665955714Skris add an LHASH database driver and add several ctx helper functions. 666055714Skris [Steve Henson] 666155714Skris 666255714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 666355714Skris fail when they extended the size of a BIGNUM. 666455714Skris [Steve Henson] 666555714Skris 666655714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 666755714Skris support typesafe stack. 666855714Skris [Steve Henson] 666955714Skris 667055714Skris *) Fix typo in SSL_[gs]et_options(). 667155714Skris [Nils Frostberg <nils@medcom.se>] 667255714Skris 667355714Skris *) Delete various functions and files that belonged to the (now obsolete) 667455714Skris old X509V3 handling code. 667555714Skris [Steve Henson] 667655714Skris 667755714Skris *) New Configure option "rsaref". 667855714Skris [Ulf M�ller] 667955714Skris 668055714Skris *) Don't auto-generate pem.h. 668155714Skris [Bodo Moeller] 668255714Skris 668355714Skris *) Introduce type-safe ASN.1 SETs. 668455714Skris [Ben Laurie] 668555714Skris 668655714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 668755714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 668855714Skris 668955714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 669055714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 669155714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 669255714Skris few STACKed things have been converted already. Feel free to convert more. 669355714Skris In the fullness of time, I'll do away with the STACK type altogether. 669455714Skris [Ben Laurie] 669555714Skris 669655714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 669755714Skris specified in <certfile> by updating the entry in the index.txt file. 669855714Skris This way one no longer has to edit the index.txt file manually for 669955714Skris revoking a certificate. The -revoke option does the gory details now. 670055714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 670155714Skris 670255714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 670355714Skris `-text' option at all and this way the `-noout -text' combination was 670455714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 670555714Skris [Ralf S. Engelschall] 670655714Skris 670755714Skris *) Make sure a corresponding plain text error message exists for the 670855714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 670955714Skris verify callback function determined that a certificate was revoked. 671055714Skris [Ralf S. Engelschall] 671155714Skris 671255714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 671355714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 671455714Skris all available cipers including rc5, which was forgotten until now. 671555714Skris In order to let the testing shell script know which algorithms 671655714Skris are available, a new (up to now undocumented) command 671755714Skris "openssl list-cipher-commands" is used. 671855714Skris [Bodo Moeller] 671955714Skris 672055714Skris *) Bugfix: s_client occasionally would sleep in select() when 672155714Skris it should have checked SSL_pending() first. 672255714Skris [Bodo Moeller] 672355714Skris 672455714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 672555714Skris the raw DSA values prior to ASN.1 encoding. 672655714Skris [Ulf M�ller] 672755714Skris 672855714Skris *) Tweaks to Configure 672955714Skris [Niels Poppe <niels@netbox.org>] 673055714Skris 673155714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 673255714Skris yet... 673355714Skris [Steve Henson] 673455714Skris 673555714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 673655714Skris [Ulf M�ller] 673755714Skris 673855714Skris *) New config option to avoid instructions that are illegal on the 80386. 673955714Skris The default code is faster, but requires at least a 486. 674055714Skris [Ulf M�ller] 674155714Skris 674255714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 674355714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 674455714Skris same as SSL2_VERSION anyway. 674555714Skris [Bodo Moeller] 674655714Skris 674755714Skris *) New "-showcerts" option for s_client. 674855714Skris [Bodo Moeller] 674955714Skris 675055714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 675155714Skris application. Various cleanups and fixes. 675255714Skris [Steve Henson] 675355714Skris 675455714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 675555714Skris modify error routines to work internally. Add error codes and PBE init 675655714Skris to library startup routines. 675755714Skris [Steve Henson] 675855714Skris 675955714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 676055714Skris packing functions to asn1 and evp. Changed function names and error 676155714Skris codes along the way. 676255714Skris [Steve Henson] 676355714Skris 676455714Skris *) PKCS12 integration: and so it begins... First of several patches to 676555714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 676655714Skris objects to objects.h 676755714Skris [Steve Henson] 676855714Skris 676955714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 677055714Skris and display support for Thawte strong extranet extension. 677155714Skris [Steve Henson] 677255714Skris 677355714Skris *) Add LinuxPPC support. 677455714Skris [Jeff Dubrule <igor@pobox.org>] 677555714Skris 677655714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 677755714Skris bn_div_words in alpha.s. 677855714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 677955714Skris 678055714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 678155714Skris OAEP isn't supported when OpenSSL is built with RSAref. 