CHANGES revision 120631
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 6120631Snectar 7120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 8120631Snectar 9120631Snectar Stop out of bounds reads in the ASN1 code when presented with 10120631Snectar invalid tags (CAN-2003-0543 and CAN-2003-0544). 11120631Snectar 12120631Snectar Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545). 13120631Snectar 14120631Snectar If verify callback ignores invalid public key errors don't try to check 15120631Snectar certificate signature with the NULL public key. 16120631Snectar 17120631Snectar [Steve Henson] 18120631Snectar 19120631Snectar *) New -ignore_err option in ocsp application to stop the server 20120631Snectar exiting on the first error in a request. 21120631Snectar [Steve Henson] 22120631Snectar 23120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 24120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 25120631Snectar specifications. 26120631Snectar [Steve Henson] 27120631Snectar 28120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 29120631Snectar extra data after the compression methods not only for TLS 1.0 30120631Snectar but also for SSL 3.0 (as required by the specification). 31120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 32120631Snectar 33120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 34120631Snectar when it's 512 *bits* long, not 512 bytes. 35120631Snectar [Richard Levitte] 36120631Snectar 37120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 38120631Snectar blocks during encryption. 39120631Snectar [Richard Levitte] 40120631Snectar 41120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 42120631Snectar flushes were not handled properly if the BIO retried. On read 43120631Snectar data was not being buffered properly and had various logic bugs. 44120631Snectar This also affects blocking I/O when the data being decoded is a 45120631Snectar certain size. 46120631Snectar [Steve Henson] 47120631Snectar 48120631Snectar *) Various S/MIME bugfixes and compatibility changes: 49120631Snectar output correct application/pkcs7 MIME type if 50120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 51120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 52120631Snectar of files as .eml work). Correctly handle very long lines in MIME 53120631Snectar parser. 54120631Snectar [Steve Henson] 55120631Snectar 56120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 57120631Snectar 58120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 59120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 60120631Snectar a protocol version number mismatch like a decryption error 61120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 62120631Snectar [Bodo Moeller] 63120631Snectar 64120631Snectar *) Turn on RSA blinding by default in the default implementation 65120631Snectar to avoid a timing attack. Applications that don't want it can call 66120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 67120631Snectar They would be ill-advised to do so in most cases. 68120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 69120631Snectar 70120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 71120631Snectar seeded (in this case, the secret RSA exponent is abused as 72120631Snectar an unpredictable seed -- if it is not unpredictable, there 73120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 74120631Snectar by remembering the creator's thread ID in rsa->blinding and 75120631Snectar having all other threads use local one-time blinding factors 76120631Snectar (this requires more computation than sharing rsa->blinding, but 77120631Snectar avoids excessive locking; and if an RSA object is not shared 78120631Snectar between threads, blinding will still be very fast). 79120631Snectar [Bodo Moeller] 80120631Snectar 81120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 82120631Snectar ENGINE as defaults for all supported algorithms irrespective of 83120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 84120631Snectar should make sure they are passing it correctly. 85120631Snectar [Geoff Thorpe] 86120631Snectar 87120631Snectar *) Target "mingw" now allows native Windows code to be generated in 88120631Snectar the Cygwin environment as well as with the MinGW compiler. 89120631Snectar [Ulf Moeller] 90120631Snectar 91111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 92111147Snectar 93111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 94111147Snectar via timing by performing a MAC computation even if incorrrect 95111147Snectar block cipher padding has been found. This is a countermeasure 96111147Snectar against active attacks where the attacker has to distinguish 97111147Snectar between bad padding and a MAC verification error. (CAN-2003-0078) 98111147Snectar 99111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 100111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 101111147Snectar Martin Vuagnoux (EPFL, Ilion)] 102111147Snectar 103111147Snectar *) Make the no-err option work as intended. The intention with no-err 104111147Snectar is not to have the whole error stack handling routines removed from 105111147Snectar libcrypto, it's only intended to remove all the function name and 106111147Snectar reason texts, thereby removing some of the footprint that may not 107111147Snectar be interesting if those errors aren't displayed anyway. 108111147Snectar 109111147Snectar NOTE: it's still possible for any application or module to have it's 110111147Snectar own set of error texts inserted. The routines are there, just not 111111147Snectar used by default when no-err is given. 112111147Snectar [Richard Levitte] 113111147Snectar 114111147Snectar *) Add support for FreeBSD on IA64. 115111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 116111147Snectar 117111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 118111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 119111147Snectar the value returned by DES_cbc_cksum() was like the one from 120111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 121111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 122111147Snectar 123111147Snectar *) Allow an application to disable the automatic SSL chain building. 124111147Snectar Before this a rather primitive chain build was always performed in 125111147Snectar ssl3_output_cert_chain(): an application had no way to send the 126111147Snectar correct chain if the automatic operation produced an incorrect result. 127111147Snectar 128111147Snectar Now the chain builder is disabled if either: 129111147Snectar 130111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 131111147Snectar 132111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 133111147Snectar 134111147Snectar The reasoning behind this is that an application would not want the 135111147Snectar auto chain building to take place if extra chain certificates are 136111147Snectar present and it might also want a means of sending no additional 137111147Snectar certificates (for example the chain has two certificates and the 138111147Snectar root is omitted). 139111147Snectar [Steve Henson] 140111147Snectar 141111147Snectar *) Add the possibility to build without the ENGINE framework. 142111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 143111147Snectar 144111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 145111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 146111147Snectar [Steve Henson] 147111147Snectar 148111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 149111147Snectar could be freed. Solution: make sure initialization is performed early 150111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 151111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 152111147Snectar [Lutz Jaenicke] 153111147Snectar 154111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 155111147Snectar checked on reconnect on the client side, therefore session resumption 156111147Snectar could still fail with a "ssl session id is different" error. This 157111147Snectar behaviour is masked when SSL_OP_ALL is used due to 158111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 159111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 160111147Snectar followup to PR #377. 161111147Snectar [Lutz Jaenicke] 162111147Snectar 163111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 164111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 165111147Snectar [Andy Polyakov] 166111147Snectar 167111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 168111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 169111147Snectar the config script, much like the NetBSD support. 170111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 171111147Snectar 172109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 173109998Smarkm 174120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 175120631Snectar OpenSSL 0.9.7.] 176120631Snectar 177109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 178109998Smarkm code (06) was taken as the first octet of the session ID and the last 179109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 180109998Smarkm caching could not have worked due to the session ID mismatch between 181109998Smarkm client and server. 182109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 183109998Smarkm PR #377. 184109998Smarkm [Lutz Jaenicke] 185109998Smarkm 186109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 187109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 188109998Smarkm removed entirely. 189109998Smarkm [Richard Levitte] 190109998Smarkm 191109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 192109998Smarkm seems that in spite of existing for more than a year, many application 193109998Smarkm author have done nothing to provide the necessary callbacks, which 194109998Smarkm means that this particular engine will not work properly anywhere. 195109998Smarkm This is a very unfortunate situation which forces us, in the name 196109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 197109998Smarkm of libcrypto. 198109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 199109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 200109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 201109998Smarkm make such changes in the libcrypto locking code that changes will 202109998Smarkm have to be made anyway). 203109998Smarkm [Richard Levitte] 204109998Smarkm 205109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 206109998Smarkm octets have been read, EOF or an error occurs. Without this change 207109998Smarkm some truncated ASN1 structures will not produce an error. 208109998Smarkm [Steve Henson] 209109998Smarkm 210109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 211109998Smarkm Still give the possibility to force the use of Heimdal, but with 212109998Smarkm warnings and a request that patches get sent to openssl-dev. 213109998Smarkm [Richard Levitte] 214109998Smarkm 215109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 216109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 217109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 218109998Smarkm 219109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 220109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 221109998Smarkm edit numbers of the version. 222109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 223109998Smarkm 224109998Smarkm *) Introduce safe string copy and catenation functions 225109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 226109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 227109998Smarkm 228109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 229109998Smarkm [Ben Laurie (CHATS)] 230109998Smarkm 231109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 232109998Smarkm resizing buffers containing secrets, and use where appropriate. 233109998Smarkm [Ben Laurie (CHATS)] 234109998Smarkm 235109998Smarkm *) Avoid using fixed size buffers for configuration file location. 236109998Smarkm [Ben Laurie (CHATS)] 237109998Smarkm 238109998Smarkm *) Avoid filename truncation for various CA files. 239109998Smarkm [Ben Laurie (CHATS)] 240109998Smarkm 241109998Smarkm *) Use sizeof in preference to magic numbers. 242109998Smarkm [Ben Laurie (CHATS)] 243109998Smarkm 244109998Smarkm *) Avoid filename truncation in cert requests. 245109998Smarkm [Ben Laurie (CHATS)] 246109998Smarkm 247109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 248109998Smarkm overflows. 249109998Smarkm [Ben Laurie (CHATS)] 250109998Smarkm 251109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 252109998Smarkm potentially lead to a spoofing attack). 253109998Smarkm [Ben Laurie (CHATS)] 254109998Smarkm 255109998Smarkm *) Fix various buffers to be large enough for hex/decimal 256109998Smarkm representations in a platform independent manner. 257109998Smarkm [Ben Laurie (CHATS)] 258109998Smarkm 259109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 260109998Smarkm resizing buffers containing secrets, and use where appropriate. 261109998Smarkm [Ben Laurie (CHATS)] 262109998Smarkm 263109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 264109998Smarkm indents. 265109998Smarkm [Ben Laurie (CHATS)] 266109998Smarkm 267109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 268109998Smarkm [Ben Laurie (CHATS)] 269109998Smarkm 270109998Smarkm *) buffer_gets() could terminate with the buffer only half 271109998Smarkm full. Fixed. 272109998Smarkm [Ben Laurie (CHATS)] 273109998Smarkm 274109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 275109998Smarkm overflowing internal buffers by having large block sizes, etc. 276109998Smarkm [Ben Laurie (CHATS)] 277109998Smarkm 278109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 279109998Smarkm unconditionally). 280109998Smarkm [Ben Laurie (CHATS)] 281109998Smarkm 282109998Smarkm *) Eliminate unused copy of key in RC4. 283109998Smarkm [Ben Laurie (CHATS)] 284109998Smarkm 285109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 286109998Smarkm [Ben Laurie (CHATS)] 287109998Smarkm 288109998Smarkm *) Fix off-by-one error in EGD path. 289109998Smarkm [Ben Laurie (CHATS)] 290109998Smarkm 291109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 292109998Smarkm [Ben Laurie (CHATS)] 293109998Smarkm 294109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 295109998Smarkm CBCParameter. 296109998Smarkm [Ben Laurie (CHATS)] 297109998Smarkm 298109998Smarkm *) Eliminate unused and dangerous function knumber(). 299109998Smarkm [Ben Laurie (CHATS)] 300109998Smarkm 301109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 302109998Smarkm [Ben Laurie (CHATS)] 303109998Smarkm 304109998Smarkm *) Protect against overlong session ID context length in an encoded 305109998Smarkm session object. Since these are local, this does not appear to be 306109998Smarkm exploitable. 307109998Smarkm [Ben Laurie (CHATS)] 308109998Smarkm 309109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 310109998Smarkm the 0.9.6 release series: 311109998Smarkm 312109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 313109998Smarkm supply an oversized master key in Kerberos-enabled versions. 314109998Smarkm (CAN-2002-0657) 315109998Smarkm [Ben Laurie (CHATS)] 316109998Smarkm 317109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 318109998Smarkm [Richard Levitte] 319109998Smarkm 320109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 321109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 322109998Smarkm 323109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 324109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 325109998Smarkm 326109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 327109998Smarkm have been removed entirely. This was also the last step to make 328109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 329109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 330109998Smarkm 331109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 332109998Smarkm to allow version independent disabling of normally unselected ciphers, 333109998Smarkm which may be activated as a side-effect of selecting a single cipher. 334109998Smarkm 335109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 336109998Smarkm out of "ALL" because they do not provide symmetric encryption. 337109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 338109998Smarkm [Lutz Jaenicke, Bodo Moeller] 339109998Smarkm 340109998Smarkm *) Add appropriate support for separate platform-dependent build 341109998Smarkm directories. The recommended way to make a platform-dependent 342109998Smarkm build directory is the following (tested on Linux), maybe with 343109998Smarkm some local tweaks: 344109998Smarkm 345109998Smarkm # Place yourself outside of the OpenSSL source tree. In 346109998Smarkm # this example, the environment variable OPENSSL_SOURCE 347109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 348109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 349109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 350111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 351109998Smarkm mkdir -p `dirname $F` 352109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 353109998Smarkm done 354109998Smarkm 355109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 356109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 357109998Smarkm it probably means the source directory is very clean. 358109998Smarkm [Richard Levitte] 359109998Smarkm 360109998Smarkm *) Make sure any ENGINE control commands make local copies of string 361109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 362109998Smarkm the caller may have overwritten (or deallocated) the original string 363109998Smarkm data when a later ENGINE operation tries to use the stored values. 364109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 365109998Smarkm 366109998Smarkm *) Improve diagnostics in file reading and command-line digests. 367109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 368109998Smarkm 369109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 370109998Smarkm error in AES-CFB decryption. 371109998Smarkm [Richard Levitte] 372109998Smarkm 373109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 374109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 375109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 376109998Smarkm BIOs and some applications. This has the side effect that 377109998Smarkm applications must explicitly clean up cipher contexts with 378109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 379109998Smarkm [Steve Henson] 380109998Smarkm 381109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 382109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 383109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 384109998Smarkm [Steve Henson] 385109998Smarkm 386109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 387109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 388109998Smarkm [Lutz Jaenicke] 389109998Smarkm 390109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 391109998Smarkm form for "surname", serialNumber has no short form. 392109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 393109998Smarkm therefore remove "mail" short name for "internet 7". 394109998Smarkm The OID for unique identifiers in X509 certificates is 395109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 396109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 397109998Smarkm [Lutz Jaenicke] 398109998Smarkm 399109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 400109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 401109998Smarkm after all ctrl commands have been executed on it. If init=1 the 402109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 403109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 404109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 405109998Smarkm [Steve Henson] 406109998Smarkm 407109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 408109998Smarkm argument is actually passed to the callback: In the 409109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 410109998Smarkm declaration has been changed from 411109998Smarkm int (*cb)() 412109998Smarkm into 413109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 414109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 415109998Smarkm i=s->ctx->app_verify_callback(&ctx) 416109998Smarkm has been changed into 417109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 418109998Smarkm 419109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 420109998Smarkm a dummy argument can be added to their callback functions. 421109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 422109998Smarkm 423109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 424109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 425109998Smarkm 426109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 427109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 428109998Smarkm This allows older applications to transparently support certain 429109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 430109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 431109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 432109998Smarkm always load it have also been added. 433109998Smarkm [Steve Henson] 434109998Smarkm 435109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 436109998Smarkm Adjust NIDs and EVP layer. 437109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 438109998Smarkm 439109998Smarkm *) Config modules support in openssl utility. 440109998Smarkm 441109998Smarkm Most commands now load modules from the config file, 442109998Smarkm though in a few (such as version) this isn't done 443109998Smarkm because it couldn't be used for anything. 444109998Smarkm 445109998Smarkm In the case of ca and req the config file used is 446109998Smarkm the same as the utility itself: that is the -config 447109998Smarkm command line option can be used to specify an 448109998Smarkm alternative file. 449109998Smarkm [Steve Henson] 450109998Smarkm 451109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 452109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 453109998Smarkm [Steve Henson] 454109998Smarkm 455109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 456109998Smarkm config section name. Add a new flag to tolerate a missing config file 457109998Smarkm and move code to CONF_modules_load_file(). 458109998Smarkm [Steve Henson] 459109998Smarkm 460109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 461109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 462109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 463109998Smarkm to work with the new engine framework. 464109998Smarkm [AEP Inc. and Richard Levitte] 465109998Smarkm 466109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 467109998Smarkm Technologies. (Use engine 'sureware') 468109998Smarkm The support was copied from 0.9.6c [engine] and adapted 469109998Smarkm to work with the new engine framework. 470109998Smarkm [Richard Levitte] 471109998Smarkm 472109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 473109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 474109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 475109998Smarkm 476109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 477109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 478109998Smarkm 479109998Smarkm *) Add the configuration target debug-linux-ppro. 480109998Smarkm Make 'openssl rsa' use the general key loading routines 481109998Smarkm implemented in apps.c, and make those routines able to 482109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 483109998Smarkm FORMAT_IISSGC. 484109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 485109998Smarkm 486109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 487109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 488109998Smarkm 489109998Smarkm *) Add -keyform to rsautl, and document -engine. 490109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 491109998Smarkm 492109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 493109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 494109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 495109998Smarkm [Ben Laurie] 496109998Smarkm 497109998Smarkm *) Add new functions 498109998Smarkm ERR_peek_last_error 499109998Smarkm ERR_peek_last_error_line 500109998Smarkm ERR_peek_last_error_line_data. 501109998Smarkm These are similar to 502109998Smarkm ERR_peek_error 503109998Smarkm ERR_peek_error_line 504109998Smarkm ERR_peek_error_line_data, 505109998Smarkm but report on the latest error recorded rather than the first one 506109998Smarkm still in the error queue. 507109998Smarkm [Ben Laurie, Bodo Moeller] 508109998Smarkm 509109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 510109998Smarkm like: 511109998Smarkm default_algorithms = ALL 512109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 513109998Smarkm [Steve Henson] 514109998Smarkm 515109998Smarkm *) Prelminary ENGINE config module. 516109998Smarkm [Steve Henson] 517109998Smarkm 518109998Smarkm *) New experimental application configuration code. 519109998Smarkm [Steve Henson] 520109998Smarkm 521109998Smarkm *) Change the AES code to follow the same name structure as all other 522109998Smarkm symmetric ciphers, and behave the same way. Move everything to 523109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 524109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 525109998Smarkm 526109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 527109998Smarkm [Ben Laurie and Theo de Raadt] 528109998Smarkm 529109998Smarkm *) Add option to output public keys in req command. 530109998Smarkm [Massimiliano Pala madwolf@openca.org] 531109998Smarkm 532109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 533109998Smarkm (up to about 10% better than before for P-192 and P-224). 534109998Smarkm [Bodo Moeller] 535109998Smarkm 536109998Smarkm *) New functions/macros 537109998Smarkm 538109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 539109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 540109998Smarkm SSL_set_msg_callback(ssl, cb) 541109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 542109998Smarkm 543109998Smarkm to request calling a callback function 544109998Smarkm 545109998Smarkm void cb(int write_p, int version, int content_type, 546109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 547109998Smarkm 548109998Smarkm whenever a protocol message has been completely received 549109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 550109998Smarkm protocol version according to which the SSL library interprets 551109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 552109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 553109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 554109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 555109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 556109998Smarkm SSL object, and 'arg' is the application-defined value set by 557109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 558109998Smarkm 559109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 560109998Smarkm to enable a callback that displays all protocol messages. 561109998Smarkm [Bodo Moeller] 562109998Smarkm 563109998Smarkm *) Change the shared library support so shared libraries are built as 564109998Smarkm soon as the corresponding static library is finished, and thereby get 565109998Smarkm openssl and the test programs linked against the shared library. 566109998Smarkm This still only happens when the keyword "shard" has been given to 567109998Smarkm the configuration scripts. 568109998Smarkm 569109998Smarkm NOTE: shared library support is still an experimental thing, and 570109998Smarkm backward binary compatibility is still not guaranteed. 571109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 572109998Smarkm 573109998Smarkm *) Add support for Subject Information Access extension. 574109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 575109998Smarkm 576109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 577109998Smarkm additional bytes when new memory had to be allocated, not just 578109998Smarkm when reusing an existing buffer. 579109998Smarkm [Bodo Moeller] 580109998Smarkm 581109998Smarkm *) New command line and configuration option 'utf8' for the req command. 582109998Smarkm This allows field values to be specified as UTF8 strings. 583109998Smarkm [Steve Henson] 584109998Smarkm 585109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 586109998Smarkm runs for the former and machine-readable output for the latter. 587109998Smarkm [Ben Laurie] 588109998Smarkm 589109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 590109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 591109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 592109998Smarkm has the same effect. 593109998Smarkm [Massimiliano Pala madwolf@openca.org] 594109998Smarkm 595109998Smarkm *) Change all functions with names starting with des_ to be starting 596109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 597109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 598109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 599109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 600109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 601109998Smarkm exception. 602109998Smarkm 603109998Smarkm Since we provide two compatibility mappings, the user needs to 604109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 605109998Smarkm compatibility is desired. The default (i.e., when that macro 606109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 607109998Smarkm 608109998Smarkm There are also macros that enable and disable the support of old 609109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 610109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 611109998Smarkm are defined, the default will apply: to support the old des routines. 612109998Smarkm 613109998Smarkm In either case, one must include openssl/des.h to get the correct 614109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 615109998Smarkm won't work. 616109998Smarkm 617109998Smarkm NOTE: This is a major break of an old API into a new one. Software 618109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 619109998Smarkm time in the future, des_old.h and the libdes compatibility functions 620109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 621109998Smarkm default), and then completely removed. 622109998Smarkm [Richard Levitte] 623109998Smarkm 624109998Smarkm *) Test for certificates which contain unsupported critical extensions. 625109998Smarkm If such a certificate is found during a verify operation it is 626109998Smarkm rejected by default: this behaviour can be overridden by either 627109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 628109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 629109998Smarkm X509_supported_extension() has also been added which returns 1 if a 630109998Smarkm particular extension is supported. 631109998Smarkm [Steve Henson] 632109998Smarkm 633109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 634109998Smarkm to retain compatibility with existing code. 635109998Smarkm [Steve Henson] 636109998Smarkm 637109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 638109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 639109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 640109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 641109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 642109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 643109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 644109998Smarkm requires the destination to be valid. 645109998Smarkm 646109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 647109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 648109998Smarkm [Steve Henson] 649109998Smarkm 650109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 651109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 652109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 653109998Smarkm [Bodo Moeller] 654109998Smarkm 655109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 656109998Smarkm [Massimo Santin via Richard Levitte] 657109998Smarkm 658109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 659109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 660109998Smarkm (initialisation, etc) from functionality dealing with implementations 661109998Smarkm of specific crypto iterfaces. This change also introduces integrated 662109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 663109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 664109998Smarkm implementations of their own. This is detailed in crypto/engine/README 665109998Smarkm as it couldn't be adequately described here. However, there are a few 666109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 667109998Smarkm were changed in the original introduction of ENGINE code have now 668109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 669109998Smarkm deal more passive and at run-time, operations deal directly with 670109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 671109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 672109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 673109998Smarkm they were not being used by the framework as there is no concept of a 674109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 675109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 676109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 677109998Smarkm the new code. 678109998Smarkm [Geoff Thorpe] 679109998Smarkm 680109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 681109998Smarkm [Steve Henson] 682109998Smarkm 683109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 684109998Smarkm and make sure the automatically generated functions ERR_load_* 685109998Smarkm become part of libeay.num as well. 686109998Smarkm [Richard Levitte] 687109998Smarkm 688109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 689109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 690109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 691109998Smarkm false once a handshake has been completed. 692109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 693109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 694109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 695109998Smarkm client has followed the request.) 696109998Smarkm [Bodo Moeller] 697109998Smarkm 698109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 699109998Smarkm By default, clients may request session resumption even during 700109998Smarkm renegotiation (if session ID contexts permit); with this option, 701109998Smarkm session resumption is possible only in the first handshake. 702109998Smarkm 703109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 704109998Smarkm more bits available for options that should not be part of 705109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 706109998Smarkm [Bodo Moeller] 707109998Smarkm 708109998Smarkm *) Add some demos for certificate and certificate request creation. 709109998Smarkm [Steve Henson] 710109998Smarkm 711109998Smarkm *) Make maximum certificate chain size accepted from the peer application 712109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 713109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 714109998Smarkm [Lutz Jaenicke] 715109998Smarkm 716109998Smarkm *) Add support for shared libraries for Unixware-7 717109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 718109998Smarkm [Lutz Jaenicke] 719109998Smarkm 720109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 721109998Smarkm be done prior to destruction. Use this to unload error strings from 722109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 723109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 724109998Smarkm [Geoff Thorpe] 725109998Smarkm 726109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 727109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 728109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 729109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 730109998Smarkm Also, add stub code to each that makes building them as self-contained 731109998Smarkm shared-libraries easier (see README.ENGINE). 732109998Smarkm [Geoff Thorpe] 733109998Smarkm 734109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 735109998Smarkm implementations into applications that are completely implemented in 736109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 737109998Smarkm commands that can be used to configure what shared-library to load and 738109998Smarkm to control aspects of the way it is handled. Also, made an update to 739109998Smarkm the README.ENGINE file that brings its information up-to-date and 740109998Smarkm provides some information and instructions on the "dynamic" ENGINE 741109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 742109998Smarkm [Geoff Thorpe] 743109998Smarkm 744109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 745109998Smarkm "ERR_unload_strings" function. 746109998Smarkm [Geoff Thorpe] 747109998Smarkm 748109998Smarkm *) Add a copy() function to EVP_MD. 749109998Smarkm [Ben Laurie] 750109998Smarkm 751109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 752109998Smarkm md_data void pointer. 753109998Smarkm [Ben Laurie] 754109998Smarkm 755109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 756109998Smarkm that the digest can only process a single chunk of data 757109998Smarkm (typically because it is provided by a piece of 758109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 759109998Smarkm is only going to provide a single chunk of data, and hence the 760109998Smarkm framework needn't accumulate the data for oneshot drivers. 761109998Smarkm [Ben Laurie] 762109998Smarkm 763109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 764109998Smarkm functions. This change also alters the storage and management of global 765109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 766109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 767109998Smarkm index counters. The API functions that use this state have been changed 768109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 769109998Smarkm and counter, and there is now an API function to dynamically create new 770109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 771109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 772109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 773109998Smarkm such data would previously have always leaked in application code and 774109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 775109998Smarkm to it. Application code that doesn't use this new function will still 776109998Smarkm leak as before, but their memory debugging output will announce it now 777109998Smarkm rather than letting it slide. 778109998Smarkm 779109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 780109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 781109998Smarkm has a return value to indicate success or failure. 782109998Smarkm [Geoff Thorpe] 783109998Smarkm 784109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 785109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 786109998Smarkm implementation. This change also adds two functions to "get" and "set" 787109998Smarkm the implementation prior to it being automatically set the first time 788109998Smarkm any other ERR function takes place. Ie. an application can call "get", 789109998Smarkm pass the return value to a module it has just loaded, and that module 790109998Smarkm can call its own "set" function using that value. This means the 791109998Smarkm module's "ERR" operations will use (and modify) the error state in the 792109998Smarkm application and not in its own statically linked copy of OpenSSL code. 793109998Smarkm [Geoff Thorpe] 794109998Smarkm 795109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 796109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 797109998Smarkm the operation, and provides a more encapsulated way for external code 798109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 799109998Smarkm to use these functions rather than manually incrementing the counts. 800109998Smarkm 801109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 802109998Smarkm [Geoff Thorpe] 803109998Smarkm 804109998Smarkm *) Add EVP test program. 805109998Smarkm [Ben Laurie] 806109998Smarkm 807109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 808109998Smarkm [Ben Laurie] 809109998Smarkm 810109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 811109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 812109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 813109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 814109998Smarkm directly. Modify 'ca' application to use new functions. 