CHANGES revision 127128
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 6127128Snectar 7127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 8127128Snectar by using the Codenomicon TLS Test Tool (CAN-2004-0079) 9127128Snectar [Joe Orton, Steve Henson] 10127128Snectar 11127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 12127128Snectar (CAN-2004-0112) 13127128Snectar [Joe Orton, Steve Henson] 14127128Snectar 15127128Snectar *) Make it possible to have multiple active certificates with the same 16127128Snectar subject in the CA index file. This is done only if the keyword 17127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 18127128Snectar if 'CA_default') of the configuration file. The value is saved 19127128Snectar with the database itself in a separate index attribute file, 20127128Snectar named like the index file with '.attr' appended to the name. 21127128Snectar [Richard Levitte] 22127128Snectar 23127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 24127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 25127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 26127128Snectar extensions: since verify currently doesn't process CRL extensions this 27127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 28127128Snectar for these cases. 29127128Snectar [Steve Henson] 30127128Snectar 31127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 32127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 33127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 34127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 35127128Snectar parsing them this should not create any compatibility issues. 36127128Snectar [Steve Henson] 37127128Snectar 38127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 39127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 40127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 41127128Snectar < 0.9.7. 42127128Snectar [Steve Henson] 43127128Snectar 44127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 45127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 46127128Snectar 47127128Snectar *) Use the correct content when signing type "other". 48127128Snectar [Steve Henson] 49127128Snectar 50120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 51120631Snectar 52120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 53120631Snectar 54120631Snectar Stop out of bounds reads in the ASN1 code when presented with 55120631Snectar invalid tags (CAN-2003-0543 and CAN-2003-0544). 56120631Snectar 57120631Snectar Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545). 58120631Snectar 59120631Snectar If verify callback ignores invalid public key errors don't try to check 60120631Snectar certificate signature with the NULL public key. 61120631Snectar 62120631Snectar [Steve Henson] 63120631Snectar 64120631Snectar *) New -ignore_err option in ocsp application to stop the server 65120631Snectar exiting on the first error in a request. 66120631Snectar [Steve Henson] 67120631Snectar 68120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 69120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 70120631Snectar specifications. 71120631Snectar [Steve Henson] 72120631Snectar 73120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 74120631Snectar extra data after the compression methods not only for TLS 1.0 75120631Snectar but also for SSL 3.0 (as required by the specification). 76120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 77120631Snectar 78120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 79120631Snectar when it's 512 *bits* long, not 512 bytes. 80120631Snectar [Richard Levitte] 81120631Snectar 82120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 83120631Snectar blocks during encryption. 84120631Snectar [Richard Levitte] 85120631Snectar 86120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 87120631Snectar flushes were not handled properly if the BIO retried. On read 88120631Snectar data was not being buffered properly and had various logic bugs. 89120631Snectar This also affects blocking I/O when the data being decoded is a 90120631Snectar certain size. 91120631Snectar [Steve Henson] 92120631Snectar 93120631Snectar *) Various S/MIME bugfixes and compatibility changes: 94120631Snectar output correct application/pkcs7 MIME type if 95120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 96120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 97120631Snectar of files as .eml work). Correctly handle very long lines in MIME 98120631Snectar parser. 99120631Snectar [Steve Henson] 100120631Snectar 101120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 102120631Snectar 103120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 104120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 105120631Snectar a protocol version number mismatch like a decryption error 106120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 107120631Snectar [Bodo Moeller] 108120631Snectar 109120631Snectar *) Turn on RSA blinding by default in the default implementation 110120631Snectar to avoid a timing attack. Applications that don't want it can call 111120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 112120631Snectar They would be ill-advised to do so in most cases. 113120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 114120631Snectar 115120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 116120631Snectar seeded (in this case, the secret RSA exponent is abused as 117120631Snectar an unpredictable seed -- if it is not unpredictable, there 118120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 119120631Snectar by remembering the creator's thread ID in rsa->blinding and 120120631Snectar having all other threads use local one-time blinding factors 121120631Snectar (this requires more computation than sharing rsa->blinding, but 122120631Snectar avoids excessive locking; and if an RSA object is not shared 123120631Snectar between threads, blinding will still be very fast). 124120631Snectar [Bodo Moeller] 125120631Snectar 126120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 127120631Snectar ENGINE as defaults for all supported algorithms irrespective of 128120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 129120631Snectar should make sure they are passing it correctly. 130120631Snectar [Geoff Thorpe] 131120631Snectar 132120631Snectar *) Target "mingw" now allows native Windows code to be generated in 133120631Snectar the Cygwin environment as well as with the MinGW compiler. 134120631Snectar [Ulf Moeller] 135120631Snectar 136111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 137111147Snectar 138111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 139111147Snectar via timing by performing a MAC computation even if incorrrect 140111147Snectar block cipher padding has been found. This is a countermeasure 141111147Snectar against active attacks where the attacker has to distinguish 142111147Snectar between bad padding and a MAC verification error. (CAN-2003-0078) 143111147Snectar 144111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 145111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 146111147Snectar Martin Vuagnoux (EPFL, Ilion)] 147111147Snectar 148111147Snectar *) Make the no-err option work as intended. The intention with no-err 149111147Snectar is not to have the whole error stack handling routines removed from 150111147Snectar libcrypto, it's only intended to remove all the function name and 151111147Snectar reason texts, thereby removing some of the footprint that may not 152111147Snectar be interesting if those errors aren't displayed anyway. 153111147Snectar 154111147Snectar NOTE: it's still possible for any application or module to have it's 155111147Snectar own set of error texts inserted. The routines are there, just not 156111147Snectar used by default when no-err is given. 157111147Snectar [Richard Levitte] 158111147Snectar 159111147Snectar *) Add support for FreeBSD on IA64. 160111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 161111147Snectar 162111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 163111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 164111147Snectar the value returned by DES_cbc_cksum() was like the one from 165111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 166111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 167111147Snectar 168111147Snectar *) Allow an application to disable the automatic SSL chain building. 169111147Snectar Before this a rather primitive chain build was always performed in 170111147Snectar ssl3_output_cert_chain(): an application had no way to send the 171111147Snectar correct chain if the automatic operation produced an incorrect result. 172111147Snectar 173111147Snectar Now the chain builder is disabled if either: 174111147Snectar 175111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 176111147Snectar 177111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 178111147Snectar 179111147Snectar The reasoning behind this is that an application would not want the 180111147Snectar auto chain building to take place if extra chain certificates are 181111147Snectar present and it might also want a means of sending no additional 182111147Snectar certificates (for example the chain has two certificates and the 183111147Snectar root is omitted). 184111147Snectar [Steve Henson] 185111147Snectar 186111147Snectar *) Add the possibility to build without the ENGINE framework. 187111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 188111147Snectar 189111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 190111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 191111147Snectar [Steve Henson] 192111147Snectar 193111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 194111147Snectar could be freed. Solution: make sure initialization is performed early 195111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 196111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 197111147Snectar [Lutz Jaenicke] 198111147Snectar 199111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 200111147Snectar checked on reconnect on the client side, therefore session resumption 201111147Snectar could still fail with a "ssl session id is different" error. This 202111147Snectar behaviour is masked when SSL_OP_ALL is used due to 203111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 204111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 205111147Snectar followup to PR #377. 206111147Snectar [Lutz Jaenicke] 207111147Snectar 208111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 209111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 210111147Snectar [Andy Polyakov] 211111147Snectar 212111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 213111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 214111147Snectar the config script, much like the NetBSD support. 215111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 216111147Snectar 217109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 218109998Smarkm 219120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 220120631Snectar OpenSSL 0.9.7.] 221120631Snectar 222109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 223109998Smarkm code (06) was taken as the first octet of the session ID and the last 224109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 225109998Smarkm caching could not have worked due to the session ID mismatch between 226109998Smarkm client and server. 227109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 228109998Smarkm PR #377. 229109998Smarkm [Lutz Jaenicke] 230109998Smarkm 231109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 232109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 233109998Smarkm removed entirely. 234109998Smarkm [Richard Levitte] 235109998Smarkm 236109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 237109998Smarkm seems that in spite of existing for more than a year, many application 238109998Smarkm author have done nothing to provide the necessary callbacks, which 239109998Smarkm means that this particular engine will not work properly anywhere. 240109998Smarkm This is a very unfortunate situation which forces us, in the name 241109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 242109998Smarkm of libcrypto. 243109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 244109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 245109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 246109998Smarkm make such changes in the libcrypto locking code that changes will 247109998Smarkm have to be made anyway). 248109998Smarkm [Richard Levitte] 249109998Smarkm 250109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 251109998Smarkm octets have been read, EOF or an error occurs. Without this change 252109998Smarkm some truncated ASN1 structures will not produce an error. 253109998Smarkm [Steve Henson] 254109998Smarkm 255109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 256109998Smarkm Still give the possibility to force the use of Heimdal, but with 257109998Smarkm warnings and a request that patches get sent to openssl-dev. 258109998Smarkm [Richard Levitte] 259109998Smarkm 260109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 261109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 262109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 263109998Smarkm 264109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 265109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 266109998Smarkm edit numbers of the version. 267109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 268109998Smarkm 269109998Smarkm *) Introduce safe string copy and catenation functions 270109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 271109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 272109998Smarkm 273109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 274109998Smarkm [Ben Laurie (CHATS)] 275109998Smarkm 276109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 277109998Smarkm resizing buffers containing secrets, and use where appropriate. 278109998Smarkm [Ben Laurie (CHATS)] 279109998Smarkm 280109998Smarkm *) Avoid using fixed size buffers for configuration file location. 281109998Smarkm [Ben Laurie (CHATS)] 282109998Smarkm 283109998Smarkm *) Avoid filename truncation for various CA files. 284109998Smarkm [Ben Laurie (CHATS)] 285109998Smarkm 286109998Smarkm *) Use sizeof in preference to magic numbers. 287109998Smarkm [Ben Laurie (CHATS)] 288109998Smarkm 289109998Smarkm *) Avoid filename truncation in cert requests. 290109998Smarkm [Ben Laurie (CHATS)] 291109998Smarkm 292109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 293109998Smarkm overflows. 294109998Smarkm [Ben Laurie (CHATS)] 295109998Smarkm 296109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 297109998Smarkm potentially lead to a spoofing attack). 298109998Smarkm [Ben Laurie (CHATS)] 299109998Smarkm 300109998Smarkm *) Fix various buffers to be large enough for hex/decimal 301109998Smarkm representations in a platform independent manner. 302109998Smarkm [Ben Laurie (CHATS)] 303109998Smarkm 304109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 305109998Smarkm resizing buffers containing secrets, and use where appropriate. 306109998Smarkm [Ben Laurie (CHATS)] 307109998Smarkm 308109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 309109998Smarkm indents. 310109998Smarkm [Ben Laurie (CHATS)] 311109998Smarkm 312109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 313109998Smarkm [Ben Laurie (CHATS)] 314109998Smarkm 315109998Smarkm *) buffer_gets() could terminate with the buffer only half 316109998Smarkm full. Fixed. 317109998Smarkm [Ben Laurie (CHATS)] 318109998Smarkm 319109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 320109998Smarkm overflowing internal buffers by having large block sizes, etc. 321109998Smarkm [Ben Laurie (CHATS)] 322109998Smarkm 323109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 324109998Smarkm unconditionally). 325109998Smarkm [Ben Laurie (CHATS)] 326109998Smarkm 327109998Smarkm *) Eliminate unused copy of key in RC4. 328109998Smarkm [Ben Laurie (CHATS)] 329109998Smarkm 330109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 331109998Smarkm [Ben Laurie (CHATS)] 332109998Smarkm 333109998Smarkm *) Fix off-by-one error in EGD path. 334109998Smarkm [Ben Laurie (CHATS)] 335109998Smarkm 336109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 337109998Smarkm [Ben Laurie (CHATS)] 338109998Smarkm 339109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 340109998Smarkm CBCParameter. 341109998Smarkm [Ben Laurie (CHATS)] 342109998Smarkm 343109998Smarkm *) Eliminate unused and dangerous function knumber(). 344109998Smarkm [Ben Laurie (CHATS)] 345109998Smarkm 346109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 347109998Smarkm [Ben Laurie (CHATS)] 348109998Smarkm 349109998Smarkm *) Protect against overlong session ID context length in an encoded 350109998Smarkm session object. Since these are local, this does not appear to be 351109998Smarkm exploitable. 352109998Smarkm [Ben Laurie (CHATS)] 353109998Smarkm 354109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 355109998Smarkm the 0.9.6 release series: 356109998Smarkm 357109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 358109998Smarkm supply an oversized master key in Kerberos-enabled versions. 359109998Smarkm (CAN-2002-0657) 360109998Smarkm [Ben Laurie (CHATS)] 361109998Smarkm 362109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 363109998Smarkm [Richard Levitte] 364109998Smarkm 365109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 366109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 367109998Smarkm 368109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 369109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 370109998Smarkm 371109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 372109998Smarkm have been removed entirely. This was also the last step to make 373109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 374109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 375109998Smarkm 376109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 377109998Smarkm to allow version independent disabling of normally unselected ciphers, 378109998Smarkm which may be activated as a side-effect of selecting a single cipher. 379109998Smarkm 380109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 381109998Smarkm out of "ALL" because they do not provide symmetric encryption. 382109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 383109998Smarkm [Lutz Jaenicke, Bodo Moeller] 384109998Smarkm 385109998Smarkm *) Add appropriate support for separate platform-dependent build 386109998Smarkm directories. The recommended way to make a platform-dependent 387109998Smarkm build directory is the following (tested on Linux), maybe with 388109998Smarkm some local tweaks: 389109998Smarkm 390109998Smarkm # Place yourself outside of the OpenSSL source tree. In 391109998Smarkm # this example, the environment variable OPENSSL_SOURCE 392109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 393109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 394109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 395111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 396109998Smarkm mkdir -p `dirname $F` 397109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 398109998Smarkm done 399109998Smarkm 400109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 401109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 402109998Smarkm it probably means the source directory is very clean. 403109998Smarkm [Richard Levitte] 404109998Smarkm 405109998Smarkm *) Make sure any ENGINE control commands make local copies of string 406109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 407109998Smarkm the caller may have overwritten (or deallocated) the original string 408109998Smarkm data when a later ENGINE operation tries to use the stored values. 409109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 410109998Smarkm 411109998Smarkm *) Improve diagnostics in file reading and command-line digests. 412109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 413109998Smarkm 414109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 415109998Smarkm error in AES-CFB decryption. 416109998Smarkm [Richard Levitte] 417109998Smarkm 418109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 419109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 420109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 421109998Smarkm BIOs and some applications. This has the side effect that 422109998Smarkm applications must explicitly clean up cipher contexts with 423109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 424109998Smarkm [Steve Henson] 425109998Smarkm 426109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 427109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 428109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 429109998Smarkm [Steve Henson] 430109998Smarkm 431109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 432109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 433109998Smarkm [Lutz Jaenicke] 434109998Smarkm 435109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 436109998Smarkm form for "surname", serialNumber has no short form. 437109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 438109998Smarkm therefore remove "mail" short name for "internet 7". 439109998Smarkm The OID for unique identifiers in X509 certificates is 440109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 441109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 442109998Smarkm [Lutz Jaenicke] 443109998Smarkm 444109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 445109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 446109998Smarkm after all ctrl commands have been executed on it. If init=1 the 447109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 448109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 449109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 450109998Smarkm [Steve Henson] 451109998Smarkm 452109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 453109998Smarkm argument is actually passed to the callback: In the 454109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 455109998Smarkm declaration has been changed from 456109998Smarkm int (*cb)() 457109998Smarkm into 458109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 459109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 460109998Smarkm i=s->ctx->app_verify_callback(&ctx) 461109998Smarkm has been changed into 462109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 463109998Smarkm 464109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 465109998Smarkm a dummy argument can be added to their callback functions. 466109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 467109998Smarkm 468109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 469109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 470109998Smarkm 471109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 472109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 473109998Smarkm This allows older applications to transparently support certain 474109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 475109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 476109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 477109998Smarkm always load it have also been added. 478109998Smarkm [Steve Henson] 479109998Smarkm 480109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 481109998Smarkm Adjust NIDs and EVP layer. 482109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 483109998Smarkm 484109998Smarkm *) Config modules support in openssl utility. 485109998Smarkm 486109998Smarkm Most commands now load modules from the config file, 487109998Smarkm though in a few (such as version) this isn't done 488109998Smarkm because it couldn't be used for anything. 489109998Smarkm 490109998Smarkm In the case of ca and req the config file used is 491109998Smarkm the same as the utility itself: that is the -config 492109998Smarkm command line option can be used to specify an 493109998Smarkm alternative file. 494109998Smarkm [Steve Henson] 495109998Smarkm 496109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 497109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 498109998Smarkm [Steve Henson] 499109998Smarkm 500109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 501109998Smarkm config section name. Add a new flag to tolerate a missing config file 502109998Smarkm and move code to CONF_modules_load_file(). 503109998Smarkm [Steve Henson] 504109998Smarkm 505109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 506109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 507109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 508109998Smarkm to work with the new engine framework. 509109998Smarkm [AEP Inc. and Richard Levitte] 510109998Smarkm 511109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 512109998Smarkm Technologies. (Use engine 'sureware') 513109998Smarkm The support was copied from 0.9.6c [engine] and adapted 514109998Smarkm to work with the new engine framework. 515109998Smarkm [Richard Levitte] 516109998Smarkm 517109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 518109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 519109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 520109998Smarkm 521109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 522109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 523109998Smarkm 524109998Smarkm *) Add the configuration target debug-linux-ppro. 525109998Smarkm Make 'openssl rsa' use the general key loading routines 526109998Smarkm implemented in apps.c, and make those routines able to 527109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 528109998Smarkm FORMAT_IISSGC. 529109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 530109998Smarkm 531109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 532109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 533109998Smarkm 534109998Smarkm *) Add -keyform to rsautl, and document -engine. 535109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 536109998Smarkm 537109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 538109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 539109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 540109998Smarkm [Ben Laurie] 541109998Smarkm 542109998Smarkm *) Add new functions 543109998Smarkm ERR_peek_last_error 544109998Smarkm ERR_peek_last_error_line 545109998Smarkm ERR_peek_last_error_line_data. 546109998Smarkm These are similar to 547109998Smarkm ERR_peek_error 548109998Smarkm ERR_peek_error_line 549109998Smarkm ERR_peek_error_line_data, 550109998Smarkm but report on the latest error recorded rather than the first one 551109998Smarkm still in the error queue. 552109998Smarkm [Ben Laurie, Bodo Moeller] 553109998Smarkm 554109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 555109998Smarkm like: 556109998Smarkm default_algorithms = ALL 557109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 558109998Smarkm [Steve Henson] 559109998Smarkm 560109998Smarkm *) Prelminary ENGINE config module. 561109998Smarkm [Steve Henson] 562109998Smarkm 563109998Smarkm *) New experimental application configuration code. 564109998Smarkm [Steve Henson] 565109998Smarkm 566109998Smarkm *) Change the AES code to follow the same name structure as all other 567109998Smarkm symmetric ciphers, and behave the same way. Move everything to 568109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 569109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 570109998Smarkm 571109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 572109998Smarkm [Ben Laurie and Theo de Raadt] 573109998Smarkm 574109998Smarkm *) Add option to output public keys in req command. 575109998Smarkm [Massimiliano Pala madwolf@openca.org] 576109998Smarkm 577109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 578109998Smarkm (up to about 10% better than before for P-192 and P-224). 579109998Smarkm [Bodo Moeller] 580109998Smarkm 581109998Smarkm *) New functions/macros 582109998Smarkm 583109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 584109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 585109998Smarkm SSL_set_msg_callback(ssl, cb) 586109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 587109998Smarkm 588109998Smarkm to request calling a callback function 589109998Smarkm 590109998Smarkm void cb(int write_p, int version, int content_type, 591109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 592109998Smarkm 593109998Smarkm whenever a protocol message has been completely received 594109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 595109998Smarkm protocol version according to which the SSL library interprets 596109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 597109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 598109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 599109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 600109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 601109998Smarkm SSL object, and 'arg' is the application-defined value set by 602109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 603109998Smarkm 604109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 605109998Smarkm to enable a callback that displays all protocol messages. 606109998Smarkm [Bodo Moeller] 607109998Smarkm 608109998Smarkm *) Change the shared library support so shared libraries are built as 609109998Smarkm soon as the corresponding static library is finished, and thereby get 610109998Smarkm openssl and the test programs linked against the shared library. 611109998Smarkm This still only happens when the keyword "shard" has been given to 612109998Smarkm the configuration scripts. 613109998Smarkm 614109998Smarkm NOTE: shared library support is still an experimental thing, and 615109998Smarkm backward binary compatibility is still not guaranteed. 616109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 617109998Smarkm 618109998Smarkm *) Add support for Subject Information Access extension. 619109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 620109998Smarkm 621109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 622109998Smarkm additional bytes when new memory had to be allocated, not just 623109998Smarkm when reusing an existing buffer. 624109998Smarkm [Bodo Moeller] 625109998Smarkm 626109998Smarkm *) New command line and configuration option 'utf8' for the req command. 627109998Smarkm This allows field values to be specified as UTF8 strings. 628109998Smarkm [Steve Henson] 629109998Smarkm 630109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 631109998Smarkm runs for the former and machine-readable output for the latter. 632109998Smarkm [Ben Laurie] 633109998Smarkm 634109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 635109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 636109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 637109998Smarkm has the same effect. 638109998Smarkm [Massimiliano Pala madwolf@openca.org] 639109998Smarkm 640109998Smarkm *) Change all functions with names starting with des_ to be starting 641109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 642109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 643109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 644109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 645109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 646109998Smarkm exception. 647109998Smarkm 648109998Smarkm Since we provide two compatibility mappings, the user needs to 649109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 650109998Smarkm compatibility is desired. The default (i.e., when that macro 651109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 652109998Smarkm 653109998Smarkm There are also macros that enable and disable the support of old 654109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 655109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 656109998Smarkm are defined, the default will apply: to support the old des routines. 657109998Smarkm 658109998Smarkm In either case, one must include openssl/des.h to get the correct 659109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 660109998Smarkm won't work. 661109998Smarkm 662109998Smarkm NOTE: This is a major break of an old API into a new one. Software 663109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 664109998Smarkm time in the future, des_old.h and the libdes compatibility functions 665109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 666109998Smarkm default), and then completely removed. 667109998Smarkm [Richard Levitte] 668109998Smarkm 669109998Smarkm *) Test for certificates which contain unsupported critical extensions. 670109998Smarkm If such a certificate is found during a verify operation it is 671109998Smarkm rejected by default: this behaviour can be overridden by either 672109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 673109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 674109998Smarkm X509_supported_extension() has also been added which returns 1 if a 675109998Smarkm particular extension is supported. 676109998Smarkm [Steve Henson] 677109998Smarkm 678109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 679109998Smarkm to retain compatibility with existing code. 680109998Smarkm [Steve Henson] 681109998Smarkm 682109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 683109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 684109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 685109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 686109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 687109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 688109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 689109998Smarkm requires the destination to be valid. 690109998Smarkm 691109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 692109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 693109998Smarkm [Steve Henson] 694109998Smarkm 695109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 696109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 697109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 698109998Smarkm [Bodo Moeller] 699109998Smarkm 700109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 701109998Smarkm [Massimo Santin via Richard Levitte] 702109998Smarkm 703109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 704109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 705109998Smarkm (initialisation, etc) from functionality dealing with implementations 706109998Smarkm of specific crypto iterfaces. This change also introduces integrated 707109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 708109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 709109998Smarkm implementations of their own. This is detailed in crypto/engine/README 710109998Smarkm as it couldn't be adequately described here. However, there are a few 711109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 712109998Smarkm were changed in the original introduction of ENGINE code have now 713109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 714109998Smarkm deal more passive and at run-time, operations deal directly with 715109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 716109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 717109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 718109998Smarkm they were not being used by the framework as there is no concept of a 719109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 720109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 721109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 722109998Smarkm the new code. 723109998Smarkm [Geoff Thorpe] 724109998Smarkm 725109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 726109998Smarkm [Steve Henson] 727109998Smarkm 728109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 729109998Smarkm and make sure the automatically generated functions ERR_load_* 730109998Smarkm become part of libeay.num as well. 731109998Smarkm [Richard Levitte] 732109998Smarkm 733109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 734109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 735109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 736109998Smarkm false once a handshake has been completed. 737109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 738109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 739109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 740109998Smarkm client has followed the request.) 741109998Smarkm [Bodo Moeller] 742109998Smarkm 743109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 744109998Smarkm By default, clients may request session resumption even during 745109998Smarkm renegotiation (if session ID contexts permit); with this option, 746109998Smarkm session resumption is possible only in the first handshake. 747109998Smarkm 748109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 749109998Smarkm more bits available for options that should not be part of 750109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 751109998Smarkm [Bodo Moeller] 752109998Smarkm 753109998Smarkm *) Add some demos for certificate and certificate request creation. 754109998Smarkm [Steve Henson] 755109998Smarkm 756109998Smarkm *) Make maximum certificate chain size accepted from the peer application 757109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 758109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 759109998Smarkm [Lutz Jaenicke] 760109998Smarkm 761109998Smarkm *) Add support for shared libraries for Unixware-7 762109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 763109998Smarkm [Lutz Jaenicke] 764109998Smarkm 765109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 766109998Smarkm be done prior to destruction. Use this to unload error strings from 767109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 768109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 769109998Smarkm [Geoff Thorpe] 770109998Smarkm 771109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 772109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 773109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 774109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 775109998Smarkm Also, add stub code to each that makes building them as self-contained 776109998Smarkm shared-libraries easier (see README.ENGINE). 777109998Smarkm [Geoff Thorpe] 778109998Smarkm 779109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 780109998Smarkm implementations into applications that are completely implemented in 781109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 782109998Smarkm commands that can be used to configure what shared-library to load and 783109998Smarkm to control aspects of the way it is handled. Also, made an update to 784109998Smarkm the README.ENGINE file that brings its information up-to-date and 785109998Smarkm provides some information and instructions on the "dynamic" ENGINE 786109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 787109998Smarkm [Geoff Thorpe] 788109998Smarkm 789109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 790109998Smarkm "ERR_unload_strings" function. 791109998Smarkm [Geoff Thorpe] 792109998Smarkm 793109998Smarkm *) Add a copy() function to EVP_MD. 794109998Smarkm [Ben Laurie] 795109998Smarkm 796109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 797109998Smarkm md_data void pointer. 798109998Smarkm [Ben Laurie] 799109998Smarkm 800109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 801109998Smarkm that the digest can only process a single chunk of data 802109998Smarkm (typically because it is provided by a piece of 803109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 804109998Smarkm is only going to provide a single chunk of data, and hence the 805109998Smarkm framework needn't accumulate the data for oneshot drivers. 806109998Smarkm [Ben Laurie] 807109998Smarkm 808109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 809109998Smarkm functions. This change also alters the storage and management of global 810109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 811109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 812109998Smarkm index counters. The API functions that use this state have been changed 813109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 814109998Smarkm and counter, and there is now an API function to dynamically create new 815109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 816109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 817109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 818109998Smarkm such data would previously have always leaked in application code and 819109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 820109998Smarkm to it. Application code that doesn't use this new function will still 821109998Smarkm leak as before, but their memory debugging output will announce it now 822109998Smarkm rather than letting it slide. 