678255714Skris [Ulf Moeller <ulf@fitug.de>] 678355714Skris 678455714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 678555714Skris so they no longer are missing under -DNOPROTO. 678655714Skris [Soren S. Jorvang <soren@t.dk>] 678755714Skris 678855714Skris 678955714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 679055714Skris 679155714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 679255714Skris doesn't work when the session is reused. Coming soon! 679355714Skris [Ben Laurie] 679455714Skris 679555714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 679655714Skris context thus bypassing client cert protection! All software that uses 679755714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 679855714Skris allow session reuse! A fuller solution is in the works. 679955714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 680055714Skris 680155714Skris *) Some more source tree cleanups (removed obsolete files 680255714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 680355714Skris permission on "config" script to be executable) and a fix for the INSTALL 680455714Skris document. 680555714Skris [Ulf Moeller <ulf@fitug.de>] 680655714Skris 680755714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 680855714Skris Malloc, Free. 680955714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 681055714Skris 681155714Skris *) Make rsa_oaep_test return non-zero on error. 681255714Skris [Ulf Moeller <ulf@fitug.de>] 681355714Skris 681455714Skris *) Add support for native Solaris shared libraries. Configure 681555714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 681655714Skris if someone would make that last step automatic. 681755714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 681855714Skris 681955714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 682055714Skris [Ben Laurie] 682155714Skris 682255714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 682355714Skris except NULL ciphers". This means the default cipher list will no longer 682455714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 682555714Skris the string "DEFAULT:eNULL". 682655714Skris [Steve Henson] 682755714Skris 682855714Skris *) Fix to RSA private encryption routines: if p < q then it would 682955714Skris occasionally produce an invalid result. This will only happen with 683055714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 683155714Skris [Steve Henson] 683255714Skris 683355714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 683455714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 683555714Skris because this way one can also use an interpreter named `perl5' (which is 683655714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 683755714Skris installed as `perl'). 683855714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 683955714Skris 684055714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 684155714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 684255714Skris 684355714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 684455714Skris advapi32.lib to Win32 build and change the pem test comparision 684555714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 684655714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 684755714Skris and crypto/des/ede_cbcm_enc.c. 684855714Skris [Steve Henson] 684955714Skris 685055714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 685155714Skris [Ben Laurie] 685255714Skris 685355714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 685455714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 685555714Skris is horrible: I feel ill.... 685655714Skris [Steve Henson] 685755714Skris 685855714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 685955714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 686055714Skris sections: 10 functions were absent from non ANSI section and not exported 686155714Skris from Windows DLLs. Fixed up libeay.num for new functions. 686255714Skris [Steve Henson] 686355714Skris 686455714Skris *) Make `openssl version' output lines consistent. 686555714Skris [Ralf S. Engelschall] 686655714Skris 686755714Skris *) Fix Win32 symbol export lists for BIO functions: Added 686855714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 686955714Skris to ms/libeay{16,32}.def. 687055714Skris [Ralf S. Engelschall] 687155714Skris 687255714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 687355714Skris fine under Unix and passes some trivial tests I've now added. But the 687455714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 687555714Skris added to make sure no one expects that this stuff really works in the 687655714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 687755714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 687855714Skris openssl_bio.xs. 687955714Skris [Ralf S. Engelschall] 688055714Skris 688155714Skris *) Fix the generation of two part addresses in perl. 688255714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 688355714Skris 688455714Skris *) Add config entry for Linux on MIPS. 688555714Skris [John Tobey <jtobey@channel1.com>] 688655714Skris 688755714Skris *) Make links whenever Configure is run, unless we are on Windoze. 