815109998Smarkm [Steve Henson] 816109998Smarkm 817109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 818109998Smarkm bug workarounds. Rollback attack detection is a security feature. 819109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 820109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 821109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 822109998Smarkm for their choice and can explicitly enable this option. 823109998Smarkm [Bodo Moeller, Lutz Jaenicke] 824109998Smarkm 825109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 826109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 827109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 828109998Smarkm Usage example: 829109998Smarkm 830109998Smarkm EVP_MD_CTX md; 831109998Smarkm 832109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 833109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 834109998Smarkm EVP_DigestUpdate(&md, in, len); 835109998Smarkm EVP_DigestFinal(&md, out, NULL); 836109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 837109998Smarkm 838109998Smarkm [Ben Laurie] 839109998Smarkm 840109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 841109998Smarkm correcting its structure. This means that calls to DES functions 842109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 843109998Smarkm plain des_key_schedule (which was actually always a pointer 844109998Smarkm anyway): E.g., 845109998Smarkm 846109998Smarkm des_key_schedule ks; 847109998Smarkm 848109998Smarkm des_set_key_checked(..., &ks); 849109998Smarkm des_ncbc_encrypt(..., &ks, ...); 850109998Smarkm 851109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 852109998Smarkm [Ben Laurie] 853109998Smarkm 854109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 855109998Smarkm PEM causes large amounts of unused functions to be linked in due to 856109998Smarkm poor organisation. For example pem_all.c contains every PEM function 857109998Smarkm which has a knock on effect of linking in large amounts of (unused) 858109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 859109998Smarkm functions prevents this. 860109998Smarkm [Steve Henson] 861109998Smarkm 862109998Smarkm *) Cleanup of EVP macros. 863109998Smarkm [Ben Laurie] 864109998Smarkm 865109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 866109998Smarkm correct _ecb suffix. 867109998Smarkm [Ben Laurie] 868109998Smarkm 869109998Smarkm *) Add initial OCSP responder support to ocsp application. The 870109998Smarkm revocation information is handled using the text based index 871109998Smarkm use by the ca application. The responder can either handle 872109998Smarkm requests generated internally, supplied in files (for example 873109998Smarkm via a CGI script) or using an internal minimal server. 874109998Smarkm [Steve Henson] 875109998Smarkm 876109998Smarkm *) Add configuration choices to get zlib compression for TLS. 877109998Smarkm [Richard Levitte] 878109998Smarkm 879109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 880109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 881109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 882109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 883109998Smarkm 884109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 885109998Smarkm and authenticator structs; see crypto/krb5/. 886109998Smarkm 887109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 888109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 889109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 890109998Smarkm via Richard Levitte] 891109998Smarkm 892109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 893109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 894109998Smarkm values for each of the key sizes rather than having just 895109998Smarkm parameters (and 'speed' generating keys each time). 896109998Smarkm [Geoff Thorpe] 897109998Smarkm 898109998Smarkm *) Speed up EVP routines. 899109998Smarkm Before: 900109998Smarkmencrypt 901109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 902109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 903109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 904109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 905109998Smarkmdecrypt 906109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 907109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 908109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 909109998Smarkm After: 910109998Smarkmencrypt 911109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 912109998Smarkmdecrypt 913109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 914109998Smarkm [Ben Laurie] 915109998Smarkm 916109998Smarkm *) Added the OS2-EMX target. 917109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 918109998Smarkm 919109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 920109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 921109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 922109998Smarkm structure: this means that the old CONF compatible routines can be 923109998Smarkm retained (in particular wrt extensions) without having to duplicate the 924109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 925109998Smarkm [Steve Henson] 926109998Smarkm 927109998Smarkm *) Enhance the general user interface with mechanisms for inner control 928109998Smarkm and with possibilities to have yes/no kind of prompts. 929109998Smarkm [Richard Levitte] 930109998Smarkm 931109998Smarkm *) Change all calls to low level digest routines in the library and 932109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 933109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 934109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 935109998Smarkm 936109998Smarkm *) Add the possibility to control engines through control names but with 937109998Smarkm arbitrary arguments instead of just a string. 938109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 939109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 940109998Smarkm versions of OpenSSL [engine]. 941109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 942109998Smarkm callback. 943109998Smarkm [Richard Levitte] 944109998Smarkm 945109998Smarkm *) Enhance the general user interface with mechanisms to better support 946109998Smarkm dialog box interfaces, application-defined prompts, the possibility 947109998Smarkm to use defaults (for example default passwords from somewhere else) 948109998Smarkm and interrupts/cancellations. 949109998Smarkm [Richard Levitte] 950109998Smarkm 951109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 952109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 953109998Smarkm [Steve Henson] 954109998Smarkm 955109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 956109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 957109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 958109998Smarkm 959109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 960109998Smarkm callback (pem_password_cb) as all other routines that need this 961109998Smarkm kind of callback. 962109998Smarkm [Richard Levitte] 963109998Smarkm 964109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 965109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 966109998Smarkm than this minimum value is recommended. 967109998Smarkm [Lutz Jaenicke] 968109998Smarkm 969109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 970109998Smarkm that are easily reachable. 971109998Smarkm [Richard Levitte] 972109998Smarkm 973109998Smarkm *) Windows apparently can't transparently handle global 974109998Smarkm variables defined in DLLs. Initialisations such as: 975109998Smarkm 976109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 977109998Smarkm 978109998Smarkm wont compile. This is used by the any applications that need to 979109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 980109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 981109998Smarkm needed for static libraries under Win32. 982109998Smarkm [Steve Henson] 983109998Smarkm 984109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 985109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 986109998Smarkm purpose functions and tidy up setting in other SSL functions. 987109998Smarkm [Steve Henson] 988109998Smarkm 989109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 990109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 991109998Smarkm initialised. This allows various defaults to be set in the 992109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 993109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 994109998Smarkm internally such as S/MIME. 995109998Smarkm 996109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 997109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 998109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 999109998Smarkm 1000109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 1001109998Smarkm applications. 1002109998Smarkm [Steve Henson] 1003109998Smarkm 1004109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 1005109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 1006109998Smarkm its validity and signature checked, then if the certificate is found 1007109998Smarkm in the CRL the verify fails with a revoked error. 1008109998Smarkm 1009109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 1010109998Smarkm 1011109998Smarkm Command line options added to 'verify' application to support this. 1012109998Smarkm 1013109998Smarkm This needs some additional work, such as being able to handle multiple 1014109998Smarkm CRLs with different times, extension based lookup (rather than just 1015109998Smarkm by subject name) and ultimately more complete V2 CRL extension 1016109998Smarkm handling. 1017109998Smarkm [Steve Henson] 1018109998Smarkm 1019109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 1020109998Smarkm to replace things like des_read_password and friends (backward 1021109998Smarkm compatibility functions using this new API are provided). 1022109998Smarkm The purpose is to remove prompting functions from the DES code 1023109998Smarkm section as well as provide for prompting through dialog boxes in 1024109998Smarkm a window system and the like. 1025109998Smarkm [Richard Levitte] 1026109998Smarkm 1027109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 1028109998Smarkm per-structure level rather than having to store it globally. 1029109998Smarkm [Geoff] 1030109998Smarkm 1031109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 1032109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 1033109998Smarkm This causes the "original" ENGINE structure to act like a template, 1034109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 1035109998Smarkm operational state can be localised to each ENGINE structure, despite the 1036109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 1037109998Smarkm this case have no functional references and the return value is the single 1038109998Smarkm structural reference. This matches the single structural reference returned 1039109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 1040109998Smarkm ENGINE structure. 1041109998Smarkm [Geoff] 1042109998Smarkm 1043109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 1044109998Smarkm needs to match any other type at all we need to manually clear the 1045109998Smarkm tag cache. 1046109998Smarkm [Steve Henson] 1047109998Smarkm 1048109998Smarkm *) Changes to the "openssl engine" utility to include; 1049109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 1050109998Smarkm about an ENGINE's available control commands. 1051109998Smarkm - executing control commands from command line arguments using the 1052109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 1053109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 1054109998Smarkm the individual commands are colon-separated, for example; 1055109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 1056109998Smarkm [Geoff] 1057109998Smarkm 1058109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 1059109998Smarkm declare their own commands (numbers), names (strings), descriptions, 1060109998Smarkm and input types for run-time discovery by calling applications. A 1061109998Smarkm subset of these commands are implicitly classed as "executable" 1062109998Smarkm depending on their input type, and only these can be invoked through 1063109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 1064109998Smarkm can be based on user input, config files, etc). The distinction is 1065109998Smarkm that "executable" commands cannot return anything other than a boolean 1066109998Smarkm result and can only support numeric or string input, whereas some 1067109998Smarkm discoverable commands may only be for direct use through 1068109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 1069109998Smarkm pointers, or other custom uses. The "executable" commands are to 1070109998Smarkm support parameterisations of ENGINE behaviour that can be 1071109998Smarkm unambiguously defined by ENGINEs and used consistently across any 1072109998Smarkm OpenSSL-based application. Commands have been added to all the 1073109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 1074109998Smarkm control over shared-library paths without source code alterations. 1075109998Smarkm [Geoff] 1076109998Smarkm 1077109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 1078109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 1079109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 1080109998Smarkm this also allows the implementations to compile without using the 1081109998Smarkm internal engine_int.h header. 1082109998Smarkm [Geoff] 1083109998Smarkm 1084109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 1085109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 1086109998Smarkm should already have non-const pointers to it (ie. they should only 1087109998Smarkm modify their own ones). 1088109998Smarkm [Geoff] 1089109998Smarkm 1090109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 1091109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 1092109998Smarkm to C code. "nuron" string definitions were placed in variables 1093109998Smarkm rather than hard-coded - allowing parameterisation of these values 1094109998Smarkm later on via ctrl() commands. 1095109998Smarkm - Removed unused "#if 0"'d code. 1096109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 1097109998Smarkm structural references. 1098109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 1099109998Smarkm - Constified various get/set functions as appropriate and added 1100109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 1101109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 1102109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 1103109998Smarkm or function to NULL is a way of cancelling out a previously set 1104109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 1105109998Smarkm and doesn't justify the extra error symbols and code. 1106109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 1107109998Smarkm flags from engine_int.h to engine.h. 1108109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 1109109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 1110109998Smarkm [Geoff] 1111109998Smarkm 1112109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 1113109998Smarkm to the algorithm using long division. The binary algorithm can be 1114109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 1115109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 1116109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 1117109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 1118109998Smarkm appears to be advantageous for much longer moduli; here we use it 1119109998Smarkm for moduli up to 2048 bits. 1120109998Smarkm [Bodo Moeller] 1121109998Smarkm 1122109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 1123109998Smarkm could not support the combine flag in choice fields. 1124109998Smarkm [Steve Henson] 1125109998Smarkm 1126109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 1127109998Smarkm extensions from a certificate request to the certificate. 1128109998Smarkm [Steve Henson] 1129109998Smarkm 1130109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 1131109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 1132109998Smarkm file: this allows the display of the certificate about to be 1133109998Smarkm signed to be customised, to allow certain fields to be included 1134109998Smarkm or excluded and extension details. The old system didn't display 1135109998Smarkm multicharacter strings properly, omitted fields not in the policy 1136109998Smarkm and couldn't display additional details such as extensions. 1137109998Smarkm [Steve Henson] 1138109998Smarkm 1139109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 1140109998Smarkm of an arbitrary number of elliptic curve points 1141109998Smarkm \sum scalars[i]*points[i], 1142109998Smarkm optionally including the generator defined for the EC_GROUP: 1143109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 1144109998Smarkm 1145109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 1146109998Smarkm that the point list has just one item (besides the optional 1147109998Smarkm generator). 1148109998Smarkm [Bodo Moeller] 1149109998Smarkm 1150109998Smarkm *) First EC_METHODs for curves over GF(p): 1151109998Smarkm 1152109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 1153109998Smarkm operations and provides various method functions that can also 1154109998Smarkm operate with faster implementations of modular arithmetic. 1155109998Smarkm 1156109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 1157109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 1158109998Smarkm 1159109998Smarkm [Bodo Moeller; point addition and point doubling 1160109998Smarkm implementation directly derived from source code provided by 1161109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 1162109998Smarkm 1163109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 1164109998Smarkm crypto/ec/ec_lib.c): 1165109998Smarkm 1166109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 1167109998Smarkm based on EC_METHODs that are built into the library. 1168109998Smarkm 1169109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 1170109998Smarkm 1171109998Smarkm Most of the framework would be able to handle curves over arbitrary 1172109998Smarkm finite fields, but as there are no obvious types for fields other 1173109998Smarkm than GF(p), some functions are limited to that for now. 1174109998Smarkm [Bodo Moeller] 1175109998Smarkm 1176109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 1177109998Smarkm that the file contains a complete HTTP response. 1178109998Smarkm [Richard Levitte] 1179109998Smarkm 1180109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 1181109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 1182109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 1183109998Smarkm field while the former will cause them to run together if the field 1184109998Smarkm is 40 of more characters long. 1185109998Smarkm [Steve Henson] 1186109998Smarkm 1187109998Smarkm *) Constify the cipher and digest 'method' functions and structures 1188109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 1189109998Smarkm pointers. 1190109998Smarkm [Steve Henson] 1191109998Smarkm 1192109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 1193109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 1194109998Smarkm [Bodo Moeller] 1195109998Smarkm 1196109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 1197109998Smarkm internal software routines can never fail additional hardware versions 1198109998Smarkm might. 1199109998Smarkm [Steve Henson] 1200109998Smarkm 1201109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 1202109998Smarkm 1203109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 1204109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 1205109998Smarkm 1206109998Smarkm ASN1 error codes 1207109998Smarkm ERR_R_NESTED_ASN1_ERROR 1208109998Smarkm ... 1209109998Smarkm ERR_R_MISSING_ASN1_EOS 1210109998Smarkm were 4 .. 9, conflicting with 1211109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 1212109998Smarkm ... 1213109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 1214109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 1215109998Smarkm 1216109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 1217109998Smarkm [Bodo Moeller] 1218109998Smarkm 1219109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 1220109998Smarkm suffices. 1221109998Smarkm [Bodo Moeller] 1222109998Smarkm 1223109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 1224109998Smarkm sets the subject name for a new request or supersedes the 1225109998Smarkm subject name in a given request. Formats that can be parsed are 1226109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 1227109998Smarkm and 1228109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 1229109998Smarkm 1230109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 1231109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 1232109998Smarkm 1233109998Smarkm *) Introduce the possibility to access global variables through 1234109998Smarkm functions on platform were that's the best way to handle exporting 1235109998Smarkm global variables in shared libraries. To enable this functionality, 1236109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 1237109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 1238109998Smarkm is normally done by Configure or something similar). 1239109998Smarkm 1240109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 1241109998Smarkm in the source file (foo.c) like this: 1242109998Smarkm 1243109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 1244109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 1245109998Smarkm 1246109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 1247109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 1248109998Smarkm 1249109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 1250109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 1251109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 1252109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 1253109998Smarkm 1254109998Smarkm The #defines are very important, and therefore so is including the 1255109998Smarkm header file everywhere where the defined globals are used. 1256109998Smarkm 1257109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 1258109998Smarkm of ASN.1 items, but that structure is a bit different. 1259109998Smarkm 1260109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 1261109998Smarkm better and easier to understand logic to choose which symbols should 1262109998Smarkm go into the Windows .def files as well as a number of fixes and code 1263109998Smarkm cleanup (among others, algorithm keywords are now sorted 1264109998Smarkm lexicographically to avoid constant rewrites). 1265109998Smarkm [Richard Levitte] 1266109998Smarkm 1267109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 1268109998Smarkm result to 'rm' because if rm==num the value will be overwritten 1269109998Smarkm and produce the wrong result if 'num' is negative: this caused 1270109998Smarkm problems with BN_mod() and BN_nnmod(). 1271109998Smarkm [Steve Henson] 1272109998Smarkm 1273109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 1274109998Smarkm OCSP request and verifies the signer certificate. The signer 1275109998Smarkm certificate is just checked for a generic purpose and OCSP request 1276109998Smarkm trust settings. 1277109998Smarkm [Steve Henson] 1278109998Smarkm 1279109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 1280109998Smarkm responses. OCSP responses are prepared in real time and may only 1281109998Smarkm be a few seconds old. Simply checking that the current time lies 1282109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 1283109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 1284109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 1285109998Smarkm the current time. The age of the response can also optionally be 1286109998Smarkm checked. Two new options -validity_period and -status_age added to 1287109998Smarkm ocsp utility. 1288109998Smarkm [Steve Henson] 1289109998Smarkm 1290109998Smarkm *) If signature or public key algorithm is unrecognized print out its 1291109998Smarkm OID rather that just UNKNOWN. 1292109998Smarkm [Steve Henson] 1293109998Smarkm 1294109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 1295109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 1296109998Smarkm ID to be generated from the issuer certificate alone which can then be 1297109998Smarkm passed to OCSP_id_issuer_cmp(). 1298109998Smarkm [Steve Henson] 1299109998Smarkm 1300109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 1301109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 1302109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 1303109998Smarkm new macros which allow the underlying ASN1 function/structure to 1304109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 1305109998Smarkm references directly (such as &X509_it) but instead use the relevant 1306109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 1307109998Smarkm use of the new ASN1 code on platforms where exporting structures 1308109998Smarkm is problematical (for example in shared libraries) but exporting 1309109998Smarkm functions returning pointers to structures is not. 1310109998Smarkm [Steve Henson] 1311109998Smarkm 1312109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 1313109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 1314109998Smarkm The purpose of this is to allow applications to control, if they wish, 1315109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 1316109998Smarkm can be useful for session caching in multiple-server environments. A 1317109998Smarkm command-line switch for testing this (and any client code that wishes 1318109998Smarkm to use such a feature) has been added to "s_server". 1319109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 1320109998Smarkm 1321109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 1322109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 1323109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 1324109998Smarkm the growing number of special cases it was previously handling. 1325109998Smarkm [Richard Levitte] 1326109998Smarkm 1327109998Smarkm *) Make all configuration macros available for application by making 1328109998Smarkm sure they are available in opensslconf.h, by giving them names starting 1329109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 1330109998Smarkm sure e_os2.h will cover all platform-specific cases together with 1331109998Smarkm opensslconf.h. 1332109998Smarkm Additionally, it is now possible to define configuration/platform- 1333109998Smarkm specific names (called "system identities"). In the C code, these 1334109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 1335109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 1336109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 1337109998Smarkm what is available. 1338109998Smarkm [Richard Levitte] 1339109998Smarkm 1340109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 1341109998Smarkm number to use to be specified on the command line. Previously self 1342109998Smarkm signed certificates were hard coded with serial number 0 and the 1343109998Smarkm CA options of 'x509' had to use a serial number in a file which was 1344109998Smarkm auto incremented. 1345109998Smarkm [Steve Henson] 1346109998Smarkm 1347109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 1348109998Smarkm Currently CRL reason, invalidity date and hold instruction are 1349109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 1350109998Smarkm [Steve Henson] 1351109998Smarkm 1352109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 1353109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 1354109998Smarkm API, which was previously mandatory. This means that the data is 1355109998Smarkm not padded in any way and so the total length much be a multiple 1356109998Smarkm of the block size, otherwise an error occurs. 1357109998Smarkm [Steve Henson] 1358109998Smarkm 1359109998Smarkm *) Initial (incomplete) OCSP SSL support. 1360109998Smarkm [Steve Henson] 1361109998Smarkm 1362109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 1363109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 1364109998Smarkm option to ocsp utility. 1365109998Smarkm [Steve Henson] 1366109998Smarkm 1367109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 1368109998Smarkm reflects the various checks performed. Applications can decide 1369109998Smarkm whether to tolerate certain situations such as an absent nonce 1370109998Smarkm in a response when one was present in a request: the ocsp application 1371109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 1372109998Smarkm this is to allow responders to include a nonce in a response even if 1373109998Smarkm the request is nonce-less. 1374109998Smarkm [Steve Henson] 1375109998Smarkm 1376109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 1377109998Smarkm skipped when using openssl x509 multiple times on a single input file, 1378109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 1379109998Smarkm [Bodo Moeller] 1380109998Smarkm 1381109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 1382109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 1383109998Smarkm utility to correctly initialize revocation date of CRLs. 1384109998Smarkm [Steve Henson] 1385109998Smarkm 1386109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 1387109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 1388109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 1389109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 1390109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 1391109998Smarkm [Lutz Jaenicke] 1392109998Smarkm 1393109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 1394109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 1395109998Smarkm appear to exist. 1396109998Smarkm [Steve Henson] 1397109998Smarkm 1398109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 1399109998Smarkm additional certificates supplied. 1400109998Smarkm [Steve Henson] 1401109998Smarkm 1402109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 1403109998Smarkm OCSP client a number of certificate to only verify the response 1404109998Smarkm signature against. 1405109998Smarkm [Richard Levitte] 1406109998Smarkm 1407109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 1408109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 1409109998Smarkm AES OIDs. 1410109998Smarkm 1411109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 1412109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 1413109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 1414109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 1415109998Smarkm alias because they were not yet official; they could be 1416109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 1417109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 1418109998Smarkm alias is called "AES" and is part of "ALL".) 1419109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 1420109998Smarkm 1421109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 1422109998Smarkm request to response. 1423109998Smarkm [Steve Henson] 1424109998Smarkm 1425109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 1426109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 1427109998Smarkm extract information from a certificate request. OCSP_response_create() 1428109998Smarkm creates a response and optionally adds a basic response structure. 1429109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 1430109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 1431109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 1432109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 1433109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 1434109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 1435109998Smarkm (converts ASN1_TIME to GeneralizedTime). 1436109998Smarkm [Steve Henson] 1437109998Smarkm 1438109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 1439109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 1440109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 1441109998Smarkm contents: this is used in various key identifiers. 1442109998Smarkm [Steve Henson] 1443109998Smarkm 1444109998Smarkm *) Make sk_sort() tolerate a NULL argument. 1445109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 1446109998Smarkm 1447109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 1448109998Smarkm passed by the function are trusted implicitly. If any of them signed the 1449109998Smarkm response then it is assumed to be valid and is not verified. 1450109998Smarkm [Steve Henson] 1451109998Smarkm 1452109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 1453109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 1454109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 1455109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 1456109998Smarkm <support@securenetterm.com>] 1457109998Smarkm 1458109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 1459109998Smarkm routines: without these tracing memory leaks is very painful. 1460109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 1461109998Smarkm [Steve Henson] 1462109998Smarkm 1463109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 1464109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 1465109998Smarkm effectively meant GeneralizedTime would never be used. Now it 1466109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 1467109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 1468109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 1469109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 1470109998Smarkm <support@securenetterm.com>] 1471109998Smarkm 1472109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 1473109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 1474109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 1475109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 1476109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 1477109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 1478109998Smarkm [Steve Henson] 1479109998Smarkm 1480109998Smarkm *) Add summary printout to ocsp utility. The various functions which 1481109998Smarkm convert status values to strings have been renamed to: 1482109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 1483109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 1484109998Smarkm to verify nonce values and to disable verification. OCSP response 1485109998Smarkm printout format cleaned up. 1486109998Smarkm [Steve Henson] 1487109998Smarkm 1488109998Smarkm *) Add additional OCSP certificate checks. These are those specified 1489109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 1490109998Smarkm certificate being checked must either be the OCSP signer certificate 1491109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 1492109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 1493109998Smarkm usage. This check is performed by attempting to match the OCSP 1494109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 1495109998Smarkm in the OCSP_CERTID structures of the response. 1496109998Smarkm [Steve Henson] 1497109998Smarkm 1498109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 1499109998Smarkm and related routines. This uses the standard OpenSSL certificate 1500109998Smarkm verify routines to perform initial checks (just CA validity) and 1501109998Smarkm to obtain the certificate chain. Then additional checks will be 1502109998Smarkm performed on the chain. Currently the root CA is checked to see 1503109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 1504109998Smarkm a root CA as a global signing root: that is any certificate that 1505109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 1506109998Smarkm [Steve Henson] 1507109998Smarkm 1508109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 1509109998Smarkm extensions from a separate configuration file. 1510109998Smarkm As when reading extensions from the main configuration file, 1511109998Smarkm the '-extensions ...' option may be used for specifying the 1512109998Smarkm section to use. 1513109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 1514109998Smarkm 1515109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 1516109998Smarkm read. The request can be sent to a responder and the output 1517109998Smarkm parsed, outputed or printed in text form. Not complete yet: 1518109998Smarkm still needs to check the OCSP response validity. 1519109998Smarkm [Steve Henson] 1520109998Smarkm 1521109998Smarkm *) New subcommands for 'openssl ca': 1522109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 1523109998Smarkm the given serial number (according to the index file). 1524109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 1525109998Smarkm in the index file. 1526109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 1527109998Smarkm 1528109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 1529109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 1530109998Smarkm so that the resulting key is not encrypted. 1531109998Smarkm [Damien Miller <djm@mindrot.org>] 1532109998Smarkm 1533109998Smarkm *) New configuration for the GNU Hurd. 1534109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 1535109998Smarkm 1536109998Smarkm *) Initial code to implement OCSP basic response verify. This 1537109998Smarkm is currently incomplete. Currently just finds the signer's 1538109998Smarkm certificate and verifies the signature on the response. 1539109998Smarkm [Steve Henson] 1540109998Smarkm 1541109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 1542109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 1543109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 1544109998Smarkm [Bodo Moeller] 1545109998Smarkm 1546109998Smarkm *) Allowing defining memory allocation callbacks that will be given 1547109998Smarkm file name and line number information in additional arguments 1548109998Smarkm (a const char* and an int). The basic functionality remains, as 1549109998Smarkm well as the original possibility to just replace malloc(), 1550109998Smarkm realloc() and free() by functions that do not know about these 1551109998Smarkm additional arguments. To register and find out the current 1552109998Smarkm settings for extended allocation functions, the following 1553109998Smarkm functions are provided: 1554109998Smarkm 1555109998Smarkm CRYPTO_set_mem_ex_functions 1556109998Smarkm CRYPTO_set_locked_mem_ex_functions 1557109998Smarkm CRYPTO_get_mem_ex_functions 1558109998Smarkm CRYPTO_get_locked_mem_ex_functions 1559109998Smarkm 1560109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 1561109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 1562109998Smarkm extended allocation function is enabled. 1563109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 1564109998Smarkm a conventional allocation function is enabled. 1565109998Smarkm [Richard Levitte, Bodo Moeller] 1566109998Smarkm 1567109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 1568109998Smarkm There should no longer be any prototype-casting required when using 1569109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 1570109998Smarkm the callback types and macros at the head of lhash.h for details 1571109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 1572109998Smarkm [Geoff Thorpe] 1573109998Smarkm 1574109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 1575109998Smarkm If /dev/[u]random devices are not available or do not return enough 1576109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 1577109998Smarkm be queried. 1578109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 1579109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 1580109998Smarkm when enough entropy was collected without querying more sockets. 1581109998Smarkm [Lutz Jaenicke] 1582109998Smarkm 1583109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 1584109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 1585109998Smarkm of data has been collected. We spend at most 10 ms on each file 1586109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 1587109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 1588109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 1589109998Smarkm platforms the 10 ms delay will never occur. 1590109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 1591109998Smarkm For VMS, there's a currently-empty rand_vms.c. 1592109998Smarkm [Richard Levitte] 1593109998Smarkm 1594109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 1595109998Smarkm provide utility functions which an application needing 1596109998Smarkm to issue a request to an OCSP responder and analyse the 1597109998Smarkm response will typically need: as opposed to those which an 1598109998Smarkm OCSP responder itself would need which will be added later. 