1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 6 7 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 8 server and client random values. Previously 9 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 10 less random data when sizeof(time_t) > 4 (some 64 bit platforms). 11 12 This change has negligible security impact because: 13 14 1. Server and client random values still have 24 bytes of pseudo random 15 data. 16 17 2. Server and client random values are sent in the clear in the initial 18 handshake. 19 20 3. The master secret is derived using the premaster secret (48 bytes in 21 size for static RSA ciphersuites) as well as client server and random 22 values. 23 24 The OpenSSL team would like to thank the UK NISCC for bringing this issue 25 to our attention. 26 27 [Stephen Henson, reported by UK NISCC] 28 29 *) Use Windows randomness collection on Cygwin. 30 [Ulf M�ller] 31 32 *) Fix hang in EGD/PRNGD query when communication socket is closed 33 prematurely by EGD/PRNGD. 34 [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 35 36 *) Prompt for pass phrases when appropriate for PKCS12 input format. 37 [Steve Henson] 38 39 *) Back-port of selected performance improvements from development 40 branch, as well as improved support for PowerPC platforms. 41 [Andy Polyakov] 42 43 *) Add lots of checks for memory allocation failure, error codes to indicate 44 failure and freeing up memory if a failure occurs. 45 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 46 47 *) Add new -passin argument to dgst. 48 [Steve Henson] 49 50 *) Perform some character comparisons of different types in X509_NAME_cmp: 51 this is needed for some certificates that reencode DNs into UTF8Strings 52 (in violation of RFC3280) and can't or wont issue name rollover 53 certificates. 54 [Steve Henson] 55 56 *) Make an explicit check during certificate validation to see that 57 the CA setting in each certificate on the chain is correct. As a 58 side effect always do the following basic checks on extensions, 59 not just when there's an associated purpose to the check: 60 61 - if there is an unhandled critical extension (unless the user 62 has chosen to ignore this fault) 63 - if the path length has been exceeded (if one is set at all) 64 - that certain extensions fit the associated purpose (if one has 65 been given) 66 [Richard Levitte] 67 68 Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 69 70 *) Avoid a race condition when CRLs are checked in a multi threaded 71 environment. This would happen due to the reordering of the revoked 72 entries during signature checking and serial number lookup. Now the 73 encoding is cached and the serial number sort performed under a lock. 74 Add new STACK function sk_is_sorted(). 75 [Steve Henson] 76 77 *) Add Delta CRL to the extension code. 78 [Steve Henson] 79 80 *) Various fixes to s3_pkt.c so alerts are sent properly. 81 [David Holmes <d.holmes@f5.com>] 82 83 *) Reduce the chances of duplicate issuer name and serial numbers (in 84 violation of RFC3280) using the OpenSSL certificate creation utilities. 85 This is done by creating a random 64 bit value for the initial serial 86 number when a serial number file is created or when a self signed 87 certificate is created using 'openssl req -x509'. The initial serial 88 number file is created using 'openssl x509 -next_serial' in CA.pl 89 rather than being initialized to 1. 90 [Steve Henson] 91 92 Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 93 94 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 95 by using the Codenomicon TLS Test Tool (CAN-2004-0079) 96 [Joe Orton, Steve Henson] 97 98 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 99 (CAN-2004-0112) 100 [Joe Orton, Steve Henson] 101 102 *) Make it possible to have multiple active certificates with the same 103 subject in the CA index file. This is done only if the keyword 104 'unique_subject' is set to 'no' in the main CA section (default 105 if 'CA_default') of the configuration file. The value is saved 106 with the database itself in a separate index attribute file, 107 named like the index file with '.attr' appended to the name. 108 [Richard Levitte] 109 110 *) X509 verify fixes. Disable broken certificate workarounds when 111 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 112 keyUsage extension present. Don't accept CRLs with unhandled critical 113 extensions: since verify currently doesn't process CRL extensions this 114 rejects a CRL with *any* critical extensions. Add new verify error codes 115 for these cases. 116 [Steve Henson] 117 118 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 119 A clarification of RFC2560 will require the use of OCTET STRINGs and 120 some implementations cannot handle the current raw format. Since OpenSSL 121 copies and compares OCSP nonces as opaque blobs without any attempt at 122 parsing them this should not create any compatibility issues. 123 [Steve Henson] 124 125 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 126 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 127 this HMAC (and other) operations are several times slower than OpenSSL 128 < 0.9.7. 129 [Steve Henson] 130 131 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 132 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 133 134 *) Use the correct content when signing type "other". 135 [Steve Henson] 136 137 Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 138 139 *) Fix various bugs revealed by running the NISCC test suite: 140 141 Stop out of bounds reads in the ASN1 code when presented with 142 invalid tags (CAN-2003-0543 and CAN-2003-0544). 143 144 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545). 145 146 If verify callback ignores invalid public key errors don't try to check 147 certificate signature with the NULL public key. 148 149 [Steve Henson] 150 151 *) New -ignore_err option in ocsp application to stop the server 152 exiting on the first error in a request. 153 [Steve Henson] 154 155 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 156 if the server requested one: as stated in TLS 1.0 and SSL 3.0 157 specifications. 158 [Steve Henson] 159 160 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 161 extra data after the compression methods not only for TLS 1.0 162 but also for SSL 3.0 (as required by the specification). 163 [Bodo Moeller; problem pointed out by Matthias Loepfe] 164 165 *) Change X509_certificate_type() to mark the key as exported/exportable 166 when it's 512 *bits* long, not 512 bytes. 167 [Richard Levitte] 168 169 *) Change AES_cbc_encrypt() so it outputs exact multiple of 170 blocks during encryption. 171 [Richard Levitte] 172 173 *) Various fixes to base64 BIO and non blocking I/O. On write 174 flushes were not handled properly if the BIO retried. On read 175 data was not being buffered properly and had various logic bugs. 176 This also affects blocking I/O when the data being decoded is a 177 certain size. 178 [Steve Henson] 179 180 *) Various S/MIME bugfixes and compatibility changes: 181 output correct application/pkcs7 MIME type if 182 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 183 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 184 of files as .eml work). Correctly handle very long lines in MIME 185 parser. 186 [Steve Henson] 187 188 Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 189 190 *) Countermeasure against the Klima-Pokorny-Rosa extension of 191 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 192 a protocol version number mismatch like a decryption error 193 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 194 [Bodo Moeller] 195 196 *) Turn on RSA blinding by default in the default implementation 197 to avoid a timing attack. Applications that don't want it can call 198 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 199 They would be ill-advised to do so in most cases. 200 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 201 202 *) Change RSA blinding code so that it works when the PRNG is not 203 seeded (in this case, the secret RSA exponent is abused as 204 an unpredictable seed -- if it is not unpredictable, there 205 is no point in blinding anyway). Make RSA blinding thread-safe 206 by remembering the creator's thread ID in rsa->blinding and 207 having all other threads use local one-time blinding factors 208 (this requires more computation than sharing rsa->blinding, but 209 avoids excessive locking; and if an RSA object is not shared 210 between threads, blinding will still be very fast). 211 [Bodo Moeller] 212 213 *) Fixed a typo bug that would cause ENGINE_set_default() to set an 214 ENGINE as defaults for all supported algorithms irrespective of 215 the 'flags' parameter. 'flags' is now honoured, so applications 216 should make sure they are passing it correctly. 217 [Geoff Thorpe] 218 219 *) Target "mingw" now allows native Windows code to be generated in 220 the Cygwin environment as well as with the MinGW compiler. 221 [Ulf Moeller] 222 223 Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 224 225 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 226 via timing by performing a MAC computation even if incorrrect 227 block cipher padding has been found. This is a countermeasure 228 against active attacks where the attacker has to distinguish 229 between bad padding and a MAC verification error. (CAN-2003-0078) 230 231 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 232 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 233 Martin Vuagnoux (EPFL, Ilion)] 234 235 *) Make the no-err option work as intended. The intention with no-err 236 is not to have the whole error stack handling routines removed from 237 libcrypto, it's only intended to remove all the function name and 238 reason texts, thereby removing some of the footprint that may not 239 be interesting if those errors aren't displayed anyway. 240 241 NOTE: it's still possible for any application or module to have it's 242 own set of error texts inserted. The routines are there, just not 243 used by default when no-err is given. 244 [Richard Levitte] 245 246 *) Add support for FreeBSD on IA64. 247 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 248 249 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 250 Kerberos function mit_des_cbc_cksum(). Before this change, 251 the value returned by DES_cbc_cksum() was like the one from 252 mit_des_cbc_cksum(), except the bytes were swapped. 253 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 254 255 *) Allow an application to disable the automatic SSL chain building. 256 Before this a rather primitive chain build was always performed in 257 ssl3_output_cert_chain(): an application had no way to send the 258 correct chain if the automatic operation produced an incorrect result. 259 260 Now the chain builder is disabled if either: 261 262 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 263 264 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 265 266 The reasoning behind this is that an application would not want the 267 auto chain building to take place if extra chain certificates are 268 present and it might also want a means of sending no additional 269 certificates (for example the chain has two certificates and the 270 root is omitted). 271 [Steve Henson] 272 273 *) Add the possibility to build without the ENGINE framework. 274 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 275 276 *) Under Win32 gmtime() can return NULL: check return value in 277 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 278 [Steve Henson] 279 280 *) DSA routines: under certain error conditions uninitialized BN objects 281 could be freed. Solution: make sure initialization is performed early 282 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 283 Nils Larsch <nla@trustcenter.de> via PR#459) 284 [Lutz Jaenicke] 285 286 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 287 checked on reconnect on the client side, therefore session resumption 288 could still fail with a "ssl session id is different" error. This 289 behaviour is masked when SSL_OP_ALL is used due to 290 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 291 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 292 followup to PR #377. 293 [Lutz Jaenicke] 294 295 *) IA-32 assembler support enhancements: unified ELF targets, support 296 for SCO/Caldera platforms, fix for Cygwin shared build. 297 [Andy Polyakov] 298 299 *) Add support for FreeBSD on sparc64. As a consequence, support for 300 FreeBSD on non-x86 processors is separate from x86 processors on 301 the config script, much like the NetBSD support. 302 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 303 304 Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 305 306 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 307 OpenSSL 0.9.7.] 308 309 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 310 code (06) was taken as the first octet of the session ID and the last 311 octet was ignored consequently. As a result SSLv2 client side session 312 caching could not have worked due to the session ID mismatch between 313 client and server. 314 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 315 PR #377. 316 [Lutz Jaenicke] 317 318 *) Change the declaration of needed Kerberos libraries to use EX_LIBS 319 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 320 removed entirely. 321 [Richard Levitte] 322 323 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 324 seems that in spite of existing for more than a year, many application 325 author have done nothing to provide the necessary callbacks, which 326 means that this particular engine will not work properly anywhere. 327 This is a very unfortunate situation which forces us, in the name 328 of usability, to give the hw_ncipher.c a static lock, which is part 329 of libcrypto. 330 NOTE: This is for the 0.9.7 series ONLY. This hack will never 331 appear in 0.9.8 or later. We EXPECT application authors to have 332 dealt properly with this when 0.9.8 is released (unless we actually 333 make such changes in the libcrypto locking code that changes will 334 have to be made anyway). 335 [Richard Levitte] 336 337 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 338 octets have been read, EOF or an error occurs. Without this change 339 some truncated ASN1 structures will not produce an error. 340 [Steve Henson] 341 342 *) Disable Heimdal support, since it hasn't been fully implemented. 343 Still give the possibility to force the use of Heimdal, but with 344 warnings and a request that patches get sent to openssl-dev. 345 [Richard Levitte] 346 347 *) Add the VC-CE target, introduce the WINCE sysname, and add 348 INSTALL.WCE and appropriate conditionals to make it build. 349 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 350 351 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 352 cygssl-x.y.z.dll, where x, y and z are the major, minor and 353 edit numbers of the version. 354 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 355 356 *) Introduce safe string copy and catenation functions 357 (BUF_strlcpy() and BUF_strlcat()). 358 [Ben Laurie (CHATS) and Richard Levitte] 359 360 *) Avoid using fixed-size buffers for one-line DNs. 361 [Ben Laurie (CHATS)] 362 363 *) Add BUF_MEM_grow_clean() to avoid information leakage when 364 resizing buffers containing secrets, and use where appropriate. 365 [Ben Laurie (CHATS)] 366 367 *) Avoid using fixed size buffers for configuration file location. 368 [Ben Laurie (CHATS)] 369 370 *) Avoid filename truncation for various CA files. 371 [Ben Laurie (CHATS)] 372 373 *) Use sizeof in preference to magic numbers. 374 [Ben Laurie (CHATS)] 375 376 *) Avoid filename truncation in cert requests. 377 [Ben Laurie (CHATS)] 378 379 *) Add assertions to check for (supposedly impossible) buffer 380 overflows. 381 [Ben Laurie (CHATS)] 382 383 *) Don't cache truncated DNS entries in the local cache (this could 384 potentially lead to a spoofing attack). 385 [Ben Laurie (CHATS)] 386 387 *) Fix various buffers to be large enough for hex/decimal 388 representations in a platform independent manner. 389 [Ben Laurie (CHATS)] 390 391 *) Add CRYPTO_realloc_clean() to avoid information leakage when 392 resizing buffers containing secrets, and use where appropriate. 393 [Ben Laurie (CHATS)] 394 395 *) Add BIO_indent() to avoid much slightly worrying code to do 396 indents. 397 [Ben Laurie (CHATS)] 398 399 *) Convert sprintf()/BIO_puts() to BIO_printf(). 400 [Ben Laurie (CHATS)] 401 402 *) buffer_gets() could terminate with the buffer only half 403 full. Fixed. 404 [Ben Laurie (CHATS)] 405 406 *) Add assertions to prevent user-supplied crypto functions from 407 overflowing internal buffers by having large block sizes, etc. 408 [Ben Laurie (CHATS)] 409 410 *) New OPENSSL_assert() macro (similar to assert(), but enabled 411 unconditionally). 412 [Ben Laurie (CHATS)] 413 414 *) Eliminate unused copy of key in RC4. 415 [Ben Laurie (CHATS)] 416 417 *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 418 [Ben Laurie (CHATS)] 419 420 *) Fix off-by-one error in EGD path. 421 [Ben Laurie (CHATS)] 422 423 *) If RANDFILE path is too long, ignore instead of truncating. 424 [Ben Laurie (CHATS)] 425 426 *) Eliminate unused and incorrectly sized X.509 structure 427 CBCParameter. 428 [Ben Laurie (CHATS)] 429 430 *) Eliminate unused and dangerous function knumber(). 431 [Ben Laurie (CHATS)] 432 433 *) Eliminate unused and dangerous structure, KSSL_ERR. 434 [Ben Laurie (CHATS)] 435 436 *) Protect against overlong session ID context length in an encoded 437 session object. Since these are local, this does not appear to be 438 exploitable. 439 [Ben Laurie (CHATS)] 440 441 *) Change from security patch (see 0.9.6e below) that did not affect 442 the 0.9.6 release series: 443 444 Remote buffer overflow in SSL3 protocol - an attacker could 445 supply an oversized master key in Kerberos-enabled versions. 446 (CAN-2002-0657) 447 [Ben Laurie (CHATS)] 448 449 *) Change the SSL kerb5 codes to match RFC 2712. 450 [Richard Levitte] 451 452 *) Make -nameopt work fully for req and add -reqopt switch. 453 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 454 455 *) The "block size" for block ciphers in CFB and OFB mode should be 1. 456 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 457 458 *) Make sure tests can be performed even if the corresponding algorithms 459 have been removed entirely. This was also the last step to make 460 OpenSSL compilable with DJGPP under all reasonable conditions. 461 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 462 463 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 464 to allow version independent disabling of normally unselected ciphers, 465 which may be activated as a side-effect of selecting a single cipher. 466 467 (E.g., cipher list string "RSA" enables ciphersuites that are left 468 out of "ALL" because they do not provide symmetric encryption. 469 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 470 [Lutz Jaenicke, Bodo Moeller] 471 472 *) Add appropriate support for separate platform-dependent build 473 directories. The recommended way to make a platform-dependent 474 build directory is the following (tested on Linux), maybe with 475 some local tweaks: 476 477 # Place yourself outside of the OpenSSL source tree. In 478 # this example, the environment variable OPENSSL_SOURCE 479 # is assumed to contain the absolute OpenSSL source directory. 480 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 481 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 482 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 483 mkdir -p `dirname $F` 484 ln -s $OPENSSL_SOURCE/$F $F 485 done 486 487 To be absolutely sure not to disturb the source tree, a "make clean" 488 is a good thing. If it isn't successfull, don't worry about it, 489 it probably means the source directory is very clean. 490 [Richard Levitte] 491 492 *) Make sure any ENGINE control commands make local copies of string 493 pointers passed to them whenever necessary. Otherwise it is possible 494 the caller may have overwritten (or deallocated) the original string 495 data when a later ENGINE operation tries to use the stored values. 496 [G�tz Babin-Ebell <babinebell@trustcenter.de>] 497 498 *) Improve diagnostics in file reading and command-line digests. 499 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 500 501 *) Add AES modes CFB and OFB to the object database. Correct an 502 error in AES-CFB decryption. 503 [Richard Levitte] 504 505 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 506 allows existing EVP_CIPHER_CTX structures to be reused after 507 calling EVP_*Final(). This behaviour is used by encryption 508 BIOs and some applications. This has the side effect that 509 applications must explicitly clean up cipher contexts with 510 EVP_CIPHER_CTX_cleanup() or they will leak memory. 511 [Steve Henson] 512 513 *) Check the values of dna and dnb in bn_mul_recursive before calling 514 bn_mul_comba (a non zero value means the a or b arrays do not contain 515 n2 elements) and fallback to bn_mul_normal if either is not zero. 516 [Steve Henson] 517 518 *) Fix escaping of non-ASCII characters when using the -subj option 519 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 520 [Lutz Jaenicke] 521 522 *) Make object definitions compliant to LDAP (RFC2256): SN is the short 523 form for "surname", serialNumber has no short form. 524 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 525 therefore remove "mail" short name for "internet 7". 526 The OID for unique identifiers in X509 certificates is 527 x500UniqueIdentifier, not uniqueIdentifier. 528 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 529 [Lutz Jaenicke] 530 531 *) Add an "init" command to the ENGINE config module and auto initialize 532 ENGINEs. Without any "init" command the ENGINE will be initialized 533 after all ctrl commands have been executed on it. If init=1 the 534 ENGINE is initailized at that point (ctrls before that point are run 535 on the uninitialized ENGINE and after on the initialized one). If 536 init=0 then the ENGINE will not be iniatialized at all. 537 [Steve Henson] 538 539 *) Fix the 'app_verify_callback' interface so that the user-defined 540 argument is actually passed to the callback: In the 541 SSL_CTX_set_cert_verify_callback() prototype, the callback 542 declaration has been changed from 543 int (*cb)() 544 into 545 int (*cb)(X509_STORE_CTX *,void *); 546 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 547 i=s->ctx->app_verify_callback(&ctx) 548 has been changed into 549 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 550 551 To update applications using SSL_CTX_set_cert_verify_callback(), 552 a dummy argument can be added to their callback functions. 553 [D. K. Smetters <smetters@parc.xerox.com>] 554 555 *) Added the '4758cca' ENGINE to support IBM 4758 cards. 556 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 557 558 *) Add and OPENSSL_LOAD_CONF define which will cause 559 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 560 This allows older applications to transparently support certain 561 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 562 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 563 load the config file and OPENSSL_add_all_algorithms_conf() which will 564 always load it have also been added. 565 [Steve Henson] 566 567 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 568 Adjust NIDs and EVP layer. 569 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 570 571 *) Config modules support in openssl utility. 572 573 Most commands now load modules from the config file, 574 though in a few (such as version) this isn't done 575 because it couldn't be used for anything. 576 577 In the case of ca and req the config file used is 578 the same as the utility itself: that is the -config 579 command line option can be used to specify an 580 alternative file. 581 [Steve Henson] 582 583 *) Move default behaviour from OPENSSL_config(). If appname is NULL 584 use "openssl_conf" if filename is NULL use default openssl config file. 585 [Steve Henson] 586 587 *) Add an argument to OPENSSL_config() to allow the use of an alternative 588 config section name. Add a new flag to tolerate a missing config file 589 and move code to CONF_modules_load_file(). 590 [Steve Henson] 591 592 *) Support for crypto accelerator cards from Accelerated Encryption 593 Processing, www.aep.ie. (Use engine 'aep') 594 The support was copied from 0.9.6c [engine] and adapted/corrected 595 to work with the new engine framework. 596 [AEP Inc. and Richard Levitte] 597 598 *) Support for SureWare crypto accelerator cards from Baltimore 599 Technologies. (Use engine 'sureware') 600 The support was copied from 0.9.6c [engine] and adapted 601 to work with the new engine framework. 602 [Richard Levitte] 603 604 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 605 make the newer ENGINE framework commands for the CHIL engine work. 606 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 607 608 *) Make it possible to produce shared libraries on ReliantUNIX. 609 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 610 611 *) Add the configuration target debug-linux-ppro. 612 Make 'openssl rsa' use the general key loading routines 613 implemented in apps.c, and make those routines able to 614 handle the key format FORMAT_NETSCAPE and the variant 615 FORMAT_IISSGC. 616 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 617 618 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 619 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 620 621 *) Add -keyform to rsautl, and document -engine. 622 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 623 624 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 625 BIO_R_NO_SUCH_FILE error code rather than the generic 626 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 627 [Ben Laurie] 628 629 *) Add new functions 630 ERR_peek_last_error 631 ERR_peek_last_error_line 632 ERR_peek_last_error_line_data. 633 These are similar to 634 ERR_peek_error 635 ERR_peek_error_line 636 ERR_peek_error_line_data, 637 but report on the latest error recorded rather than the first one 638 still in the error queue. 639 [Ben Laurie, Bodo Moeller] 640 641 *) default_algorithms option in ENGINE config module. This allows things 642 like: 643 default_algorithms = ALL 644 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 645 [Steve Henson] 646 647 *) Prelminary ENGINE config module. 648 [Steve Henson] 649 650 *) New experimental application configuration code. 651 [Steve Henson] 652 653 *) Change the AES code to follow the same name structure as all other 654 symmetric ciphers, and behave the same way. Move everything to 655 the directory crypto/aes, thereby obsoleting crypto/rijndael. 656 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 657 658 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 659 [Ben Laurie and Theo de Raadt] 660 661 *) Add option to output public keys in req command. 662 [Massimiliano Pala madwolf@openca.org] 663 664 *) Use wNAFs in EC_POINTs_mul() for improved efficiency 665 (up to about 10% better than before for P-192 and P-224). 666 [Bodo Moeller] 667 668 *) New functions/macros 669 670 SSL_CTX_set_msg_callback(ctx, cb) 671 SSL_CTX_set_msg_callback_arg(ctx, arg) 672 SSL_set_msg_callback(ssl, cb) 673 SSL_set_msg_callback_arg(ssl, arg) 674 675 to request calling a callback function 676 677 void cb(int write_p, int version, int content_type, 678 const void *buf, size_t len, SSL *ssl, void *arg) 679 680 whenever a protocol message has been completely received 681 (write_p == 0) or sent (write_p == 1). Here 'version' is the 682 protocol version according to which the SSL library interprets 683 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 684 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 685 the content type as defined in the SSL 3.0/TLS 1.0 protocol 686 specification (change_cipher_spec(20), alert(21), handshake(22)). 687 'buf' and 'len' point to the actual message, 'ssl' to the 688 SSL object, and 'arg' is the application-defined value set by 689 SSL[_CTX]_set_msg_callback_arg(). 690 691 'openssl s_client' and 'openssl s_server' have new '-msg' options 692 to enable a callback that displays all protocol messages. 693 [Bodo Moeller] 694 695 *) Change the shared library support so shared libraries are built as 696 soon as the corresponding static library is finished, and thereby get 697 openssl and the test programs linked against the shared library. 698 This still only happens when the keyword "shard" has been given to 699 the configuration scripts. 700 701 NOTE: shared library support is still an experimental thing, and 702 backward binary compatibility is still not guaranteed. 703 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 704 705 *) Add support for Subject Information Access extension. 706 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 707 708 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 709 additional bytes when new memory had to be allocated, not just 710 when reusing an existing buffer. 711 [Bodo Moeller] 712 713 *) New command line and configuration option 'utf8' for the req command. 714 This allows field values to be specified as UTF8 strings. 715 [Steve Henson] 716 717 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 718 runs for the former and machine-readable output for the latter. 719 [Ben Laurie] 720 721 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 722 of the e-mail address in the DN (i.e., it will go into a certificate 723 extension only). The new configuration file option 'email_in_dn = no' 724 has the same effect. 725 [Massimiliano Pala madwolf@openca.org] 726 727 *) Change all functions with names starting with des_ to be starting 728 with DES_ instead. Add wrappers that are compatible with libdes, 729 but are named _ossl_old_des_*. Finally, add macros that map the 730 des_* symbols to the corresponding _ossl_old_des_* if libdes 731 compatibility is desired. If OpenSSL 0.9.6c compatibility is 732 desired, the des_* symbols will be mapped to DES_*, with one 733 exception. 734 735 Since we provide two compatibility mappings, the user needs to 736 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 737 compatibility is desired. The default (i.e., when that macro 738 isn't defined) is OpenSSL 0.9.6c compatibility. 739 740 There are also macros that enable and disable the support of old 741 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 742 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 743 are defined, the default will apply: to support the old des routines. 744 745 In either case, one must include openssl/des.h to get the correct 746 definitions. Do not try to just include openssl/des_old.h, that 747 won't work. 748 749 NOTE: This is a major break of an old API into a new one. Software 750 authors are encouraged to switch to the DES_ style functions. Some 751 time in the future, des_old.h and the libdes compatibility functions 752 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 753 default), and then completely removed. 754 [Richard Levitte] 755 756 *) Test for certificates which contain unsupported critical extensions. 757 If such a certificate is found during a verify operation it is 758 rejected by default: this behaviour can be overridden by either 759 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 760 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 761 X509_supported_extension() has also been added which returns 1 if a 762 particular extension is supported. 763 [Steve Henson] 764 765 *) Modify the behaviour of EVP cipher functions in similar way to digests 766 to retain compatibility with existing code. 767 [Steve Henson] 768 769 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 770 compatibility with existing code. In particular the 'ctx' parameter does 771 not have to be to be initialized before the call to EVP_DigestInit() and 772 it is tidied up after a call to EVP_DigestFinal(). New function 773 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 774 EVP_MD_CTX_copy() changed to not require the destination to be 775 initialized valid and new function EVP_MD_CTX_copy_ex() added which 776 requires the destination to be valid. 