823109998Smarkm 824109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 825109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 826109998Smarkm has a return value to indicate success or failure. 827109998Smarkm [Geoff Thorpe] 828109998Smarkm 829109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 830109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 831109998Smarkm implementation. This change also adds two functions to "get" and "set" 832109998Smarkm the implementation prior to it being automatically set the first time 833109998Smarkm any other ERR function takes place. Ie. an application can call "get", 834109998Smarkm pass the return value to a module it has just loaded, and that module 835109998Smarkm can call its own "set" function using that value. This means the 836109998Smarkm module's "ERR" operations will use (and modify) the error state in the 837109998Smarkm application and not in its own statically linked copy of OpenSSL code. 838109998Smarkm [Geoff Thorpe] 839109998Smarkm 840109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 841109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 842109998Smarkm the operation, and provides a more encapsulated way for external code 843109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 844109998Smarkm to use these functions rather than manually incrementing the counts. 845109998Smarkm 846109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 847109998Smarkm [Geoff Thorpe] 848109998Smarkm 849109998Smarkm *) Add EVP test program. 850109998Smarkm [Ben Laurie] 851109998Smarkm 852109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 853109998Smarkm [Ben Laurie] 854109998Smarkm 855109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 856109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 857109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 858109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 859109998Smarkm directly. Modify 'ca' application to use new functions. 860109998Smarkm [Steve Henson] 861109998Smarkm 862109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 863109998Smarkm bug workarounds. Rollback attack detection is a security feature. 864109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 865109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 866109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 867109998Smarkm for their choice and can explicitly enable this option. 868109998Smarkm [Bodo Moeller, Lutz Jaenicke] 869109998Smarkm 870109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 871109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 872109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 873109998Smarkm Usage example: 874109998Smarkm 875109998Smarkm EVP_MD_CTX md; 876109998Smarkm 877109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 878109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 879109998Smarkm EVP_DigestUpdate(&md, in, len); 880109998Smarkm EVP_DigestFinal(&md, out, NULL); 881109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 882109998Smarkm 883109998Smarkm [Ben Laurie] 884109998Smarkm 885109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 886109998Smarkm correcting its structure. This means that calls to DES functions 887109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 888109998Smarkm plain des_key_schedule (which was actually always a pointer 889109998Smarkm anyway): E.g., 890109998Smarkm 891109998Smarkm des_key_schedule ks; 892109998Smarkm 893109998Smarkm des_set_key_checked(..., &ks); 894109998Smarkm des_ncbc_encrypt(..., &ks, ...); 895109998Smarkm 896109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 897109998Smarkm [Ben Laurie] 898109998Smarkm 899109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 900109998Smarkm PEM causes large amounts of unused functions to be linked in due to 901109998Smarkm poor organisation. For example pem_all.c contains every PEM function 902109998Smarkm which has a knock on effect of linking in large amounts of (unused) 903109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 904109998Smarkm functions prevents this. 905109998Smarkm [Steve Henson] 906109998Smarkm 907109998Smarkm *) Cleanup of EVP macros. 908109998Smarkm [Ben Laurie] 909109998Smarkm 910109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 911109998Smarkm correct _ecb suffix. 912109998Smarkm [Ben Laurie] 913109998Smarkm 914109998Smarkm *) Add initial OCSP responder support to ocsp application. The 915109998Smarkm revocation information is handled using the text based index 916109998Smarkm use by the ca application. The responder can either handle 917109998Smarkm requests generated internally, supplied in files (for example 918109998Smarkm via a CGI script) or using an internal minimal server. 919109998Smarkm [Steve Henson] 920109998Smarkm 921109998Smarkm *) Add configuration choices to get zlib compression for TLS. 922109998Smarkm [Richard Levitte] 923109998Smarkm 924109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 925109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 926109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 927109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 928109998Smarkm 929109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 930109998Smarkm and authenticator structs; see crypto/krb5/. 931109998Smarkm 932109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 933109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 934109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 935109998Smarkm via Richard Levitte] 936109998Smarkm 937109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 938109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 939109998Smarkm values for each of the key sizes rather than having just 940109998Smarkm parameters (and 'speed' generating keys each time). 941109998Smarkm [Geoff Thorpe] 942109998Smarkm 943109998Smarkm *) Speed up EVP routines. 944109998Smarkm Before: 945109998Smarkmencrypt 946109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 947109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 948109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 949109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 950109998Smarkmdecrypt 951109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 952109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 953109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 954109998Smarkm After: 955109998Smarkmencrypt 956109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 957109998Smarkmdecrypt 958109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 959109998Smarkm [Ben Laurie] 960109998Smarkm 961109998Smarkm *) Added the OS2-EMX target. 962109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 963109998Smarkm 964109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 965109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 966109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 967109998Smarkm structure: this means that the old CONF compatible routines can be 968109998Smarkm retained (in particular wrt extensions) without having to duplicate the 969109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 970109998Smarkm [Steve Henson] 971109998Smarkm 972109998Smarkm *) Enhance the general user interface with mechanisms for inner control 973109998Smarkm and with possibilities to have yes/no kind of prompts. 974109998Smarkm [Richard Levitte] 975109998Smarkm 976109998Smarkm *) Change all calls to low level digest routines in the library and 977109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 978109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 979109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 980109998Smarkm 981109998Smarkm *) Add the possibility to control engines through control names but with 982109998Smarkm arbitrary arguments instead of just a string. 983109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 984109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 985109998Smarkm versions of OpenSSL [engine]. 986109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 987109998Smarkm callback. 988109998Smarkm [Richard Levitte] 989109998Smarkm 990109998Smarkm *) Enhance the general user interface with mechanisms to better support 991109998Smarkm dialog box interfaces, application-defined prompts, the possibility 992109998Smarkm to use defaults (for example default passwords from somewhere else) 993109998Smarkm and interrupts/cancellations. 994109998Smarkm [Richard Levitte] 995109998Smarkm 996109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 997109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 998109998Smarkm [Steve Henson] 999109998Smarkm 1000109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 1001109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 1002109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 1003109998Smarkm 1004109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 1005109998Smarkm callback (pem_password_cb) as all other routines that need this 1006109998Smarkm kind of callback. 1007109998Smarkm [Richard Levitte] 1008109998Smarkm 1009109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 1010109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 1011109998Smarkm than this minimum value is recommended. 1012109998Smarkm [Lutz Jaenicke] 1013109998Smarkm 1014109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 1015109998Smarkm that are easily reachable. 1016109998Smarkm [Richard Levitte] 1017109998Smarkm 1018109998Smarkm *) Windows apparently can't transparently handle global 1019109998Smarkm variables defined in DLLs. Initialisations such as: 1020109998Smarkm 1021109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 1022109998Smarkm 1023109998Smarkm wont compile. This is used by the any applications that need to 1024109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 1025109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 1026109998Smarkm needed for static libraries under Win32. 1027109998Smarkm [Steve Henson] 1028109998Smarkm 1029109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 1030109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 1031109998Smarkm purpose functions and tidy up setting in other SSL functions. 1032109998Smarkm [Steve Henson] 1033109998Smarkm 1034109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 1035109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 1036109998Smarkm initialised. This allows various defaults to be set in the 1037109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 1038109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 1039109998Smarkm internally such as S/MIME. 1040109998Smarkm 1041109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 1042109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 1043109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 1044109998Smarkm 1045109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 1046109998Smarkm applications. 1047109998Smarkm [Steve Henson] 1048109998Smarkm 1049109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 1050109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 1051109998Smarkm its validity and signature checked, then if the certificate is found 1052109998Smarkm in the CRL the verify fails with a revoked error. 1053109998Smarkm 1054109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 1055109998Smarkm 1056109998Smarkm Command line options added to 'verify' application to support this. 1057109998Smarkm 1058109998Smarkm This needs some additional work, such as being able to handle multiple 1059109998Smarkm CRLs with different times, extension based lookup (rather than just 1060109998Smarkm by subject name) and ultimately more complete V2 CRL extension 1061109998Smarkm handling. 1062109998Smarkm [Steve Henson] 1063109998Smarkm 1064109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 1065109998Smarkm to replace things like des_read_password and friends (backward 1066109998Smarkm compatibility functions using this new API are provided). 1067109998Smarkm The purpose is to remove prompting functions from the DES code 1068109998Smarkm section as well as provide for prompting through dialog boxes in 1069109998Smarkm a window system and the like. 1070109998Smarkm [Richard Levitte] 1071109998Smarkm 1072109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 1073109998Smarkm per-structure level rather than having to store it globally. 1074109998Smarkm [Geoff] 1075109998Smarkm 1076109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 1077109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 1078109998Smarkm This causes the "original" ENGINE structure to act like a template, 1079109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 1080109998Smarkm operational state can be localised to each ENGINE structure, despite the 1081109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 1082109998Smarkm this case have no functional references and the return value is the single 1083109998Smarkm structural reference. This matches the single structural reference returned 1084109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 1085109998Smarkm ENGINE structure. 1086109998Smarkm [Geoff] 1087109998Smarkm 1088109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 1089109998Smarkm needs to match any other type at all we need to manually clear the 1090109998Smarkm tag cache. 1091109998Smarkm [Steve Henson] 1092109998Smarkm 1093109998Smarkm *) Changes to the "openssl engine" utility to include; 1094109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 1095109998Smarkm about an ENGINE's available control commands. 1096109998Smarkm - executing control commands from command line arguments using the 1097109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 1098109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 1099109998Smarkm the individual commands are colon-separated, for example; 1100109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 1101109998Smarkm [Geoff] 1102109998Smarkm 1103109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 1104109998Smarkm declare their own commands (numbers), names (strings), descriptions, 1105109998Smarkm and input types for run-time discovery by calling applications. A 1106109998Smarkm subset of these commands are implicitly classed as "executable" 1107109998Smarkm depending on their input type, and only these can be invoked through 1108109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 1109109998Smarkm can be based on user input, config files, etc). The distinction is 1110109998Smarkm that "executable" commands cannot return anything other than a boolean 1111109998Smarkm result and can only support numeric or string input, whereas some 1112109998Smarkm discoverable commands may only be for direct use through 1113109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 1114109998Smarkm pointers, or other custom uses. The "executable" commands are to 1115109998Smarkm support parameterisations of ENGINE behaviour that can be 1116109998Smarkm unambiguously defined by ENGINEs and used consistently across any 1117109998Smarkm OpenSSL-based application. Commands have been added to all the 1118109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 1119109998Smarkm control over shared-library paths without source code alterations. 1120109998Smarkm [Geoff] 1121109998Smarkm 1122109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 1123109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 1124109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 1125109998Smarkm this also allows the implementations to compile without using the 1126109998Smarkm internal engine_int.h header. 1127109998Smarkm [Geoff] 1128109998Smarkm 1129109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 1130109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 1131109998Smarkm should already have non-const pointers to it (ie. they should only 1132109998Smarkm modify their own ones). 1133109998Smarkm [Geoff] 1134109998Smarkm 1135109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 1136109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 1137109998Smarkm to C code. "nuron" string definitions were placed in variables 1138109998Smarkm rather than hard-coded - allowing parameterisation of these values 1139109998Smarkm later on via ctrl() commands. 1140109998Smarkm - Removed unused "#if 0"'d code. 1141109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 1142109998Smarkm structural references. 1143109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 1144109998Smarkm - Constified various get/set functions as appropriate and added 1145109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 1146109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 1147109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 1148109998Smarkm or function to NULL is a way of cancelling out a previously set 1149109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 1150109998Smarkm and doesn't justify the extra error symbols and code. 1151109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 1152109998Smarkm flags from engine_int.h to engine.h. 1153109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 1154109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 1155109998Smarkm [Geoff] 1156109998Smarkm 1157109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 1158109998Smarkm to the algorithm using long division. The binary algorithm can be 1159109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 1160109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 1161109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 1162109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 1163109998Smarkm appears to be advantageous for much longer moduli; here we use it 1164109998Smarkm for moduli up to 2048 bits. 1165109998Smarkm [Bodo Moeller] 1166109998Smarkm 1167109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 1168109998Smarkm could not support the combine flag in choice fields. 1169109998Smarkm [Steve Henson] 1170109998Smarkm 1171109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 1172109998Smarkm extensions from a certificate request to the certificate. 1173109998Smarkm [Steve Henson] 1174109998Smarkm 1175109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 1176109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 1177109998Smarkm file: this allows the display of the certificate about to be 1178109998Smarkm signed to be customised, to allow certain fields to be included 1179109998Smarkm or excluded and extension details. The old system didn't display 1180109998Smarkm multicharacter strings properly, omitted fields not in the policy 1181109998Smarkm and couldn't display additional details such as extensions. 1182109998Smarkm [Steve Henson] 1183109998Smarkm 1184109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 1185109998Smarkm of an arbitrary number of elliptic curve points 1186109998Smarkm \sum scalars[i]*points[i], 1187109998Smarkm optionally including the generator defined for the EC_GROUP: 1188109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 1189109998Smarkm 1190109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 1191109998Smarkm that the point list has just one item (besides the optional 1192109998Smarkm generator). 1193109998Smarkm [Bodo Moeller] 1194109998Smarkm 1195109998Smarkm *) First EC_METHODs for curves over GF(p): 1196109998Smarkm 1197109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 1198109998Smarkm operations and provides various method functions that can also 1199109998Smarkm operate with faster implementations of modular arithmetic. 1200109998Smarkm 1201109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 1202109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 1203109998Smarkm 1204109998Smarkm [Bodo Moeller; point addition and point doubling 1205109998Smarkm implementation directly derived from source code provided by 1206109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 1207109998Smarkm 1208109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 1209109998Smarkm crypto/ec/ec_lib.c): 1210109998Smarkm 1211109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 1212109998Smarkm based on EC_METHODs that are built into the library. 1213109998Smarkm 1214109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 1215109998Smarkm 1216109998Smarkm Most of the framework would be able to handle curves over arbitrary 1217109998Smarkm finite fields, but as there are no obvious types for fields other 1218109998Smarkm than GF(p), some functions are limited to that for now. 1219109998Smarkm [Bodo Moeller] 1220109998Smarkm 1221109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 1222109998Smarkm that the file contains a complete HTTP response. 1223109998Smarkm [Richard Levitte] 1224109998Smarkm 1225109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 1226109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 1227109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 1228109998Smarkm field while the former will cause them to run together if the field 1229109998Smarkm is 40 of more characters long. 1230109998Smarkm [Steve Henson] 1231109998Smarkm 1232109998Smarkm *) Constify the cipher and digest 'method' functions and structures 1233109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 1234109998Smarkm pointers. 1235109998Smarkm [Steve Henson] 1236109998Smarkm 1237109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 1238109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 1239109998Smarkm [Bodo Moeller] 1240109998Smarkm 1241109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 1242109998Smarkm internal software routines can never fail additional hardware versions 1243109998Smarkm might. 1244109998Smarkm [Steve Henson] 1245109998Smarkm 1246109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 1247109998Smarkm 1248109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 1249109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 1250109998Smarkm 1251109998Smarkm ASN1 error codes 1252109998Smarkm ERR_R_NESTED_ASN1_ERROR 1253109998Smarkm ... 1254109998Smarkm ERR_R_MISSING_ASN1_EOS 1255109998Smarkm were 4 .. 9, conflicting with 1256109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 1257109998Smarkm ... 1258109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 1259109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 1260109998Smarkm 1261109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 1262109998Smarkm [Bodo Moeller] 1263109998Smarkm 1264109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 1265109998Smarkm suffices. 1266109998Smarkm [Bodo Moeller] 1267109998Smarkm 1268109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 1269109998Smarkm sets the subject name for a new request or supersedes the 1270109998Smarkm subject name in a given request. Formats that can be parsed are 1271109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 1272109998Smarkm and 1273109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 1274109998Smarkm 1275109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 1276109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 1277109998Smarkm 1278109998Smarkm *) Introduce the possibility to access global variables through 1279109998Smarkm functions on platform were that's the best way to handle exporting 1280109998Smarkm global variables in shared libraries. To enable this functionality, 1281109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 1282109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 1283109998Smarkm is normally done by Configure or something similar). 1284109998Smarkm 1285109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 1286109998Smarkm in the source file (foo.c) like this: 1287109998Smarkm 1288109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 1289109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 1290109998Smarkm 1291109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 1292109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 1293109998Smarkm 1294109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 1295109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 1296109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 1297109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 1298109998Smarkm 1299109998Smarkm The #defines are very important, and therefore so is including the 1300109998Smarkm header file everywhere where the defined globals are used. 1301109998Smarkm 1302109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 1303109998Smarkm of ASN.1 items, but that structure is a bit different. 1304109998Smarkm 1305109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 1306109998Smarkm better and easier to understand logic to choose which symbols should 1307109998Smarkm go into the Windows .def files as well as a number of fixes and code 1308109998Smarkm cleanup (among others, algorithm keywords are now sorted 1309109998Smarkm lexicographically to avoid constant rewrites). 1310109998Smarkm [Richard Levitte] 1311109998Smarkm 1312109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 1313109998Smarkm result to 'rm' because if rm==num the value will be overwritten 1314109998Smarkm and produce the wrong result if 'num' is negative: this caused 1315109998Smarkm problems with BN_mod() and BN_nnmod(). 1316109998Smarkm [Steve Henson] 1317109998Smarkm 1318109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 1319109998Smarkm OCSP request and verifies the signer certificate. The signer 1320109998Smarkm certificate is just checked for a generic purpose and OCSP request 1321109998Smarkm trust settings. 1322109998Smarkm [Steve Henson] 1323109998Smarkm 1324109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 1325109998Smarkm responses. OCSP responses are prepared in real time and may only 1326109998Smarkm be a few seconds old. Simply checking that the current time lies 1327109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 1328109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 1329109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 1330109998Smarkm the current time. The age of the response can also optionally be 1331109998Smarkm checked. Two new options -validity_period and -status_age added to 1332109998Smarkm ocsp utility. 1333109998Smarkm [Steve Henson] 1334109998Smarkm 1335109998Smarkm *) If signature or public key algorithm is unrecognized print out its 1336109998Smarkm OID rather that just UNKNOWN. 1337109998Smarkm [Steve Henson] 1338109998Smarkm 1339109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 1340109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 1341109998Smarkm ID to be generated from the issuer certificate alone which can then be 1342109998Smarkm passed to OCSP_id_issuer_cmp(). 1343109998Smarkm [Steve Henson] 1344109998Smarkm 1345109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 1346109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 1347109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 1348109998Smarkm new macros which allow the underlying ASN1 function/structure to 1349109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 1350109998Smarkm references directly (such as &X509_it) but instead use the relevant 1351109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 1352109998Smarkm use of the new ASN1 code on platforms where exporting structures 1353109998Smarkm is problematical (for example in shared libraries) but exporting 1354109998Smarkm functions returning pointers to structures is not. 1355109998Smarkm [Steve Henson] 1356109998Smarkm 1357109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 1358109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 1359109998Smarkm The purpose of this is to allow applications to control, if they wish, 1360109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 1361109998Smarkm can be useful for session caching in multiple-server environments. A 1362109998Smarkm command-line switch for testing this (and any client code that wishes 1363109998Smarkm to use such a feature) has been added to "s_server". 1364109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 1365109998Smarkm 1366109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 1367109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 1368109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 1369109998Smarkm the growing number of special cases it was previously handling. 1370109998Smarkm [Richard Levitte] 1371109998Smarkm 1372109998Smarkm *) Make all configuration macros available for application by making 1373109998Smarkm sure they are available in opensslconf.h, by giving them names starting 1374109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 1375109998Smarkm sure e_os2.h will cover all platform-specific cases together with 1376109998Smarkm opensslconf.h. 1377109998Smarkm Additionally, it is now possible to define configuration/platform- 1378109998Smarkm specific names (called "system identities"). In the C code, these 1379109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 1380109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 1381109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 1382109998Smarkm what is available. 1383109998Smarkm [Richard Levitte] 1384109998Smarkm 1385109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 1386109998Smarkm number to use to be specified on the command line. Previously self 1387109998Smarkm signed certificates were hard coded with serial number 0 and the 1388109998Smarkm CA options of 'x509' had to use a serial number in a file which was 1389109998Smarkm auto incremented. 1390109998Smarkm [Steve Henson] 1391109998Smarkm 1392109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 1393109998Smarkm Currently CRL reason, invalidity date and hold instruction are 1394109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 1395109998Smarkm [Steve Henson] 1396109998Smarkm 1397109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 1398109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 1399109998Smarkm API, which was previously mandatory. This means that the data is 1400109998Smarkm not padded in any way and so the total length much be a multiple 1401109998Smarkm of the block size, otherwise an error occurs. 1402109998Smarkm [Steve Henson] 1403109998Smarkm 1404109998Smarkm *) Initial (incomplete) OCSP SSL support. 1405109998Smarkm [Steve Henson] 1406109998Smarkm 1407109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 1408109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 1409109998Smarkm option to ocsp utility. 1410109998Smarkm [Steve Henson] 1411109998Smarkm 1412109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 1413109998Smarkm reflects the various checks performed. Applications can decide 1414109998Smarkm whether to tolerate certain situations such as an absent nonce 1415109998Smarkm in a response when one was present in a request: the ocsp application 1416109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 1417109998Smarkm this is to allow responders to include a nonce in a response even if 1418109998Smarkm the request is nonce-less. 1419109998Smarkm [Steve Henson] 1420109998Smarkm 1421109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 1422109998Smarkm skipped when using openssl x509 multiple times on a single input file, 1423109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 1424109998Smarkm [Bodo Moeller] 1425109998Smarkm 1426109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 1427109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 1428109998Smarkm utility to correctly initialize revocation date of CRLs. 1429109998Smarkm [Steve Henson] 1430109998Smarkm 1431109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 1432109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 1433109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 1434109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 1435109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 1436109998Smarkm [Lutz Jaenicke] 1437109998Smarkm 1438109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 1439109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 1440109998Smarkm appear to exist. 1441109998Smarkm [Steve Henson] 1442109998Smarkm 1443109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 1444109998Smarkm additional certificates supplied. 1445109998Smarkm [Steve Henson] 1446109998Smarkm 1447109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 1448109998Smarkm OCSP client a number of certificate to only verify the response 1449109998Smarkm signature against. 1450109998Smarkm [Richard Levitte] 1451109998Smarkm 1452109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 1453109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 1454109998Smarkm AES OIDs. 1455109998Smarkm 1456109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 1457109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 1458109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 1459109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 1460109998Smarkm alias because they were not yet official; they could be 1461109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 1462109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 1463109998Smarkm alias is called "AES" and is part of "ALL".) 1464109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 1465109998Smarkm 1466109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 1467109998Smarkm request to response. 1468109998Smarkm [Steve Henson] 1469109998Smarkm 1470109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 1471109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 1472109998Smarkm extract information from a certificate request. OCSP_response_create() 1473109998Smarkm creates a response and optionally adds a basic response structure. 1474109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 1475109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 1476109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 1477109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 1478109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 1479109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 1480109998Smarkm (converts ASN1_TIME to GeneralizedTime). 1481109998Smarkm [Steve Henson] 1482109998Smarkm 1483109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 1484109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 1485109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 1486109998Smarkm contents: this is used in various key identifiers. 1487109998Smarkm [Steve Henson] 1488109998Smarkm 1489109998Smarkm *) Make sk_sort() tolerate a NULL argument. 1490109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 1491109998Smarkm 1492109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 1493109998Smarkm passed by the function are trusted implicitly. If any of them signed the 1494109998Smarkm response then it is assumed to be valid and is not verified. 1495109998Smarkm [Steve Henson] 1496109998Smarkm 1497109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 1498109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 1499109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 1500109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 1501109998Smarkm <support@securenetterm.com>] 1502109998Smarkm 1503109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 1504109998Smarkm routines: without these tracing memory leaks is very painful. 1505109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 1506109998Smarkm [Steve Henson] 1507109998Smarkm 1508109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 1509109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 1510109998Smarkm effectively meant GeneralizedTime would never be used. Now it 1511109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 1512109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 1513109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 1514109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 1515109998Smarkm <support@securenetterm.com>] 1516109998Smarkm 1517109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 1518109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 1519109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 1520109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 1521109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 1522109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 1523109998Smarkm [Steve Henson] 1524109998Smarkm 1525109998Smarkm *) Add summary printout to ocsp utility. The various functions which 1526109998Smarkm convert status values to strings have been renamed to: 1527109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 1528109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 1529109998Smarkm to verify nonce values and to disable verification. OCSP response 1530109998Smarkm printout format cleaned up. 1531109998Smarkm [Steve Henson] 1532109998Smarkm 1533109998Smarkm *) Add additional OCSP certificate checks. These are those specified 1534109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 1535109998Smarkm certificate being checked must either be the OCSP signer certificate 1536109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 1537109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 1538109998Smarkm usage. This check is performed by attempting to match the OCSP 1539109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 1540109998Smarkm in the OCSP_CERTID structures of the response. 1541109998Smarkm [Steve Henson] 1542109998Smarkm 1543109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 1544109998Smarkm and related routines. This uses the standard OpenSSL certificate 1545109998Smarkm verify routines to perform initial checks (just CA validity) and 1546109998Smarkm to obtain the certificate chain. Then additional checks will be 1547109998Smarkm performed on the chain. Currently the root CA is checked to see 1548109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 1549109998Smarkm a root CA as a global signing root: that is any certificate that 1550109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 1551109998Smarkm [Steve Henson] 1552109998Smarkm 1553109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 1554109998Smarkm extensions from a separate configuration file. 1555109998Smarkm As when reading extensions from the main configuration file, 1556109998Smarkm the '-extensions ...' option may be used for specifying the 1557109998Smarkm section to use. 1558109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 1559109998Smarkm 1560109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 1561109998Smarkm read. The request can be sent to a responder and the output 1562109998Smarkm parsed, outputed or printed in text form. Not complete yet: 1563109998Smarkm still needs to check the OCSP response validity. 1564109998Smarkm [Steve Henson] 1565109998Smarkm 1566109998Smarkm *) New subcommands for 'openssl ca': 1567109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 1568109998Smarkm the given serial number (according to the index file). 1569109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 1570109998Smarkm in the index file. 1571109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 1572109998Smarkm 1573109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 1574109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 1575109998Smarkm so that the resulting key is not encrypted. 1576109998Smarkm [Damien Miller <djm@mindrot.org>] 1577109998Smarkm 1578109998Smarkm *) New configuration for the GNU Hurd. 1579109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 1580109998Smarkm 1581109998Smarkm *) Initial code to implement OCSP basic response verify. This 1582109998Smarkm is currently incomplete. Currently just finds the signer's 1583109998Smarkm certificate and verifies the signature on the response. 1584109998Smarkm [Steve Henson] 1585109998Smarkm 1586109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 1587109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 1588109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 1589109998Smarkm [Bodo Moeller] 1590109998Smarkm 1591109998Smarkm *) Allowing defining memory allocation callbacks that will be given 1592109998Smarkm file name and line number information in additional arguments 1593109998Smarkm (a const char* and an int). The basic functionality remains, as 1594109998Smarkm well as the original possibility to just replace malloc(), 1595109998Smarkm realloc() and free() by functions that do not know about these 1596109998Smarkm additional arguments. To register and find out the current 1597109998Smarkm settings for extended allocation functions, the following 1598109998Smarkm functions are provided: 1599109998Smarkm 1600109998Smarkm CRYPTO_set_mem_ex_functions 1601109998Smarkm CRYPTO_set_locked_mem_ex_functions 1602109998Smarkm CRYPTO_get_mem_ex_functions 1603109998Smarkm CRYPTO_get_locked_mem_ex_functions 1604109998Smarkm 1605109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 1606109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 1607109998Smarkm extended allocation function is enabled. 1608109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 1609109998Smarkm a conventional allocation function is enabled. 