1599109998Smarkm 1600109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 1601109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 1602109998Smarkm response. OCSP_response_get1_basic() extracts basic response 1603109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 1604109998Smarkm information from an OCSP_CERTID structure (which will be created 1605109998Smarkm when the request structure is built). These are built from lower 1606109998Smarkm level functions which work on OCSP_SINGLERESP structures but 1607109998Smarkm wont normally be used unless the application wishes to examine 1608109998Smarkm extensions in the OCSP response for example. 1609109998Smarkm 1610109998Smarkm Replace nonce routines with a pair of functions. 1611109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 1612109998Smarkm generates a random value. OCSP_check_nonce() checks the 1613109998Smarkm validity of the nonce in an OCSP response. 1614109998Smarkm [Steve Henson] 1615109998Smarkm 1616109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 1617109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 1618109998Smarkm need to free up the newly created id. Change return type 1619109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 1620109998Smarkm This can then be used to add extensions to the request. 1621109998Smarkm Deleted OCSP_request_new(), since most of its functionality 1622109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 1623109998Smarkm clash) apart from the ability to set the request name which 1624109998Smarkm will be added elsewhere. 1625109998Smarkm [Steve Henson] 1626109998Smarkm 1627109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 1628109998Smarkm various functions. Extensions are now handled using the new 1629109998Smarkm OCSP extension code. New simple OCSP HTTP function which 1630109998Smarkm can be used to send requests and parse the response. 1631109998Smarkm [Steve Henson] 1632109998Smarkm 1633109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 1634109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 1635109998Smarkm uses the special reorder version of SET OF to sort the attributes 1636109998Smarkm and reorder them to match the encoded order. This resolves a long 1637109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 1638109998Smarkm it used to fail because the attribute order did not match the 1639109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 1640109998Smarkm it uses the received order. This is necessary to tolerate some broken 1641109998Smarkm software that does not order SET OF. This is handled by encoding 1642109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 1643109998Smarkm to produce the required SET OF. 1644109998Smarkm [Steve Henson] 1645109998Smarkm 1646109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 1647109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 1648109998Smarkm files to get correct declarations of the ASN.1 item variables. 1649109998Smarkm [Richard Levitte] 1650109998Smarkm 1651109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 1652109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 1653109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 1654109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 1655109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 1656109998Smarkm ASN1_ITEM and no wrapper functions. 1657109998Smarkm [Steve Henson] 1658109998Smarkm 1659109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 1660109998Smarkm replace the old function pointer based I/O routines. Change most of 1661109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 1662109998Smarkm [Steve Henson] 1663109998Smarkm 1664109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 1665109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 1666109998Smarkm it complain about algorithm deselection that isn't recognised. 1667109998Smarkm [Richard Levitte] 1668109998Smarkm 1669109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 1670109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 1671109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 1672109998Smarkm some old style ASN1 functions: this can be used to determine if old 1673109998Smarkm code will still work when these eventually go away. 1674109998Smarkm [Steve Henson] 1675109998Smarkm 1676109998Smarkm *) New extension functions for OCSP structures, these follow the 1677109998Smarkm same conventions as certificates and CRLs. 1678109998Smarkm [Steve Henson] 1679109998Smarkm 1680109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 1681109998Smarkm adds an extension. Its behaviour can be customised with various 1682109998Smarkm flags to append, replace or delete. Various wrappers added for 1683109998Smarkm certifcates and CRLs. 1684109998Smarkm [Steve Henson] 1685109998Smarkm 1686109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 1687109998Smarkm an extension cannot be parsed. Correct a typo in the 1688109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 1689109998Smarkm [Steve Henson] 1690109998Smarkm 1691109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 1692109998Smarkm entries for variables. 1693109998Smarkm [Steve Henson] 1694109998Smarkm 1695109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 1696109998Smarkm problems: As the program is single-threaded, all we have 1697109998Smarkm to do is register a locking callback using an array for 1698109998Smarkm storing which locks are currently held by the program. 1699109998Smarkm [Bodo Moeller] 1700109998Smarkm 1701109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 1702109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 1703109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 1704109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 1705109998Smarkm Unfortunately, the ex_data design is not at all suited 1706109998Smarkm for multi-threaded use, so it probably should be abolished. 1707109998Smarkm [Bodo Moeller] 1708109998Smarkm 1709109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 1710109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 1711109998Smarkm 1712109998Smarkm *) Move common extension printing code to new function 1713109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 1714109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 1715109998Smarkm [Steve Henson] 1716109998Smarkm 1717109998Smarkm *) New function X509_signature_print() to remove duplication in some 1718109998Smarkm print routines. 1719109998Smarkm [Steve Henson] 1720109998Smarkm 1721109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 1722109998Smarkm set (this was treated exactly the same as SET OF previously). This 1723109998Smarkm is used to reorder the STACK representing the structure to match the 1724109998Smarkm encoding. This will be used to get round a problem where a PKCS7 1725109998Smarkm structure which was signed could not be verified because the STACK 1726109998Smarkm order did not reflect the encoded order. 1727109998Smarkm [Steve Henson] 1728109998Smarkm 1729109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 1730109998Smarkm [Steve Henson] 1731109998Smarkm 1732109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 1733109998Smarkm for its ASN1 operations. The old style function pointers still exist 1734109998Smarkm for now but they will eventually go away. 1735109998Smarkm [Steve Henson] 1736109998Smarkm 1737109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 1738109998Smarkm completely replaces the old ASN1 functionality with a table driven 1739109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 1740109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 1741109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 1742109998Smarkm has also been converted to the new form. 1743109998Smarkm [Steve Henson] 1744109998Smarkm 1745109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 1746109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 1747109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 1748109998Smarkm for negative moduli. 1749109998Smarkm [Bodo Moeller] 1750109998Smarkm 1751109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 1752109998Smarkm of not touching the result's sign bit. 1753109998Smarkm [Bodo Moeller] 1754109998Smarkm 1755109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 1756109998Smarkm set. 1757109998Smarkm [Bodo Moeller] 1758109998Smarkm 1759109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 1760109998Smarkm macros to declare and implement thin (optionally static) functions 1761109998Smarkm that provide type-safety and avoid function pointer casting for the 1762109998Smarkm type-specific callbacks. 1763109998Smarkm [Geoff Thorpe] 1764109998Smarkm 1765109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 1766109998Smarkm RFC 2712. 1767109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 1768109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 1769109998Smarkm 1770109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 1771109998Smarkm in sections depending on the subject. 1772109998Smarkm [Richard Levitte] 1773109998Smarkm 1774109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 1775109998Smarkm Windows. 1776109998Smarkm [Richard Levitte] 1777109998Smarkm 1778109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 1779109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 1780109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 1781109998Smarkm be handled deterministically). 1782109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1783109998Smarkm 1784109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 1785109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 1786109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 1787109998Smarkm [Bodo Moeller] 1788109998Smarkm 1789109998Smarkm *) New function BN_kronecker. 1790109998Smarkm [Bodo Moeller] 1791109998Smarkm 1792109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 1793109998Smarkm positive unless both parameters are zero. 1794109998Smarkm Previously something reasonably close to an infinite loop was 1795109998Smarkm possible because numbers could be growing instead of shrinking 1796109998Smarkm in the implementation of Euclid's algorithm. 1797109998Smarkm [Bodo Moeller] 1798109998Smarkm 1799109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 1800109998Smarkm sign of the number in question. 1801109998Smarkm 1802109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 1803109998Smarkm 1804109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 1805109998Smarkm because its test if the absolute value of 'a' equals 'w'. 1806109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 1807109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 1808109998Smarkm BN_is_one(), and BN_is_word(). 1809109998Smarkm [Bodo Moeller] 1810109998Smarkm 1811109998Smarkm *) New function BN_swap. 1812109998Smarkm [Bodo Moeller] 1813109998Smarkm 1814109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 1815109998Smarkm the exponentiation functions are more likely to produce reasonable 1816109998Smarkm results on negative inputs. 1817109998Smarkm [Bodo Moeller] 1818109998Smarkm 1819109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 1820109998Smarkm Previously, it could be negative if one of the factors was negative; 1821109998Smarkm I don't think anyone really wanted that behaviour. 1822109998Smarkm [Bodo Moeller] 1823109998Smarkm 1824109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 1825109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 1826109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 1827109998Smarkm and add new functions: 1828109998Smarkm 1829109998Smarkm BN_nnmod 1830109998Smarkm BN_mod_sqr 1831109998Smarkm BN_mod_add 1832109998Smarkm BN_mod_add_quick 1833109998Smarkm BN_mod_sub 1834109998Smarkm BN_mod_sub_quick 1835109998Smarkm BN_mod_lshift1 1836109998Smarkm BN_mod_lshift1_quick 1837109998Smarkm BN_mod_lshift 1838109998Smarkm BN_mod_lshift_quick 1839109998Smarkm 1840109998Smarkm These functions always generate non-negative results. 1841109998Smarkm 1842109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 1843109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 1844109998Smarkm 1845109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 1846109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 1847109998Smarkm be reduced modulo m. 1848109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1849109998Smarkm 1850111147Snectar#if 0 1851111147Snectar The following entry accidentily appeared in the CHANGES file 1852111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 1853111147Snectar it do *not* apply to OpenSSL 0.9.7. 1854111147Snectar 1855109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1856109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 1857109998Smarkm required a small change in bn_mul_part_recursive() and the addition 1858109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 1859109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 1860109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 1861109998Smarkm differing sizes. 1862109998Smarkm [Richard Levitte] 1863111147Snectar#endif 1864109998Smarkm 1865109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 1866109998Smarkm unless the '-salt' option is used (which usually means that 1867109998Smarkm verification would just waste user's time since the resulting 1868109998Smarkm hash is going to be compared with some given password hash) 1869109998Smarkm or the new '-noverify' option is used. 1870109998Smarkm 1871109998Smarkm This is an incompatible change, but it does not affect 1872109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 1873109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 1874109998Smarkm cause any problems. 1875109998Smarkm [Bodo Moeller] 1876109998Smarkm 1877109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 1878109998Smarkm [Richard Levitte] 1879109998Smarkm 1880109998Smarkm *) Make DSO load along a path given through an environment variable 1881109998Smarkm (SHLIB_PATH) with shl_load(). 1882109998Smarkm [Richard Levitte] 1883109998Smarkm 1884109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 1885109998Smarkm Also constify the RSA code and most things related to it. In a 1886109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 1887109998Smarkm casts back to non-const were required (to be solved at a later 1888109998Smarkm time) 1889109998Smarkm [Richard Levitte] 1890109998Smarkm 1891109998Smarkm *) Make it so the openssl application has all engines loaded by default. 1892109998Smarkm [Richard Levitte] 1893109998Smarkm 1894109998Smarkm *) Constify the BIGNUM routines a little more. 1895109998Smarkm [Richard Levitte] 1896109998Smarkm 1897109998Smarkm *) Add the following functions: 1898109998Smarkm 1899109998Smarkm ENGINE_load_cswift() 1900109998Smarkm ENGINE_load_chil() 1901109998Smarkm ENGINE_load_atalla() 1902109998Smarkm ENGINE_load_nuron() 1903109998Smarkm ENGINE_load_builtin_engines() 1904109998Smarkm 1905109998Smarkm That way, an application can itself choose if external engines that 1906109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 1907109998Smarkm that applications won't have to be linked with libdl or other dso 1908109998Smarkm libraries unless it's really needed. 1909109998Smarkm 1910109998Smarkm Changed 'openssl engine' to load all engines on demand. 1911109998Smarkm Changed the engine header files to avoid the duplication of some 1912109998Smarkm declarations (they differed!). 1913109998Smarkm [Richard Levitte] 1914109998Smarkm 1915109998Smarkm *) 'openssl engine' can now list capabilities. 1916109998Smarkm [Richard Levitte] 1917109998Smarkm 1918109998Smarkm *) Better error reporting in 'openssl engine'. 1919109998Smarkm [Richard Levitte] 1920109998Smarkm 1921109998Smarkm *) Never call load_dh_param(NULL) in s_server. 1922109998Smarkm [Bodo Moeller] 1923109998Smarkm 1924109998Smarkm *) Add engine application. It can currently list engines by name and 1925109998Smarkm identity, and test if they are actually available. 1926109998Smarkm [Richard Levitte] 1927109998Smarkm 1928109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 1929109998Smarkm sure the installed documentation is also owned by root.root. 1930109998Smarkm [Damien Miller <djm@mindrot.org>] 1931109998Smarkm 1932109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 1933109998Smarkm keys (public as well as private) handled by engines. 1934109998Smarkm [Richard Levitte] 1935109998Smarkm 1936109998Smarkm *) Add OCSP code that comes from CertCo. 1937109998Smarkm [Richard Levitte] 1938109998Smarkm 1939109998Smarkm *) Add VMS support for the Rijndael code. 1940109998Smarkm [Richard Levitte] 1941109998Smarkm 1942109998Smarkm *) Added untested support for Nuron crypto accelerator. 1943109998Smarkm [Ben Laurie] 1944109998Smarkm 1945109998Smarkm *) Add support for external cryptographic devices. This code was 1946109998Smarkm previously distributed separately as the "engine" branch. 1947109998Smarkm [Geoff Thorpe, Richard Levitte] 1948109998Smarkm 1949109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 1950109998Smarkm have far greater control over how a "name" is turned into a filename 1951109998Smarkm depending on the operating environment and any oddities about the 1952109998Smarkm different shared library filenames on each system. 1953109998Smarkm [Geoff Thorpe] 1954109998Smarkm 1955109998Smarkm *) Support threads on FreeBSD-elf in Configure. 1956109998Smarkm [Richard Levitte] 1957109998Smarkm 1958109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 1959109998Smarkm warnings about corrupt line number information when assembling 1960109998Smarkm with debugging information. This is caused by the overlapping 1961109998Smarkm of two sections. 1962109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 1963109998Smarkm 1964109998Smarkm *) NCONF changes. 1965109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 1966109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 1967109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 1968109998Smarkm binary backward compatibility. 1969109998Smarkm Make it possible for methods to load from something other than a BIO, 1970109998Smarkm by providing a function pointer that is given a name instead of a BIO. 1971109998Smarkm For example, this could be used to load configuration data from an 1972109998Smarkm LDAP server. 1973109998Smarkm [Richard Levitte] 1974109998Smarkm 1975109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 1976109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 1977109998Smarkm with non blocking I/O was not possible because no retry code was 1978109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 1979109998Smarkm this case. 1980109998Smarkm [Steve Henson] 1981109998Smarkm 1982109998Smarkm *) Added the beginnings of Rijndael support. 1983109998Smarkm [Ben Laurie] 1984109998Smarkm 1985109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 1986109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 1987109998Smarkm to allow certificate printing to more controllable, additional 1988109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 1989109998Smarkm set. 1990109998Smarkm [Steve Henson] 1991109998Smarkm 1992109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 1993109998Smarkm [Richard Levitte] 1994109998Smarkm 1995120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 1996120631Snectar 1997120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 1998120631Snectar 1999120631Snectar Stop out of bounds reads in the ASN1 code when presented with 2000120631Snectar invalid tags (CAN-2003-0543 and CAN-2003-0544). 2001120631Snectar 2002120631Snectar If verify callback ignores invalid public key errors don't try to check 2003120631Snectar certificate signature with the NULL public key. 2004120631Snectar 2005120631Snectar [Steve Henson] 2006120631Snectar 2007120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 2008120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 2009120631Snectar specifications. 2010120631Snectar [Steve Henson] 2011120631Snectar 2012120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 2013120631Snectar extra data after the compression methods not only for TLS 1.0 2014120631Snectar but also for SSL 3.0 (as required by the specification). 2015120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 2016120631Snectar 2017120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 2018120631Snectar when it's 512 *bits* long, not 512 bytes. 2019120631Snectar [Richard Levitte] 2020120631Snectar 2021120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 2022120631Snectar 2023120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 2024120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 2025120631Snectar a protocol version number mismatch like a decryption error 2026120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 2027120631Snectar [Bodo Moeller] 2028120631Snectar 2029120631Snectar *) Turn on RSA blinding by default in the default implementation 2030120631Snectar to avoid a timing attack. Applications that don't want it can call 2031120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 2032120631Snectar They would be ill-advised to do so in most cases. 2033120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 2034120631Snectar 2035120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 2036120631Snectar seeded (in this case, the secret RSA exponent is abused as 2037120631Snectar an unpredictable seed -- if it is not unpredictable, there 2038120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 2039120631Snectar by remembering the creator's thread ID in rsa->blinding and 2040120631Snectar having all other threads use local one-time blinding factors 2041120631Snectar (this requires more computation than sharing rsa->blinding, but 2042120631Snectar avoids excessive locking; and if an RSA object is not shared 2043120631Snectar between threads, blinding will still be very fast). 2044120631Snectar [Bodo Moeller] 2045120631Snectar 2046111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 2047111147Snectar 2048111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 2049111147Snectar via timing by performing a MAC computation even if incorrrect 2050111147Snectar block cipher padding has been found. This is a countermeasure 2051111147Snectar against active attacks where the attacker has to distinguish 2052111147Snectar between bad padding and a MAC verification error. (CAN-2003-0078) 2053111147Snectar 2054111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 2055111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 2056111147Snectar Martin Vuagnoux (EPFL, Ilion)] 2057111147Snectar 2058109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 2059109998Smarkm 2060109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 2061109998Smarkm memory from it's contents. This is done with a counter that will 2062109998Smarkm place alternating values in each byte. This can be used to solve 2063109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 2064109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 2065109998Smarkm be read through on certain media, for example a swap space on disk. 2066109998Smarkm [Geoff Thorpe] 2067109998Smarkm 2068109998Smarkm *) Bugfix: client side session caching did not work with external caching, 2069109998Smarkm because the session->cipher setting was not restored when reloading 2070109998Smarkm from the external cache. This problem was masked, when 2071109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 2072109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 2073109998Smarkm [Lutz Jaenicke] 2074109998Smarkm 2075109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 2076109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 2077109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 2078109998Smarkm 2079109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 2080109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 2081109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 2082109998Smarkm EVP_cleanup(). 2083109998Smarkm [Richard Levitte] 2084109998Smarkm 2085109998Smarkm *) Change the default configuration reader to deal with last line not 2086109998Smarkm being properly terminated. 2087109998Smarkm [Richard Levitte] 2088109998Smarkm 2089109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 2090109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 2091109998Smarkm emailAddress where the value has the type ia5String. 2092109998Smarkm [stefank@valicert.com via Richard Levitte] 2093109998Smarkm 2094109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 2095109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 2096109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 2097109998Smarkm the bitwise-OR of the two for use by the majority of applications 2098109998Smarkm wanting this behaviour, and update the docs. The documented 2099109998Smarkm behaviour and actual behaviour were inconsistent and had been 2100109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 2101109998Smarkm change. 2102109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 2103109998Smarkm 2104109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 2105109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 2106109998Smarkm [Bodo Moeller] 2107109998Smarkm 2108109998Smarkm *) Fix initialization code race conditions in 2109109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 2110109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 2111109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 2112109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 2113109998Smarkm ssl2_get_cipher_by_char(), 2114109998Smarkm ssl3_get_cipher_by_char(). 2115109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 2116109998Smarkm 2117109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 2118109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 2119109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 2120109998Smarkm (see [openssl.org #212]). 2121109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 2122109998Smarkm 2123109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 2124109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 2125109998Smarkm [Steve Henson] 2126109998Smarkm 2127101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 2128101618Snectar 2129101618Snectar *) [In 0.9.6g-engine release:] 2130101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 2131101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 2132101618Snectar 2133101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 2134101613Snectar 2135101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 2136101613Snectar and get fix the header length calculation. 2137101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 2138101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 2139101613Snectar Steve Henson] 2140101613Snectar 2141101613Snectar *) Use proper error handling instead of 'assertions' in buffer 2142101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 2143101613Snectar assertions could call abort()). 2144101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 2145101613Snectar 2146100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 2147100936Snectar 2148109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 2149109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 2150109998Smarkm negative or the content length exceeds the length of the 2151109998Smarkm supplied buffer. 2152109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2153109998Smarkm 2154100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 2155100936Snectar for the cipher strength set and where therefore not handled correctly 2156100936Snectar by the selection routines (PR #130). 2157100936Snectar [Lutz Jaenicke] 2158100936Snectar 2159100936Snectar *) Fix EVP_dsa_sha macro. 2160100936Snectar [Nils Larsch] 2161100936Snectar 2162100936Snectar *) New option 2163100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 2164100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 2165100936Snectar that was added in OpenSSL 0.9.6d. 2166100936Snectar 2167100936Snectar As the countermeasure turned out to be incompatible with some 2168100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 2169100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 2170100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 2171100936Snectar 's_server'), so the new option is automatically set in many 2172100936Snectar applications. 2173100936Snectar [Bodo Moeller] 2174100936Snectar 2175100936Snectar *) Changes in security patch: 2176100936Snectar 2177100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 2178100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 2179100936Snectar Air Force Materiel Command, USAF, under agreement number 2180100936Snectar F30602-01-2-0537. 2181100936Snectar 2182100936Snectar *) Add various sanity checks to asn1_get_length() to reject 2183100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 2184100936Snectar negative or the content length exceeds the length of the 2185109998Smarkm supplied buffer. (CAN-2002-0659) 2186100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2187100936Snectar 2188100936Snectar *) Assertions for various potential buffer overflows, not known to 2189100936Snectar happen in practice. 2190100936Snectar [Ben Laurie (CHATS)] 2191100936Snectar 2192100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 2193100936Snectar too small for 64 bit platforms. (CAN-2002-0655) 2194100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 2195100936Snectar 2196100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 2197100936Snectar supply an oversized session ID to a client. (CAN-2002-0656) 2198100936Snectar [Ben Laurie (CHATS)] 2199100936Snectar 2200100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 2201100936Snectar supply an oversized client master key. (CAN-2002-0656) 2202100936Snectar [Ben Laurie (CHATS)] 2203100936Snectar 2204100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 2205100928Snectar 2206100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 2207100928Snectar encoded as NULL) with id-dsa-with-sha1. 2208100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 2209100928Snectar 2210100928Snectar *) Check various X509_...() return values in apps/req.c. 2211100928Snectar [Nils Larsch <nla@trustcenter.de>] 2212100928Snectar 2213100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 2214100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 2215100928Snectar was just at the end of a processed block. The bug was discovered when 2216100928Snectar processing data through a buffering memory BIO handing the data to a 2217100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 2218100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 2219100928Snectar [Lutz Jaenicke] 2220100928Snectar 2221100928Snectar *) Implement a countermeasure against a vulnerability recently found 2222100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 2223100928Snectar before application data chunks to avoid the use of known IVs 2224100928Snectar with data potentially chosen by the attacker. 2225100928Snectar [Bodo Moeller] 2226100928Snectar 2227100928Snectar *) Fix length checks in ssl3_get_client_hello(). 2228100928Snectar [Bodo Moeller] 2229100928Snectar 2230100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 2231100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 2232100928Snectar ssl3_read_bytes() found application data while handshake 2233100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 2234100928Snectar merely automatically cleared during the initial handshake. 2235100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 2236100928Snectar 2237100928Snectar *) Fix object definitions for Private and Enterprise: they were not 2238100928Snectar recognized in their shortname (=lowercase) representation. Extend 2239100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 2240100928Snectar of silently ignoring the problem (Svenning Sorensen 2241100928Snectar <sss@sss.dnsalias.net>). 2242100928Snectar [Lutz Jaenicke] 2243100928Snectar 2244100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 2245100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 2246100928Snectar code did not properly initialise the 'add' and 'rem' values to 2247100928Snectar BN_generate_prime().) 2248100928Snectar 2249100928Snectar In the new general case, we do not insist that 'generator' is 2250100928Snectar actually a primitive root: This requirement is rather pointless; 2251100928Snectar a generator of the order-q subgroup is just as good, if not 2252100928Snectar better. 2253100928Snectar [Bodo Moeller] 2254100928Snectar 2255100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 2256100928Snectar Tom Wu <tom@arcot.com>. 2257100928Snectar [Lutz Jaenicke] 2258100928Snectar 2259100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 2260100928Snectar returning non-zero before the data has been completely received 2261100928Snectar when using non-blocking I/O. 2262100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 2263100928Snectar 2264100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 2265100928Snectar [Ben Laurie, Lutz Jaenicke] 2266100928Snectar 2267100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 2268100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 2269100928Snectar [Lutz Jaenicke] 2270100928Snectar 2271100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 2272100928Snectar configuration for the versions before that. 2273100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2274100928Snectar 2275100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 2276100928Snectar check whether we deal with a copy of a session and do not delete from 2277100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 2278100928Snectar <izhar@checkpoint.com>. 2279100928Snectar [Lutz Jaenicke] 2280100928Snectar 2281100928Snectar *) Do not store session data into the internal session cache, if it 2282100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 2283100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 2284100928Snectar [Lutz Jaenicke] 2285100928Snectar 2286100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 2287100928Snectar value is 0. 2288100928Snectar [Richard Levitte] 2289100928Snectar 2290109998Smarkm *) [In 0.9.6d-engine release:] 2291109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2292100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2293100928Snectar 2294100928Snectar *) Add the configuration target linux-s390x. 2295100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 2296100928Snectar 2297100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 2298100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 2299100928Snectar variable as an indication that a ClientHello message has been 2300100928Snectar received. As the flag value will be lost between multiple 2301100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 2302100928Snectar function may not be aware that a handshake has actually taken 2303100928Snectar place, thus preventing a new session from being added to the 2304100928Snectar session cache. 2305100928Snectar 2306100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 2307100928Snectar using a local variable. 2308100928Snectar [Lutz Jaenicke, Bodo Moeller] 2309100928Snectar 2310100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 2311100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 2312100928Snectar [Geoff Thorpe, Bodo Moeller] 2313100928Snectar 2314100928Snectar *) New 'shared_ldflag' column in Configure platform table. 2315100928Snectar [Richard Levitte] 2316100928Snectar 2317100928Snectar *) Fix EVP_CIPHER_mode macro. 2318100928Snectar ["Dan S. Camper" <dan@bti.net>] 2319100928Snectar 2320100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 2321100928Snectar type, we must throw them away by setting rr->length to 0. 2322100928Snectar [D P Chang <dpc@qualys.com>] 2323100928Snectar 232489837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 232589837Skris 232689837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 232789837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 232889837Skris worked incorrectly for those cases where range = 10..._2 and 232989837Skris 3*range is two bits longer than range.) 233089837Skris [Bodo Moeller] 233189837Skris 233289837Skris *) Only add signing time to PKCS7 structures if it is not already 233389837Skris present. 233489837Skris [Steve Henson] 233589837Skris 233689837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 233789837Skris OBJ_ld_ce should be OBJ_id_ce. 233889837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 233989837Skris incorrect (cf. RFC 3039). 234089837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 234189837Skris 234289837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 234389837Skris returns early because it has nothing to do. 234489837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 234589837Skris 234689837Skris *) [In 0.9.6c-engine release:] 234789837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 234889837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 234989837Skris 235089837Skris *) [In 0.9.6c-engine release:] 235189837Skris Add support for Cryptographic Appliance's keyserver technology. 235289837Skris (Use engine 'keyclient') 235389837Skris [Cryptographic Appliances and Geoff Thorpe] 235489837Skris 235589837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 235689837Skris is called via tools/c89.sh because arguments have to be 235789837Skris rearranged (all '-L' options must appear before the first object 235889837Skris modules). 235989837Skris [Richard Shapiro <rshapiro@abinitio.com>] 236089837Skris 236189837Skris *) [In 0.9.6c-engine release:] 236289837Skris Add support for Broadcom crypto accelerator cards, backported 236389837Skris from 0.9.7. 236489837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 236589837Skris 236689837Skris *) [In 0.9.6c-engine release:] 236789837Skris Add support for SureWare crypto accelerator cards from 236889837Skris Baltimore Technologies. (Use engine 'sureware') 236989837Skris [Baltimore Technologies and Mark Cox] 237089837Skris 237189837Skris *) [In 0.9.6c-engine release:] 237289837Skris Add support for crypto accelerator cards from Accelerated 237389837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 237489837Skris [AEP Inc. and Mark Cox] 237589837Skris 237689837Skris *) Add a configuration entry for gcc on UnixWare. 237789837Skris [Gary Benson <gbenson@redhat.com>] 237889837Skris 237989837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 238089837Skris messages are stored in a single piece (fixed-length part and 238189837Skris variable-length part combined) and fix various bugs found on the way. 238289837Skris [Bodo Moeller] 238389837Skris 238489837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 238589837Skris instead. BIO_gethostbyname() does not know what timeouts are 238689837Skris appropriate, so entries would stay in cache even when they have 238789837Skris become invalid. 238889837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 238989837Skris 239089837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 239189837Skris faced with a pathologically small ClientHello fragment that does 239289837Skris not contain client_version: Instead of aborting with an error, 239389837Skris simply choose the highest available protocol version (i.e., 239489837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 239589837Skris messages are never sent like this, but this change gives us 239689837Skris strictly correct behaviour at least for TLS. 239789837Skris [Bodo Moeller] 239889837Skris 239989837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 240089837Skris never resets s->method to s->ctx->method when called from within 240189837Skris one of the SSL handshake functions. 240289837Skris [Bodo Moeller; problem pointed out by Niko Baric] 240389837Skris 240489837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 240589837Skris (sent using the client's version number) if client_version is 240689837Skris smaller than the protocol version in use. Also change 240789837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 240889837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 240989837Skris the client will at least see that alert. 241089837Skris [Bodo Moeller] 241189837Skris 241289837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 241389837Skris correctly. 241489837Skris [Bodo Moeller] 241589837Skris 241689837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 241789837Skris client receives HelloRequest while in a handshake. 