688855714Skris [Ben Laurie] 688955714Skris 689055714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 689155714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 689255714Skris in CRLs. 689355714Skris [Steve Henson] 689455714Skris 689555714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 689655714Skris other platforms details on the command line without having to patch the 689755714Skris Configure script everytime: One now can use ``perl Configure 689855714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 689955714Skris to them (seperated by colons). This is treated as there would be a static 690055714Skris pre-configured entry in Configure's %table under key <id> with value 690155714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 690255714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 690355714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 690455714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 690555714Skris [Ralf S. Engelschall] 690655714Skris 690755714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 690855714Skris [Ben Laurie] 690955714Skris 691055714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 691155714Skris on the `perl Configure ...' command line. This way one can compile 691255714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 691355714Skris for linking it into DSOs. 691455714Skris [Ralf S. Engelschall] 691555714Skris 691655714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 691755714Skris Fixed. 691855714Skris [Ben Laurie] 691955714Skris 692055714Skris *) Cleaned up the LICENSE document: The official contact for any license 692155714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 692255714Skris And add a paragraph about the dual-license situation to make sure people 692355714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 692455714Skris to the OpenSSL toolkit. 692555714Skris [Ralf S. Engelschall] 692655714Skris 692755714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 692855714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 692955714Skris Additonally cleaned up the `make links' target: Remove unnecessary 693055714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 693155714Skris to speed processing and no longer clutter the display with confusing 693255714Skris stuff. Instead only the actually done links are displayed. 693355714Skris [Ralf S. Engelschall] 693455714Skris 693555714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 693655714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 693755714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 693855714Skris encryption. 693955714Skris [Ben Laurie] 694055714Skris 694155714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 694255714Skris signed attributes when verifying signatures (this would break them), 694355714Skris the detached data encoding was wrong and public keys obtained using 694455714Skris X509_get_pubkey() weren't freed. 694555714Skris [Steve Henson] 694655714Skris 694755714Skris *) Add text documentation for the BUFFER functions. Also added a work around 694855714Skris to a Win95 console bug. This was triggered by the password read stuff: the 694955714Skris last character typed gets carried over to the next fread(). If you were 695055714Skris generating a new cert request using 'req' for example then the last 695155714Skris character of the passphrase would be CR which would then enter the first 695255714Skris field as blank. 695355714Skris [Steve Henson] 695455714Skris 695555714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 695655714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 695755714Skris button and can be used by applications based on OpenSSL to show the 695855714Skris relationship to the OpenSSL project. 695955714Skris [Ralf S. Engelschall] 696055714Skris 696155714Skris *) Remove confusing variables in function signatures in files 696255714Skris ssl/ssl_lib.c and ssl/ssl.h. 696355714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 696455714Skris 696555714Skris *) Don't install bss_file.c under PREFIX/include/ 696655714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 696755714Skris 696855714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 696955714Skris functions that return function pointers and has support for NT specific 697055714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 697155714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 697255714Skris unsigned to signed types: this was killing the Win32 compile. 697355714Skris [Steve Henson] 697455714Skris 697555714Skris *) Add new certificate file to stack functions, 697655714Skris SSL_add_dir_cert_subjects_to_stack() and 697755714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 697855714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 697955714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 698055714Skris This means that Apache-SSL and similar packages don't have to mess around 698155714Skris to add as many CAs as they want to the preferred list. 698255714Skris [Ben Laurie] 698355714Skris 698455714Skris *) Experiment with doxygen documentation. Currently only partially applied to 698555714Skris ssl/ssl_lib.c. 698655714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 698755714Skris openssl.doxy as the configuration file. 698855714Skris [Ben Laurie] 698955714Skris 699055714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 699155714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 699255714Skris 699355714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 699455714Skris compiled in by default: it has problems with large keys. 699555714Skris [Steve Henson] 699655714Skris 699755714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 699855714Skris DH private keys and/or callback functions which directly correspond to 699955714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 700055714Skris is needed for applications which have to configure certificates on a 700155714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 700255714Skris (e.g. s_server). 