777 778 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 779 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 780 [Steve Henson] 781 782 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 783 so that complete 'Handshake' protocol structures are kept in memory 784 instead of overwriting 'msg_type' and 'length' with 'body' data. 785 [Bodo Moeller] 786 787 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 788 [Massimo Santin via Richard Levitte] 789 790 *) Major restructuring to the underlying ENGINE code. This includes 791 reduction of linker bloat, separation of pure "ENGINE" manipulation 792 (initialisation, etc) from functionality dealing with implementations 793 of specific crypto iterfaces. This change also introduces integrated 794 support for symmetric ciphers and digest implementations - so ENGINEs 795 can now accelerate these by providing EVP_CIPHER and EVP_MD 796 implementations of their own. This is detailed in crypto/engine/README 797 as it couldn't be adequately described here. However, there are a few 798 API changes worth noting - some RSA, DSA, DH, and RAND functions that 799 were changed in the original introduction of ENGINE code have now 800 reverted back - the hooking from this code to ENGINE is now a good 801 deal more passive and at run-time, operations deal directly with 802 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 803 dereferencing through an ENGINE pointer any more. Also, the ENGINE 804 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 805 they were not being used by the framework as there is no concept of a 806 BIGNUM_METHOD and they could not be generalised to the new 807 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 808 ENGINE_cpy() has been removed as it cannot be consistently defined in 809 the new code. 810 [Geoff Thorpe] 811 812 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 813 [Steve Henson] 814 815 *) Change mkdef.pl to sort symbols that get the same entry number, 816 and make sure the automatically generated functions ERR_load_* 817 become part of libeay.num as well. 818 [Richard Levitte] 819 820 *) New function SSL_renegotiate_pending(). This returns true once 821 renegotiation has been requested (either SSL_renegotiate() call 822 or HelloRequest/ClientHello receveived from the peer) and becomes 823 false once a handshake has been completed. 824 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 825 sends a HelloRequest, but does not ensure that a handshake takes 826 place. SSL_renegotiate_pending() is useful for checking if the 827 client has followed the request.) 828 [Bodo Moeller] 829 830 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 831 By default, clients may request session resumption even during 832 renegotiation (if session ID contexts permit); with this option, 833 session resumption is possible only in the first handshake. 834 835 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 836 more bits available for options that should not be part of 837 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 838 [Bodo Moeller] 839 840 *) Add some demos for certificate and certificate request creation. 841 [Steve Henson] 842 843 *) Make maximum certificate chain size accepted from the peer application 844 settable (SSL*_get/set_max_cert_list()), as proposed by 845 "Douglas E. Engert" <deengert@anl.gov>. 846 [Lutz Jaenicke] 847 848 *) Add support for shared libraries for Unixware-7 849 (Boyd Lynn Gerber <gerberb@zenez.com>). 850 [Lutz Jaenicke] 851 852 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 853 be done prior to destruction. Use this to unload error strings from 854 ENGINEs that load their own error strings. NB: This adds two new API 855 functions to "get" and "set" this destroy handler in an ENGINE. 856 [Geoff Thorpe] 857 858 *) Alter all existing ENGINE implementations (except "openssl" and 859 "openbsd") to dynamically instantiate their own error strings. This 860 makes them more flexible to be built both as statically-linked ENGINEs 861 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 862 Also, add stub code to each that makes building them as self-contained 863 shared-libraries easier (see README.ENGINE). 864 [Geoff Thorpe] 865 866 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 867 implementations into applications that are completely implemented in 868 self-contained shared-libraries. The "dynamic" ENGINE exposes control 869 commands that can be used to configure what shared-library to load and 870 to control aspects of the way it is handled. Also, made an update to 871 the README.ENGINE file that brings its information up-to-date and 872 provides some information and instructions on the "dynamic" ENGINE 873 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 874 [Geoff Thorpe] 875 876 *) Make it possible to unload ranges of ERR strings with a new 877 "ERR_unload_strings" function. 878 [Geoff Thorpe] 879 880 *) Add a copy() function to EVP_MD. 881 [Ben Laurie] 882 883 *) Make EVP_MD routines take a context pointer instead of just the 884 md_data void pointer. 885 [Ben Laurie] 886 887 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 888 that the digest can only process a single chunk of data 889 (typically because it is provided by a piece of 890 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 891 is only going to provide a single chunk of data, and hence the 892 framework needn't accumulate the data for oneshot drivers. 893 [Ben Laurie] 894 895 *) As with "ERR", make it possible to replace the underlying "ex_data" 896 functions. This change also alters the storage and management of global 897 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 898 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 899 index counters. The API functions that use this state have been changed 900 to take a "class_index" rather than pointers to the class's local STACK 901 and counter, and there is now an API function to dynamically create new 902 classes. This centralisation allows us to (a) plug a lot of the 903 thread-safety problems that existed, and (b) makes it possible to clean 904 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 905 such data would previously have always leaked in application code and 906 workarounds were in place to make the memory debugging turn a blind eye 907 to it. Application code that doesn't use this new function will still 908 leak as before, but their memory debugging output will announce it now 909 rather than letting it slide. 910 911 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 912 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 913 has a return value to indicate success or failure. 914 [Geoff Thorpe] 915 916 *) Make it possible to replace the underlying "ERR" functions such that the 917 global state (2 LHASH tables and 2 locks) is only used by the "default" 918 implementation. This change also adds two functions to "get" and "set" 919 the implementation prior to it being automatically set the first time 920 any other ERR function takes place. Ie. an application can call "get", 921 pass the return value to a module it has just loaded, and that module 922 can call its own "set" function using that value. This means the 923 module's "ERR" operations will use (and modify) the error state in the 924 application and not in its own statically linked copy of OpenSSL code. 925 [Geoff Thorpe] 926 927 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 928 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 929 the operation, and provides a more encapsulated way for external code 930 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 931 to use these functions rather than manually incrementing the counts. 932 933 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 934 [Geoff Thorpe] 935 936 *) Add EVP test program. 937 [Ben Laurie] 938 939 *) Add symmetric cipher support to ENGINE. Expect the API to change! 940 [Ben Laurie] 941 942 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 943 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 944 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 945 These allow a CRL to be built without having to access X509_CRL fields 946 directly. Modify 'ca' application to use new functions. 947 [Steve Henson] 948 949 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 950 bug workarounds. Rollback attack detection is a security feature. 951 The problem will only arise on OpenSSL servers when TLSv1 is not 952 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 953 Software authors not wanting to support TLSv1 will have special reasons 954 for their choice and can explicitly enable this option. 955 [Bodo Moeller, Lutz Jaenicke] 956 957 *) Rationalise EVP so it can be extended: don't include a union of 958 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 959 (similar to those existing for EVP_CIPHER_CTX). 960 Usage example: 961 962 EVP_MD_CTX md; 963 964 EVP_MD_CTX_init(&md); /* new function call */ 965 EVP_DigestInit(&md, EVP_sha1()); 966 EVP_DigestUpdate(&md, in, len); 967 EVP_DigestFinal(&md, out, NULL); 968 EVP_MD_CTX_cleanup(&md); /* new function call */ 969 970 [Ben Laurie] 971 972 *) Make DES key schedule conform to the usual scheme, as well as 973 correcting its structure. This means that calls to DES functions 974 now have to pass a pointer to a des_key_schedule instead of a 975 plain des_key_schedule (which was actually always a pointer 976 anyway): E.g., 977 978 des_key_schedule ks; 979 980 des_set_key_checked(..., &ks); 981 des_ncbc_encrypt(..., &ks, ...); 982 983 (Note that a later change renames 'des_...' into 'DES_...'.) 984 [Ben Laurie] 985 986 *) Initial reduction of linker bloat: the use of some functions, such as 987 PEM causes large amounts of unused functions to be linked in due to 988 poor organisation. For example pem_all.c contains every PEM function 989 which has a knock on effect of linking in large amounts of (unused) 990 ASN1 code. Grouping together similar functions and splitting unrelated 991 functions prevents this. 992 [Steve Henson] 993 994 *) Cleanup of EVP macros. 995 [Ben Laurie] 996 997 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 998 correct _ecb suffix. 999 [Ben Laurie] 1000 1001 *) Add initial OCSP responder support to ocsp application. The 1002 revocation information is handled using the text based index 1003 use by the ca application. The responder can either handle 1004 requests generated internally, supplied in files (for example 1005 via a CGI script) or using an internal minimal server. 1006 [Steve Henson] 1007 1008 *) Add configuration choices to get zlib compression for TLS. 1009 [Richard Levitte] 1010 1011 *) Changes to Kerberos SSL for RFC 2712 compliance: 1012 1. Implemented real KerberosWrapper, instead of just using 1013 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 1014 2. Implemented optional authenticator field of KerberosWrapper. 1015 1016 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 1017 and authenticator structs; see crypto/krb5/. 1018 1019 Generalized Kerberos calls to support multiple Kerberos libraries. 1020 [Vern Staats <staatsvr@asc.hpc.mil>, 1021 Jeffrey Altman <jaltman@columbia.edu> 1022 via Richard Levitte] 1023 1024 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 1025 already does with RSA. testdsa.h now has 'priv_key/pub_key' 1026 values for each of the key sizes rather than having just 1027 parameters (and 'speed' generating keys each time). 1028 [Geoff Thorpe] 1029 1030 *) Speed up EVP routines. 1031 Before: 1032encrypt 1033type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 1034des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 1035des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 1036des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 1037decrypt 1038des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 1039des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 1040des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 1041 After: 1042encrypt 1043des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 1044decrypt 1045des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 1046 [Ben Laurie] 1047 1048 *) Added the OS2-EMX target. 1049 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 1050 1051 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 1052 to support NCONF routines in extension code. New function CONF_set_nconf() 1053 to allow functions which take an NCONF to also handle the old LHASH 1054 structure: this means that the old CONF compatible routines can be 1055 retained (in particular wrt extensions) without having to duplicate the 1056 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 1057 [Steve Henson] 1058 1059 *) Enhance the general user interface with mechanisms for inner control 1060 and with possibilities to have yes/no kind of prompts. 1061 [Richard Levitte] 1062 1063 *) Change all calls to low level digest routines in the library and 1064 applications to use EVP. Add missing calls to HMAC_cleanup() and 1065 don't assume HMAC_CTX can be copied using memcpy(). 1066 [Verdon Walker <VWalker@novell.com>, Steve Henson] 1067 1068 *) Add the possibility to control engines through control names but with 1069 arbitrary arguments instead of just a string. 1070 Change the key loaders to take a UI_METHOD instead of a callback 1071 function pointer. NOTE: this breaks binary compatibility with earlier 1072 versions of OpenSSL [engine]. 1073 Adapt the nCipher code for these new conditions and add a card insertion 1074 callback. 1075 [Richard Levitte] 1076 1077 *) Enhance the general user interface with mechanisms to better support 1078 dialog box interfaces, application-defined prompts, the possibility 1079 to use defaults (for example default passwords from somewhere else) 1080 and interrupts/cancellations. 1081 [Richard Levitte] 1082 1083 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 1084 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 1085 [Steve Henson] 1086 1087 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 1088 tidy up some unnecessarily weird code in 'sk_new()'). 1089 [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 1090 1091 *) Change the key loading routines for ENGINEs to use the same kind 1092 callback (pem_password_cb) as all other routines that need this 1093 kind of callback. 1094 [Richard Levitte] 1095 1096 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 1097 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 1098 than this minimum value is recommended. 1099 [Lutz Jaenicke] 1100 1101 *) New random seeder for OpenVMS, using the system process statistics 1102 that are easily reachable. 1103 [Richard Levitte] 1104 1105 *) Windows apparently can't transparently handle global 1106 variables defined in DLLs. Initialisations such as: 1107 1108 const ASN1_ITEM *it = &ASN1_INTEGER_it; 1109 1110 wont compile. This is used by the any applications that need to 1111 declare their own ASN1 modules. This was fixed by adding the option 1112 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 1113 needed for static libraries under Win32. 1114 [Steve Henson] 1115 1116 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 1117 setting of purpose and trust fields. New X509_STORE trust and 1118 purpose functions and tidy up setting in other SSL functions. 1119 [Steve Henson] 1120 1121 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 1122 structure. These are inherited by X509_STORE_CTX when it is 1123 initialised. This allows various defaults to be set in the 1124 X509_STORE structure (such as flags for CRL checking and custom 1125 purpose or trust settings) for functions which only use X509_STORE_CTX 1126 internally such as S/MIME. 1127 1128 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 1129 trust settings if they are not set in X509_STORE. This allows X509_STORE 1130 purposes and trust (in S/MIME for example) to override any set by default. 1131 1132 Add command line options for CRL checking to smime, s_client and s_server 1133 applications. 1134 [Steve Henson] 1135 1136 *) Initial CRL based revocation checking. If the CRL checking flag(s) 1137 are set then the CRL is looked up in the X509_STORE structure and 1138 its validity and signature checked, then if the certificate is found 1139 in the CRL the verify fails with a revoked error. 1140 1141 Various new CRL related callbacks added to X509_STORE_CTX structure. 1142 1143 Command line options added to 'verify' application to support this. 1144 1145 This needs some additional work, such as being able to handle multiple 1146 CRLs with different times, extension based lookup (rather than just 1147 by subject name) and ultimately more complete V2 CRL extension 1148 handling. 1149 [Steve Henson] 1150 1151 *) Add a general user interface API (crypto/ui/). This is designed 1152 to replace things like des_read_password and friends (backward 1153 compatibility functions using this new API are provided). 1154 The purpose is to remove prompting functions from the DES code 1155 section as well as provide for prompting through dialog boxes in 1156 a window system and the like. 1157 [Richard Levitte] 1158 1159 *) Add "ex_data" support to ENGINE so implementations can add state at a 1160 per-structure level rather than having to store it globally. 1161 [Geoff] 1162 1163 *) Make it possible for ENGINE structures to be copied when retrieved by 1164 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 1165 This causes the "original" ENGINE structure to act like a template, 1166 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 1167 operational state can be localised to each ENGINE structure, despite the 1168 fact they all share the same "methods". New ENGINE structures returned in 1169 this case have no functional references and the return value is the single 1170 structural reference. This matches the single structural reference returned 1171 by ENGINE_by_id() normally, when it is incremented on the pre-existing 1172 ENGINE structure. 1173 [Geoff] 1174 1175 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 1176 needs to match any other type at all we need to manually clear the 1177 tag cache. 1178 [Steve Henson] 1179 1180 *) Changes to the "openssl engine" utility to include; 1181 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 1182 about an ENGINE's available control commands. 1183 - executing control commands from command line arguments using the 1184 '-pre' and '-post' switches. '-post' is only used if '-t' is 1185 specified and the ENGINE is successfully initialised. The syntax for 1186 the individual commands are colon-separated, for example; 1187 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 1188 [Geoff] 1189 1190 *) New dynamic control command support for ENGINEs. ENGINEs can now 1191 declare their own commands (numbers), names (strings), descriptions, 1192 and input types for run-time discovery by calling applications. A 1193 subset of these commands are implicitly classed as "executable" 1194 depending on their input type, and only these can be invoked through 1195 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 1196 can be based on user input, config files, etc). The distinction is 1197 that "executable" commands cannot return anything other than a boolean 1198 result and can only support numeric or string input, whereas some 1199 discoverable commands may only be for direct use through 1200 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 1201 pointers, or other custom uses. The "executable" commands are to 1202 support parameterisations of ENGINE behaviour that can be 1203 unambiguously defined by ENGINEs and used consistently across any 1204 OpenSSL-based application. Commands have been added to all the 1205 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 1206 control over shared-library paths without source code alterations. 1207 [Geoff] 1208 1209 *) Changed all ENGINE implementations to dynamically allocate their 1210 ENGINEs rather than declaring them statically. Apart from this being 1211 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 1212 this also allows the implementations to compile without using the 1213 internal engine_int.h header. 1214 [Geoff] 1215 1216 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 1217 'const' value. Any code that should be able to modify a RAND_METHOD 1218 should already have non-const pointers to it (ie. they should only 1219 modify their own ones). 1220 [Geoff] 1221 1222 *) Made a variety of little tweaks to the ENGINE code. 1223 - "atalla" and "ubsec" string definitions were moved from header files 1224 to C code. "nuron" string definitions were placed in variables 1225 rather than hard-coded - allowing parameterisation of these values 1226 later on via ctrl() commands. 1227 - Removed unused "#if 0"'d code. 1228 - Fixed engine list iteration code so it uses ENGINE_free() to release 1229 structural references. 1230 - Constified the RAND_METHOD element of ENGINE structures. 1231 - Constified various get/set functions as appropriate and added 1232 missing functions (including a catch-all ENGINE_cpy that duplicates 1233 all ENGINE values onto a new ENGINE except reference counts/state). 1234 - Removed NULL parameter checks in get/set functions. Setting a method 1235 or function to NULL is a way of cancelling out a previously set 1236 value. Passing a NULL ENGINE parameter is just plain stupid anyway 1237 and doesn't justify the extra error symbols and code. 1238 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 1239 flags from engine_int.h to engine.h. 1240 - Changed prototypes for ENGINE handler functions (init(), finish(), 1241 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 1242 [Geoff] 1243 1244 *) Implement binary inversion algorithm for BN_mod_inverse in addition 1245 to the algorithm using long division. The binary algorithm can be 1246 used only if the modulus is odd. On 32-bit systems, it is faster 1247 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 1248 roughly 5-15% for 256-bit moduli), so we use it only for moduli 1249 up to 450 bits. In 64-bit environments, the binary algorithm 1250 appears to be advantageous for much longer moduli; here we use it 1251 for moduli up to 2048 bits. 1252 [Bodo Moeller] 1253 1254 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 1255 could not support the combine flag in choice fields. 1256 [Steve Henson] 1257 1258 *) Add a 'copy_extensions' option to the 'ca' utility. This copies 1259 extensions from a certificate request to the certificate. 1260 [Steve Henson] 1261 1262 *) Allow multiple 'certopt' and 'nameopt' options to be separated 1263 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 1264 file: this allows the display of the certificate about to be 1265 signed to be customised, to allow certain fields to be included 1266 or excluded and extension details. The old system didn't display 1267 multicharacter strings properly, omitted fields not in the policy 1268 and couldn't display additional details such as extensions. 1269 [Steve Henson] 1270 1271 *) Function EC_POINTs_mul for multiple scalar multiplication 1272 of an arbitrary number of elliptic curve points 1273 \sum scalars[i]*points[i], 1274 optionally including the generator defined for the EC_GROUP: 1275 scalar*generator + \sum scalars[i]*points[i]. 1276 1277 EC_POINT_mul is a simple wrapper function for the typical case 1278 that the point list has just one item (besides the optional 1279 generator). 1280 [Bodo Moeller] 1281 1282 *) First EC_METHODs for curves over GF(p): 1283 1284 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 1285 operations and provides various method functions that can also 1286 operate with faster implementations of modular arithmetic. 1287 1288 EC_GFp_mont_method() reuses most functions that are part of 1289 EC_GFp_simple_method, but uses Montgomery arithmetic. 1290 1291 [Bodo Moeller; point addition and point doubling 1292 implementation directly derived from source code provided by 1293 Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 1294 1295 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 1296 crypto/ec/ec_lib.c): 1297 1298 Curves are EC_GROUP objects (with an optional group generator) 1299 based on EC_METHODs that are built into the library. 1300 1301 Points are EC_POINT objects based on EC_GROUP objects. 1302 1303 Most of the framework would be able to handle curves over arbitrary 1304 finite fields, but as there are no obvious types for fields other 1305 than GF(p), some functions are limited to that for now. 1306 [Bodo Moeller] 1307 1308 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 1309 that the file contains a complete HTTP response. 1310 [Richard Levitte] 1311 1312 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 1313 change the def and num file printf format specifier from "%-40sXXX" 1314 to "%-39s XXX". The latter will always guarantee a space after the 1315 field while the former will cause them to run together if the field 1316 is 40 of more characters long. 1317 [Steve Henson] 1318 1319 *) Constify the cipher and digest 'method' functions and structures 1320 and modify related functions to take constant EVP_MD and EVP_CIPHER 1321 pointers. 1322 [Steve Henson] 1323 1324 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 1325 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 1326 [Bodo Moeller] 1327 1328 *) Modify EVP_Digest*() routines so they now return values. Although the 1329 internal software routines can never fail additional hardware versions 1330 might. 1331 [Steve Henson] 1332 1333 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 1334 1335 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 1336 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 1337 1338 ASN1 error codes 1339 ERR_R_NESTED_ASN1_ERROR 1340 ... 1341 ERR_R_MISSING_ASN1_EOS 1342 were 4 .. 9, conflicting with 1343 ERR_LIB_RSA (= ERR_R_RSA_LIB) 1344 ... 1345 ERR_LIB_PEM (= ERR_R_PEM_LIB). 1346 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 1347 1348 Add new error code 'ERR_R_INTERNAL_ERROR'. 1349 [Bodo Moeller] 1350 1351 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 1352 suffices. 1353 [Bodo Moeller] 1354 1355 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 1356 sets the subject name for a new request or supersedes the 1357 subject name in a given request. Formats that can be parsed are 1358 'CN=Some Name, OU=myOU, C=IT' 1359 and 1360 'CN=Some Name/OU=myOU/C=IT'. 1361 1362 Add options '-batch' and '-verbose' to 'openssl req'. 1363 [Massimiliano Pala <madwolf@hackmasters.net>] 1364 1365 *) Introduce the possibility to access global variables through 1366 functions on platform were that's the best way to handle exporting 1367 global variables in shared libraries. To enable this functionality, 1368 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 1369 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 1370 is normally done by Configure or something similar). 1371 1372 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 1373 in the source file (foo.c) like this: 1374 1375 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 1376 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 1377 1378 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 1379 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 1380 1381 OPENSSL_DECLARE_GLOBAL(int,foo); 1382 #define foo OPENSSL_GLOBAL_REF(foo) 1383 OPENSSL_DECLARE_GLOBAL(double,bar); 1384 #define bar OPENSSL_GLOBAL_REF(bar) 1385 1386 The #defines are very important, and therefore so is including the 1387 header file everywhere where the defined globals are used. 1388 1389 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 1390 of ASN.1 items, but that structure is a bit different. 1391 1392 The largest change is in util/mkdef.pl which has been enhanced with 1393 better and easier to understand logic to choose which symbols should 1394 go into the Windows .def files as well as a number of fixes and code 1395 cleanup (among others, algorithm keywords are now sorted 1396 lexicographically to avoid constant rewrites). 1397 [Richard Levitte] 1398 1399 *) In BN_div() keep a copy of the sign of 'num' before writing the 1400 result to 'rm' because if rm==num the value will be overwritten 1401 and produce the wrong result if 'num' is negative: this caused 1402 problems with BN_mod() and BN_nnmod(). 1403 [Steve Henson] 1404 1405 *) Function OCSP_request_verify(). This checks the signature on an 1406 OCSP request and verifies the signer certificate. The signer 1407 certificate is just checked for a generic purpose and OCSP request 1408 trust settings. 1409 [Steve Henson] 1410 1411 *) Add OCSP_check_validity() function to check the validity of OCSP 1412 responses. OCSP responses are prepared in real time and may only 1413 be a few seconds old. Simply checking that the current time lies 1414 between thisUpdate and nextUpdate max reject otherwise valid responses 1415 caused by either OCSP responder or client clock inaccuracy. Instead 1416 we allow thisUpdate and nextUpdate to fall within a certain period of 1417 the current time. The age of the response can also optionally be 1418 checked. Two new options -validity_period and -status_age added to 1419 ocsp utility. 1420 [Steve Henson] 1421 1422 *) If signature or public key algorithm is unrecognized print out its 1423 OID rather that just UNKNOWN. 1424 [Steve Henson] 1425 1426 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 1427 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 1428 ID to be generated from the issuer certificate alone which can then be 1429 passed to OCSP_id_issuer_cmp(). 1430 [Steve Henson] 1431 1432 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 1433 ASN1 modules to export functions returning ASN1_ITEM pointers 1434 instead of the ASN1_ITEM structures themselves. This adds several 1435 new macros which allow the underlying ASN1 function/structure to 1436 be accessed transparently. As a result code should not use ASN1_ITEM 1437 references directly (such as &X509_it) but instead use the relevant 1438 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 1439 use of the new ASN1 code on platforms where exporting structures 1440 is problematical (for example in shared libraries) but exporting 1441 functions returning pointers to structures is not. 1442 [Steve Henson] 1443 1444 *) Add support for overriding the generation of SSL/TLS session IDs. 1445 These callbacks can be registered either in an SSL_CTX or per SSL. 1446 The purpose of this is to allow applications to control, if they wish, 1447 the arbitrary values chosen for use as session IDs, particularly as it 1448 can be useful for session caching in multiple-server environments. A 1449 command-line switch for testing this (and any client code that wishes 1450 to use such a feature) has been added to "s_server". 1451 [Geoff Thorpe, Lutz Jaenicke] 1452 1453 *) Modify mkdef.pl to recognise and parse preprocessor conditionals 1454 of the form '#if defined(...) || defined(...) || ...' and 1455 '#if !defined(...) && !defined(...) && ...'. This also avoids 1456 the growing number of special cases it was previously handling. 1457 [Richard Levitte] 1458 1459 *) Make all configuration macros available for application by making 1460 sure they are available in opensslconf.h, by giving them names starting 1461 with "OPENSSL_" to avoid conflicts with other packages and by making 1462 sure e_os2.h will cover all platform-specific cases together with 1463 opensslconf.h. 1464 Additionally, it is now possible to define configuration/platform- 1465 specific names (called "system identities"). In the C code, these 1466 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 1467 macro with the name beginning with "OPENSSL_SYS_", which is determined 1468 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 1469 what is available. 1470 [Richard Levitte] 1471 1472 *) New option -set_serial to 'req' and 'x509' this allows the serial 1473 number to use to be specified on the command line. Previously self 1474 signed certificates were hard coded with serial number 0 and the 1475 CA options of 'x509' had to use a serial number in a file which was 1476 auto incremented. 