241889837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 241989837Skris 242089837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 242189837Skris should end in 'break', not 'goto end' which circuments various 242289837Skris cleanups done in state SSL_ST_OK. But session related stuff 242389837Skris must be disabled for SSL_ST_OK in the case that we just sent a 242489837Skris HelloRequest. 242589837Skris 242689837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 242789837Skris before just sending a HelloRequest. 242889837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 242989837Skris 243089837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 243189837Skris reveal whether illegal block cipher padding was found or a MAC 243289837Skris verification error occured. (Neither SSLerr() codes nor alerts 243389837Skris are directly visible to potential attackers, but the information 243489837Skris may leak via logfiles.) 243589837Skris 243689837Skris Similar changes are not required for the SSL 2.0 implementation 243789837Skris because the number of padding bytes is sent in clear for SSL 2.0, 243889837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 243989837Skris failed to verify that the purported number of padding bytes is in 244089837Skris the legal range. 244189837Skris [Bodo Moeller] 244289837Skris 244389837Skris *) Add OpenUNIX-8 support including shared libraries 244489837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 244589837Skris [Lutz Jaenicke] 244689837Skris 244789837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 244889837Skris 'wristwatch attack' using huge encoding parameters (cf. 244989837Skris James H. Manger's CRYPTO 2001 paper). Note that the 245089837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 245189837Skris encoding parameters and hence was not vulnerable. 245289837Skris [Bodo Moeller] 245389837Skris 245489837Skris *) BN_sqr() bug fix. 245589837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 245689837Skris 245789837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 245889837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 245989837Skris followed by modular reduction. 246089837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 246189837Skris 246289837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 246389837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 246489837Skris [Bodo Moeller] 246589837Skris 246689837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 246789837Skris This function was broken, as the check for a new client hello message 246889837Skris to handle SGC did not allow these large messages. 246989837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 247089837Skris [Lutz Jaenicke] 247189837Skris 247289837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 247389837Skris [Lutz Jaenicke] 247489837Skris 247589837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 247689837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 247789837Skris [Lutz Jaenicke] 247889837Skris 247989837Skris *) Rework the configuration and shared library support for Tru64 Unix. 248089837Skris The configuration part makes use of modern compiler features and 248189837Skris still retains old compiler behavior for those that run older versions 248289837Skris of the OS. The shared library support part includes a variant that 248389837Skris uses the RPATH feature, and is available through the special 248489837Skris configuration target "alpha-cc-rpath", which will never be selected 248589837Skris automatically. 248689837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 248789837Skris 248889837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 248989837Skris with the same message size as in ssl3_get_certificate_request(). 249089837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 249189837Skris messages might inadvertently be reject as too long. 249289837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 249389837Skris 249489837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 249589837Skris [Andy Polyakov] 249689837Skris 249789837Skris *) Modified SSL library such that the verify_callback that has been set 249889837Skris specificly for an SSL object with SSL_set_verify() is actually being 249989837Skris used. Before the change, a verify_callback set with this function was 250089837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 250189837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 250289837Skris to allow the necessary settings. 250389837Skris [Lutz Jaenicke] 250489837Skris 250589837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 250689837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 250789837Skris done automatically (in contradiction to the requirements of the C 250889837Skris standard). This made problems when used from OpenSSH. 250989837Skris [Lutz Jaenicke] 251089837Skris 251189837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 251289837Skris dh->length and always used 251389837Skris 251489837Skris BN_rand_range(priv_key, dh->p). 251589837Skris 251689837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 251789837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 251889837Skris dh->length (recommended exponent length) is much smaller than the 251989837Skris length of dh->p. We could use BN_rand_range() if the order of 252089837Skris the subgroup was stored in the DH structure, but we only have 252189837Skris dh->length. 252289837Skris 252389837Skris So switch back to 252489837Skris 252589837Skris BN_rand(priv_key, l, ...) 252689837Skris 252789837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 252889837Skris otherwise. 252989837Skris [Bodo Moeller] 253089837Skris 253189837Skris *) In 253289837Skris 253389837Skris RSA_eay_public_encrypt 253489837Skris RSA_eay_private_decrypt 253589837Skris RSA_eay_private_encrypt (signing) 253689837Skris RSA_eay_public_decrypt (signature verification) 253789837Skris 253889837Skris (default implementations for RSA_public_encrypt, 253989837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 254089837Skris always reject numbers >= n. 254189837Skris [Bodo Moeller] 254289837Skris 254389837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 254489837Skris to synchronize access to 'locking_thread'. This is necessary on 254589837Skris systems where access to 'locking_thread' (an 'unsigned long' 254689837Skris variable) is not atomic. 254789837Skris [Bodo Moeller] 254889837Skris 254989837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 255089837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 255189837Skris a race condition if 0 is a valid thread ID. 255289837Skris [Travis Vitek <vitek@roguewave.com>] 255389837Skris 255489837Skris *) Add support for shared libraries under Irix. 255589837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 255689837Skris 255789837Skris *) Add configuration option to build on Linux on both big-endian and 255889837Skris little-endian MIPS. 255989837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 256089837Skris 256189837Skris *) Add the possibility to create shared libraries on HP-UX. 256289837Skris [Richard Levitte] 256389837Skris 256479998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 256579998Skris 256679998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 256779998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 256879998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 256979998Skris PRNG state recovery was possible based on the output of 257079998Skris one PRNG request appropriately sized to gain knowledge on 257179998Skris 'md' followed by enough consecutive 1-byte PRNG requests 257279998Skris to traverse all of 'state'. 257379998Skris 257479998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 257579998Skris during PRNG output generation, hash all of the previous 257679998Skris 'md_local' value, not just the half used for PRNG output. 257779998Skris 257879998Skris 2. Make the number of bytes from 'state' included into the hash 257979998Skris independent from the number of PRNG bytes requested. 258079998Skris 258179998Skris The first measure alone would be sufficient to avoid 258279998Skris Markku-Juhani's attack. (Actually it had never occurred 258379998Skris to me that the half of 'md_local' used for chaining was the 258479998Skris half from which PRNG output bytes were taken -- I had always 258579998Skris assumed that the secret half would be used.) The second 258679998Skris measure makes sure that additional data from 'state' is never 258779998Skris mixed into 'md_local' in small portions; this heuristically 258879998Skris further strengthens the PRNG. 258979998Skris [Bodo Moeller] 259079998Skris 259179998Skris *) Fix crypto/bn/asm/mips3.s. 259279998Skris [Andy Polyakov] 259379998Skris 259479998Skris *) When only the key is given to "enc", the IV is undefined. Print out 259579998Skris an error message in this case. 259679998Skris [Lutz Jaenicke] 259779998Skris 259879998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 259979998Skris [Steve Henson] 260079998Skris 260179998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 260279998Skris positive and less than q. 260379998Skris [Bodo Moeller] 260479998Skris 260579998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 260679998Skris used: it isn't thread safe and the add_lock_callback should handle 260779998Skris that itself. 260879998Skris [Paul Rose <Paul.Rose@bridge.com>] 260979998Skris 261079998Skris *) Verify that incoming data obeys the block size in 261179998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 261279998Skris [Bodo Moeller] 261379998Skris 261479998Skris *) Fix OAEP check. 261579998Skris [Ulf M�ller, Bodo M�ller] 261679998Skris 261779998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 261889837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 261979998Skris when fixing the server behaviour for backwards-compatible 'client 262079998Skris hello' messages. (Note that the attack is impractical against 262179998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 262279998Skris means that the probability of guessing a valid ciphertext is 262379998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 262479998Skris paper.) 262579998Skris 262679998Skris Before 0.9.5, the countermeasure (hide the error by generating a 262779998Skris random 'decryption result') did not work properly because 262879998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 262979998Skris detect the supposedly ignored error. 263079998Skris 263179998Skris Both problems are now fixed. 263279998Skris [Bodo Moeller] 263379998Skris 263479998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 263579998Skris (previously it was 1024). 263679998Skris [Bodo Moeller] 263779998Skris 263879998Skris *) Fix for compatibility mode trust settings: ignore trust settings 263979998Skris unless some valid trust or reject settings are present. 264079998Skris [Steve Henson] 264179998Skris 264279998Skris *) Fix for blowfish EVP: its a variable length cipher. 264379998Skris [Steve Henson] 264479998Skris 264579998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 264679998Skris parameters in DSA public key structures and return an error in the 264779998Skris DSA routines if parameters are absent. 264879998Skris [Steve Henson] 264979998Skris 265079998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 265179998Skris in the current directory if neither $RANDFILE nor $HOME was set. 265279998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 265379998Skris caused some confusion to Windows users who haven't defined $HOME. 265479998Skris Thus RAND_file_name() is changed again: e_os.h can define a 265579998Skris DEFAULT_HOME, which will be used if $HOME is not set. 265679998Skris For Windows, we use "C:"; on other platforms, we still require 265779998Skris environment variables. 265879998Skris 265979998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 266079998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 266179998Skris having multiple threads call RAND_poll() concurrently. 266279998Skris [Bodo Moeller] 266379998Skris 266479998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 266579998Skris combination of a flag and a thread ID variable. 266679998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 266779998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 266889837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 266979998Skris that they do not hold after the first thread unsets add_do_not_lock). 267079998Skris [Bodo Moeller] 267179998Skris 267279998Skris *) Change bctest again: '-x' expressions are not available in all 267379998Skris versions of 'test'. 267479998Skris [Bodo Moeller] 267579998Skris 267676866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 267772613Skris 267876866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 267976866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 268076866Skris 268176866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 268276866Skris the default extension for executables, if any. Also, make the perl 268376866Skris scripts that use symlink() to test if it really exists and use "cp" 268476866Skris if it doesn't. All this made OpenSSL compilable and installable in 268576866Skris CygWin. 268676866Skris [Richard Levitte] 268776866Skris 268876866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 268976866Skris If SEQUENCE is length is indefinite just set c->slen to the total 269076866Skris amount of data available. 269176866Skris [Steve Henson, reported by shige@FreeBSD.org] 269276866Skris [This change does not apply to 0.9.7.] 269376866Skris 269476866Skris *) Change bctest to avoid here-documents inside command substitution 269576866Skris (workaround for FreeBSD /bin/sh bug). 269676866Skris For compatibility with Ultrix, avoid shell functions (introduced 269776866Skris in the bctest version that searches along $PATH). 269876866Skris [Bodo Moeller] 269976866Skris 270076866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 270176866Skris with des_encrypt() defined on some operating systems, like Solaris 270276866Skris and UnixWare. 270376866Skris [Richard Levitte] 270476866Skris 270576866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 270676866Skris On the Importance of Eliminating Errors in Cryptographic 270776866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 270876866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 270976866Skris [Ulf Moeller] 271076866Skris 271176866Skris *) MIPS assembler BIGNUM division bug fix. 271276866Skris [Andy Polyakov] 271376866Skris 271476866Skris *) Disabled incorrect Alpha assembler code. 271576866Skris [Richard Levitte] 271676866Skris 271776866Skris *) Fix PKCS#7 decode routines so they correctly update the length 271876866Skris after reading an EOC for the EXPLICIT tag. 271976866Skris [Steve Henson] 272076866Skris [This change does not apply to 0.9.7.] 272176866Skris 272276866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 272376866Skris if a 3DES key was generated with a 0 initial byte. Include 272476866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 272576866Skris (but broken) behaviour. 272676866Skris [Steve Henson] 272776866Skris 272876866Skris *) Enhance bctest to search for a working bc along $PATH and print 272976866Skris it when found. 273076866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 273176866Skris 273276866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 273376866Skris don't write to the wrong index in ERR_set_error_data. 273476866Skris [Bodo Moeller] 273576866Skris 273676866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 273776866Skris did not exist. 273876866Skris [Bodo Moeller] 273976866Skris 274076866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 274176866Skris [Jeremy Cooper <jeremy@baymoo.org>] 274276866Skris 274376866Skris *) Make it possible to reuse SSLv2 sessions. 274476866Skris [Richard Levitte] 274576866Skris 274676866Skris *) In copy_email() check for >= 0 as a return value for 274776866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 274876866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 274976866Skris 275076866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 275176866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 275276866Skris PKCS7_verify() fails with non detached data. 275376866Skris [Steve Henson] 275476866Skris 275576866Skris *) Don't use getenv in library functions when run as setuid/setgid. 275676866Skris New function OPENSSL_issetugid(). 275776866Skris [Ulf Moeller] 275876866Skris 275976866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 276076866Skris due to incorrect handling of multi-threading: 276176866Skris 276276866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 276376866Skris 276476866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 276576866Skris 276676866Skris 3. Count how many times MemCheck_off() has been called so that 276776866Skris nested use can be treated correctly. This also avoids 276876866Skris inband-signalling in the previous code (which relied on the 276976866Skris assumption that thread ID 0 is impossible). 277076866Skris [Bodo Moeller] 277176866Skris 277276866Skris *) Add "-rand" option also to s_client and s_server. 277376866Skris [Lutz Jaenicke] 277476866Skris 277576866Skris *) Fix CPU detection on Irix 6.x. 277676866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 277776866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 277876866Skris 277976866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 278076866Skris was empty. 278176866Skris [Steve Henson] 278276866Skris [This change does not apply to 0.9.7.] 278376866Skris 278476866Skris *) Use the cached encoding of an X509_NAME structure rather than 278576866Skris copying it. This is apparently the reason for the libsafe "errors" 278676866Skris but the code is actually correct. 278776866Skris [Steve Henson] 278876866Skris 278972613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 279072613Skris Bleichenbacher's DSA attack. 279176866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 279276866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 279376866Skris and leaves the highest bit random. 279476866Skris [Ulf Moeller, Bodo Moeller] 279572613Skris 279672613Skris *) In the NCONF_...-based implementations for CONF_... queries 279772613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 279872613Skris a temporary CONF structure with the data component set to NULL 279972613Skris (which gives segmentation faults in lh_retrieve). 280072613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 280172613Skris CONF_get_number (which may use environment variables) and directly 280272613Skris return NULL from CONF_get_section. 280372613Skris [Bodo Moeller] 280472613Skris 280572613Skris *) Fix potential buffer overrun for EBCDIC. 280672613Skris [Ulf Moeller] 280772613Skris 280872613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 280972613Skris keyUsage if basicConstraints absent for a CA. 281072613Skris [Steve Henson] 281172613Skris 281272613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 281372613Skris is more generally accepted (no spaces before the semicolon), since 281472613Skris some programs can't parse those values properly otherwise. Also make 281572613Skris sure BIO's that break lines after each write do not create invalid 281672613Skris headers. 281772613Skris [Richard Levitte] 281872613Skris 281972613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 282072613Skris macros previously used would not encode an empty SEQUENCE OF 282172613Skris and break the signature. 282272613Skris [Steve Henson] 282376866Skris [This change does not apply to 0.9.7.] 282472613Skris 282572613Skris *) Zero the premaster secret after deriving the master secret in 282672613Skris DH ciphersuites. 282772613Skris [Steve Henson] 282872613Skris 282972613Skris *) Add some EVP_add_digest_alias registrations (as found in 283072613Skris OpenSSL_add_all_digests()) to SSL_library_init() 283172613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 283272613Skris compatibility with peers using X.509 certificates 283372613Skris with unconventional AlgorithmIdentifier OIDs. 283472613Skris [Bodo Moeller] 283572613Skris 283672613Skris *) Fix for Irix with NO_ASM. 283772613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 283872613Skris 283972613Skris *) ./config script fixes. 284072613Skris [Ulf Moeller, Richard Levitte] 284172613Skris 284272613Skris *) Fix 'openssl passwd -1'. 284372613Skris [Bodo Moeller] 284472613Skris 284572613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 284672613Skris terminated strings whose length is passed in the passlen 284772613Skris parameter, for example from PEM callbacks. This was done 284872613Skris by adding an extra length parameter to asc2uni(). 284972613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 285072613Skris 285172613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 285272613Skris call failed, free the DSA structure. 285372613Skris [Bodo Moeller] 285472613Skris 285572613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 285672613Skris These are present in some PKCS#12 files. 285772613Skris [Steve Henson] 285872613Skris 285972613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 286072613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 286172613Skris when writing a 32767 byte record. 286272613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 286372613Skris 286472613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 286572613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 286672613Skris 286772613Skris (RSA objects have a reference count access to which is protected 286872613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 286972613Skris so they are meant to be shared between threads.) 287072613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 287172613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 287272613Skris 287372613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 287472613Skris [Bodo Moeller] 287572613Skris 287676866Skris *) Use better test patterns in bntest. 287776866Skris [Ulf M�ller] 287876866Skris 287972613Skris *) rand_win.c fix for Borland C. 288072613Skris [Ulf M�ller] 288172613Skris 288272613Skris *) BN_rshift bugfix for n == 0. 288372613Skris [Bodo Moeller] 288472613Skris 288576866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 288676866Skris so that 'make test' does not abort just because 'bc' is broken. 288776866Skris [Bodo Moeller] 288876866Skris 288972613Skris *) Store verify_result within SSL_SESSION also for client side to 289072613Skris avoid potential security hole. (Re-used sessions on the client side 289172613Skris always resulted in verify_result==X509_V_OK, not using the original 289272613Skris result of the server certificate verification.) 289372613Skris [Lutz Jaenicke] 289472613Skris 289572613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 289672613Skris SSL3_RT_APPLICATION_DATA, return 0. 289772613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 289872613Skris [Bodo Moeller] 289972613Skris 290072613Skris *) Fix SSL_peek: 290172613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 290272613Skris releases, have been re-implemented by renaming the previous 290372613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 290472613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 290572613Skris to them. The new ssl[23]_{read,peek} functions are calls to 290672613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 290772613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 290872613Skris does the actual work for ssl3_read_internal. 290972613Skris [Bodo Moeller] 291072613Skris 291176866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 291276866Skris the method-specific "init()" handler. Also clean up ex_data after 291376866Skris calling the method-specific "finish()" handler. Previously, this was 291476866Skris happening the other way round. 291576866Skris [Geoff Thorpe] 291676866Skris 291772613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 291872613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 291972613Skris [Bodo Moeller] 292072613Skris 292176866Skris *) Make sure that shared libraries get the internal name engine with 292276866Skris the full version number and not just 0. This should mark the 292376866Skris shared libraries as not backward compatible. Of course, this should 292476866Skris be changed again when we can guarantee backward binary compatibility. 292576866Skris [Richard Levitte] 292676866Skris 292772613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 292872613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 292972613Skris 293076866Skris *) Rework the system to generate shared libraries: 293176866Skris 293276866Skris - Make note of the expected extension for the shared libraries and 293376866Skris if there is a need for symbolic links from for example libcrypto.so.0 293476866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 293576866Skris that. 293676866Skris 293776866Skris - Make as few rebuilds of the shared libraries as possible. 293876866Skris 293976866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 294076866Skris 294176866Skris - When installing, install the shared libraries separately from the 294276866Skris static ones. 294376866Skris [Richard Levitte] 294476866Skris 294572613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 294672613Skris 294772613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 294872613Skris and not in SSL_clear because the latter is also used by the 294972613Skris accept/connect functions; previously, the settings made by 295072613Skris SSL_set_read_ahead would be lost during the handshake. 295172613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 295272613Skris 295372613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 295472613Skris Previously, it would create entries for disableed algorithms no 295572613Skris matter what. 295672613Skris [Richard Levitte] 295772613Skris 295872613Skris *) Added several new manual pages for SSL_* function. 295972613Skris [Lutz Jaenicke] 296072613Skris 296168651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 296268651Skris 296368651Skris *) In ssl23_get_client_hello, generate an error message when faced 296468651Skris with an initial SSL 3.0/TLS record that is too small to contain the 296568651Skris first two bytes of the ClientHello message, i.e. client_version. 296668651Skris (Note that this is a pathologic case that probably has never happened 296768651Skris in real life.) The previous approach was to use the version number 296868651Skris from the record header as a substitute; but our protocol choice 296968651Skris should not depend on that one because it is not authenticated 297068651Skris by the Finished messages. 297168651Skris [Bodo Moeller] 297268651Skris 297368651Skris *) More robust randomness gathering functions for Windows. 297468651Skris [Jeffrey Altman <jaltman@columbia.edu>] 297568651Skris 297668651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 297768651Skris not set then we don't setup the error code for issuer check errors 297868651Skris to avoid possibly overwriting other errors which the callback does 297968651Skris handle. If an application does set the flag then we assume it knows 298068651Skris what it is doing and can handle the new informational codes 298168651Skris appropriately. 298268651Skris [Steve Henson] 298368651Skris 298468651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 298568651Skris a general "ANY" type, as such it should be able to decode anything 298668651Skris including tagged types. However it didn't check the class so it would 298768651Skris wrongly interpret tagged types in the same way as their universal 298868651Skris counterpart and unknown types were just rejected. Changed so that the 298968651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 299068651Skris that is the encoding is stored intact. There is also a new type 299168651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 299268651Skris case we have no idea what the actual type is so we just lump them all 299368651Skris together. 299468651Skris [Steve Henson] 299568651Skris 299668651Skris *) On VMS, stdout may very well lead to a file that is written to 299768651Skris in a record-oriented fashion. That means that every write() will 299868651Skris write a separate record, which will be read separately by the 299968651Skris programs trying to read from it. This can be very confusing. 300068651Skris 300168651Skris The solution is to put a BIO filter in the way that will buffer 300268651Skris text until a linefeed is reached, and then write everything a 300368651Skris line at a time, so every record written will be an actual line, 300468651Skris not chunks of lines and not (usually doesn't happen, but I've 300568651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 300668651Skris the answer. 300768651Skris 300868651Skris Currently, it's a VMS-only method, because that's where it has 300968651Skris been tested well enough. 301068651Skris [Richard Levitte] 301168651Skris 301268651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 301368651Skris it can return incorrect results. 301468651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 301568651Skris but it was in 0.9.6-beta[12].) 301668651Skris [Bodo Moeller] 301768651Skris 301868651Skris *) Disable the check for content being present when verifying detached 301968651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 302068651Skris include zero length content when signing messages. 302168651Skris [Steve Henson] 302268651Skris 302368651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 302468651Skris BIO_ctrl (for BIO pairs). 302568651Skris [Bodo M�ller] 302668651Skris 302768651Skris *) Add DSO method for VMS. 302868651Skris [Richard Levitte] 302968651Skris 303068651Skris *) Bug fix: Montgomery multiplication could produce results with the 303168651Skris wrong sign. 303268651Skris [Ulf M�ller] 303368651Skris 303468651Skris *) Add RPM specification openssl.spec and modify it to build three 303568651Skris packages. The default package contains applications, application 303668651Skris documentation and run-time libraries. The devel package contains 303768651Skris include files, static libraries and function documentation. The 303868651Skris doc package contains the contents of the doc directory. The original 303968651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 304068651Skris [Richard Levitte] 304168651Skris 304268651Skris *) Add a large number of documentation files for many SSL routines. 304368651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 304468651Skris 304568651Skris *) Add a configuration entry for Sony News 4. 304668651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 304768651Skris 304868651Skris *) Don't set the two most significant bits to one when generating a 304968651Skris random number < q in the DSA library. 305068651Skris [Ulf M�ller] 305168651Skris 305268651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 305368651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 305468651Skris the underlying transport is blocking) if a handshake took place. 305568651Skris (The default behaviour is needed by applications such as s_client 305668651Skris and s_server that use select() to determine when to use SSL_read; 305768651Skris but for applications that know in advance when to expect data, it 305868651Skris just makes things more complicated.) 305968651Skris [Bodo Moeller] 306068651Skris 306168651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 306268651Skris from EGD. 306368651Skris [Ben Laurie] 306468651Skris 306568651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 306668651Skris work better on such systems. 306768651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 306868651Skris 306968651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 307068651Skris Update PKCS12_parse() so it copies the friendlyName and the 307168651Skris keyid to the certificates aux info. 307268651Skris [Steve Henson] 307368651Skris 307468651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 307568651Skris if there was more than one signature. 307668651Skris [Sven Uszpelkat <su@celocom.de>] 307768651Skris 307868651Skris *) Major change in util/mkdef.pl to include extra information 307968651Skris about each symbol, as well as presentig variables as well 308068651Skris as functions. This change means that there's n more need 308168651Skris to rebuild the .num files when some algorithms are excluded. 308268651Skris [Richard Levitte] 308368651Skris 308468651Skris *) Allow the verify time to be set by an application, 308568651Skris rather than always using the current time. 308668651Skris [Steve Henson] 308768651Skris 308868651Skris *) Phase 2 verify code reorganisation. The certificate 308968651Skris verify code now looks up an issuer certificate by a 309068651Skris number of criteria: subject name, authority key id 309168651Skris and key usage. It also verifies self signed certificates 309268651Skris by the same criteria. The main comparison function is 309368651Skris X509_check_issued() which performs these checks. 309468651Skris 309568651Skris Lot of changes were necessary in order to support this 309668651Skris without completely rewriting the lookup code. 309768651Skris 309868651Skris Authority and subject key identifier are now cached. 309968651Skris 310068651Skris The LHASH 'certs' is X509_STORE has now been replaced 310168651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 310268651Skris LHASH can't store or retrieve multiple objects with 310368651Skris the same hash value. 310468651Skris 310568651Skris As a result various functions (which were all internal 310668651Skris use only) have changed to handle the new X509_STORE 310768651Skris structure. This will break anything that messed round 310868651Skris with X509_STORE internally. 310968651Skris 311068651Skris The functions X509_STORE_add_cert() now checks for an 311168651Skris exact match, rather than just subject name. 311268651Skris 311368651Skris The X509_STORE API doesn't directly support the retrieval 311468651Skris of multiple certificates matching a given criteria, however 311568651Skris this can be worked round by performing a lookup first 311668651Skris (which will fill the cache with candidate certificates) 311768651Skris and then examining the cache for matches. This is probably 311868651Skris the best we can do without throwing out X509_LOOKUP 311968651Skris entirely (maybe later...). 312068651Skris 312168651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 312268651Skris 312368651Skris All certificate lookup operations now go via a get_issuer() 312468651Skris callback. Although this currently uses an X509_STORE it 312568651Skris can be replaced by custom lookups. This is a simple way 312668651Skris to bypass the X509_STORE hackery necessary to make this 312768651Skris work and makes it possible to use more efficient techniques 312868651Skris in future. A very simple version which uses a simple 312968651Skris STACK for its trusted certificate store is also provided 313068651Skris using X509_STORE_CTX_trusted_stack(). 313168651Skris 313268651Skris The verify_cb() and verify() callbacks now have equivalents 313368651Skris in the X509_STORE_CTX structure. 313468651Skris 313568651Skris X509_STORE_CTX also has a 'flags' field which can be used 313668651Skris to customise the verify behaviour. 313768651Skris [Steve Henson] 313868651Skris 313968651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 314068651Skris excludes S/MIME capabilities. 314168651Skris [Steve Henson] 314268651Skris 314368651Skris *) When a certificate request is read in keep a copy of the 314468651Skris original encoding of the signed data and use it when outputing 314568651Skris again. Signatures then use the original encoding rather than 314668651Skris a decoded, encoded version which may cause problems if the 314768651Skris request is improperly encoded. 314868651Skris [Steve Henson] 314968651Skris 315068651Skris *) For consistency with other BIO_puts implementations, call 315168651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 315268651Skris BIO_write(b, ...). 315368651Skris 315468651Skris In BIO_puts, increment b->num_write as in BIO_write. 315568651Skris [Peter.Sylvester@EdelWeb.fr] 315668651Skris 315768651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 315868651Skris BN_zero, we may not return a BIGNUM with an array consisting of 315968651Skris words set to zero.) 316068651Skris [Bodo Moeller] 316168651Skris 316268651Skris *) Avoid calling abort() from within the library when problems are 316368651Skris detected, except if preprocessor symbols have been defined 316468651Skris (such as REF_CHECK, BN_DEBUG etc.). 316568651Skris [Bodo Moeller] 316668651Skris 316768651Skris *) New openssl application 'rsautl'. This utility can be 316868651Skris used for low level RSA operations. DER public key 316968651Skris BIO/fp routines also added. 317068651Skris [Steve Henson] 317168651Skris 317268651Skris *) New Configure entry and patches for compiling on QNX 4. 317368651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 317468651Skris 317568651Skris *) A demo state-machine implementation was sponsored by 317668651Skris Nuron (http://www.nuron.com/) and is now available in 317768651Skris demos/state_machine. 317868651Skris [Ben Laurie] 317968651Skris 318068651Skris *) New options added to the 'dgst' utility for signature 318168651Skris generation and verification. 318268651Skris [Steve Henson] 318368651Skris 318468651Skris *) Unrecognized PKCS#7 content types are now handled via a 318568651Skris catch all ASN1_TYPE structure. This allows unsupported 318668651Skris types to be stored as a "blob" and an application can 318768651Skris encode and decode it manually. 318868651Skris [Steve Henson] 318968651Skris 319068651Skris *) Fix various signed/unsigned issues to make a_strex.c 319168651Skris compile under VC++. 319268651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 319368651Skris 319468651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 319568651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 319668651Skris if passed a NULL BN and its argument was negative. 319768651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 319868651Skris 319968651Skris *) Modification to PKCS#7 encoding routines to output definite 320068651Skris length encoding. Since currently the whole structures are in 320168651Skris memory there's not real point in using indefinite length 320268651Skris constructed encoding. However if OpenSSL is compiled with 320368651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 320468651Skris [Steve Henson] 320568651Skris 320668651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 320768651Skris [Richard Levitte] 320868651Skris 320968651Skris *) Added more prefixes to parse for in the the strings written 321068651Skris through a logging bio, to cover all the levels that are available 321168651Skris through syslog. The prefixes are now: 321268651Skris 321368651Skris PANIC, EMERG, EMR => LOG_EMERG 321468651Skris ALERT, ALR => LOG_ALERT 321568651Skris CRIT, CRI => LOG_CRIT 321668651Skris ERROR, ERR => LOG_ERR 321768651Skris WARNING, WARN, WAR => LOG_WARNING 321868651Skris NOTICE, NOTE, NOT => LOG_NOTICE 321968651Skris INFO, INF => LOG_INFO 322068651Skris DEBUG, DBG => LOG_DEBUG 322168651Skris 322268651Skris and as before, if none of those prefixes are present at the 322368651Skris beginning of the string, LOG_ERR is chosen. 322468651Skris 322568651Skris On Win32, the LOG_* levels are mapped according to this: 322668651Skris 322768651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 322868651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 322968651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 323068651Skris 323168651Skris [Richard Levitte] 323268651Skris 323368651Skris *) Made it possible to reconfigure with just the configuration 323468651Skris argument "reconf" or "reconfigure". The command line arguments 323568651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 323668651Skris and are retrieved from there when reconfiguring. 323768651Skris [Richard Levitte] 323868651Skris 323968651Skris *) MD4 implemented. 324068651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 324168651Skris 324268651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 324368651Skris [Richard Levitte] 324468651Skris 324568651Skris *) The obj_dat.pl script was messing up the sorting of object 324668651Skris names. The reason was that it compared the quoted version 324768651Skris of strings as a result "OCSP" > "OCSP Signing" because 324868651Skris " > SPACE. Changed script to store unquoted versions of 324968651Skris names and add quotes on output. It was also omitting some 325068651Skris names from the lookup table if they were given a default 325168651Skris value (that is if SN is missing it is given the same 325268651Skris value as LN and vice versa), these are now added on the 325368651Skris grounds that if an object has a name we should be able to 325468651Skris look it up. Finally added warning output when duplicate 325568651Skris short or long names are found. 