700355714Skris For the RSA certificate situation is makes no difference, but 700455714Skris for the DSA certificate situation this fixes the "no shared cipher" 700555714Skris problem where the OpenSSL cipher selection procedure failed because the 700655714Skris temporary keys were not overtaken from the context and the API provided 700755714Skris no way to reconfigure them. 700855714Skris The new functions now let applications reconfigure the stuff and they 700955714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 701055714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 701155714Skris non-public-API function ssl_cert_instantiate() is used as a helper 701255714Skris function and also to reduce code redundancy inside ssl_rsa.c. 701355714Skris [Ralf S. Engelschall] 701455714Skris 701555714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 701655714Skris area because they are useful for the DSA situation and should be 701755714Skris recognized by the users. 701855714Skris [Ralf S. Engelschall] 701955714Skris 702055714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 702155714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 702255714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 702355714Skris already masked variable. 702455714Skris [Richard Levitte <levitte@stacken.kth.se>] 702555714Skris 702655714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 702755714Skris [Richard Levitte <levitte@stacken.kth.se>] 702855714Skris 702955714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 703055714Skris from `int' to `unsigned int' because it's a length and initialized by 703155714Skris EVP_DigestFinal() which expects an `unsigned int *'. 703255714Skris [Richard Levitte <levitte@stacken.kth.se>] 703355714Skris 703455714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 703555714Skris script. Instead use the usual Shell->Perl transition trick. 703655714Skris [Ralf S. Engelschall] 703755714Skris 703855714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 703955714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 704055714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 704155714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 704255714Skris currently the public key is printed (a decision which was already done by 704355714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 704455714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 704555714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 704655714Skris now, too. 704755714Skris [Ralf S. Engelschall] 704855714Skris 704955714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 705055714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 705155714Skris [Arne Ansper <arne@ats.cyber.ee>] 705255714Skris 705355714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 705455714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 705555714Skris config file. 705655714Skris [Steve Henson] 705755714Skris 705855714Skris *) Add cool BIO that does syslog (or event log on NT). 705955714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 706055714Skris 706155714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 706255714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 706355714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 706455714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 706555714Skris [Ben Laurie] 706655714Skris 706755714Skris *) Add preliminary config info for new extension code. 706855714Skris [Steve Henson] 706955714Skris 707055714Skris *) Make RSA_NO_PADDING really use no padding. 707155714Skris [Ulf Moeller <ulf@fitug.de>] 707255714Skris 707355714Skris *) Generate errors when private/public key check is done. 707455714Skris [Ben Laurie] 707555714Skris 707655714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 707755714Skris for some CRL extensions and new objects added. 707855714Skris [Steve Henson] 707955714Skris 708055714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 708155714Skris key usage extension and fuller support for authority key id. 708255714Skris [Steve Henson] 708355714Skris 708455714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 708555714Skris padding method for RSA, which is recommended for new applications in PKCS 708655714Skris #1 v2.0 (RFC 2437, October 1998). 708755714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 708855714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 708955714Skris against Bleichbacher's attack on RSA. 709055714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 709155714Skris Ben Laurie] 709255714Skris 709355714Skris *) Updates to the new SSL compression code 709455714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 709555714Skris 709655714Skris *) Fix so that the version number in the master secret, when passed 709755714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 709855714Skris (because the server will not accept higher), that the version number 709955714Skris is 0x03,0x01, not 0x03,0x00 710055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 710155714Skris 710255714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 710355714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 710455714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 710555714Skris [Steve Henson] 710655714Skris 710755714Skris *) Support for RAW extensions where an arbitrary extension can be 710855714Skris created by including its DER encoding. See apps/openssl.cnf for 710955714Skris an example. 711055714Skris [Steve Henson] 711155714Skris 711255714Skris *) Make sure latest Perl versions don't interpret some generated C array 711355714Skris code as Perl array code in the crypto/err/err_genc.pl script. 711455714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 711555714Skris 711655714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 711755714Skris not many people have the assembler. Various Win32 compilation fixes and 711855714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 711955714Skris build instructions. 