1477 [Steve Henson] 1478 1479 *) New options to 'ca' utility to support V2 CRL entry extensions. 1480 Currently CRL reason, invalidity date and hold instruction are 1481 supported. Add new CRL extensions to V3 code and some new objects. 1482 [Steve Henson] 1483 1484 *) New function EVP_CIPHER_CTX_set_padding() this is used to 1485 disable standard block padding (aka PKCS#5 padding) in the EVP 1486 API, which was previously mandatory. This means that the data is 1487 not padded in any way and so the total length much be a multiple 1488 of the block size, otherwise an error occurs. 1489 [Steve Henson] 1490 1491 *) Initial (incomplete) OCSP SSL support. 1492 [Steve Henson] 1493 1494 *) New function OCSP_parse_url(). This splits up a URL into its host, 1495 port and path components: primarily to parse OCSP URLs. New -url 1496 option to ocsp utility. 1497 [Steve Henson] 1498 1499 *) New nonce behavior. The return value of OCSP_check_nonce() now 1500 reflects the various checks performed. Applications can decide 1501 whether to tolerate certain situations such as an absent nonce 1502 in a response when one was present in a request: the ocsp application 1503 just prints out a warning. New function OCSP_add1_basic_nonce() 1504 this is to allow responders to include a nonce in a response even if 1505 the request is nonce-less. 1506 [Steve Henson] 1507 1508 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 1509 skipped when using openssl x509 multiple times on a single input file, 1510 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 1511 [Bodo Moeller] 1512 1513 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 1514 set string type: to handle setting ASN1_TIME structures. Fix ca 1515 utility to correctly initialize revocation date of CRLs. 1516 [Steve Henson] 1517 1518 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 1519 the clients preferred ciphersuites and rather use its own preferences. 1520 Should help to work around M$ SGC (Server Gated Cryptography) bug in 1521 Internet Explorer by ensuring unchanged hash method during stepup. 1522 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 1523 [Lutz Jaenicke] 1524 1525 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 1526 to aes and add a new 'exist' option to print out symbols that don't 1527 appear to exist. 1528 [Steve Henson] 1529 1530 *) Additional options to ocsp utility to allow flags to be set and 1531 additional certificates supplied. 1532 [Steve Henson] 1533 1534 *) Add the option -VAfile to 'openssl ocsp', so the user can give the 1535 OCSP client a number of certificate to only verify the response 1536 signature against. 1537 [Richard Levitte] 1538 1539 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 1540 handle the new API. Currently only ECB, CBC modes supported. Add new 1541 AES OIDs. 1542 1543 Add TLS AES ciphersuites as described in RFC3268, "Advanced 1544 Encryption Standard (AES) Ciphersuites for Transport Layer 1545 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 1546 not enabled by default and were not part of the "ALL" ciphersuite 1547 alias because they were not yet official; they could be 1548 explicitly requested by specifying the "AESdraft" ciphersuite 1549 group alias. In the final release of OpenSSL 0.9.7, the group 1550 alias is called "AES" and is part of "ALL".) 1551 [Ben Laurie, Steve Henson, Bodo Moeller] 1552 1553 *) New function OCSP_copy_nonce() to copy nonce value (if present) from 1554 request to response. 1555 [Steve Henson] 1556 1557 *) Functions for OCSP responders. OCSP_request_onereq_count(), 1558 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 1559 extract information from a certificate request. OCSP_response_create() 1560 creates a response and optionally adds a basic response structure. 1561 OCSP_basic_add1_status() adds a complete single response to a basic 1562 response and returns the OCSP_SINGLERESP structure just added (to allow 1563 extensions to be included for example). OCSP_basic_add1_cert() adds a 1564 certificate to a basic response and OCSP_basic_sign() signs a basic 1565 response with various flags. New helper functions ASN1_TIME_check() 1566 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 1567 (converts ASN1_TIME to GeneralizedTime). 1568 [Steve Henson] 1569 1570 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 1571 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 1572 structure from a certificate. X509_pubkey_digest() digests the public_key 1573 contents: this is used in various key identifiers. 1574 [Steve Henson] 1575 1576 *) Make sk_sort() tolerate a NULL argument. 1577 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 1578 1579 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 1580 passed by the function are trusted implicitly. If any of them signed the 1581 response then it is assumed to be valid and is not verified. 1582 [Steve Henson] 1583 1584 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 1585 to data. This was previously part of the PKCS7 ASN1 code. This 1586 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 1587 [Steve Henson, reported by Kenneth R. Robinette 1588 <support@securenetterm.com>] 1589 1590 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 1591 routines: without these tracing memory leaks is very painful. 1592 Fix leaks in PKCS12 and PKCS7 routines. 1593 [Steve Henson] 1594 1595 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 1596 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 1597 effectively meant GeneralizedTime would never be used. Now it 1598 is initialised to -1 but X509_time_adj() now has to check the value 1599 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 1600 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 1601 [Steve Henson, reported by Kenneth R. Robinette 1602 <support@securenetterm.com>] 1603 1604 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 1605 result in a zero length in the ASN1_INTEGER structure which was 1606 not consistent with the structure when d2i_ASN1_INTEGER() was used 1607 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 1608 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 1609 where it did not print out a minus for negative ASN1_INTEGER. 1610 [Steve Henson] 1611 1612 *) Add summary printout to ocsp utility. The various functions which 1613 convert status values to strings have been renamed to: 1614 OCSP_response_status_str(), OCSP_cert_status_str() and 1615 OCSP_crl_reason_str() and are no longer static. New options 1616 to verify nonce values and to disable verification. OCSP response 1617 printout format cleaned up. 1618 [Steve Henson] 1619 1620 *) Add additional OCSP certificate checks. These are those specified 1621 in RFC2560. This consists of two separate checks: the CA of the 1622 certificate being checked must either be the OCSP signer certificate 1623 or the issuer of the OCSP signer certificate. In the latter case the 1624 OCSP signer certificate must contain the OCSP signing extended key 1625 usage. This check is performed by attempting to match the OCSP 1626 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 1627 in the OCSP_CERTID structures of the response. 1628 [Steve Henson] 1629 1630 *) Initial OCSP certificate verification added to OCSP_basic_verify() 1631 and related routines. This uses the standard OpenSSL certificate 1632 verify routines to perform initial checks (just CA validity) and 1633 to obtain the certificate chain. Then additional checks will be 1634 performed on the chain. Currently the root CA is checked to see 1635 if it is explicitly trusted for OCSP signing. This is used to set 1636 a root CA as a global signing root: that is any certificate that 1637 chains to that CA is an acceptable OCSP signing certificate. 1638 [Steve Henson] 1639 1640 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 1641 extensions from a separate configuration file. 1642 As when reading extensions from the main configuration file, 1643 the '-extensions ...' option may be used for specifying the 1644 section to use. 1645 [Massimiliano Pala <madwolf@comune.modena.it>] 1646 1647 *) New OCSP utility. Allows OCSP requests to be generated or 1648 read. The request can be sent to a responder and the output 1649 parsed, outputed or printed in text form. Not complete yet: 1650 still needs to check the OCSP response validity. 1651 [Steve Henson] 1652 1653 *) New subcommands for 'openssl ca': 1654 'openssl ca -status <serial>' prints the status of the cert with 1655 the given serial number (according to the index file). 1656 'openssl ca -updatedb' updates the expiry status of certificates 1657 in the index file. 1658 [Massimiliano Pala <madwolf@comune.modena.it>] 1659 1660 *) New '-newreq-nodes' command option to CA.pl. This is like 1661 '-newreq', but calls 'openssl req' with the '-nodes' option 1662 so that the resulting key is not encrypted. 1663 [Damien Miller <djm@mindrot.org>] 1664 1665 *) New configuration for the GNU Hurd. 1666 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 1667 1668 *) Initial code to implement OCSP basic response verify. This 1669 is currently incomplete. Currently just finds the signer's 1670 certificate and verifies the signature on the response. 1671 [Steve Henson] 1672 1673 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 1674 value of OPENSSLDIR. This is available via the new '-d' option 1675 to 'openssl version', and is also included in 'openssl version -a'. 1676 [Bodo Moeller] 1677 1678 *) Allowing defining memory allocation callbacks that will be given 1679 file name and line number information in additional arguments 1680 (a const char* and an int). The basic functionality remains, as 1681 well as the original possibility to just replace malloc(), 1682 realloc() and free() by functions that do not know about these 1683 additional arguments. To register and find out the current 1684 settings for extended allocation functions, the following 1685 functions are provided: 1686 1687 CRYPTO_set_mem_ex_functions 1688 CRYPTO_set_locked_mem_ex_functions 1689 CRYPTO_get_mem_ex_functions 1690 CRYPTO_get_locked_mem_ex_functions 1691 1692 These work the same way as CRYPTO_set_mem_functions and friends. 1693 CRYPTO_get_[locked_]mem_functions now writes 0 where such an 1694 extended allocation function is enabled. 1695 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 1696 a conventional allocation function is enabled. 1697 [Richard Levitte, Bodo Moeller] 1698 1699 *) Finish off removing the remaining LHASH function pointer casts. 1700 There should no longer be any prototype-casting required when using 1701 the LHASH abstraction, and any casts that remain are "bugs". See 1702 the callback types and macros at the head of lhash.h for details 1703 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 1704 [Geoff Thorpe] 1705 1706 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 1707 If /dev/[u]random devices are not available or do not return enough 1708 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 1709 be queried. 1710 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 1711 /etc/entropy will be queried once each in this sequence, quering stops 1712 when enough entropy was collected without querying more sockets. 1713 [Lutz Jaenicke] 1714 1715 *) Change the Unix RAND_poll() variant to be able to poll several 1716 random devices, as specified by DEVRANDOM, until a sufficient amount 1717 of data has been collected. We spend at most 10 ms on each file 1718 (select timeout) and read in non-blocking mode. DEVRANDOM now 1719 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 1720 (previously it was just the string "/dev/urandom"), so on typical 1721 platforms the 10 ms delay will never occur. 1722 Also separate out the Unix variant to its own file, rand_unix.c. 1723 For VMS, there's a currently-empty rand_vms.c. 1724 [Richard Levitte] 1725 1726 *) Move OCSP client related routines to ocsp_cl.c. These 1727 provide utility functions which an application needing 1728 to issue a request to an OCSP responder and analyse the 1729 response will typically need: as opposed to those which an 1730 OCSP responder itself would need which will be added later. 1731 1732 OCSP_request_sign() signs an OCSP request with an API similar 1733 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 1734 response. OCSP_response_get1_basic() extracts basic response 1735 from response. OCSP_resp_find_status(): finds and extracts status 1736 information from an OCSP_CERTID structure (which will be created 1737 when the request structure is built). These are built from lower 1738 level functions which work on OCSP_SINGLERESP structures but 1739 wont normally be used unless the application wishes to examine 1740 extensions in the OCSP response for example. 1741 1742 Replace nonce routines with a pair of functions. 1743 OCSP_request_add1_nonce() adds a nonce value and optionally 1744 generates a random value. OCSP_check_nonce() checks the 1745 validity of the nonce in an OCSP response. 1746 [Steve Henson] 1747 1748 *) Change function OCSP_request_add() to OCSP_request_add0_id(). 1749 This doesn't copy the supplied OCSP_CERTID and avoids the 1750 need to free up the newly created id. Change return type 1751 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 1752 This can then be used to add extensions to the request. 1753 Deleted OCSP_request_new(), since most of its functionality 1754 is now in OCSP_REQUEST_new() (and the case insensitive name 1755 clash) apart from the ability to set the request name which 1756 will be added elsewhere. 1757 [Steve Henson] 1758 1759 *) Update OCSP API. Remove obsolete extensions argument from 1760 various functions. Extensions are now handled using the new 1761 OCSP extension code. New simple OCSP HTTP function which 1762 can be used to send requests and parse the response. 1763 [Steve Henson] 1764 1765 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 1766 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 1767 uses the special reorder version of SET OF to sort the attributes 1768 and reorder them to match the encoded order. This resolves a long 1769 standing problem: a verify on a PKCS7 structure just after signing 1770 it used to fail because the attribute order did not match the 1771 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 1772 it uses the received order. This is necessary to tolerate some broken 1773 software that does not order SET OF. This is handled by encoding 1774 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 1775 to produce the required SET OF. 1776 [Steve Henson] 1777 1778 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 1779 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 1780 files to get correct declarations of the ASN.1 item variables. 1781 [Richard Levitte] 1782 1783 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 1784 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 1785 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 1786 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 1787 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 1788 ASN1_ITEM and no wrapper functions. 1789 [Steve Henson] 1790 1791 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 1792 replace the old function pointer based I/O routines. Change most of 1793 the *_d2i_bio() and *_d2i_fp() functions to use these. 1794 [Steve Henson] 1795 1796 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 1797 lines, recognice more "algorithms" that can be deselected, and make 1798 it complain about algorithm deselection that isn't recognised. 1799 [Richard Levitte] 1800 1801 *) New ASN1 functions to handle dup, sign, verify, digest, pack and 1802 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 1803 to use new functions. Add NO_ASN1_OLD which can be set to remove 1804 some old style ASN1 functions: this can be used to determine if old 1805 code will still work when these eventually go away. 1806 [Steve Henson] 1807 1808 *) New extension functions for OCSP structures, these follow the 1809 same conventions as certificates and CRLs. 1810 [Steve Henson] 1811 1812 *) New function X509V3_add1_i2d(). This automatically encodes and 1813 adds an extension. Its behaviour can be customised with various 1814 flags to append, replace or delete. Various wrappers added for 1815 certifcates and CRLs. 1816 [Steve Henson] 1817 1818 *) Fix to avoid calling the underlying ASN1 print routine when 1819 an extension cannot be parsed. Correct a typo in the 1820 OCSP_SERVICELOC extension. Tidy up print OCSP format. 1821 [Steve Henson] 1822 1823 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 1824 entries for variables. 1825 [Steve Henson] 1826 1827 *) Add functionality to apps/openssl.c for detecting locking 1828 problems: As the program is single-threaded, all we have 1829 to do is register a locking callback using an array for 1830 storing which locks are currently held by the program. 1831 [Bodo Moeller] 1832 1833 *) Use a lock around the call to CRYPTO_get_ex_new_index() in 1834 SSL_get_ex_data_X509_STORE_idx(), which is used in 1835 ssl_verify_cert_chain() and thus can be called at any time 1836 during TLS/SSL handshakes so that thread-safety is essential. 1837 Unfortunately, the ex_data design is not at all suited 1838 for multi-threaded use, so it probably should be abolished. 1839 [Bodo Moeller] 1840 1841 *) Added Broadcom "ubsec" ENGINE to OpenSSL. 1842 [Broadcom, tweaked and integrated by Geoff Thorpe] 1843 1844 *) Move common extension printing code to new function 1845 X509V3_print_extensions(). Reorganise OCSP print routines and 1846 implement some needed OCSP ASN1 functions. Add OCSP extensions. 1847 [Steve Henson] 1848 1849 *) New function X509_signature_print() to remove duplication in some 1850 print routines. 1851 [Steve Henson] 1852 1853 *) Add a special meaning when SET OF and SEQUENCE OF flags are both 1854 set (this was treated exactly the same as SET OF previously). This 1855 is used to reorder the STACK representing the structure to match the 1856 encoding. This will be used to get round a problem where a PKCS7 1857 structure which was signed could not be verified because the STACK 1858 order did not reflect the encoded order. 1859 [Steve Henson] 1860 1861 *) Reimplement the OCSP ASN1 module using the new code. 1862 [Steve Henson] 1863 1864 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 1865 for its ASN1 operations. The old style function pointers still exist 1866 for now but they will eventually go away. 1867 [Steve Henson] 1868 1869 *) Merge in replacement ASN1 code from the ASN1 branch. This almost 1870 completely replaces the old ASN1 functionality with a table driven 1871 encoder and decoder which interprets an ASN1_ITEM structure describing 1872 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 1873 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 1874 has also been converted to the new form. 1875 [Steve Henson] 1876 1877 *) Change BN_mod_exp_recp so that negative moduli are tolerated 1878 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 1879 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 1880 for negative moduli. 1881 [Bodo Moeller] 1882 1883 *) Fix BN_uadd and BN_usub: Always return non-negative results instead 1884 of not touching the result's sign bit. 1885 [Bodo Moeller] 1886 1887 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 1888 set. 1889 [Bodo Moeller] 1890 1891 *) Changed the LHASH code to use prototypes for callbacks, and created 1892 macros to declare and implement thin (optionally static) functions 1893 that provide type-safety and avoid function pointer casting for the 1894 type-specific callbacks. 1895 [Geoff Thorpe] 1896 1897 *) Added Kerberos Cipher Suites to be used with TLS, as written in 1898 RFC 2712. 1899 [Veers Staats <staatsvr@asc.hpc.mil>, 1900 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 1901 1902 *) Reformat the FAQ so the different questions and answers can be divided 1903 in sections depending on the subject. 1904 [Richard Levitte] 1905 1906 *) Have the zlib compression code load ZLIB.DLL dynamically under 1907 Windows. 1908 [Richard Levitte] 1909 1910 *) New function BN_mod_sqrt for computing square roots modulo a prime 1911 (using the probabilistic Tonelli-Shanks algorithm unless 1912 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 1913 be handled deterministically). 1914 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1915 1916 *) Make BN_mod_inverse faster by explicitly handling small quotients 1917 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 1918 512 bits], about 30% for larger ones [1024 or 2048 bits].) 1919 [Bodo Moeller] 1920 1921 *) New function BN_kronecker. 1922 [Bodo Moeller] 1923 1924 *) Fix BN_gcd so that it works on negative inputs; the result is 1925 positive unless both parameters are zero. 1926 Previously something reasonably close to an infinite loop was 1927 possible because numbers could be growing instead of shrinking 1928 in the implementation of Euclid's algorithm. 1929 [Bodo Moeller] 1930 1931 *) Fix BN_is_word() and BN_is_one() macros to take into account the 1932 sign of the number in question. 1933 1934 Fix BN_is_word(a,w) to work correctly for w == 0. 1935 1936 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 1937 because its test if the absolute value of 'a' equals 'w'. 1938 Note that BN_abs_is_word does *not* handle w == 0 reliably; 1939 it exists mostly for use in the implementations of BN_is_zero(), 1940 BN_is_one(), and BN_is_word(). 1941 [Bodo Moeller] 1942 1943 *) New function BN_swap. 1944 [Bodo Moeller] 1945 1946 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 1947 the exponentiation functions are more likely to produce reasonable 1948 results on negative inputs. 1949 [Bodo Moeller] 1950 1951 *) Change BN_mod_mul so that the result is always non-negative. 1952 Previously, it could be negative if one of the factors was negative; 1953 I don't think anyone really wanted that behaviour. 1954 [Bodo Moeller] 1955 1956 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 1957 (except for exponentiation, which stays in crypto/bn/bn_exp.c, 1958 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 1959 and add new functions: 1960 1961 BN_nnmod 1962 BN_mod_sqr 1963 BN_mod_add 1964 BN_mod_add_quick 1965 BN_mod_sub 1966 BN_mod_sub_quick 1967 BN_mod_lshift1 1968 BN_mod_lshift1_quick 1969 BN_mod_lshift 1970 BN_mod_lshift_quick 1971 1972 These functions always generate non-negative results. 1973 1974 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 1975 such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 1976 1977 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 1978 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 1979 be reduced modulo m. 1980 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1981 1982#if 0 1983 The following entry accidentily appeared in the CHANGES file 1984 distributed with OpenSSL 0.9.7. The modifications described in 1985 it do *not* apply to OpenSSL 0.9.7. 1986 1987 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1988 was actually never needed) and in BN_mul(). The removal in BN_mul() 1989 required a small change in bn_mul_part_recursive() and the addition 1990 of the functions bn_cmp_part_words(), bn_sub_part_words() and 1991 bn_add_part_words(), which do the same thing as bn_cmp_words(), 1992 bn_sub_words() and bn_add_words() except they take arrays with 1993 differing sizes. 1994 [Richard Levitte] 1995#endif 1996 1997 *) In 'openssl passwd', verify passwords read from the terminal 1998 unless the '-salt' option is used (which usually means that 1999 verification would just waste user's time since the resulting 2000 hash is going to be compared with some given password hash) 2001 or the new '-noverify' option is used. 2002 2003 This is an incompatible change, but it does not affect 2004 non-interactive use of 'openssl passwd' (passwords on the command 2005 line, '-stdin' option, '-in ...' option) and thus should not 2006 cause any problems. 2007 [Bodo Moeller] 2008 2009 *) Remove all references to RSAref, since there's no more need for it. 2010 [Richard Levitte] 2011 2012 *) Make DSO load along a path given through an environment variable 2013 (SHLIB_PATH) with shl_load(). 2014 [Richard Levitte] 2015 2016 *) Constify the ENGINE code as a result of BIGNUM constification. 2017 Also constify the RSA code and most things related to it. In a 2018 few places, most notable in the depth of the ASN.1 code, ugly 2019 casts back to non-const were required (to be solved at a later 2020 time) 2021 [Richard Levitte] 2022 2023 *) Make it so the openssl application has all engines loaded by default. 2024 [Richard Levitte] 2025 2026 *) Constify the BIGNUM routines a little more. 2027 [Richard Levitte] 2028 2029 *) Add the following functions: 2030 2031 ENGINE_load_cswift() 2032 ENGINE_load_chil() 2033 ENGINE_load_atalla() 2034 ENGINE_load_nuron() 2035 ENGINE_load_builtin_engines() 2036 2037 That way, an application can itself choose if external engines that 2038 are built-in in OpenSSL shall ever be used or not. The benefit is 2039 that applications won't have to be linked with libdl or other dso 2040 libraries unless it's really needed. 2041 2042 Changed 'openssl engine' to load all engines on demand. 2043 Changed the engine header files to avoid the duplication of some 2044 declarations (they differed!). 2045 [Richard Levitte] 2046 2047 *) 'openssl engine' can now list capabilities. 2048 [Richard Levitte] 2049 2050 *) Better error reporting in 'openssl engine'. 2051 [Richard Levitte] 2052 2053 *) Never call load_dh_param(NULL) in s_server. 2054 [Bodo Moeller] 2055 2056 *) Add engine application. It can currently list engines by name and 2057 identity, and test if they are actually available. 2058 [Richard Levitte] 2059 2060 *) Improve RPM specification file by forcing symbolic linking and making 2061 sure the installed documentation is also owned by root.root. 2062 [Damien Miller <djm@mindrot.org>] 2063 2064 *) Give the OpenSSL applications more possibilities to make use of 2065 keys (public as well as private) handled by engines. 2066 [Richard Levitte] 2067 2068 *) Add OCSP code that comes from CertCo. 2069 [Richard Levitte] 2070 2071 *) Add VMS support for the Rijndael code. 2072 [Richard Levitte] 2073 2074 *) Added untested support for Nuron crypto accelerator. 2075 [Ben Laurie] 2076 2077 *) Add support for external cryptographic devices. This code was 2078 previously distributed separately as the "engine" branch. 2079 [Geoff Thorpe, Richard Levitte] 2080 2081 *) Rework the filename-translation in the DSO code. It is now possible to 2082 have far greater control over how a "name" is turned into a filename 2083 depending on the operating environment and any oddities about the 2084 different shared library filenames on each system. 2085 [Geoff Thorpe] 2086 2087 *) Support threads on FreeBSD-elf in Configure. 2088 [Richard Levitte] 2089 2090 *) Fix for SHA1 assembly problem with MASM: it produces 2091 warnings about corrupt line number information when assembling 2092 with debugging information. This is caused by the overlapping 2093 of two sections. 2094 [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 2095 2096 *) NCONF changes. 2097 NCONF_get_number() has no error checking at all. As a replacement, 2098 NCONF_get_number_e() is defined (_e for "error checking") and is 2099 promoted strongly. The old NCONF_get_number is kept around for 2100 binary backward compatibility. 2101 Make it possible for methods to load from something other than a BIO, 2102 by providing a function pointer that is given a name instead of a BIO. 2103 For example, this could be used to load configuration data from an 2104 LDAP server. 2105 [Richard Levitte] 2106 2107 *) Fix for non blocking accept BIOs. Added new I/O special reason 2108 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 2109 with non blocking I/O was not possible because no retry code was 2110 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 2111 this case. 2112 [Steve Henson] 2113 2114 *) Added the beginnings of Rijndael support. 2115 [Ben Laurie] 2116 2117 *) Fix for bug in DirectoryString mask setting. Add support for 2118 X509_NAME_print_ex() in 'req' and X509_print_ex() function 2119 to allow certificate printing to more controllable, additional 2120 'certopt' option to 'x509' to allow new printing options to be 2121 set. 2122 [Steve Henson] 2123 2124 *) Clean old EAY MD5 hack from e_os.h. 2125 [Richard Levitte] 2126 2127 Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 2128 2129 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 2130 by using the Codenomicon TLS Test Tool (CAN-2004-0079) 2131 [Joe Orton, Steve Henson] 2132 2133 Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 2134 2135 *) Fix additional bug revealed by the NISCC test suite: 2136 2137 Stop bug triggering large recursion when presented with 2138 certain ASN.1 tags (CAN-2003-0851) 2139 [Steve Henson] 2140 2141 Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 2142 2143 *) Fix various bugs revealed by running the NISCC test suite: 2144 2145 Stop out of bounds reads in the ASN1 code when presented with 2146 invalid tags (CAN-2003-0543 and CAN-2003-0544). 2147 2148 If verify callback ignores invalid public key errors don't try to check 2149 certificate signature with the NULL public key. 2150 2151 [Steve Henson] 2152 2153 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 2154 if the server requested one: as stated in TLS 1.0 and SSL 3.0 2155 specifications. 2156 [Steve Henson] 2157 2158 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 2159 extra data after the compression methods not only for TLS 1.0 2160 but also for SSL 3.0 (as required by the specification). 2161 [Bodo Moeller; problem pointed out by Matthias Loepfe] 2162 2163 *) Change X509_certificate_type() to mark the key as exported/exportable 2164 when it's 512 *bits* long, not 512 bytes. 2165 [Richard Levitte] 2166 2167 Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 2168 2169 *) Countermeasure against the Klima-Pokorny-Rosa extension of 2170 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 2171 a protocol version number mismatch like a decryption error 2172 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 2173 [Bodo Moeller] 2174 2175 *) Turn on RSA blinding by default in the default implementation 2176 to avoid a timing attack. Applications that don't want it can call 2177 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 2178 They would be ill-advised to do so in most cases. 2179 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 2180 2181 *) Change RSA blinding code so that it works when the PRNG is not 2182 seeded (in this case, the secret RSA exponent is abused as 2183 an unpredictable seed -- if it is not unpredictable, there 2184 is no point in blinding anyway). Make RSA blinding thread-safe 2185 by remembering the creator's thread ID in rsa->blinding and 2186 having all other threads use local one-time blinding factors 2187 (this requires more computation than sharing rsa->blinding, but 2188 avoids excessive locking; and if an RSA object is not shared 2189 between threads, blinding will still be very fast). 