325668651Skris [Steve Henson] 325768651Skris 325868651Skris *) Changes needed for Tandem NSK. 325968651Skris [Scott Uroff <scott@xypro.com>] 326068651Skris 326168651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 326268651Skris RSA_padding_check_SSLv23(), special padding was never detected 326368651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 326468651Skris version rollback attacks was not effective. 326568651Skris 326668651Skris In s23_clnt.c, don't use special rollback-attack detection padding 326768651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 326868651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 326968651Skris SSL 2.0 is the only protocol enabled in the server. 327068651Skris [Bodo Moeller] 327168651Skris 327268651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 327368651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 327468651Skris BIO_dump_indent() are added. 327568651Skris [Richard Levitte] 327668651Skris 327768651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 327868651Skris these print out strings and name structures based on various 327968651Skris flags including RFC2253 support and proper handling of 328068651Skris multibyte characters. Added options to the 'x509' utility 328168651Skris to allow the various flags to be set. 328268651Skris [Steve Henson] 328368651Skris 328468651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 328568651Skris Also change the functions X509_cmp_current_time() and 328668651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 328768651Skris this will enable certificates using GeneralizedTime in validity 328868651Skris dates to be checked. 328968651Skris [Steve Henson] 329068651Skris 329168651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 329268651Skris negative public key encodings) on by default, 329368651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 329468651Skris [Steve Henson] 329568651Skris 329668651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 329768651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 329868651Skris the encoding can be trivially obtained from the structure. 329968651Skris [Steve Henson] 330068651Skris 330168651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 330268651Skris not read locks (CRYPTO_r_[un]lock). 330368651Skris [Bodo Moeller] 330468651Skris 330568651Skris *) A first attempt at creating official support for shared 330668651Skris libraries through configuration. I've kept it so the 330768651Skris default is static libraries only, and the OpenSSL programs 330868651Skris are always statically linked for now, but there are 330968651Skris preparations for dynamic linking in place. 331089837Skris This has been tested on Linux and Tru64. 331168651Skris [Richard Levitte] 331268651Skris 331368651Skris *) Randomness polling function for Win9x, as described in: 331468651Skris Peter Gutmann, Software Generation of Practically Strong 331568651Skris Random Numbers. 331668651Skris [Ulf M�ller] 331768651Skris 331868651Skris *) Fix so PRNG is seeded in req if using an already existing 331968651Skris DSA key. 332068651Skris [Steve Henson] 332168651Skris 332268651Skris *) New options to smime application. -inform and -outform 332368651Skris allow alternative formats for the S/MIME message including 332468651Skris PEM and DER. The -content option allows the content to be 332568651Skris specified separately. This should allow things like Netscape 332668651Skris form signing output easier to verify. 332768651Skris [Steve Henson] 332868651Skris 332968651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 333068651Skris [Steve Henson] 333168651Skris 333268651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 333368651Skris STRING types. These convert content octets to and from the 333468651Skris underlying type. The actual tag and length octets are 333568651Skris already assumed to have been read in and checked. These 333668651Skris are needed because all other string types have virtually 333768651Skris identical handling apart from the tag. By having versions 333868651Skris of the ASN1 functions that just operate on content octets 333968651Skris IMPLICIT tagging can be handled properly. It also allows 334068651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 334168651Skris and ASN1_INTEGER are identical apart from the tag. 334268651Skris [Steve Henson] 334368651Skris 334468651Skris *) Change the handling of OID objects as follows: 334568651Skris 334668651Skris - New object identifiers are inserted in objects.txt, following 334768651Skris the syntax given in objects.README. 334868651Skris - objects.pl is used to process obj_mac.num and create a new 334968651Skris obj_mac.h. 335068651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 335168651Skris obj_mac.h. 335268651Skris 335368651Skris This is currently kind of a hack, and the perl code in objects.pl 335468651Skris isn't very elegant, but it works as I intended. The simplest way 335568651Skris to check that it worked correctly is to look in obj_dat.h and 335668651Skris check the array nid_objs and make sure the objects haven't moved 335768651Skris around (this is important!). Additions are OK, as well as 335868651Skris consistent name changes. 335968651Skris [Richard Levitte] 336068651Skris 336168651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 336268651Skris [Bodo Moeller] 336368651Skris 336468651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 336568651Skris The given file adds to whatever has already been seeded into the 336668651Skris random pool through the RANDFILE configuration file option or 336768651Skris environment variable, or the default random state file. 336868651Skris [Richard Levitte] 336968651Skris 337068651Skris *) mkstack.pl now sorts each macro group into lexical order. 337168651Skris Previously the output order depended on the order the files 337268651Skris appeared in the directory, resulting in needless rewriting 337368651Skris of safestack.h . 337468651Skris [Steve Henson] 337568651Skris 337668651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 337768651Skris work arounds for the VC++ problem that it treats func() as 337868651Skris func(void). Also stripped out the parts of mkdef.pl that 337968651Skris added extra typesafe functions: these no longer exist. 338068651Skris [Steve Henson] 338168651Skris 338268651Skris *) Reorganisation of the stack code. The macros are now all 338368651Skris collected in safestack.h . Each macro is defined in terms of 338468651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 338568651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 338668651Skris this has the advantage of retaining type safety without the 338768651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 338868651Skris then the non typesafe macros are used instead. Also modified the 338968651Skris mkstack.pl script to handle the new form. Needs testing to see 339068651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 339168651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 339268651Skris and PKCS12_STACK_OF. 339368651Skris [Steve Henson] 339468651Skris 339568651Skris *) When some versions of IIS use the 'NET' form of private key the 339668651Skris key derivation algorithm is different. Normally MD5(password) is 339768651Skris used as a 128 bit RC4 key. In the modified case 339868651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 339968651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 340068651Skris as the old Netscape_RSA functions except they have an additional 340168651Skris 'sgckey' parameter which uses the modified algorithm. Also added 340268651Skris an -sgckey command line option to the rsa utility. Thanks to 340368651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 340468651Skris algorithm to openssl-dev. 340568651Skris [Steve Henson] 340668651Skris 340768651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 340868651Skris invalid expansion on some systems (SCO 5.0.5 for example). 340968651Skris Corrected to 'c.kname'. 341068651Skris [Phillip Porch <root@theporch.com>] 341168651Skris 341268651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 341368651Skris a STACK of email addresses from a certificate or request, these look 341468651Skris in the subject name and the subject alternative name extensions and 341568651Skris omit any duplicate addresses. 341668651Skris [Steve Henson] 341768651Skris 341868651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 341968651Skris This makes DSA verification about 2 % faster. 342068651Skris [Bodo Moeller] 342168651Skris 342268651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 342368651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 342468651Skris plus overhead for 1024 bit moduli). 342568651Skris This makes exponentiations about 0.5 % faster for 1024 bit 342668651Skris exponents (as measured by "openssl speed rsa2048"). 342768651Skris [Bodo Moeller] 342868651Skris 342968651Skris *) Rename memory handling macros to avoid conflicts with other 343068651Skris software: 343168651Skris Malloc => OPENSSL_malloc 343268651Skris Malloc_locked => OPENSSL_malloc_locked 343368651Skris Realloc => OPENSSL_realloc 343468651Skris Free => OPENSSL_free 343568651Skris [Richard Levitte] 343668651Skris 343768651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 343868651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 343968651Skris [Bodo Moeller] 344068651Skris 344168651Skris *) CygWin32 support. 344268651Skris [John Jarvie <jjarvie@newsguy.com>] 344368651Skris 344468651Skris *) The type-safe stack code has been rejigged. It is now only compiled 344568651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 344668651Skris by default all type-specific stack functions are "#define"d back to 344768651Skris standard stack functions. This results in more streamlined output 344868651Skris but retains the type-safety checking possibilities of the original 344968651Skris approach. 345068651Skris [Geoff Thorpe] 345168651Skris 345268651Skris *) The STACK code has been cleaned up, and certain type declarations 345368651Skris that didn't make a lot of sense have been brought in line. This has 345468651Skris also involved a cleanup of sorts in safestack.h to more correctly 345568651Skris map type-safe stack functions onto their plain stack counterparts. 345668651Skris This work has also resulted in a variety of "const"ifications of 345768651Skris lots of the code, especially "_cmp" operations which should normally 345868651Skris be prototyped with "const" parameters anyway. 345968651Skris [Geoff Thorpe] 346068651Skris 346168651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 346268651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 346368651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 346468651Skris where all of 'md' is used each time the PRNG is used, but 'state' 346568651Skris is used only indexed by a cyclic counter. As entropy may not be 346668651Skris well distributed from the beginning, 'md' is important as a 346768651Skris chaining variable. However, the output function chains only half 346868651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 346968651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 347068651Skris in all of 'state' being rewritten, with the new values depending 347168651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 347268651Skris [Bodo Moeller] 347368651Skris 347468651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 347568651Skris the handshake is continued after ssl_verify_cert_chain(); 347668651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 347768651Skris can lead to 'unexplainable' connection aborts later. 347868651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 347968651Skris 348068651Skris *) Major EVP API cipher revision. 348168651Skris Add hooks for extra EVP features. This allows various cipher 348268651Skris parameters to be set in the EVP interface. Support added for variable 348368651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 348468651Skris setting of RC2 and RC5 parameters. 348568651Skris 348668651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 348768651Skris ciphers. 348868651Skris 348968651Skris Remove lots of duplicated code from the EVP library. For example *every* 349068651Skris cipher init() function handles the 'iv' in the same way according to the 349168651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 349268651Skris for CFB and OFB modes they zero ctx->num. 349368651Skris 349468651Skris New functionality allows removal of S/MIME code RC2 hack. 349568651Skris 349668651Skris Most of the routines have the same form and so can be declared in terms 349768651Skris of macros. 349868651Skris 349968651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 350068651Skris all individual ciphers. If the cipher wants to handle IVs or keys 350168651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 350268651Skris flags. 350368651Skris 350468651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 350568651Skris value: although software versions of the algorithms cannot fail 350668651Skris any installed hardware versions can. 350768651Skris [Steve Henson] 350868651Skris 350968651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 351068651Skris this option is set, tolerate broken clients that send the negotiated 351168651Skris protocol version number instead of the requested protocol version 351268651Skris number. 351368651Skris [Bodo Moeller] 351468651Skris 351568651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 351668651Skris i.e. non-zero for export ciphersuites, zero otherwise. 351768651Skris Previous versions had this flag inverted, inconsistent with 351868651Skris rsa_tmp_cb (..._TMP_RSA_CB). 351968651Skris [Bodo Moeller; problem reported by Amit Chopra] 352068651Skris 352168651Skris *) Add missing DSA library text string. Work around for some IIS 352268651Skris key files with invalid SEQUENCE encoding. 352368651Skris [Steve Henson] 352468651Skris 352568651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 352668651Skris and so on that are implemented in OpenSSL. 352768651Skris [Richard Levitte] 352868651Skris 352968651Skris *) Enhance c_rehash script. Old version would mishandle certificates 353068651Skris with the same subject name hash and wouldn't handle CRLs at all. 353168651Skris Added -fingerprint option to crl utility, to support new c_rehash 353268651Skris features. 353368651Skris [Steve Henson] 353468651Skris 353568651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 353668651Skris [Ulf M�ller] 353768651Skris 353868651Skris *) Fix for SSL server purpose checking. Server checking was 353968651Skris rejecting certificates which had extended key usage present 354068651Skris but no ssl client purpose. 354168651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 354268651Skris 354368651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 354468651Skris is a little unclear about how a blank password is handled. 354568651Skris Since the password in encoded as a BMPString with terminating 354668651Skris double NULL a zero length password would end up as just the 354768651Skris double NULL. However no password at all is different and is 354868651Skris handled differently in the PKCS#12 key generation code. NS 354968651Skris treats a blank password as zero length. MSIE treats it as no 355068651Skris password on export: but it will try both on import. We now do 355168651Skris the same: PKCS12_parse() tries zero length and no password if 355268651Skris the password is set to "" or NULL (NULL is now a valid password: 355368651Skris it wasn't before) as does the pkcs12 application. 355468651Skris [Steve Henson] 355568651Skris 355668651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 355768651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 355868651Skris be obtained from the error queue. 355968651Skris [Bodo Moeller] 356068651Skris 356168651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 356268651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 356368651Skris accordingly to avoid race conditions (this is necessary because 356468651Skris thread_hash is no longer constant once set). 356568651Skris [Bodo Moeller] 356668651Skris 356768651Skris *) Bugfix for linux-elf makefile.one. 356868651Skris [Ulf M�ller] 356968651Skris 357068651Skris *) RSA_get_default_method() will now cause a default 357168651Skris RSA_METHOD to be chosen if one doesn't exist already. 357268651Skris Previously this was only set during a call to RSA_new() 357368651Skris or RSA_new_method(NULL) meaning it was possible for 357468651Skris RSA_get_default_method() to return NULL. 357568651Skris [Geoff Thorpe] 357668651Skris 357768651Skris *) Added native name translation to the existing DSO code 357868651Skris that will convert (if the flag to do so is set) filenames 357968651Skris that are sufficiently small and have no path information 358068651Skris into a canonical native form. Eg. "blah" converted to 358168651Skris "libblah.so" or "blah.dll" etc. 358268651Skris [Geoff Thorpe] 358368651Skris 358468651Skris *) New function ERR_error_string_n(e, buf, len) which is like 358568651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 358668651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 358768651Skris may not be NULL. 358868651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 358968651Skris 359068651Skris *) CONF library reworked to become more general. A new CONF 359168651Skris configuration file reader "class" is implemented as well as a 359268651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 359368651Skris old CONF_* functions are still there, but are reimplemented to 359468651Skris work in terms of the new functions. Also, a set of functions 359568651Skris to handle the internal storage of the configuration data is 359668651Skris provided to make it easier to write new configuration file 359768651Skris reader "classes" (I can definitely see something reading a 359868651Skris configuration file in XML format, for example), called _CONF_*, 359968651Skris or "the configuration storage API"... 360068651Skris 360168651Skris The new configuration file reading functions are: 360268651Skris 360368651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 360468651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 360568651Skris 360668651Skris NCONF_default, NCONF_WIN32 360768651Skris 360868651Skris NCONF_dump_fp, NCONF_dump_bio 360968651Skris 361068651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 361168651Skris NCONF_new creates a new CONF object. This works in the same way 361268651Skris as other interfaces in OpenSSL, like the BIO interface. 361368651Skris NCONF_dump_* dump the internal storage of the configuration file, 361468651Skris which is useful for debugging. All other functions take the same 361568651Skris arguments as the old CONF_* functions wth the exception of the 361668651Skris first that must be a `CONF *' instead of a `LHASH *'. 361768651Skris 361868651Skris To make it easer to use the new classes with the old CONF_* functions, 361968651Skris the function CONF_set_default_method is provided. 362068651Skris [Richard Levitte] 362168651Skris 362268651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 362368651Skris mentioned in the documentation but had not been implemented. 362468651Skris (This option is not yet really useful because even the additional 362568651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 362668651Skris [Bodo Moeller] 362768651Skris 362868651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 362968651Skris OpenSSL-based applications) load shared libraries and bind to 363068651Skris them in a portable way. 363168651Skris [Geoff Thorpe, with contributions from Richard Levitte] 363268651Skris 363359191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 363459191Skris 363559191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 363659191Skris 363759191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 363859191Skris (the default implementation of RAND_status). 363959191Skris 364059191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 364159191Skris to '-clrext' (= clear extensions), as intended and documented. 364259191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 364359191Skris <attili@amaxo.com>] 364459191Skris 364559191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 364659191Skris was larger than the MD block size. 364759191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 364859191Skris 364959191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 365059191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 365159191Skris using the passed key: if the passed key was a private key the result 365259191Skris of X509_print(), for example, would be to print out all the private key 365359191Skris components. 365459191Skris [Steve Henson] 365559191Skris 365659191Skris *) des_quad_cksum() byte order bug fix. 365759191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 365859191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 365959191Skris 366059191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 366159191Skris discouraged. 366259191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 366359191Skris 366459191Skris *) For easily testing in shell scripts whether some command 366559191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 366659191Skris returns with exit code 0 iff no command of the given name is available. 366759191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 366859191Skris the output goes to stdout and nothing is printed to stderr. 366959191Skris Additional arguments are always ignored. 367059191Skris 367159191Skris Since for each cipher there is a command of the same name, 367259191Skris the 'no-cipher' compilation switches can be tested this way. 367359191Skris 367459191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 367559191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 367659191Skris [Bodo Moeller] 367759191Skris 367859191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 367959191Skris [Bodo Moeller] 368059191Skris 368159191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 368259191Skris is set; it will be thrown away anyway because each handshake creates 368359191Skris its own key. 368459191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 368559191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 368659191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 368759191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 368859191Skris [Bodo Moeller] 368959191Skris 369059191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 369159191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 369259191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 369359191Skris does not suppress any output. 369459191Skris [Richard Levitte] 369559191Skris 369659191Skris *) Add compatibility options to the purpose and trust code. The 369759191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 369859191Skris accepts a certificate or CA, this was the previous behaviour, 369959191Skris with all the associated security issues. 370059191Skris 370159191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 370259191Skris automatically trust self signed roots in certificate store. A 370359191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 370459191Skris a purpose has no associated trust setting and it should instead 370559191Skris use the value in the default purpose. 370659191Skris [Steve Henson] 370759191Skris 370859191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 370959191Skris and fix a memory leak. 371059191Skris [Steve Henson] 371159191Skris 371259191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 371359191Skris reason strings from the previous version of the .c file, as 371459191Skris the default to have only downcase letters (and digits) in 371559191Skris automatically generated reasons codes is not always appropriate. 371659191Skris [Bodo Moeller] 371759191Skris 371859191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 371959191Skris using strerror. Previously, ERR_reason_error_string() returned 372059191Skris library names as reason strings for SYSerr; but SYSerr is a special 372159191Skris case where small numbers are errno values, not library numbers. 372259191Skris [Bodo Moeller] 372359191Skris 372459191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 372559191Skris converts DSA parameters into DH parameters. (When creating parameters, 372659191Skris DSA_generate_parameters is used.) 372759191Skris [Bodo Moeller] 372859191Skris 372959191Skris *) Include 'length' (recommended exponent length) in C code generated 373059191Skris by 'openssl dhparam -C'. 373159191Skris [Bodo Moeller] 373259191Skris 373359191Skris *) The second argument to set_label in perlasm was already being used 373459191Skris so couldn't be used as a "file scope" flag. Moved to third argument 373559191Skris which was free. 373659191Skris [Steve Henson] 373759191Skris 373859191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 373959191Skris instead of RAND_bytes for encryption IVs and salts. 374059191Skris [Bodo Moeller] 374159191Skris 374259191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 374359191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 374459191Skris RAND_set_rand_method would be impossible. 374559191Skris [Bodo Moeller] 374659191Skris 374759191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 374859191Skris number generation fails. 374959191Skris [Bodo Moeller] 375059191Skris 375159191Skris *) New 'rand' application for creating pseudo-random output. 375259191Skris [Bodo Moeller] 375359191Skris 375459191Skris *) Added configuration support for Linux/IA64 375559191Skris [Rolf Haberrecker <rolf@suse.de>] 375659191Skris 375759191Skris *) Assembler module support for Mingw32. 375859191Skris [Ulf M�ller] 375959191Skris 376059191Skris *) Shared library support for HPUX (in shlib/). 376159191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 376259191Skris 376359191Skris *) Shared library support for Solaris gcc. 376459191Skris [Lutz Behnke <behnke@trustcenter.de>] 376559191Skris 376659191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 376759191Skris 376859191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 376959191Skris were added manually and by SMIME_crlf_copy(). 377059191Skris [Steve Henson] 377159191Skris 377259191Skris *) In bntest.c don't call BN_rand with zero bits argument. 377359191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 377459191Skris 377559191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 377659191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 377759191Skris [Ulf M�ller] 377859191Skris 377959191Skris *) Add an optional second argument to the set_label() in the perl 378059191Skris assembly language builder. If this argument exists and is set 378159191Skris to 1 it signals that the assembler should use a symbol whose 378259191Skris scope is the entire file, not just the current function. This 378359191Skris is needed with MASM which uses the format label:: for this scope. 378459191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 378559191Skris 378659191Skris *) Change the ASN1 types so they are typedefs by default. Before 378759191Skris almost all types were #define'd to ASN1_STRING which was causing 378859191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 378959191Skris for example. 379059191Skris [Steve Henson] 379159191Skris 379259191Skris *) Change names of new functions to the new get1/get0 naming 379359191Skris convention: After 'get1', the caller owns a reference count 379459191Skris and has to call ..._free; 'get0' returns a pointer to some 379559191Skris data structure without incrementing reference counters. 379659191Skris (Some of the existing 'get' functions increment a reference 379759191Skris counter, some don't.) 379859191Skris Similarly, 'set1' and 'add1' functions increase reference 379959191Skris counters or duplicate objects. 380059191Skris [Steve Henson] 380159191Skris 380259191Skris *) Allow for the possibility of temp RSA key generation failure: 380359191Skris the code used to assume it always worked and crashed on failure. 380459191Skris [Steve Henson] 380559191Skris 380659191Skris *) Fix potential buffer overrun problem in BIO_printf(). 380759191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 380859191Skris pointed out by David Sacerdote <das33@cornell.edu>] 380959191Skris 381059191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 381159191Skris RAND_egd() and RAND_status(). In the command line application, 381259191Skris the EGD socket can be specified like a seed file using RANDFILE 381359191Skris or -rand. 381459191Skris [Ulf M�ller] 381559191Skris 381659191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 381759191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 381859191Skris [Steve Henson] 381959191Skris 382059191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 382159191Skris list to exclude them. This means that no special compilation option 382259191Skris is needed to use anonymous DH: it just needs to be included in the 382359191Skris cipher list. 382459191Skris [Steve Henson] 382559191Skris 382659191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 382759191Skris EVP_MD_type. The old functionality is available in a new macro called 382859191Skris EVP_MD_md(). Change code that uses it and update docs. 382959191Skris [Steve Henson] 383059191Skris 383159191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 383259191Skris where the 'void *' argument is replaced by a function pointer argument. 383359191Skris Previously 'void *' was abused to point to functions, which works on 383459191Skris many platforms, but is not correct. As these functions are usually 383559191Skris called by macros defined in OpenSSL header files, most source code 383659191Skris should work without changes. 383759191Skris [Richard Levitte] 383859191Skris 383959191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 384059191Skris sections with information on -D... compiler switches used for 384159191Skris compiling the library so that applications can see them. To enable 384259191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 384359191Skris must be defined. E.g., 384459191Skris #define OPENSSL_ALGORITHM_DEFINES 384559191Skris #include <openssl/opensslconf.h> 384659191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 384759191Skris [Richard Levitte, Ulf and Bodo M�ller] 384859191Skris 384959191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 385059191Skris record layer. 385159191Skris [Bodo Moeller] 385259191Skris 385359191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 385459191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 385559191Skris the required ASN1 format: arbitrary types determined by an OID. 385659191Skris [Steve Henson] 385759191Skris 385859191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 385959191Skris argument to 'req'. This is not because the function is newer or 386059191Skris better than others it just uses the work 'NEW' in the certificate 386159191Skris request header lines. Some software needs this. 386259191Skris [Steve Henson] 386359191Skris 386459191Skris *) Reorganise password command line arguments: now passwords can be 386559191Skris obtained from various sources. Delete the PEM_cb function and make 386659191Skris it the default behaviour: i.e. if the callback is NULL and the 386759191Skris usrdata argument is not NULL interpret it as a null terminated pass 386859191Skris phrase. If usrdata and the callback are NULL then the pass phrase 386959191Skris is prompted for as usual. 387059191Skris [Steve Henson] 387159191Skris 387259191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 387359191Skris the support is automatically enabled. The resulting binaries will 387459191Skris autodetect the card and use it if present. 387559191Skris [Ben Laurie and Compaq Inc.] 387659191Skris 387759191Skris *) Work around for Netscape hang bug. This sends certificate request 387859191Skris and server done in one record. Since this is perfectly legal in the 387959191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 388059191Skris the bugs/SSLv3 entry for more info. 388159191Skris [Steve Henson] 388259191Skris 388359191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 388459191Skris [Andy Polyakov] 388559191Skris 388659191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 388759191Skris of seed file. 388859191Skris [Steve Henson] 388959191Skris 389059191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 389159191Skris [Bodo Moeller] 389259191Skris 389359191Skris *) Add command line password options to the remaining applications. 389459191Skris [Steve Henson] 389559191Skris 389659191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 389759191Skris bits. 389859191Skris [Ulf M�ller] 389959191Skris 390059191Skris *) More tests in bntest.c, and changed test_bn output. 390159191Skris [Ulf M�ller] 390259191Skris 390359191Skris *) ./config recognizes MacOS X now. 390459191Skris [Andy Polyakov] 390559191Skris 390659191Skris *) Bug fix for BN_div() when the first words of num and divsor are 390759191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 390859191Skris [Ulf M�ller] 390959191Skris 391059191Skris *) Add support for various broken PKCS#8 formats, and command line 391159191Skris options to produce them. 391259191Skris [Steve Henson] 391359191Skris 391459191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 391559191Skris get temporary BIGNUMs from a BN_CTX. 391659191Skris [Ulf M�ller] 391759191Skris 391859191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 391959191Skris for p == 0. 392059191Skris [Ulf M�ller] 392159191Skris 392259191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 392359191Skris include a #define from the old name to the new. The original intent 392459191Skris was that statically linked binaries could for example just call 392559191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 392659191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 392759191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 392859191Skris one would link with the other. They are now in separate source files. 392959191Skris [Steve Henson] 393059191Skris 393159191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 393259191Skris [Steve Henson] 393359191Skris 393459191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 393559191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 393659191Skris loop, our standard modexp algorithms are faster). 393759191Skris [Bodo Moeller] 393859191Skris 393959191Skris *) Support for the EBCDIC character set completed. 394059191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 394159191Skris 394259191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 394359191Skris use void * instead of char * in lhash. 394459191Skris [Ulf M�ller] 394559191Skris 394659191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 394759191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 394859191Skris this the server could overwrite ephemeral keys that the client 394959191Skris has already seen). 395059191Skris [Bodo Moeller] 395159191Skris 395259191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 395359191Skris using 50 iterations of the Rabin-Miller test. 395459191Skris 395559191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 395659191Skris iterations of the Rabin-Miller test as required by the appendix 395759191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 395859191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 395959191Skris generation becomes much faster. 396059191Skris 396159191Skris This implies a change for the callback functions in DSA_is_prime 396259191Skris and DSA_generate_parameters: The callback function is called once 396359191Skris for each positive witness in the Rabin-Miller test, not just 396459191Skris occasionally in the inner loop; and the parameters to the 396559191Skris callback function now provide an iteration count for the outer 396659191Skris loop rather than for the current invocation of the inner loop. 396759191Skris DSA_generate_parameters additionally can call the callback 396859191Skris function with an 'iteration count' of -1, meaning that a 396959191Skris candidate has passed the trial division test (when q is generated 397059191Skris from an application-provided seed, trial division is skipped). 397159191Skris [Bodo Moeller] 397259191Skris 397359191Skris *) New function BN_is_prime_fasttest that optionally does trial 397459191Skris division before starting the Rabin-Miller test and has 397559191Skris an additional BN_CTX * argument (whereas BN_is_prime always 397659191Skris has to allocate at least one BN_CTX). 397759191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 397859191Skris trial division stage. 397959191Skris [Bodo Moeller] 398059191Skris 398159191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 398259191Skris as ASN1_TIME. 398359191Skris [Steve Henson] 398459191Skris 398559191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 398659191Skris [Steve Henson] 398759191Skris 398859191Skris *) New function BN_pseudo_rand(). 398959191Skris [Ulf M�ller] 399059191Skris 399159191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 399259191Skris bignum version of BN_from_montgomery() with the working code from 399359191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 399459191Skris the comments. 399559191Skris [Ulf M�ller] 399659191Skris 399759191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 399859191Skris made it impossible to use the same SSL_SESSION data structure in 399959191Skris SSL2 clients in multiple threads. 400059191Skris [Bodo Moeller] 400159191Skris 400259191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 400359191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 400459191Skris to seed the PRNG (previously an explicit byte count was required). 400559191Skris [Ulf M�ller, Bodo M�ller] 400659191Skris 400759191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 400859191Skris used (char *) instead of (void *) and had casts all over the place. 400959191Skris [Steve Henson] 401059191Skris 401159191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 401259191Skris [Ulf M�ller] 401359191Skris 401459191Skris *) Retain source code compatibility for BN_prime_checks macro: 401559191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 401659191Skris BN_prime_checks_for_size to determine the appropriate number of 401759191Skris Rabin-Miller iterations. 401859191Skris [Ulf M�ller] 401959191Skris 402059191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 402159191Skris DH_CHECK_P_NOT_SAFE_PRIME. 402259191Skris (Check if this is true? OpenPGP calls them "strong".) 402359191Skris [Ulf M�ller] 402459191Skris 402559191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 402659191Skris "dhparam". The old programs are retained for now but will handle DH keys 402759191Skris (instead of parameters) in future. 402859191Skris [Steve Henson] 402959191Skris 403059191Skris *) Make the ciphers, s_server and s_client programs check the return values 403159191Skris when a new cipher list is set. 403259191Skris [Steve Henson] 403359191Skris 403459191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 403559191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 403659191Skris wrong. 403759191Skris 403859191Skris The syntax for the cipher sorting has been extended to support sorting by 403959191Skris cipher-strength (using the strength_bits hard coded in the tables). 404059191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 404159191Skris 404259191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 404359191Skris string with an "undefined" symbol (neither command nor alphanumeric 404459191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 404559191Skris an error is flagged. 404659191Skris 404759191Skris Due to the strength-sorting extension, the code of the 404859191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 404959191Skris the readability was also increased :-) 405059191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 405159191Skris 405259191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 405359191Skris for the first serial number and places 2 in the serial number file. This 405459191Skris avoids problems when the root CA is created with serial number zero and 405559191Skris the first user certificate has the same issuer name and serial number 405659191Skris as the root CA. 405759191Skris [Steve Henson] 405859191Skris 405959191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 406059191Skris the new code. Add documentation for this stuff. 