712055714Skris [Steve Henson] 712155714Skris 712255714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 712355714Skris file under Win32 and also build pem.h from pem.org. New script 712455714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 712555714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 712655714Skris [Steve Henson] 712755714Skris 712855714Skris *) Major rework of DES function declarations, in the pursuit of correctness 712955714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 713055714Skris too. You may find this causes warnings in your code. Zapping your evil 713155714Skris casts will probably fix them. Mostly. 713255714Skris [Ben Laurie] 713355714Skris 713455714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 713555714Skris obj_dat.pl. It considered a zero in an object definition to mean 713655714Skris "end of object": none of the objects in objects.h have any zeros 713755714Skris so it wasn't spotted. 713855714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 713955714Skris 714055714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 714155714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 714255714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 714355714Skris vectors if you have them. 714455714Skris [Ben Laurie] 714555714Skris 714655714Skris *) Correct calculation of key length for export ciphers (too much space was 714755714Skris allocated for null ciphers). This has not been tested! 714855714Skris [Ben Laurie] 714955714Skris 715055714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 715155714Skris message is now correct (it understands "crypto" and "ssl" on its 715255714Skris command line). There is also now an "update" option. This will update 715355714Skris the util/ssleay.num and util/libeay.num files with any new functions. 715455714Skris If you do a: 715555714Skris perl util/mkdef.pl crypto ssl update 715655714Skris it will update them. 715755714Skris [Steve Henson] 715855714Skris 715955714Skris *) Overhauled the Perl interface (perl/*): 716055714Skris - ported BN stuff to OpenSSL's different BN library 716155714Skris - made the perl/ source tree CVS-aware 716255714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 716355714Skris their history because I've copied them in the repository) 716455714Skris - removed obsolete files (the test scripts will be replaced 716555714Skris by better Test::Harness variants in the future) 716655714Skris [Ralf S. Engelschall] 716755714Skris 716855714Skris *) First cut for a very conservative source tree cleanup: 716955714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 717055714Skris where we collect the old documents and readme texts. 717155714Skris 2. remove the first part of files where I'm already sure that we no 717255714Skris longer need them because of three reasons: either they are just temporary 717355714Skris files which were left by Eric or they are preserved original files where 717455714Skris I've verified that the diff is also available in the CVS via "cvs diff 717555714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 717655714Skris the crypto/md/ stuff). 717755714Skris [Ralf S. Engelschall] 717855714Skris 717955714Skris *) More extension code. Incomplete support for subject and issuer alt 718055714Skris name, issuer and authority key id. Change the i2v function parameters 718155714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 718255714Skris what that's for :-) Fix to ASN1 macro which messed up 718355714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 718455714Skris [Steve Henson] 718555714Skris 718655714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 718755714Skris INTEGER code. 718855714Skris [Steve Henson] 718955714Skris 719055714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 719155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 719255714Skris 719355714Skris *) Make sure `make rehash' target really finds the `openssl' program. 719455714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 719555714Skris 719655714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 719755714Skris like to hear about it if this slows down other processors. 719855714Skris [Ben Laurie] 719955714Skris 720055714Skris *) Add CygWin32 platform information to Configure script. 720155714Skris [Alan Batie <batie@aahz.jf.intel.com>] 720255714Skris 720355714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 720455714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 720555714Skris 720655714Skris *) New program nseq to manipulate netscape certificate sequences 720755714Skris [Steve Henson] 720855714Skris 720955714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 721055714Skris few typos. 721155714Skris [Steve Henson] 721255714Skris 721355714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 721455714Skris but the BN code had some problems that would cause failures when 721555714Skris doing certificate verification and some other functions. 721655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 721755714Skris 721855714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 721955714Skris [Steve Henson] 722055714Skris 722155714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 722255714Skris [Steve Henson] 722355714Skris 722455714Skris *) Add several PKIX and private extended key usage OIDs. 722555714Skris [Steve Henson] 722655714Skris 722755714Skris *) Modify the 'ca' program to handle the new extension code. Modify 722855714Skris openssl.cnf for new extension format, add comments. 