2190 [Bodo Moeller] 2191 2192 Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 2193 2194 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 2195 via timing by performing a MAC computation even if incorrrect 2196 block cipher padding has been found. This is a countermeasure 2197 against active attacks where the attacker has to distinguish 2198 between bad padding and a MAC verification error. (CAN-2003-0078) 2199 2200 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 2201 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 2202 Martin Vuagnoux (EPFL, Ilion)] 2203 2204 Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 2205 2206 *) New function OPENSSL_cleanse(), which is used to cleanse a section of 2207 memory from it's contents. This is done with a counter that will 2208 place alternating values in each byte. This can be used to solve 2209 two issues: 1) the removal of calls to memset() by highly optimizing 2210 compilers, and 2) cleansing with other values than 0, since those can 2211 be read through on certain media, for example a swap space on disk. 2212 [Geoff Thorpe] 2213 2214 *) Bugfix: client side session caching did not work with external caching, 2215 because the session->cipher setting was not restored when reloading 2216 from the external cache. This problem was masked, when 2217 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 2218 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 2219 [Lutz Jaenicke] 2220 2221 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 2222 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 2223 [Zeev Lieber <zeev-l@yahoo.com>] 2224 2225 *) Undo an undocumented change introduced in 0.9.6e which caused 2226 repeated calls to OpenSSL_add_all_ciphers() and 2227 OpenSSL_add_all_digests() to be ignored, even after calling 2228 EVP_cleanup(). 2229 [Richard Levitte] 2230 2231 *) Change the default configuration reader to deal with last line not 2232 being properly terminated. 2233 [Richard Levitte] 2234 2235 *) Change X509_NAME_cmp() so it applies the special rules on handling 2236 DN values that are of type PrintableString, as well as RDNs of type 2237 emailAddress where the value has the type ia5String. 2238 [stefank@valicert.com via Richard Levitte] 2239 2240 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 2241 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 2242 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 2243 the bitwise-OR of the two for use by the majority of applications 2244 wanting this behaviour, and update the docs. The documented 2245 behaviour and actual behaviour were inconsistent and had been 2246 changing anyway, so this is more a bug-fix than a behavioural 2247 change. 2248 [Geoff Thorpe, diagnosed by Nadav Har'El] 2249 2250 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 2251 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 2252 [Bodo Moeller] 2253 2254 *) Fix initialization code race conditions in 2255 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 2256 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 2257 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 2258 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 2259 ssl2_get_cipher_by_char(), 2260 ssl3_get_cipher_by_char(). 2261 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 2262 2263 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 2264 the cached sessions are flushed, as the remove_cb() might use ex_data 2265 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 2266 (see [openssl.org #212]). 2267 [Geoff Thorpe, Lutz Jaenicke] 2268 2269 *) Fix typo in OBJ_txt2obj which incorrectly passed the content 2270 length, instead of the encoding length to d2i_ASN1_OBJECT. 2271 [Steve Henson] 2272 2273 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 2274 2275 *) [In 0.9.6g-engine release:] 2276 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 2277 [Lynn Gazis <lgazis@rainbow.com>] 2278 2279 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 2280 2281 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 2282 and get fix the header length calculation. 2283 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 2284 Alon Kantor <alonk@checkpoint.com> (and others), 2285 Steve Henson] 2286 2287 *) Use proper error handling instead of 'assertions' in buffer 2288 overflow checks added in 0.9.6e. This prevents DoS (the 2289 assertions could call abort()). 2290 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 2291 2292 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 2293 2294 *) Add various sanity checks to asn1_get_length() to reject 2295 the ASN1 length bytes if they exceed sizeof(long), will appear 2296 negative or the content length exceeds the length of the 2297 supplied buffer. 2298 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2299 2300 *) Fix cipher selection routines: ciphers without encryption had no flags 2301 for the cipher strength set and where therefore not handled correctly 2302 by the selection routines (PR #130). 2303 [Lutz Jaenicke] 2304 2305 *) Fix EVP_dsa_sha macro. 2306 [Nils Larsch] 2307 2308 *) New option 2309 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 2310 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 2311 that was added in OpenSSL 0.9.6d. 2312 2313 As the countermeasure turned out to be incompatible with some 2314 broken SSL implementations, the new option is part of SSL_OP_ALL. 2315 SSL_OP_ALL is usually employed when compatibility with weird SSL 2316 implementations is desired (e.g. '-bugs' option to 's_client' and 2317 's_server'), so the new option is automatically set in many 2318 applications. 2319 [Bodo Moeller] 2320 2321 *) Changes in security patch: 2322 2323 Changes marked "(CHATS)" were sponsored by the Defense Advanced 2324 Research Projects Agency (DARPA) and Air Force Research Laboratory, 2325 Air Force Materiel Command, USAF, under agreement number 2326 F30602-01-2-0537. 2327 2328 *) Add various sanity checks to asn1_get_length() to reject 2329 the ASN1 length bytes if they exceed sizeof(long), will appear 2330 negative or the content length exceeds the length of the 2331 supplied buffer. (CAN-2002-0659) 2332 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2333 2334 *) Assertions for various potential buffer overflows, not known to 2335 happen in practice. 2336 [Ben Laurie (CHATS)] 2337 2338 *) Various temporary buffers to hold ASCII versions of integers were 2339 too small for 64 bit platforms. (CAN-2002-0655) 2340 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 2341 2342 *) Remote buffer overflow in SSL3 protocol - an attacker could 2343 supply an oversized session ID to a client. (CAN-2002-0656) 2344 [Ben Laurie (CHATS)] 2345 2346 *) Remote buffer overflow in SSL2 protocol - an attacker could 2347 supply an oversized client master key. (CAN-2002-0656) 2348 [Ben Laurie (CHATS)] 2349 2350 Changes between 0.9.6c and 0.9.6d [9 May 2002] 2351 2352 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 2353 encoded as NULL) with id-dsa-with-sha1. 2354 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 2355 2356 *) Check various X509_...() return values in apps/req.c. 2357 [Nils Larsch <nla@trustcenter.de>] 2358 2359 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 2360 an end-of-file condition would erronously be flagged, when the CRLF 2361 was just at the end of a processed block. The bug was discovered when 2362 processing data through a buffering memory BIO handing the data to a 2363 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 2364 <ptsekov@syntrex.com> and Nedelcho Stanev. 2365 [Lutz Jaenicke] 2366 2367 *) Implement a countermeasure against a vulnerability recently found 2368 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 2369 before application data chunks to avoid the use of known IVs 2370 with data potentially chosen by the attacker. 2371 [Bodo Moeller] 2372 2373 *) Fix length checks in ssl3_get_client_hello(). 2374 [Bodo Moeller] 2375 2376 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 2377 to prevent ssl3_read_internal() from incorrectly assuming that 2378 ssl3_read_bytes() found application data while handshake 2379 processing was enabled when in fact s->s3->in_read_app_data was 2380 merely automatically cleared during the initial handshake. 2381 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 2382 2383 *) Fix object definitions for Private and Enterprise: they were not 2384 recognized in their shortname (=lowercase) representation. Extend 2385 obj_dat.pl to issue an error when using undefined keywords instead 2386 of silently ignoring the problem (Svenning Sorensen 2387 <sss@sss.dnsalias.net>). 2388 [Lutz Jaenicke] 2389 2390 *) Fix DH_generate_parameters() so that it works for 'non-standard' 2391 generators, i.e. generators other than 2 and 5. (Previously, the 2392 code did not properly initialise the 'add' and 'rem' values to 2393 BN_generate_prime().) 2394 2395 In the new general case, we do not insist that 'generator' is 2396 actually a primitive root: This requirement is rather pointless; 2397 a generator of the order-q subgroup is just as good, if not 2398 better. 2399 [Bodo Moeller] 2400 2401 *) Map new X509 verification errors to alerts. Discovered and submitted by 2402 Tom Wu <tom@arcot.com>. 2403 [Lutz Jaenicke] 2404 2405 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 2406 returning non-zero before the data has been completely received 2407 when using non-blocking I/O. 2408 [Bodo Moeller; problem pointed out by John Hughes] 2409 2410 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 2411 [Ben Laurie, Lutz Jaenicke] 2412 2413 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 2414 Yoram Zahavi <YoramZ@gilian.com>). 2415 [Lutz Jaenicke] 2416 2417 *) Add information about CygWin 1.3 and on, and preserve proper 2418 configuration for the versions before that. 2419 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2420 2421 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 2422 check whether we deal with a copy of a session and do not delete from 2423 the cache in this case. Problem reported by "Izhar Shoshani Levi" 2424 <izhar@checkpoint.com>. 2425 [Lutz Jaenicke] 2426 2427 *) Do not store session data into the internal session cache, if it 2428 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 2429 flag is set). Proposed by Aslam <aslam@funk.com>. 2430 [Lutz Jaenicke] 2431 2432 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 2433 value is 0. 2434 [Richard Levitte] 2435 2436 *) [In 0.9.6d-engine release:] 2437 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2438 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2439 2440 *) Add the configuration target linux-s390x. 2441 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 2442 2443 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 2444 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 2445 variable as an indication that a ClientHello message has been 2446 received. As the flag value will be lost between multiple 2447 invocations of ssl3_accept when using non-blocking I/O, the 2448 function may not be aware that a handshake has actually taken 2449 place, thus preventing a new session from being added to the 2450 session cache. 2451 2452 To avoid this problem, we now set s->new_session to 2 instead of 2453 using a local variable. 2454 [Lutz Jaenicke, Bodo Moeller] 2455 2456 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 2457 if the SSL_R_LENGTH_MISMATCH error is detected. 2458 [Geoff Thorpe, Bodo Moeller] 2459 2460 *) New 'shared_ldflag' column in Configure platform table. 2461 [Richard Levitte] 2462 2463 *) Fix EVP_CIPHER_mode macro. 2464 ["Dan S. Camper" <dan@bti.net>] 2465 2466 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 2467 type, we must throw them away by setting rr->length to 0. 2468 [D P Chang <dpc@qualys.com>] 2469 2470 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 2471 2472 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 2473 <Dominikus.Scherkl@biodata.com>. (The previous implementation 2474 worked incorrectly for those cases where range = 10..._2 and 2475 3*range is two bits longer than range.) 2476 [Bodo Moeller] 2477 2478 *) Only add signing time to PKCS7 structures if it is not already 2479 present. 2480 [Steve Henson] 2481 2482 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 2483 OBJ_ld_ce should be OBJ_id_ce. 2484 Also some ip-pda OIDs in crypto/objects/objects.txt were 2485 incorrect (cf. RFC 3039). 2486 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 2487 2488 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 2489 returns early because it has nothing to do. 2490 [Andy Schneider <andy.schneider@bjss.co.uk>] 2491 2492 *) [In 0.9.6c-engine release:] 2493 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 2494 [Andy Schneider <andy.schneider@bjss.co.uk>] 2495 2496 *) [In 0.9.6c-engine release:] 2497 Add support for Cryptographic Appliance's keyserver technology. 2498 (Use engine 'keyclient') 2499 [Cryptographic Appliances and Geoff Thorpe] 2500 2501 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 2502 is called via tools/c89.sh because arguments have to be 2503 rearranged (all '-L' options must appear before the first object 2504 modules). 2505 [Richard Shapiro <rshapiro@abinitio.com>] 2506 2507 *) [In 0.9.6c-engine release:] 2508 Add support for Broadcom crypto accelerator cards, backported 2509 from 0.9.7. 2510 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 2511 2512 *) [In 0.9.6c-engine release:] 2513 Add support for SureWare crypto accelerator cards from 2514 Baltimore Technologies. (Use engine 'sureware') 2515 [Baltimore Technologies and Mark Cox] 2516 2517 *) [In 0.9.6c-engine release:] 2518 Add support for crypto accelerator cards from Accelerated 2519 Encryption Processing, www.aep.ie. (Use engine 'aep') 2520 [AEP Inc. and Mark Cox] 2521 2522 *) Add a configuration entry for gcc on UnixWare. 2523 [Gary Benson <gbenson@redhat.com>] 2524 2525 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 2526 messages are stored in a single piece (fixed-length part and 2527 variable-length part combined) and fix various bugs found on the way. 2528 [Bodo Moeller] 2529 2530 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 2531 instead. BIO_gethostbyname() does not know what timeouts are 2532 appropriate, so entries would stay in cache even when they have 2533 become invalid. 2534 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 2535 2536 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 2537 faced with a pathologically small ClientHello fragment that does 2538 not contain client_version: Instead of aborting with an error, 2539 simply choose the highest available protocol version (i.e., 2540 TLS 1.0 unless it is disabled). In practice, ClientHello 2541 messages are never sent like this, but this change gives us 2542 strictly correct behaviour at least for TLS. 2543 [Bodo Moeller] 2544 2545 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 2546 never resets s->method to s->ctx->method when called from within 2547 one of the SSL handshake functions. 2548 [Bodo Moeller; problem pointed out by Niko Baric] 2549 2550 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 2551 (sent using the client's version number) if client_version is 2552 smaller than the protocol version in use. Also change 2553 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 2554 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 2555 the client will at least see that alert. 2556 [Bodo Moeller] 2557 2558 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 2559 correctly. 2560 [Bodo Moeller] 2561 2562 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 2563 client receives HelloRequest while in a handshake. 2564 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 2565 2566 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 2567 should end in 'break', not 'goto end' which circuments various 2568 cleanups done in state SSL_ST_OK. But session related stuff 2569 must be disabled for SSL_ST_OK in the case that we just sent a 2570 HelloRequest. 2571 2572 Also avoid some overhead by not calling ssl_init_wbio_buffer() 2573 before just sending a HelloRequest. 2574 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 2575 2576 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 2577 reveal whether illegal block cipher padding was found or a MAC 2578 verification error occured. (Neither SSLerr() codes nor alerts 2579 are directly visible to potential attackers, but the information 2580 may leak via logfiles.) 2581 2582 Similar changes are not required for the SSL 2.0 implementation 2583 because the number of padding bytes is sent in clear for SSL 2.0, 2584 and the extra bytes are just ignored. However ssl/s2_pkt.c 2585 failed to verify that the purported number of padding bytes is in 2586 the legal range. 2587 [Bodo Moeller] 2588 2589 *) Add OpenUNIX-8 support including shared libraries 2590 (Boyd Lynn Gerber <gerberb@zenez.com>). 2591 [Lutz Jaenicke] 2592 2593 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 2594 'wristwatch attack' using huge encoding parameters (cf. 2595 James H. Manger's CRYPTO 2001 paper). Note that the 2596 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 2597 encoding parameters and hence was not vulnerable. 2598 [Bodo Moeller] 2599 2600 *) BN_sqr() bug fix. 2601 [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 2602 2603 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 2604 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 2605 followed by modular reduction. 2606 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 2607 2608 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 2609 equivalent based on BN_pseudo_rand() instead of BN_rand(). 2610 [Bodo Moeller] 2611 2612 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 2613 This function was broken, as the check for a new client hello message 2614 to handle SGC did not allow these large messages. 2615 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 2616 [Lutz Jaenicke] 2617 2618 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 2619 [Lutz Jaenicke] 2620 2621 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 2622 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 2623 [Lutz Jaenicke] 2624 2625 *) Rework the configuration and shared library support for Tru64 Unix. 2626 The configuration part makes use of modern compiler features and 2627 still retains old compiler behavior for those that run older versions 2628 of the OS. The shared library support part includes a variant that 2629 uses the RPATH feature, and is available through the special 2630 configuration target "alpha-cc-rpath", which will never be selected 2631 automatically. 2632 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 2633 2634 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 2635 with the same message size as in ssl3_get_certificate_request(). 2636 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 2637 messages might inadvertently be reject as too long. 2638 [Petr Lampa <lampa@fee.vutbr.cz>] 2639 2640 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 2641 [Andy Polyakov] 2642 2643 *) Modified SSL library such that the verify_callback that has been set 2644 specificly for an SSL object with SSL_set_verify() is actually being 2645 used. Before the change, a verify_callback set with this function was 2646 ignored and the verify_callback() set in the SSL_CTX at the time of 2647 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 2648 to allow the necessary settings. 2649 [Lutz Jaenicke] 2650 2651 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 2652 explicitly to NULL, as at least on Solaris 8 this seems not always to be 2653 done automatically (in contradiction to the requirements of the C 2654 standard). This made problems when used from OpenSSH. 2655 [Lutz Jaenicke] 2656 2657 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 2658 dh->length and always used 2659 2660 BN_rand_range(priv_key, dh->p). 2661 2662 BN_rand_range() is not necessary for Diffie-Hellman, and this 2663 specific range makes Diffie-Hellman unnecessarily inefficient if 2664 dh->length (recommended exponent length) is much smaller than the 2665 length of dh->p. We could use BN_rand_range() if the order of 2666 the subgroup was stored in the DH structure, but we only have 2667 dh->length. 2668 2669 So switch back to 2670 2671 BN_rand(priv_key, l, ...) 2672 2673 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 2674 otherwise. 2675 [Bodo Moeller] 2676 2677 *) In 2678 2679 RSA_eay_public_encrypt 2680 RSA_eay_private_decrypt 2681 RSA_eay_private_encrypt (signing) 2682 RSA_eay_public_decrypt (signature verification) 2683 2684 (default implementations for RSA_public_encrypt, 2685 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 2686 always reject numbers >= n. 2687 [Bodo Moeller] 2688 2689 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 2690 to synchronize access to 'locking_thread'. This is necessary on 2691 systems where access to 'locking_thread' (an 'unsigned long' 2692 variable) is not atomic. 2693 [Bodo Moeller] 2694 2695 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 2696 *before* setting the 'crypto_lock_rand' flag. The previous code had 2697 a race condition if 0 is a valid thread ID. 2698 [Travis Vitek <vitek@roguewave.com>] 2699 2700 *) Add support for shared libraries under Irix. 2701 [Albert Chin-A-Young <china@thewrittenword.com>] 2702 2703 *) Add configuration option to build on Linux on both big-endian and 2704 little-endian MIPS. 2705 [Ralf Baechle <ralf@uni-koblenz.de>] 2706 2707 *) Add the possibility to create shared libraries on HP-UX. 2708 [Richard Levitte] 2709 2710 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 2711 2712 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 2713 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 2714 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 2715 PRNG state recovery was possible based on the output of 2716 one PRNG request appropriately sized to gain knowledge on 2717 'md' followed by enough consecutive 1-byte PRNG requests 2718 to traverse all of 'state'. 2719 2720 1. When updating 'md_local' (the current thread's copy of 'md') 2721 during PRNG output generation, hash all of the previous 2722 'md_local' value, not just the half used for PRNG output. 2723 2724 2. Make the number of bytes from 'state' included into the hash 2725 independent from the number of PRNG bytes requested. 2726 2727 The first measure alone would be sufficient to avoid 2728 Markku-Juhani's attack. (Actually it had never occurred 2729 to me that the half of 'md_local' used for chaining was the 2730 half from which PRNG output bytes were taken -- I had always 2731 assumed that the secret half would be used.) The second 2732 measure makes sure that additional data from 'state' is never 2733 mixed into 'md_local' in small portions; this heuristically 2734 further strengthens the PRNG. 2735 [Bodo Moeller] 2736 2737 *) Fix crypto/bn/asm/mips3.s. 2738 [Andy Polyakov] 2739 2740 *) When only the key is given to "enc", the IV is undefined. Print out 2741 an error message in this case. 2742 [Lutz Jaenicke] 2743 2744 *) Handle special case when X509_NAME is empty in X509 printing routines. 2745 [Steve Henson] 2746 2747 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 2748 positive and less than q. 2749 [Bodo Moeller] 2750 2751 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 2752 used: it isn't thread safe and the add_lock_callback should handle 2753 that itself. 2754 [Paul Rose <Paul.Rose@bridge.com>] 2755 2756 *) Verify that incoming data obeys the block size in 2757 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 2758 [Bodo Moeller] 2759 2760 *) Fix OAEP check. 2761 [Ulf M�ller, Bodo M�ller] 2762 2763 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 2764 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 2765 when fixing the server behaviour for backwards-compatible 'client 2766 hello' messages. (Note that the attack is impractical against 2767 SSL 3.0 and TLS 1.0 anyway because length and version checking 2768 means that the probability of guessing a valid ciphertext is 2769 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 2770 paper.) 2771 2772 Before 0.9.5, the countermeasure (hide the error by generating a 2773 random 'decryption result') did not work properly because 2774 ERR_clear_error() was missing, meaning that SSL_get_error() would 2775 detect the supposedly ignored error. 2776 2777 Both problems are now fixed. 2778 [Bodo Moeller] 2779 2780 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 2781 (previously it was 1024). 2782 [Bodo Moeller] 2783 2784 *) Fix for compatibility mode trust settings: ignore trust settings 2785 unless some valid trust or reject settings are present. 2786 [Steve Henson] 2787 2788 *) Fix for blowfish EVP: its a variable length cipher. 2789 [Steve Henson] 2790 2791 *) Fix various bugs related to DSA S/MIME verification. Handle missing 2792 parameters in DSA public key structures and return an error in the 2793 DSA routines if parameters are absent. 2794 [Steve Henson] 2795 2796 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 2797 in the current directory if neither $RANDFILE nor $HOME was set. 2798 RAND_file_name() in 0.9.6a returned NULL in this case. This has 2799 caused some confusion to Windows users who haven't defined $HOME. 2800 Thus RAND_file_name() is changed again: e_os.h can define a 2801 DEFAULT_HOME, which will be used if $HOME is not set. 2802 For Windows, we use "C:"; on other platforms, we still require 2803 environment variables. 2804 2805 *) Move 'if (!initialized) RAND_poll()' into regions protected by 2806 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 2807 having multiple threads call RAND_poll() concurrently. 2808 [Bodo Moeller] 2809 2810 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 2811 combination of a flag and a thread ID variable. 2812 Otherwise while one thread is in ssleay_rand_bytes (which sets the 2813 flag), *other* threads can enter ssleay_add_bytes without obeying 2814 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 2815 that they do not hold after the first thread unsets add_do_not_lock). 2816 [Bodo Moeller] 2817 2818 *) Change bctest again: '-x' expressions are not available in all 2819 versions of 'test'. 2820 [Bodo Moeller] 2821 2822 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 2823 2824 *) Fix a couple of memory leaks in PKCS7_dataDecode() 2825 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 2826 2827 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 2828 the default extension for executables, if any. Also, make the perl 2829 scripts that use symlink() to test if it really exists and use "cp" 2830 if it doesn't. All this made OpenSSL compilable and installable in 2831 CygWin. 2832 [Richard Levitte] 2833 2834 *) Fix for asn1_GetSequence() for indefinite length constructed data. 2835 If SEQUENCE is length is indefinite just set c->slen to the total 2836 amount of data available. 2837 [Steve Henson, reported by shige@FreeBSD.org] 2838 [This change does not apply to 0.9.7.] 2839 2840 *) Change bctest to avoid here-documents inside command substitution 2841 (workaround for FreeBSD /bin/sh bug). 2842 For compatibility with Ultrix, avoid shell functions (introduced 2843 in the bctest version that searches along $PATH). 2844 [Bodo Moeller] 2845 2846 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 2847 with des_encrypt() defined on some operating systems, like Solaris 2848 and UnixWare. 2849 [Richard Levitte] 2850 2851 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 2852 On the Importance of Eliminating Errors in Cryptographic 2853 Computations, J. Cryptology 14 (2001) 2, 101-119, 2854 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 2855 [Ulf Moeller] 2856 2857 *) MIPS assembler BIGNUM division bug fix. 2858 [Andy Polyakov] 2859 2860 *) Disabled incorrect Alpha assembler code. 2861 [Richard Levitte] 2862 2863 *) Fix PKCS#7 decode routines so they correctly update the length 2864 after reading an EOC for the EXPLICIT tag. 2865 [Steve Henson] 2866 [This change does not apply to 0.9.7.] 2867 2868 *) Fix bug in PKCS#12 key generation routines. This was triggered 2869 if a 3DES key was generated with a 0 initial byte. Include 2870 PKCS12_BROKEN_KEYGEN compilation option to retain the old 2871 (but broken) behaviour. 2872 [Steve Henson] 2873 2874 *) Enhance bctest to search for a working bc along $PATH and print 2875 it when found. 2876 [Tim Rice <tim@multitalents.net> via Richard Levitte] 2877 2878 *) Fix memory leaks in err.c: free err_data string if necessary; 2879 don't write to the wrong index in ERR_set_error_data. 2880 [Bodo Moeller] 2881 2882 *) Implement ssl23_peek (analogous to ssl23_read), which previously 2883 did not exist. 2884 [Bodo Moeller] 2885 2886 *) Replace rdtsc with _emit statements for VC++ version 5. 2887 [Jeremy Cooper <jeremy@baymoo.org>] 2888 2889 *) Make it possible to reuse SSLv2 sessions. 2890 [Richard Levitte] 2891 2892 *) In copy_email() check for >= 0 as a return value for 2893 X509_NAME_get_index_by_NID() since 0 is a valid index. 2894 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 2895 2896 *) Avoid coredump with unsupported or invalid public keys by checking if 2897 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 2898 PKCS7_verify() fails with non detached data. 2899 [Steve Henson] 2900 2901 *) Don't use getenv in library functions when run as setuid/setgid. 2902 New function OPENSSL_issetugid(). 2903 [Ulf Moeller] 2904 2905 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 2906 due to incorrect handling of multi-threading: 2907 2908 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 2909 2910 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 2911 2912 3. Count how many times MemCheck_off() has been called so that 2913 nested use can be treated correctly. This also avoids 2914 inband-signalling in the previous code (which relied on the 2915 assumption that thread ID 0 is impossible). 2916 [Bodo Moeller] 2917 2918 *) Add "-rand" option also to s_client and s_server. 2919 [Lutz Jaenicke] 2920 2921 *) Fix CPU detection on Irix 6.x. 2922 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 2923 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 2924 2925 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 2926 was empty. 2927 [Steve Henson] 2928 [This change does not apply to 0.9.7.] 2929 2930 *) Use the cached encoding of an X509_NAME structure rather than 2931 copying it. This is apparently the reason for the libsafe "errors" 2932 but the code is actually correct. 2933 [Steve Henson] 2934 2935 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 2936 Bleichenbacher's DSA attack. 2937 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 2938 to be set and top=0 forces the highest bit to be set; top=-1 is new 2939 and leaves the highest bit random. 2940 [Ulf Moeller, Bodo Moeller] 2941 2942 *) In the NCONF_...-based implementations for CONF_... queries 2943 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 2944 a temporary CONF structure with the data component set to NULL 2945 (which gives segmentation faults in lh_retrieve). 2946 Instead, use NULL for the CONF pointer in CONF_get_string and 2947 CONF_get_number (which may use environment variables) and directly 2948 return NULL from CONF_get_section. 2949 [Bodo Moeller] 2950 2951 *) Fix potential buffer overrun for EBCDIC. 2952 [Ulf Moeller] 2953 2954 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 2955 keyUsage if basicConstraints absent for a CA. 2956 [Steve Henson] 2957 2958 *) Make SMIME_write_PKCS7() write mail header values with a format that 2959 is more generally accepted (no spaces before the semicolon), since 2960 some programs can't parse those values properly otherwise. Also make 2961 sure BIO's that break lines after each write do not create invalid 2962 headers. 2963 [Richard Levitte] 2964 2965 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 2966 macros previously used would not encode an empty SEQUENCE OF 2967 and break the signature. 2968 [Steve Henson] 2969 [This change does not apply to 0.9.7.] 2970 2971 *) Zero the premaster secret after deriving the master secret in 2972 DH ciphersuites. 2973 [Steve Henson] 2974 2975 *) Add some EVP_add_digest_alias registrations (as found in 2976 OpenSSL_add_all_digests()) to SSL_library_init() 2977 aka OpenSSL_add_ssl_algorithms(). This provides improved 2978 compatibility with peers using X.509 certificates 2979 with unconventional AlgorithmIdentifier OIDs. 2980 [Bodo Moeller] 2981 2982 *) Fix for Irix with NO_ASM. 2983 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 2984 2985 *) ./config script fixes. 2986 [Ulf Moeller, Richard Levitte] 2987 2988 *) Fix 'openssl passwd -1'. 2989 [Bodo Moeller] 2990 2991 *) Change PKCS12_key_gen_asc() so it can cope with non null 2992 terminated strings whose length is passed in the passlen 2993 parameter, for example from PEM callbacks. This was done 2994 by adding an extra length parameter to asc2uni(). 