406159191Skris [Steve Henson] 406259191Skris 406359191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 406459191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 406559191Skris structures and behave in an analagous way to the X509v3 functions: 406659191Skris they shouldn't be called directly but wrapper functions should be used 406759191Skris instead. 406859191Skris 406959191Skris So we also now have some wrapper functions that call the X509at functions 407059191Skris when passed certificate requests. (TO DO: similar things can be done with 407159191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 407259191Skris things. Some of these need some d2i or i2d and print functionality 407359191Skris because they handle more complex structures.) 407459191Skris [Steve Henson] 407559191Skris 407659191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 407759191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 407859191Skris NO_RSA in ssl/s2*.c. 407959191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 408059191Skris 408159191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 408259191Skris has a return value which indicates the quality of the random data 408359191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 408459191Skris error queue. New function RAND_pseudo_bytes() generates output that is 408559191Skris guaranteed to be unique but not unpredictable. RAND_add is like 408659191Skris RAND_seed, but takes an extra argument for an entropy estimate 408759191Skris (RAND_seed always assumes full entropy). 408859191Skris [Ulf M�ller] 408959191Skris 409059191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 409159191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 409259191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 409359191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 409459191Skris false-positive rate of at most 2^-80 for random input. 409559191Skris [Bodo Moeller] 409659191Skris 409759191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 409859191Skris [Bodo Moeller] 409959191Skris 410059191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 410159191Skris in the 0.9.5 release), this returns the chain 410259191Skris from an X509_CTX structure with a dup of the stack and all 410359191Skris the X509 reference counts upped: so the stack will exist 410459191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 410559191Skris to use this. 410659191Skris 410759191Skris Also make SSL_SESSION_print() print out the verify return 410859191Skris code. 410959191Skris [Steve Henson] 411059191Skris 411159191Skris *) Add manpage for the pkcs12 command. Also change the default 411259191Skris behaviour so MAC iteration counts are used unless the new 411359191Skris -nomaciter option is used. This improves file security and 411459191Skris only older versions of MSIE (4.0 for example) need it. 411559191Skris [Steve Henson] 411659191Skris 411759191Skris *) Honor the no-xxx Configure options when creating .DEF files. 411859191Skris [Ulf M�ller] 411959191Skris 412059191Skris *) Add PKCS#10 attributes to field table: challengePassword, 412159191Skris unstructuredName and unstructuredAddress. These are taken from 412259191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 412359191Skris international characters are used. 412459191Skris 412559191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 412659191Skris based on strings. Remove the 'loc' parameter when adding 412759191Skris attributes because these will be a SET OF encoding which is sorted 412859191Skris in ASN1 order. 412959191Skris [Steve Henson] 413059191Skris 413159191Skris *) Initial changes to the 'req' utility to allow request generation 413259191Skris automation. This will allow an application to just generate a template 413359191Skris file containing all the field values and have req construct the 413459191Skris request. 413559191Skris 413659191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 413759191Skris used all over the place including certificate requests and PKCS#7 413859191Skris structures. They are currently handled manually where necessary with 413959191Skris some primitive wrappers for PKCS#7. The new functions behave in a 414059191Skris manner analogous to the X509 extension functions: they allow 414159191Skris attributes to be looked up by NID and added. 414259191Skris 414359191Skris Later something similar to the X509V3 code would be desirable to 414459191Skris automatically handle the encoding, decoding and printing of the 414559191Skris more complex types. The string types like challengePassword can 414659191Skris be handled by the string table functions. 414759191Skris 414859191Skris Also modified the multi byte string table handling. Now there is 414959191Skris a 'global mask' which masks out certain types. The table itself 415059191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 415159191Skris is useful when for example there is only one permissible type 415259191Skris (as in countryName) and using the mask might result in no valid 415359191Skris types at all. 415459191Skris [Steve Henson] 415559191Skris 415659191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 415759191Skris SSL_get_peer_finished to allow applications to obtain the latest 415859191Skris Finished messages sent to the peer or expected from the peer, 415959191Skris respectively. (SSL_get_peer_finished is usually the Finished message 416059191Skris actually received from the peer, otherwise the protocol will be aborted.) 416159191Skris 416259191Skris As the Finished message are message digests of the complete handshake 416359191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 416459191Skris be used for external authentication procedures when the authentication 416559191Skris provided by SSL/TLS is not desired or is not enough. 416659191Skris [Bodo Moeller] 416759191Skris 416859191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 416959191Skris the host supports BWX extension and if Compaq C is present on the 417059191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 417159191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 417259191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 417359191Skris SHA1. 417459191Skris [Andy Polyakov] 417559191Skris 417659191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 417759191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 417859191Skris weak crypto and after checking the certificate is SGC a second one 417959191Skris with strong crypto. MS SGC stops the first handshake after receiving 418059191Skris the server certificate message and sends a second client hello. Since 418159191Skris a server will typically do all the time consuming operations before 418259191Skris expecting any further messages from the client (server key exchange 418359191Skris is the most expensive) there is little difference between the two. 418459191Skris 418559191Skris To get OpenSSL to support MS SGC we have to permit a second client 418659191Skris hello message after we have sent server done. In addition we have to 418759191Skris reset the MAC if we do get this second client hello. 418859191Skris [Steve Henson] 418959191Skris 419059191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 419159191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 419259191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 419359191Skris format DER encoded private key. Newer code should use PKCS#8 format which 419459191Skris has the key type encoded in the ASN1 structure. Added DER private key 419559191Skris support to pkcs8 application. 419659191Skris [Steve Henson] 419759191Skris 419859191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 419959191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 420059191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 420159191Skris is set, we interpret this as a request to violate the specification 420259191Skris (the worst that can happen is a handshake failure, and 'correct' 420359191Skris behaviour would result in a handshake failure anyway). 420459191Skris [Bodo Moeller] 420559191Skris 420659191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 420759191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 420859191Skris concurrently obtain them from an external cache). 420959191Skris The internal cache can handle only one SSL_SESSION with a given ID, 421059191Skris so if there's a conflict, we now throw out the old one to achieve 421159191Skris consistency. 421259191Skris [Bodo Moeller] 421359191Skris 421459191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 421559191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 421659191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 421759191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 421859191Skris example. 421959191Skris [Steve Henson] 422059191Skris 422159191Skris *) Simplify the trust setting structure and code. Now we just have 422259191Skris two sequences of OIDs for trusted and rejected settings. These will 422359191Skris typically have values the same as the extended key usage extension 422459191Skris and any application specific purposes. 422559191Skris 422659191Skris The trust checking code now has a default behaviour: it will just 422759191Skris check for an object with the same NID as the passed id. Functions can 422859191Skris be provided to override either the default behaviour or the behaviour 422959191Skris for a given id. SSL client, server and email already have functions 423059191Skris in place for compatibility: they check the NID and also return "trusted" 423159191Skris if the certificate is self signed. 423259191Skris [Steve Henson] 423359191Skris 423459191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 423559191Skris traditional format into an EVP_PKEY structure. 423659191Skris [Steve Henson] 423759191Skris 423859191Skris *) Add a password callback function PEM_cb() which either prompts for 423959191Skris a password if usr_data is NULL or otherwise assumes it is a null 424059191Skris terminated password. Allow passwords to be passed on command line 424159191Skris environment or config files in a few more utilities. 424259191Skris [Steve Henson] 424359191Skris 424459191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 424559191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 424659191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 424759191Skris Update documentation. 424859191Skris [Steve Henson] 424959191Skris 425059191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 425159191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 425259191Skris and produce an error if it couldn't. For compatibility we also have 425359191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 425459191Skris don't allocate anything because they don't need to. 425559191Skris [Steve Henson] 425659191Skris 425759191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 425859191Skris for details. 425959191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 426059191Skris 426159191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 426259191Skris possibly others as well. The purpose is to make an interface that 426359191Skris provide hooks so anyone can build a separate set of allocation and 426459191Skris deallocation routines to be used by OpenSSL, for example memory 426559191Skris pool implementations, or something else, which was previously hard 426659191Skris since Malloc(), Realloc() and Free() were defined as macros having 426759191Skris the values malloc, realloc and free, respectively (except for Win32 426859191Skris compilations). The same is provided for memory debugging code. 426959191Skris OpenSSL already comes with functionality to find memory leaks, but 427059191Skris this gives people a chance to debug other memory problems. 427159191Skris 427259191Skris With these changes, a new set of functions and macros have appeared: 427359191Skris 427468651Skris CRYPTO_set_mem_debug_functions() [F] 427568651Skris CRYPTO_get_mem_debug_functions() [F] 427668651Skris CRYPTO_dbg_set_options() [F] 427768651Skris CRYPTO_dbg_get_options() [F] 427868651Skris CRYPTO_malloc_debug_init() [M] 427959191Skris 428059191Skris The memory debug functions are NULL by default, unless the library 428159191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 428259191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 428359191Skris gives the standard debugging functions that come with OpenSSL) or 428459191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 428559191Skris provided by the library user) must be used. When the standard 428659191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 428759191Skris request additional information: 428859191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 428959191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 429059191Skris 429159191Skris Also, things like CRYPTO_set_mem_functions will always give the 429259191Skris expected result (the new set of functions is used for allocation 429359191Skris and deallocation) at all times, regardless of platform and compiler 429459191Skris options. 429559191Skris 429659191Skris To finish it up, some functions that were never use in any other 429759191Skris way than through macros have a new API and new semantic: 429859191Skris 429959191Skris CRYPTO_dbg_malloc() 430059191Skris CRYPTO_dbg_realloc() 430159191Skris CRYPTO_dbg_free() 430259191Skris 430359191Skris All macros of value have retained their old syntax. 430459191Skris [Richard Levitte and Bodo Moeller] 430559191Skris 430659191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 430759191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 430859191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 430959191Skris algorithm. 431059191Skris [Steve Henson] 431159191Skris 431259191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 431359191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 431459191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 431559191Skris 431659191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 431759191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 431859191Skris functionality to handle multipart/signed properly) and a utility 431959191Skris called 'smime' to call all this stuff. This is based on code I 432059191Skris originally wrote for Celo who have kindly allowed it to be 432159191Skris included in OpenSSL. 432259191Skris [Steve Henson] 432359191Skris 432459191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 432559191Skris des_set_key (aka des_key_sched). Global variable des_check_key 432659191Skris decides which of these is called by des_set_key; this way 432759191Skris des_check_key behaves as it always did, but applications and 432859191Skris the library itself, which was buggy for des_check_key == 1, 432959191Skris have a cleaner way to pick the version they need. 433059191Skris [Bodo Moeller] 433159191Skris 433259191Skris *) New function PKCS12_newpass() which changes the password of a 433359191Skris PKCS12 structure. 433459191Skris [Steve Henson] 433559191Skris 433659191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 433759191Skris dynamic mix. In both cases the ids can be used as an index into the 433859191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 433959191Skris functions so they accept a list of the field values and the 434059191Skris application doesn't need to directly manipulate the X509_TRUST 434159191Skris structure. 434259191Skris [Steve Henson] 434359191Skris 434459191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 434559191Skris need initialising. 434659191Skris [Steve Henson] 434759191Skris 434859191Skris *) Modify the way the V3 extension code looks up extensions. This now 434959191Skris works in a similar way to the object code: we have some "standard" 435059191Skris extensions in a static table which is searched with OBJ_bsearch() 435159191Skris and the application can add dynamic ones if needed. The file 435259191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 435359191Skris updated whenever a new extension is added to the core code and kept 435459191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 435559191Skris this. New extensions are not added too often so this file can readily 435659191Skris be maintained manually. 435759191Skris 435859191Skris There are two big advantages in doing things this way. The extensions 435959191Skris can be looked up immediately and no longer need to be "added" using 436059191Skris X509V3_add_standard_extensions(): this function now does nothing. 436159191Skris [Side note: I get *lots* of email saying the extension code doesn't 436259191Skris work because people forget to call this function] 436359191Skris Also no dynamic allocation is done unless new extensions are added: 436459191Skris so if we don't add custom extensions there is no need to call 436559191Skris X509V3_EXT_cleanup(). 436659191Skris [Steve Henson] 436759191Skris 436859191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 436959191Skris magic header, so unsalted files fail gracefully instead of just decrypting 437059191Skris to garbage. This is because not salting is a big security hole, so people 437159191Skris should be discouraged from doing it. 437259191Skris [Ben Laurie] 437359191Skris 437459191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 437559191Skris digest to be passed on the command line but it only used this 437659191Skris parameter when signing a certificate. Modified so all relevant 437759191Skris operations are affected by the digest parameter including the 437859191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 437959191Skris DSA key was used because it didn't fix the digest. 438059191Skris [Steve Henson] 438159191Skris 438259191Skris *) Initial certificate chain verify code. Currently tests the untrusted 438359191Skris certificates for consistency with the verify purpose (which is set 438459191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 438559191Skris 438659191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 438759191Skris this is because it will reject chains with invalid extensions whereas 438859191Skris every previous version of OpenSSL and SSLeay made no checks at all. 438959191Skris 439059191Skris Trust code: checks the root CA for the relevant trust settings. Trust 439159191Skris settings have an initial value consistent with the verify purpose: e.g. 439259191Skris if the verify purpose is for SSL client use it expects the CA to be 439359191Skris trusted for SSL client use. However the default value can be changed to 439459191Skris permit custom trust settings: one example of this would be to only trust 439559191Skris certificates from a specific "secure" set of CAs. 439659191Skris 439759191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 439859191Skris which should be used for version portability: especially since the 439959191Skris verify structure is likely to change more often now. 440059191Skris 440159191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 440259191Skris to set them. If not set then assume SSL clients will verify SSL servers 440359191Skris and vice versa. 440459191Skris 440559191Skris Two new options to the verify program: -untrusted allows a set of 440659191Skris untrusted certificates to be passed in and -purpose which sets the 440759191Skris intended purpose of the certificate. If a purpose is set then the 440859191Skris new chain verify code is used to check extension consistency. 440959191Skris [Steve Henson] 441059191Skris 441159191Skris *) Support for the authority information access extension. 441259191Skris [Steve Henson] 441359191Skris 441459191Skris *) Modify RSA and DSA PEM read routines to transparently handle 441559191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 441659191Skris public keys in a format compatible with certificate 441759191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 441859191Skris functions called *_PublicKey_* which used various odd formats so 441959191Skris these are retained for compatibility: however the DSA variants were 442059191Skris never in a public release so they have been deleted. Changed dsa/rsa 442159191Skris utilities to handle the new format: note no releases ever handled public 442259191Skris keys so we should be OK. 442359191Skris 442459191Skris The primary motivation for this change is to avoid the same fiasco 442559191Skris that dogs private keys: there are several incompatible private key 442659191Skris formats some of which are standard and some OpenSSL specific and 442759191Skris require various evil hacks to allow partial transparent handling and 442859191Skris even then it doesn't work with DER formats. Given the option anything 442959191Skris other than PKCS#8 should be dumped: but the other formats have to 443059191Skris stay in the name of compatibility. 443159191Skris 443259191Skris With public keys and the benefit of hindsight one standard format 443359191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 443459191Skris it clearly returns an error if you try to read the wrong kind of key. 443559191Skris 443659191Skris Added a -pubkey option to the 'x509' utility to output the public key. 443759191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 443859191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 443959191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 444059191Skris that do the same as the EVP_PKEY_assign_*() except they up the 444159191Skris reference count of the added key (they don't "swallow" the 444259191Skris supplied key). 444359191Skris [Steve Henson] 444459191Skris 444559191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 444659191Skris CRLs would fail if the file contained no certificates or no CRLs: 444759191Skris added a new function to read in both types and return the number 444859191Skris read: this means that if none are read it will be an error. The 444959191Skris DER versions of the certificate and CRL reader would always fail 445059191Skris because it isn't possible to mix certificates and CRLs in DER format 445159191Skris without choking one or the other routine. Changed this to just read 445259191Skris a certificate: this is the best we can do. Also modified the code 445359191Skris in apps/verify.c to take notice of return codes: it was previously 445459191Skris attempting to read in certificates from NULL pointers and ignoring 445559191Skris any errors: this is one reason why the cert and CRL reader seemed 445659191Skris to work. It doesn't check return codes from the default certificate 445759191Skris routines: these may well fail if the certificates aren't installed. 445859191Skris [Steve Henson] 445959191Skris 446059191Skris *) Code to support otherName option in GeneralName. 446159191Skris [Steve Henson] 446259191Skris 446359191Skris *) First update to verify code. Change the verify utility 446459191Skris so it warns if it is passed a self signed certificate: 446559191Skris for consistency with the normal behaviour. X509_verify 446659191Skris has been modified to it will now verify a self signed 446759191Skris certificate if *exactly* the same certificate appears 446859191Skris in the store: it was previously impossible to trust a 446959191Skris single self signed certificate. This means that: 447059191Skris openssl verify ss.pem 447159191Skris now gives a warning about a self signed certificate but 447259191Skris openssl verify -CAfile ss.pem ss.pem 447359191Skris is OK. 447459191Skris [Steve Henson] 447559191Skris 447659191Skris *) For servers, store verify_result in SSL_SESSION data structure 447759191Skris (and add it to external session representation). 447859191Skris This is needed when client certificate verifications fails, 447959191Skris but an application-provided verification callback (set by 448059191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 448159191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 448259191Skris but returns 1): When the session is reused, we have to set 448359191Skris ssl->verify_result to the appropriate error code to avoid 448459191Skris security holes. 448559191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 448659191Skris 448759191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 448859191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 448959191Skris didn't contain any existing data because it was being created. 449059191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 449159191Skris 449259191Skris *) Add a salt to the key derivation routines in enc.c. This 449359191Skris forms the first 8 bytes of the encrypted file. Also add a 449459191Skris -S option to allow a salt to be input on the command line. 449559191Skris [Steve Henson] 449659191Skris 449759191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 449859191Skris to compare two certificates. We do this by working out the SHA1 449959191Skris hash and comparing that. X509_cmp() will be needed by the trust 450059191Skris code. 450159191Skris [Steve Henson] 450259191Skris 450359191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 450459191Skris the reference count in the SSL_SESSION returned. 450559191Skris [Geoff Thorpe <geoff@eu.c2.net>] 450659191Skris 450759191Skris *) Fix for 'req': it was adding a null to request attributes. 450859191Skris Also change the X509_LOOKUP and X509_INFO code to handle 450959191Skris certificate auxiliary information. 451059191Skris [Steve Henson] 451159191Skris 451259191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 451359191Skris the 'enc' command. 451459191Skris [Steve Henson] 451559191Skris 451659191Skris *) Add the possibility to add extra information to the memory leak 451759191Skris detecting output, to form tracebacks, showing from where each 451859191Skris allocation was originated: CRYPTO_push_info("constant string") adds 451959191Skris the string plus current file name and line number to a per-thread 452059191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 452159191Skris is like calling CYRPTO_pop_info() until the stack is empty. 452259191Skris Also updated memory leak detection code to be multi-thread-safe. 452359191Skris [Richard Levitte] 452459191Skris 452559191Skris *) Add options -text and -noout to pkcs7 utility and delete the 452659191Skris encryption options which never did anything. Update docs. 452759191Skris [Steve Henson] 452859191Skris 452959191Skris *) Add options to some of the utilities to allow the pass phrase 453059191Skris to be included on either the command line (not recommended on 453159191Skris OSes like Unix) or read from the environment. Update the 453259191Skris manpages and fix a few bugs. 453359191Skris [Steve Henson] 453459191Skris 453559191Skris *) Add a few manpages for some of the openssl commands. 453659191Skris [Steve Henson] 453759191Skris 453859191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 453959191Skris leaking and not finding already revoked certificates. 454059191Skris [Steve Henson] 454159191Skris 454259191Skris *) Extensive changes to support certificate auxiliary information. 454359191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 454459191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 454559191Skris can still read in a certificate file in the usual way but it 454659191Skris will also read in any additional "auxiliary information". By 454759191Skris doing things this way a fair degree of compatibility can be 454859191Skris retained: existing certificates can have this information added 454959191Skris using the new 'x509' options. 455059191Skris 455159191Skris Current auxiliary information includes an "alias" and some trust 455259191Skris settings. The trust settings will ultimately be used in enhanced 455359191Skris certificate chain verification routines: currently a certificate 455459191Skris can only be trusted if it is self signed and then it is trusted 455559191Skris for all purposes. 455659191Skris [Steve Henson] 455759191Skris 455859191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 455959191Skris The problem was that one of the replacement routines had not been working 456059191Skris since SSLeay releases. For now the offending routine has been replaced 456159191Skris with non-optimised assembler. Even so, this now gives around 95% 456259191Skris performance improvement for 1024 bit RSA signs. 456359191Skris [Mark Cox] 456459191Skris 456559191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 456659191Skris handling. Most clients have the effective key size in bits equal to 456759191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 456859191Skris A few however don't do this and instead use the size of the decrypted key 456959191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 457059191Skris the effective key length. In this case the effective key length can still 457159191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 457259191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 457359191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 457459191Skris the key length and effective key length are equal. 457559191Skris [Steve Henson] 457659191Skris 457759191Skris *) Add a bunch of functions that should simplify the creation of 457859191Skris X509_NAME structures. Now you should be able to do: 457959191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 458059191Skris and have it automatically work out the correct field type and fill in 458159191Skris the structures. The more adventurous can try: 458259191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 458359191Skris and it will (hopefully) work out the correct multibyte encoding. 458459191Skris [Steve Henson] 458559191Skris 458659191Skris *) Change the 'req' utility to use the new field handling and multibyte 458759191Skris copy routines. Before the DN field creation was handled in an ad hoc 458859191Skris way in req, ca, and x509 which was rather broken and didn't support 458959191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 459059191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 459159191Skris using the dirstring_type option. See the new comment in the default 459259191Skris openssl.cnf for more info. 459359191Skris [Steve Henson] 459459191Skris 459559191Skris *) Make crypto/rand/md_rand.c more robust: 459659191Skris - Assure unique random numbers after fork(). 459759191Skris - Make sure that concurrent threads access the global counter and 459859191Skris md serializably so that we never lose entropy in them 459959191Skris or use exactly the same state in multiple threads. 460059191Skris Access to the large state is not always serializable because 460159191Skris the additional locking could be a performance killer, and 460259191Skris md should be large enough anyway. 460359191Skris [Bodo Moeller] 460459191Skris 460559191Skris *) New file apps/app_rand.c with commonly needed functionality 460659191Skris for handling the random seed file. 460759191Skris 460859191Skris Use the random seed file in some applications that previously did not: 460959191Skris ca, 461059191Skris dsaparam -genkey (which also ignored its '-rand' option), 461159191Skris s_client, 461259191Skris s_server, 461359191Skris x509 (when signing). 461459191Skris Except on systems with /dev/urandom, it is crucial to have a random 461559191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 461659191Skris for RSA signatures we could do without one. 461759191Skris 461859191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 461959191Skris of each file listed in the '-rand' option. The function as previously 462059191Skris found in genrsa is now in app_rand.c and is used by all programs 462159191Skris that support '-rand'. 462259191Skris [Bodo Moeller] 462359191Skris 462459191Skris *) In RAND_write_file, use mode 0600 for creating files; 462559191Skris don't just chmod when it may be too late. 462659191Skris [Bodo Moeller] 462759191Skris 462859191Skris *) Report an error from X509_STORE_load_locations 462959191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 463059191Skris [Bill Perry] 463159191Skris 463259191Skris *) New function ASN1_mbstring_copy() this copies a string in either 463359191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 463459191Skris into an ASN1_STRING type. A mask of permissible types is passed 463559191Skris and it chooses the "minimal" type to use or an error if not type 463659191Skris is suitable. 463759191Skris [Steve Henson] 463859191Skris 463959191Skris *) Add function equivalents to the various macros in asn1.h. The old 464059191Skris macros are retained with an M_ prefix. Code inside the library can 464159191Skris use the M_ macros. External code (including the openssl utility) 464259191Skris should *NOT* in order to be "shared library friendly". 464359191Skris [Steve Henson] 464459191Skris 464559191Skris *) Add various functions that can check a certificate's extensions 464659191Skris to see if it usable for various purposes such as SSL client, 464759191Skris server or S/MIME and CAs of these types. This is currently 464859191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 464959191Skris verification. Also added a -purpose flag to x509 utility to 465059191Skris print out all the purposes. 465159191Skris [Steve Henson] 465259191Skris 465359191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 465459191Skris functions. 465559191Skris [Steve Henson] 465659191Skris 465759191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 465859191Skris for, obtain and decode and extension and obtain its critical flag. 465959191Skris This allows all the necessary extension code to be handled in a 466059191Skris single function call. 466159191Skris [Steve Henson] 466259191Skris 466359191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 466459191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 466559191Skris [Andy Polyakov] 466659191Skris 466759191Skris *) New -noout option to asn1parse. This causes no output to be produced 466859191Skris its main use is when combined with -strparse and -out to extract data 466959191Skris from a file (which may not be in ASN.1 format). 467059191Skris [Steve Henson] 467159191Skris 467259191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 467359191Skris when producing the local key id. 467459191Skris [Richard Levitte <levitte@stacken.kth.se>] 467559191Skris 467659191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 467759191Skris stated explicitly. If it is not stated then it tries the first server 467859191Skris certificate file. The previous behaviour hard coded the filename 467959191Skris "server.pem". 468059191Skris [Steve Henson] 468159191Skris 468259191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 468359191Skris a public key to be input or output. For example: 468459191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 468559191Skris Also added necessary DSA public key functions to handle this. 468659191Skris [Steve Henson] 468759191Skris 468859191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 468959191Skris in the message. This was handled by allowing 469059191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 469159191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 469259191Skris 469359191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 469459191Skris to the end of the strings whereas this didn't. This would cause problems 469559191Skris if strings read with d2i_ASN1_bytes() were later modified. 469659191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 469759191Skris 469859191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 469959191Skris data and it contains EOF it will end up returning an error. This is 470059191Skris caused by input 46 bytes long. The cause is due to the way base64 470159191Skris BIOs find the start of base64 encoded data. They do this by trying a 470259191Skris trial decode on each line until they find one that works. When they 470359191Skris do a flag is set and it starts again knowing it can pass all the 470459191Skris data directly through the decoder. Unfortunately it doesn't reset 470559191Skris the context it uses. This means that if EOF is reached an attempt 470659191Skris is made to pass two EOFs through the context and this causes the 470759191Skris resulting error. This can also cause other problems as well. As is 470859191Skris usual with these problems it takes *ages* to find and the fix is 470959191Skris trivial: move one line. 471059191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 471159191Skris 471259191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 471359191Skris old code wouldn't work because it needed to select() on sockets and the 471459191Skris tty (for keypresses and to see if data could be written). Win32 only 471559191Skris supports select() on sockets so we select() with a 1s timeout on the 471659191Skris sockets and then see if any characters are waiting to be read, if none 471759191Skris are present then we retry, we also assume we can always write data to 471859191Skris the tty. This isn't nice because the code then blocks until we've 471959191Skris received a complete line of data and it is effectively polling the 472059191Skris keyboard at 1s intervals: however it's quite a bit better than not 472159191Skris working at all :-) A dedicated Windows application might handle this 472259191Skris with an event loop for example. 472359191Skris [Steve Henson] 472459191Skris 472559191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 472659191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 472759191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 472859191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 472959191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 473059191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 473159191Skris This necessitated the support of an extra signature type NID_md5_sha1 473259191Skris for SSL signatures and modifications to the SSL library to use it instead 473359191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 473459191Skris [Steve Henson] 473559191Skris 473659191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 473759191Skris will lookup a CRL issuers certificate and verify the signature in a 473859191Skris similar way to the verify program. Tidy up the crl program so it 473959191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 474059191Skris less strict. It will now permit CRL extensions even if it is not 474159191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 474259191Skris [Steve Henson] 474359191Skris 474459191Skris *) Initialize all non-automatic variables each time one of the openssl 474559191Skris sub-programs is started (this is necessary as they may be started 474659191Skris multiple times from the "OpenSSL>" prompt). 474759191Skris [Lennart Bang, Bodo Moeller] 474859191Skris 474959191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 475059191Skris removing all other RSA functionality (this is what NO_RSA does). This 475159191Skris is so (for example) those in the US can disable those operations covered 475259191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 475359191Skris key generation. 475459191Skris [Steve Henson] 475559191Skris 475659191Skris *) Non-copying interface to BIO pairs. 475759191Skris (still largely untested) 475859191Skris [Bodo Moeller] 475959191Skris 476059191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 476159191Skris ASCII string. This was handled independently in various places before. 