722955714Skris [Steve Henson] 723055714Skris 723155714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 723255714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 723355714Skris CA extensions. 723455714Skris [Steve Henson] 723555714Skris 723655714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 723755714Skris error code, add initial support to X509_print() and x509 application. 723855714Skris [Steve Henson] 723955714Skris 724055714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 724155714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 724255714Skris stuff is currently isolated and isn't even compiled yet. 724355714Skris [Steve Henson] 724455714Skris 724555714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 724655714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 724755714Skris Removed the versions check from X509 routines when loading extensions: 724855714Skris this allows certain broken certificates that don't set the version 724955714Skris properly to be processed. 725055714Skris [Steve Henson] 725155714Skris 725255714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 725355714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 725455714Skris can still be regenerated with "make depend". 725555714Skris [Ben Laurie] 725655714Skris 725755714Skris *) Spelling mistake in C version of CAST-128. 725855714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 725955714Skris 726055714Skris *) Changes to the error generation code. The perl script err-code.pl 726155714Skris now reads in the old error codes and retains the old numbers, only 726255714Skris adding new ones if necessary. It also only changes the .err files if new 726355714Skris codes are added. The makefiles have been modified to only insert errors 726455714Skris when needed (to avoid needlessly modifying header files). This is done 726555714Skris by only inserting errors if the .err file is newer than the auto generated 726655714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 726755714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 726855714Skris or delete all the .err files. 726955714Skris [Steve Henson] 727055714Skris 727155714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 727255714Skris been fixed, but is untested. The assembler versions are also fixed, but 727355714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 727455714Skris to regenerate it if needed. 727555714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 727655714Skris Hagino <itojun@kame.net>] 727755714Skris 727855714Skris *) File was opened incorrectly in randfile.c. 727955714Skris [Ulf M�ller <ulf@fitug.de>] 728055714Skris 728155714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 728255714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 728355714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 728455714Skris al: it's just almost always a UTCTime. Note this patch adds new error 728555714Skris codes so do a "make errors" if there are problems. 728655714Skris [Steve Henson] 728755714Skris 728855714Skris *) Correct Linux 1 recognition in config. 728955714Skris [Ulf M�ller <ulf@fitug.de>] 729055714Skris 729155714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 729255714Skris [Anonymous <nobody@replay.com>] 729355714Skris 729455714Skris *) Generate an error if given an empty string as a cert directory. Also 729555714Skris generate an error if handed NULL (previously returned 0 to indicate an 729655714Skris error, but didn't set one). 729755714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 729855714Skris 729955714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 730055714Skris [Ben Laurie] 730155714Skris 730255714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 730355714Skris parameters. This was causing a warning which killed off the Win32 compile. 730455714Skris [Steve Henson] 730555714Skris 730655714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 730755714Skris [Neil Costigan <neil.costigan@celocom.com>] 730855714Skris 730955714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 731055714Skris based on a text string, looking up short and long names and finally 731155714Skris "dot" format. The "dot" format stuff didn't work. Added new function 731255714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 731355714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 731455714Skris OID is not part of the table. 731555714Skris [Steve Henson] 731655714Skris 731755714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 731855714Skris X509_LOOKUP_by_alias(). 731955714Skris [Ben Laurie] 732055714Skris 732155714Skris *) Sort openssl functions by name. 732255714Skris [Ben Laurie] 732355714Skris 732455714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 732555714Skris encryption from sample DSA keys (in case anyone is interested the password 732655714Skris was "1234"). 732755714Skris [Steve Henson] 732855714Skris 732955714Skris *) Make _all_ *_free functions accept a NULL pointer. 733055714Skris [Frans Heymans <fheymans@isaserver.be>] 733155714Skris 733255714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 733355714Skris NULL pointers. 733455714Skris [Anonymous <nobody@replay.com>] 733555714Skris 733655714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 733755714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 733855714Skris 733955714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 734055714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 734155714Skris 734255714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 734355714Skris [Anonymous <nobody@replay.com>] 734455714Skris 734555714Skris *) Add prototype for temp key callback functions 734655714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 734755714Skris [Ben Laurie] 734855714Skris 734955714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 735055714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 735155714Skris [Steve Henson] 735255714Skris 735355714Skris *) X509_name_add_entry() freed the wrong thing after an error. 