1610109998Smarkm [Richard Levitte, Bodo Moeller] 1611109998Smarkm 1612109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 1613109998Smarkm There should no longer be any prototype-casting required when using 1614109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 1615109998Smarkm the callback types and macros at the head of lhash.h for details 1616109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 1617109998Smarkm [Geoff Thorpe] 1618109998Smarkm 1619109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 1620109998Smarkm If /dev/[u]random devices are not available or do not return enough 1621109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 1622109998Smarkm be queried. 1623109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 1624109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 1625109998Smarkm when enough entropy was collected without querying more sockets. 1626109998Smarkm [Lutz Jaenicke] 1627109998Smarkm 1628109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 1629109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 1630109998Smarkm of data has been collected. We spend at most 10 ms on each file 1631109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 1632109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 1633109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 1634109998Smarkm platforms the 10 ms delay will never occur. 1635109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 1636109998Smarkm For VMS, there's a currently-empty rand_vms.c. 1637109998Smarkm [Richard Levitte] 1638109998Smarkm 1639109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 1640109998Smarkm provide utility functions which an application needing 1641109998Smarkm to issue a request to an OCSP responder and analyse the 1642109998Smarkm response will typically need: as opposed to those which an 1643109998Smarkm OCSP responder itself would need which will be added later. 1644109998Smarkm 1645109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 1646109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 1647109998Smarkm response. OCSP_response_get1_basic() extracts basic response 1648109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 1649109998Smarkm information from an OCSP_CERTID structure (which will be created 1650109998Smarkm when the request structure is built). These are built from lower 1651109998Smarkm level functions which work on OCSP_SINGLERESP structures but 1652109998Smarkm wont normally be used unless the application wishes to examine 1653109998Smarkm extensions in the OCSP response for example. 1654109998Smarkm 1655109998Smarkm Replace nonce routines with a pair of functions. 1656109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 1657109998Smarkm generates a random value. OCSP_check_nonce() checks the 1658109998Smarkm validity of the nonce in an OCSP response. 1659109998Smarkm [Steve Henson] 1660109998Smarkm 1661109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 1662109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 1663109998Smarkm need to free up the newly created id. Change return type 1664109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 1665109998Smarkm This can then be used to add extensions to the request. 1666109998Smarkm Deleted OCSP_request_new(), since most of its functionality 1667109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 1668109998Smarkm clash) apart from the ability to set the request name which 1669109998Smarkm will be added elsewhere. 1670109998Smarkm [Steve Henson] 1671109998Smarkm 1672109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 1673109998Smarkm various functions. Extensions are now handled using the new 1674109998Smarkm OCSP extension code. New simple OCSP HTTP function which 1675109998Smarkm can be used to send requests and parse the response. 1676109998Smarkm [Steve Henson] 1677109998Smarkm 1678109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 1679109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 1680109998Smarkm uses the special reorder version of SET OF to sort the attributes 1681109998Smarkm and reorder them to match the encoded order. This resolves a long 1682109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 1683109998Smarkm it used to fail because the attribute order did not match the 1684109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 1685109998Smarkm it uses the received order. This is necessary to tolerate some broken 1686109998Smarkm software that does not order SET OF. This is handled by encoding 1687109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 1688109998Smarkm to produce the required SET OF. 1689109998Smarkm [Steve Henson] 1690109998Smarkm 1691109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 1692109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 1693109998Smarkm files to get correct declarations of the ASN.1 item variables. 1694109998Smarkm [Richard Levitte] 1695109998Smarkm 1696109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 1697109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 1698109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 1699109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 1700109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 1701109998Smarkm ASN1_ITEM and no wrapper functions. 1702109998Smarkm [Steve Henson] 1703109998Smarkm 1704109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 1705109998Smarkm replace the old function pointer based I/O routines. Change most of 1706109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 1707109998Smarkm [Steve Henson] 1708109998Smarkm 1709109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 1710109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 1711109998Smarkm it complain about algorithm deselection that isn't recognised. 1712109998Smarkm [Richard Levitte] 1713109998Smarkm 1714109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 1715109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 1716109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 1717109998Smarkm some old style ASN1 functions: this can be used to determine if old 1718109998Smarkm code will still work when these eventually go away. 1719109998Smarkm [Steve Henson] 1720109998Smarkm 1721109998Smarkm *) New extension functions for OCSP structures, these follow the 1722109998Smarkm same conventions as certificates and CRLs. 1723109998Smarkm [Steve Henson] 1724109998Smarkm 1725109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 1726109998Smarkm adds an extension. Its behaviour can be customised with various 1727109998Smarkm flags to append, replace or delete. Various wrappers added for 1728109998Smarkm certifcates and CRLs. 1729109998Smarkm [Steve Henson] 1730109998Smarkm 1731109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 1732109998Smarkm an extension cannot be parsed. Correct a typo in the 1733109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 1734109998Smarkm [Steve Henson] 1735109998Smarkm 1736109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 1737109998Smarkm entries for variables. 1738109998Smarkm [Steve Henson] 1739109998Smarkm 1740109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 1741109998Smarkm problems: As the program is single-threaded, all we have 1742109998Smarkm to do is register a locking callback using an array for 1743109998Smarkm storing which locks are currently held by the program. 1744109998Smarkm [Bodo Moeller] 1745109998Smarkm 1746109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 1747109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 1748109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 1749109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 1750109998Smarkm Unfortunately, the ex_data design is not at all suited 1751109998Smarkm for multi-threaded use, so it probably should be abolished. 1752109998Smarkm [Bodo Moeller] 1753109998Smarkm 1754109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 1755109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 1756109998Smarkm 1757109998Smarkm *) Move common extension printing code to new function 1758109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 1759109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 1760109998Smarkm [Steve Henson] 1761109998Smarkm 1762109998Smarkm *) New function X509_signature_print() to remove duplication in some 1763109998Smarkm print routines. 1764109998Smarkm [Steve Henson] 1765109998Smarkm 1766109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 1767109998Smarkm set (this was treated exactly the same as SET OF previously). This 1768109998Smarkm is used to reorder the STACK representing the structure to match the 1769109998Smarkm encoding. This will be used to get round a problem where a PKCS7 1770109998Smarkm structure which was signed could not be verified because the STACK 1771109998Smarkm order did not reflect the encoded order. 1772109998Smarkm [Steve Henson] 1773109998Smarkm 1774109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 1775109998Smarkm [Steve Henson] 1776109998Smarkm 1777109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 1778109998Smarkm for its ASN1 operations. The old style function pointers still exist 1779109998Smarkm for now but they will eventually go away. 1780109998Smarkm [Steve Henson] 1781109998Smarkm 1782109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 1783109998Smarkm completely replaces the old ASN1 functionality with a table driven 1784109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 1785109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 1786109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 1787109998Smarkm has also been converted to the new form. 1788109998Smarkm [Steve Henson] 1789109998Smarkm 1790109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 1791109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 1792109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 1793109998Smarkm for negative moduli. 1794109998Smarkm [Bodo Moeller] 1795109998Smarkm 1796109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 1797109998Smarkm of not touching the result's sign bit. 1798109998Smarkm [Bodo Moeller] 1799109998Smarkm 1800109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 1801109998Smarkm set. 1802109998Smarkm [Bodo Moeller] 1803109998Smarkm 1804109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 1805109998Smarkm macros to declare and implement thin (optionally static) functions 1806109998Smarkm that provide type-safety and avoid function pointer casting for the 1807109998Smarkm type-specific callbacks. 1808109998Smarkm [Geoff Thorpe] 1809109998Smarkm 1810109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 1811109998Smarkm RFC 2712. 1812109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 1813109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 1814109998Smarkm 1815109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 1816109998Smarkm in sections depending on the subject. 1817109998Smarkm [Richard Levitte] 1818109998Smarkm 1819109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 1820109998Smarkm Windows. 1821109998Smarkm [Richard Levitte] 1822109998Smarkm 1823109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 1824109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 1825109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 1826109998Smarkm be handled deterministically). 1827109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1828109998Smarkm 1829109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 1830109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 1831109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 1832109998Smarkm [Bodo Moeller] 1833109998Smarkm 1834109998Smarkm *) New function BN_kronecker. 1835109998Smarkm [Bodo Moeller] 1836109998Smarkm 1837109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 1838109998Smarkm positive unless both parameters are zero. 1839109998Smarkm Previously something reasonably close to an infinite loop was 1840109998Smarkm possible because numbers could be growing instead of shrinking 1841109998Smarkm in the implementation of Euclid's algorithm. 1842109998Smarkm [Bodo Moeller] 1843109998Smarkm 1844109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 1845109998Smarkm sign of the number in question. 1846109998Smarkm 1847109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 1848109998Smarkm 1849109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 1850109998Smarkm because its test if the absolute value of 'a' equals 'w'. 1851109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 1852109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 1853109998Smarkm BN_is_one(), and BN_is_word(). 1854109998Smarkm [Bodo Moeller] 1855109998Smarkm 1856109998Smarkm *) New function BN_swap. 1857109998Smarkm [Bodo Moeller] 1858109998Smarkm 1859109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 1860109998Smarkm the exponentiation functions are more likely to produce reasonable 1861109998Smarkm results on negative inputs. 1862109998Smarkm [Bodo Moeller] 1863109998Smarkm 1864109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 1865109998Smarkm Previously, it could be negative if one of the factors was negative; 1866109998Smarkm I don't think anyone really wanted that behaviour. 1867109998Smarkm [Bodo Moeller] 1868109998Smarkm 1869109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 1870109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 1871109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 1872109998Smarkm and add new functions: 1873109998Smarkm 1874109998Smarkm BN_nnmod 1875109998Smarkm BN_mod_sqr 1876109998Smarkm BN_mod_add 1877109998Smarkm BN_mod_add_quick 1878109998Smarkm BN_mod_sub 1879109998Smarkm BN_mod_sub_quick 1880109998Smarkm BN_mod_lshift1 1881109998Smarkm BN_mod_lshift1_quick 1882109998Smarkm BN_mod_lshift 1883109998Smarkm BN_mod_lshift_quick 1884109998Smarkm 1885109998Smarkm These functions always generate non-negative results. 1886109998Smarkm 1887109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 1888109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 1889109998Smarkm 1890109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 1891109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 1892109998Smarkm be reduced modulo m. 1893109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1894109998Smarkm 1895111147Snectar#if 0 1896111147Snectar The following entry accidentily appeared in the CHANGES file 1897111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 1898111147Snectar it do *not* apply to OpenSSL 0.9.7. 1899111147Snectar 1900109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1901109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 1902109998Smarkm required a small change in bn_mul_part_recursive() and the addition 1903109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 1904109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 1905109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 1906109998Smarkm differing sizes. 1907109998Smarkm [Richard Levitte] 1908111147Snectar#endif 1909109998Smarkm 1910109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 1911109998Smarkm unless the '-salt' option is used (which usually means that 1912109998Smarkm verification would just waste user's time since the resulting 1913109998Smarkm hash is going to be compared with some given password hash) 1914109998Smarkm or the new '-noverify' option is used. 1915109998Smarkm 1916109998Smarkm This is an incompatible change, but it does not affect 1917109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 1918109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 1919109998Smarkm cause any problems. 1920109998Smarkm [Bodo Moeller] 1921109998Smarkm 1922109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 1923109998Smarkm [Richard Levitte] 1924109998Smarkm 1925109998Smarkm *) Make DSO load along a path given through an environment variable 1926109998Smarkm (SHLIB_PATH) with shl_load(). 1927109998Smarkm [Richard Levitte] 1928109998Smarkm 1929109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 1930109998Smarkm Also constify the RSA code and most things related to it. In a 1931109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 1932109998Smarkm casts back to non-const were required (to be solved at a later 1933109998Smarkm time) 1934109998Smarkm [Richard Levitte] 1935109998Smarkm 1936109998Smarkm *) Make it so the openssl application has all engines loaded by default. 1937109998Smarkm [Richard Levitte] 1938109998Smarkm 1939109998Smarkm *) Constify the BIGNUM routines a little more. 1940109998Smarkm [Richard Levitte] 1941109998Smarkm 1942109998Smarkm *) Add the following functions: 1943109998Smarkm 1944109998Smarkm ENGINE_load_cswift() 1945109998Smarkm ENGINE_load_chil() 1946109998Smarkm ENGINE_load_atalla() 1947109998Smarkm ENGINE_load_nuron() 1948109998Smarkm ENGINE_load_builtin_engines() 1949109998Smarkm 1950109998Smarkm That way, an application can itself choose if external engines that 1951109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 1952109998Smarkm that applications won't have to be linked with libdl or other dso 1953109998Smarkm libraries unless it's really needed. 1954109998Smarkm 1955109998Smarkm Changed 'openssl engine' to load all engines on demand. 1956109998Smarkm Changed the engine header files to avoid the duplication of some 1957109998Smarkm declarations (they differed!). 1958109998Smarkm [Richard Levitte] 1959109998Smarkm 1960109998Smarkm *) 'openssl engine' can now list capabilities. 1961109998Smarkm [Richard Levitte] 1962109998Smarkm 1963109998Smarkm *) Better error reporting in 'openssl engine'. 1964109998Smarkm [Richard Levitte] 1965109998Smarkm 1966109998Smarkm *) Never call load_dh_param(NULL) in s_server. 1967109998Smarkm [Bodo Moeller] 1968109998Smarkm 1969109998Smarkm *) Add engine application. It can currently list engines by name and 1970109998Smarkm identity, and test if they are actually available. 1971109998Smarkm [Richard Levitte] 1972109998Smarkm 1973109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 1974109998Smarkm sure the installed documentation is also owned by root.root. 1975109998Smarkm [Damien Miller <djm@mindrot.org>] 1976109998Smarkm 1977109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 1978109998Smarkm keys (public as well as private) handled by engines. 1979109998Smarkm [Richard Levitte] 1980109998Smarkm 1981109998Smarkm *) Add OCSP code that comes from CertCo. 1982109998Smarkm [Richard Levitte] 1983109998Smarkm 1984109998Smarkm *) Add VMS support for the Rijndael code. 1985109998Smarkm [Richard Levitte] 1986109998Smarkm 1987109998Smarkm *) Added untested support for Nuron crypto accelerator. 1988109998Smarkm [Ben Laurie] 1989109998Smarkm 1990109998Smarkm *) Add support for external cryptographic devices. This code was 1991109998Smarkm previously distributed separately as the "engine" branch. 1992109998Smarkm [Geoff Thorpe, Richard Levitte] 1993109998Smarkm 1994109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 1995109998Smarkm have far greater control over how a "name" is turned into a filename 1996109998Smarkm depending on the operating environment and any oddities about the 1997109998Smarkm different shared library filenames on each system. 1998109998Smarkm [Geoff Thorpe] 1999109998Smarkm 2000109998Smarkm *) Support threads on FreeBSD-elf in Configure. 2001109998Smarkm [Richard Levitte] 2002109998Smarkm 2003109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 2004109998Smarkm warnings about corrupt line number information when assembling 2005109998Smarkm with debugging information. This is caused by the overlapping 2006109998Smarkm of two sections. 2007109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 2008109998Smarkm 2009109998Smarkm *) NCONF changes. 2010109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 2011109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 2012109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 2013109998Smarkm binary backward compatibility. 2014109998Smarkm Make it possible for methods to load from something other than a BIO, 2015109998Smarkm by providing a function pointer that is given a name instead of a BIO. 2016109998Smarkm For example, this could be used to load configuration data from an 2017109998Smarkm LDAP server. 2018109998Smarkm [Richard Levitte] 2019109998Smarkm 2020109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 2021109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 2022109998Smarkm with non blocking I/O was not possible because no retry code was 2023109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 2024109998Smarkm this case. 2025109998Smarkm [Steve Henson] 2026109998Smarkm 2027109998Smarkm *) Added the beginnings of Rijndael support. 2028109998Smarkm [Ben Laurie] 2029109998Smarkm 2030109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 2031109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 2032109998Smarkm to allow certificate printing to more controllable, additional 2033109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 2034109998Smarkm set. 2035109998Smarkm [Steve Henson] 2036109998Smarkm 2037109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 2038109998Smarkm [Richard Levitte] 2039109998Smarkm 2040120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 2041120631Snectar 2042120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 2043120631Snectar 2044120631Snectar Stop out of bounds reads in the ASN1 code when presented with 2045120631Snectar invalid tags (CAN-2003-0543 and CAN-2003-0544). 2046120631Snectar 2047120631Snectar If verify callback ignores invalid public key errors don't try to check 2048120631Snectar certificate signature with the NULL public key. 2049120631Snectar 2050120631Snectar [Steve Henson] 2051120631Snectar 2052120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 2053120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 2054120631Snectar specifications. 2055120631Snectar [Steve Henson] 2056120631Snectar 2057120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 2058120631Snectar extra data after the compression methods not only for TLS 1.0 2059120631Snectar but also for SSL 3.0 (as required by the specification). 2060120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 2061120631Snectar 2062120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 2063120631Snectar when it's 512 *bits* long, not 512 bytes. 2064120631Snectar [Richard Levitte] 2065120631Snectar 2066120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 2067120631Snectar 2068120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 2069120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 2070120631Snectar a protocol version number mismatch like a decryption error 2071120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 2072120631Snectar [Bodo Moeller] 2073120631Snectar 2074120631Snectar *) Turn on RSA blinding by default in the default implementation 2075120631Snectar to avoid a timing attack. Applications that don't want it can call 2076120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 2077120631Snectar They would be ill-advised to do so in most cases. 2078120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 2079120631Snectar 2080120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 2081120631Snectar seeded (in this case, the secret RSA exponent is abused as 2082120631Snectar an unpredictable seed -- if it is not unpredictable, there 2083120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 2084120631Snectar by remembering the creator's thread ID in rsa->blinding and 2085120631Snectar having all other threads use local one-time blinding factors 2086120631Snectar (this requires more computation than sharing rsa->blinding, but 2087120631Snectar avoids excessive locking; and if an RSA object is not shared 2088120631Snectar between threads, blinding will still be very fast). 2089120631Snectar [Bodo Moeller] 2090120631Snectar 2091111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 2092111147Snectar 2093111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 2094111147Snectar via timing by performing a MAC computation even if incorrrect 2095111147Snectar block cipher padding has been found. This is a countermeasure 2096111147Snectar against active attacks where the attacker has to distinguish 2097111147Snectar between bad padding and a MAC verification error. (CAN-2003-0078) 2098111147Snectar 2099111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 2100111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 2101111147Snectar Martin Vuagnoux (EPFL, Ilion)] 2102111147Snectar 2103109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 2104109998Smarkm 2105109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 2106109998Smarkm memory from it's contents. This is done with a counter that will 2107109998Smarkm place alternating values in each byte. This can be used to solve 2108109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 2109109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 2110109998Smarkm be read through on certain media, for example a swap space on disk. 2111109998Smarkm [Geoff Thorpe] 2112109998Smarkm 2113109998Smarkm *) Bugfix: client side session caching did not work with external caching, 2114109998Smarkm because the session->cipher setting was not restored when reloading 2115109998Smarkm from the external cache. This problem was masked, when 2116109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 2117109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 2118109998Smarkm [Lutz Jaenicke] 2119109998Smarkm 2120109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 2121109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 2122109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 2123109998Smarkm 2124109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 2125109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 2126109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 2127109998Smarkm EVP_cleanup(). 2128109998Smarkm [Richard Levitte] 2129109998Smarkm 2130109998Smarkm *) Change the default configuration reader to deal with last line not 2131109998Smarkm being properly terminated. 2132109998Smarkm [Richard Levitte] 2133109998Smarkm 2134109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 2135109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 2136109998Smarkm emailAddress where the value has the type ia5String. 2137109998Smarkm [stefank@valicert.com via Richard Levitte] 2138109998Smarkm 2139109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 2140109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 2141109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 2142109998Smarkm the bitwise-OR of the two for use by the majority of applications 2143109998Smarkm wanting this behaviour, and update the docs. The documented 2144109998Smarkm behaviour and actual behaviour were inconsistent and had been 2145109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 2146109998Smarkm change. 2147109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 2148109998Smarkm 2149109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 2150109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 2151109998Smarkm [Bodo Moeller] 2152109998Smarkm 2153109998Smarkm *) Fix initialization code race conditions in 2154109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 2155109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 2156109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 2157109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 2158109998Smarkm ssl2_get_cipher_by_char(), 2159109998Smarkm ssl3_get_cipher_by_char(). 2160109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 2161109998Smarkm 2162109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 2163109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 2164109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 2165109998Smarkm (see [openssl.org #212]). 2166109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 2167109998Smarkm 2168109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 2169109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 2170109998Smarkm [Steve Henson] 2171109998Smarkm 2172101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 2173101618Snectar 2174101618Snectar *) [In 0.9.6g-engine release:] 2175101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 2176101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 2177101618Snectar 2178101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 2179101613Snectar 2180101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 2181101613Snectar and get fix the header length calculation. 2182101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 2183101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 2184101613Snectar Steve Henson] 2185101613Snectar 2186101613Snectar *) Use proper error handling instead of 'assertions' in buffer 2187101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 2188101613Snectar assertions could call abort()). 2189101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 2190101613Snectar 2191100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 2192100936Snectar 2193109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 2194109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 2195109998Smarkm negative or the content length exceeds the length of the 2196109998Smarkm supplied buffer. 2197109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2198109998Smarkm 2199100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 2200100936Snectar for the cipher strength set and where therefore not handled correctly 2201100936Snectar by the selection routines (PR #130). 2202100936Snectar [Lutz Jaenicke] 2203100936Snectar 2204100936Snectar *) Fix EVP_dsa_sha macro. 2205100936Snectar [Nils Larsch] 2206100936Snectar 2207100936Snectar *) New option 2208100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 2209100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 2210100936Snectar that was added in OpenSSL 0.9.6d. 2211100936Snectar 2212100936Snectar As the countermeasure turned out to be incompatible with some 2213100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 2214100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 2215100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 2216100936Snectar 's_server'), so the new option is automatically set in many 2217100936Snectar applications. 2218100936Snectar [Bodo Moeller] 2219100936Snectar 2220100936Snectar *) Changes in security patch: 2221100936Snectar 2222100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 2223100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 2224100936Snectar Air Force Materiel Command, USAF, under agreement number 2225100936Snectar F30602-01-2-0537. 2226100936Snectar 2227100936Snectar *) Add various sanity checks to asn1_get_length() to reject 2228100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 2229100936Snectar negative or the content length exceeds the length of the 2230109998Smarkm supplied buffer. (CAN-2002-0659) 2231100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2232100936Snectar 2233100936Snectar *) Assertions for various potential buffer overflows, not known to 2234100936Snectar happen in practice. 2235100936Snectar [Ben Laurie (CHATS)] 2236100936Snectar 2237100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 2238100936Snectar too small for 64 bit platforms. (CAN-2002-0655) 2239100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 2240100936Snectar 2241100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 2242100936Snectar supply an oversized session ID to a client. (CAN-2002-0656) 2243100936Snectar [Ben Laurie (CHATS)] 2244100936Snectar 2245100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 2246100936Snectar supply an oversized client master key. (CAN-2002-0656) 2247100936Snectar [Ben Laurie (CHATS)] 2248100936Snectar 2249100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 2250100928Snectar 2251100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 2252100928Snectar encoded as NULL) with id-dsa-with-sha1. 2253100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 2254100928Snectar 2255100928Snectar *) Check various X509_...() return values in apps/req.c. 2256100928Snectar [Nils Larsch <nla@trustcenter.de>] 2257100928Snectar 2258100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 2259100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 2260100928Snectar was just at the end of a processed block. The bug was discovered when 2261100928Snectar processing data through a buffering memory BIO handing the data to a 2262100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 2263100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 2264100928Snectar [Lutz Jaenicke] 2265100928Snectar 2266100928Snectar *) Implement a countermeasure against a vulnerability recently found 2267100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 2268100928Snectar before application data chunks to avoid the use of known IVs 2269100928Snectar with data potentially chosen by the attacker. 2270100928Snectar [Bodo Moeller] 2271100928Snectar 2272100928Snectar *) Fix length checks in ssl3_get_client_hello(). 2273100928Snectar [Bodo Moeller] 2274100928Snectar 2275100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 2276100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 2277100928Snectar ssl3_read_bytes() found application data while handshake 2278100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 2279100928Snectar merely automatically cleared during the initial handshake. 2280100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 2281100928Snectar 2282100928Snectar *) Fix object definitions for Private and Enterprise: they were not 2283100928Snectar recognized in their shortname (=lowercase) representation. Extend 2284100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 2285100928Snectar of silently ignoring the problem (Svenning Sorensen 2286100928Snectar <sss@sss.dnsalias.net>). 2287100928Snectar [Lutz Jaenicke] 2288100928Snectar 2289100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 2290100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 2291100928Snectar code did not properly initialise the 'add' and 'rem' values to 2292100928Snectar BN_generate_prime().) 2293100928Snectar 2294100928Snectar In the new general case, we do not insist that 'generator' is 2295100928Snectar actually a primitive root: This requirement is rather pointless; 2296100928Snectar a generator of the order-q subgroup is just as good, if not 2297100928Snectar better. 2298100928Snectar [Bodo Moeller] 2299100928Snectar 2300100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 2301100928Snectar Tom Wu <tom@arcot.com>. 2302100928Snectar [Lutz Jaenicke] 2303100928Snectar 2304100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 2305100928Snectar returning non-zero before the data has been completely received 2306100928Snectar when using non-blocking I/O. 2307100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 2308100928Snectar 2309100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 2310100928Snectar [Ben Laurie, Lutz Jaenicke] 2311100928Snectar 2312100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 2313100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 2314100928Snectar [Lutz Jaenicke] 2315100928Snectar 2316100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 2317100928Snectar configuration for the versions before that. 2318100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2319100928Snectar 2320100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 2321100928Snectar check whether we deal with a copy of a session and do not delete from 2322100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 2323100928Snectar <izhar@checkpoint.com>. 2324100928Snectar [Lutz Jaenicke] 2325100928Snectar 2326100928Snectar *) Do not store session data into the internal session cache, if it 2327100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 2328100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 2329100928Snectar [Lutz Jaenicke] 2330100928Snectar 2331100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 2332100928Snectar value is 0. 2333100928Snectar [Richard Levitte] 2334100928Snectar 2335109998Smarkm *) [In 0.9.6d-engine release:] 2336109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2337100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2338100928Snectar 2339100928Snectar *) Add the configuration target linux-s390x. 2340100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 2341100928Snectar 2342100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 2343100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 2344100928Snectar variable as an indication that a ClientHello message has been 2345100928Snectar received. As the flag value will be lost between multiple 2346100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 2347100928Snectar function may not be aware that a handshake has actually taken 2348100928Snectar place, thus preventing a new session from being added to the 2349100928Snectar session cache. 2350100928Snectar 2351100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 2352100928Snectar using a local variable. 2353100928Snectar [Lutz Jaenicke, Bodo Moeller] 2354100928Snectar 2355100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 2356100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 2357100928Snectar [Geoff Thorpe, Bodo Moeller] 2358100928Snectar 2359100928Snectar *) New 'shared_ldflag' column in Configure platform table. 2360100928Snectar [Richard Levitte] 2361100928Snectar 2362100928Snectar *) Fix EVP_CIPHER_mode macro. 2363100928Snectar ["Dan S. Camper" <dan@bti.net>] 2364100928Snectar 2365100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 2366100928Snectar type, we must throw them away by setting rr->length to 0. 2367100928Snectar [D P Chang <dpc@qualys.com>] 2368100928Snectar 236989837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 237089837Skris 237189837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 237289837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 237389837Skris worked incorrectly for those cases where range = 10..._2 and 237489837Skris 3*range is two bits longer than range.) 237589837Skris [Bodo Moeller] 237689837Skris 237789837Skris *) Only add signing time to PKCS7 structures if it is not already 237889837Skris present. 237989837Skris [Steve Henson] 238089837Skris 238189837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 238289837Skris OBJ_ld_ce should be OBJ_id_ce. 238389837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 238489837Skris incorrect (cf. RFC 3039). 238589837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 238689837Skris 238789837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 238889837Skris returns early because it has nothing to do. 238989837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 239089837Skris 239189837Skris *) [In 0.9.6c-engine release:] 239289837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 239389837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 239489837Skris 239589837Skris *) [In 0.9.6c-engine release:] 239689837Skris Add support for Cryptographic Appliance's keyserver technology. 239789837Skris (Use engine 'keyclient') 239889837Skris [Cryptographic Appliances and Geoff Thorpe] 239989837Skris 240089837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 240189837Skris is called via tools/c89.sh because arguments have to be 240289837Skris rearranged (all '-L' options must appear before the first object 240389837Skris modules). 240489837Skris [Richard Shapiro <rshapiro@abinitio.com>] 240589837Skris 240689837Skris *) [In 0.9.6c-engine release:] 240789837Skris Add support for Broadcom crypto accelerator cards, backported 240889837Skris from 0.9.7. 240989837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 241089837Skris 241189837Skris *) [In 0.9.6c-engine release:] 241289837Skris Add support for SureWare crypto accelerator cards from 241389837Skris Baltimore Technologies. (Use engine 'sureware') 241489837Skris [Baltimore Technologies and Mark Cox] 241589837Skris 241689837Skris *) [In 0.9.6c-engine release:] 241789837Skris Add support for crypto accelerator cards from Accelerated 241889837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 241989837Skris [AEP Inc. and Mark Cox] 242089837Skris 242189837Skris *) Add a configuration entry for gcc on UnixWare. 242289837Skris [Gary Benson <gbenson@redhat.com>] 242389837Skris 242489837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 242589837Skris messages are stored in a single piece (fixed-length part and 242689837Skris variable-length part combined) and fix various bugs found on the way. 242789837Skris [Bodo Moeller] 242889837Skris 242989837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 243089837Skris instead. BIO_gethostbyname() does not know what timeouts are 243189837Skris appropriate, so entries would stay in cache even when they have 243289837Skris become invalid. 