2995 [Steve Henson, reported by <oddissey@samsung.co.kr>] 2996 2997 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 2998 call failed, free the DSA structure. 2999 [Bodo Moeller] 3000 3001 *) Fix to uni2asc() to cope with zero length Unicode strings. 3002 These are present in some PKCS#12 files. 3003 [Steve Henson] 3004 3005 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 3006 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 3007 when writing a 32767 byte record. 3008 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 3009 3010 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 3011 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 3012 3013 (RSA objects have a reference count access to which is protected 3014 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 3015 so they are meant to be shared between threads.) 3016 [Bodo Moeller, Geoff Thorpe; original patch submitted by 3017 "Reddie, Steven" <Steven.Reddie@ca.com>] 3018 3019 *) Fix a deadlock in CRYPTO_mem_leaks(). 3020 [Bodo Moeller] 3021 3022 *) Use better test patterns in bntest. 3023 [Ulf M�ller] 3024 3025 *) rand_win.c fix for Borland C. 3026 [Ulf M�ller] 3027 3028 *) BN_rshift bugfix for n == 0. 3029 [Bodo Moeller] 3030 3031 *) Add a 'bctest' script that checks for some known 'bc' bugs 3032 so that 'make test' does not abort just because 'bc' is broken. 3033 [Bodo Moeller] 3034 3035 *) Store verify_result within SSL_SESSION also for client side to 3036 avoid potential security hole. (Re-used sessions on the client side 3037 always resulted in verify_result==X509_V_OK, not using the original 3038 result of the server certificate verification.) 3039 [Lutz Jaenicke] 3040 3041 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 3042 SSL3_RT_APPLICATION_DATA, return 0. 3043 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 3044 [Bodo Moeller] 3045 3046 *) Fix SSL_peek: 3047 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 3048 releases, have been re-implemented by renaming the previous 3049 implementations of ssl2_read and ssl3_read to ssl2_read_internal 3050 and ssl3_read_internal, respectively, and adding 'peek' parameters 3051 to them. The new ssl[23]_{read,peek} functions are calls to 3052 ssl[23]_read_internal with the 'peek' flag set appropriately. 3053 A 'peek' parameter has also been added to ssl3_read_bytes, which 3054 does the actual work for ssl3_read_internal. 3055 [Bodo Moeller] 3056 3057 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 3058 the method-specific "init()" handler. Also clean up ex_data after 3059 calling the method-specific "finish()" handler. Previously, this was 3060 happening the other way round. 3061 [Geoff Thorpe] 3062 3063 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 3064 The previous value, 12, was not always sufficient for BN_mod_exp(). 3065 [Bodo Moeller] 3066 3067 *) Make sure that shared libraries get the internal name engine with 3068 the full version number and not just 0. This should mark the 3069 shared libraries as not backward compatible. Of course, this should 3070 be changed again when we can guarantee backward binary compatibility. 3071 [Richard Levitte] 3072 3073 *) Fix typo in get_cert_by_subject() in by_dir.c 3074 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 3075 3076 *) Rework the system to generate shared libraries: 3077 3078 - Make note of the expected extension for the shared libraries and 3079 if there is a need for symbolic links from for example libcrypto.so.0 3080 to libcrypto.so.0.9.7. There is extended info in Configure for 3081 that. 3082 3083 - Make as few rebuilds of the shared libraries as possible. 3084 3085 - Still avoid linking the OpenSSL programs with the shared libraries. 3086 3087 - When installing, install the shared libraries separately from the 3088 static ones. 3089 [Richard Levitte] 3090 3091 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 3092 3093 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 3094 and not in SSL_clear because the latter is also used by the 3095 accept/connect functions; previously, the settings made by 3096 SSL_set_read_ahead would be lost during the handshake. 3097 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 3098 3099 *) Correct util/mkdef.pl to be selective about disabled algorithms. 3100 Previously, it would create entries for disableed algorithms no 3101 matter what. 3102 [Richard Levitte] 3103 3104 *) Added several new manual pages for SSL_* function. 3105 [Lutz Jaenicke] 3106 3107 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 3108 3109 *) In ssl23_get_client_hello, generate an error message when faced 3110 with an initial SSL 3.0/TLS record that is too small to contain the 3111 first two bytes of the ClientHello message, i.e. client_version. 3112 (Note that this is a pathologic case that probably has never happened 3113 in real life.) The previous approach was to use the version number 3114 from the record header as a substitute; but our protocol choice 3115 should not depend on that one because it is not authenticated 3116 by the Finished messages. 3117 [Bodo Moeller] 3118 3119 *) More robust randomness gathering functions for Windows. 3120 [Jeffrey Altman <jaltman@columbia.edu>] 3121 3122 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 3123 not set then we don't setup the error code for issuer check errors 3124 to avoid possibly overwriting other errors which the callback does 3125 handle. If an application does set the flag then we assume it knows 3126 what it is doing and can handle the new informational codes 3127 appropriately. 3128 [Steve Henson] 3129 3130 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 3131 a general "ANY" type, as such it should be able to decode anything 3132 including tagged types. However it didn't check the class so it would 3133 wrongly interpret tagged types in the same way as their universal 3134 counterpart and unknown types were just rejected. Changed so that the 3135 tagged and unknown types are handled in the same way as a SEQUENCE: 3136 that is the encoding is stored intact. There is also a new type 3137 "V_ASN1_OTHER" which is used when the class is not universal, in this 3138 case we have no idea what the actual type is so we just lump them all 3139 together. 3140 [Steve Henson] 3141 3142 *) On VMS, stdout may very well lead to a file that is written to 3143 in a record-oriented fashion. That means that every write() will 3144 write a separate record, which will be read separately by the 3145 programs trying to read from it. This can be very confusing. 3146 3147 The solution is to put a BIO filter in the way that will buffer 3148 text until a linefeed is reached, and then write everything a 3149 line at a time, so every record written will be an actual line, 3150 not chunks of lines and not (usually doesn't happen, but I've 3151 seen it once) several lines in one record. BIO_f_linebuffer() is 3152 the answer. 3153 3154 Currently, it's a VMS-only method, because that's where it has 3155 been tested well enough. 3156 [Richard Levitte] 3157 3158 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 3159 it can return incorrect results. 3160 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 3161 but it was in 0.9.6-beta[12].) 3162 [Bodo Moeller] 3163 3164 *) Disable the check for content being present when verifying detached 3165 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 3166 include zero length content when signing messages. 3167 [Steve Henson] 3168 3169 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 3170 BIO_ctrl (for BIO pairs). 3171 [Bodo M�ller] 3172 3173 *) Add DSO method for VMS. 3174 [Richard Levitte] 3175 3176 *) Bug fix: Montgomery multiplication could produce results with the 3177 wrong sign. 3178 [Ulf M�ller] 3179 3180 *) Add RPM specification openssl.spec and modify it to build three 3181 packages. The default package contains applications, application 3182 documentation and run-time libraries. The devel package contains 3183 include files, static libraries and function documentation. The 3184 doc package contains the contents of the doc directory. The original 3185 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 3186 [Richard Levitte] 3187 3188 *) Add a large number of documentation files for many SSL routines. 3189 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 3190 3191 *) Add a configuration entry for Sony News 4. 3192 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 3193 3194 *) Don't set the two most significant bits to one when generating a 3195 random number < q in the DSA library. 3196 [Ulf M�ller] 3197 3198 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 3199 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 3200 the underlying transport is blocking) if a handshake took place. 3201 (The default behaviour is needed by applications such as s_client 3202 and s_server that use select() to determine when to use SSL_read; 3203 but for applications that know in advance when to expect data, it 3204 just makes things more complicated.) 3205 [Bodo Moeller] 3206 3207 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 3208 from EGD. 3209 [Ben Laurie] 3210 3211 *) Add a few more EBCDIC conditionals that make `req' and `x509' 3212 work better on such systems. 3213 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 3214 3215 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 3216 Update PKCS12_parse() so it copies the friendlyName and the 3217 keyid to the certificates aux info. 3218 [Steve Henson] 3219 3220 *) Fix bug in PKCS7_verify() which caused an infinite loop 3221 if there was more than one signature. 3222 [Sven Uszpelkat <su@celocom.de>] 3223 3224 *) Major change in util/mkdef.pl to include extra information 3225 about each symbol, as well as presentig variables as well 3226 as functions. This change means that there's n more need 3227 to rebuild the .num files when some algorithms are excluded. 3228 [Richard Levitte] 3229 3230 *) Allow the verify time to be set by an application, 3231 rather than always using the current time. 3232 [Steve Henson] 3233 3234 *) Phase 2 verify code reorganisation. The certificate 3235 verify code now looks up an issuer certificate by a 3236 number of criteria: subject name, authority key id 3237 and key usage. It also verifies self signed certificates 3238 by the same criteria. The main comparison function is 3239 X509_check_issued() which performs these checks. 3240 3241 Lot of changes were necessary in order to support this 3242 without completely rewriting the lookup code. 3243 3244 Authority and subject key identifier are now cached. 3245 3246 The LHASH 'certs' is X509_STORE has now been replaced 3247 by a STACK_OF(X509_OBJECT). This is mainly because an 3248 LHASH can't store or retrieve multiple objects with 3249 the same hash value. 3250 3251 As a result various functions (which were all internal 3252 use only) have changed to handle the new X509_STORE 3253 structure. This will break anything that messed round 3254 with X509_STORE internally. 3255 3256 The functions X509_STORE_add_cert() now checks for an 3257 exact match, rather than just subject name. 3258 3259 The X509_STORE API doesn't directly support the retrieval 3260 of multiple certificates matching a given criteria, however 3261 this can be worked round by performing a lookup first 3262 (which will fill the cache with candidate certificates) 3263 and then examining the cache for matches. This is probably 3264 the best we can do without throwing out X509_LOOKUP 3265 entirely (maybe later...). 3266 3267 The X509_VERIFY_CTX structure has been enhanced considerably. 3268 3269 All certificate lookup operations now go via a get_issuer() 3270 callback. Although this currently uses an X509_STORE it 3271 can be replaced by custom lookups. This is a simple way 3272 to bypass the X509_STORE hackery necessary to make this 3273 work and makes it possible to use more efficient techniques 3274 in future. A very simple version which uses a simple 3275 STACK for its trusted certificate store is also provided 3276 using X509_STORE_CTX_trusted_stack(). 3277 3278 The verify_cb() and verify() callbacks now have equivalents 3279 in the X509_STORE_CTX structure. 3280 3281 X509_STORE_CTX also has a 'flags' field which can be used 3282 to customise the verify behaviour. 3283 [Steve Henson] 3284 3285 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 3286 excludes S/MIME capabilities. 3287 [Steve Henson] 3288 3289 *) When a certificate request is read in keep a copy of the 3290 original encoding of the signed data and use it when outputing 3291 again. Signatures then use the original encoding rather than 3292 a decoded, encoded version which may cause problems if the 3293 request is improperly encoded. 3294 [Steve Henson] 3295 3296 *) For consistency with other BIO_puts implementations, call 3297 buffer_write(b, ...) directly in buffer_puts instead of calling 3298 BIO_write(b, ...). 3299 3300 In BIO_puts, increment b->num_write as in BIO_write. 3301 [Peter.Sylvester@EdelWeb.fr] 3302 3303 *) Fix BN_mul_word for the case where the word is 0. (We have to use 3304 BN_zero, we may not return a BIGNUM with an array consisting of 3305 words set to zero.) 3306 [Bodo Moeller] 3307 3308 *) Avoid calling abort() from within the library when problems are 3309 detected, except if preprocessor symbols have been defined 3310 (such as REF_CHECK, BN_DEBUG etc.). 3311 [Bodo Moeller] 3312 3313 *) New openssl application 'rsautl'. This utility can be 3314 used for low level RSA operations. DER public key 3315 BIO/fp routines also added. 3316 [Steve Henson] 3317 3318 *) New Configure entry and patches for compiling on QNX 4. 3319 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 3320 3321 *) A demo state-machine implementation was sponsored by 3322 Nuron (http://www.nuron.com/) and is now available in 3323 demos/state_machine. 3324 [Ben Laurie] 3325 3326 *) New options added to the 'dgst' utility for signature 3327 generation and verification. 3328 [Steve Henson] 3329 3330 *) Unrecognized PKCS#7 content types are now handled via a 3331 catch all ASN1_TYPE structure. This allows unsupported 3332 types to be stored as a "blob" and an application can 3333 encode and decode it manually. 3334 [Steve Henson] 3335 3336 *) Fix various signed/unsigned issues to make a_strex.c 3337 compile under VC++. 3338 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 3339 3340 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 3341 length if passed a buffer. ASN1_INTEGER_to_BN failed 3342 if passed a NULL BN and its argument was negative. 3343 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 3344 3345 *) Modification to PKCS#7 encoding routines to output definite 3346 length encoding. Since currently the whole structures are in 3347 memory there's not real point in using indefinite length 3348 constructed encoding. However if OpenSSL is compiled with 3349 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 3350 [Steve Henson] 3351 3352 *) Added BIO_vprintf() and BIO_vsnprintf(). 3353 [Richard Levitte] 3354 3355 *) Added more prefixes to parse for in the the strings written 3356 through a logging bio, to cover all the levels that are available 3357 through syslog. The prefixes are now: 3358 3359 PANIC, EMERG, EMR => LOG_EMERG 3360 ALERT, ALR => LOG_ALERT 3361 CRIT, CRI => LOG_CRIT 3362 ERROR, ERR => LOG_ERR 3363 WARNING, WARN, WAR => LOG_WARNING 3364 NOTICE, NOTE, NOT => LOG_NOTICE 3365 INFO, INF => LOG_INFO 3366 DEBUG, DBG => LOG_DEBUG 3367 3368 and as before, if none of those prefixes are present at the 3369 beginning of the string, LOG_ERR is chosen. 3370 3371 On Win32, the LOG_* levels are mapped according to this: 3372 3373 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 3374 LOG_WARNING => EVENTLOG_WARNING_TYPE 3375 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 3376 3377 [Richard Levitte] 3378 3379 *) Made it possible to reconfigure with just the configuration 3380 argument "reconf" or "reconfigure". The command line arguments 3381 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 3382 and are retrieved from there when reconfiguring. 3383 [Richard Levitte] 3384 3385 *) MD4 implemented. 3386 [Assar Westerlund <assar@sics.se>, Richard Levitte] 3387 3388 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 3389 [Richard Levitte] 3390 3391 *) The obj_dat.pl script was messing up the sorting of object 3392 names. The reason was that it compared the quoted version 3393 of strings as a result "OCSP" > "OCSP Signing" because 3394 " > SPACE. Changed script to store unquoted versions of 3395 names and add quotes on output. It was also omitting some 3396 names from the lookup table if they were given a default 3397 value (that is if SN is missing it is given the same 3398 value as LN and vice versa), these are now added on the 3399 grounds that if an object has a name we should be able to 3400 look it up. Finally added warning output when duplicate 3401 short or long names are found. 3402 [Steve Henson] 3403 3404 *) Changes needed for Tandem NSK. 3405 [Scott Uroff <scott@xypro.com>] 3406 3407 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 3408 RSA_padding_check_SSLv23(), special padding was never detected 3409 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 3410 version rollback attacks was not effective. 3411 3412 In s23_clnt.c, don't use special rollback-attack detection padding 3413 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 3414 client; similarly, in s23_srvr.c, don't do the rollback check if 3415 SSL 2.0 is the only protocol enabled in the server. 3416 [Bodo Moeller] 3417 3418 *) Make it possible to get hexdumps of unprintable data with 'openssl 3419 asn1parse'. By implication, the functions ASN1_parse_dump() and 3420 BIO_dump_indent() are added. 3421 [Richard Levitte] 3422 3423 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 3424 these print out strings and name structures based on various 3425 flags including RFC2253 support and proper handling of 3426 multibyte characters. Added options to the 'x509' utility 3427 to allow the various flags to be set. 3428 [Steve Henson] 3429 3430 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 3431 Also change the functions X509_cmp_current_time() and 3432 X509_gmtime_adj() work with an ASN1_TIME structure, 3433 this will enable certificates using GeneralizedTime in validity 3434 dates to be checked. 3435 [Steve Henson] 3436 3437 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 3438 negative public key encodings) on by default, 3439 NO_NEG_PUBKEY_BUG can be set to disable it. 3440 [Steve Henson] 3441 3442 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 3443 content octets. An i2c_ASN1_OBJECT is unnecessary because 3444 the encoding can be trivially obtained from the structure. 3445 [Steve Henson] 3446 3447 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 3448 not read locks (CRYPTO_r_[un]lock). 3449 [Bodo Moeller] 3450 3451 *) A first attempt at creating official support for shared 3452 libraries through configuration. I've kept it so the 3453 default is static libraries only, and the OpenSSL programs 3454 are always statically linked for now, but there are 3455 preparations for dynamic linking in place. 3456 This has been tested on Linux and Tru64. 3457 [Richard Levitte] 3458 3459 *) Randomness polling function for Win9x, as described in: 3460 Peter Gutmann, Software Generation of Practically Strong 3461 Random Numbers. 3462 [Ulf M�ller] 3463 3464 *) Fix so PRNG is seeded in req if using an already existing 3465 DSA key. 3466 [Steve Henson] 3467 3468 *) New options to smime application. -inform and -outform 3469 allow alternative formats for the S/MIME message including 3470 PEM and DER. The -content option allows the content to be 3471 specified separately. This should allow things like Netscape 3472 form signing output easier to verify. 3473 [Steve Henson] 3474 3475 *) Fix the ASN1 encoding of tags using the 'long form'. 3476 [Steve Henson] 3477 3478 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 3479 STRING types. These convert content octets to and from the 3480 underlying type. The actual tag and length octets are 3481 already assumed to have been read in and checked. These 3482 are needed because all other string types have virtually 3483 identical handling apart from the tag. By having versions 3484 of the ASN1 functions that just operate on content octets 3485 IMPLICIT tagging can be handled properly. It also allows 3486 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 3487 and ASN1_INTEGER are identical apart from the tag. 3488 [Steve Henson] 3489 3490 *) Change the handling of OID objects as follows: 3491 3492 - New object identifiers are inserted in objects.txt, following 3493 the syntax given in objects.README. 3494 - objects.pl is used to process obj_mac.num and create a new 3495 obj_mac.h. 3496 - obj_dat.pl is used to create a new obj_dat.h, using the data in 3497 obj_mac.h. 3498 3499 This is currently kind of a hack, and the perl code in objects.pl 3500 isn't very elegant, but it works as I intended. The simplest way 3501 to check that it worked correctly is to look in obj_dat.h and 3502 check the array nid_objs and make sure the objects haven't moved 3503 around (this is important!). Additions are OK, as well as 3504 consistent name changes. 3505 [Richard Levitte] 3506 3507 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 3508 [Bodo Moeller] 3509 3510 *) Addition of the command line parameter '-rand file' to 'openssl req'. 3511 The given file adds to whatever has already been seeded into the 3512 random pool through the RANDFILE configuration file option or 3513 environment variable, or the default random state file. 3514 [Richard Levitte] 3515 3516 *) mkstack.pl now sorts each macro group into lexical order. 3517 Previously the output order depended on the order the files 3518 appeared in the directory, resulting in needless rewriting 3519 of safestack.h . 3520 [Steve Henson] 3521 3522 *) Patches to make OpenSSL compile under Win32 again. Mostly 3523 work arounds for the VC++ problem that it treats func() as 3524 func(void). Also stripped out the parts of mkdef.pl that 3525 added extra typesafe functions: these no longer exist. 3526 [Steve Henson] 3527 3528 *) Reorganisation of the stack code. The macros are now all 3529 collected in safestack.h . Each macro is defined in terms of 3530 a "stack macro" of the form SKM_<name>(type, a, b). The 3531 DEBUG_SAFESTACK is now handled in terms of function casts, 3532 this has the advantage of retaining type safety without the 3533 use of additional functions. If DEBUG_SAFESTACK is not defined 3534 then the non typesafe macros are used instead. Also modified the 3535 mkstack.pl script to handle the new form. Needs testing to see 3536 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 3537 the default if no major problems. Similar behaviour for ASN1_SET_OF 3538 and PKCS12_STACK_OF. 3539 [Steve Henson] 3540 3541 *) When some versions of IIS use the 'NET' form of private key the 3542 key derivation algorithm is different. Normally MD5(password) is 3543 used as a 128 bit RC4 key. In the modified case 3544 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 3545 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 3546 as the old Netscape_RSA functions except they have an additional 3547 'sgckey' parameter which uses the modified algorithm. Also added 3548 an -sgckey command line option to the rsa utility. Thanks to 3549 Adrian Peck <bertie@ncipher.com> for posting details of the modified 3550 algorithm to openssl-dev. 3551 [Steve Henson] 3552 3553 *) The evp_local.h macros were using 'c.##kname' which resulted in 3554 invalid expansion on some systems (SCO 5.0.5 for example). 3555 Corrected to 'c.kname'. 3556 [Phillip Porch <root@theporch.com>] 3557 3558 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 3559 a STACK of email addresses from a certificate or request, these look 3560 in the subject name and the subject alternative name extensions and 3561 omit any duplicate addresses. 3562 [Steve Henson] 3563 3564 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 3565 This makes DSA verification about 2 % faster. 3566 [Bodo Moeller] 3567 3568 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 3569 (meaning that now 2^5 values will be precomputed, which is only 4 KB 3570 plus overhead for 1024 bit moduli). 3571 This makes exponentiations about 0.5 % faster for 1024 bit 3572 exponents (as measured by "openssl speed rsa2048"). 3573 [Bodo Moeller] 3574 3575 *) Rename memory handling macros to avoid conflicts with other 3576 software: 3577 Malloc => OPENSSL_malloc 3578 Malloc_locked => OPENSSL_malloc_locked 3579 Realloc => OPENSSL_realloc 3580 Free => OPENSSL_free 3581 [Richard Levitte] 3582 3583 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 3584 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 3585 [Bodo Moeller] 3586 3587 *) CygWin32 support. 3588 [John Jarvie <jjarvie@newsguy.com>] 3589 3590 *) The type-safe stack code has been rejigged. It is now only compiled 3591 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 3592 by default all type-specific stack functions are "#define"d back to 3593 standard stack functions. This results in more streamlined output 3594 but retains the type-safety checking possibilities of the original 3595 approach. 3596 [Geoff Thorpe] 3597 3598 *) The STACK code has been cleaned up, and certain type declarations 3599 that didn't make a lot of sense have been brought in line. This has 3600 also involved a cleanup of sorts in safestack.h to more correctly 3601 map type-safe stack functions onto their plain stack counterparts. 3602 This work has also resulted in a variety of "const"ifications of 3603 lots of the code, especially "_cmp" operations which should normally 3604 be prototyped with "const" parameters anyway. 3605 [Geoff Thorpe] 3606 3607 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 3608 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 3609 (The PRNG state consists of two parts, the large pool 'state' and 'md', 3610 where all of 'md' is used each time the PRNG is used, but 'state' 3611 is used only indexed by a cyclic counter. As entropy may not be 3612 well distributed from the beginning, 'md' is important as a 3613 chaining variable. However, the output function chains only half 3614 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 3615 all of 'md', and seeding with STATE_SIZE dummy bytes will result 3616 in all of 'state' being rewritten, with the new values depending 3617 on virtually all of 'md'. This overcomes the 80 bit limitation.) 3618 [Bodo Moeller] 3619 3620 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 3621 the handshake is continued after ssl_verify_cert_chain(); 3622 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 3623 can lead to 'unexplainable' connection aborts later. 3624 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 3625 3626 *) Major EVP API cipher revision. 3627 Add hooks for extra EVP features. This allows various cipher 3628 parameters to be set in the EVP interface. Support added for variable 3629 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 3630 setting of RC2 and RC5 parameters. 3631 3632 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 3633 ciphers. 3634 3635 Remove lots of duplicated code from the EVP library. For example *every* 3636 cipher init() function handles the 'iv' in the same way according to the 3637 cipher mode. They also all do nothing if the 'key' parameter is NULL and 3638 for CFB and OFB modes they zero ctx->num. 3639 3640 New functionality allows removal of S/MIME code RC2 hack. 3641 3642 Most of the routines have the same form and so can be declared in terms 3643 of macros. 3644 3645 By shifting this to the top level EVP_CipherInit() it can be removed from 3646 all individual ciphers. If the cipher wants to handle IVs or keys 3647 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 3648 flags. 3649 3650 Change lots of functions like EVP_EncryptUpdate() to now return a 3651 value: although software versions of the algorithms cannot fail 3652 any installed hardware versions can. 3653 [Steve Henson] 3654 3655 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 3656 this option is set, tolerate broken clients that send the negotiated 3657 protocol version number instead of the requested protocol version 3658 number. 3659 [Bodo Moeller] 3660 3661 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 3662 i.e. non-zero for export ciphersuites, zero otherwise. 3663 Previous versions had this flag inverted, inconsistent with 3664 rsa_tmp_cb (..._TMP_RSA_CB). 3665 [Bodo Moeller; problem reported by Amit Chopra] 3666 3667 *) Add missing DSA library text string. Work around for some IIS 3668 key files with invalid SEQUENCE encoding. 3669 [Steve Henson] 3670 3671 *) Add a document (doc/standards.txt) that list all kinds of standards 3672 and so on that are implemented in OpenSSL. 3673 [Richard Levitte] 3674 3675 *) Enhance c_rehash script. Old version would mishandle certificates 3676 with the same subject name hash and wouldn't handle CRLs at all. 3677 Added -fingerprint option to crl utility, to support new c_rehash 3678 features. 3679 [Steve Henson] 3680 3681 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 3682 [Ulf M�ller] 3683 3684 *) Fix for SSL server purpose checking. Server checking was 3685 rejecting certificates which had extended key usage present 3686 but no ssl client purpose. 3687 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 3688 3689 *) Make PKCS#12 code work with no password. The PKCS#12 spec 3690 is a little unclear about how a blank password is handled. 3691 Since the password in encoded as a BMPString with terminating 3692 double NULL a zero length password would end up as just the 3693 double NULL. However no password at all is different and is 3694 handled differently in the PKCS#12 key generation code. NS 3695 treats a blank password as zero length. MSIE treats it as no 3696 password on export: but it will try both on import. We now do 3697 the same: PKCS12_parse() tries zero length and no password if 3698 the password is set to "" or NULL (NULL is now a valid password: 3699 it wasn't before) as does the pkcs12 application. 3700 [Steve Henson] 3701 3702 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 3703 perror when PEM_read_bio_X509_REQ fails, the error message must 3704 be obtained from the error queue. 3705 [Bodo Moeller] 3706 3707 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 3708 it in ERR_remove_state if appropriate, and change ERR_get_state 3709 accordingly to avoid race conditions (this is necessary because 3710 thread_hash is no longer constant once set). 3711 [Bodo Moeller] 3712 3713 *) Bugfix for linux-elf makefile.one. 3714 [Ulf M�ller] 3715 3716 *) RSA_get_default_method() will now cause a default 3717 RSA_METHOD to be chosen if one doesn't exist already. 3718 Previously this was only set during a call to RSA_new() 3719 or RSA_new_method(NULL) meaning it was possible for 3720 RSA_get_default_method() to return NULL. 3721 [Geoff Thorpe] 3722 3723 *) Added native name translation to the existing DSO code 3724 that will convert (if the flag to do so is set) filenames 3725 that are sufficiently small and have no path information 3726 into a canonical native form. Eg. "blah" converted to 3727 "libblah.so" or "blah.dll" etc. 3728 [Geoff Thorpe] 3729 3730 *) New function ERR_error_string_n(e, buf, len) which is like 3731 ERR_error_string(e, buf), but writes at most 'len' bytes 3732 including the 0 terminator. For ERR_error_string_n, 'buf' 3733 may not be NULL. 3734 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 3735 3736 *) CONF library reworked to become more general. A new CONF 3737 configuration file reader "class" is implemented as well as a 3738 new functions (NCONF_*, for "New CONF") to handle it. The now 3739 old CONF_* functions are still there, but are reimplemented to 3740 work in terms of the new functions. Also, a set of functions 3741 to handle the internal storage of the configuration data is 3742 provided to make it easier to write new configuration file 3743 reader "classes" (I can definitely see something reading a 3744 configuration file in XML format, for example), called _CONF_*, 3745 or "the configuration storage API"... 