476259191Skris [Steve Henson] 476359191Skris 476459191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 476559191Skris UTF8 strings a character at a time. 476659191Skris [Steve Henson] 476759191Skris 476859191Skris *) Use client_version from client hello to select the protocol 476959191Skris (s23_srvr.c) and for RSA client key exchange verification 477059191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 477159191Skris [Bodo Moeller] 477259191Skris 477359191Skris *) Add various utility functions to handle SPKACs, these were previously 477459191Skris handled by poking round in the structure internals. Added new function 477559191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 477659191Skris print, verify and generate SPKACs. Based on an original idea from 477759191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 477859191Skris [Steve Henson] 477959191Skris 478059191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 478159191Skris [Andy Polyakov] 478259191Skris 478359191Skris *) Allow the config file extension section to be overwritten on the 478459191Skris command line. Based on an original idea from Massimiliano Pala 478559191Skris <madwolf@comune.modena.it>. The new option is called -extensions 478659191Skris and can be applied to ca, req and x509. Also -reqexts to override 478759191Skris the request extensions in req and -crlexts to override the crl extensions 478859191Skris in ca. 478959191Skris [Steve Henson] 479059191Skris 479159191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 479259191Skris the same field multiple times by preceding it by "XXXX." for example: 479359191Skris 1.OU="Unit name 1" 479459191Skris 2.OU="Unit name 2" 479559191Skris this is the same syntax as used in the req config file. 479659191Skris [Steve Henson] 479759191Skris 479859191Skris *) Allow certificate extensions to be added to certificate requests. These 479959191Skris are specified in a 'req_extensions' option of the req section of the 480059191Skris config file. They can be printed out with the -text option to req but 480159191Skris are otherwise ignored at present. 480259191Skris [Steve Henson] 480359191Skris 480459191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 480559191Skris data read consists of only the final block it would not decrypted because 480659191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 480759191Skris A misplaced 'break' also meant the decrypted final block might not be 480859191Skris copied until the next read. 480959191Skris [Steve Henson] 481059191Skris 481159191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 481259191Skris a few extra parameters to the DH structure: these will be useful if 481359191Skris for example we want the value of 'q' or implement X9.42 DH. 481459191Skris [Steve Henson] 481559191Skris 481659191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 481759191Skris provides hooks that allow the default DSA functions or functions on a 481859191Skris "per key" basis to be replaced. This allows hardware acceleration and 481959191Skris hardware key storage to be handled without major modification to the 482059191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 482159191Skris associated functions. 482259191Skris [Steve Henson] 482359191Skris 482459191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 482559191Skris as "read only": it can't be written to and the buffer it points to will 482659191Skris not be freed. Reading from a read only BIO is much more efficient than 482759191Skris a normal memory BIO. This was added because there are several times when 482859191Skris an area of memory needs to be read from a BIO. The previous method was 482959191Skris to create a memory BIO and write the data to it, this results in two 483059191Skris copies of the data and an O(n^2) reading algorithm. There is a new 483159191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 483259191Skris an area of memory. Also modified the PKCS#7 routines to use read only 483359191Skris memory BIOs. 483459191Skris [Steve Henson] 483559191Skris 483659191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 483759191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 483859191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 483959191Skris but a retry condition occured while trying to read the rest. 484059191Skris [Bodo Moeller] 484159191Skris 484259191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 484359191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 484459191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 484559191Skris the encrypted data type: this is a more sensible place to put it and it 484659191Skris allows the PKCS#12 code to be tidied up that duplicated this 484759191Skris functionality. 484859191Skris [Steve Henson] 484959191Skris 485059191Skris *) Changed obj_dat.pl script so it takes its input and output files on 485159191Skris the command line. This should avoid shell escape redirection problems 485259191Skris under Win32. 485359191Skris [Steve Henson] 485459191Skris 485559191Skris *) Initial support for certificate extension requests, these are included 485659191Skris in things like Xenroll certificate requests. Included functions to allow 485759191Skris extensions to be obtained and added. 485859191Skris [Steve Henson] 485959191Skris 486059191Skris *) -crlf option to s_client and s_server for sending newlines as 486159191Skris CRLF (as required by many protocols). 486259191Skris [Bodo Moeller] 486359191Skris 486455714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 486555714Skris 486655714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 486755714Skris [Ralf S. Engelschall] 486855714Skris 486955714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 487055714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 487155714Skris 487255714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 487355714Skris program. 487455714Skris [Steve Henson] 487555714Skris 487655714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 487755714Skris DH parameters/keys (q is lost during that conversion, but the resulting 487855714Skris DH parameters contain its length). 487955714Skris 488055714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 488155714Skris much faster than DH_generate_parameters (which creates parameters 488255714Skris where p = 2*q + 1), and also the smaller q makes DH computations 488355714Skris much more efficient (160-bit exponentiation instead of 1024-bit 488455714Skris exponentiation); so this provides a convenient way to support DHE 488555714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 488655714Skris utter importance to use 488755714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 488855714Skris or 488955714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 489055714Skris when such DH parameters are used, because otherwise small subgroup 489155714Skris attacks may become possible! 489255714Skris [Bodo Moeller] 489355714Skris 489455714Skris *) Avoid memory leak in i2d_DHparams. 489555714Skris [Bodo Moeller] 489655714Skris 489755714Skris *) Allow the -k option to be used more than once in the enc program: 489855714Skris this allows the same encrypted message to be read by multiple recipients. 489955714Skris [Steve Henson] 490055714Skris 490155714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 490255714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 490355714Skris it will always use the numerical form of the OID, even if it has a short 490455714Skris or long name. 490555714Skris [Steve Henson] 490655714Skris 490755714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 490855714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 490955714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 491055714Skris no private key components need be present and it might store extra data 491159191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 491259191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 491359191Skris private key operations. 491455714Skris [Steve Henson] 491555714Skris 491655714Skris *) Added support for SPARC Linux. 491755714Skris [Andy Polyakov] 491855714Skris 491955714Skris *) pem_password_cb function type incompatibly changed from 492055714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 492155714Skris to 492255714Skris ....(char *buf, int size, int rwflag, void *userdata); 492355714Skris so that applications can pass data to their callbacks: 492455714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 492555714Skris additional void * argument, which is just handed through whenever 492655714Skris the password callback is called. 492759191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 492855714Skris 492955714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 493055714Skris 493155714Skris Compatibility note: As many C implementations push function arguments 493255714Skris onto the stack in reverse order, the new library version is likely to 493355714Skris interoperate with programs that have been compiled with the old 493455714Skris pem_password_cb definition (PEM_whatever takes some data that 493555714Skris happens to be on the stack as its last argument, and the callback 493655714Skris just ignores this garbage); but there is no guarantee whatsoever that 493755714Skris this will work. 493855714Skris 493955714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 494055714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 494155714Skris problems not only on Windows, but also on some Unix platforms. 494255714Skris To avoid problematic command lines, these definitions are now in an 494355714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 494455714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 494555714Skris [Bodo Moeller] 494655714Skris 494755714Skris *) MIPS III/IV assembler module is reimplemented. 494855714Skris [Andy Polyakov] 494955714Skris 495055714Skris *) More DES library cleanups: remove references to srand/rand and 495155714Skris delete an unused file. 495255714Skris [Ulf M�ller] 495355714Skris 495455714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 495555714Skris since not many people have MASM (ml) and it can be hard to obtain. 495655714Skris This is currently experimental but it seems to work OK and pass all 495755714Skris the tests. Check out INSTALL.W32 for info. 495855714Skris [Steve Henson] 495955714Skris 496055714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 496155714Skris without temporary keys kept an extra copy of the server key, 496255714Skris and connections with temporary keys did not free everything in case 496355714Skris of an error. 496455714Skris [Bodo Moeller] 496555714Skris 496655714Skris *) New function RSA_check_key and new openssl rsa option -check 496755714Skris for verifying the consistency of RSA keys. 496855714Skris [Ulf Moeller, Bodo Moeller] 496955714Skris 497055714Skris *) Various changes to make Win32 compile work: 497155714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 497255714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 497355714Skris comparison" warnings. 497455714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 497555714Skris [Steve Henson] 497655714Skris 497755714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 497855714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 497955714Skris derived keys are printed to stderr. 498055714Skris [Steve Henson] 498155714Skris 498255714Skris *) Copy the flags in ASN1_STRING_dup(). 498355714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 498455714Skris 498555714Skris *) The x509 application mishandled signing requests containing DSA 498655714Skris keys when the signing key was also DSA and the parameters didn't match. 498755714Skris 498855714Skris It was supposed to omit the parameters when they matched the signing key: 498955714Skris the verifying software was then supposed to automatically use the CA's 499055714Skris parameters if they were absent from the end user certificate. 499155714Skris 499255714Skris Omitting parameters is no longer recommended. The test was also 499355714Skris the wrong way round! This was probably due to unusual behaviour in 499455714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 499555714Skris This meant that parameters were omitted when they *didn't* match and 499655714Skris the certificate was useless. Certificates signed with 'ca' didn't have 499755714Skris this bug. 499855714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 499955714Skris 500055714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 500155714Skris The interface is as follows: 500255714Skris Applications can use 500355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 500455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 500555714Skris "off" is now the default. 500655714Skris The library internally uses 500755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 500855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 500955714Skris to disable memory-checking temporarily. 501055714Skris 501155714Skris Some inconsistent states that previously were possible (and were 501255714Skris even the default) are now avoided. 501355714Skris 501455714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 501555714Skris with each memory chunk allocated; this is occasionally more helpful 501655714Skris than just having a counter. 501755714Skris 501855714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 501955714Skris 502055714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 502155714Skris extensions. 502255714Skris [Bodo Moeller] 502355714Skris 502455714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 502555714Skris which largely parallels "options", but is for changing API behaviour, 502655714Skris whereas "options" are about protocol behaviour. 502755714Skris Initial "mode" flags are: 502855714Skris 502955714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 503055714Skris a single record has been written. 503155714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 503255714Skris retries use the same buffer location. 503355714Skris (But all of the contents must be 503455714Skris copied!) 503555714Skris [Bodo Moeller] 503655714Skris 503779998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 503855714Skris worked. 503955714Skris 504055714Skris *) Fix problems with no-hmac etc. 504155714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 504255714Skris 504355714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 504455714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 504555714Skris to mess around with the internals of an RSA structure. 504655714Skris [Steve Henson] 504755714Skris 504855714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 504955714Skris Also really enable memory leak checks in openssl.c and in some 505055714Skris test programs. 505155714Skris [Chad C. Mulligan, Bodo Moeller] 505255714Skris 505355714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 505455714Skris up the length of negative integers. This has now been simplified to just 505555714Skris store the length when it is first determined and use it later, rather 505655714Skris than trying to keep track of where data is copied and updating it to 505755714Skris point to the end. 505855714Skris [Steve Henson, reported by Brien Wheeler 505955714Skris <bwheeler@authentica-security.com>] 506055714Skris 506155714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 506255714Skris of a PKCS#7 signature but with the signing certificate passed to the 506355714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 506455714Skris certificate is present in the PKCS#7 structure. This isn't always the 506555714Skris case: certificates can be omitted from a PKCS#7 structure and be 506655714Skris distributed by "out of band" means (such as a certificate database). 506755714Skris [Steve Henson] 506855714Skris 506955714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 507055714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 507155714Skris necessary function names. 507255714Skris [Steve Henson] 507355714Skris 507455714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 507555714Skris options set by Configure in the top level Makefile, and Configure 507655714Skris was not even able to write more than one option correctly. 507755714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 507855714Skris [Bodo Moeller] 507955714Skris 508055714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 508155714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 508255714Skris for example allow memory BIOs to contain config info. 508355714Skris [Steve Henson] 508455714Skris 508555714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 508655714Skris Whoever hopes to achieve shared-library compatibility across versions 508755714Skris must use this, not the compile-time macro. 508855714Skris (Exercise 0.9.4: Which is the minimum library version required by 508955714Skris such programs?) 509055714Skris Note: All this applies only to multi-threaded programs, others don't 509155714Skris need locks. 509255714Skris [Bodo Moeller] 509355714Skris 509455714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 509555714Skris through a BIO pair triggered the default case, i.e. 509655714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 509755714Skris [Bodo Moeller] 509855714Skris 509955714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 510055714Skris can use the SSL library even if none of the specific BIOs is 510155714Skris appropriate. 510255714Skris [Bodo Moeller] 510355714Skris 510455714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 510555714Skris for the encoded length. 510655714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 510755714Skris 510855714Skris *) Add initial documentation of the X509V3 functions. 510955714Skris [Steve Henson] 511055714Skris 511155714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 511255714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 511355714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 511455714Skris secure PKCS#8 private key format with a high iteration count. 511555714Skris [Steve Henson] 511655714Skris 511755714Skris *) Fix determination of Perl interpreter: A perl or perl5 511855714Skris _directory_ in $PATH was also accepted as the interpreter. 511955714Skris [Ralf S. Engelschall] 512055714Skris 512155714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 512255714Skris wrong with it but it was very old and did things like calling 512355714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 512455714Skris unusual formatting. 512555714Skris [Steve Henson] 512655714Skris 512755714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 512855714Skris to use the new extension code. 512955714Skris [Steve Henson] 513055714Skris 513155714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 513255714Skris with macros. This should make it easier to change their form, add extra 513355714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 513455714Skris constant. 513555714Skris [Steve Henson] 513655714Skris 513755714Skris *) Add to configuration table a new entry that can specify an alternative 513855714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 513955714Skris according to Mark Crispin <MRC@Panda.COM>. 514055714Skris [Bodo Moeller] 514155714Skris 514255714Skris#if 0 514355714Skris *) DES CBC did not update the IV. Weird. 514455714Skris [Ben Laurie] 514555714Skris#else 514655714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 514755714Skris Changing the behaviour of the former might break existing programs -- 514855714Skris where IV updating is needed, des_ncbc_encrypt can be used. 514955714Skris#endif 515055714Skris 515155714Skris *) When bntest is run from "make test" it drives bc to check its 515255714Skris calculations, as well as internally checking them. If an internal check 515355714Skris fails, it needs to cause bc to give a non-zero result or make test carries 515455714Skris on without noticing the failure. Fixed. 515555714Skris [Ben Laurie] 515655714Skris 515755714Skris *) DES library cleanups. 515855714Skris [Ulf M�ller] 515955714Skris 516055714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 516155714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 516255714Skris ciphers. NOTE: although the key derivation function has been verified 516355714Skris against some published test vectors it has not been extensively tested 516455714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 516555714Skris of v2.0. 516655714Skris [Steve Henson] 516755714Skris 516855714Skris *) Instead of "mkdir -p", which is not fully portable, use new 516955714Skris Perl script "util/mkdir-p.pl". 517055714Skris [Bodo Moeller] 517155714Skris 517255714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 517355714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 517455714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 517555714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 517655714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 517755714Skris underlying key generation function so it must do its own ASN1 parsing. 517855714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 517955714Skris 'parameter' argument instead of literal salt and iteration count values 518055714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 518155714Skris [Steve Henson] 518255714Skris 518355714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 518455714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 518555714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 518655714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 518755714Skris value was just used as a "magic string" and not used directly its 518855714Skris value doesn't matter. 518955714Skris [Steve Henson] 519055714Skris 519155714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 519255714Skris support mutable. 519355714Skris [Ben Laurie] 519455714Skris 519555714Skris *) "linux-sparc64" configuration (ultrapenguin). 519655714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 519755714Skris "linux-sparc" configuration. 519855714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 519955714Skris 520055714Skris *) config now generates no-xxx options for missing ciphers. 520155714Skris [Ulf M�ller] 520255714Skris 520355714Skris *) Support the EBCDIC character set (work in progress). 520455714Skris File ebcdic.c not yet included because it has a different license. 520555714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 520655714Skris 520755714Skris *) Support BS2000/OSD-POSIX. 520855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 520955714Skris 521055714Skris *) Make callbacks for key generation use void * instead of char *. 521155714Skris [Ben Laurie] 521255714Skris 521355714Skris *) Make S/MIME samples compile (not yet tested). 521455714Skris [Ben Laurie] 521555714Skris 521655714Skris *) Additional typesafe stacks. 521755714Skris [Ben Laurie] 521855714Skris 521955714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 522055714Skris [Bodo Moeller] 522155714Skris 522255714Skris 522355714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 522455714Skris 522555714Skris *) New configuration variant "sco5-gcc". 522655714Skris 522755714Skris *) Updated some demos. 522855714Skris [Sean O Riordain, Wade Scholine] 522955714Skris 523055714Skris *) Add missing BIO_free at exit of pkcs12 application. 523155714Skris [Wu Zhigang] 523255714Skris 523355714Skris *) Fix memory leak in conf.c. 523455714Skris [Steve Henson] 523555714Skris 523655714Skris *) Updates for Win32 to assembler version of MD5. 523755714Skris [Steve Henson] 523855714Skris 523955714Skris *) Set #! path to perl in apps/der_chop to where we found it 524055714Skris instead of using a fixed path. 524155714Skris [Bodo Moeller] 524255714Skris 524355714Skris *) SHA library changes for irix64-mips4-cc. 524455714Skris [Andy Polyakov] 524555714Skris 524655714Skris *) Improvements for VMS support. 524755714Skris [Richard Levitte] 524855714Skris 524955714Skris 525055714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 525155714Skris 525255714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 525355714Skris This also avoids the problems with SC4.2 and unpatched SC5. 525455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 525555714Skris 525655714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 525755714Skris These are required because of the typesafe stack would otherwise break 525855714Skris existing code. If old code used a structure member which used to be STACK 525955714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 526055714Skris sk_num or sk_value it would produce an error because the num, data members 526155714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 526255714Skris replaces the old method of assigning a value to sk_value 526355714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 526455714Skris that does this will no longer work (and should use sk_set instead) but 526555714Skris this could be regarded as a "questionable" behaviour anyway. 526655714Skris [Steve Henson] 526755714Skris 526855714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 526955714Skris correctly handle encrypted S/MIME data. 527055714Skris [Steve Henson] 527155714Skris 527255714Skris *) Change type of various DES function arguments from des_cblock 527355714Skris (which means, in function argument declarations, pointer to char) 527455714Skris to des_cblock * (meaning pointer to array with 8 char elements), 527555714Skris which allows the compiler to do more typechecking; it was like 527655714Skris that back in SSLeay, but with lots of ugly casts. 527755714Skris 527855714Skris Introduce new type const_des_cblock. 527955714Skris [Bodo Moeller] 528055714Skris 528155714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 528255714Skris problems: find RecipientInfo structure that matches recipient certificate 528355714Skris and initialise the ASN1 structures properly based on passed cipher. 528455714Skris [Steve Henson] 528555714Skris 528655714Skris *) Belatedly make the BN tests actually check the results. 528755714Skris [Ben Laurie] 528855714Skris 528955714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 529055714Skris to and from BNs: it was completely broken. New compilation option 529155714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 529255714Skris key elements as negative integers. 529355714Skris [Steve Henson] 529455714Skris 529555714Skris *) Reorganize and speed up MD5. 529655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 529755714Skris 529855714Skris *) VMS support. 529955714Skris [Richard Levitte <richard@levitte.org>] 530055714Skris 530155714Skris *) New option -out to asn1parse to allow the parsed structure to be 530255714Skris output to a file. This is most useful when combined with the -strparse 530355714Skris option to examine the output of things like OCTET STRINGS. 530455714Skris [Steve Henson] 530555714Skris 530655714Skris *) Make SSL library a little more fool-proof by not requiring any longer 530755714Skris that SSL_set_{accept,connect}_state be called before 530855714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 530955714Skris in many applications because usually everything *appeared* to work as 531055714Skris intended anyway -- now it really works as intended). 531155714Skris [Bodo Moeller] 531255714Skris 531355714Skris *) Move openssl.cnf out of lib/. 531455714Skris [Ulf M�ller] 531555714Skris 531655714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 531755714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 531855714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 531955714Skris [Ralf S. Engelschall] 532055714Skris 532155714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 532255714Skris handle PKCS#7 enveloped data properly. 532355714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 532455714Skris 532555714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 532655714Skris copying pointers. The cert_st handling is changed by this in 532755714Skris various ways (and thus what used to be known as ctx->default_cert 532855714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 532955714Skris any longer when s->cert does not give us what we need). 533055714Skris ssl_cert_instantiate becomes obsolete by this change. 533155714Skris As soon as we've got the new code right (possibly it already is?), 533255714Skris we have solved a couple of bugs of the earlier code where s->cert 533355714Skris was used as if it could not have been shared with other SSL structures. 533455714Skris 533555714Skris Note that using the SSL API in certain dirty ways now will result 533655714Skris in different behaviour than observed with earlier library versions: 533755714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 533855714Skris does not influence s as it used to. 533955714Skris 534055714Skris In order to clean up things more thoroughly, inside SSL_SESSION 534155714Skris we don't use CERT any longer, but a new structure SESS_CERT 534255714Skris that holds per-session data (if available); currently, this is 534355714Skris the peer's certificate chain and, for clients, the server's certificate 534455714Skris and temporary key. CERT holds only those values that can have 534555714Skris meaningful defaults in an SSL_CTX. 534655714Skris [Bodo Moeller] 534755714Skris 534855714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 534955714Skris from the internal representation. Various PKCS#7 fixes: remove some 535055714Skris evil casts and set the enc_dig_alg field properly based on the signing 535155714Skris key type. 535255714Skris [Steve Henson] 535355714Skris 535455714Skris *) Allow PKCS#12 password to be set from the command line or the 535555714Skris environment. Let 'ca' get its config file name from the environment 535655714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 535755714Skris and 'x509'). 535855714Skris [Steve Henson] 535955714Skris 536055714Skris *) Allow certificate policies extension to use an IA5STRING for the 536155714Skris organization field. This is contrary to the PKIX definition but 536255714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 536355714Skris extension option. 536455714Skris [Steve Henson] 536555714Skris 536655714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 536755714Skris without disallowing inline assembler and the like for non-pedantic builds. 536855714Skris [Ben Laurie] 536955714Skris 537055714Skris *) Support Borland C++ builder. 537155714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 537255714Skris 537355714Skris *) Support Mingw32. 537455714Skris [Ulf M�ller] 537555714Skris 537655714Skris *) SHA-1 cleanups and performance enhancements. 537755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 537855714Skris 537955714Skris *) Sparc v8plus assembler for the bignum library. 538055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 538155714Skris 538255714Skris *) Accept any -xxx and +xxx compiler options in Configure. 538355714Skris [Ulf M�ller] 538455714Skris 538555714Skris *) Update HPUX configuration. 538655714Skris [Anonymous] 538755714Skris 538855714Skris *) Add missing sk_<type>_unshift() function to safestack.h 538955714Skris [Ralf S. Engelschall] 539055714Skris 539155714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 539255714Skris "extra_cert"s in addition to the certificate. (This makes sense 539355714Skris only for "PEM" format files, as chains as a whole are not 539455714Skris DER-encoded.) 539555714Skris [Bodo Moeller] 539655714Skris 539755714Skris *) Support verify_depth from the SSL API. 539855714Skris x509_vfy.c had what can be considered an off-by-one-error: 539955714Skris Its depth (which was not part of the external interface) 540055714Skris was actually counting the number of certificates in a chain; 540155714Skris now it really counts the depth. 540255714Skris [Bodo Moeller] 540355714Skris 540455714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 540555714Skris instead of X509err, which often resulted in confusing error 540655714Skris messages since the error codes are not globally unique 540755714Skris (e.g. an alleged error in ssl3_accept when a certificate 540855714Skris didn't match the private key). 540955714Skris 541055714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 541155714Skris value (so that you don't need SSL_set_session_id_context for each 541255714Skris connection using the SSL_CTX). 541355714Skris [Bodo Moeller] 541455714Skris 541555714Skris *) OAEP decoding bug fix. 541655714Skris [Ulf M�ller] 541755714Skris 541855714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 541955714Skris David Harris. 542055714Skris [Bodo Moeller] 542155714Skris 542255714Skris *) New Configure options "threads" and "no-threads". For systems 542355714Skris where the proper compiler options are known (currently Solaris 542455714Skris and Linux), "threads" is the default. 542555714Skris [Bodo Moeller] 542655714Skris 542755714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 542855714Skris [Bodo Moeller] 542955714Skris 543055714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 543155714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 543255714Skris such as /usr/local/bin. 543355714Skris [Bodo Moeller] 543455714Skris 543555714Skris *) "make linux-shared" to build shared libraries. 543655714Skris [Niels Poppe <niels@netbox.org>] 543755714Skris 543855714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 543955714Skris [Ulf M�ller] 544055714Skris 544155714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 544255714Skris extension adding in x509 utility. 544355714Skris [Steve Henson] 544455714Skris 544555714Skris *) Remove NOPROTO sections and error code comments. 544655714Skris [Ulf M�ller] 544755714Skris 544855714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 544955714Skris prototypes. 545055714Skris [Steve Henson] 545155714Skris 545255714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 545355714Skris [Ulf M�ller] 545455714Skris 545555714Skris *) Complete rewrite of the error code script(s). It is all now handled 545655714Skris by one script at the top level which handles error code gathering, 545755714Skris header rewriting and C source file generation. It should be much better 545855714Skris than the old method: it now uses a modified version of Ulf's parser to 545955714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 546055714Skris aren't needed for error creation any more) and do a better job of 546155714Skris translating function codes into names. The old 'ASN1 error code imbedded 546255714Skris in a comment' is no longer necessary and it doesn't use .err files which 546355714Skris have now been deleted. Also the error code call doesn't have to appear all 546455714Skris on one line (which resulted in some large lines...). 546555714Skris [Steve Henson] 546655714Skris 546755714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 546855714Skris [Bodo Moeller] 546955714Skris 547055714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 547155714Skris 0 (which usually indicates a closed connection), but continue reading. 547255714Skris [Bodo Moeller] 547355714Skris 547455714Skris *) Fix some race conditions. 547555714Skris [Bodo Moeller] 547655714Skris 547755714Skris *) Add support for CRL distribution points extension. Add Certificate 547855714Skris Policies and CRL distribution points documentation. 547955714Skris [Steve Henson] 548055714Skris 548155714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 548255714Skris [Ulf M�ller] 548355714Skris 548455714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 548555714Skris 8 of keying material. Merlin has also confirmed interop with this fix 548655714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 548755714Skris [Merlin Hughes <merlin@baltimore.ie>] 548855714Skris 548955714Skris *) Fix lots of warnings. 549055714Skris [Richard Levitte <levitte@stacken.kth.se>] 549155714Skris 549255714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 549355714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 549455714Skris [Richard Levitte <levitte@stacken.kth.se>] 549555714Skris 549655714Skris *) Fix problems with sizeof(long) == 8. 549755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 549855714Skris 549955714Skris *) Change functions to ANSI C. 550055714Skris [Ulf M�ller] 550155714Skris 550255714Skris *) Fix typos in error codes. 550355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 550455714Skris 550555714Skris *) Remove defunct assembler files from Configure. 550655714Skris [Ulf M�ller] 550755714Skris 550855714Skris *) SPARC v8 assembler BIGNUM implementation. 550955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 551055714Skris 551155714Skris *) Support for Certificate Policies extension: both print and set. 551255714Skris Various additions to support the r2i method this uses. 551355714Skris [Steve Henson] 551455714Skris 551555714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 551655714Skris return a const string when you are expecting an allocated buffer. 551755714Skris [Ben Laurie] 551855714Skris 551955714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 552055714Skris types DirectoryString and DisplayText. 552155714Skris [Steve Henson] 552255714Skris 552355714Skris *) Add code to allow r2i extensions to access the configuration database, 552455714Skris add an LHASH database driver and add several ctx helper functions. 552555714Skris [Steve Henson] 552655714Skris 552755714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 552855714Skris fail when they extended the size of a BIGNUM. 552955714Skris [Steve Henson] 553055714Skris 553155714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 553255714Skris support typesafe stack. 553355714Skris [Steve Henson] 553455714Skris 553555714Skris *) Fix typo in SSL_[gs]et_options(). 553655714Skris [Nils Frostberg <nils@medcom.se>] 553755714Skris 553855714Skris *) Delete various functions and files that belonged to the (now obsolete) 553955714Skris old X509V3 handling code. 554055714Skris [Steve Henson] 554155714Skris 554255714Skris *) New Configure option "rsaref". 554355714Skris [Ulf M�ller] 554455714Skris 554555714Skris *) Don't auto-generate pem.h. 554655714Skris [Bodo Moeller] 554755714Skris 554855714Skris *) Introduce type-safe ASN.1 SETs. 554955714Skris [Ben Laurie] 555055714Skris 555155714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 555255714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 555355714Skris 555455714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 555555714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 555655714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 555755714Skris few STACKed things have been converted already. Feel free to convert more. 555855714Skris In the fullness of time, I'll do away with the STACK type altogether. 555955714Skris [Ben Laurie] 556055714Skris 556155714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 556255714Skris specified in <certfile> by updating the entry in the index.txt file. 556355714Skris This way one no longer has to edit the index.txt file manually for 556455714Skris revoking a certificate. The -revoke option does the gory details now. 556555714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 556655714Skris 556755714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 556855714Skris `-text' option at all and this way the `-noout -text' combination was 556955714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 557055714Skris [Ralf S. Engelschall] 557155714Skris 557255714Skris *) Make sure a corresponding plain text error message exists for the 557355714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 557455714Skris verify callback function determined that a certificate was revoked. 557555714Skris [Ralf S. Engelschall] 557655714Skris 557755714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 557855714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 557955714Skris all available cipers including rc5, which was forgotten until now. 558055714Skris In order to let the testing shell script know which algorithms 558155714Skris are available, a new (up to now undocumented) command 558255714Skris "openssl list-cipher-commands" is used. 558355714Skris [Bodo Moeller] 558455714Skris 558555714Skris *) Bugfix: s_client occasionally would sleep in select() when 558655714Skris it should have checked SSL_pending() first. 558755714Skris [Bodo Moeller] 558855714Skris 558955714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 559055714Skris the raw DSA values prior to ASN.1 encoding. 559155714Skris [Ulf M�ller] 559255714Skris 559355714Skris *) Tweaks to Configure 559455714Skris [Niels Poppe <niels@netbox.org>] 559555714Skris 559655714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 559755714Skris yet... 559855714Skris [Steve Henson] 559955714Skris 560055714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 560155714Skris [Ulf M�ller] 560255714Skris 560355714Skris *) New config option to avoid instructions that are illegal on the 80386. 560455714Skris The default code is faster, but requires at least a 486. 560555714Skris [Ulf M�ller] 560655714Skris 560755714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 560855714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 560955714Skris same as SSL2_VERSION anyway. 561055714Skris [Bodo Moeller] 561155714Skris 561255714Skris *) New "-showcerts" option for s_client. 561355714Skris [Bodo Moeller] 561455714Skris 561555714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 561655714Skris application. Various cleanups and fixes. 561755714Skris [Steve Henson] 561855714Skris 561955714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 562055714Skris modify error routines to work internally. Add error codes and PBE init 562155714Skris to library startup routines. 