243389837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 243489837Skris 243589837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 243689837Skris faced with a pathologically small ClientHello fragment that does 243789837Skris not contain client_version: Instead of aborting with an error, 243889837Skris simply choose the highest available protocol version (i.e., 243989837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 244089837Skris messages are never sent like this, but this change gives us 244189837Skris strictly correct behaviour at least for TLS. 244289837Skris [Bodo Moeller] 244389837Skris 244489837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 244589837Skris never resets s->method to s->ctx->method when called from within 244689837Skris one of the SSL handshake functions. 244789837Skris [Bodo Moeller; problem pointed out by Niko Baric] 244889837Skris 244989837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 245089837Skris (sent using the client's version number) if client_version is 245189837Skris smaller than the protocol version in use. Also change 245289837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 245389837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 245489837Skris the client will at least see that alert. 245589837Skris [Bodo Moeller] 245689837Skris 245789837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 245889837Skris correctly. 245989837Skris [Bodo Moeller] 246089837Skris 246189837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 246289837Skris client receives HelloRequest while in a handshake. 246389837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 246489837Skris 246589837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 246689837Skris should end in 'break', not 'goto end' which circuments various 246789837Skris cleanups done in state SSL_ST_OK. But session related stuff 246889837Skris must be disabled for SSL_ST_OK in the case that we just sent a 246989837Skris HelloRequest. 247089837Skris 247189837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 247289837Skris before just sending a HelloRequest. 247389837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 247489837Skris 247589837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 247689837Skris reveal whether illegal block cipher padding was found or a MAC 247789837Skris verification error occured. (Neither SSLerr() codes nor alerts 247889837Skris are directly visible to potential attackers, but the information 247989837Skris may leak via logfiles.) 248089837Skris 248189837Skris Similar changes are not required for the SSL 2.0 implementation 248289837Skris because the number of padding bytes is sent in clear for SSL 2.0, 248389837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 248489837Skris failed to verify that the purported number of padding bytes is in 248589837Skris the legal range. 248689837Skris [Bodo Moeller] 248789837Skris 248889837Skris *) Add OpenUNIX-8 support including shared libraries 248989837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 249089837Skris [Lutz Jaenicke] 249189837Skris 249289837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 249389837Skris 'wristwatch attack' using huge encoding parameters (cf. 249489837Skris James H. Manger's CRYPTO 2001 paper). Note that the 249589837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 249689837Skris encoding parameters and hence was not vulnerable. 249789837Skris [Bodo Moeller] 249889837Skris 249989837Skris *) BN_sqr() bug fix. 250089837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 250189837Skris 250289837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 250389837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 250489837Skris followed by modular reduction. 250589837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 250689837Skris 250789837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 250889837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 250989837Skris [Bodo Moeller] 251089837Skris 251189837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 251289837Skris This function was broken, as the check for a new client hello message 251389837Skris to handle SGC did not allow these large messages. 251489837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 251589837Skris [Lutz Jaenicke] 251689837Skris 251789837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 251889837Skris [Lutz Jaenicke] 251989837Skris 252089837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 252189837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 252289837Skris [Lutz Jaenicke] 252389837Skris 252489837Skris *) Rework the configuration and shared library support for Tru64 Unix. 252589837Skris The configuration part makes use of modern compiler features and 252689837Skris still retains old compiler behavior for those that run older versions 252789837Skris of the OS. The shared library support part includes a variant that 252889837Skris uses the RPATH feature, and is available through the special 252989837Skris configuration target "alpha-cc-rpath", which will never be selected 253089837Skris automatically. 253189837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 253289837Skris 253389837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 253489837Skris with the same message size as in ssl3_get_certificate_request(). 253589837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 253689837Skris messages might inadvertently be reject as too long. 253789837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 253889837Skris 253989837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 254089837Skris [Andy Polyakov] 254189837Skris 254289837Skris *) Modified SSL library such that the verify_callback that has been set 254389837Skris specificly for an SSL object with SSL_set_verify() is actually being 254489837Skris used. Before the change, a verify_callback set with this function was 254589837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 254689837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 254789837Skris to allow the necessary settings. 254889837Skris [Lutz Jaenicke] 254989837Skris 255089837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 255189837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 255289837Skris done automatically (in contradiction to the requirements of the C 255389837Skris standard). This made problems when used from OpenSSH. 255489837Skris [Lutz Jaenicke] 255589837Skris 255689837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 255789837Skris dh->length and always used 255889837Skris 255989837Skris BN_rand_range(priv_key, dh->p). 256089837Skris 256189837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 256289837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 256389837Skris dh->length (recommended exponent length) is much smaller than the 256489837Skris length of dh->p. We could use BN_rand_range() if the order of 256589837Skris the subgroup was stored in the DH structure, but we only have 256689837Skris dh->length. 256789837Skris 256889837Skris So switch back to 256989837Skris 257089837Skris BN_rand(priv_key, l, ...) 257189837Skris 257289837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 257389837Skris otherwise. 257489837Skris [Bodo Moeller] 257589837Skris 257689837Skris *) In 257789837Skris 257889837Skris RSA_eay_public_encrypt 257989837Skris RSA_eay_private_decrypt 258089837Skris RSA_eay_private_encrypt (signing) 258189837Skris RSA_eay_public_decrypt (signature verification) 258289837Skris 258389837Skris (default implementations for RSA_public_encrypt, 258489837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 258589837Skris always reject numbers >= n. 258689837Skris [Bodo Moeller] 258789837Skris 258889837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 258989837Skris to synchronize access to 'locking_thread'. This is necessary on 259089837Skris systems where access to 'locking_thread' (an 'unsigned long' 259189837Skris variable) is not atomic. 259289837Skris [Bodo Moeller] 259389837Skris 259489837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 259589837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 259689837Skris a race condition if 0 is a valid thread ID. 259789837Skris [Travis Vitek <vitek@roguewave.com>] 259889837Skris 259989837Skris *) Add support for shared libraries under Irix. 260089837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 260189837Skris 260289837Skris *) Add configuration option to build on Linux on both big-endian and 260389837Skris little-endian MIPS. 260489837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 260589837Skris 260689837Skris *) Add the possibility to create shared libraries on HP-UX. 260789837Skris [Richard Levitte] 260889837Skris 260979998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 261079998Skris 261179998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 261279998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 261379998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 261479998Skris PRNG state recovery was possible based on the output of 261579998Skris one PRNG request appropriately sized to gain knowledge on 261679998Skris 'md' followed by enough consecutive 1-byte PRNG requests 261779998Skris to traverse all of 'state'. 261879998Skris 261979998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 262079998Skris during PRNG output generation, hash all of the previous 262179998Skris 'md_local' value, not just the half used for PRNG output. 262279998Skris 262379998Skris 2. Make the number of bytes from 'state' included into the hash 262479998Skris independent from the number of PRNG bytes requested. 262579998Skris 262679998Skris The first measure alone would be sufficient to avoid 262779998Skris Markku-Juhani's attack. (Actually it had never occurred 262879998Skris to me that the half of 'md_local' used for chaining was the 262979998Skris half from which PRNG output bytes were taken -- I had always 263079998Skris assumed that the secret half would be used.) The second 263179998Skris measure makes sure that additional data from 'state' is never 263279998Skris mixed into 'md_local' in small portions; this heuristically 263379998Skris further strengthens the PRNG. 263479998Skris [Bodo Moeller] 263579998Skris 263679998Skris *) Fix crypto/bn/asm/mips3.s. 263779998Skris [Andy Polyakov] 263879998Skris 263979998Skris *) When only the key is given to "enc", the IV is undefined. Print out 264079998Skris an error message in this case. 264179998Skris [Lutz Jaenicke] 264279998Skris 264379998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 264479998Skris [Steve Henson] 264579998Skris 264679998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 264779998Skris positive and less than q. 264879998Skris [Bodo Moeller] 264979998Skris 265079998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 265179998Skris used: it isn't thread safe and the add_lock_callback should handle 265279998Skris that itself. 265379998Skris [Paul Rose <Paul.Rose@bridge.com>] 265479998Skris 265579998Skris *) Verify that incoming data obeys the block size in 265679998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 265779998Skris [Bodo Moeller] 265879998Skris 265979998Skris *) Fix OAEP check. 266079998Skris [Ulf M�ller, Bodo M�ller] 266179998Skris 266279998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 266389837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 266479998Skris when fixing the server behaviour for backwards-compatible 'client 266579998Skris hello' messages. (Note that the attack is impractical against 266679998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 266779998Skris means that the probability of guessing a valid ciphertext is 266879998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 266979998Skris paper.) 267079998Skris 267179998Skris Before 0.9.5, the countermeasure (hide the error by generating a 267279998Skris random 'decryption result') did not work properly because 267379998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 267479998Skris detect the supposedly ignored error. 267579998Skris 267679998Skris Both problems are now fixed. 267779998Skris [Bodo Moeller] 267879998Skris 267979998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 268079998Skris (previously it was 1024). 268179998Skris [Bodo Moeller] 268279998Skris 268379998Skris *) Fix for compatibility mode trust settings: ignore trust settings 268479998Skris unless some valid trust or reject settings are present. 268579998Skris [Steve Henson] 268679998Skris 268779998Skris *) Fix for blowfish EVP: its a variable length cipher. 268879998Skris [Steve Henson] 268979998Skris 269079998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 269179998Skris parameters in DSA public key structures and return an error in the 269279998Skris DSA routines if parameters are absent. 269379998Skris [Steve Henson] 269479998Skris 269579998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 269679998Skris in the current directory if neither $RANDFILE nor $HOME was set. 269779998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 269879998Skris caused some confusion to Windows users who haven't defined $HOME. 269979998Skris Thus RAND_file_name() is changed again: e_os.h can define a 270079998Skris DEFAULT_HOME, which will be used if $HOME is not set. 270179998Skris For Windows, we use "C:"; on other platforms, we still require 270279998Skris environment variables. 270379998Skris 270479998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 270579998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 270679998Skris having multiple threads call RAND_poll() concurrently. 270779998Skris [Bodo Moeller] 270879998Skris 270979998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 271079998Skris combination of a flag and a thread ID variable. 271179998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 271279998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 271389837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 271479998Skris that they do not hold after the first thread unsets add_do_not_lock). 271579998Skris [Bodo Moeller] 271679998Skris 271779998Skris *) Change bctest again: '-x' expressions are not available in all 271879998Skris versions of 'test'. 271979998Skris [Bodo Moeller] 272079998Skris 272176866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 272272613Skris 272376866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 272476866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 272576866Skris 272676866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 272776866Skris the default extension for executables, if any. Also, make the perl 272876866Skris scripts that use symlink() to test if it really exists and use "cp" 272976866Skris if it doesn't. All this made OpenSSL compilable and installable in 273076866Skris CygWin. 273176866Skris [Richard Levitte] 273276866Skris 273376866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 273476866Skris If SEQUENCE is length is indefinite just set c->slen to the total 273576866Skris amount of data available. 273676866Skris [Steve Henson, reported by shige@FreeBSD.org] 273776866Skris [This change does not apply to 0.9.7.] 273876866Skris 273976866Skris *) Change bctest to avoid here-documents inside command substitution 274076866Skris (workaround for FreeBSD /bin/sh bug). 274176866Skris For compatibility with Ultrix, avoid shell functions (introduced 274276866Skris in the bctest version that searches along $PATH). 274376866Skris [Bodo Moeller] 274476866Skris 274576866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 274676866Skris with des_encrypt() defined on some operating systems, like Solaris 274776866Skris and UnixWare. 274876866Skris [Richard Levitte] 274976866Skris 275076866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 275176866Skris On the Importance of Eliminating Errors in Cryptographic 275276866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 275376866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 275476866Skris [Ulf Moeller] 275576866Skris 275676866Skris *) MIPS assembler BIGNUM division bug fix. 275776866Skris [Andy Polyakov] 275876866Skris 275976866Skris *) Disabled incorrect Alpha assembler code. 276076866Skris [Richard Levitte] 276176866Skris 276276866Skris *) Fix PKCS#7 decode routines so they correctly update the length 276376866Skris after reading an EOC for the EXPLICIT tag. 276476866Skris [Steve Henson] 276576866Skris [This change does not apply to 0.9.7.] 276676866Skris 276776866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 276876866Skris if a 3DES key was generated with a 0 initial byte. Include 276976866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 277076866Skris (but broken) behaviour. 277176866Skris [Steve Henson] 277276866Skris 277376866Skris *) Enhance bctest to search for a working bc along $PATH and print 277476866Skris it when found. 277576866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 277676866Skris 277776866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 277876866Skris don't write to the wrong index in ERR_set_error_data. 277976866Skris [Bodo Moeller] 278076866Skris 278176866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 278276866Skris did not exist. 278376866Skris [Bodo Moeller] 278476866Skris 278576866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 278676866Skris [Jeremy Cooper <jeremy@baymoo.org>] 278776866Skris 278876866Skris *) Make it possible to reuse SSLv2 sessions. 278976866Skris [Richard Levitte] 279076866Skris 279176866Skris *) In copy_email() check for >= 0 as a return value for 279276866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 279376866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 279476866Skris 279576866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 279676866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 279776866Skris PKCS7_verify() fails with non detached data. 279876866Skris [Steve Henson] 279976866Skris 280076866Skris *) Don't use getenv in library functions when run as setuid/setgid. 280176866Skris New function OPENSSL_issetugid(). 280276866Skris [Ulf Moeller] 280376866Skris 280476866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 280576866Skris due to incorrect handling of multi-threading: 280676866Skris 280776866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 280876866Skris 280976866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 281076866Skris 281176866Skris 3. Count how many times MemCheck_off() has been called so that 281276866Skris nested use can be treated correctly. This also avoids 281376866Skris inband-signalling in the previous code (which relied on the 281476866Skris assumption that thread ID 0 is impossible). 281576866Skris [Bodo Moeller] 281676866Skris 281776866Skris *) Add "-rand" option also to s_client and s_server. 281876866Skris [Lutz Jaenicke] 281976866Skris 282076866Skris *) Fix CPU detection on Irix 6.x. 282176866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 282276866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 282376866Skris 282476866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 282576866Skris was empty. 282676866Skris [Steve Henson] 282776866Skris [This change does not apply to 0.9.7.] 282876866Skris 282976866Skris *) Use the cached encoding of an X509_NAME structure rather than 283076866Skris copying it. This is apparently the reason for the libsafe "errors" 283176866Skris but the code is actually correct. 283276866Skris [Steve Henson] 283376866Skris 283472613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 283572613Skris Bleichenbacher's DSA attack. 283676866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 283776866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 283876866Skris and leaves the highest bit random. 283976866Skris [Ulf Moeller, Bodo Moeller] 284072613Skris 284172613Skris *) In the NCONF_...-based implementations for CONF_... queries 284272613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 284372613Skris a temporary CONF structure with the data component set to NULL 284472613Skris (which gives segmentation faults in lh_retrieve). 284572613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 284672613Skris CONF_get_number (which may use environment variables) and directly 284772613Skris return NULL from CONF_get_section. 284872613Skris [Bodo Moeller] 284972613Skris 285072613Skris *) Fix potential buffer overrun for EBCDIC. 285172613Skris [Ulf Moeller] 285272613Skris 285372613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 285472613Skris keyUsage if basicConstraints absent for a CA. 285572613Skris [Steve Henson] 285672613Skris 285772613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 285872613Skris is more generally accepted (no spaces before the semicolon), since 285972613Skris some programs can't parse those values properly otherwise. Also make 286072613Skris sure BIO's that break lines after each write do not create invalid 286172613Skris headers. 286272613Skris [Richard Levitte] 286372613Skris 286472613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 286572613Skris macros previously used would not encode an empty SEQUENCE OF 286672613Skris and break the signature. 286772613Skris [Steve Henson] 286876866Skris [This change does not apply to 0.9.7.] 286972613Skris 287072613Skris *) Zero the premaster secret after deriving the master secret in 287172613Skris DH ciphersuites. 287272613Skris [Steve Henson] 287372613Skris 287472613Skris *) Add some EVP_add_digest_alias registrations (as found in 287572613Skris OpenSSL_add_all_digests()) to SSL_library_init() 287672613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 287772613Skris compatibility with peers using X.509 certificates 287872613Skris with unconventional AlgorithmIdentifier OIDs. 287972613Skris [Bodo Moeller] 288072613Skris 288172613Skris *) Fix for Irix with NO_ASM. 288272613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 288372613Skris 288472613Skris *) ./config script fixes. 288572613Skris [Ulf Moeller, Richard Levitte] 288672613Skris 288772613Skris *) Fix 'openssl passwd -1'. 288872613Skris [Bodo Moeller] 288972613Skris 289072613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 289172613Skris terminated strings whose length is passed in the passlen 289272613Skris parameter, for example from PEM callbacks. This was done 289372613Skris by adding an extra length parameter to asc2uni(). 289472613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 289572613Skris 289672613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 289772613Skris call failed, free the DSA structure. 289872613Skris [Bodo Moeller] 289972613Skris 290072613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 290172613Skris These are present in some PKCS#12 files. 290272613Skris [Steve Henson] 290372613Skris 290472613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 290572613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 290672613Skris when writing a 32767 byte record. 290772613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 290872613Skris 290972613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 291072613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 291172613Skris 291272613Skris (RSA objects have a reference count access to which is protected 291372613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 291472613Skris so they are meant to be shared between threads.) 291572613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 291672613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 291772613Skris 291872613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 291972613Skris [Bodo Moeller] 292072613Skris 292176866Skris *) Use better test patterns in bntest. 292276866Skris [Ulf M�ller] 292376866Skris 292472613Skris *) rand_win.c fix for Borland C. 292572613Skris [Ulf M�ller] 292672613Skris 292772613Skris *) BN_rshift bugfix for n == 0. 292872613Skris [Bodo Moeller] 292972613Skris 293076866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 293176866Skris so that 'make test' does not abort just because 'bc' is broken. 293276866Skris [Bodo Moeller] 293376866Skris 293472613Skris *) Store verify_result within SSL_SESSION also for client side to 293572613Skris avoid potential security hole. (Re-used sessions on the client side 293672613Skris always resulted in verify_result==X509_V_OK, not using the original 293772613Skris result of the server certificate verification.) 293872613Skris [Lutz Jaenicke] 293972613Skris 294072613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 294172613Skris SSL3_RT_APPLICATION_DATA, return 0. 294272613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 294372613Skris [Bodo Moeller] 294472613Skris 294572613Skris *) Fix SSL_peek: 294672613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 294772613Skris releases, have been re-implemented by renaming the previous 294872613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 294972613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 295072613Skris to them. The new ssl[23]_{read,peek} functions are calls to 295172613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 295272613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 295372613Skris does the actual work for ssl3_read_internal. 295472613Skris [Bodo Moeller] 295572613Skris 295676866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 295776866Skris the method-specific "init()" handler. Also clean up ex_data after 295876866Skris calling the method-specific "finish()" handler. Previously, this was 295976866Skris happening the other way round. 296076866Skris [Geoff Thorpe] 296176866Skris 296272613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 296372613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 296472613Skris [Bodo Moeller] 296572613Skris 296676866Skris *) Make sure that shared libraries get the internal name engine with 296776866Skris the full version number and not just 0. This should mark the 296876866Skris shared libraries as not backward compatible. Of course, this should 296976866Skris be changed again when we can guarantee backward binary compatibility. 297076866Skris [Richard Levitte] 297176866Skris 297272613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 297372613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 297472613Skris 297576866Skris *) Rework the system to generate shared libraries: 297676866Skris 297776866Skris - Make note of the expected extension for the shared libraries and 297876866Skris if there is a need for symbolic links from for example libcrypto.so.0 297976866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 298076866Skris that. 298176866Skris 298276866Skris - Make as few rebuilds of the shared libraries as possible. 298376866Skris 298476866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 298576866Skris 298676866Skris - When installing, install the shared libraries separately from the 298776866Skris static ones. 298876866Skris [Richard Levitte] 298976866Skris 299072613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 299172613Skris 299272613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 299372613Skris and not in SSL_clear because the latter is also used by the 299472613Skris accept/connect functions; previously, the settings made by 299572613Skris SSL_set_read_ahead would be lost during the handshake. 299672613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 299772613Skris 299872613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 299972613Skris Previously, it would create entries for disableed algorithms no 300072613Skris matter what. 300172613Skris [Richard Levitte] 300272613Skris 300372613Skris *) Added several new manual pages for SSL_* function. 300472613Skris [Lutz Jaenicke] 300572613Skris 300668651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 300768651Skris 300868651Skris *) In ssl23_get_client_hello, generate an error message when faced 300968651Skris with an initial SSL 3.0/TLS record that is too small to contain the 301068651Skris first two bytes of the ClientHello message, i.e. client_version. 301168651Skris (Note that this is a pathologic case that probably has never happened 301268651Skris in real life.) The previous approach was to use the version number 301368651Skris from the record header as a substitute; but our protocol choice 301468651Skris should not depend on that one because it is not authenticated 301568651Skris by the Finished messages. 301668651Skris [Bodo Moeller] 301768651Skris 301868651Skris *) More robust randomness gathering functions for Windows. 301968651Skris [Jeffrey Altman <jaltman@columbia.edu>] 302068651Skris 302168651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 302268651Skris not set then we don't setup the error code for issuer check errors 302368651Skris to avoid possibly overwriting other errors which the callback does 302468651Skris handle. If an application does set the flag then we assume it knows 302568651Skris what it is doing and can handle the new informational codes 302668651Skris appropriately. 302768651Skris [Steve Henson] 302868651Skris 302968651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 303068651Skris a general "ANY" type, as such it should be able to decode anything 303168651Skris including tagged types. However it didn't check the class so it would 303268651Skris wrongly interpret tagged types in the same way as their universal 303368651Skris counterpart and unknown types were just rejected. Changed so that the 303468651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 303568651Skris that is the encoding is stored intact. There is also a new type 303668651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 303768651Skris case we have no idea what the actual type is so we just lump them all 303868651Skris together. 303968651Skris [Steve Henson] 304068651Skris 304168651Skris *) On VMS, stdout may very well lead to a file that is written to 304268651Skris in a record-oriented fashion. That means that every write() will 304368651Skris write a separate record, which will be read separately by the 304468651Skris programs trying to read from it. This can be very confusing. 304568651Skris 304668651Skris The solution is to put a BIO filter in the way that will buffer 304768651Skris text until a linefeed is reached, and then write everything a 304868651Skris line at a time, so every record written will be an actual line, 304968651Skris not chunks of lines and not (usually doesn't happen, but I've 305068651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 305168651Skris the answer. 305268651Skris 305368651Skris Currently, it's a VMS-only method, because that's where it has 305468651Skris been tested well enough. 305568651Skris [Richard Levitte] 305668651Skris 305768651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 305868651Skris it can return incorrect results. 305968651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 306068651Skris but it was in 0.9.6-beta[12].) 306168651Skris [Bodo Moeller] 306268651Skris 306368651Skris *) Disable the check for content being present when verifying detached 306468651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 306568651Skris include zero length content when signing messages. 306668651Skris [Steve Henson] 306768651Skris 306868651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 306968651Skris BIO_ctrl (for BIO pairs). 307068651Skris [Bodo M�ller] 307168651Skris 307268651Skris *) Add DSO method for VMS. 307368651Skris [Richard Levitte] 307468651Skris 307568651Skris *) Bug fix: Montgomery multiplication could produce results with the 307668651Skris wrong sign. 307768651Skris [Ulf M�ller] 307868651Skris 307968651Skris *) Add RPM specification openssl.spec and modify it to build three 308068651Skris packages. The default package contains applications, application 308168651Skris documentation and run-time libraries. The devel package contains 308268651Skris include files, static libraries and function documentation. The 308368651Skris doc package contains the contents of the doc directory. The original 308468651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 308568651Skris [Richard Levitte] 308668651Skris 308768651Skris *) Add a large number of documentation files for many SSL routines. 308868651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 308968651Skris 309068651Skris *) Add a configuration entry for Sony News 4. 309168651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 309268651Skris 309368651Skris *) Don't set the two most significant bits to one when generating a 309468651Skris random number < q in the DSA library. 309568651Skris [Ulf M�ller] 309668651Skris 309768651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 309868651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 309968651Skris the underlying transport is blocking) if a handshake took place. 310068651Skris (The default behaviour is needed by applications such as s_client 310168651Skris and s_server that use select() to determine when to use SSL_read; 310268651Skris but for applications that know in advance when to expect data, it 310368651Skris just makes things more complicated.) 310468651Skris [Bodo Moeller] 310568651Skris 310668651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 310768651Skris from EGD. 310868651Skris [Ben Laurie] 310968651Skris 311068651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 311168651Skris work better on such systems. 311268651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 311368651Skris 311468651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 311568651Skris Update PKCS12_parse() so it copies the friendlyName and the 311668651Skris keyid to the certificates aux info. 311768651Skris [Steve Henson] 311868651Skris 311968651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 312068651Skris if there was more than one signature. 312168651Skris [Sven Uszpelkat <su@celocom.de>] 312268651Skris 312368651Skris *) Major change in util/mkdef.pl to include extra information 312468651Skris about each symbol, as well as presentig variables as well 312568651Skris as functions. This change means that there's n more need 312668651Skris to rebuild the .num files when some algorithms are excluded. 312768651Skris [Richard Levitte] 312868651Skris 312968651Skris *) Allow the verify time to be set by an application, 313068651Skris rather than always using the current time. 313168651Skris [Steve Henson] 313268651Skris 313368651Skris *) Phase 2 verify code reorganisation. The certificate 313468651Skris verify code now looks up an issuer certificate by a 313568651Skris number of criteria: subject name, authority key id 313668651Skris and key usage. It also verifies self signed certificates 313768651Skris by the same criteria. The main comparison function is 313868651Skris X509_check_issued() which performs these checks. 313968651Skris 314068651Skris Lot of changes were necessary in order to support this 314168651Skris without completely rewriting the lookup code. 314268651Skris 314368651Skris Authority and subject key identifier are now cached. 314468651Skris 314568651Skris The LHASH 'certs' is X509_STORE has now been replaced 314668651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 314768651Skris LHASH can't store or retrieve multiple objects with 314868651Skris the same hash value. 314968651Skris 315068651Skris As a result various functions (which were all internal 315168651Skris use only) have changed to handle the new X509_STORE 315268651Skris structure. This will break anything that messed round 315368651Skris with X509_STORE internally. 315468651Skris 315568651Skris The functions X509_STORE_add_cert() now checks for an 315668651Skris exact match, rather than just subject name. 315768651Skris 315868651Skris The X509_STORE API doesn't directly support the retrieval 315968651Skris of multiple certificates matching a given criteria, however 316068651Skris this can be worked round by performing a lookup first 316168651Skris (which will fill the cache with candidate certificates) 316268651Skris and then examining the cache for matches. This is probably 316368651Skris the best we can do without throwing out X509_LOOKUP 316468651Skris entirely (maybe later...). 316568651Skris 316668651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 316768651Skris 316868651Skris All certificate lookup operations now go via a get_issuer() 316968651Skris callback. Although this currently uses an X509_STORE it 317068651Skris can be replaced by custom lookups. This is a simple way 317168651Skris to bypass the X509_STORE hackery necessary to make this 317268651Skris work and makes it possible to use more efficient techniques 317368651Skris in future. A very simple version which uses a simple 317468651Skris STACK for its trusted certificate store is also provided 317568651Skris using X509_STORE_CTX_trusted_stack(). 317668651Skris 317768651Skris The verify_cb() and verify() callbacks now have equivalents 317868651Skris in the X509_STORE_CTX structure. 317968651Skris 318068651Skris X509_STORE_CTX also has a 'flags' field which can be used 318168651Skris to customise the verify behaviour. 318268651Skris [Steve Henson] 318368651Skris 318468651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 318568651Skris excludes S/MIME capabilities. 318668651Skris [Steve Henson] 318768651Skris 318868651Skris *) When a certificate request is read in keep a copy of the 318968651Skris original encoding of the signed data and use it when outputing 319068651Skris again. Signatures then use the original encoding rather than 319168651Skris a decoded, encoded version which may cause problems if the 319268651Skris request is improperly encoded. 319368651Skris [Steve Henson] 319468651Skris 319568651Skris *) For consistency with other BIO_puts implementations, call 319668651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 319768651Skris BIO_write(b, ...). 319868651Skris 319968651Skris In BIO_puts, increment b->num_write as in BIO_write. 320068651Skris [Peter.Sylvester@EdelWeb.fr] 320168651Skris 320268651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 320368651Skris BN_zero, we may not return a BIGNUM with an array consisting of 320468651Skris words set to zero.) 320568651Skris [Bodo Moeller] 320668651Skris 320768651Skris *) Avoid calling abort() from within the library when problems are 320868651Skris detected, except if preprocessor symbols have been defined 320968651Skris (such as REF_CHECK, BN_DEBUG etc.). 321068651Skris [Bodo Moeller] 321168651Skris 321268651Skris *) New openssl application 'rsautl'. This utility can be 321368651Skris used for low level RSA operations. DER public key 321468651Skris BIO/fp routines also added. 321568651Skris [Steve Henson] 321668651Skris 321768651Skris *) New Configure entry and patches for compiling on QNX 4. 321868651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 321968651Skris 322068651Skris *) A demo state-machine implementation was sponsored by 322168651Skris Nuron (http://www.nuron.com/) and is now available in 322268651Skris demos/state_machine. 322368651Skris [Ben Laurie] 322468651Skris 322568651Skris *) New options added to the 'dgst' utility for signature 322668651Skris generation and verification. 322768651Skris [Steve Henson] 322868651Skris 322968651Skris *) Unrecognized PKCS#7 content types are now handled via a 323068651Skris catch all ASN1_TYPE structure. This allows unsupported 323168651Skris types to be stored as a "blob" and an application can 323268651Skris encode and decode it manually. 323368651Skris [Steve Henson] 323468651Skris 323568651Skris *) Fix various signed/unsigned issues to make a_strex.c 323668651Skris compile under VC++. 323768651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 323868651Skris 323968651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 324068651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 324168651Skris if passed a NULL BN and its argument was negative. 324268651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 324368651Skris 324468651Skris *) Modification to PKCS#7 encoding routines to output definite 324568651Skris length encoding. Since currently the whole structures are in 324668651Skris memory there's not real point in using indefinite length 324768651Skris constructed encoding. However if OpenSSL is compiled with 324868651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 324968651Skris [Steve Henson] 325068651Skris 325168651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 325268651Skris [Richard Levitte] 325368651Skris 325468651Skris *) Added more prefixes to parse for in the the strings written 325568651Skris through a logging bio, to cover all the levels that are available 325668651Skris through syslog. The prefixes are now: 325768651Skris 325868651Skris PANIC, EMERG, EMR => LOG_EMERG 325968651Skris ALERT, ALR => LOG_ALERT 326068651Skris CRIT, CRI => LOG_CRIT 326168651Skris ERROR, ERR => LOG_ERR 326268651Skris WARNING, WARN, WAR => LOG_WARNING 326368651Skris NOTICE, NOTE, NOT => LOG_NOTICE 326468651Skris INFO, INF => LOG_INFO 326568651Skris DEBUG, DBG => LOG_DEBUG 326668651Skris 326768651Skris and as before, if none of those prefixes are present at the 326868651Skris beginning of the string, LOG_ERR is chosen. 