3746 3747 The new configuration file reading functions are: 3748 3749 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 3750 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 3751 3752 NCONF_default, NCONF_WIN32 3753 3754 NCONF_dump_fp, NCONF_dump_bio 3755 3756 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 3757 NCONF_new creates a new CONF object. This works in the same way 3758 as other interfaces in OpenSSL, like the BIO interface. 3759 NCONF_dump_* dump the internal storage of the configuration file, 3760 which is useful for debugging. All other functions take the same 3761 arguments as the old CONF_* functions wth the exception of the 3762 first that must be a `CONF *' instead of a `LHASH *'. 3763 3764 To make it easer to use the new classes with the old CONF_* functions, 3765 the function CONF_set_default_method is provided. 3766 [Richard Levitte] 3767 3768 *) Add '-tls1' option to 'openssl ciphers', which was already 3769 mentioned in the documentation but had not been implemented. 3770 (This option is not yet really useful because even the additional 3771 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 3772 [Bodo Moeller] 3773 3774 *) Initial DSO code added into libcrypto for letting OpenSSL (and 3775 OpenSSL-based applications) load shared libraries and bind to 3776 them in a portable way. 3777 [Geoff Thorpe, with contributions from Richard Levitte] 3778 3779 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 3780 3781 *) Make sure _lrotl and _lrotr are only used with MSVC. 3782 3783 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 3784 (the default implementation of RAND_status). 3785 3786 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 3787 to '-clrext' (= clear extensions), as intended and documented. 3788 [Bodo Moeller; inconsistency pointed out by Michael Attili 3789 <attili@amaxo.com>] 3790 3791 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 3792 was larger than the MD block size. 3793 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 3794 3795 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 3796 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 3797 using the passed key: if the passed key was a private key the result 3798 of X509_print(), for example, would be to print out all the private key 3799 components. 3800 [Steve Henson] 3801 3802 *) des_quad_cksum() byte order bug fix. 3803 [Ulf M�ller, using the problem description in krb4-0.9.7, where 3804 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 3805 3806 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 3807 discouraged. 3808 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 3809 3810 *) For easily testing in shell scripts whether some command 3811 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 3812 returns with exit code 0 iff no command of the given name is available. 3813 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 3814 the output goes to stdout and nothing is printed to stderr. 3815 Additional arguments are always ignored. 3816 3817 Since for each cipher there is a command of the same name, 3818 the 'no-cipher' compilation switches can be tested this way. 3819 3820 ('openssl no-XXX' is not able to detect pseudo-commands such 3821 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 3822 [Bodo Moeller] 3823 3824 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 3825 [Bodo Moeller] 3826 3827 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 3828 is set; it will be thrown away anyway because each handshake creates 3829 its own key. 3830 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 3831 to parameters -- in previous versions (since OpenSSL 0.9.3) the 3832 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 3833 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 3834 [Bodo Moeller] 3835 3836 *) New s_client option -ign_eof: EOF at stdin is ignored, and 3837 'Q' and 'R' lose their special meanings (quit/renegotiate). 3838 This is part of what -quiet does; unlike -quiet, -ign_eof 3839 does not suppress any output. 3840 [Richard Levitte] 3841 3842 *) Add compatibility options to the purpose and trust code. The 3843 purpose X509_PURPOSE_ANY is "any purpose" which automatically 3844 accepts a certificate or CA, this was the previous behaviour, 3845 with all the associated security issues. 3846 3847 X509_TRUST_COMPAT is the old trust behaviour: only and 3848 automatically trust self signed roots in certificate store. A 3849 new trust setting X509_TRUST_DEFAULT is used to specify that 3850 a purpose has no associated trust setting and it should instead 3851 use the value in the default purpose. 3852 [Steve Henson] 3853 3854 *) Fix the PKCS#8 DSA private key code so it decodes keys again 3855 and fix a memory leak. 3856 [Steve Henson] 3857 3858 *) In util/mkerr.pl (which implements 'make errors'), preserve 3859 reason strings from the previous version of the .c file, as 3860 the default to have only downcase letters (and digits) in 3861 automatically generated reasons codes is not always appropriate. 3862 [Bodo Moeller] 3863 3864 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 3865 using strerror. Previously, ERR_reason_error_string() returned 3866 library names as reason strings for SYSerr; but SYSerr is a special 3867 case where small numbers are errno values, not library numbers. 3868 [Bodo Moeller] 3869 3870 *) Add '-dsaparam' option to 'openssl dhparam' application. This 3871 converts DSA parameters into DH parameters. (When creating parameters, 3872 DSA_generate_parameters is used.) 3873 [Bodo Moeller] 3874 3875 *) Include 'length' (recommended exponent length) in C code generated 3876 by 'openssl dhparam -C'. 3877 [Bodo Moeller] 3878 3879 *) The second argument to set_label in perlasm was already being used 3880 so couldn't be used as a "file scope" flag. Moved to third argument 3881 which was free. 3882 [Steve Henson] 3883 3884 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 3885 instead of RAND_bytes for encryption IVs and salts. 3886 [Bodo Moeller] 3887 3888 *) Include RAND_status() into RAND_METHOD instead of implementing 3889 it only for md_rand.c Otherwise replacing the PRNG by calling 3890 RAND_set_rand_method would be impossible. 3891 [Bodo Moeller] 3892 3893 *) Don't let DSA_generate_key() enter an infinite loop if the random 3894 number generation fails. 3895 [Bodo Moeller] 3896 3897 *) New 'rand' application for creating pseudo-random output. 3898 [Bodo Moeller] 3899 3900 *) Added configuration support for Linux/IA64 3901 [Rolf Haberrecker <rolf@suse.de>] 3902 3903 *) Assembler module support for Mingw32. 3904 [Ulf M�ller] 3905 3906 *) Shared library support for HPUX (in shlib/). 3907 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 3908 3909 *) Shared library support for Solaris gcc. 3910 [Lutz Behnke <behnke@trustcenter.de>] 3911 3912 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 3913 3914 *) PKCS7_encrypt() was adding text MIME headers twice because they 3915 were added manually and by SMIME_crlf_copy(). 3916 [Steve Henson] 3917 3918 *) In bntest.c don't call BN_rand with zero bits argument. 3919 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 3920 3921 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 3922 case was implemented. This caused BN_div_recp() to fail occasionally. 3923 [Ulf M�ller] 3924 3925 *) Add an optional second argument to the set_label() in the perl 3926 assembly language builder. If this argument exists and is set 3927 to 1 it signals that the assembler should use a symbol whose 3928 scope is the entire file, not just the current function. This 3929 is needed with MASM which uses the format label:: for this scope. 3930 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 3931 3932 *) Change the ASN1 types so they are typedefs by default. Before 3933 almost all types were #define'd to ASN1_STRING which was causing 3934 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 3935 for example. 3936 [Steve Henson] 3937 3938 *) Change names of new functions to the new get1/get0 naming 3939 convention: After 'get1', the caller owns a reference count 3940 and has to call ..._free; 'get0' returns a pointer to some 3941 data structure without incrementing reference counters. 3942 (Some of the existing 'get' functions increment a reference 3943 counter, some don't.) 3944 Similarly, 'set1' and 'add1' functions increase reference 3945 counters or duplicate objects. 3946 [Steve Henson] 3947 3948 *) Allow for the possibility of temp RSA key generation failure: 3949 the code used to assume it always worked and crashed on failure. 3950 [Steve Henson] 3951 3952 *) Fix potential buffer overrun problem in BIO_printf(). 3953 [Ulf M�ller, using public domain code by Patrick Powell; problem 3954 pointed out by David Sacerdote <das33@cornell.edu>] 3955 3956 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 3957 RAND_egd() and RAND_status(). In the command line application, 3958 the EGD socket can be specified like a seed file using RANDFILE 3959 or -rand. 3960 [Ulf M�ller] 3961 3962 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 3963 Some CAs (e.g. Verisign) distribute certificates in this form. 3964 [Steve Henson] 3965 3966 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 3967 list to exclude them. This means that no special compilation option 3968 is needed to use anonymous DH: it just needs to be included in the 3969 cipher list. 3970 [Steve Henson] 3971 3972 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 3973 EVP_MD_type. The old functionality is available in a new macro called 3974 EVP_MD_md(). Change code that uses it and update docs. 3975 [Steve Henson] 3976 3977 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 3978 where the 'void *' argument is replaced by a function pointer argument. 3979 Previously 'void *' was abused to point to functions, which works on 3980 many platforms, but is not correct. As these functions are usually 3981 called by macros defined in OpenSSL header files, most source code 3982 should work without changes. 3983 [Richard Levitte] 3984 3985 *) <openssl/opensslconf.h> (which is created by Configure) now contains 3986 sections with information on -D... compiler switches used for 3987 compiling the library so that applications can see them. To enable 3988 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 3989 must be defined. E.g., 3990 #define OPENSSL_ALGORITHM_DEFINES 3991 #include <openssl/opensslconf.h> 3992 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 3993 [Richard Levitte, Ulf and Bodo M�ller] 3994 3995 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 3996 record layer. 3997 [Bodo Moeller] 3998 3999 *) Change the 'other' type in certificate aux info to a STACK_OF 4000 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 4001 the required ASN1 format: arbitrary types determined by an OID. 4002 [Steve Henson] 4003 4004 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 4005 argument to 'req'. This is not because the function is newer or 4006 better than others it just uses the work 'NEW' in the certificate 4007 request header lines. Some software needs this. 4008 [Steve Henson] 4009 4010 *) Reorganise password command line arguments: now passwords can be 4011 obtained from various sources. Delete the PEM_cb function and make 4012 it the default behaviour: i.e. if the callback is NULL and the 4013 usrdata argument is not NULL interpret it as a null terminated pass 4014 phrase. If usrdata and the callback are NULL then the pass phrase 4015 is prompted for as usual. 4016 [Steve Henson] 4017 4018 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 4019 the support is automatically enabled. The resulting binaries will 4020 autodetect the card and use it if present. 4021 [Ben Laurie and Compaq Inc.] 4022 4023 *) Work around for Netscape hang bug. This sends certificate request 4024 and server done in one record. Since this is perfectly legal in the 4025 SSL/TLS protocol it isn't a "bug" option and is on by default. See 4026 the bugs/SSLv3 entry for more info. 4027 [Steve Henson] 4028 4029 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 4030 [Andy Polyakov] 4031 4032 *) Add -rand argument to smime and pkcs12 applications and read/write 4033 of seed file. 4034 [Steve Henson] 4035 4036 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 4037 [Bodo Moeller] 4038 4039 *) Add command line password options to the remaining applications. 4040 [Steve Henson] 4041 4042 *) Bug fix for BN_div_recp() for numerators with an even number of 4043 bits. 4044 [Ulf M�ller] 4045 4046 *) More tests in bntest.c, and changed test_bn output. 4047 [Ulf M�ller] 4048 4049 *) ./config recognizes MacOS X now. 4050 [Andy Polyakov] 4051 4052 *) Bug fix for BN_div() when the first words of num and divsor are 4053 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 4054 [Ulf M�ller] 4055 4056 *) Add support for various broken PKCS#8 formats, and command line 4057 options to produce them. 4058 [Steve Henson] 4059 4060 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 4061 get temporary BIGNUMs from a BN_CTX. 4062 [Ulf M�ller] 4063 4064 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 4065 for p == 0. 4066 [Ulf M�ller] 4067 4068 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 4069 include a #define from the old name to the new. The original intent 4070 was that statically linked binaries could for example just call 4071 SSLeay_add_all_ciphers() to just add ciphers to the table and not 4072 link with digests. This never worked becayse SSLeay_add_all_digests() 4073 and SSLeay_add_all_ciphers() were in the same source file so calling 4074 one would link with the other. They are now in separate source files. 4075 [Steve Henson] 4076 4077 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 4078 [Steve Henson] 4079 4080 *) Use a less unusual form of the Miller-Rabin primality test (it used 4081 a binary algorithm for exponentiation integrated into the Miller-Rabin 4082 loop, our standard modexp algorithms are faster). 4083 [Bodo Moeller] 4084 4085 *) Support for the EBCDIC character set completed. 4086 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 4087 4088 *) Source code cleanups: use const where appropriate, eliminate casts, 4089 use void * instead of char * in lhash. 4090 [Ulf M�ller] 4091 4092 *) Bugfix: ssl3_send_server_key_exchange was not restartable 4093 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 4094 this the server could overwrite ephemeral keys that the client 4095 has already seen). 4096 [Bodo Moeller] 4097 4098 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 4099 using 50 iterations of the Rabin-Miller test. 4100 4101 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 4102 iterations of the Rabin-Miller test as required by the appendix 4103 to FIPS PUB 186[-1]) instead of DSA_is_prime. 4104 As BN_is_prime_fasttest includes trial division, DSA parameter 4105 generation becomes much faster. 4106 4107 This implies a change for the callback functions in DSA_is_prime 4108 and DSA_generate_parameters: The callback function is called once 4109 for each positive witness in the Rabin-Miller test, not just 4110 occasionally in the inner loop; and the parameters to the 4111 callback function now provide an iteration count for the outer 4112 loop rather than for the current invocation of the inner loop. 4113 DSA_generate_parameters additionally can call the callback 4114 function with an 'iteration count' of -1, meaning that a 4115 candidate has passed the trial division test (when q is generated 4116 from an application-provided seed, trial division is skipped). 4117 [Bodo Moeller] 4118 4119 *) New function BN_is_prime_fasttest that optionally does trial 4120 division before starting the Rabin-Miller test and has 4121 an additional BN_CTX * argument (whereas BN_is_prime always 4122 has to allocate at least one BN_CTX). 4123 'callback(1, -1, cb_arg)' is called when a number has passed the 4124 trial division stage. 4125 [Bodo Moeller] 4126 4127 *) Fix for bug in CRL encoding. The validity dates weren't being handled 4128 as ASN1_TIME. 4129 [Steve Henson] 4130 4131 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 4132 [Steve Henson] 4133 4134 *) New function BN_pseudo_rand(). 4135 [Ulf M�ller] 4136 4137 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 4138 bignum version of BN_from_montgomery() with the working code from 4139 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 4140 the comments. 4141 [Ulf M�ller] 4142 4143 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 4144 made it impossible to use the same SSL_SESSION data structure in 4145 SSL2 clients in multiple threads. 4146 [Bodo Moeller] 4147 4148 *) The return value of RAND_load_file() no longer counts bytes obtained 4149 by stat(). RAND_load_file(..., -1) is new and uses the complete file 4150 to seed the PRNG (previously an explicit byte count was required). 4151 [Ulf M�ller, Bodo M�ller] 4152 4153 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 4154 used (char *) instead of (void *) and had casts all over the place. 4155 [Steve Henson] 4156 4157 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 4158 [Ulf M�ller] 4159 4160 *) Retain source code compatibility for BN_prime_checks macro: 4161 BN_is_prime(..., BN_prime_checks, ...) now uses 4162 BN_prime_checks_for_size to determine the appropriate number of 4163 Rabin-Miller iterations. 4164 [Ulf M�ller] 4165 4166 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 4167 DH_CHECK_P_NOT_SAFE_PRIME. 4168 (Check if this is true? OpenPGP calls them "strong".) 4169 [Ulf M�ller] 4170 4171 *) Merge the functionality of "dh" and "gendh" programs into a new program 4172 "dhparam". The old programs are retained for now but will handle DH keys 4173 (instead of parameters) in future. 4174 [Steve Henson] 4175 4176 *) Make the ciphers, s_server and s_client programs check the return values 4177 when a new cipher list is set. 4178 [Steve Henson] 4179 4180 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 4181 ciphers. Before when the 56bit ciphers were enabled the sorting was 4182 wrong. 4183 4184 The syntax for the cipher sorting has been extended to support sorting by 4185 cipher-strength (using the strength_bits hard coded in the tables). 4186 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 4187 4188 Fix a bug in the cipher-command parser: when supplying a cipher command 4189 string with an "undefined" symbol (neither command nor alphanumeric 4190 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 4191 an error is flagged. 4192 4193 Due to the strength-sorting extension, the code of the 4194 ssl_create_cipher_list() function was completely rearranged. I hope that 4195 the readability was also increased :-) 4196 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 4197 4198 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 4199 for the first serial number and places 2 in the serial number file. This 4200 avoids problems when the root CA is created with serial number zero and 4201 the first user certificate has the same issuer name and serial number 4202 as the root CA. 4203 [Steve Henson] 4204 4205 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 4206 the new code. Add documentation for this stuff. 4207 [Steve Henson] 4208 4209 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 4210 X509_*() to X509at_*() on the grounds that they don't handle X509 4211 structures and behave in an analagous way to the X509v3 functions: 4212 they shouldn't be called directly but wrapper functions should be used 4213 instead. 4214 4215 So we also now have some wrapper functions that call the X509at functions 4216 when passed certificate requests. (TO DO: similar things can be done with 4217 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 4218 things. Some of these need some d2i or i2d and print functionality 4219 because they handle more complex structures.) 4220 [Steve Henson] 4221 4222 *) Add missing #ifndefs that caused missing symbols when building libssl 4223 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 4224 NO_RSA in ssl/s2*.c. 4225 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 4226 4227 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 4228 has a return value which indicates the quality of the random data 4229 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 4230 error queue. New function RAND_pseudo_bytes() generates output that is 4231 guaranteed to be unique but not unpredictable. RAND_add is like 4232 RAND_seed, but takes an extra argument for an entropy estimate 4233 (RAND_seed always assumes full entropy). 4234 [Ulf M�ller] 4235 4236 *) Do more iterations of Rabin-Miller probable prime test (specifically, 4237 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 4238 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 4239 in crypto/bn/bn_prime.c for the complete table). This guarantees a 4240 false-positive rate of at most 2^-80 for random input. 4241 [Bodo Moeller] 4242 4243 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 4244 [Bodo Moeller] 4245 4246 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 4247 in the 0.9.5 release), this returns the chain 4248 from an X509_CTX structure with a dup of the stack and all 4249 the X509 reference counts upped: so the stack will exist 4250 after X509_CTX_cleanup() has been called. Modify pkcs12.c 4251 to use this. 4252 4253 Also make SSL_SESSION_print() print out the verify return 4254 code. 4255 [Steve Henson] 4256 4257 *) Add manpage for the pkcs12 command. Also change the default 4258 behaviour so MAC iteration counts are used unless the new 4259 -nomaciter option is used. This improves file security and 4260 only older versions of MSIE (4.0 for example) need it. 4261 [Steve Henson] 4262 4263 *) Honor the no-xxx Configure options when creating .DEF files. 4264 [Ulf M�ller] 4265 4266 *) Add PKCS#10 attributes to field table: challengePassword, 4267 unstructuredName and unstructuredAddress. These are taken from 4268 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 4269 international characters are used. 4270 4271 More changes to X509_ATTRIBUTE code: allow the setting of types 4272 based on strings. Remove the 'loc' parameter when adding 4273 attributes because these will be a SET OF encoding which is sorted 4274 in ASN1 order. 4275 [Steve Henson] 4276 4277 *) Initial changes to the 'req' utility to allow request generation 4278 automation. This will allow an application to just generate a template 4279 file containing all the field values and have req construct the 4280 request. 4281 4282 Initial support for X509_ATTRIBUTE handling. Stacks of these are 4283 used all over the place including certificate requests and PKCS#7 4284 structures. They are currently handled manually where necessary with 4285 some primitive wrappers for PKCS#7. The new functions behave in a 4286 manner analogous to the X509 extension functions: they allow 4287 attributes to be looked up by NID and added. 4288 4289 Later something similar to the X509V3 code would be desirable to 4290 automatically handle the encoding, decoding and printing of the 4291 more complex types. The string types like challengePassword can 4292 be handled by the string table functions. 4293 4294 Also modified the multi byte string table handling. Now there is 4295 a 'global mask' which masks out certain types. The table itself 4296 can use the flag STABLE_NO_MASK to ignore the mask setting: this 4297 is useful when for example there is only one permissible type 4298 (as in countryName) and using the mask might result in no valid 4299 types at all. 4300 [Steve Henson] 4301 4302 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 4303 SSL_get_peer_finished to allow applications to obtain the latest 4304 Finished messages sent to the peer or expected from the peer, 4305 respectively. (SSL_get_peer_finished is usually the Finished message 4306 actually received from the peer, otherwise the protocol will be aborted.) 4307 4308 As the Finished message are message digests of the complete handshake 4309 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 4310 be used for external authentication procedures when the authentication 4311 provided by SSL/TLS is not desired or is not enough. 4312 [Bodo Moeller] 4313 4314 *) Enhanced support for Alpha Linux is added. Now ./config checks if 4315 the host supports BWX extension and if Compaq C is present on the 4316 $PATH. Just exploiting of the BWX extension results in 20-30% 4317 performance kick for some algorithms, e.g. DES and RC4 to mention 4318 a couple. Compaq C in turn generates ~20% faster code for MD5 and 4319 SHA1. 4320 [Andy Polyakov] 4321 4322 *) Add support for MS "fast SGC". This is arguably a violation of the 4323 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 4324 weak crypto and after checking the certificate is SGC a second one 4325 with strong crypto. MS SGC stops the first handshake after receiving 4326 the server certificate message and sends a second client hello. Since 4327 a server will typically do all the time consuming operations before 4328 expecting any further messages from the client (server key exchange 4329 is the most expensive) there is little difference between the two. 4330 4331 To get OpenSSL to support MS SGC we have to permit a second client 4332 hello message after we have sent server done. In addition we have to 4333 reset the MAC if we do get this second client hello. 4334 [Steve Henson] 4335 4336 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 4337 if a DER encoded private key is RSA or DSA traditional format. Changed 4338 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 4339 format DER encoded private key. Newer code should use PKCS#8 format which 4340 has the key type encoded in the ASN1 structure. Added DER private key 4341 support to pkcs8 application. 4342 [Steve Henson] 4343 4344 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 4345 ciphersuites has been selected (as required by the SSL 3/TLS 1 4346 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 4347 is set, we interpret this as a request to violate the specification 4348 (the worst that can happen is a handshake failure, and 'correct' 4349 behaviour would result in a handshake failure anyway). 4350 [Bodo Moeller] 4351 4352 *) In SSL_CTX_add_session, take into account that there might be multiple 4353 SSL_SESSION structures with the same session ID (e.g. when two threads 4354 concurrently obtain them from an external cache). 4355 The internal cache can handle only one SSL_SESSION with a given ID, 4356 so if there's a conflict, we now throw out the old one to achieve 4357 consistency. 4358 [Bodo Moeller] 4359 4360 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 4361 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 4362 some routines that use cipher OIDs: some ciphers do not have OIDs 4363 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 4364 example. 4365 [Steve Henson] 4366 4367 *) Simplify the trust setting structure and code. Now we just have 4368 two sequences of OIDs for trusted and rejected settings. These will 4369 typically have values the same as the extended key usage extension 4370 and any application specific purposes. 4371 4372 The trust checking code now has a default behaviour: it will just 4373 check for an object with the same NID as the passed id. Functions can 4374 be provided to override either the default behaviour or the behaviour 4375 for a given id. SSL client, server and email already have functions 4376 in place for compatibility: they check the NID and also return "trusted" 4377 if the certificate is self signed. 4378 [Steve Henson] 4379 4380 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 4381 traditional format into an EVP_PKEY structure. 4382 [Steve Henson] 4383 4384 *) Add a password callback function PEM_cb() which either prompts for 4385 a password if usr_data is NULL or otherwise assumes it is a null 4386 terminated password. Allow passwords to be passed on command line 4387 environment or config files in a few more utilities. 4388 [Steve Henson] 4389 4390 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 4391 keys. Add some short names for PKCS#8 PBE algorithms and allow them 4392 to be specified on the command line for the pkcs8 and pkcs12 utilities. 4393 Update documentation. 4394 [Steve Henson] 4395 4396 *) Support for ASN1 "NULL" type. This could be handled before by using 4397 ASN1_TYPE but there wasn't any function that would try to read a NULL 4398 and produce an error if it couldn't. For compatibility we also have 4399 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 4400 don't allocate anything because they don't need to. 4401 [Steve Henson] 4402 4403 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 4404 for details. 4405 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 4406 4407 *) Rebuild of the memory allocation routines used by OpenSSL code and 4408 possibly others as well. The purpose is to make an interface that 4409 provide hooks so anyone can build a separate set of allocation and 4410 deallocation routines to be used by OpenSSL, for example memory 4411 pool implementations, or something else, which was previously hard 4412 since Malloc(), Realloc() and Free() were defined as macros having 4413 the values malloc, realloc and free, respectively (except for Win32 4414 compilations). The same is provided for memory debugging code. 4415 OpenSSL already comes with functionality to find memory leaks, but 4416 this gives people a chance to debug other memory problems. 4417 4418 With these changes, a new set of functions and macros have appeared: 4419 4420 CRYPTO_set_mem_debug_functions() [F] 4421 CRYPTO_get_mem_debug_functions() [F] 4422 CRYPTO_dbg_set_options() [F] 4423 CRYPTO_dbg_get_options() [F] 4424 CRYPTO_malloc_debug_init() [M] 4425 4426 The memory debug functions are NULL by default, unless the library 4427 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 4428 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 4429 gives the standard debugging functions that come with OpenSSL) or 4430 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 4431 provided by the library user) must be used. When the standard 4432 debugging functions are used, CRYPTO_dbg_set_options can be used to 4433 request additional information: 4434 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 4435 the CRYPTO_MDEBUG_xxx macro when compiling the library. 4436 4437 Also, things like CRYPTO_set_mem_functions will always give the 4438 expected result (the new set of functions is used for allocation 4439 and deallocation) at all times, regardless of platform and compiler 4440 options. 4441 4442 To finish it up, some functions that were never use in any other 4443 way than through macros have a new API and new semantic: 4444 4445 CRYPTO_dbg_malloc() 4446 CRYPTO_dbg_realloc() 4447 CRYPTO_dbg_free() 4448 4449 All macros of value have retained their old syntax. 4450 [Richard Levitte and Bodo Moeller] 4451 4452 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 4453 ordering of SMIMECapabilities wasn't in "strength order" and there 4454 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 4455 algorithm. 4456 [Steve Henson] 4457 4458 *) Some ASN1 types with illegal zero length encoding (INTEGER, 4459 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 4460 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 4461 4462 *) Merge in my S/MIME library for OpenSSL. This provides a simple 4463 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 4464 functionality to handle multipart/signed properly) and a utility 4465 called 'smime' to call all this stuff. This is based on code I 4466 originally wrote for Celo who have kindly allowed it to be 4467 included in OpenSSL. 