735455714Skris [Arne Ansper <arne@ats.cyber.ee>] 735555714Skris 735655714Skris *) rsa_eay.c would attempt to free a NULL context. 735755714Skris [Arne Ansper <arne@ats.cyber.ee>] 735855714Skris 735955714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 736055714Skris [Arne Ansper <arne@ats.cyber.ee>] 736155714Skris 736255714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 736355714Skris [Arne Ansper <arne@ats.cyber.ee>] 736455714Skris 736555714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 736655714Skris in X509_STORE_new(), but document the fact that this variable is still 736755714Skris unused in the certificate verification process. 736855714Skris [Ralf S. Engelschall] 736955714Skris 737055714Skris *) Fix the various library and apps files to free up pkeys obtained from 737155714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 737255714Skris [Steve Henson] 737355714Skris 737455714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 737555714Skris demos/maurice/example2.c work, amongst others, probably. 737655714Skris [Steve Henson and Ben Laurie] 737755714Skris 737855714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 737955714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 738055714Skris are no longer created. This way we have a single and consistent command 738155714Skris line interface `openssl <command>', similar to `cvs <command>'. 738255714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 738355714Skris 738455714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 738555714Skris BIT STRING wrapper always have zero unused bits. 738655714Skris [Steve Henson] 738755714Skris 738855714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 738955714Skris [Steve Henson] 739055714Skris 739155714Skris *) Make the top-level INSTALL documentation easier to understand. 739255714Skris [Paul Sutton] 739355714Skris 739455714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 739555714Skris make (including if user presses ^C) [Paul Sutton] 739655714Skris 739755714Skris *) Make Montgomery context stuff explicit in RSA data structure. 739855714Skris [Ben Laurie] 739955714Skris 740055714Skris *) Fix build order of pem and err to allow for generated pem.h. 740155714Skris [Ben Laurie] 740255714Skris 740355714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 740455714Skris [Ben Laurie] 740555714Skris 740655714Skris *) Enhanced the err-ins.pl script so it makes the error library number 740755714Skris global and can add a library name. This is needed for external ASN1 and 740855714Skris other error libraries. 740955714Skris [Steve Henson] 741055714Skris 741155714Skris *) Fixed sk_insert which never worked properly. 741255714Skris [Steve Henson] 741355714Skris 741455714Skris *) Fix ASN1 macros so they can handle indefinite length construted 741555714Skris EXPLICIT tags. Some non standard certificates use these: they can now 741655714Skris be read in. 741755714Skris [Steve Henson] 741855714Skris 741955714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 742055714Skris into a single doc/ssleay.txt bundle. This way the information is still 742155714Skris preserved but no longer messes up this directory. Now it's new room for 742255714Skris the new set of documenation files. 742355714Skris [Ralf S. Engelschall] 742455714Skris 742555714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 742655714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 742755714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 742855714Skris number of arguments. 742955714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 743055714Skris 743155714Skris *) Fix test data to work with the above. 743255714Skris [Ben Laurie] 743355714Skris 743455714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 743555714Skris was already fixed by Eric for 0.9.1 it seems. 743655714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 743755714Skris 743855714Skris *) Autodetect FreeBSD3. 743955714Skris [Ben Laurie] 744055714Skris 744155714Skris *) Fix various bugs in Configure. This affects the following platforms: 744255714Skris nextstep 744355714Skris ncr-scde 744455714Skris unixware-2.0 744555714Skris unixware-2.0-pentium 744655714Skris sco5-cc. 744755714Skris [Ben Laurie] 744855714Skris 744955714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 745055714Skris before they are needed. 745155714Skris [Ben Laurie] 745255714Skris 745355714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 745455714Skris [Ben Laurie] 745555714Skris 745655714Skris 745755714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 745855714Skris 745955714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 746055714Skris changed SSLeay to OpenSSL in version strings. 746155714Skris [Ralf S. Engelschall] 746255714Skris 746355714Skris *) Some fixups to the top-level documents. 746455714Skris [Paul Sutton] 746555714Skris 746655714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 746755714Skris because the symlink to include/ was missing. 746855714Skris [Ralf S. Engelschall] 746955714Skris 747055714Skris *) Incorporated the popular no-RSA/DSA-only patches 747155714Skris which allow to compile a RSA-free SSLeay. 747255714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 747355714Skris 747455714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 747555714Skris when "ssleay" is still not found. 