326968651Skris 327068651Skris On Win32, the LOG_* levels are mapped according to this: 327168651Skris 327268651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 327368651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 327468651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 327568651Skris 327668651Skris [Richard Levitte] 327768651Skris 327868651Skris *) Made it possible to reconfigure with just the configuration 327968651Skris argument "reconf" or "reconfigure". The command line arguments 328068651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 328168651Skris and are retrieved from there when reconfiguring. 328268651Skris [Richard Levitte] 328368651Skris 328468651Skris *) MD4 implemented. 328568651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 328668651Skris 328768651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 328868651Skris [Richard Levitte] 328968651Skris 329068651Skris *) The obj_dat.pl script was messing up the sorting of object 329168651Skris names. The reason was that it compared the quoted version 329268651Skris of strings as a result "OCSP" > "OCSP Signing" because 329368651Skris " > SPACE. Changed script to store unquoted versions of 329468651Skris names and add quotes on output. It was also omitting some 329568651Skris names from the lookup table if they were given a default 329668651Skris value (that is if SN is missing it is given the same 329768651Skris value as LN and vice versa), these are now added on the 329868651Skris grounds that if an object has a name we should be able to 329968651Skris look it up. Finally added warning output when duplicate 330068651Skris short or long names are found. 330168651Skris [Steve Henson] 330268651Skris 330368651Skris *) Changes needed for Tandem NSK. 330468651Skris [Scott Uroff <scott@xypro.com>] 330568651Skris 330668651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 330768651Skris RSA_padding_check_SSLv23(), special padding was never detected 330868651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 330968651Skris version rollback attacks was not effective. 331068651Skris 331168651Skris In s23_clnt.c, don't use special rollback-attack detection padding 331268651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 331368651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 331468651Skris SSL 2.0 is the only protocol enabled in the server. 331568651Skris [Bodo Moeller] 331668651Skris 331768651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 331868651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 331968651Skris BIO_dump_indent() are added. 332068651Skris [Richard Levitte] 332168651Skris 332268651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 332368651Skris these print out strings and name structures based on various 332468651Skris flags including RFC2253 support and proper handling of 332568651Skris multibyte characters. Added options to the 'x509' utility 332668651Skris to allow the various flags to be set. 332768651Skris [Steve Henson] 332868651Skris 332968651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 333068651Skris Also change the functions X509_cmp_current_time() and 333168651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 333268651Skris this will enable certificates using GeneralizedTime in validity 333368651Skris dates to be checked. 333468651Skris [Steve Henson] 333568651Skris 333668651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 333768651Skris negative public key encodings) on by default, 333868651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 333968651Skris [Steve Henson] 334068651Skris 334168651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 334268651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 334368651Skris the encoding can be trivially obtained from the structure. 334468651Skris [Steve Henson] 334568651Skris 334668651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 334768651Skris not read locks (CRYPTO_r_[un]lock). 334868651Skris [Bodo Moeller] 334968651Skris 335068651Skris *) A first attempt at creating official support for shared 335168651Skris libraries through configuration. I've kept it so the 335268651Skris default is static libraries only, and the OpenSSL programs 335368651Skris are always statically linked for now, but there are 335468651Skris preparations for dynamic linking in place. 335589837Skris This has been tested on Linux and Tru64. 335668651Skris [Richard Levitte] 335768651Skris 335868651Skris *) Randomness polling function for Win9x, as described in: 335968651Skris Peter Gutmann, Software Generation of Practically Strong 336068651Skris Random Numbers. 336168651Skris [Ulf M�ller] 336268651Skris 336368651Skris *) Fix so PRNG is seeded in req if using an already existing 336468651Skris DSA key. 336568651Skris [Steve Henson] 336668651Skris 336768651Skris *) New options to smime application. -inform and -outform 336868651Skris allow alternative formats for the S/MIME message including 336968651Skris PEM and DER. The -content option allows the content to be 337068651Skris specified separately. This should allow things like Netscape 337168651Skris form signing output easier to verify. 337268651Skris [Steve Henson] 337368651Skris 337468651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 337568651Skris [Steve Henson] 337668651Skris 337768651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 337868651Skris STRING types. These convert content octets to and from the 337968651Skris underlying type. The actual tag and length octets are 338068651Skris already assumed to have been read in and checked. These 338168651Skris are needed because all other string types have virtually 338268651Skris identical handling apart from the tag. By having versions 338368651Skris of the ASN1 functions that just operate on content octets 338468651Skris IMPLICIT tagging can be handled properly. It also allows 338568651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 338668651Skris and ASN1_INTEGER are identical apart from the tag. 338768651Skris [Steve Henson] 338868651Skris 338968651Skris *) Change the handling of OID objects as follows: 339068651Skris 339168651Skris - New object identifiers are inserted in objects.txt, following 339268651Skris the syntax given in objects.README. 339368651Skris - objects.pl is used to process obj_mac.num and create a new 339468651Skris obj_mac.h. 339568651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 339668651Skris obj_mac.h. 339768651Skris 339868651Skris This is currently kind of a hack, and the perl code in objects.pl 339968651Skris isn't very elegant, but it works as I intended. The simplest way 340068651Skris to check that it worked correctly is to look in obj_dat.h and 340168651Skris check the array nid_objs and make sure the objects haven't moved 340268651Skris around (this is important!). Additions are OK, as well as 340368651Skris consistent name changes. 340468651Skris [Richard Levitte] 340568651Skris 340668651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 340768651Skris [Bodo Moeller] 340868651Skris 340968651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 341068651Skris The given file adds to whatever has already been seeded into the 341168651Skris random pool through the RANDFILE configuration file option or 341268651Skris environment variable, or the default random state file. 341368651Skris [Richard Levitte] 341468651Skris 341568651Skris *) mkstack.pl now sorts each macro group into lexical order. 341668651Skris Previously the output order depended on the order the files 341768651Skris appeared in the directory, resulting in needless rewriting 341868651Skris of safestack.h . 341968651Skris [Steve Henson] 342068651Skris 342168651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 342268651Skris work arounds for the VC++ problem that it treats func() as 342368651Skris func(void). Also stripped out the parts of mkdef.pl that 342468651Skris added extra typesafe functions: these no longer exist. 342568651Skris [Steve Henson] 342668651Skris 342768651Skris *) Reorganisation of the stack code. The macros are now all 342868651Skris collected in safestack.h . Each macro is defined in terms of 342968651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 343068651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 343168651Skris this has the advantage of retaining type safety without the 343268651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 343368651Skris then the non typesafe macros are used instead. Also modified the 343468651Skris mkstack.pl script to handle the new form. Needs testing to see 343568651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 343668651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 343768651Skris and PKCS12_STACK_OF. 343868651Skris [Steve Henson] 343968651Skris 344068651Skris *) When some versions of IIS use the 'NET' form of private key the 344168651Skris key derivation algorithm is different. Normally MD5(password) is 344268651Skris used as a 128 bit RC4 key. In the modified case 344368651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 344468651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 344568651Skris as the old Netscape_RSA functions except they have an additional 344668651Skris 'sgckey' parameter which uses the modified algorithm. Also added 344768651Skris an -sgckey command line option to the rsa utility. Thanks to 344868651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 344968651Skris algorithm to openssl-dev. 345068651Skris [Steve Henson] 345168651Skris 345268651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 345368651Skris invalid expansion on some systems (SCO 5.0.5 for example). 345468651Skris Corrected to 'c.kname'. 345568651Skris [Phillip Porch <root@theporch.com>] 345668651Skris 345768651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 345868651Skris a STACK of email addresses from a certificate or request, these look 345968651Skris in the subject name and the subject alternative name extensions and 346068651Skris omit any duplicate addresses. 346168651Skris [Steve Henson] 346268651Skris 346368651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 346468651Skris This makes DSA verification about 2 % faster. 346568651Skris [Bodo Moeller] 346668651Skris 346768651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 346868651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 346968651Skris plus overhead for 1024 bit moduli). 347068651Skris This makes exponentiations about 0.5 % faster for 1024 bit 347168651Skris exponents (as measured by "openssl speed rsa2048"). 347268651Skris [Bodo Moeller] 347368651Skris 347468651Skris *) Rename memory handling macros to avoid conflicts with other 347568651Skris software: 347668651Skris Malloc => OPENSSL_malloc 347768651Skris Malloc_locked => OPENSSL_malloc_locked 347868651Skris Realloc => OPENSSL_realloc 347968651Skris Free => OPENSSL_free 348068651Skris [Richard Levitte] 348168651Skris 348268651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 348368651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 348468651Skris [Bodo Moeller] 348568651Skris 348668651Skris *) CygWin32 support. 348768651Skris [John Jarvie <jjarvie@newsguy.com>] 348868651Skris 348968651Skris *) The type-safe stack code has been rejigged. It is now only compiled 349068651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 349168651Skris by default all type-specific stack functions are "#define"d back to 349268651Skris standard stack functions. This results in more streamlined output 349368651Skris but retains the type-safety checking possibilities of the original 349468651Skris approach. 349568651Skris [Geoff Thorpe] 349668651Skris 349768651Skris *) The STACK code has been cleaned up, and certain type declarations 349868651Skris that didn't make a lot of sense have been brought in line. This has 349968651Skris also involved a cleanup of sorts in safestack.h to more correctly 350068651Skris map type-safe stack functions onto their plain stack counterparts. 350168651Skris This work has also resulted in a variety of "const"ifications of 350268651Skris lots of the code, especially "_cmp" operations which should normally 350368651Skris be prototyped with "const" parameters anyway. 350468651Skris [Geoff Thorpe] 350568651Skris 350668651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 350768651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 350868651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 350968651Skris where all of 'md' is used each time the PRNG is used, but 'state' 351068651Skris is used only indexed by a cyclic counter. As entropy may not be 351168651Skris well distributed from the beginning, 'md' is important as a 351268651Skris chaining variable. However, the output function chains only half 351368651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 351468651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 351568651Skris in all of 'state' being rewritten, with the new values depending 351668651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 351768651Skris [Bodo Moeller] 351868651Skris 351968651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 352068651Skris the handshake is continued after ssl_verify_cert_chain(); 352168651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 352268651Skris can lead to 'unexplainable' connection aborts later. 352368651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 352468651Skris 352568651Skris *) Major EVP API cipher revision. 352668651Skris Add hooks for extra EVP features. This allows various cipher 352768651Skris parameters to be set in the EVP interface. Support added for variable 352868651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 352968651Skris setting of RC2 and RC5 parameters. 353068651Skris 353168651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 353268651Skris ciphers. 353368651Skris 353468651Skris Remove lots of duplicated code from the EVP library. For example *every* 353568651Skris cipher init() function handles the 'iv' in the same way according to the 353668651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 353768651Skris for CFB and OFB modes they zero ctx->num. 353868651Skris 353968651Skris New functionality allows removal of S/MIME code RC2 hack. 354068651Skris 354168651Skris Most of the routines have the same form and so can be declared in terms 354268651Skris of macros. 354368651Skris 354468651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 354568651Skris all individual ciphers. If the cipher wants to handle IVs or keys 354668651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 354768651Skris flags. 354868651Skris 354968651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 355068651Skris value: although software versions of the algorithms cannot fail 355168651Skris any installed hardware versions can. 355268651Skris [Steve Henson] 355368651Skris 355468651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 355568651Skris this option is set, tolerate broken clients that send the negotiated 355668651Skris protocol version number instead of the requested protocol version 355768651Skris number. 355868651Skris [Bodo Moeller] 355968651Skris 356068651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 356168651Skris i.e. non-zero for export ciphersuites, zero otherwise. 356268651Skris Previous versions had this flag inverted, inconsistent with 356368651Skris rsa_tmp_cb (..._TMP_RSA_CB). 356468651Skris [Bodo Moeller; problem reported by Amit Chopra] 356568651Skris 356668651Skris *) Add missing DSA library text string. Work around for some IIS 356768651Skris key files with invalid SEQUENCE encoding. 356868651Skris [Steve Henson] 356968651Skris 357068651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 357168651Skris and so on that are implemented in OpenSSL. 357268651Skris [Richard Levitte] 357368651Skris 357468651Skris *) Enhance c_rehash script. Old version would mishandle certificates 357568651Skris with the same subject name hash and wouldn't handle CRLs at all. 357668651Skris Added -fingerprint option to crl utility, to support new c_rehash 357768651Skris features. 357868651Skris [Steve Henson] 357968651Skris 358068651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 358168651Skris [Ulf M�ller] 358268651Skris 358368651Skris *) Fix for SSL server purpose checking. Server checking was 358468651Skris rejecting certificates which had extended key usage present 358568651Skris but no ssl client purpose. 358668651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 358768651Skris 358868651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 358968651Skris is a little unclear about how a blank password is handled. 359068651Skris Since the password in encoded as a BMPString with terminating 359168651Skris double NULL a zero length password would end up as just the 359268651Skris double NULL. However no password at all is different and is 359368651Skris handled differently in the PKCS#12 key generation code. NS 359468651Skris treats a blank password as zero length. MSIE treats it as no 359568651Skris password on export: but it will try both on import. We now do 359668651Skris the same: PKCS12_parse() tries zero length and no password if 359768651Skris the password is set to "" or NULL (NULL is now a valid password: 359868651Skris it wasn't before) as does the pkcs12 application. 359968651Skris [Steve Henson] 360068651Skris 360168651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 360268651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 360368651Skris be obtained from the error queue. 360468651Skris [Bodo Moeller] 360568651Skris 360668651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 360768651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 360868651Skris accordingly to avoid race conditions (this is necessary because 360968651Skris thread_hash is no longer constant once set). 361068651Skris [Bodo Moeller] 361168651Skris 361268651Skris *) Bugfix for linux-elf makefile.one. 361368651Skris [Ulf M�ller] 361468651Skris 361568651Skris *) RSA_get_default_method() will now cause a default 361668651Skris RSA_METHOD to be chosen if one doesn't exist already. 361768651Skris Previously this was only set during a call to RSA_new() 361868651Skris or RSA_new_method(NULL) meaning it was possible for 361968651Skris RSA_get_default_method() to return NULL. 362068651Skris [Geoff Thorpe] 362168651Skris 362268651Skris *) Added native name translation to the existing DSO code 362368651Skris that will convert (if the flag to do so is set) filenames 362468651Skris that are sufficiently small and have no path information 362568651Skris into a canonical native form. Eg. "blah" converted to 362668651Skris "libblah.so" or "blah.dll" etc. 362768651Skris [Geoff Thorpe] 362868651Skris 362968651Skris *) New function ERR_error_string_n(e, buf, len) which is like 363068651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 363168651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 363268651Skris may not be NULL. 363368651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 363468651Skris 363568651Skris *) CONF library reworked to become more general. A new CONF 363668651Skris configuration file reader "class" is implemented as well as a 363768651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 363868651Skris old CONF_* functions are still there, but are reimplemented to 363968651Skris work in terms of the new functions. Also, a set of functions 364068651Skris to handle the internal storage of the configuration data is 364168651Skris provided to make it easier to write new configuration file 364268651Skris reader "classes" (I can definitely see something reading a 364368651Skris configuration file in XML format, for example), called _CONF_*, 364468651Skris or "the configuration storage API"... 364568651Skris 364668651Skris The new configuration file reading functions are: 364768651Skris 364868651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 364968651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 365068651Skris 365168651Skris NCONF_default, NCONF_WIN32 365268651Skris 365368651Skris NCONF_dump_fp, NCONF_dump_bio 365468651Skris 365568651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 365668651Skris NCONF_new creates a new CONF object. This works in the same way 365768651Skris as other interfaces in OpenSSL, like the BIO interface. 365868651Skris NCONF_dump_* dump the internal storage of the configuration file, 365968651Skris which is useful for debugging. All other functions take the same 366068651Skris arguments as the old CONF_* functions wth the exception of the 366168651Skris first that must be a `CONF *' instead of a `LHASH *'. 366268651Skris 366368651Skris To make it easer to use the new classes with the old CONF_* functions, 366468651Skris the function CONF_set_default_method is provided. 366568651Skris [Richard Levitte] 366668651Skris 366768651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 366868651Skris mentioned in the documentation but had not been implemented. 366968651Skris (This option is not yet really useful because even the additional 367068651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 367168651Skris [Bodo Moeller] 367268651Skris 367368651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 367468651Skris OpenSSL-based applications) load shared libraries and bind to 367568651Skris them in a portable way. 367668651Skris [Geoff Thorpe, with contributions from Richard Levitte] 367768651Skris 367859191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 367959191Skris 368059191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 368159191Skris 368259191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 368359191Skris (the default implementation of RAND_status). 368459191Skris 368559191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 368659191Skris to '-clrext' (= clear extensions), as intended and documented. 368759191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 368859191Skris <attili@amaxo.com>] 368959191Skris 369059191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 369159191Skris was larger than the MD block size. 369259191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 369359191Skris 369459191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 369559191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 369659191Skris using the passed key: if the passed key was a private key the result 369759191Skris of X509_print(), for example, would be to print out all the private key 369859191Skris components. 369959191Skris [Steve Henson] 370059191Skris 370159191Skris *) des_quad_cksum() byte order bug fix. 370259191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 370359191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 370459191Skris 370559191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 370659191Skris discouraged. 370759191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 370859191Skris 370959191Skris *) For easily testing in shell scripts whether some command 371059191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 371159191Skris returns with exit code 0 iff no command of the given name is available. 371259191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 371359191Skris the output goes to stdout and nothing is printed to stderr. 371459191Skris Additional arguments are always ignored. 371559191Skris 371659191Skris Since for each cipher there is a command of the same name, 371759191Skris the 'no-cipher' compilation switches can be tested this way. 371859191Skris 371959191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 372059191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 372159191Skris [Bodo Moeller] 372259191Skris 372359191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 372459191Skris [Bodo Moeller] 372559191Skris 372659191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 372759191Skris is set; it will be thrown away anyway because each handshake creates 372859191Skris its own key. 372959191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 373059191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 373159191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 373259191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 373359191Skris [Bodo Moeller] 373459191Skris 373559191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 373659191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 373759191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 373859191Skris does not suppress any output. 373959191Skris [Richard Levitte] 374059191Skris 374159191Skris *) Add compatibility options to the purpose and trust code. The 374259191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 374359191Skris accepts a certificate or CA, this was the previous behaviour, 374459191Skris with all the associated security issues. 374559191Skris 374659191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 374759191Skris automatically trust self signed roots in certificate store. A 374859191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 374959191Skris a purpose has no associated trust setting and it should instead 375059191Skris use the value in the default purpose. 375159191Skris [Steve Henson] 375259191Skris 375359191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 375459191Skris and fix a memory leak. 375559191Skris [Steve Henson] 375659191Skris 375759191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 375859191Skris reason strings from the previous version of the .c file, as 375959191Skris the default to have only downcase letters (and digits) in 376059191Skris automatically generated reasons codes is not always appropriate. 376159191Skris [Bodo Moeller] 376259191Skris 376359191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 376459191Skris using strerror. Previously, ERR_reason_error_string() returned 376559191Skris library names as reason strings for SYSerr; but SYSerr is a special 376659191Skris case where small numbers are errno values, not library numbers. 376759191Skris [Bodo Moeller] 376859191Skris 376959191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 377059191Skris converts DSA parameters into DH parameters. (When creating parameters, 377159191Skris DSA_generate_parameters is used.) 377259191Skris [Bodo Moeller] 377359191Skris 377459191Skris *) Include 'length' (recommended exponent length) in C code generated 377559191Skris by 'openssl dhparam -C'. 377659191Skris [Bodo Moeller] 377759191Skris 377859191Skris *) The second argument to set_label in perlasm was already being used 377959191Skris so couldn't be used as a "file scope" flag. Moved to third argument 378059191Skris which was free. 378159191Skris [Steve Henson] 378259191Skris 378359191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 378459191Skris instead of RAND_bytes for encryption IVs and salts. 378559191Skris [Bodo Moeller] 378659191Skris 378759191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 378859191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 378959191Skris RAND_set_rand_method would be impossible. 379059191Skris [Bodo Moeller] 379159191Skris 379259191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 379359191Skris number generation fails. 379459191Skris [Bodo Moeller] 379559191Skris 379659191Skris *) New 'rand' application for creating pseudo-random output. 379759191Skris [Bodo Moeller] 379859191Skris 379959191Skris *) Added configuration support for Linux/IA64 380059191Skris [Rolf Haberrecker <rolf@suse.de>] 380159191Skris 380259191Skris *) Assembler module support for Mingw32. 380359191Skris [Ulf M�ller] 380459191Skris 380559191Skris *) Shared library support for HPUX (in shlib/). 380659191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 380759191Skris 380859191Skris *) Shared library support for Solaris gcc. 380959191Skris [Lutz Behnke <behnke@trustcenter.de>] 381059191Skris 381159191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 381259191Skris 381359191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 381459191Skris were added manually and by SMIME_crlf_copy(). 381559191Skris [Steve Henson] 381659191Skris 381759191Skris *) In bntest.c don't call BN_rand with zero bits argument. 381859191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 381959191Skris 382059191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 382159191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 382259191Skris [Ulf M�ller] 382359191Skris 382459191Skris *) Add an optional second argument to the set_label() in the perl 382559191Skris assembly language builder. If this argument exists and is set 382659191Skris to 1 it signals that the assembler should use a symbol whose 382759191Skris scope is the entire file, not just the current function. This 382859191Skris is needed with MASM which uses the format label:: for this scope. 382959191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 383059191Skris 383159191Skris *) Change the ASN1 types so they are typedefs by default. Before 383259191Skris almost all types were #define'd to ASN1_STRING which was causing 383359191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 383459191Skris for example. 383559191Skris [Steve Henson] 383659191Skris 383759191Skris *) Change names of new functions to the new get1/get0 naming 383859191Skris convention: After 'get1', the caller owns a reference count 383959191Skris and has to call ..._free; 'get0' returns a pointer to some 384059191Skris data structure without incrementing reference counters. 384159191Skris (Some of the existing 'get' functions increment a reference 384259191Skris counter, some don't.) 384359191Skris Similarly, 'set1' and 'add1' functions increase reference 384459191Skris counters or duplicate objects. 384559191Skris [Steve Henson] 384659191Skris 384759191Skris *) Allow for the possibility of temp RSA key generation failure: 384859191Skris the code used to assume it always worked and crashed on failure. 384959191Skris [Steve Henson] 385059191Skris 385159191Skris *) Fix potential buffer overrun problem in BIO_printf(). 385259191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 385359191Skris pointed out by David Sacerdote <das33@cornell.edu>] 385459191Skris 385559191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 385659191Skris RAND_egd() and RAND_status(). In the command line application, 385759191Skris the EGD socket can be specified like a seed file using RANDFILE 385859191Skris or -rand. 385959191Skris [Ulf M�ller] 386059191Skris 386159191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 386259191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 386359191Skris [Steve Henson] 386459191Skris 386559191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 386659191Skris list to exclude them. This means that no special compilation option 386759191Skris is needed to use anonymous DH: it just needs to be included in the 386859191Skris cipher list. 386959191Skris [Steve Henson] 387059191Skris 387159191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 387259191Skris EVP_MD_type. The old functionality is available in a new macro called 387359191Skris EVP_MD_md(). Change code that uses it and update docs. 387459191Skris [Steve Henson] 387559191Skris 387659191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 387759191Skris where the 'void *' argument is replaced by a function pointer argument. 387859191Skris Previously 'void *' was abused to point to functions, which works on 387959191Skris many platforms, but is not correct. As these functions are usually 388059191Skris called by macros defined in OpenSSL header files, most source code 388159191Skris should work without changes. 388259191Skris [Richard Levitte] 388359191Skris 388459191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 388559191Skris sections with information on -D... compiler switches used for 388659191Skris compiling the library so that applications can see them. To enable 388759191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 388859191Skris must be defined. E.g., 388959191Skris #define OPENSSL_ALGORITHM_DEFINES 389059191Skris #include <openssl/opensslconf.h> 389159191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 389259191Skris [Richard Levitte, Ulf and Bodo M�ller] 389359191Skris 389459191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 389559191Skris record layer. 389659191Skris [Bodo Moeller] 389759191Skris 389859191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 389959191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 390059191Skris the required ASN1 format: arbitrary types determined by an OID. 390159191Skris [Steve Henson] 390259191Skris 390359191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 390459191Skris argument to 'req'. This is not because the function is newer or 390559191Skris better than others it just uses the work 'NEW' in the certificate 390659191Skris request header lines. Some software needs this. 390759191Skris [Steve Henson] 390859191Skris 390959191Skris *) Reorganise password command line arguments: now passwords can be 391059191Skris obtained from various sources. Delete the PEM_cb function and make 391159191Skris it the default behaviour: i.e. if the callback is NULL and the 391259191Skris usrdata argument is not NULL interpret it as a null terminated pass 391359191Skris phrase. If usrdata and the callback are NULL then the pass phrase 391459191Skris is prompted for as usual. 391559191Skris [Steve Henson] 391659191Skris 391759191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 391859191Skris the support is automatically enabled. The resulting binaries will 391959191Skris autodetect the card and use it if present. 392059191Skris [Ben Laurie and Compaq Inc.] 392159191Skris 392259191Skris *) Work around for Netscape hang bug. This sends certificate request 392359191Skris and server done in one record. Since this is perfectly legal in the 392459191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 392559191Skris the bugs/SSLv3 entry for more info. 392659191Skris [Steve Henson] 392759191Skris 392859191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 392959191Skris [Andy Polyakov] 393059191Skris 393159191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 393259191Skris of seed file. 393359191Skris [Steve Henson] 393459191Skris 393559191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 393659191Skris [Bodo Moeller] 393759191Skris 393859191Skris *) Add command line password options to the remaining applications. 393959191Skris [Steve Henson] 394059191Skris 394159191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 394259191Skris bits. 394359191Skris [Ulf M�ller] 394459191Skris 394559191Skris *) More tests in bntest.c, and changed test_bn output. 394659191Skris [Ulf M�ller] 394759191Skris 394859191Skris *) ./config recognizes MacOS X now. 394959191Skris [Andy Polyakov] 395059191Skris 395159191Skris *) Bug fix for BN_div() when the first words of num and divsor are 395259191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 395359191Skris [Ulf M�ller] 395459191Skris 395559191Skris *) Add support for various broken PKCS#8 formats, and command line 395659191Skris options to produce them. 395759191Skris [Steve Henson] 395859191Skris 395959191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 396059191Skris get temporary BIGNUMs from a BN_CTX. 396159191Skris [Ulf M�ller] 396259191Skris 396359191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 396459191Skris for p == 0. 396559191Skris [Ulf M�ller] 396659191Skris 396759191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 396859191Skris include a #define from the old name to the new. The original intent 396959191Skris was that statically linked binaries could for example just call 397059191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 397159191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 397259191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 397359191Skris one would link with the other. They are now in separate source files. 397459191Skris [Steve Henson] 397559191Skris 397659191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 397759191Skris [Steve Henson] 397859191Skris 397959191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 398059191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 398159191Skris loop, our standard modexp algorithms are faster). 398259191Skris [Bodo Moeller] 398359191Skris 398459191Skris *) Support for the EBCDIC character set completed. 398559191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 398659191Skris 398759191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 398859191Skris use void * instead of char * in lhash. 398959191Skris [Ulf M�ller] 399059191Skris 399159191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 399259191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 399359191Skris this the server could overwrite ephemeral keys that the client 399459191Skris has already seen). 399559191Skris [Bodo Moeller] 399659191Skris 399759191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 399859191Skris using 50 iterations of the Rabin-Miller test. 399959191Skris 400059191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 400159191Skris iterations of the Rabin-Miller test as required by the appendix 400259191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 400359191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 400459191Skris generation becomes much faster. 400559191Skris 400659191Skris This implies a change for the callback functions in DSA_is_prime 400759191Skris and DSA_generate_parameters: The callback function is called once 400859191Skris for each positive witness in the Rabin-Miller test, not just 400959191Skris occasionally in the inner loop; and the parameters to the 401059191Skris callback function now provide an iteration count for the outer 401159191Skris loop rather than for the current invocation of the inner loop. 401259191Skris DSA_generate_parameters additionally can call the callback 401359191Skris function with an 'iteration count' of -1, meaning that a 401459191Skris candidate has passed the trial division test (when q is generated 401559191Skris from an application-provided seed, trial division is skipped). 401659191Skris [Bodo Moeller] 401759191Skris 401859191Skris *) New function BN_is_prime_fasttest that optionally does trial 401959191Skris division before starting the Rabin-Miller test and has 402059191Skris an additional BN_CTX * argument (whereas BN_is_prime always 402159191Skris has to allocate at least one BN_CTX). 402259191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 402359191Skris trial division stage. 402459191Skris [Bodo Moeller] 402559191Skris 402659191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 402759191Skris as ASN1_TIME. 402859191Skris [Steve Henson] 402959191Skris 403059191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 403159191Skris [Steve Henson] 403259191Skris 403359191Skris *) New function BN_pseudo_rand(). 403459191Skris [Ulf M�ller] 403559191Skris 403659191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 403759191Skris bignum version of BN_from_montgomery() with the working code from 403859191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 403959191Skris the comments. 404059191Skris [Ulf M�ller] 404159191Skris 404259191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 404359191Skris made it impossible to use the same SSL_SESSION data structure in 404459191Skris SSL2 clients in multiple threads. 404559191Skris [Bodo Moeller] 404659191Skris 404759191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 404859191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 404959191Skris to seed the PRNG (previously an explicit byte count was required). 405059191Skris [Ulf M�ller, Bodo M�ller] 405159191Skris 405259191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 405359191Skris used (char *) instead of (void *) and had casts all over the place. 405459191Skris [Steve Henson] 405559191Skris 405659191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 405759191Skris [Ulf M�ller] 405859191Skris 405959191Skris *) Retain source code compatibility for BN_prime_checks macro: 406059191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 406159191Skris BN_prime_checks_for_size to determine the appropriate number of 406259191Skris Rabin-Miller iterations. 406359191Skris [Ulf M�ller] 406459191Skris 406559191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 406659191Skris DH_CHECK_P_NOT_SAFE_PRIME. 406759191Skris (Check if this is true? OpenPGP calls them "strong".) 406859191Skris [Ulf M�ller] 406959191Skris 407059191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 407159191Skris "dhparam". The old programs are retained for now but will handle DH keys 407259191Skris (instead of parameters) in future. 407359191Skris [Steve Henson] 407459191Skris 407559191Skris *) Make the ciphers, s_server and s_client programs check the return values 407659191Skris when a new cipher list is set. 407759191Skris [Steve Henson] 407859191Skris 407959191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 408059191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 408159191Skris wrong. 