4468 [Steve Henson] 4469 4470 *) Add variants des_set_key_checked and des_set_key_unchecked of 4471 des_set_key (aka des_key_sched). Global variable des_check_key 4472 decides which of these is called by des_set_key; this way 4473 des_check_key behaves as it always did, but applications and 4474 the library itself, which was buggy for des_check_key == 1, 4475 have a cleaner way to pick the version they need. 4476 [Bodo Moeller] 4477 4478 *) New function PKCS12_newpass() which changes the password of a 4479 PKCS12 structure. 4480 [Steve Henson] 4481 4482 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 4483 dynamic mix. In both cases the ids can be used as an index into the 4484 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 4485 functions so they accept a list of the field values and the 4486 application doesn't need to directly manipulate the X509_TRUST 4487 structure. 4488 [Steve Henson] 4489 4490 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 4491 need initialising. 4492 [Steve Henson] 4493 4494 *) Modify the way the V3 extension code looks up extensions. This now 4495 works in a similar way to the object code: we have some "standard" 4496 extensions in a static table which is searched with OBJ_bsearch() 4497 and the application can add dynamic ones if needed. The file 4498 crypto/x509v3/ext_dat.h now has the info: this file needs to be 4499 updated whenever a new extension is added to the core code and kept 4500 in ext_nid order. There is a simple program 'tabtest.c' which checks 4501 this. New extensions are not added too often so this file can readily 4502 be maintained manually. 4503 4504 There are two big advantages in doing things this way. The extensions 4505 can be looked up immediately and no longer need to be "added" using 4506 X509V3_add_standard_extensions(): this function now does nothing. 4507 [Side note: I get *lots* of email saying the extension code doesn't 4508 work because people forget to call this function] 4509 Also no dynamic allocation is done unless new extensions are added: 4510 so if we don't add custom extensions there is no need to call 4511 X509V3_EXT_cleanup(). 4512 [Steve Henson] 4513 4514 *) Modify enc utility's salting as follows: make salting the default. Add a 4515 magic header, so unsalted files fail gracefully instead of just decrypting 4516 to garbage. This is because not salting is a big security hole, so people 4517 should be discouraged from doing it. 4518 [Ben Laurie] 4519 4520 *) Fixes and enhancements to the 'x509' utility. It allowed a message 4521 digest to be passed on the command line but it only used this 4522 parameter when signing a certificate. Modified so all relevant 4523 operations are affected by the digest parameter including the 4524 -fingerprint and -x509toreq options. Also -x509toreq choked if a 4525 DSA key was used because it didn't fix the digest. 4526 [Steve Henson] 4527 4528 *) Initial certificate chain verify code. Currently tests the untrusted 4529 certificates for consistency with the verify purpose (which is set 4530 when the X509_STORE_CTX structure is set up) and checks the pathlength. 4531 4532 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 4533 this is because it will reject chains with invalid extensions whereas 4534 every previous version of OpenSSL and SSLeay made no checks at all. 4535 4536 Trust code: checks the root CA for the relevant trust settings. Trust 4537 settings have an initial value consistent with the verify purpose: e.g. 4538 if the verify purpose is for SSL client use it expects the CA to be 4539 trusted for SSL client use. However the default value can be changed to 4540 permit custom trust settings: one example of this would be to only trust 4541 certificates from a specific "secure" set of CAs. 4542 4543 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 4544 which should be used for version portability: especially since the 4545 verify structure is likely to change more often now. 4546 4547 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 4548 to set them. If not set then assume SSL clients will verify SSL servers 4549 and vice versa. 4550 4551 Two new options to the verify program: -untrusted allows a set of 4552 untrusted certificates to be passed in and -purpose which sets the 4553 intended purpose of the certificate. If a purpose is set then the 4554 new chain verify code is used to check extension consistency. 4555 [Steve Henson] 4556 4557 *) Support for the authority information access extension. 4558 [Steve Henson] 4559 4560 *) Modify RSA and DSA PEM read routines to transparently handle 4561 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 4562 public keys in a format compatible with certificate 4563 SubjectPublicKeyInfo structures. Unfortunately there were already 4564 functions called *_PublicKey_* which used various odd formats so 4565 these are retained for compatibility: however the DSA variants were 4566 never in a public release so they have been deleted. Changed dsa/rsa 4567 utilities to handle the new format: note no releases ever handled public 4568 keys so we should be OK. 4569 4570 The primary motivation for this change is to avoid the same fiasco 4571 that dogs private keys: there are several incompatible private key 4572 formats some of which are standard and some OpenSSL specific and 4573 require various evil hacks to allow partial transparent handling and 4574 even then it doesn't work with DER formats. Given the option anything 4575 other than PKCS#8 should be dumped: but the other formats have to 4576 stay in the name of compatibility. 4577 4578 With public keys and the benefit of hindsight one standard format 4579 is used which works with EVP_PKEY, RSA or DSA structures: though 4580 it clearly returns an error if you try to read the wrong kind of key. 4581 4582 Added a -pubkey option to the 'x509' utility to output the public key. 4583 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 4584 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 4585 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 4586 that do the same as the EVP_PKEY_assign_*() except they up the 4587 reference count of the added key (they don't "swallow" the 4588 supplied key). 4589 [Steve Henson] 4590 4591 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 4592 CRLs would fail if the file contained no certificates or no CRLs: 4593 added a new function to read in both types and return the number 4594 read: this means that if none are read it will be an error. The 4595 DER versions of the certificate and CRL reader would always fail 4596 because it isn't possible to mix certificates and CRLs in DER format 4597 without choking one or the other routine. Changed this to just read 4598 a certificate: this is the best we can do. Also modified the code 4599 in apps/verify.c to take notice of return codes: it was previously 4600 attempting to read in certificates from NULL pointers and ignoring 4601 any errors: this is one reason why the cert and CRL reader seemed 4602 to work. It doesn't check return codes from the default certificate 4603 routines: these may well fail if the certificates aren't installed. 4604 [Steve Henson] 4605 4606 *) Code to support otherName option in GeneralName. 4607 [Steve Henson] 4608 4609 *) First update to verify code. Change the verify utility 4610 so it warns if it is passed a self signed certificate: 4611 for consistency with the normal behaviour. X509_verify 4612 has been modified to it will now verify a self signed 4613 certificate if *exactly* the same certificate appears 4614 in the store: it was previously impossible to trust a 4615 single self signed certificate. This means that: 4616 openssl verify ss.pem 4617 now gives a warning about a self signed certificate but 4618 openssl verify -CAfile ss.pem ss.pem 4619 is OK. 4620 [Steve Henson] 4621 4622 *) For servers, store verify_result in SSL_SESSION data structure 4623 (and add it to external session representation). 4624 This is needed when client certificate verifications fails, 4625 but an application-provided verification callback (set by 4626 SSL_CTX_set_cert_verify_callback) allows accepting the session 4627 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 4628 but returns 1): When the session is reused, we have to set 4629 ssl->verify_result to the appropriate error code to avoid 4630 security holes. 4631 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 4632 4633 *) Fix a bug in the new PKCS#7 code: it didn't consider the 4634 case in PKCS7_dataInit() where the signed PKCS7 structure 4635 didn't contain any existing data because it was being created. 4636 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 4637 4638 *) Add a salt to the key derivation routines in enc.c. This 4639 forms the first 8 bytes of the encrypted file. Also add a 4640 -S option to allow a salt to be input on the command line. 4641 [Steve Henson] 4642 4643 *) New function X509_cmp(). Oddly enough there wasn't a function 4644 to compare two certificates. We do this by working out the SHA1 4645 hash and comparing that. X509_cmp() will be needed by the trust 4646 code. 4647 [Steve Henson] 4648 4649 *) SSL_get1_session() is like SSL_get_session(), but increments 4650 the reference count in the SSL_SESSION returned. 4651 [Geoff Thorpe <geoff@eu.c2.net>] 4652 4653 *) Fix for 'req': it was adding a null to request attributes. 4654 Also change the X509_LOOKUP and X509_INFO code to handle 4655 certificate auxiliary information. 4656 [Steve Henson] 4657 4658 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 4659 the 'enc' command. 4660 [Steve Henson] 4661 4662 *) Add the possibility to add extra information to the memory leak 4663 detecting output, to form tracebacks, showing from where each 4664 allocation was originated: CRYPTO_push_info("constant string") adds 4665 the string plus current file name and line number to a per-thread 4666 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 4667 is like calling CYRPTO_pop_info() until the stack is empty. 4668 Also updated memory leak detection code to be multi-thread-safe. 4669 [Richard Levitte] 4670 4671 *) Add options -text and -noout to pkcs7 utility and delete the 4672 encryption options which never did anything. Update docs. 4673 [Steve Henson] 4674 4675 *) Add options to some of the utilities to allow the pass phrase 4676 to be included on either the command line (not recommended on 4677 OSes like Unix) or read from the environment. Update the 4678 manpages and fix a few bugs. 4679 [Steve Henson] 4680 4681 *) Add a few manpages for some of the openssl commands. 4682 [Steve Henson] 4683 4684 *) Fix the -revoke option in ca. It was freeing up memory twice, 4685 leaking and not finding already revoked certificates. 4686 [Steve Henson] 4687 4688 *) Extensive changes to support certificate auxiliary information. 4689 This involves the use of X509_CERT_AUX structure and X509_AUX 4690 functions. An X509_AUX function such as PEM_read_X509_AUX() 4691 can still read in a certificate file in the usual way but it 4692 will also read in any additional "auxiliary information". By 4693 doing things this way a fair degree of compatibility can be 4694 retained: existing certificates can have this information added 4695 using the new 'x509' options. 4696 4697 Current auxiliary information includes an "alias" and some trust 4698 settings. The trust settings will ultimately be used in enhanced 4699 certificate chain verification routines: currently a certificate 4700 can only be trusted if it is self signed and then it is trusted 4701 for all purposes. 4702 [Steve Henson] 4703 4704 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 4705 The problem was that one of the replacement routines had not been working 4706 since SSLeay releases. For now the offending routine has been replaced 4707 with non-optimised assembler. Even so, this now gives around 95% 4708 performance improvement for 1024 bit RSA signs. 4709 [Mark Cox] 4710 4711 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 4712 handling. Most clients have the effective key size in bits equal to 4713 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 4714 A few however don't do this and instead use the size of the decrypted key 4715 to determine the RC2 key length and the AlgorithmIdentifier to determine 4716 the effective key length. In this case the effective key length can still 4717 be 40 bits but the key length can be 168 bits for example. This is fixed 4718 by manually forcing an RC2 key into the EVP_PKEY structure because the 4719 EVP code can't currently handle unusual RC2 key sizes: it always assumes 4720 the key length and effective key length are equal. 4721 [Steve Henson] 4722 4723 *) Add a bunch of functions that should simplify the creation of 4724 X509_NAME structures. Now you should be able to do: 4725 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 4726 and have it automatically work out the correct field type and fill in 4727 the structures. The more adventurous can try: 4728 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 4729 and it will (hopefully) work out the correct multibyte encoding. 4730 [Steve Henson] 4731 4732 *) Change the 'req' utility to use the new field handling and multibyte 4733 copy routines. Before the DN field creation was handled in an ad hoc 4734 way in req, ca, and x509 which was rather broken and didn't support 4735 BMPStrings or UTF8Strings. Since some software doesn't implement 4736 BMPStrings or UTF8Strings yet, they can be enabled using the config file 4737 using the dirstring_type option. See the new comment in the default 4738 openssl.cnf for more info. 4739 [Steve Henson] 4740 4741 *) Make crypto/rand/md_rand.c more robust: 4742 - Assure unique random numbers after fork(). 4743 - Make sure that concurrent threads access the global counter and 4744 md serializably so that we never lose entropy in them 4745 or use exactly the same state in multiple threads. 4746 Access to the large state is not always serializable because 4747 the additional locking could be a performance killer, and 4748 md should be large enough anyway. 4749 [Bodo Moeller] 4750 4751 *) New file apps/app_rand.c with commonly needed functionality 4752 for handling the random seed file. 4753 4754 Use the random seed file in some applications that previously did not: 4755 ca, 4756 dsaparam -genkey (which also ignored its '-rand' option), 4757 s_client, 4758 s_server, 4759 x509 (when signing). 4760 Except on systems with /dev/urandom, it is crucial to have a random 4761 seed file at least for key creation, DSA signing, and for DH exchanges; 4762 for RSA signatures we could do without one. 4763 4764 gendh and gendsa (unlike genrsa) used to read only the first byte 4765 of each file listed in the '-rand' option. The function as previously 4766 found in genrsa is now in app_rand.c and is used by all programs 4767 that support '-rand'. 4768 [Bodo Moeller] 4769 4770 *) In RAND_write_file, use mode 0600 for creating files; 4771 don't just chmod when it may be too late. 4772 [Bodo Moeller] 4773 4774 *) Report an error from X509_STORE_load_locations 4775 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 4776 [Bill Perry] 4777 4778 *) New function ASN1_mbstring_copy() this copies a string in either 4779 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 4780 into an ASN1_STRING type. A mask of permissible types is passed 4781 and it chooses the "minimal" type to use or an error if not type 4782 is suitable. 4783 [Steve Henson] 4784 4785 *) Add function equivalents to the various macros in asn1.h. The old 4786 macros are retained with an M_ prefix. Code inside the library can 4787 use the M_ macros. External code (including the openssl utility) 4788 should *NOT* in order to be "shared library friendly". 4789 [Steve Henson] 4790 4791 *) Add various functions that can check a certificate's extensions 4792 to see if it usable for various purposes such as SSL client, 4793 server or S/MIME and CAs of these types. This is currently 4794 VERY EXPERIMENTAL but will ultimately be used for certificate chain 4795 verification. Also added a -purpose flag to x509 utility to 4796 print out all the purposes. 4797 [Steve Henson] 4798 4799 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 4800 functions. 4801 [Steve Henson] 4802 4803 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 4804 for, obtain and decode and extension and obtain its critical flag. 4805 This allows all the necessary extension code to be handled in a 4806 single function call. 4807 [Steve Henson] 4808 4809 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 4810 platforms. See crypto/rc4/rc4_enc.c for further details. 4811 [Andy Polyakov] 4812 4813 *) New -noout option to asn1parse. This causes no output to be produced 4814 its main use is when combined with -strparse and -out to extract data 4815 from a file (which may not be in ASN.1 format). 4816 [Steve Henson] 4817 4818 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 4819 when producing the local key id. 4820 [Richard Levitte <levitte@stacken.kth.se>] 4821 4822 *) New option -dhparam in s_server. This allows a DH parameter file to be 4823 stated explicitly. If it is not stated then it tries the first server 4824 certificate file. The previous behaviour hard coded the filename 4825 "server.pem". 4826 [Steve Henson] 4827 4828 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 4829 a public key to be input or output. For example: 4830 openssl rsa -in key.pem -pubout -out pubkey.pem 4831 Also added necessary DSA public key functions to handle this. 4832 [Steve Henson] 4833 4834 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 4835 in the message. This was handled by allowing 4836 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 4837 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 4838 4839 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 4840 to the end of the strings whereas this didn't. This would cause problems 4841 if strings read with d2i_ASN1_bytes() were later modified. 4842 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 4843 4844 *) Fix for base64 decode bug. When a base64 bio reads only one line of 4845 data and it contains EOF it will end up returning an error. This is 4846 caused by input 46 bytes long. The cause is due to the way base64 4847 BIOs find the start of base64 encoded data. They do this by trying a 4848 trial decode on each line until they find one that works. When they 4849 do a flag is set and it starts again knowing it can pass all the 4850 data directly through the decoder. Unfortunately it doesn't reset 4851 the context it uses. This means that if EOF is reached an attempt 4852 is made to pass two EOFs through the context and this causes the 4853 resulting error. This can also cause other problems as well. As is 4854 usual with these problems it takes *ages* to find and the fix is 4855 trivial: move one line. 4856 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 4857 4858 *) Ugly workaround to get s_client and s_server working under Windows. The 4859 old code wouldn't work because it needed to select() on sockets and the 4860 tty (for keypresses and to see if data could be written). Win32 only 4861 supports select() on sockets so we select() with a 1s timeout on the 4862 sockets and then see if any characters are waiting to be read, if none 4863 are present then we retry, we also assume we can always write data to 4864 the tty. This isn't nice because the code then blocks until we've 4865 received a complete line of data and it is effectively polling the 4866 keyboard at 1s intervals: however it's quite a bit better than not 4867 working at all :-) A dedicated Windows application might handle this 4868 with an event loop for example. 4869 [Steve Henson] 4870 4871 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 4872 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 4873 will be called when RSA_sign() and RSA_verify() are used. This is useful 4874 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 4875 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 4876 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 4877 This necessitated the support of an extra signature type NID_md5_sha1 4878 for SSL signatures and modifications to the SSL library to use it instead 4879 of calling RSA_public_decrypt() and RSA_private_encrypt(). 4880 [Steve Henson] 4881 4882 *) Add new -verify -CAfile and -CApath options to the crl program, these 4883 will lookup a CRL issuers certificate and verify the signature in a 4884 similar way to the verify program. Tidy up the crl program so it 4885 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 4886 less strict. It will now permit CRL extensions even if it is not 4887 a V2 CRL: this will allow it to tolerate some broken CRLs. 4888 [Steve Henson] 4889 4890 *) Initialize all non-automatic variables each time one of the openssl 4891 sub-programs is started (this is necessary as they may be started 4892 multiple times from the "OpenSSL>" prompt). 4893 [Lennart Bang, Bodo Moeller] 4894 4895 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 4896 removing all other RSA functionality (this is what NO_RSA does). This 4897 is so (for example) those in the US can disable those operations covered 4898 by the RSA patent while allowing storage and parsing of RSA keys and RSA 4899 key generation. 4900 [Steve Henson] 4901 4902 *) Non-copying interface to BIO pairs. 4903 (still largely untested) 4904 [Bodo Moeller] 4905 4906 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 4907 ASCII string. This was handled independently in various places before. 4908 [Steve Henson] 4909 4910 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 4911 UTF8 strings a character at a time. 4912 [Steve Henson] 4913 4914 *) Use client_version from client hello to select the protocol 4915 (s23_srvr.c) and for RSA client key exchange verification 4916 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 4917 [Bodo Moeller] 4918 4919 *) Add various utility functions to handle SPKACs, these were previously 4920 handled by poking round in the structure internals. Added new function 4921 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 4922 print, verify and generate SPKACs. Based on an original idea from 4923 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 4924 [Steve Henson] 4925 4926 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 4927 [Andy Polyakov] 4928 4929 *) Allow the config file extension section to be overwritten on the 4930 command line. Based on an original idea from Massimiliano Pala 4931 <madwolf@comune.modena.it>. The new option is called -extensions 4932 and can be applied to ca, req and x509. Also -reqexts to override 4933 the request extensions in req and -crlexts to override the crl extensions 4934 in ca. 4935 [Steve Henson] 4936 4937 *) Add new feature to the SPKAC handling in ca. Now you can include 4938 the same field multiple times by preceding it by "XXXX." for example: 4939 1.OU="Unit name 1" 4940 2.OU="Unit name 2" 4941 this is the same syntax as used in the req config file. 4942 [Steve Henson] 4943 4944 *) Allow certificate extensions to be added to certificate requests. These 4945 are specified in a 'req_extensions' option of the req section of the 4946 config file. They can be printed out with the -text option to req but 4947 are otherwise ignored at present. 4948 [Steve Henson] 4949 4950 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 4951 data read consists of only the final block it would not decrypted because 4952 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 4953 A misplaced 'break' also meant the decrypted final block might not be 4954 copied until the next read. 4955 [Steve Henson] 4956 4957 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 4958 a few extra parameters to the DH structure: these will be useful if 4959 for example we want the value of 'q' or implement X9.42 DH. 4960 [Steve Henson] 4961 4962 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 4963 provides hooks that allow the default DSA functions or functions on a 4964 "per key" basis to be replaced. This allows hardware acceleration and 4965 hardware key storage to be handled without major modification to the 4966 library. Also added low level modexp hooks and CRYPTO_EX structure and 4967 associated functions. 4968 [Steve Henson] 4969 4970 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 4971 as "read only": it can't be written to and the buffer it points to will 4972 not be freed. Reading from a read only BIO is much more efficient than 4973 a normal memory BIO. This was added because there are several times when 4974 an area of memory needs to be read from a BIO. The previous method was 4975 to create a memory BIO and write the data to it, this results in two 4976 copies of the data and an O(n^2) reading algorithm. There is a new 4977 function BIO_new_mem_buf() which creates a read only memory BIO from 4978 an area of memory. Also modified the PKCS#7 routines to use read only 4979 memory BIOs. 4980 [Steve Henson] 4981 4982 *) Bugfix: ssl23_get_client_hello did not work properly when called in 4983 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 4984 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 4985 but a retry condition occured while trying to read the rest. 4986 [Bodo Moeller] 4987 4988 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 4989 NID_pkcs7_encrypted by default: this was wrong since this should almost 4990 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 4991 the encrypted data type: this is a more sensible place to put it and it 4992 allows the PKCS#12 code to be tidied up that duplicated this 4993 functionality. 4994 [Steve Henson] 4995 4996 *) Changed obj_dat.pl script so it takes its input and output files on 4997 the command line. This should avoid shell escape redirection problems 4998 under Win32. 4999 [Steve Henson] 5000 5001 *) Initial support for certificate extension requests, these are included 5002 in things like Xenroll certificate requests. Included functions to allow 5003 extensions to be obtained and added. 5004 [Steve Henson] 5005 5006 *) -crlf option to s_client and s_server for sending newlines as 5007 CRLF (as required by many protocols). 5008 [Bodo Moeller] 5009 5010 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 5011 5012 *) Install libRSAglue.a when OpenSSL is built with RSAref. 5013 [Ralf S. Engelschall] 5014 5015 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 5016 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 5017 5018 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 5019 program. 5020 [Steve Henson] 5021 5022 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 5023 DH parameters/keys (q is lost during that conversion, but the resulting 5024 DH parameters contain its length). 5025 5026 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 5027 much faster than DH_generate_parameters (which creates parameters 5028 where p = 2*q + 1), and also the smaller q makes DH computations 5029 much more efficient (160-bit exponentiation instead of 1024-bit 5030 exponentiation); so this provides a convenient way to support DHE 5031 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 5032 utter importance to use 5033 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 5034 or 5035 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 5036 when such DH parameters are used, because otherwise small subgroup 5037 attacks may become possible! 5038 [Bodo Moeller] 5039 5040 *) Avoid memory leak in i2d_DHparams. 5041 [Bodo Moeller] 5042 5043 *) Allow the -k option to be used more than once in the enc program: 5044 this allows the same encrypted message to be read by multiple recipients. 5045 [Steve Henson] 5046 5047 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 5048 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 5049 it will always use the numerical form of the OID, even if it has a short 5050 or long name. 5051 [Steve Henson] 5052 5053 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 5054 method only got called if p,q,dmp1,dmq1,iqmp components were present, 5055 otherwise bn_mod_exp was called. In the case of hardware keys for example 5056 no private key components need be present and it might store extra data 5057 in the RSA structure, which cannot be accessed from bn_mod_exp. 5058 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 5059 private key operations. 5060 [Steve Henson] 5061 5062 *) Added support for SPARC Linux. 5063 [Andy Polyakov] 5064 5065 *) pem_password_cb function type incompatibly changed from 5066 typedef int pem_password_cb(char *buf, int size, int rwflag); 5067 to 5068 ....(char *buf, int size, int rwflag, void *userdata); 5069 so that applications can pass data to their callbacks: 5070 The PEM[_ASN1]_{read,write}... functions and macros now take an 5071 additional void * argument, which is just handed through whenever 5072 the password callback is called. 5073 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 5074 5075 New function SSL_CTX_set_default_passwd_cb_userdata. 5076 5077 Compatibility note: As many C implementations push function arguments 5078 onto the stack in reverse order, the new library version is likely to 5079 interoperate with programs that have been compiled with the old 5080 pem_password_cb definition (PEM_whatever takes some data that 5081 happens to be on the stack as its last argument, and the callback 5082 just ignores this garbage); but there is no guarantee whatsoever that 5083 this will work. 5084 5085 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 5086 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 5087 problems not only on Windows, but also on some Unix platforms. 5088 To avoid problematic command lines, these definitions are now in an 5089 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 5090 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 5091 [Bodo Moeller] 5092 5093 *) MIPS III/IV assembler module is reimplemented. 5094 [Andy Polyakov] 5095 5096 *) More DES library cleanups: remove references to srand/rand and 5097 delete an unused file. 5098 [Ulf M�ller] 5099 5100 *) Add support for the the free Netwide assembler (NASM) under Win32, 5101 since not many people have MASM (ml) and it can be hard to obtain. 5102 This is currently experimental but it seems to work OK and pass all 5103 the tests. Check out INSTALL.W32 for info. 5104 [Steve Henson] 5105 5106 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 5107 without temporary keys kept an extra copy of the server key, 5108 and connections with temporary keys did not free everything in case 5109 of an error. 5110 [Bodo Moeller] 5111 5112 *) New function RSA_check_key and new openssl rsa option -check 5113 for verifying the consistency of RSA keys. 5114 [Ulf Moeller, Bodo Moeller] 5115 5116 *) Various changes to make Win32 compile work: 5117 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 5118 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 5119 comparison" warnings. 5120 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 5121 [Steve Henson] 5122 5123 *) Add a debugging option to PKCS#5 v2 key generation function: when 5124 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 5125 derived keys are printed to stderr. 5126 [Steve Henson] 5127 5128 *) Copy the flags in ASN1_STRING_dup(). 5129 [Roman E. Pavlov <pre@mo.msk.ru>] 5130 5131 *) The x509 application mishandled signing requests containing DSA 5132 keys when the signing key was also DSA and the parameters didn't match. 5133 5134 It was supposed to omit the parameters when they matched the signing key: 5135 the verifying software was then supposed to automatically use the CA's 5136 parameters if they were absent from the end user certificate. 5137 5138 Omitting parameters is no longer recommended. The test was also 5139 the wrong way round! This was probably due to unusual behaviour in 5140 EVP_cmp_parameters() which returns 1 if the parameters match. 