747655714Skris [Ralf S. Engelschall] 747755714Skris 747855714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 747955714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 748055714Skris 748155714Skris *) Updated the README file. 748255714Skris [Ralf S. Engelschall] 748355714Skris 748455714Skris *) Added various .cvsignore files in the CVS repository subdirs 748555714Skris to make a "cvs update" really silent. 748655714Skris [Ralf S. Engelschall] 748755714Skris 748855714Skris *) Recompiled the error-definition header files and added 748955714Skris missing symbols to the Win32 linker tables. 749055714Skris [Ralf S. Engelschall] 749155714Skris 749255714Skris *) Cleaned up the top-level documents; 749355714Skris o new files: CHANGES and LICENSE 749455714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 749555714Skris o merged COPYRIGHT into LICENSE 749655714Skris o removed obsolete TODO file 749755714Skris o renamed MICROSOFT to INSTALL.W32 749855714Skris [Ralf S. Engelschall] 749955714Skris 750055714Skris *) Removed dummy files from the 0.9.1b source tree: 750155714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 750255714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 750355714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 750455714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 750555714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 750655714Skris [Ralf S. Engelschall] 750755714Skris 750855714Skris *) Added various platform portability fixes. 750955714Skris [Mark J. Cox] 751055714Skris 751155714Skris *) The Genesis of the OpenSSL rpject: 751255714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 751355714Skris Young and Tim J. Hudson created while they were working for C2Net until 751455714Skris summer 1998. 751555714Skris [The OpenSSL Project] 751655714Skris 751755714Skris 751855714Skris Changes between 0.9.0b and 0.9.1b [not released] 751955714Skris 752055714Skris *) Updated a few CA certificates under certs/ 752155714Skris [Eric A. Young] 752255714Skris 752355714Skris *) Changed some BIGNUM api stuff. 752455714Skris [Eric A. Young] 752555714Skris 752655714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 752755714Skris DGUX x86, Linux Alpha, etc. 752855714Skris [Eric A. Young] 752955714Skris 753055714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 753155714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 753255714Skris available). 753355714Skris [Eric A. Young] 753455714Skris 753555714Skris *) Add -strparse option to asn1pars program which parses nested 753655714Skris binary structures 753755714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 753855714Skris 753955714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 754055714Skris [Eric A. Young] 754155714Skris 754255714Skris *) DSA fix for "ca" program. 754355714Skris [Eric A. Young] 754455714Skris 754555714Skris *) Added "-genkey" option to "dsaparam" program. 754655714Skris [Eric A. Young] 754755714Skris 754855714Skris *) Added RIPE MD160 (rmd160) message digest. 754955714Skris [Eric A. Young] 755055714Skris 755155714Skris *) Added -a (all) option to "ssleay version" command. 755255714Skris [Eric A. Young] 755355714Skris 755455714Skris *) Added PLATFORM define which is the id given to Configure. 755555714Skris [Eric A. Young] 755655714Skris 755755714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 755855714Skris [Eric A. Young] 755955714Skris 756055714Skris *) Extended the ASN.1 parser routines. 756155714Skris [Eric A. Young] 756255714Skris 756355714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 756455714Skris [Eric A. Young] 756555714Skris 756655714Skris *) Added a BN_CTX to the BN library. 756755714Skris [Eric A. Young] 756855714Skris 756955714Skris *) Fixed the weak key values in DES library 757055714Skris [Eric A. Young] 757155714Skris 757255714Skris *) Changed API in EVP library for cipher aliases. 757355714Skris [Eric A. Young] 757455714Skris 757555714Skris *) Added support for RC2/64bit cipher. 757655714Skris [Eric A. Young] 757755714Skris 757855714Skris *) Converted the lhash library to the crypto/mem.c functions. 757955714Skris [Eric A. Young] 758055714Skris 758155714Skris *) Added more recognized ASN.1 object ids. 758255714Skris [Eric A. Young] 758355714Skris 758455714Skris *) Added more RSA padding checks for SSL/TLS. 758555714Skris [Eric A. Young] 758655714Skris 758755714Skris *) Added BIO proxy/filter functionality. 758855714Skris [Eric A. Young] 758955714Skris 759055714Skris *) Added extra_certs to SSL_CTX which can be used 759155714Skris send extra CA certificates to the client in the CA cert chain sending 759255714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 759355714Skris [Eric A. Young] 759455714Skris 759555714Skris *) Now Fortezza is denied in the authentication phase because 759655714Skris this is key exchange mechanism is not supported by SSLeay at all. 759755714Skris [Eric A. Young] 759855714Skris 759955714Skris *) Additional PKCS1 checks. 760055714Skris [Eric A. Young] 760155714Skris 760255714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 760355714Skris [Eric A. Young] 760455714Skris 760555714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 760655714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 760755714Skris [Eric A. Young] 760855714Skris 760955714Skris *) Fixed a few memory leaks. 761055714Skris [Eric A. Young] 761155714Skris 761255714Skris *) Fixed various code and comment typos. 761355714Skris [Eric A. Young] 761455714Skris 761555714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 761655714Skris bytes sent in the client random. 761755714Skris [Edward Bishop <ebishop@spyglass.com>] 761855714Skris 7619