408259191Skris 408359191Skris The syntax for the cipher sorting has been extended to support sorting by 408459191Skris cipher-strength (using the strength_bits hard coded in the tables). 408559191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 408659191Skris 408759191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 408859191Skris string with an "undefined" symbol (neither command nor alphanumeric 408959191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 409059191Skris an error is flagged. 409159191Skris 409259191Skris Due to the strength-sorting extension, the code of the 409359191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 409459191Skris the readability was also increased :-) 409559191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 409659191Skris 409759191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 409859191Skris for the first serial number and places 2 in the serial number file. This 409959191Skris avoids problems when the root CA is created with serial number zero and 410059191Skris the first user certificate has the same issuer name and serial number 410159191Skris as the root CA. 410259191Skris [Steve Henson] 410359191Skris 410459191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 410559191Skris the new code. Add documentation for this stuff. 410659191Skris [Steve Henson] 410759191Skris 410859191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 410959191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 411059191Skris structures and behave in an analagous way to the X509v3 functions: 411159191Skris they shouldn't be called directly but wrapper functions should be used 411259191Skris instead. 411359191Skris 411459191Skris So we also now have some wrapper functions that call the X509at functions 411559191Skris when passed certificate requests. (TO DO: similar things can be done with 411659191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 411759191Skris things. Some of these need some d2i or i2d and print functionality 411859191Skris because they handle more complex structures.) 411959191Skris [Steve Henson] 412059191Skris 412159191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 412259191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 412359191Skris NO_RSA in ssl/s2*.c. 412459191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 412559191Skris 412659191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 412759191Skris has a return value which indicates the quality of the random data 412859191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 412959191Skris error queue. New function RAND_pseudo_bytes() generates output that is 413059191Skris guaranteed to be unique but not unpredictable. RAND_add is like 413159191Skris RAND_seed, but takes an extra argument for an entropy estimate 413259191Skris (RAND_seed always assumes full entropy). 413359191Skris [Ulf M�ller] 413459191Skris 413559191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 413659191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 413759191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 413859191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 413959191Skris false-positive rate of at most 2^-80 for random input. 414059191Skris [Bodo Moeller] 414159191Skris 414259191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 414359191Skris [Bodo Moeller] 414459191Skris 414559191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 414659191Skris in the 0.9.5 release), this returns the chain 414759191Skris from an X509_CTX structure with a dup of the stack and all 414859191Skris the X509 reference counts upped: so the stack will exist 414959191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 415059191Skris to use this. 415159191Skris 415259191Skris Also make SSL_SESSION_print() print out the verify return 415359191Skris code. 415459191Skris [Steve Henson] 415559191Skris 415659191Skris *) Add manpage for the pkcs12 command. Also change the default 415759191Skris behaviour so MAC iteration counts are used unless the new 415859191Skris -nomaciter option is used. This improves file security and 415959191Skris only older versions of MSIE (4.0 for example) need it. 416059191Skris [Steve Henson] 416159191Skris 416259191Skris *) Honor the no-xxx Configure options when creating .DEF files. 416359191Skris [Ulf M�ller] 416459191Skris 416559191Skris *) Add PKCS#10 attributes to field table: challengePassword, 416659191Skris unstructuredName and unstructuredAddress. These are taken from 416759191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 416859191Skris international characters are used. 416959191Skris 417059191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 417159191Skris based on strings. Remove the 'loc' parameter when adding 417259191Skris attributes because these will be a SET OF encoding which is sorted 417359191Skris in ASN1 order. 417459191Skris [Steve Henson] 417559191Skris 417659191Skris *) Initial changes to the 'req' utility to allow request generation 417759191Skris automation. This will allow an application to just generate a template 417859191Skris file containing all the field values and have req construct the 417959191Skris request. 418059191Skris 418159191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 418259191Skris used all over the place including certificate requests and PKCS#7 418359191Skris structures. They are currently handled manually where necessary with 418459191Skris some primitive wrappers for PKCS#7. The new functions behave in a 418559191Skris manner analogous to the X509 extension functions: they allow 418659191Skris attributes to be looked up by NID and added. 418759191Skris 418859191Skris Later something similar to the X509V3 code would be desirable to 418959191Skris automatically handle the encoding, decoding and printing of the 419059191Skris more complex types. The string types like challengePassword can 419159191Skris be handled by the string table functions. 419259191Skris 419359191Skris Also modified the multi byte string table handling. Now there is 419459191Skris a 'global mask' which masks out certain types. The table itself 419559191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 419659191Skris is useful when for example there is only one permissible type 419759191Skris (as in countryName) and using the mask might result in no valid 419859191Skris types at all. 419959191Skris [Steve Henson] 420059191Skris 420159191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 420259191Skris SSL_get_peer_finished to allow applications to obtain the latest 420359191Skris Finished messages sent to the peer or expected from the peer, 420459191Skris respectively. (SSL_get_peer_finished is usually the Finished message 420559191Skris actually received from the peer, otherwise the protocol will be aborted.) 420659191Skris 420759191Skris As the Finished message are message digests of the complete handshake 420859191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 420959191Skris be used for external authentication procedures when the authentication 421059191Skris provided by SSL/TLS is not desired or is not enough. 421159191Skris [Bodo Moeller] 421259191Skris 421359191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 421459191Skris the host supports BWX extension and if Compaq C is present on the 421559191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 421659191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 421759191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 421859191Skris SHA1. 421959191Skris [Andy Polyakov] 422059191Skris 422159191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 422259191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 422359191Skris weak crypto and after checking the certificate is SGC a second one 422459191Skris with strong crypto. MS SGC stops the first handshake after receiving 422559191Skris the server certificate message and sends a second client hello. Since 422659191Skris a server will typically do all the time consuming operations before 422759191Skris expecting any further messages from the client (server key exchange 422859191Skris is the most expensive) there is little difference between the two. 422959191Skris 423059191Skris To get OpenSSL to support MS SGC we have to permit a second client 423159191Skris hello message after we have sent server done. In addition we have to 423259191Skris reset the MAC if we do get this second client hello. 423359191Skris [Steve Henson] 423459191Skris 423559191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 423659191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 423759191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 423859191Skris format DER encoded private key. Newer code should use PKCS#8 format which 423959191Skris has the key type encoded in the ASN1 structure. Added DER private key 424059191Skris support to pkcs8 application. 424159191Skris [Steve Henson] 424259191Skris 424359191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 424459191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 424559191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 424659191Skris is set, we interpret this as a request to violate the specification 424759191Skris (the worst that can happen is a handshake failure, and 'correct' 424859191Skris behaviour would result in a handshake failure anyway). 424959191Skris [Bodo Moeller] 425059191Skris 425159191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 425259191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 425359191Skris concurrently obtain them from an external cache). 425459191Skris The internal cache can handle only one SSL_SESSION with a given ID, 425559191Skris so if there's a conflict, we now throw out the old one to achieve 425659191Skris consistency. 425759191Skris [Bodo Moeller] 425859191Skris 425959191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 426059191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 426159191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 426259191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 426359191Skris example. 426459191Skris [Steve Henson] 426559191Skris 426659191Skris *) Simplify the trust setting structure and code. Now we just have 426759191Skris two sequences of OIDs for trusted and rejected settings. These will 426859191Skris typically have values the same as the extended key usage extension 426959191Skris and any application specific purposes. 427059191Skris 427159191Skris The trust checking code now has a default behaviour: it will just 427259191Skris check for an object with the same NID as the passed id. Functions can 427359191Skris be provided to override either the default behaviour or the behaviour 427459191Skris for a given id. SSL client, server and email already have functions 427559191Skris in place for compatibility: they check the NID and also return "trusted" 427659191Skris if the certificate is self signed. 427759191Skris [Steve Henson] 427859191Skris 427959191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 428059191Skris traditional format into an EVP_PKEY structure. 428159191Skris [Steve Henson] 428259191Skris 428359191Skris *) Add a password callback function PEM_cb() which either prompts for 428459191Skris a password if usr_data is NULL or otherwise assumes it is a null 428559191Skris terminated password. Allow passwords to be passed on command line 428659191Skris environment or config files in a few more utilities. 428759191Skris [Steve Henson] 428859191Skris 428959191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 429059191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 429159191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 429259191Skris Update documentation. 429359191Skris [Steve Henson] 429459191Skris 429559191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 429659191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 429759191Skris and produce an error if it couldn't. For compatibility we also have 429859191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 429959191Skris don't allocate anything because they don't need to. 430059191Skris [Steve Henson] 430159191Skris 430259191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 430359191Skris for details. 430459191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 430559191Skris 430659191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 430759191Skris possibly others as well. The purpose is to make an interface that 430859191Skris provide hooks so anyone can build a separate set of allocation and 430959191Skris deallocation routines to be used by OpenSSL, for example memory 431059191Skris pool implementations, or something else, which was previously hard 431159191Skris since Malloc(), Realloc() and Free() were defined as macros having 431259191Skris the values malloc, realloc and free, respectively (except for Win32 431359191Skris compilations). The same is provided for memory debugging code. 431459191Skris OpenSSL already comes with functionality to find memory leaks, but 431559191Skris this gives people a chance to debug other memory problems. 431659191Skris 431759191Skris With these changes, a new set of functions and macros have appeared: 431859191Skris 431968651Skris CRYPTO_set_mem_debug_functions() [F] 432068651Skris CRYPTO_get_mem_debug_functions() [F] 432168651Skris CRYPTO_dbg_set_options() [F] 432268651Skris CRYPTO_dbg_get_options() [F] 432368651Skris CRYPTO_malloc_debug_init() [M] 432459191Skris 432559191Skris The memory debug functions are NULL by default, unless the library 432659191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 432759191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 432859191Skris gives the standard debugging functions that come with OpenSSL) or 432959191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 433059191Skris provided by the library user) must be used. When the standard 433159191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 433259191Skris request additional information: 433359191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 433459191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 433559191Skris 433659191Skris Also, things like CRYPTO_set_mem_functions will always give the 433759191Skris expected result (the new set of functions is used for allocation 433859191Skris and deallocation) at all times, regardless of platform and compiler 433959191Skris options. 434059191Skris 434159191Skris To finish it up, some functions that were never use in any other 434259191Skris way than through macros have a new API and new semantic: 434359191Skris 434459191Skris CRYPTO_dbg_malloc() 434559191Skris CRYPTO_dbg_realloc() 434659191Skris CRYPTO_dbg_free() 434759191Skris 434859191Skris All macros of value have retained their old syntax. 434959191Skris [Richard Levitte and Bodo Moeller] 435059191Skris 435159191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 435259191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 435359191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 435459191Skris algorithm. 435559191Skris [Steve Henson] 435659191Skris 435759191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 435859191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 435959191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 436059191Skris 436159191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 436259191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 436359191Skris functionality to handle multipart/signed properly) and a utility 436459191Skris called 'smime' to call all this stuff. This is based on code I 436559191Skris originally wrote for Celo who have kindly allowed it to be 436659191Skris included in OpenSSL. 436759191Skris [Steve Henson] 436859191Skris 436959191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 437059191Skris des_set_key (aka des_key_sched). Global variable des_check_key 437159191Skris decides which of these is called by des_set_key; this way 437259191Skris des_check_key behaves as it always did, but applications and 437359191Skris the library itself, which was buggy for des_check_key == 1, 437459191Skris have a cleaner way to pick the version they need. 437559191Skris [Bodo Moeller] 437659191Skris 437759191Skris *) New function PKCS12_newpass() which changes the password of a 437859191Skris PKCS12 structure. 437959191Skris [Steve Henson] 438059191Skris 438159191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 438259191Skris dynamic mix. In both cases the ids can be used as an index into the 438359191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 438459191Skris functions so they accept a list of the field values and the 438559191Skris application doesn't need to directly manipulate the X509_TRUST 438659191Skris structure. 438759191Skris [Steve Henson] 438859191Skris 438959191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 439059191Skris need initialising. 439159191Skris [Steve Henson] 439259191Skris 439359191Skris *) Modify the way the V3 extension code looks up extensions. This now 439459191Skris works in a similar way to the object code: we have some "standard" 439559191Skris extensions in a static table which is searched with OBJ_bsearch() 439659191Skris and the application can add dynamic ones if needed. The file 439759191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 439859191Skris updated whenever a new extension is added to the core code and kept 439959191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 440059191Skris this. New extensions are not added too often so this file can readily 440159191Skris be maintained manually. 440259191Skris 440359191Skris There are two big advantages in doing things this way. The extensions 440459191Skris can be looked up immediately and no longer need to be "added" using 440559191Skris X509V3_add_standard_extensions(): this function now does nothing. 440659191Skris [Side note: I get *lots* of email saying the extension code doesn't 440759191Skris work because people forget to call this function] 440859191Skris Also no dynamic allocation is done unless new extensions are added: 440959191Skris so if we don't add custom extensions there is no need to call 441059191Skris X509V3_EXT_cleanup(). 441159191Skris [Steve Henson] 441259191Skris 441359191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 441459191Skris magic header, so unsalted files fail gracefully instead of just decrypting 441559191Skris to garbage. This is because not salting is a big security hole, so people 441659191Skris should be discouraged from doing it. 441759191Skris [Ben Laurie] 441859191Skris 441959191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 442059191Skris digest to be passed on the command line but it only used this 442159191Skris parameter when signing a certificate. Modified so all relevant 442259191Skris operations are affected by the digest parameter including the 442359191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 442459191Skris DSA key was used because it didn't fix the digest. 442559191Skris [Steve Henson] 442659191Skris 442759191Skris *) Initial certificate chain verify code. Currently tests the untrusted 442859191Skris certificates for consistency with the verify purpose (which is set 442959191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 443059191Skris 443159191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 443259191Skris this is because it will reject chains with invalid extensions whereas 443359191Skris every previous version of OpenSSL and SSLeay made no checks at all. 443459191Skris 443559191Skris Trust code: checks the root CA for the relevant trust settings. Trust 443659191Skris settings have an initial value consistent with the verify purpose: e.g. 443759191Skris if the verify purpose is for SSL client use it expects the CA to be 443859191Skris trusted for SSL client use. However the default value can be changed to 443959191Skris permit custom trust settings: one example of this would be to only trust 444059191Skris certificates from a specific "secure" set of CAs. 444159191Skris 444259191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 444359191Skris which should be used for version portability: especially since the 444459191Skris verify structure is likely to change more often now. 444559191Skris 444659191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 444759191Skris to set them. If not set then assume SSL clients will verify SSL servers 444859191Skris and vice versa. 444959191Skris 445059191Skris Two new options to the verify program: -untrusted allows a set of 445159191Skris untrusted certificates to be passed in and -purpose which sets the 445259191Skris intended purpose of the certificate. If a purpose is set then the 445359191Skris new chain verify code is used to check extension consistency. 445459191Skris [Steve Henson] 445559191Skris 445659191Skris *) Support for the authority information access extension. 445759191Skris [Steve Henson] 445859191Skris 445959191Skris *) Modify RSA and DSA PEM read routines to transparently handle 446059191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 446159191Skris public keys in a format compatible with certificate 446259191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 446359191Skris functions called *_PublicKey_* which used various odd formats so 446459191Skris these are retained for compatibility: however the DSA variants were 446559191Skris never in a public release so they have been deleted. Changed dsa/rsa 446659191Skris utilities to handle the new format: note no releases ever handled public 446759191Skris keys so we should be OK. 446859191Skris 446959191Skris The primary motivation for this change is to avoid the same fiasco 447059191Skris that dogs private keys: there are several incompatible private key 447159191Skris formats some of which are standard and some OpenSSL specific and 447259191Skris require various evil hacks to allow partial transparent handling and 447359191Skris even then it doesn't work with DER formats. Given the option anything 447459191Skris other than PKCS#8 should be dumped: but the other formats have to 447559191Skris stay in the name of compatibility. 447659191Skris 447759191Skris With public keys and the benefit of hindsight one standard format 447859191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 447959191Skris it clearly returns an error if you try to read the wrong kind of key. 448059191Skris 448159191Skris Added a -pubkey option to the 'x509' utility to output the public key. 448259191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 448359191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 448459191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 448559191Skris that do the same as the EVP_PKEY_assign_*() except they up the 448659191Skris reference count of the added key (they don't "swallow" the 448759191Skris supplied key). 448859191Skris [Steve Henson] 448959191Skris 449059191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 449159191Skris CRLs would fail if the file contained no certificates or no CRLs: 449259191Skris added a new function to read in both types and return the number 449359191Skris read: this means that if none are read it will be an error. The 449459191Skris DER versions of the certificate and CRL reader would always fail 449559191Skris because it isn't possible to mix certificates and CRLs in DER format 449659191Skris without choking one or the other routine. Changed this to just read 449759191Skris a certificate: this is the best we can do. Also modified the code 449859191Skris in apps/verify.c to take notice of return codes: it was previously 449959191Skris attempting to read in certificates from NULL pointers and ignoring 450059191Skris any errors: this is one reason why the cert and CRL reader seemed 450159191Skris to work. It doesn't check return codes from the default certificate 450259191Skris routines: these may well fail if the certificates aren't installed. 450359191Skris [Steve Henson] 450459191Skris 450559191Skris *) Code to support otherName option in GeneralName. 450659191Skris [Steve Henson] 450759191Skris 450859191Skris *) First update to verify code. Change the verify utility 450959191Skris so it warns if it is passed a self signed certificate: 451059191Skris for consistency with the normal behaviour. X509_verify 451159191Skris has been modified to it will now verify a self signed 451259191Skris certificate if *exactly* the same certificate appears 451359191Skris in the store: it was previously impossible to trust a 451459191Skris single self signed certificate. This means that: 451559191Skris openssl verify ss.pem 451659191Skris now gives a warning about a self signed certificate but 451759191Skris openssl verify -CAfile ss.pem ss.pem 451859191Skris is OK. 451959191Skris [Steve Henson] 452059191Skris 452159191Skris *) For servers, store verify_result in SSL_SESSION data structure 452259191Skris (and add it to external session representation). 452359191Skris This is needed when client certificate verifications fails, 452459191Skris but an application-provided verification callback (set by 452559191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 452659191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 452759191Skris but returns 1): When the session is reused, we have to set 452859191Skris ssl->verify_result to the appropriate error code to avoid 452959191Skris security holes. 453059191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 453159191Skris 453259191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 453359191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 453459191Skris didn't contain any existing data because it was being created. 453559191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 453659191Skris 453759191Skris *) Add a salt to the key derivation routines in enc.c. This 453859191Skris forms the first 8 bytes of the encrypted file. Also add a 453959191Skris -S option to allow a salt to be input on the command line. 454059191Skris [Steve Henson] 454159191Skris 454259191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 454359191Skris to compare two certificates. We do this by working out the SHA1 454459191Skris hash and comparing that. X509_cmp() will be needed by the trust 454559191Skris code. 454659191Skris [Steve Henson] 454759191Skris 454859191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 454959191Skris the reference count in the SSL_SESSION returned. 455059191Skris [Geoff Thorpe <geoff@eu.c2.net>] 455159191Skris 455259191Skris *) Fix for 'req': it was adding a null to request attributes. 455359191Skris Also change the X509_LOOKUP and X509_INFO code to handle 455459191Skris certificate auxiliary information. 455559191Skris [Steve Henson] 455659191Skris 455759191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 455859191Skris the 'enc' command. 455959191Skris [Steve Henson] 456059191Skris 456159191Skris *) Add the possibility to add extra information to the memory leak 456259191Skris detecting output, to form tracebacks, showing from where each 456359191Skris allocation was originated: CRYPTO_push_info("constant string") adds 456459191Skris the string plus current file name and line number to a per-thread 456559191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 456659191Skris is like calling CYRPTO_pop_info() until the stack is empty. 456759191Skris Also updated memory leak detection code to be multi-thread-safe. 456859191Skris [Richard Levitte] 456959191Skris 457059191Skris *) Add options -text and -noout to pkcs7 utility and delete the 457159191Skris encryption options which never did anything. Update docs. 457259191Skris [Steve Henson] 457359191Skris 457459191Skris *) Add options to some of the utilities to allow the pass phrase 457559191Skris to be included on either the command line (not recommended on 457659191Skris OSes like Unix) or read from the environment. Update the 457759191Skris manpages and fix a few bugs. 457859191Skris [Steve Henson] 457959191Skris 458059191Skris *) Add a few manpages for some of the openssl commands. 458159191Skris [Steve Henson] 458259191Skris 458359191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 458459191Skris leaking and not finding already revoked certificates. 458559191Skris [Steve Henson] 458659191Skris 458759191Skris *) Extensive changes to support certificate auxiliary information. 458859191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 458959191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 459059191Skris can still read in a certificate file in the usual way but it 459159191Skris will also read in any additional "auxiliary information". By 459259191Skris doing things this way a fair degree of compatibility can be 459359191Skris retained: existing certificates can have this information added 459459191Skris using the new 'x509' options. 459559191Skris 459659191Skris Current auxiliary information includes an "alias" and some trust 459759191Skris settings. The trust settings will ultimately be used in enhanced 459859191Skris certificate chain verification routines: currently a certificate 459959191Skris can only be trusted if it is self signed and then it is trusted 460059191Skris for all purposes. 460159191Skris [Steve Henson] 460259191Skris 460359191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 460459191Skris The problem was that one of the replacement routines had not been working 460559191Skris since SSLeay releases. For now the offending routine has been replaced 460659191Skris with non-optimised assembler. Even so, this now gives around 95% 460759191Skris performance improvement for 1024 bit RSA signs. 460859191Skris [Mark Cox] 460959191Skris 461059191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 461159191Skris handling. Most clients have the effective key size in bits equal to 461259191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 461359191Skris A few however don't do this and instead use the size of the decrypted key 461459191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 461559191Skris the effective key length. In this case the effective key length can still 461659191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 461759191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 461859191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 461959191Skris the key length and effective key length are equal. 462059191Skris [Steve Henson] 462159191Skris 462259191Skris *) Add a bunch of functions that should simplify the creation of 462359191Skris X509_NAME structures. Now you should be able to do: 462459191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 462559191Skris and have it automatically work out the correct field type and fill in 462659191Skris the structures. The more adventurous can try: 462759191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 462859191Skris and it will (hopefully) work out the correct multibyte encoding. 462959191Skris [Steve Henson] 463059191Skris 463159191Skris *) Change the 'req' utility to use the new field handling and multibyte 463259191Skris copy routines. Before the DN field creation was handled in an ad hoc 463359191Skris way in req, ca, and x509 which was rather broken and didn't support 463459191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 463559191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 463659191Skris using the dirstring_type option. See the new comment in the default 463759191Skris openssl.cnf for more info. 463859191Skris [Steve Henson] 463959191Skris 464059191Skris *) Make crypto/rand/md_rand.c more robust: 464159191Skris - Assure unique random numbers after fork(). 464259191Skris - Make sure that concurrent threads access the global counter and 464359191Skris md serializably so that we never lose entropy in them 464459191Skris or use exactly the same state in multiple threads. 464559191Skris Access to the large state is not always serializable because 464659191Skris the additional locking could be a performance killer, and 464759191Skris md should be large enough anyway. 464859191Skris [Bodo Moeller] 464959191Skris 465059191Skris *) New file apps/app_rand.c with commonly needed functionality 465159191Skris for handling the random seed file. 465259191Skris 465359191Skris Use the random seed file in some applications that previously did not: 465459191Skris ca, 465559191Skris dsaparam -genkey (which also ignored its '-rand' option), 465659191Skris s_client, 465759191Skris s_server, 465859191Skris x509 (when signing). 465959191Skris Except on systems with /dev/urandom, it is crucial to have a random 466059191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 466159191Skris for RSA signatures we could do without one. 466259191Skris 466359191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 466459191Skris of each file listed in the '-rand' option. The function as previously 466559191Skris found in genrsa is now in app_rand.c and is used by all programs 466659191Skris that support '-rand'. 466759191Skris [Bodo Moeller] 466859191Skris 466959191Skris *) In RAND_write_file, use mode 0600 for creating files; 467059191Skris don't just chmod when it may be too late. 467159191Skris [Bodo Moeller] 467259191Skris 467359191Skris *) Report an error from X509_STORE_load_locations 467459191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 467559191Skris [Bill Perry] 467659191Skris 467759191Skris *) New function ASN1_mbstring_copy() this copies a string in either 467859191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 467959191Skris into an ASN1_STRING type. A mask of permissible types is passed 468059191Skris and it chooses the "minimal" type to use or an error if not type 468159191Skris is suitable. 468259191Skris [Steve Henson] 468359191Skris 468459191Skris *) Add function equivalents to the various macros in asn1.h. The old 468559191Skris macros are retained with an M_ prefix. Code inside the library can 468659191Skris use the M_ macros. External code (including the openssl utility) 468759191Skris should *NOT* in order to be "shared library friendly". 468859191Skris [Steve Henson] 468959191Skris 469059191Skris *) Add various functions that can check a certificate's extensions 469159191Skris to see if it usable for various purposes such as SSL client, 469259191Skris server or S/MIME and CAs of these types. This is currently 469359191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 469459191Skris verification. Also added a -purpose flag to x509 utility to 469559191Skris print out all the purposes. 469659191Skris [Steve Henson] 469759191Skris 469859191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 469959191Skris functions. 470059191Skris [Steve Henson] 470159191Skris 470259191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 470359191Skris for, obtain and decode and extension and obtain its critical flag. 470459191Skris This allows all the necessary extension code to be handled in a 470559191Skris single function call. 470659191Skris [Steve Henson] 470759191Skris 470859191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 470959191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 471059191Skris [Andy Polyakov] 471159191Skris 471259191Skris *) New -noout option to asn1parse. This causes no output to be produced 471359191Skris its main use is when combined with -strparse and -out to extract data 471459191Skris from a file (which may not be in ASN.1 format). 471559191Skris [Steve Henson] 471659191Skris 471759191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 471859191Skris when producing the local key id. 471959191Skris [Richard Levitte <levitte@stacken.kth.se>] 472059191Skris 472159191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 472259191Skris stated explicitly. If it is not stated then it tries the first server 472359191Skris certificate file. The previous behaviour hard coded the filename 472459191Skris "server.pem". 472559191Skris [Steve Henson] 472659191Skris 472759191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 472859191Skris a public key to be input or output. For example: 472959191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 473059191Skris Also added necessary DSA public key functions to handle this. 473159191Skris [Steve Henson] 473259191Skris 473359191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 473459191Skris in the message. This was handled by allowing 473559191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 473659191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 473759191Skris 473859191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 473959191Skris to the end of the strings whereas this didn't. This would cause problems 474059191Skris if strings read with d2i_ASN1_bytes() were later modified. 474159191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 474259191Skris 474359191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 474459191Skris data and it contains EOF it will end up returning an error. This is 474559191Skris caused by input 46 bytes long. The cause is due to the way base64 474659191Skris BIOs find the start of base64 encoded data. They do this by trying a 474759191Skris trial decode on each line until they find one that works. When they 474859191Skris do a flag is set and it starts again knowing it can pass all the 474959191Skris data directly through the decoder. Unfortunately it doesn't reset 475059191Skris the context it uses. This means that if EOF is reached an attempt 475159191Skris is made to pass two EOFs through the context and this causes the 475259191Skris resulting error. This can also cause other problems as well. As is 475359191Skris usual with these problems it takes *ages* to find and the fix is 475459191Skris trivial: move one line. 475559191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 475659191Skris 475759191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 475859191Skris old code wouldn't work because it needed to select() on sockets and the 475959191Skris tty (for keypresses and to see if data could be written). Win32 only 476059191Skris supports select() on sockets so we select() with a 1s timeout on the 476159191Skris sockets and then see if any characters are waiting to be read, if none 476259191Skris are present then we retry, we also assume we can always write data to 476359191Skris the tty. This isn't nice because the code then blocks until we've 476459191Skris received a complete line of data and it is effectively polling the 476559191Skris keyboard at 1s intervals: however it's quite a bit better than not 476659191Skris working at all :-) A dedicated Windows application might handle this 476759191Skris with an event loop for example. 476859191Skris [Steve Henson] 476959191Skris 477059191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 477159191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 477259191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 477359191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 477459191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 477559191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 477659191Skris This necessitated the support of an extra signature type NID_md5_sha1 477759191Skris for SSL signatures and modifications to the SSL library to use it instead 477859191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 477959191Skris [Steve Henson] 478059191Skris 478159191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 478259191Skris will lookup a CRL issuers certificate and verify the signature in a 478359191Skris similar way to the verify program. Tidy up the crl program so it 478459191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 478559191Skris less strict. It will now permit CRL extensions even if it is not 478659191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 478759191Skris [Steve Henson] 478859191Skris 478959191Skris *) Initialize all non-automatic variables each time one of the openssl 479059191Skris sub-programs is started (this is necessary as they may be started 479159191Skris multiple times from the "OpenSSL>" prompt). 479259191Skris [Lennart Bang, Bodo Moeller] 479359191Skris 479459191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 479559191Skris removing all other RSA functionality (this is what NO_RSA does). This 479659191Skris is so (for example) those in the US can disable those operations covered 479759191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 479859191Skris key generation. 479959191Skris [Steve Henson] 480059191Skris 480159191Skris *) Non-copying interface to BIO pairs. 480259191Skris (still largely untested) 480359191Skris [Bodo Moeller] 480459191Skris 480559191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 480659191Skris ASCII string. This was handled independently in various places before. 