562255714Skris [Steve Henson] 562355714Skris 562455714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 562555714Skris packing functions to asn1 and evp. Changed function names and error 562655714Skris codes along the way. 562755714Skris [Steve Henson] 562855714Skris 562955714Skris *) PKCS12 integration: and so it begins... First of several patches to 563055714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 563155714Skris objects to objects.h 563255714Skris [Steve Henson] 563355714Skris 563455714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 563555714Skris and display support for Thawte strong extranet extension. 563655714Skris [Steve Henson] 563755714Skris 563855714Skris *) Add LinuxPPC support. 563955714Skris [Jeff Dubrule <igor@pobox.org>] 564055714Skris 564155714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 564255714Skris bn_div_words in alpha.s. 564355714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 564455714Skris 564555714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 564655714Skris OAEP isn't supported when OpenSSL is built with RSAref. 564755714Skris [Ulf Moeller <ulf@fitug.de>] 564855714Skris 564955714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 565055714Skris so they no longer are missing under -DNOPROTO. 565155714Skris [Soren S. Jorvang <soren@t.dk>] 565255714Skris 565355714Skris 565455714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 565555714Skris 565655714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 565755714Skris doesn't work when the session is reused. Coming soon! 565855714Skris [Ben Laurie] 565955714Skris 566055714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 566155714Skris context thus bypassing client cert protection! All software that uses 566255714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 566355714Skris allow session reuse! A fuller solution is in the works. 566455714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 566555714Skris 566655714Skris *) Some more source tree cleanups (removed obsolete files 566755714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 566855714Skris permission on "config" script to be executable) and a fix for the INSTALL 566955714Skris document. 567055714Skris [Ulf Moeller <ulf@fitug.de>] 567155714Skris 567255714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 567355714Skris Malloc, Free. 567455714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 567555714Skris 567655714Skris *) Make rsa_oaep_test return non-zero on error. 567755714Skris [Ulf Moeller <ulf@fitug.de>] 567855714Skris 567955714Skris *) Add support for native Solaris shared libraries. Configure 568055714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 568155714Skris if someone would make that last step automatic. 568255714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 568355714Skris 568455714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 568555714Skris [Ben Laurie] 568655714Skris 568755714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 568855714Skris except NULL ciphers". This means the default cipher list will no longer 568955714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 569055714Skris the string "DEFAULT:eNULL". 569155714Skris [Steve Henson] 569255714Skris 569355714Skris *) Fix to RSA private encryption routines: if p < q then it would 569455714Skris occasionally produce an invalid result. This will only happen with 569555714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 569655714Skris [Steve Henson] 569755714Skris 569855714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 569955714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 570055714Skris because this way one can also use an interpreter named `perl5' (which is 570155714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 570255714Skris installed as `perl'). 570355714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 570455714Skris 570555714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 570655714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 570755714Skris 570855714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 570955714Skris advapi32.lib to Win32 build and change the pem test comparision 571055714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 571155714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 571255714Skris and crypto/des/ede_cbcm_enc.c. 571355714Skris [Steve Henson] 571455714Skris 571555714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 571655714Skris [Ben Laurie] 571755714Skris 571855714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 571955714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 572055714Skris is horrible: I feel ill.... 572155714Skris [Steve Henson] 572255714Skris 572355714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 572455714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 572555714Skris sections: 10 functions were absent from non ANSI section and not exported 572655714Skris from Windows DLLs. Fixed up libeay.num for new functions. 572755714Skris [Steve Henson] 572855714Skris 572955714Skris *) Make `openssl version' output lines consistent. 573055714Skris [Ralf S. Engelschall] 573155714Skris 573255714Skris *) Fix Win32 symbol export lists for BIO functions: Added 573355714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 573455714Skris to ms/libeay{16,32}.def. 573555714Skris [Ralf S. Engelschall] 573655714Skris 573755714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 573855714Skris fine under Unix and passes some trivial tests I've now added. But the 573955714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 574055714Skris added to make sure no one expects that this stuff really works in the 574155714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 574255714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 574355714Skris openssl_bio.xs. 574455714Skris [Ralf S. Engelschall] 574555714Skris 574655714Skris *) Fix the generation of two part addresses in perl. 574755714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 574855714Skris 574955714Skris *) Add config entry for Linux on MIPS. 575055714Skris [John Tobey <jtobey@channel1.com>] 575155714Skris 575255714Skris *) Make links whenever Configure is run, unless we are on Windoze. 575355714Skris [Ben Laurie] 575455714Skris 575555714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 575655714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 575755714Skris in CRLs. 575855714Skris [Steve Henson] 575955714Skris 576055714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 576155714Skris other platforms details on the command line without having to patch the 576255714Skris Configure script everytime: One now can use ``perl Configure 576355714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 576455714Skris to them (seperated by colons). This is treated as there would be a static 576555714Skris pre-configured entry in Configure's %table under key <id> with value 576655714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 576755714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 576855714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 576955714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 577055714Skris [Ralf S. Engelschall] 577155714Skris 577255714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 577355714Skris [Ben Laurie] 577455714Skris 577555714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 577655714Skris on the `perl Configure ...' command line. This way one can compile 577755714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 577855714Skris for linking it into DSOs. 577955714Skris [Ralf S. Engelschall] 578055714Skris 578155714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 578255714Skris Fixed. 578355714Skris [Ben Laurie] 578455714Skris 578555714Skris *) Cleaned up the LICENSE document: The official contact for any license 578655714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 578755714Skris And add a paragraph about the dual-license situation to make sure people 578855714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 578955714Skris to the OpenSSL toolkit. 579055714Skris [Ralf S. Engelschall] 579155714Skris 579255714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 579355714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 579455714Skris Additonally cleaned up the `make links' target: Remove unnecessary 579555714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 579655714Skris to speed processing and no longer clutter the display with confusing 579755714Skris stuff. Instead only the actually done links are displayed. 579855714Skris [Ralf S. Engelschall] 579955714Skris 580055714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 580155714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 580255714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 580355714Skris encryption. 580455714Skris [Ben Laurie] 580555714Skris 580655714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 580755714Skris signed attributes when verifying signatures (this would break them), 580855714Skris the detached data encoding was wrong and public keys obtained using 580955714Skris X509_get_pubkey() weren't freed. 581055714Skris [Steve Henson] 581155714Skris 581255714Skris *) Add text documentation for the BUFFER functions. Also added a work around 581355714Skris to a Win95 console bug. This was triggered by the password read stuff: the 581455714Skris last character typed gets carried over to the next fread(). If you were 581555714Skris generating a new cert request using 'req' for example then the last 581655714Skris character of the passphrase would be CR which would then enter the first 581755714Skris field as blank. 581855714Skris [Steve Henson] 581955714Skris 582055714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 582155714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 582255714Skris button and can be used by applications based on OpenSSL to show the 582355714Skris relationship to the OpenSSL project. 582455714Skris [Ralf S. Engelschall] 582555714Skris 582655714Skris *) Remove confusing variables in function signatures in files 582755714Skris ssl/ssl_lib.c and ssl/ssl.h. 582855714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 582955714Skris 583055714Skris *) Don't install bss_file.c under PREFIX/include/ 583155714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 583255714Skris 583355714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 583455714Skris functions that return function pointers and has support for NT specific 583555714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 583655714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 583755714Skris unsigned to signed types: this was killing the Win32 compile. 583855714Skris [Steve Henson] 583955714Skris 584055714Skris *) Add new certificate file to stack functions, 584155714Skris SSL_add_dir_cert_subjects_to_stack() and 584255714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 584355714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 584455714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 584555714Skris This means that Apache-SSL and similar packages don't have to mess around 584655714Skris to add as many CAs as they want to the preferred list. 584755714Skris [Ben Laurie] 584855714Skris 584955714Skris *) Experiment with doxygen documentation. Currently only partially applied to 585055714Skris ssl/ssl_lib.c. 585155714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 585255714Skris openssl.doxy as the configuration file. 585355714Skris [Ben Laurie] 585455714Skris 585555714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 585655714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 585755714Skris 585855714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 585955714Skris compiled in by default: it has problems with large keys. 586055714Skris [Steve Henson] 586155714Skris 586255714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 586355714Skris DH private keys and/or callback functions which directly correspond to 586455714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 586555714Skris is needed for applications which have to configure certificates on a 586655714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 586755714Skris (e.g. s_server). 586855714Skris For the RSA certificate situation is makes no difference, but 586955714Skris for the DSA certificate situation this fixes the "no shared cipher" 587055714Skris problem where the OpenSSL cipher selection procedure failed because the 587155714Skris temporary keys were not overtaken from the context and the API provided 587255714Skris no way to reconfigure them. 587355714Skris The new functions now let applications reconfigure the stuff and they 587455714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 587555714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 587655714Skris non-public-API function ssl_cert_instantiate() is used as a helper 587755714Skris function and also to reduce code redundancy inside ssl_rsa.c. 587855714Skris [Ralf S. Engelschall] 587955714Skris 588055714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 588155714Skris area because they are useful for the DSA situation and should be 588255714Skris recognized by the users. 588355714Skris [Ralf S. Engelschall] 588455714Skris 588555714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 588655714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 588755714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 588855714Skris already masked variable. 588955714Skris [Richard Levitte <levitte@stacken.kth.se>] 589055714Skris 589155714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 589255714Skris [Richard Levitte <levitte@stacken.kth.se>] 589355714Skris 589455714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 589555714Skris from `int' to `unsigned int' because it's a length and initialized by 589655714Skris EVP_DigestFinal() which expects an `unsigned int *'. 589755714Skris [Richard Levitte <levitte@stacken.kth.se>] 589855714Skris 589955714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 590055714Skris script. Instead use the usual Shell->Perl transition trick. 590155714Skris [Ralf S. Engelschall] 590255714Skris 590355714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 590455714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 590555714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 590655714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 590755714Skris currently the public key is printed (a decision which was already done by 590855714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 590955714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 591055714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 591155714Skris now, too. 591255714Skris [Ralf S. Engelschall] 591355714Skris 591455714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 591555714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 591655714Skris [Arne Ansper <arne@ats.cyber.ee>] 591755714Skris 591855714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 591955714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 592055714Skris config file. 592155714Skris [Steve Henson] 592255714Skris 592355714Skris *) Add cool BIO that does syslog (or event log on NT). 592455714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 592555714Skris 592655714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 592755714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 592855714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 592955714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 593055714Skris [Ben Laurie] 593155714Skris 593255714Skris *) Add preliminary config info for new extension code. 593355714Skris [Steve Henson] 593455714Skris 593555714Skris *) Make RSA_NO_PADDING really use no padding. 593655714Skris [Ulf Moeller <ulf@fitug.de>] 593755714Skris 593855714Skris *) Generate errors when private/public key check is done. 593955714Skris [Ben Laurie] 594055714Skris 594155714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 594255714Skris for some CRL extensions and new objects added. 594355714Skris [Steve Henson] 594455714Skris 594555714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 594655714Skris key usage extension and fuller support for authority key id. 594755714Skris [Steve Henson] 594855714Skris 594955714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 595055714Skris padding method for RSA, which is recommended for new applications in PKCS 595155714Skris #1 v2.0 (RFC 2437, October 1998). 595255714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 595355714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 595455714Skris against Bleichbacher's attack on RSA. 595555714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 595655714Skris Ben Laurie] 595755714Skris 595855714Skris *) Updates to the new SSL compression code 595955714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 596055714Skris 596155714Skris *) Fix so that the version number in the master secret, when passed 596255714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 596355714Skris (because the server will not accept higher), that the version number 596455714Skris is 0x03,0x01, not 0x03,0x00 596555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 596655714Skris 596755714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 596855714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 596955714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 597055714Skris [Steve Henson] 597155714Skris 597255714Skris *) Support for RAW extensions where an arbitrary extension can be 597355714Skris created by including its DER encoding. See apps/openssl.cnf for 597455714Skris an example. 597555714Skris [Steve Henson] 597655714Skris 597755714Skris *) Make sure latest Perl versions don't interpret some generated C array 597855714Skris code as Perl array code in the crypto/err/err_genc.pl script. 597955714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 598055714Skris 598155714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 598255714Skris not many people have the assembler. Various Win32 compilation fixes and 598355714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 598455714Skris build instructions. 598555714Skris [Steve Henson] 598655714Skris 598755714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 598855714Skris file under Win32 and also build pem.h from pem.org. New script 598955714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 599055714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 599155714Skris [Steve Henson] 599255714Skris 599355714Skris *) Major rework of DES function declarations, in the pursuit of correctness 599455714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 599555714Skris too. You may find this causes warnings in your code. Zapping your evil 599655714Skris casts will probably fix them. Mostly. 599755714Skris [Ben Laurie] 599855714Skris 599955714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 600055714Skris obj_dat.pl. It considered a zero in an object definition to mean 600155714Skris "end of object": none of the objects in objects.h have any zeros 600255714Skris so it wasn't spotted. 600355714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 600455714Skris 600555714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 600655714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 600755714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 600855714Skris vectors if you have them. 600955714Skris [Ben Laurie] 601055714Skris 601155714Skris *) Correct calculation of key length for export ciphers (too much space was 601255714Skris allocated for null ciphers). This has not been tested! 601355714Skris [Ben Laurie] 601455714Skris 601555714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 601655714Skris message is now correct (it understands "crypto" and "ssl" on its 601755714Skris command line). There is also now an "update" option. This will update 601855714Skris the util/ssleay.num and util/libeay.num files with any new functions. 601955714Skris If you do a: 602055714Skris perl util/mkdef.pl crypto ssl update 602155714Skris it will update them. 602255714Skris [Steve Henson] 602355714Skris 602455714Skris *) Overhauled the Perl interface (perl/*): 602555714Skris - ported BN stuff to OpenSSL's different BN library 602655714Skris - made the perl/ source tree CVS-aware 602755714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 602855714Skris their history because I've copied them in the repository) 602955714Skris - removed obsolete files (the test scripts will be replaced 603055714Skris by better Test::Harness variants in the future) 603155714Skris [Ralf S. Engelschall] 603255714Skris 603355714Skris *) First cut for a very conservative source tree cleanup: 603455714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 603555714Skris where we collect the old documents and readme texts. 603655714Skris 2. remove the first part of files where I'm already sure that we no 603755714Skris longer need them because of three reasons: either they are just temporary 603855714Skris files which were left by Eric or they are preserved original files where 603955714Skris I've verified that the diff is also available in the CVS via "cvs diff 604055714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 604155714Skris the crypto/md/ stuff). 604255714Skris [Ralf S. Engelschall] 604355714Skris 604455714Skris *) More extension code. Incomplete support for subject and issuer alt 604555714Skris name, issuer and authority key id. Change the i2v function parameters 604655714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 604755714Skris what that's for :-) Fix to ASN1 macro which messed up 604855714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 604955714Skris [Steve Henson] 605055714Skris 605155714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 605255714Skris INTEGER code. 605355714Skris [Steve Henson] 605455714Skris 605555714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 605655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 605755714Skris 605855714Skris *) Make sure `make rehash' target really finds the `openssl' program. 605955714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 606055714Skris 606155714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 606255714Skris like to hear about it if this slows down other processors. 606355714Skris [Ben Laurie] 606455714Skris 606555714Skris *) Add CygWin32 platform information to Configure script. 606655714Skris [Alan Batie <batie@aahz.jf.intel.com>] 606755714Skris 606855714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 606955714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 607055714Skris 607155714Skris *) New program nseq to manipulate netscape certificate sequences 607255714Skris [Steve Henson] 607355714Skris 607455714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 607555714Skris few typos. 607655714Skris [Steve Henson] 607755714Skris 607855714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 607955714Skris but the BN code had some problems that would cause failures when 608055714Skris doing certificate verification and some other functions. 608155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 608255714Skris 608355714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 608455714Skris [Steve Henson] 608555714Skris 608655714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 608755714Skris [Steve Henson] 608855714Skris 608955714Skris *) Add several PKIX and private extended key usage OIDs. 609055714Skris [Steve Henson] 609155714Skris 609255714Skris *) Modify the 'ca' program to handle the new extension code. Modify 609355714Skris openssl.cnf for new extension format, add comments. 609455714Skris [Steve Henson] 609555714Skris 609655714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 609755714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 609855714Skris CA extensions. 609955714Skris [Steve Henson] 610055714Skris 610155714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 610255714Skris error code, add initial support to X509_print() and x509 application. 610355714Skris [Steve Henson] 610455714Skris 610555714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 610655714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 610755714Skris stuff is currently isolated and isn't even compiled yet. 610855714Skris [Steve Henson] 610955714Skris 611055714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 611155714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 611255714Skris Removed the versions check from X509 routines when loading extensions: 611355714Skris this allows certain broken certificates that don't set the version 611455714Skris properly to be processed. 611555714Skris [Steve Henson] 611655714Skris 611755714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 611855714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 611955714Skris can still be regenerated with "make depend". 612055714Skris [Ben Laurie] 612155714Skris 612255714Skris *) Spelling mistake in C version of CAST-128. 612355714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 612455714Skris 612555714Skris *) Changes to the error generation code. The perl script err-code.pl 612655714Skris now reads in the old error codes and retains the old numbers, only 612755714Skris adding new ones if necessary. It also only changes the .err files if new 612855714Skris codes are added. The makefiles have been modified to only insert errors 612955714Skris when needed (to avoid needlessly modifying header files). This is done 613055714Skris by only inserting errors if the .err file is newer than the auto generated 613155714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 613255714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 613355714Skris or delete all the .err files. 613455714Skris [Steve Henson] 613555714Skris 613655714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 613755714Skris been fixed, but is untested. The assembler versions are also fixed, but 613855714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 613955714Skris to regenerate it if needed. 614055714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 614155714Skris Hagino <itojun@kame.net>] 614255714Skris 614355714Skris *) File was opened incorrectly in randfile.c. 614455714Skris [Ulf M�ller <ulf@fitug.de>] 614555714Skris 614655714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 614755714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 614855714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 614955714Skris al: it's just almost always a UTCTime. Note this patch adds new error 615055714Skris codes so do a "make errors" if there are problems. 615155714Skris [Steve Henson] 615255714Skris 615355714Skris *) Correct Linux 1 recognition in config. 615455714Skris [Ulf M�ller <ulf@fitug.de>] 615555714Skris 615655714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 615755714Skris [Anonymous <nobody@replay.com>] 615855714Skris 615955714Skris *) Generate an error if given an empty string as a cert directory. Also 616055714Skris generate an error if handed NULL (previously returned 0 to indicate an 616155714Skris error, but didn't set one). 616255714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 616355714Skris 616455714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 616555714Skris [Ben Laurie] 616655714Skris 616755714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 616855714Skris parameters. This was causing a warning which killed off the Win32 compile. 616955714Skris [Steve Henson] 617055714Skris 617155714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 617255714Skris [Neil Costigan <neil.costigan@celocom.com>] 617355714Skris 617455714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 617555714Skris based on a text string, looking up short and long names and finally 617655714Skris "dot" format. The "dot" format stuff didn't work. Added new function 617755714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 617855714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 617955714Skris OID is not part of the table. 618055714Skris [Steve Henson] 618155714Skris 618255714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 618355714Skris X509_LOOKUP_by_alias(). 618455714Skris [Ben Laurie] 618555714Skris 618655714Skris *) Sort openssl functions by name. 618755714Skris [Ben Laurie] 618855714Skris 618955714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 619055714Skris encryption from sample DSA keys (in case anyone is interested the password 619155714Skris was "1234"). 619255714Skris [Steve Henson] 619355714Skris 619455714Skris *) Make _all_ *_free functions accept a NULL pointer. 619555714Skris [Frans Heymans <fheymans@isaserver.be>] 619655714Skris 619755714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 619855714Skris NULL pointers. 619955714Skris [Anonymous <nobody@replay.com>] 620055714Skris 620155714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 620255714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 620355714Skris 620455714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 620555714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 620655714Skris 620755714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 620855714Skris [Anonymous <nobody@replay.com>] 620955714Skris 621055714Skris *) Add prototype for temp key callback functions 621155714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 621255714Skris [Ben Laurie] 621355714Skris 621455714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 621555714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 621655714Skris [Steve Henson] 621755714Skris 621855714Skris *) X509_name_add_entry() freed the wrong thing after an error. 621955714Skris [Arne Ansper <arne@ats.cyber.ee>] 622055714Skris 622155714Skris *) rsa_eay.c would attempt to free a NULL context. 622255714Skris [Arne Ansper <arne@ats.cyber.ee>] 622355714Skris 622455714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 622555714Skris [Arne Ansper <arne@ats.cyber.ee>] 622655714Skris 622755714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 622855714Skris [Arne Ansper <arne@ats.cyber.ee>] 622955714Skris 623055714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 623155714Skris in X509_STORE_new(), but document the fact that this variable is still 623255714Skris unused in the certificate verification process. 623355714Skris [Ralf S. Engelschall] 623455714Skris 623555714Skris *) Fix the various library and apps files to free up pkeys obtained from 623655714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 623755714Skris [Steve Henson] 623855714Skris 623955714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 624055714Skris demos/maurice/example2.c work, amongst others, probably. 624155714Skris [Steve Henson and Ben Laurie] 624255714Skris 624355714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 624455714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 624555714Skris are no longer created. This way we have a single and consistent command 624655714Skris line interface `openssl <command>', similar to `cvs <command>'. 624755714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 624855714Skris 624955714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 625055714Skris BIT STRING wrapper always have zero unused bits. 625155714Skris [Steve Henson] 625255714Skris 625355714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 625455714Skris [Steve Henson] 625555714Skris 625655714Skris *) Make the top-level INSTALL documentation easier to understand. 625755714Skris [Paul Sutton] 625855714Skris 625955714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 626055714Skris make (including if user presses ^C) [Paul Sutton] 626155714Skris 626255714Skris *) Make Montgomery context stuff explicit in RSA data structure. 626355714Skris [Ben Laurie] 626455714Skris 626555714Skris *) Fix build order of pem and err to allow for generated pem.h. 626655714Skris [Ben Laurie] 626755714Skris 626855714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 626955714Skris [Ben Laurie] 627055714Skris 627155714Skris *) Enhanced the err-ins.pl script so it makes the error library number 627255714Skris global and can add a library name. This is needed for external ASN1 and 627355714Skris other error libraries. 627455714Skris [Steve Henson] 627555714Skris 627655714Skris *) Fixed sk_insert which never worked properly. 627755714Skris [Steve Henson] 627855714Skris 627955714Skris *) Fix ASN1 macros so they can handle indefinite length construted 628055714Skris EXPLICIT tags. Some non standard certificates use these: they can now 628155714Skris be read in. 628255714Skris [Steve Henson] 628355714Skris 628455714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 628555714Skris into a single doc/ssleay.txt bundle. This way the information is still 628655714Skris preserved but no longer messes up this directory. Now it's new room for 628755714Skris the new set of documenation files. 628855714Skris [Ralf S. Engelschall] 628955714Skris 629055714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 629155714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 629255714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 629355714Skris number of arguments. 629455714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 629555714Skris 629655714Skris *) Fix test data to work with the above. 629755714Skris [Ben Laurie] 629855714Skris 629955714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 630055714Skris was already fixed by Eric for 0.9.1 it seems. 630155714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 630255714Skris 630355714Skris *) Autodetect FreeBSD3. 630455714Skris [Ben Laurie] 630555714Skris 630655714Skris *) Fix various bugs in Configure. This affects the following platforms: 630755714Skris nextstep 630855714Skris ncr-scde 630955714Skris unixware-2.0 631055714Skris unixware-2.0-pentium 631155714Skris sco5-cc. 631255714Skris [Ben Laurie] 631355714Skris 631455714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 631555714Skris before they are needed. 631655714Skris [Ben Laurie] 631755714Skris 631855714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 631955714Skris [Ben Laurie] 632055714Skris 632155714Skris 632255714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 632355714Skris 632455714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 632555714Skris changed SSLeay to OpenSSL in version strings. 632655714Skris [Ralf S. Engelschall] 632755714Skris 632855714Skris *) Some fixups to the top-level documents. 632955714Skris [Paul Sutton] 633055714Skris 633155714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 633255714Skris because the symlink to include/ was missing. 633355714Skris [Ralf S. Engelschall] 633455714Skris 633555714Skris *) Incorporated the popular no-RSA/DSA-only patches 633655714Skris which allow to compile a RSA-free SSLeay. 633755714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 633855714Skris 633955714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 634055714Skris when "ssleay" is still not found. 634155714Skris [Ralf S. Engelschall] 634255714Skris 634355714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 634455714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 634555714Skris 634655714Skris *) Updated the README file. 634755714Skris [Ralf S. Engelschall] 634855714Skris 634955714Skris *) Added various .cvsignore files in the CVS repository subdirs 635055714Skris to make a "cvs update" really silent. 635155714Skris [Ralf S. Engelschall] 635255714Skris 635355714Skris *) Recompiled the error-definition header files and added 635455714Skris missing symbols to the Win32 linker tables. 635555714Skris [Ralf S. Engelschall] 635655714Skris 635755714Skris *) Cleaned up the top-level documents; 635855714Skris o new files: CHANGES and LICENSE 635955714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 636055714Skris o merged COPYRIGHT into LICENSE 636155714Skris o removed obsolete TODO file 636255714Skris o renamed MICROSOFT to INSTALL.W32 636355714Skris [Ralf S. Engelschall] 636455714Skris 636555714Skris *) Removed dummy files from the 0.9.1b source tree: 636655714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 636755714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 636855714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 636955714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 637055714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 637155714Skris [Ralf S. Engelschall] 637255714Skris 637355714Skris *) Added various platform portability fixes. 637455714Skris [Mark J. Cox] 637555714Skris 637655714Skris *) The Genesis of the OpenSSL rpject: 637755714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 637855714Skris Young and Tim J. Hudson created while they were working for C2Net until 637955714Skris summer 1998. 638055714Skris [The OpenSSL Project] 638155714Skris 638255714Skris 638355714Skris Changes between 0.9.0b and 0.9.1b [not released] 638455714Skris 638555714Skris *) Updated a few CA certificates under certs/ 638655714Skris [Eric A. Young] 638755714Skris 638855714Skris *) Changed some BIGNUM api stuff. 638955714Skris [Eric A. Young] 639055714Skris 639155714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 639255714Skris DGUX x86, Linux Alpha, etc. 639355714Skris [Eric A. Young] 639455714Skris 639555714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 639655714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 639755714Skris available). 639855714Skris [Eric A. Young] 639955714Skris 640055714Skris *) Add -strparse option to asn1pars program which parses nested 640155714Skris binary structures 640255714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 640355714Skris 640455714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 640555714Skris [Eric A. Young] 640655714Skris 640755714Skris *) DSA fix for "ca" program. 640855714Skris [Eric A. Young] 640955714Skris 641055714Skris *) Added "-genkey" option to "dsaparam" program. 641155714Skris [Eric A. Young] 641255714Skris 641355714Skris *) Added RIPE MD160 (rmd160) message digest. 641455714Skris [Eric A. Young] 641555714Skris 641655714Skris *) Added -a (all) option to "ssleay version" command. 641755714Skris [Eric A. Young] 641855714Skris 641955714Skris *) Added PLATFORM define which is the id given to Configure. 642055714Skris [Eric A. Young] 642155714Skris 642255714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 642355714Skris [Eric A. Young] 642455714Skris 642555714Skris *) Extended the ASN.1 parser routines. 642655714Skris [Eric A. Young] 642755714Skris 642855714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 642955714Skris [Eric A. Young] 643055714Skris 643155714Skris *) Added a BN_CTX to the BN library. 643255714Skris [Eric A. Young] 643355714Skris 643455714Skris *) Fixed the weak key values in DES library 643555714Skris [Eric A. Young] 643655714Skris 643755714Skris *) Changed API in EVP library for cipher aliases. 643855714Skris [Eric A. Young] 643955714Skris 644055714Skris *) Added support for RC2/64bit cipher. 644155714Skris [Eric A. Young] 644255714Skris 644355714Skris *) Converted the lhash library to the crypto/mem.c functions. 644455714Skris [Eric A. Young] 644555714Skris 644655714Skris *) Added more recognized ASN.1 object ids. 644755714Skris [Eric A. Young] 644855714Skris 644955714Skris *) Added more RSA padding checks for SSL/TLS. 645055714Skris [Eric A. Young] 645155714Skris 645255714Skris *) Added BIO proxy/filter functionality. 645355714Skris [Eric A. Young] 645455714Skris 645555714Skris *) Added extra_certs to SSL_CTX which can be used 645655714Skris send extra CA certificates to the client in the CA cert chain sending 645755714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 645855714Skris [Eric A. Young] 645955714Skris 646055714Skris *) Now Fortezza is denied in the authentication phase because 646155714Skris this is key exchange mechanism is not supported by SSLeay at all. 646255714Skris [Eric A. Young] 646355714Skris 646455714Skris *) Additional PKCS1 checks. 646555714Skris [Eric A. Young] 646655714Skris 646755714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 646855714Skris [Eric A. Young] 646955714Skris 647055714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 647155714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 647255714Skris [Eric A. Young] 647355714Skris 647455714Skris *) Fixed a few memory leaks. 647555714Skris [Eric A. Young] 647655714Skris 647755714Skris *) Fixed various code and comment typos. 647855714Skris [Eric A. Young] 647955714Skris 648055714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 648155714Skris bytes sent in the client random. 648255714Skris [Edward Bishop <ebishop@spyglass.com>] 648355714Skris 6484