480759191Skris [Steve Henson] 480859191Skris 480959191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 481059191Skris UTF8 strings a character at a time. 481159191Skris [Steve Henson] 481259191Skris 481359191Skris *) Use client_version from client hello to select the protocol 481459191Skris (s23_srvr.c) and for RSA client key exchange verification 481559191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 481659191Skris [Bodo Moeller] 481759191Skris 481859191Skris *) Add various utility functions to handle SPKACs, these were previously 481959191Skris handled by poking round in the structure internals. Added new function 482059191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 482159191Skris print, verify and generate SPKACs. Based on an original idea from 482259191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 482359191Skris [Steve Henson] 482459191Skris 482559191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 482659191Skris [Andy Polyakov] 482759191Skris 482859191Skris *) Allow the config file extension section to be overwritten on the 482959191Skris command line. Based on an original idea from Massimiliano Pala 483059191Skris <madwolf@comune.modena.it>. The new option is called -extensions 483159191Skris and can be applied to ca, req and x509. Also -reqexts to override 483259191Skris the request extensions in req and -crlexts to override the crl extensions 483359191Skris in ca. 483459191Skris [Steve Henson] 483559191Skris 483659191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 483759191Skris the same field multiple times by preceding it by "XXXX." for example: 483859191Skris 1.OU="Unit name 1" 483959191Skris 2.OU="Unit name 2" 484059191Skris this is the same syntax as used in the req config file. 484159191Skris [Steve Henson] 484259191Skris 484359191Skris *) Allow certificate extensions to be added to certificate requests. These 484459191Skris are specified in a 'req_extensions' option of the req section of the 484559191Skris config file. They can be printed out with the -text option to req but 484659191Skris are otherwise ignored at present. 484759191Skris [Steve Henson] 484859191Skris 484959191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 485059191Skris data read consists of only the final block it would not decrypted because 485159191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 485259191Skris A misplaced 'break' also meant the decrypted final block might not be 485359191Skris copied until the next read. 485459191Skris [Steve Henson] 485559191Skris 485659191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 485759191Skris a few extra parameters to the DH structure: these will be useful if 485859191Skris for example we want the value of 'q' or implement X9.42 DH. 485959191Skris [Steve Henson] 486059191Skris 486159191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 486259191Skris provides hooks that allow the default DSA functions or functions on a 486359191Skris "per key" basis to be replaced. This allows hardware acceleration and 486459191Skris hardware key storage to be handled without major modification to the 486559191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 486659191Skris associated functions. 486759191Skris [Steve Henson] 486859191Skris 486959191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 487059191Skris as "read only": it can't be written to and the buffer it points to will 487159191Skris not be freed. Reading from a read only BIO is much more efficient than 487259191Skris a normal memory BIO. This was added because there are several times when 487359191Skris an area of memory needs to be read from a BIO. The previous method was 487459191Skris to create a memory BIO and write the data to it, this results in two 487559191Skris copies of the data and an O(n^2) reading algorithm. There is a new 487659191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 487759191Skris an area of memory. Also modified the PKCS#7 routines to use read only 487859191Skris memory BIOs. 487959191Skris [Steve Henson] 488059191Skris 488159191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 488259191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 488359191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 488459191Skris but a retry condition occured while trying to read the rest. 488559191Skris [Bodo Moeller] 488659191Skris 488759191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 488859191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 488959191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 489059191Skris the encrypted data type: this is a more sensible place to put it and it 489159191Skris allows the PKCS#12 code to be tidied up that duplicated this 489259191Skris functionality. 489359191Skris [Steve Henson] 489459191Skris 489559191Skris *) Changed obj_dat.pl script so it takes its input and output files on 489659191Skris the command line. This should avoid shell escape redirection problems 489759191Skris under Win32. 489859191Skris [Steve Henson] 489959191Skris 490059191Skris *) Initial support for certificate extension requests, these are included 490159191Skris in things like Xenroll certificate requests. Included functions to allow 490259191Skris extensions to be obtained and added. 490359191Skris [Steve Henson] 490459191Skris 490559191Skris *) -crlf option to s_client and s_server for sending newlines as 490659191Skris CRLF (as required by many protocols). 490759191Skris [Bodo Moeller] 490859191Skris 490955714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 491055714Skris 491155714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 491255714Skris [Ralf S. Engelschall] 491355714Skris 491455714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 491555714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 491655714Skris 491755714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 491855714Skris program. 491955714Skris [Steve Henson] 492055714Skris 492155714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 492255714Skris DH parameters/keys (q is lost during that conversion, but the resulting 492355714Skris DH parameters contain its length). 492455714Skris 492555714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 492655714Skris much faster than DH_generate_parameters (which creates parameters 492755714Skris where p = 2*q + 1), and also the smaller q makes DH computations 492855714Skris much more efficient (160-bit exponentiation instead of 1024-bit 492955714Skris exponentiation); so this provides a convenient way to support DHE 493055714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 493155714Skris utter importance to use 493255714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 493355714Skris or 493455714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 493555714Skris when such DH parameters are used, because otherwise small subgroup 493655714Skris attacks may become possible! 493755714Skris [Bodo Moeller] 493855714Skris 493955714Skris *) Avoid memory leak in i2d_DHparams. 494055714Skris [Bodo Moeller] 494155714Skris 494255714Skris *) Allow the -k option to be used more than once in the enc program: 494355714Skris this allows the same encrypted message to be read by multiple recipients. 494455714Skris [Steve Henson] 494555714Skris 494655714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 494755714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 494855714Skris it will always use the numerical form of the OID, even if it has a short 494955714Skris or long name. 495055714Skris [Steve Henson] 495155714Skris 495255714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 495355714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 495455714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 495555714Skris no private key components need be present and it might store extra data 495659191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 495759191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 495859191Skris private key operations. 495955714Skris [Steve Henson] 496055714Skris 496155714Skris *) Added support for SPARC Linux. 496255714Skris [Andy Polyakov] 496355714Skris 496455714Skris *) pem_password_cb function type incompatibly changed from 496555714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 496655714Skris to 496755714Skris ....(char *buf, int size, int rwflag, void *userdata); 496855714Skris so that applications can pass data to their callbacks: 496955714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 497055714Skris additional void * argument, which is just handed through whenever 497155714Skris the password callback is called. 497259191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 497355714Skris 497455714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 497555714Skris 497655714Skris Compatibility note: As many C implementations push function arguments 497755714Skris onto the stack in reverse order, the new library version is likely to 497855714Skris interoperate with programs that have been compiled with the old 497955714Skris pem_password_cb definition (PEM_whatever takes some data that 498055714Skris happens to be on the stack as its last argument, and the callback 498155714Skris just ignores this garbage); but there is no guarantee whatsoever that 498255714Skris this will work. 498355714Skris 498455714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 498555714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 498655714Skris problems not only on Windows, but also on some Unix platforms. 498755714Skris To avoid problematic command lines, these definitions are now in an 498855714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 498955714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 499055714Skris [Bodo Moeller] 499155714Skris 499255714Skris *) MIPS III/IV assembler module is reimplemented. 499355714Skris [Andy Polyakov] 499455714Skris 499555714Skris *) More DES library cleanups: remove references to srand/rand and 499655714Skris delete an unused file. 499755714Skris [Ulf M�ller] 499855714Skris 499955714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 500055714Skris since not many people have MASM (ml) and it can be hard to obtain. 500155714Skris This is currently experimental but it seems to work OK and pass all 500255714Skris the tests. Check out INSTALL.W32 for info. 500355714Skris [Steve Henson] 500455714Skris 500555714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 500655714Skris without temporary keys kept an extra copy of the server key, 500755714Skris and connections with temporary keys did not free everything in case 500855714Skris of an error. 500955714Skris [Bodo Moeller] 501055714Skris 501155714Skris *) New function RSA_check_key and new openssl rsa option -check 501255714Skris for verifying the consistency of RSA keys. 501355714Skris [Ulf Moeller, Bodo Moeller] 501455714Skris 501555714Skris *) Various changes to make Win32 compile work: 501655714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 501755714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 501855714Skris comparison" warnings. 501955714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 502055714Skris [Steve Henson] 502155714Skris 502255714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 502355714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 502455714Skris derived keys are printed to stderr. 502555714Skris [Steve Henson] 502655714Skris 502755714Skris *) Copy the flags in ASN1_STRING_dup(). 502855714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 502955714Skris 503055714Skris *) The x509 application mishandled signing requests containing DSA 503155714Skris keys when the signing key was also DSA and the parameters didn't match. 503255714Skris 503355714Skris It was supposed to omit the parameters when they matched the signing key: 503455714Skris the verifying software was then supposed to automatically use the CA's 503555714Skris parameters if they were absent from the end user certificate. 503655714Skris 503755714Skris Omitting parameters is no longer recommended. The test was also 503855714Skris the wrong way round! This was probably due to unusual behaviour in 503955714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 504055714Skris This meant that parameters were omitted when they *didn't* match and 504155714Skris the certificate was useless. Certificates signed with 'ca' didn't have 504255714Skris this bug. 504355714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 504455714Skris 504555714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 504655714Skris The interface is as follows: 504755714Skris Applications can use 504855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 504955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 505055714Skris "off" is now the default. 505155714Skris The library internally uses 505255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 505355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 505455714Skris to disable memory-checking temporarily. 505555714Skris 505655714Skris Some inconsistent states that previously were possible (and were 505755714Skris even the default) are now avoided. 505855714Skris 505955714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 506055714Skris with each memory chunk allocated; this is occasionally more helpful 506155714Skris than just having a counter. 506255714Skris 506355714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 506455714Skris 506555714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 506655714Skris extensions. 506755714Skris [Bodo Moeller] 506855714Skris 506955714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 507055714Skris which largely parallels "options", but is for changing API behaviour, 507155714Skris whereas "options" are about protocol behaviour. 507255714Skris Initial "mode" flags are: 507355714Skris 507455714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 507555714Skris a single record has been written. 507655714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 507755714Skris retries use the same buffer location. 507855714Skris (But all of the contents must be 507955714Skris copied!) 508055714Skris [Bodo Moeller] 508155714Skris 508279998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 508355714Skris worked. 508455714Skris 508555714Skris *) Fix problems with no-hmac etc. 508655714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 508755714Skris 508855714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 508955714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 509055714Skris to mess around with the internals of an RSA structure. 509155714Skris [Steve Henson] 509255714Skris 509355714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 509455714Skris Also really enable memory leak checks in openssl.c and in some 509555714Skris test programs. 509655714Skris [Chad C. Mulligan, Bodo Moeller] 509755714Skris 509855714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 509955714Skris up the length of negative integers. This has now been simplified to just 510055714Skris store the length when it is first determined and use it later, rather 510155714Skris than trying to keep track of where data is copied and updating it to 510255714Skris point to the end. 510355714Skris [Steve Henson, reported by Brien Wheeler 510455714Skris <bwheeler@authentica-security.com>] 510555714Skris 510655714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 510755714Skris of a PKCS#7 signature but with the signing certificate passed to the 510855714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 510955714Skris certificate is present in the PKCS#7 structure. This isn't always the 511055714Skris case: certificates can be omitted from a PKCS#7 structure and be 511155714Skris distributed by "out of band" means (such as a certificate database). 511255714Skris [Steve Henson] 511355714Skris 511455714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 511555714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 511655714Skris necessary function names. 511755714Skris [Steve Henson] 511855714Skris 511955714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 512055714Skris options set by Configure in the top level Makefile, and Configure 512155714Skris was not even able to write more than one option correctly. 512255714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 512355714Skris [Bodo Moeller] 512455714Skris 512555714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 512655714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 512755714Skris for example allow memory BIOs to contain config info. 512855714Skris [Steve Henson] 512955714Skris 513055714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 513155714Skris Whoever hopes to achieve shared-library compatibility across versions 513255714Skris must use this, not the compile-time macro. 513355714Skris (Exercise 0.9.4: Which is the minimum library version required by 513455714Skris such programs?) 513555714Skris Note: All this applies only to multi-threaded programs, others don't 513655714Skris need locks. 513755714Skris [Bodo Moeller] 513855714Skris 513955714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 514055714Skris through a BIO pair triggered the default case, i.e. 514155714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 514255714Skris [Bodo Moeller] 514355714Skris 514455714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 514555714Skris can use the SSL library even if none of the specific BIOs is 514655714Skris appropriate. 514755714Skris [Bodo Moeller] 514855714Skris 514955714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 515055714Skris for the encoded length. 515155714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 515255714Skris 515355714Skris *) Add initial documentation of the X509V3 functions. 515455714Skris [Steve Henson] 515555714Skris 515655714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 515755714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 515855714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 515955714Skris secure PKCS#8 private key format with a high iteration count. 516055714Skris [Steve Henson] 516155714Skris 516255714Skris *) Fix determination of Perl interpreter: A perl or perl5 516355714Skris _directory_ in $PATH was also accepted as the interpreter. 516455714Skris [Ralf S. Engelschall] 516555714Skris 516655714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 516755714Skris wrong with it but it was very old and did things like calling 516855714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 516955714Skris unusual formatting. 517055714Skris [Steve Henson] 517155714Skris 517255714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 517355714Skris to use the new extension code. 517455714Skris [Steve Henson] 517555714Skris 517655714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 517755714Skris with macros. This should make it easier to change their form, add extra 517855714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 517955714Skris constant. 518055714Skris [Steve Henson] 518155714Skris 518255714Skris *) Add to configuration table a new entry that can specify an alternative 518355714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 518455714Skris according to Mark Crispin <MRC@Panda.COM>. 518555714Skris [Bodo Moeller] 518655714Skris 518755714Skris#if 0 518855714Skris *) DES CBC did not update the IV. Weird. 518955714Skris [Ben Laurie] 519055714Skris#else 519155714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 519255714Skris Changing the behaviour of the former might break existing programs -- 519355714Skris where IV updating is needed, des_ncbc_encrypt can be used. 519455714Skris#endif 519555714Skris 519655714Skris *) When bntest is run from "make test" it drives bc to check its 519755714Skris calculations, as well as internally checking them. If an internal check 519855714Skris fails, it needs to cause bc to give a non-zero result or make test carries 519955714Skris on without noticing the failure. Fixed. 520055714Skris [Ben Laurie] 520155714Skris 520255714Skris *) DES library cleanups. 520355714Skris [Ulf M�ller] 520455714Skris 520555714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 520655714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 520755714Skris ciphers. NOTE: although the key derivation function has been verified 520855714Skris against some published test vectors it has not been extensively tested 520955714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 521055714Skris of v2.0. 521155714Skris [Steve Henson] 521255714Skris 521355714Skris *) Instead of "mkdir -p", which is not fully portable, use new 521455714Skris Perl script "util/mkdir-p.pl". 521555714Skris [Bodo Moeller] 521655714Skris 521755714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 521855714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 521955714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 522055714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 522155714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 522255714Skris underlying key generation function so it must do its own ASN1 parsing. 522355714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 522455714Skris 'parameter' argument instead of literal salt and iteration count values 522555714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 522655714Skris [Steve Henson] 522755714Skris 522855714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 522955714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 523055714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 523155714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 523255714Skris value was just used as a "magic string" and not used directly its 523355714Skris value doesn't matter. 523455714Skris [Steve Henson] 523555714Skris 523655714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 523755714Skris support mutable. 523855714Skris [Ben Laurie] 523955714Skris 524055714Skris *) "linux-sparc64" configuration (ultrapenguin). 524155714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 524255714Skris "linux-sparc" configuration. 524355714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 524455714Skris 524555714Skris *) config now generates no-xxx options for missing ciphers. 524655714Skris [Ulf M�ller] 524755714Skris 524855714Skris *) Support the EBCDIC character set (work in progress). 524955714Skris File ebcdic.c not yet included because it has a different license. 525055714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 525155714Skris 525255714Skris *) Support BS2000/OSD-POSIX. 525355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 525455714Skris 525555714Skris *) Make callbacks for key generation use void * instead of char *. 525655714Skris [Ben Laurie] 525755714Skris 525855714Skris *) Make S/MIME samples compile (not yet tested). 525955714Skris [Ben Laurie] 526055714Skris 526155714Skris *) Additional typesafe stacks. 526255714Skris [Ben Laurie] 526355714Skris 526455714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 526555714Skris [Bodo Moeller] 526655714Skris 526755714Skris 526855714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 526955714Skris 527055714Skris *) New configuration variant "sco5-gcc". 527155714Skris 527255714Skris *) Updated some demos. 527355714Skris [Sean O Riordain, Wade Scholine] 527455714Skris 527555714Skris *) Add missing BIO_free at exit of pkcs12 application. 527655714Skris [Wu Zhigang] 527755714Skris 527855714Skris *) Fix memory leak in conf.c. 527955714Skris [Steve Henson] 528055714Skris 528155714Skris *) Updates for Win32 to assembler version of MD5. 528255714Skris [Steve Henson] 528355714Skris 528455714Skris *) Set #! path to perl in apps/der_chop to where we found it 528555714Skris instead of using a fixed path. 528655714Skris [Bodo Moeller] 528755714Skris 528855714Skris *) SHA library changes for irix64-mips4-cc. 528955714Skris [Andy Polyakov] 529055714Skris 529155714Skris *) Improvements for VMS support. 529255714Skris [Richard Levitte] 529355714Skris 529455714Skris 529555714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 529655714Skris 529755714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 529855714Skris This also avoids the problems with SC4.2 and unpatched SC5. 529955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 530055714Skris 530155714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 530255714Skris These are required because of the typesafe stack would otherwise break 530355714Skris existing code. If old code used a structure member which used to be STACK 530455714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 530555714Skris sk_num or sk_value it would produce an error because the num, data members 530655714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 530755714Skris replaces the old method of assigning a value to sk_value 530855714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 530955714Skris that does this will no longer work (and should use sk_set instead) but 531055714Skris this could be regarded as a "questionable" behaviour anyway. 531155714Skris [Steve Henson] 531255714Skris 531355714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 531455714Skris correctly handle encrypted S/MIME data. 531555714Skris [Steve Henson] 531655714Skris 531755714Skris *) Change type of various DES function arguments from des_cblock 531855714Skris (which means, in function argument declarations, pointer to char) 531955714Skris to des_cblock * (meaning pointer to array with 8 char elements), 532055714Skris which allows the compiler to do more typechecking; it was like 532155714Skris that back in SSLeay, but with lots of ugly casts. 532255714Skris 532355714Skris Introduce new type const_des_cblock. 532455714Skris [Bodo Moeller] 532555714Skris 532655714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 532755714Skris problems: find RecipientInfo structure that matches recipient certificate 532855714Skris and initialise the ASN1 structures properly based on passed cipher. 532955714Skris [Steve Henson] 533055714Skris 533155714Skris *) Belatedly make the BN tests actually check the results. 533255714Skris [Ben Laurie] 533355714Skris 533455714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 533555714Skris to and from BNs: it was completely broken. New compilation option 533655714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 533755714Skris key elements as negative integers. 533855714Skris [Steve Henson] 533955714Skris 534055714Skris *) Reorganize and speed up MD5. 534155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 534255714Skris 534355714Skris *) VMS support. 534455714Skris [Richard Levitte <richard@levitte.org>] 534555714Skris 534655714Skris *) New option -out to asn1parse to allow the parsed structure to be 534755714Skris output to a file. This is most useful when combined with the -strparse 534855714Skris option to examine the output of things like OCTET STRINGS. 534955714Skris [Steve Henson] 535055714Skris 535155714Skris *) Make SSL library a little more fool-proof by not requiring any longer 535255714Skris that SSL_set_{accept,connect}_state be called before 535355714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 535455714Skris in many applications because usually everything *appeared* to work as 535555714Skris intended anyway -- now it really works as intended). 535655714Skris [Bodo Moeller] 535755714Skris 535855714Skris *) Move openssl.cnf out of lib/. 535955714Skris [Ulf M�ller] 536055714Skris 536155714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 536255714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 536355714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 536455714Skris [Ralf S. Engelschall] 536555714Skris 536655714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 536755714Skris handle PKCS#7 enveloped data properly. 536855714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 536955714Skris 537055714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 537155714Skris copying pointers. The cert_st handling is changed by this in 537255714Skris various ways (and thus what used to be known as ctx->default_cert 537355714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 537455714Skris any longer when s->cert does not give us what we need). 537555714Skris ssl_cert_instantiate becomes obsolete by this change. 537655714Skris As soon as we've got the new code right (possibly it already is?), 537755714Skris we have solved a couple of bugs of the earlier code where s->cert 537855714Skris was used as if it could not have been shared with other SSL structures. 537955714Skris 538055714Skris Note that using the SSL API in certain dirty ways now will result 538155714Skris in different behaviour than observed with earlier library versions: 538255714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 538355714Skris does not influence s as it used to. 538455714Skris 538555714Skris In order to clean up things more thoroughly, inside SSL_SESSION 538655714Skris we don't use CERT any longer, but a new structure SESS_CERT 538755714Skris that holds per-session data (if available); currently, this is 538855714Skris the peer's certificate chain and, for clients, the server's certificate 538955714Skris and temporary key. CERT holds only those values that can have 539055714Skris meaningful defaults in an SSL_CTX. 539155714Skris [Bodo Moeller] 539255714Skris 539355714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 539455714Skris from the internal representation. Various PKCS#7 fixes: remove some 539555714Skris evil casts and set the enc_dig_alg field properly based on the signing 539655714Skris key type. 539755714Skris [Steve Henson] 539855714Skris 539955714Skris *) Allow PKCS#12 password to be set from the command line or the 540055714Skris environment. Let 'ca' get its config file name from the environment 540155714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 540255714Skris and 'x509'). 540355714Skris [Steve Henson] 540455714Skris 540555714Skris *) Allow certificate policies extension to use an IA5STRING for the 540655714Skris organization field. This is contrary to the PKIX definition but 540755714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 540855714Skris extension option. 540955714Skris [Steve Henson] 541055714Skris 541155714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 541255714Skris without disallowing inline assembler and the like for non-pedantic builds. 541355714Skris [Ben Laurie] 541455714Skris 541555714Skris *) Support Borland C++ builder. 541655714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 541755714Skris 541855714Skris *) Support Mingw32. 541955714Skris [Ulf M�ller] 542055714Skris 542155714Skris *) SHA-1 cleanups and performance enhancements. 542255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 542355714Skris 542455714Skris *) Sparc v8plus assembler for the bignum library. 542555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 542655714Skris 542755714Skris *) Accept any -xxx and +xxx compiler options in Configure. 542855714Skris [Ulf M�ller] 542955714Skris 543055714Skris *) Update HPUX configuration. 543155714Skris [Anonymous] 543255714Skris 543355714Skris *) Add missing sk_<type>_unshift() function to safestack.h 543455714Skris [Ralf S. Engelschall] 543555714Skris 543655714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 543755714Skris "extra_cert"s in addition to the certificate. (This makes sense 543855714Skris only for "PEM" format files, as chains as a whole are not 543955714Skris DER-encoded.) 544055714Skris [Bodo Moeller] 544155714Skris 544255714Skris *) Support verify_depth from the SSL API. 544355714Skris x509_vfy.c had what can be considered an off-by-one-error: 544455714Skris Its depth (which was not part of the external interface) 544555714Skris was actually counting the number of certificates in a chain; 544655714Skris now it really counts the depth. 544755714Skris [Bodo Moeller] 544855714Skris 544955714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 545055714Skris instead of X509err, which often resulted in confusing error 545155714Skris messages since the error codes are not globally unique 545255714Skris (e.g. an alleged error in ssl3_accept when a certificate 545355714Skris didn't match the private key). 545455714Skris 545555714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 545655714Skris value (so that you don't need SSL_set_session_id_context for each 545755714Skris connection using the SSL_CTX). 545855714Skris [Bodo Moeller] 545955714Skris 546055714Skris *) OAEP decoding bug fix. 546155714Skris [Ulf M�ller] 546255714Skris 546355714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 546455714Skris David Harris. 546555714Skris [Bodo Moeller] 546655714Skris 546755714Skris *) New Configure options "threads" and "no-threads". For systems 546855714Skris where the proper compiler options are known (currently Solaris 546955714Skris and Linux), "threads" is the default. 547055714Skris [Bodo Moeller] 547155714Skris 547255714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 547355714Skris [Bodo Moeller] 547455714Skris 547555714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 547655714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 547755714Skris such as /usr/local/bin. 547855714Skris [Bodo Moeller] 547955714Skris 548055714Skris *) "make linux-shared" to build shared libraries. 548155714Skris [Niels Poppe <niels@netbox.org>] 548255714Skris 548355714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 548455714Skris [Ulf M�ller] 548555714Skris 548655714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 548755714Skris extension adding in x509 utility. 548855714Skris [Steve Henson] 548955714Skris 549055714Skris *) Remove NOPROTO sections and error code comments. 549155714Skris [Ulf M�ller] 549255714Skris 549355714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 549455714Skris prototypes. 549555714Skris [Steve Henson] 549655714Skris 549755714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 549855714Skris [Ulf M�ller] 549955714Skris 550055714Skris *) Complete rewrite of the error code script(s). It is all now handled 550155714Skris by one script at the top level which handles error code gathering, 550255714Skris header rewriting and C source file generation. It should be much better 550355714Skris than the old method: it now uses a modified version of Ulf's parser to 550455714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 550555714Skris aren't needed for error creation any more) and do a better job of 550655714Skris translating function codes into names. The old 'ASN1 error code imbedded 550755714Skris in a comment' is no longer necessary and it doesn't use .err files which 550855714Skris have now been deleted. Also the error code call doesn't have to appear all 550955714Skris on one line (which resulted in some large lines...). 551055714Skris [Steve Henson] 551155714Skris 551255714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 551355714Skris [Bodo Moeller] 551455714Skris 551555714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 551655714Skris 0 (which usually indicates a closed connection), but continue reading. 551755714Skris [Bodo Moeller] 551855714Skris 551955714Skris *) Fix some race conditions. 552055714Skris [Bodo Moeller] 552155714Skris 552255714Skris *) Add support for CRL distribution points extension. Add Certificate 552355714Skris Policies and CRL distribution points documentation. 552455714Skris [Steve Henson] 552555714Skris 552655714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 552755714Skris [Ulf M�ller] 552855714Skris 552955714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 553055714Skris 8 of keying material. Merlin has also confirmed interop with this fix 553155714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 553255714Skris [Merlin Hughes <merlin@baltimore.ie>] 553355714Skris 553455714Skris *) Fix lots of warnings. 553555714Skris [Richard Levitte <levitte@stacken.kth.se>] 553655714Skris 553755714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 553855714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 553955714Skris [Richard Levitte <levitte@stacken.kth.se>] 554055714Skris 554155714Skris *) Fix problems with sizeof(long) == 8. 554255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 554355714Skris 554455714Skris *) Change functions to ANSI C. 554555714Skris [Ulf M�ller] 554655714Skris 554755714Skris *) Fix typos in error codes. 554855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 554955714Skris 555055714Skris *) Remove defunct assembler files from Configure. 555155714Skris [Ulf M�ller] 555255714Skris 555355714Skris *) SPARC v8 assembler BIGNUM implementation. 555455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 555555714Skris 555655714Skris *) Support for Certificate Policies extension: both print and set. 555755714Skris Various additions to support the r2i method this uses. 555855714Skris [Steve Henson] 555955714Skris 556055714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 556155714Skris return a const string when you are expecting an allocated buffer. 556255714Skris [Ben Laurie] 556355714Skris 556455714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 556555714Skris types DirectoryString and DisplayText. 556655714Skris [Steve Henson] 556755714Skris 556855714Skris *) Add code to allow r2i extensions to access the configuration database, 556955714Skris add an LHASH database driver and add several ctx helper functions. 557055714Skris [Steve Henson] 557155714Skris 557255714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 557355714Skris fail when they extended the size of a BIGNUM. 557455714Skris [Steve Henson] 557555714Skris 557655714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 557755714Skris support typesafe stack. 557855714Skris [Steve Henson] 557955714Skris 558055714Skris *) Fix typo in SSL_[gs]et_options(). 558155714Skris [Nils Frostberg <nils@medcom.se>] 558255714Skris 558355714Skris *) Delete various functions and files that belonged to the (now obsolete) 558455714Skris old X509V3 handling code. 558555714Skris [Steve Henson] 558655714Skris 558755714Skris *) New Configure option "rsaref". 558855714Skris [Ulf M�ller] 558955714Skris 559055714Skris *) Don't auto-generate pem.h. 559155714Skris [Bodo Moeller] 559255714Skris 559355714Skris *) Introduce type-safe ASN.1 SETs. 559455714Skris [Ben Laurie] 559555714Skris 559655714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 559755714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 559855714Skris 559955714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 560055714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 560155714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 560255714Skris few STACKed things have been converted already. Feel free to convert more. 560355714Skris In the fullness of time, I'll do away with the STACK type altogether. 560455714Skris [Ben Laurie] 560555714Skris 560655714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 560755714Skris specified in <certfile> by updating the entry in the index.txt file. 560855714Skris This way one no longer has to edit the index.txt file manually for 560955714Skris revoking a certificate. The -revoke option does the gory details now. 561055714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 561155714Skris 561255714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 561355714Skris `-text' option at all and this way the `-noout -text' combination was 561455714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 561555714Skris [Ralf S. Engelschall] 561655714Skris 561755714Skris *) Make sure a corresponding plain text error message exists for the 561855714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 561955714Skris verify callback function determined that a certificate was revoked. 562055714Skris [Ralf S. Engelschall] 562155714Skris 562255714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 562355714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 562455714Skris all available cipers including rc5, which was forgotten until now. 562555714Skris In order to let the testing shell script know which algorithms 562655714Skris are available, a new (up to now undocumented) command 562755714Skris "openssl list-cipher-commands" is used. 562855714Skris [Bodo Moeller] 562955714Skris 563055714Skris *) Bugfix: s_client occasionally would sleep in select() when 563155714Skris it should have checked SSL_pending() first. 563255714Skris [Bodo Moeller] 563355714Skris 563455714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 563555714Skris the raw DSA values prior to ASN.1 encoding. 563655714Skris [Ulf M�ller] 563755714Skris 563855714Skris *) Tweaks to Configure 563955714Skris [Niels Poppe <niels@netbox.org>] 564055714Skris 564155714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 564255714Skris yet... 564355714Skris [Steve Henson] 564455714Skris 564555714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 564655714Skris [Ulf M�ller] 564755714Skris 564855714Skris *) New config option to avoid instructions that are illegal on the 80386. 564955714Skris The default code is faster, but requires at least a 486. 565055714Skris [Ulf M�ller] 565155714Skris 565255714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 565355714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 565455714Skris same as SSL2_VERSION anyway. 565555714Skris [Bodo Moeller] 565655714Skris 565755714Skris *) New "-showcerts" option for s_client. 565855714Skris [Bodo Moeller] 565955714Skris 566055714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 566155714Skris application. Various cleanups and fixes. 566255714Skris [Steve Henson] 566355714Skris 566455714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 566555714Skris modify error routines to work internally. Add error codes and PBE init 566655714Skris to library startup routines. 566755714Skris [Steve Henson] 566855714Skris 566955714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 567055714Skris packing functions to asn1 and evp. Changed function names and error 567155714Skris codes along the way. 567255714Skris [Steve Henson] 567355714Skris 567455714Skris *) PKCS12 integration: and so it begins... First of several patches to 567555714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 567655714Skris objects to objects.h 567755714Skris [Steve Henson] 567855714Skris 567955714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 568055714Skris and display support for Thawte strong extranet extension. 568155714Skris [Steve Henson] 568255714Skris 568355714Skris *) Add LinuxPPC support. 568455714Skris [Jeff Dubrule <igor@pobox.org>] 568555714Skris 568655714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 568755714Skris bn_div_words in alpha.s. 568855714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 568955714Skris 569055714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 569155714Skris OAEP isn't supported when OpenSSL is built with RSAref. 569255714Skris [Ulf Moeller <ulf@fitug.de>] 569355714Skris 569455714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 569555714Skris so they no longer are missing under -DNOPROTO. 569655714Skris [Soren S. Jorvang <soren@t.dk>] 569755714Skris 569855714Skris 569955714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 570055714Skris 570155714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 570255714Skris doesn't work when the session is reused. Coming soon! 570355714Skris [Ben Laurie] 570455714Skris 570555714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 570655714Skris context thus bypassing client cert protection! All software that uses 570755714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 570855714Skris allow session reuse! A fuller solution is in the works. 570955714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 571055714Skris 571155714Skris *) Some more source tree cleanups (removed obsolete files 571255714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 571355714Skris permission on "config" script to be executable) and a fix for the INSTALL 571455714Skris document. 571555714Skris [Ulf Moeller <ulf@fitug.de>] 571655714Skris 571755714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 571855714Skris Malloc, Free. 571955714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 572055714Skris 572155714Skris *) Make rsa_oaep_test return non-zero on error. 572255714Skris [Ulf Moeller <ulf@fitug.de>] 572355714Skris 572455714Skris *) Add support for native Solaris shared libraries. Configure 572555714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 572655714Skris if someone would make that last step automatic. 572755714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 572855714Skris 572955714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 573055714Skris [Ben Laurie] 573155714Skris 573255714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 573355714Skris except NULL ciphers". This means the default cipher list will no longer 573455714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 573555714Skris the string "DEFAULT:eNULL". 573655714Skris [Steve Henson] 573755714Skris 573855714Skris *) Fix to RSA private encryption routines: if p < q then it would 573955714Skris occasionally produce an invalid result. This will only happen with 574055714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 574155714Skris [Steve Henson] 574255714Skris 574355714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 574455714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 574555714Skris because this way one can also use an interpreter named `perl5' (which is 574655714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 574755714Skris installed as `perl'). 574855714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 574955714Skris 575055714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 575155714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 575255714Skris 575355714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 575455714Skris advapi32.lib to Win32 build and change the pem test comparision 575555714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 575655714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 575755714Skris and crypto/des/ede_cbcm_enc.c. 575855714Skris [Steve Henson] 575955714Skris 576055714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 576155714Skris [Ben Laurie] 576255714Skris 576355714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 576455714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 576555714Skris is horrible: I feel ill.... 576655714Skris [Steve Henson] 576755714Skris 576855714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 576955714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 577055714Skris sections: 10 functions were absent from non ANSI section and not exported 577155714Skris from Windows DLLs. Fixed up libeay.num for new functions. 577255714Skris [Steve Henson] 577355714Skris 577455714Skris *) Make `openssl version' output lines consistent. 577555714Skris [Ralf S. Engelschall] 577655714Skris 577755714Skris *) Fix Win32 symbol export lists for BIO functions: Added 577855714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 577955714Skris to ms/libeay{16,32}.def. 578055714Skris [Ralf S. Engelschall] 578155714Skris 578255714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 578355714Skris fine under Unix and passes some trivial tests I've now added. But the 578455714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 578555714Skris added to make sure no one expects that this stuff really works in the 578655714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 578755714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 578855714Skris openssl_bio.xs. 578955714Skris [Ralf S. Engelschall] 579055714Skris 579155714Skris *) Fix the generation of two part addresses in perl. 579255714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 579355714Skris 579455714Skris *) Add config entry for Linux on MIPS. 579555714Skris [John Tobey <jtobey@channel1.com>] 579655714Skris 579755714Skris *) Make links whenever Configure is run, unless we are on Windoze. 579855714Skris [Ben Laurie] 579955714Skris 580055714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 580155714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 580255714Skris in CRLs. 580355714Skris [Steve Henson] 580455714Skris 580555714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 580655714Skris other platforms details on the command line without having to patch the 580755714Skris Configure script everytime: One now can use ``perl Configure 580855714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 580955714Skris to them (seperated by colons). This is treated as there would be a static 581055714Skris pre-configured entry in Configure's %table under key <id> with value 581155714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 581255714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 581355714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 581455714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 581555714Skris [Ralf S. Engelschall] 581655714Skris 581755714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 581855714Skris [Ben Laurie] 581955714Skris 582055714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 582155714Skris on the `perl Configure ...' command line. This way one can compile 582255714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 582355714Skris for linking it into DSOs. 582455714Skris [Ralf S. Engelschall] 582555714Skris 582655714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 582755714Skris Fixed. 582855714Skris [Ben Laurie] 582955714Skris 583055714Skris *) Cleaned up the LICENSE document: The official contact for any license 583155714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 583255714Skris And add a paragraph about the dual-license situation to make sure people 583355714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 583455714Skris to the OpenSSL toolkit. 583555714Skris [Ralf S. Engelschall] 583655714Skris 583755714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 583855714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 583955714Skris Additonally cleaned up the `make links' target: Remove unnecessary 584055714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 584155714Skris to speed processing and no longer clutter the display with confusing 584255714Skris stuff. Instead only the actually done links are displayed. 584355714Skris [Ralf S. Engelschall] 584455714Skris 584555714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 584655714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 584755714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 584855714Skris encryption. 584955714Skris [Ben Laurie] 585055714Skris 585155714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 585255714Skris signed attributes when verifying signatures (this would break them), 585355714Skris the detached data encoding was wrong and public keys obtained using 585455714Skris X509_get_pubkey() weren't freed. 585555714Skris [Steve Henson] 585655714Skris 585755714Skris *) Add text documentation for the BUFFER functions. Also added a work around 585855714Skris to a Win95 console bug. This was triggered by the password read stuff: the 585955714Skris last character typed gets carried over to the next fread(). If you were 586055714Skris generating a new cert request using 'req' for example then the last 586155714Skris character of the passphrase would be CR which would then enter the first 586255714Skris field as blank. 586355714Skris [Steve Henson] 586455714Skris 586555714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 586655714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 586755714Skris button and can be used by applications based on OpenSSL to show the 586855714Skris relationship to the OpenSSL project. 586955714Skris [Ralf S. Engelschall] 587055714Skris 587155714Skris *) Remove confusing variables in function signatures in files 587255714Skris ssl/ssl_lib.c and ssl/ssl.h. 587355714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 587455714Skris 587555714Skris *) Don't install bss_file.c under PREFIX/include/ 587655714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 587755714Skris 587855714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 587955714Skris functions that return function pointers and has support for NT specific 588055714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 588155714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 588255714Skris unsigned to signed types: this was killing the Win32 compile. 588355714Skris [Steve Henson] 588455714Skris 588555714Skris *) Add new certificate file to stack functions, 588655714Skris SSL_add_dir_cert_subjects_to_stack() and 588755714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 588855714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 588955714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 589055714Skris This means that Apache-SSL and similar packages don't have to mess around 589155714Skris to add as many CAs as they want to the preferred list. 589255714Skris [Ben Laurie] 589355714Skris 589455714Skris *) Experiment with doxygen documentation. Currently only partially applied to 589555714Skris ssl/ssl_lib.c. 589655714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 589755714Skris openssl.doxy as the configuration file. 589855714Skris [Ben Laurie] 589955714Skris 590055714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 590155714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 590255714Skris 590355714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 590455714Skris compiled in by default: it has problems with large keys. 590555714Skris [Steve Henson] 590655714Skris 590755714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 590855714Skris DH private keys and/or callback functions which directly correspond to 590955714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 591055714Skris is needed for applications which have to configure certificates on a 591155714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 591255714Skris (e.g. s_server). 591355714Skris For the RSA certificate situation is makes no difference, but 591455714Skris for the DSA certificate situation this fixes the "no shared cipher" 591555714Skris problem where the OpenSSL cipher selection procedure failed because the 591655714Skris temporary keys were not overtaken from the context and the API provided 591755714Skris no way to reconfigure them. 591855714Skris The new functions now let applications reconfigure the stuff and they 591955714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 592055714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 592155714Skris non-public-API function ssl_cert_instantiate() is used as a helper 592255714Skris function and also to reduce code redundancy inside ssl_rsa.c. 592355714Skris [Ralf S. Engelschall] 592455714Skris 592555714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 592655714Skris area because they are useful for the DSA situation and should be 592755714Skris recognized by the users. 592855714Skris [Ralf S. Engelschall] 592955714Skris 593055714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 593155714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 593255714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 593355714Skris already masked variable. 593455714Skris [Richard Levitte <levitte@stacken.kth.se>] 593555714Skris 593655714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 593755714Skris [Richard Levitte <levitte@stacken.kth.se>] 593855714Skris 593955714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 594055714Skris from `int' to `unsigned int' because it's a length and initialized by 594155714Skris EVP_DigestFinal() which expects an `unsigned int *'. 594255714Skris [Richard Levitte <levitte@stacken.kth.se>] 594355714Skris 594455714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 594555714Skris script. Instead use the usual Shell->Perl transition trick. 594655714Skris [Ralf S. Engelschall] 594755714Skris 594855714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 594955714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 595055714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 595155714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 595255714Skris currently the public key is printed (a decision which was already done by 595355714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 595455714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 595555714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 595655714Skris now, too. 595755714Skris [Ralf S. Engelschall] 595855714Skris 595955714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 596055714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 596155714Skris [Arne Ansper <arne@ats.cyber.ee>] 596255714Skris 596355714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 596455714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 596555714Skris config file. 596655714Skris [Steve Henson] 596755714Skris 596855714Skris *) Add cool BIO that does syslog (or event log on NT). 596955714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 597055714Skris 597155714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 597255714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 597355714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 597455714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 597555714Skris [Ben Laurie] 597655714Skris 597755714Skris *) Add preliminary config info for new extension code. 597855714Skris [Steve Henson] 597955714Skris 598055714Skris *) Make RSA_NO_PADDING really use no padding. 598155714Skris [Ulf Moeller <ulf@fitug.de>] 598255714Skris 598355714Skris *) Generate errors when private/public key check is done. 598455714Skris [Ben Laurie] 598555714Skris 598655714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 598755714Skris for some CRL extensions and new objects added. 598855714Skris [Steve Henson] 598955714Skris 599055714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 599155714Skris key usage extension and fuller support for authority key id. 599255714Skris [Steve Henson] 599355714Skris 599455714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 599555714Skris padding method for RSA, which is recommended for new applications in PKCS 599655714Skris #1 v2.0 (RFC 2437, October 1998). 599755714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 599855714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 599955714Skris against Bleichbacher's attack on RSA. 600055714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 600155714Skris Ben Laurie] 600255714Skris 600355714Skris *) Updates to the new SSL compression code 600455714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 600555714Skris 600655714Skris *) Fix so that the version number in the master secret, when passed 600755714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 600855714Skris (because the server will not accept higher), that the version number 600955714Skris is 0x03,0x01, not 0x03,0x00 601055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 601155714Skris 601255714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 601355714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 601455714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 601555714Skris [Steve Henson] 601655714Skris 601755714Skris *) Support for RAW extensions where an arbitrary extension can be 601855714Skris created by including its DER encoding. See apps/openssl.cnf for 601955714Skris an example. 602055714Skris [Steve Henson] 602155714Skris 602255714Skris *) Make sure latest Perl versions don't interpret some generated C array 602355714Skris code as Perl array code in the crypto/err/err_genc.pl script. 602455714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 602555714Skris 602655714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 602755714Skris not many people have the assembler. Various Win32 compilation fixes and 602855714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 602955714Skris build instructions. 603055714Skris [Steve Henson] 603155714Skris 603255714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 603355714Skris file under Win32 and also build pem.h from pem.org. New script 603455714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 603555714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 603655714Skris [Steve Henson] 603755714Skris 603855714Skris *) Major rework of DES function declarations, in the pursuit of correctness 603955714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 604055714Skris too. You may find this causes warnings in your code. Zapping your evil 604155714Skris casts will probably fix them. Mostly. 604255714Skris [Ben Laurie] 604355714Skris 604455714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 604555714Skris obj_dat.pl. It considered a zero in an object definition to mean 604655714Skris "end of object": none of the objects in objects.h have any zeros 604755714Skris so it wasn't spotted. 604855714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 604955714Skris 605055714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 605155714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 605255714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 605355714Skris vectors if you have them. 605455714Skris [Ben Laurie] 605555714Skris 605655714Skris *) Correct calculation of key length for export ciphers (too much space was 605755714Skris allocated for null ciphers). This has not been tested! 605855714Skris [Ben Laurie] 605955714Skris 606055714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 606155714Skris message is now correct (it understands "crypto" and "ssl" on its 606255714Skris command line). There is also now an "update" option. This will update 606355714Skris the util/ssleay.num and util/libeay.num files with any new functions. 606455714Skris If you do a: 606555714Skris perl util/mkdef.pl crypto ssl update 606655714Skris it will update them. 606755714Skris [Steve Henson] 606855714Skris 606955714Skris *) Overhauled the Perl interface (perl/*): 607055714Skris - ported BN stuff to OpenSSL's different BN library 607155714Skris - made the perl/ source tree CVS-aware 607255714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 607355714Skris their history because I've copied them in the repository) 607455714Skris - removed obsolete files (the test scripts will be replaced 607555714Skris by better Test::Harness variants in the future) 607655714Skris [Ralf S. Engelschall] 607755714Skris 607855714Skris *) First cut for a very conservative source tree cleanup: 607955714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 608055714Skris where we collect the old documents and readme texts. 608155714Skris 2. remove the first part of files where I'm already sure that we no 608255714Skris longer need them because of three reasons: either they are just temporary 608355714Skris files which were left by Eric or they are preserved original files where 608455714Skris I've verified that the diff is also available in the CVS via "cvs diff 608555714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 608655714Skris the crypto/md/ stuff). 608755714Skris [Ralf S. Engelschall] 608855714Skris 608955714Skris *) More extension code. Incomplete support for subject and issuer alt 609055714Skris name, issuer and authority key id. Change the i2v function parameters 609155714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 609255714Skris what that's for :-) Fix to ASN1 macro which messed up 609355714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 609455714Skris [Steve Henson] 609555714Skris 609655714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 609755714Skris INTEGER code. 609855714Skris [Steve Henson] 609955714Skris 610055714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 610155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 610255714Skris 610355714Skris *) Make sure `make rehash' target really finds the `openssl' program. 610455714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 610555714Skris 610655714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 610755714Skris like to hear about it if this slows down other processors. 610855714Skris [Ben Laurie] 610955714Skris 611055714Skris *) Add CygWin32 platform information to Configure script. 611155714Skris [Alan Batie <batie@aahz.jf.intel.com>] 611255714Skris 611355714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 611455714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 611555714Skris 611655714Skris *) New program nseq to manipulate netscape certificate sequences 611755714Skris [Steve Henson] 611855714Skris 611955714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 612055714Skris few typos. 612155714Skris [Steve Henson] 612255714Skris 612355714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 612455714Skris but the BN code had some problems that would cause failures when 612555714Skris doing certificate verification and some other functions. 612655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 612755714Skris 612855714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 612955714Skris [Steve Henson] 613055714Skris 613155714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 613255714Skris [Steve Henson] 613355714Skris 613455714Skris *) Add several PKIX and private extended key usage OIDs. 613555714Skris [Steve Henson] 613655714Skris 613755714Skris *) Modify the 'ca' program to handle the new extension code. Modify 613855714Skris openssl.cnf for new extension format, add comments. 613955714Skris [Steve Henson] 614055714Skris 614155714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 614255714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 614355714Skris CA extensions. 614455714Skris [Steve Henson] 614555714Skris 614655714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 614755714Skris error code, add initial support to X509_print() and x509 application. 614855714Skris [Steve Henson] 614955714Skris 615055714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 615155714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 615255714Skris stuff is currently isolated and isn't even compiled yet. 615355714Skris [Steve Henson] 615455714Skris 615555714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 615655714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 615755714Skris Removed the versions check from X509 routines when loading extensions: 615855714Skris this allows certain broken certificates that don't set the version 615955714Skris properly to be processed. 616055714Skris [Steve Henson] 616155714Skris 616255714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 616355714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 616455714Skris can still be regenerated with "make depend". 616555714Skris [Ben Laurie] 616655714Skris 616755714Skris *) Spelling mistake in C version of CAST-128. 616855714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 616955714Skris 617055714Skris *) Changes to the error generation code. The perl script err-code.pl 617155714Skris now reads in the old error codes and retains the old numbers, only 617255714Skris adding new ones if necessary. It also only changes the .err files if new 617355714Skris codes are added. The makefiles have been modified to only insert errors 617455714Skris when needed (to avoid needlessly modifying header files). This is done 617555714Skris by only inserting errors if the .err file is newer than the auto generated 617655714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 617755714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 617855714Skris or delete all the .err files. 617955714Skris [Steve Henson] 618055714Skris 618155714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 618255714Skris been fixed, but is untested. The assembler versions are also fixed, but 618355714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 618455714Skris to regenerate it if needed. 618555714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 618655714Skris Hagino <itojun@kame.net>] 618755714Skris 618855714Skris *) File was opened incorrectly in randfile.c. 618955714Skris [Ulf M�ller <ulf@fitug.de>] 619055714Skris 619155714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 619255714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 619355714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 619455714Skris al: it's just almost always a UTCTime. Note this patch adds new error 619555714Skris codes so do a "make errors" if there are problems. 619655714Skris [Steve Henson] 619755714Skris 619855714Skris *) Correct Linux 1 recognition in config. 619955714Skris [Ulf M�ller <ulf@fitug.de>] 620055714Skris 620155714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 620255714Skris [Anonymous <nobody@replay.com>] 620355714Skris 620455714Skris *) Generate an error if given an empty string as a cert directory. Also 620555714Skris generate an error if handed NULL (previously returned 0 to indicate an 620655714Skris error, but didn't set one). 620755714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 620855714Skris 620955714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 621055714Skris [Ben Laurie] 621155714Skris 621255714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 621355714Skris parameters. This was causing a warning which killed off the Win32 compile. 621455714Skris [Steve Henson] 621555714Skris 621655714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 621755714Skris [Neil Costigan <neil.costigan@celocom.com>] 621855714Skris 621955714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 622055714Skris based on a text string, looking up short and long names and finally 622155714Skris "dot" format. The "dot" format stuff didn't work. Added new function 622255714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 622355714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 622455714Skris OID is not part of the table. 622555714Skris [Steve Henson] 622655714Skris 622755714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 622855714Skris X509_LOOKUP_by_alias(). 622955714Skris [Ben Laurie] 623055714Skris 623155714Skris *) Sort openssl functions by name. 623255714Skris [Ben Laurie] 623355714Skris 623455714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 623555714Skris encryption from sample DSA keys (in case anyone is interested the password 623655714Skris was "1234"). 623755714Skris [Steve Henson] 623855714Skris 623955714Skris *) Make _all_ *_free functions accept a NULL pointer. 624055714Skris [Frans Heymans <fheymans@isaserver.be>] 624155714Skris 624255714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 624355714Skris NULL pointers. 624455714Skris [Anonymous <nobody@replay.com>] 624555714Skris 624655714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 624755714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 624855714Skris 624955714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 625055714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 625155714Skris 625255714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 625355714Skris [Anonymous <nobody@replay.com>] 625455714Skris 625555714Skris *) Add prototype for temp key callback functions 625655714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 625755714Skris [Ben Laurie] 625855714Skris 625955714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 626055714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 626155714Skris [Steve Henson] 626255714Skris 626355714Skris *) X509_name_add_entry() freed the wrong thing after an error. 626455714Skris [Arne Ansper <arne@ats.cyber.ee>] 626555714Skris 626655714Skris *) rsa_eay.c would attempt to free a NULL context. 626755714Skris [Arne Ansper <arne@ats.cyber.ee>] 626855714Skris 626955714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 627055714Skris [Arne Ansper <arne@ats.cyber.ee>] 627155714Skris 627255714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 627355714Skris [Arne Ansper <arne@ats.cyber.ee>] 627455714Skris 627555714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 627655714Skris in X509_STORE_new(), but document the fact that this variable is still 627755714Skris unused in the certificate verification process. 627855714Skris [Ralf S. Engelschall] 627955714Skris 628055714Skris *) Fix the various library and apps files to free up pkeys obtained from 628155714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 628255714Skris [Steve Henson] 628355714Skris 628455714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 628555714Skris demos/maurice/example2.c work, amongst others, probably. 628655714Skris [Steve Henson and Ben Laurie] 628755714Skris 628855714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 628955714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 629055714Skris are no longer created. This way we have a single and consistent command 629155714Skris line interface `openssl <command>', similar to `cvs <command>'. 629255714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 629355714Skris 629455714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 629555714Skris BIT STRING wrapper always have zero unused bits. 629655714Skris [Steve Henson] 629755714Skris 629855714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 629955714Skris [Steve Henson] 630055714Skris 630155714Skris *) Make the top-level INSTALL documentation easier to understand. 630255714Skris [Paul Sutton] 630355714Skris 630455714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 630555714Skris make (including if user presses ^C) [Paul Sutton] 630655714Skris 630755714Skris *) Make Montgomery context stuff explicit in RSA data structure. 630855714Skris [Ben Laurie] 630955714Skris 631055714Skris *) Fix build order of pem and err to allow for generated pem.h. 631155714Skris [Ben Laurie] 631255714Skris 631355714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 631455714Skris [Ben Laurie] 631555714Skris 631655714Skris *) Enhanced the err-ins.pl script so it makes the error library number 631755714Skris global and can add a library name. This is needed for external ASN1 and 631855714Skris other error libraries. 631955714Skris [Steve Henson] 632055714Skris 632155714Skris *) Fixed sk_insert which never worked properly. 632255714Skris [Steve Henson] 632355714Skris 632455714Skris *) Fix ASN1 macros so they can handle indefinite length construted 632555714Skris EXPLICIT tags. Some non standard certificates use these: they can now 632655714Skris be read in. 632755714Skris [Steve Henson] 632855714Skris 632955714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 633055714Skris into a single doc/ssleay.txt bundle. This way the information is still 633155714Skris preserved but no longer messes up this directory. Now it's new room for 633255714Skris the new set of documenation files. 633355714Skris [Ralf S. Engelschall] 633455714Skris 633555714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 633655714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 633755714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 633855714Skris number of arguments. 633955714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 634055714Skris 634155714Skris *) Fix test data to work with the above. 634255714Skris [Ben Laurie] 634355714Skris 634455714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 634555714Skris was already fixed by Eric for 0.9.1 it seems. 634655714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 634755714Skris 634855714Skris *) Autodetect FreeBSD3. 634955714Skris [Ben Laurie] 635055714Skris 635155714Skris *) Fix various bugs in Configure. This affects the following platforms: 635255714Skris nextstep 635355714Skris ncr-scde 635455714Skris unixware-2.0 635555714Skris unixware-2.0-pentium 635655714Skris sco5-cc. 635755714Skris [Ben Laurie] 635855714Skris 635955714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 636055714Skris before they are needed. 636155714Skris [Ben Laurie] 636255714Skris 636355714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 636455714Skris [Ben Laurie] 636555714Skris 636655714Skris 636755714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 636855714Skris 636955714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 637055714Skris changed SSLeay to OpenSSL in version strings. 637155714Skris [Ralf S. Engelschall] 637255714Skris 637355714Skris *) Some fixups to the top-level documents. 637455714Skris [Paul Sutton] 637555714Skris 637655714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 637755714Skris because the symlink to include/ was missing. 637855714Skris [Ralf S. Engelschall] 637955714Skris 638055714Skris *) Incorporated the popular no-RSA/DSA-only patches 638155714Skris which allow to compile a RSA-free SSLeay. 638255714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 638355714Skris 638455714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 638555714Skris when "ssleay" is still not found. 638655714Skris [Ralf S. Engelschall] 638755714Skris 638855714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 638955714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 639055714Skris 639155714Skris *) Updated the README file. 639255714Skris [Ralf S. Engelschall] 639355714Skris 639455714Skris *) Added various .cvsignore files in the CVS repository subdirs 639555714Skris to make a "cvs update" really silent. 639655714Skris [Ralf S. Engelschall] 639755714Skris 639855714Skris *) Recompiled the error-definition header files and added 639955714Skris missing symbols to the Win32 linker tables. 640055714Skris [Ralf S. Engelschall] 640155714Skris 640255714Skris *) Cleaned up the top-level documents; 640355714Skris o new files: CHANGES and LICENSE 640455714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 640555714Skris o merged COPYRIGHT into LICENSE 640655714Skris o removed obsolete TODO file 640755714Skris o renamed MICROSOFT to INSTALL.W32 640855714Skris [Ralf S. Engelschall] 640955714Skris 641055714Skris *) Removed dummy files from the 0.9.1b source tree: 641155714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 641255714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 641355714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 641455714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 641555714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 641655714Skris [Ralf S. Engelschall] 641755714Skris 641855714Skris *) Added various platform portability fixes. 641955714Skris [Mark J. Cox] 642055714Skris 642155714Skris *) The Genesis of the OpenSSL rpject: 642255714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 642355714Skris Young and Tim J. Hudson created while they were working for C2Net until 642455714Skris summer 1998. 642555714Skris [The OpenSSL Project] 642655714Skris 642755714Skris 642855714Skris Changes between 0.9.0b and 0.9.1b [not released] 642955714Skris 643055714Skris *) Updated a few CA certificates under certs/ 643155714Skris [Eric A. Young] 643255714Skris 643355714Skris *) Changed some BIGNUM api stuff. 643455714Skris [Eric A. Young] 643555714Skris 643655714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 643755714Skris DGUX x86, Linux Alpha, etc. 643855714Skris [Eric A. Young] 643955714Skris 644055714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 644155714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 644255714Skris available). 644355714Skris [Eric A. Young] 644455714Skris 644555714Skris *) Add -strparse option to asn1pars program which parses nested 644655714Skris binary structures 644755714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 644855714Skris 644955714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 645055714Skris [Eric A. Young] 645155714Skris 645255714Skris *) DSA fix for "ca" program. 645355714Skris [Eric A. Young] 645455714Skris 645555714Skris *) Added "-genkey" option to "dsaparam" program. 645655714Skris [Eric A. Young] 645755714Skris 645855714Skris *) Added RIPE MD160 (rmd160) message digest. 645955714Skris [Eric A. Young] 646055714Skris 646155714Skris *) Added -a (all) option to "ssleay version" command. 646255714Skris [Eric A. Young] 646355714Skris 646455714Skris *) Added PLATFORM define which is the id given to Configure. 646555714Skris [Eric A. Young] 646655714Skris 646755714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 646855714Skris [Eric A. Young] 646955714Skris 647055714Skris *) Extended the ASN.1 parser routines. 647155714Skris [Eric A. Young] 647255714Skris 647355714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 647455714Skris [Eric A. Young] 647555714Skris 647655714Skris *) Added a BN_CTX to the BN library. 647755714Skris [Eric A. Young] 647855714Skris 647955714Skris *) Fixed the weak key values in DES library 648055714Skris [Eric A. Young] 648155714Skris 648255714Skris *) Changed API in EVP library for cipher aliases. 648355714Skris [Eric A. Young] 648455714Skris 648555714Skris *) Added support for RC2/64bit cipher. 648655714Skris [Eric A. Young] 648755714Skris 648855714Skris *) Converted the lhash library to the crypto/mem.c functions. 648955714Skris [Eric A. Young] 649055714Skris 649155714Skris *) Added more recognized ASN.1 object ids. 649255714Skris [Eric A. Young] 649355714Skris 649455714Skris *) Added more RSA padding checks for SSL/TLS. 649555714Skris [Eric A. Young] 649655714Skris 649755714Skris *) Added BIO proxy/filter functionality. 649855714Skris [Eric A. Young] 649955714Skris 650055714Skris *) Added extra_certs to SSL_CTX which can be used 650155714Skris send extra CA certificates to the client in the CA cert chain sending 650255714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 650355714Skris [Eric A. Young] 650455714Skris 650555714Skris *) Now Fortezza is denied in the authentication phase because 650655714Skris this is key exchange mechanism is not supported by SSLeay at all. 650755714Skris [Eric A. Young] 650855714Skris 650955714Skris *) Additional PKCS1 checks. 651055714Skris [Eric A. Young] 651155714Skris 651255714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 651355714Skris [Eric A. Young] 651455714Skris 651555714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 651655714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 651755714Skris [Eric A. Young] 651855714Skris 651955714Skris *) Fixed a few memory leaks. 652055714Skris [Eric A. Young] 652155714Skris 652255714Skris *) Fixed various code and comment typos. 652355714Skris [Eric A. Young] 652455714Skris 652555714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 652655714Skris bytes sent in the client random. 652755714Skris [Edward Bishop <ebishop@spyglass.com>] 652855714Skris 6529