5141 This meant that parameters were omitted when they *didn't* match and 5142 the certificate was useless. Certificates signed with 'ca' didn't have 5143 this bug. 5144 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 5145 5146 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 5147 The interface is as follows: 5148 Applications can use 5149 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 5150 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 5151 "off" is now the default. 5152 The library internally uses 5153 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 5154 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 5155 to disable memory-checking temporarily. 5156 5157 Some inconsistent states that previously were possible (and were 5158 even the default) are now avoided. 5159 5160 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 5161 with each memory chunk allocated; this is occasionally more helpful 5162 than just having a counter. 5163 5164 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 5165 5166 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 5167 extensions. 5168 [Bodo Moeller] 5169 5170 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 5171 which largely parallels "options", but is for changing API behaviour, 5172 whereas "options" are about protocol behaviour. 5173 Initial "mode" flags are: 5174 5175 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 5176 a single record has been written. 5177 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 5178 retries use the same buffer location. 5179 (But all of the contents must be 5180 copied!) 5181 [Bodo Moeller] 5182 5183 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 5184 worked. 5185 5186 *) Fix problems with no-hmac etc. 5187 [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 5188 5189 *) New functions RSA_get_default_method(), RSA_set_method() and 5190 RSA_get_method(). These allows replacement of RSA_METHODs without having 5191 to mess around with the internals of an RSA structure. 5192 [Steve Henson] 5193 5194 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 5195 Also really enable memory leak checks in openssl.c and in some 5196 test programs. 5197 [Chad C. Mulligan, Bodo Moeller] 5198 5199 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 5200 up the length of negative integers. This has now been simplified to just 5201 store the length when it is first determined and use it later, rather 5202 than trying to keep track of where data is copied and updating it to 5203 point to the end. 5204 [Steve Henson, reported by Brien Wheeler 5205 <bwheeler@authentica-security.com>] 5206 5207 *) Add a new function PKCS7_signatureVerify. This allows the verification 5208 of a PKCS#7 signature but with the signing certificate passed to the 5209 function itself. This contrasts with PKCS7_dataVerify which assumes the 5210 certificate is present in the PKCS#7 structure. This isn't always the 5211 case: certificates can be omitted from a PKCS#7 structure and be 5212 distributed by "out of band" means (such as a certificate database). 5213 [Steve Henson] 5214 5215 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 5216 function prototypes in pem.h, also change util/mkdef.pl to add the 5217 necessary function names. 5218 [Steve Henson] 5219 5220 *) mk1mf.pl (used by Windows builds) did not properly read the 5221 options set by Configure in the top level Makefile, and Configure 5222 was not even able to write more than one option correctly. 5223 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 5224 [Bodo Moeller] 5225 5226 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 5227 file to be loaded from a BIO or FILE pointer. The BIO version will 5228 for example allow memory BIOs to contain config info. 5229 [Steve Henson] 5230 5231 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 5232 Whoever hopes to achieve shared-library compatibility across versions 5233 must use this, not the compile-time macro. 5234 (Exercise 0.9.4: Which is the minimum library version required by 5235 such programs?) 5236 Note: All this applies only to multi-threaded programs, others don't 5237 need locks. 5238 [Bodo Moeller] 5239 5240 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 5241 through a BIO pair triggered the default case, i.e. 5242 SSLerr(...,SSL_R_UNKNOWN_STATE). 5243 [Bodo Moeller] 5244 5245 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 5246 can use the SSL library even if none of the specific BIOs is 5247 appropriate. 5248 [Bodo Moeller] 5249 5250 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 5251 for the encoded length. 5252 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 5253 5254 *) Add initial documentation of the X509V3 functions. 5255 [Steve Henson] 5256 5257 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 5258 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 5259 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 5260 secure PKCS#8 private key format with a high iteration count. 5261 [Steve Henson] 5262 5263 *) Fix determination of Perl interpreter: A perl or perl5 5264 _directory_ in $PATH was also accepted as the interpreter. 5265 [Ralf S. Engelschall] 5266 5267 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 5268 wrong with it but it was very old and did things like calling 5269 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 5270 unusual formatting. 5271 [Steve Henson] 5272 5273 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 5274 to use the new extension code. 5275 [Steve Henson] 5276 5277 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 5278 with macros. This should make it easier to change their form, add extra 5279 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 5280 constant. 5281 [Steve Henson] 5282 5283 *) Add to configuration table a new entry that can specify an alternative 5284 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 5285 according to Mark Crispin <MRC@Panda.COM>. 5286 [Bodo Moeller] 5287 5288#if 0 5289 *) DES CBC did not update the IV. Weird. 5290 [Ben Laurie] 5291#else 5292 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 5293 Changing the behaviour of the former might break existing programs -- 5294 where IV updating is needed, des_ncbc_encrypt can be used. 5295#endif 5296 5297 *) When bntest is run from "make test" it drives bc to check its 5298 calculations, as well as internally checking them. If an internal check 5299 fails, it needs to cause bc to give a non-zero result or make test carries 5300 on without noticing the failure. Fixed. 5301 [Ben Laurie] 5302 5303 *) DES library cleanups. 5304 [Ulf M�ller] 5305 5306 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 5307 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 5308 ciphers. NOTE: although the key derivation function has been verified 5309 against some published test vectors it has not been extensively tested 5310 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 5311 of v2.0. 5312 [Steve Henson] 5313 5314 *) Instead of "mkdir -p", which is not fully portable, use new 5315 Perl script "util/mkdir-p.pl". 5316 [Bodo Moeller] 5317 5318 *) Rewrite the way password based encryption (PBE) is handled. It used to 5319 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 5320 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 5321 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 5322 the 'parameter' field of the AlgorithmIdentifier is passed to the 5323 underlying key generation function so it must do its own ASN1 parsing. 5324 This has also changed the EVP_PBE_CipherInit() function which now has a 5325 'parameter' argument instead of literal salt and iteration count values 5326 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 5327 [Steve Henson] 5328 5329 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 5330 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 5331 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 5332 KEY" because this clashed with PKCS#8 unencrypted string. Since this 5333 value was just used as a "magic string" and not used directly its 5334 value doesn't matter. 5335 [Steve Henson] 5336 5337 *) Introduce some semblance of const correctness to BN. Shame C doesn't 5338 support mutable. 5339 [Ben Laurie] 5340 5341 *) "linux-sparc64" configuration (ultrapenguin). 5342 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 5343 "linux-sparc" configuration. 5344 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 5345 5346 *) config now generates no-xxx options for missing ciphers. 5347 [Ulf M�ller] 5348 5349 *) Support the EBCDIC character set (work in progress). 5350 File ebcdic.c not yet included because it has a different license. 5351 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 5352 5353 *) Support BS2000/OSD-POSIX. 5354 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 5355 5356 *) Make callbacks for key generation use void * instead of char *. 5357 [Ben Laurie] 5358 5359 *) Make S/MIME samples compile (not yet tested). 5360 [Ben Laurie] 5361 5362 *) Additional typesafe stacks. 5363 [Ben Laurie] 5364 5365 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 5366 [Bodo Moeller] 5367 5368 5369 Changes between 0.9.3 and 0.9.3a [29 May 1999] 5370 5371 *) New configuration variant "sco5-gcc". 5372 5373 *) Updated some demos. 5374 [Sean O Riordain, Wade Scholine] 5375 5376 *) Add missing BIO_free at exit of pkcs12 application. 5377 [Wu Zhigang] 5378 5379 *) Fix memory leak in conf.c. 5380 [Steve Henson] 5381 5382 *) Updates for Win32 to assembler version of MD5. 5383 [Steve Henson] 5384 5385 *) Set #! path to perl in apps/der_chop to where we found it 5386 instead of using a fixed path. 5387 [Bodo Moeller] 5388 5389 *) SHA library changes for irix64-mips4-cc. 5390 [Andy Polyakov] 5391 5392 *) Improvements for VMS support. 5393 [Richard Levitte] 5394 5395 5396 Changes between 0.9.2b and 0.9.3 [24 May 1999] 5397 5398 *) Bignum library bug fix. IRIX 6 passes "make test" now! 5399 This also avoids the problems with SC4.2 and unpatched SC5. 5400 [Andy Polyakov <appro@fy.chalmers.se>] 5401 5402 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 5403 These are required because of the typesafe stack would otherwise break 5404 existing code. If old code used a structure member which used to be STACK 5405 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 5406 sk_num or sk_value it would produce an error because the num, data members 5407 are not present in STACK_OF. Now it just produces a warning. sk_set 5408 replaces the old method of assigning a value to sk_value 5409 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 5410 that does this will no longer work (and should use sk_set instead) but 5411 this could be regarded as a "questionable" behaviour anyway. 5412 [Steve Henson] 5413 5414 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 5415 correctly handle encrypted S/MIME data. 5416 [Steve Henson] 5417 5418 *) Change type of various DES function arguments from des_cblock 5419 (which means, in function argument declarations, pointer to char) 5420 to des_cblock * (meaning pointer to array with 8 char elements), 5421 which allows the compiler to do more typechecking; it was like 5422 that back in SSLeay, but with lots of ugly casts. 5423 5424 Introduce new type const_des_cblock. 5425 [Bodo Moeller] 5426 5427 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 5428 problems: find RecipientInfo structure that matches recipient certificate 5429 and initialise the ASN1 structures properly based on passed cipher. 5430 [Steve Henson] 5431 5432 *) Belatedly make the BN tests actually check the results. 5433 [Ben Laurie] 5434 5435 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 5436 to and from BNs: it was completely broken. New compilation option 5437 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 5438 key elements as negative integers. 5439 [Steve Henson] 5440 5441 *) Reorganize and speed up MD5. 5442 [Andy Polyakov <appro@fy.chalmers.se>] 5443 5444 *) VMS support. 5445 [Richard Levitte <richard@levitte.org>] 5446 5447 *) New option -out to asn1parse to allow the parsed structure to be 5448 output to a file. This is most useful when combined with the -strparse 5449 option to examine the output of things like OCTET STRINGS. 5450 [Steve Henson] 5451 5452 *) Make SSL library a little more fool-proof by not requiring any longer 5453 that SSL_set_{accept,connect}_state be called before 5454 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 5455 in many applications because usually everything *appeared* to work as 5456 intended anyway -- now it really works as intended). 5457 [Bodo Moeller] 5458 5459 *) Move openssl.cnf out of lib/. 5460 [Ulf M�ller] 5461 5462 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 5463 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 5464 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 5465 [Ralf S. Engelschall] 5466 5467 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 5468 handle PKCS#7 enveloped data properly. 5469 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 5470 5471 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 5472 copying pointers. The cert_st handling is changed by this in 5473 various ways (and thus what used to be known as ctx->default_cert 5474 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 5475 any longer when s->cert does not give us what we need). 5476 ssl_cert_instantiate becomes obsolete by this change. 5477 As soon as we've got the new code right (possibly it already is?), 5478 we have solved a couple of bugs of the earlier code where s->cert 5479 was used as if it could not have been shared with other SSL structures. 5480 5481 Note that using the SSL API in certain dirty ways now will result 5482 in different behaviour than observed with earlier library versions: 5483 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 5484 does not influence s as it used to. 5485 5486 In order to clean up things more thoroughly, inside SSL_SESSION 5487 we don't use CERT any longer, but a new structure SESS_CERT 5488 that holds per-session data (if available); currently, this is 5489 the peer's certificate chain and, for clients, the server's certificate 5490 and temporary key. CERT holds only those values that can have 5491 meaningful defaults in an SSL_CTX. 5492 [Bodo Moeller] 5493 5494 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 5495 from the internal representation. Various PKCS#7 fixes: remove some 5496 evil casts and set the enc_dig_alg field properly based on the signing 5497 key type. 5498 [Steve Henson] 5499 5500 *) Allow PKCS#12 password to be set from the command line or the 5501 environment. Let 'ca' get its config file name from the environment 5502 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 5503 and 'x509'). 5504 [Steve Henson] 5505 5506 *) Allow certificate policies extension to use an IA5STRING for the 5507 organization field. This is contrary to the PKIX definition but 5508 VeriSign uses it and IE5 only recognises this form. Document 'x509' 5509 extension option. 5510 [Steve Henson] 5511 5512 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 5513 without disallowing inline assembler and the like for non-pedantic builds. 5514 [Ben Laurie] 5515 5516 *) Support Borland C++ builder. 5517 [Janez Jere <jj@void.si>, modified by Ulf M�ller] 5518 5519 *) Support Mingw32. 5520 [Ulf M�ller] 5521 5522 *) SHA-1 cleanups and performance enhancements. 5523 [Andy Polyakov <appro@fy.chalmers.se>] 5524 5525 *) Sparc v8plus assembler for the bignum library. 5526 [Andy Polyakov <appro@fy.chalmers.se>] 5527 5528 *) Accept any -xxx and +xxx compiler options in Configure. 5529 [Ulf M�ller] 5530 5531 *) Update HPUX configuration. 5532 [Anonymous] 5533 5534 *) Add missing sk_<type>_unshift() function to safestack.h 5535 [Ralf S. Engelschall] 5536 5537 *) New function SSL_CTX_use_certificate_chain_file that sets the 5538 "extra_cert"s in addition to the certificate. (This makes sense 5539 only for "PEM" format files, as chains as a whole are not 5540 DER-encoded.) 5541 [Bodo Moeller] 5542 5543 *) Support verify_depth from the SSL API. 5544 x509_vfy.c had what can be considered an off-by-one-error: 5545 Its depth (which was not part of the external interface) 5546 was actually counting the number of certificates in a chain; 5547 now it really counts the depth. 5548 [Bodo Moeller] 5549 5550 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 5551 instead of X509err, which often resulted in confusing error 5552 messages since the error codes are not globally unique 5553 (e.g. an alleged error in ssl3_accept when a certificate 5554 didn't match the private key). 5555 5556 *) New function SSL_CTX_set_session_id_context that allows to set a default 5557 value (so that you don't need SSL_set_session_id_context for each 5558 connection using the SSL_CTX). 5559 [Bodo Moeller] 5560 5561 *) OAEP decoding bug fix. 5562 [Ulf M�ller] 5563 5564 *) Support INSTALL_PREFIX for package builders, as proposed by 5565 David Harris. 5566 [Bodo Moeller] 5567 5568 *) New Configure options "threads" and "no-threads". For systems 5569 where the proper compiler options are known (currently Solaris 5570 and Linux), "threads" is the default. 5571 [Bodo Moeller] 5572 5573 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 5574 [Bodo Moeller] 5575 5576 *) Install various scripts to $(OPENSSLDIR)/misc, not to 5577 $(INSTALLTOP)/bin -- they shouldn't clutter directories 5578 such as /usr/local/bin. 5579 [Bodo Moeller] 5580 5581 *) "make linux-shared" to build shared libraries. 5582 [Niels Poppe <niels@netbox.org>] 5583 5584 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 5585 [Ulf M�ller] 5586 5587 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 5588 extension adding in x509 utility. 5589 [Steve Henson] 5590 5591 *) Remove NOPROTO sections and error code comments. 5592 [Ulf M�ller] 5593 5594 *) Partial rewrite of the DEF file generator to now parse the ANSI 5595 prototypes. 5596 [Steve Henson] 5597 5598 *) New Configure options --prefix=DIR and --openssldir=DIR. 5599 [Ulf M�ller] 5600 5601 *) Complete rewrite of the error code script(s). It is all now handled 5602 by one script at the top level which handles error code gathering, 5603 header rewriting and C source file generation. It should be much better 5604 than the old method: it now uses a modified version of Ulf's parser to 5605 read the ANSI prototypes in all header files (thus the old K&R definitions 5606 aren't needed for error creation any more) and do a better job of 5607 translating function codes into names. The old 'ASN1 error code imbedded 5608 in a comment' is no longer necessary and it doesn't use .err files which 5609 have now been deleted. Also the error code call doesn't have to appear all 5610 on one line (which resulted in some large lines...). 5611 [Steve Henson] 5612 5613 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 5614 [Bodo Moeller] 5615 5616 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 5617 0 (which usually indicates a closed connection), but continue reading. 5618 [Bodo Moeller] 5619 5620 *) Fix some race conditions. 5621 [Bodo Moeller] 5622 5623 *) Add support for CRL distribution points extension. Add Certificate 5624 Policies and CRL distribution points documentation. 5625 [Steve Henson] 5626 5627 *) Move the autogenerated header file parts to crypto/opensslconf.h. 5628 [Ulf M�ller] 5629 5630 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 5631 8 of keying material. Merlin has also confirmed interop with this fix 5632 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 5633 [Merlin Hughes <merlin@baltimore.ie>] 5634 5635 *) Fix lots of warnings. 5636 [Richard Levitte <levitte@stacken.kth.se>] 5637 5638 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 5639 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 5640 [Richard Levitte <levitte@stacken.kth.se>] 5641 5642 *) Fix problems with sizeof(long) == 8. 5643 [Andy Polyakov <appro@fy.chalmers.se>] 5644 5645 *) Change functions to ANSI C. 5646 [Ulf M�ller] 5647 5648 *) Fix typos in error codes. 5649 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 5650 5651 *) Remove defunct assembler files from Configure. 5652 [Ulf M�ller] 5653 5654 *) SPARC v8 assembler BIGNUM implementation. 5655 [Andy Polyakov <appro@fy.chalmers.se>] 5656 5657 *) Support for Certificate Policies extension: both print and set. 5658 Various additions to support the r2i method this uses. 5659 [Steve Henson] 5660 5661 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 5662 return a const string when you are expecting an allocated buffer. 5663 [Ben Laurie] 5664 5665 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 5666 types DirectoryString and DisplayText. 5667 [Steve Henson] 5668 5669 *) Add code to allow r2i extensions to access the configuration database, 5670 add an LHASH database driver and add several ctx helper functions. 5671 [Steve Henson] 5672 5673 *) Fix an evil bug in bn_expand2() which caused various BN functions to 5674 fail when they extended the size of a BIGNUM. 5675 [Steve Henson] 5676 5677 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 5678 support typesafe stack. 5679 [Steve Henson] 5680 5681 *) Fix typo in SSL_[gs]et_options(). 5682 [Nils Frostberg <nils@medcom.se>] 5683 5684 *) Delete various functions and files that belonged to the (now obsolete) 5685 old X509V3 handling code. 5686 [Steve Henson] 5687 5688 *) New Configure option "rsaref". 5689 [Ulf M�ller] 5690 5691 *) Don't auto-generate pem.h. 5692 [Bodo Moeller] 5693 5694 *) Introduce type-safe ASN.1 SETs. 5695 [Ben Laurie] 5696 5697 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 5698 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 5699 5700 *) Introduce type-safe STACKs. This will almost certainly break lots of code 5701 that links with OpenSSL (well at least cause lots of warnings), but fear 5702 not: the conversion is trivial, and it eliminates loads of evil casts. A 5703 few STACKed things have been converted already. Feel free to convert more. 5704 In the fullness of time, I'll do away with the STACK type altogether. 5705 [Ben Laurie] 5706 5707 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 5708 specified in <certfile> by updating the entry in the index.txt file. 5709 This way one no longer has to edit the index.txt file manually for 5710 revoking a certificate. The -revoke option does the gory details now. 5711 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 5712 5713 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 5714 `-text' option at all and this way the `-noout -text' combination was 5715 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 5716 [Ralf S. Engelschall] 5717 5718 *) Make sure a corresponding plain text error message exists for the 5719 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 5720 verify callback function determined that a certificate was revoked. 5721 [Ralf S. Engelschall] 5722 5723 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 5724 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 5725 all available cipers including rc5, which was forgotten until now. 5726 In order to let the testing shell script know which algorithms 5727 are available, a new (up to now undocumented) command 5728 "openssl list-cipher-commands" is used. 5729 [Bodo Moeller] 5730 5731 *) Bugfix: s_client occasionally would sleep in select() when 5732 it should have checked SSL_pending() first. 5733 [Bodo Moeller] 5734 5735 *) New functions DSA_do_sign and DSA_do_verify to provide access to 5736 the raw DSA values prior to ASN.1 encoding. 5737 [Ulf M�ller] 5738 5739 *) Tweaks to Configure 5740 [Niels Poppe <niels@netbox.org>] 5741 5742 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 5743 yet... 5744 [Steve Henson] 5745 5746 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 5747 [Ulf M�ller] 5748 5749 *) New config option to avoid instructions that are illegal on the 80386. 5750 The default code is faster, but requires at least a 486. 5751 [Ulf M�ller] 5752 5753 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 5754 SSL2_SERVER_VERSION (not used at all) macros, which are now the 5755 same as SSL2_VERSION anyway. 5756 [Bodo Moeller] 5757 5758 *) New "-showcerts" option for s_client. 5759 [Bodo Moeller] 5760 5761 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 5762 application. Various cleanups and fixes. 5763 [Steve Henson] 5764 5765 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 5766 modify error routines to work internally. Add error codes and PBE init 5767 to library startup routines. 5768 [Steve Henson] 5769 5770 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 5771 packing functions to asn1 and evp. Changed function names and error 5772 codes along the way. 5773 [Steve Henson] 5774 5775 *) PKCS12 integration: and so it begins... First of several patches to 5776 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 5777 objects to objects.h 5778 [Steve Henson] 5779 5780 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 5781 and display support for Thawte strong extranet extension. 5782 [Steve Henson] 5783 5784 *) Add LinuxPPC support. 5785 [Jeff Dubrule <igor@pobox.org>] 5786 5787 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 5788 bn_div_words in alpha.s. 5789 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 5790 5791 *) Make sure the RSA OAEP test is skipped under -DRSAref because 5792 OAEP isn't supported when OpenSSL is built with RSAref. 5793 [Ulf Moeller <ulf@fitug.de>] 5794 5795 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 5796 so they no longer are missing under -DNOPROTO. 5797 [Soren S. Jorvang <soren@t.dk>] 5798 5799 5800 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 5801 5802 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 5803 doesn't work when the session is reused. Coming soon! 5804 [Ben Laurie] 5805 5806 *) Fix a security hole, that allows sessions to be reused in the wrong 5807 context thus bypassing client cert protection! All software that uses 5808 client certs and session caches in multiple contexts NEEDS PATCHING to 5809 allow session reuse! A fuller solution is in the works. 5810 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 5811 5812 *) Some more source tree cleanups (removed obsolete files 5813 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 5814 permission on "config" script to be executable) and a fix for the INSTALL 5815 document. 5816 [Ulf Moeller <ulf@fitug.de>] 5817 5818 *) Remove some legacy and erroneous uses of malloc, free instead of 5819 Malloc, Free. 5820 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 5821 5822 *) Make rsa_oaep_test return non-zero on error. 5823 [Ulf Moeller <ulf@fitug.de>] 5824 5825 *) Add support for native Solaris shared libraries. Configure 5826 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 5827 if someone would make that last step automatic. 5828 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 5829 5830 *) ctx_size was not built with the right compiler during "make links". Fixed. 5831 [Ben Laurie] 5832 5833 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 5834 except NULL ciphers". This means the default cipher list will no longer 5835 enable NULL ciphers. They need to be specifically enabled e.g. with 5836 the string "DEFAULT:eNULL". 5837 [Steve Henson] 5838 5839 *) Fix to RSA private encryption routines: if p < q then it would 5840 occasionally produce an invalid result. This will only happen with 5841 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 5842 [Steve Henson] 5843 5844 *) Be less restrictive and allow also `perl util/perlpath.pl 5845 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 5846 because this way one can also use an interpreter named `perl5' (which is 5847 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 5848 installed as `perl'). 5849 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 5850 5851 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 5852 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 5853 5854 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 5855 advapi32.lib to Win32 build and change the pem test comparision 5856 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 5857 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 5858 and crypto/des/ede_cbcm_enc.c. 5859 [Steve Henson] 5860 5861 *) DES quad checksum was broken on big-endian architectures. Fixed. 5862 [Ben Laurie] 5863 5864 *) Comment out two functions in bio.h that aren't implemented. Fix up the 5865 Win32 test batch file so it (might) work again. The Win32 test batch file 5866 is horrible: I feel ill.... 5867 [Steve Henson] 5868 5869 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 5870 in e_os.h. Audit of header files to check ANSI and non ANSI 5871 sections: 10 functions were absent from non ANSI section and not exported 5872 from Windows DLLs. Fixed up libeay.num for new functions. 5873 [Steve Henson] 5874 5875 *) Make `openssl version' output lines consistent. 5876 [Ralf S. Engelschall] 5877 5878 *) Fix Win32 symbol export lists for BIO functions: Added 5879 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 5880 to ms/libeay{16,32}.def. 5881 [Ralf S. Engelschall] 5882 5883 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 5884 fine under Unix and passes some trivial tests I've now added. But the 5885 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 5886 added to make sure no one expects that this stuff really works in the 5887 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 5888 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 5889 openssl_bio.xs. 5890 [Ralf S. Engelschall] 5891 5892 *) Fix the generation of two part addresses in perl. 5893 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 5894 5895 *) Add config entry for Linux on MIPS. 5896 [John Tobey <jtobey@channel1.com>] 5897 5898 *) Make links whenever Configure is run, unless we are on Windoze. 5899 [Ben Laurie] 5900 5901 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 5902 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 5903 in CRLs. 5904 [Steve Henson] 5905 5906 *) Add a useful kludge to allow package maintainers to specify compiler and 5907 other platforms details on the command line without having to patch the 5908 Configure script everytime: One now can use ``perl Configure 5909 <id>:<details>'', i.e. platform ids are allowed to have details appended 5910 to them (seperated by colons). This is treated as there would be a static 5911 pre-configured entry in Configure's %table under key <id> with value 5912 <details> and ``perl Configure <id>'' is called. So, when you want to 5913 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 5914 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 5915 now, which overrides the FreeBSD-elf entry on-the-fly. 5916 [Ralf S. Engelschall] 5917 5918 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 5919 [Ben Laurie] 5920 5921 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 5922 on the `perl Configure ...' command line. This way one can compile 5923 OpenSSL libraries with Position Independent Code (PIC) which is needed 5924 for linking it into DSOs. 5925 [Ralf S. Engelschall] 5926 5927 *) Remarkably, export ciphers were totally broken and no-one had noticed! 5928 Fixed. 5929 [Ben Laurie] 5930 5931 *) Cleaned up the LICENSE document: The official contact for any license 5932 questions now is the OpenSSL core team under openssl-core@openssl.org. 5933 And add a paragraph about the dual-license situation to make sure people 5934 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 5935 to the OpenSSL toolkit. 5936 [Ralf S. Engelschall] 5937 5938 *) General source tree makefile cleanups: Made `making xxx in yyy...' 5939 display consistent in the source tree and replaced `/bin/rm' by `rm'. 5940 Additonally cleaned up the `make links' target: Remove unnecessary 5941 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 5942 to speed processing and no longer clutter the display with confusing 5943 stuff. Instead only the actually done links are displayed. 5944 [Ralf S. Engelschall] 5945 5946 *) Permit null encryption ciphersuites, used for authentication only. It used 5947 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 5948 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 5949 encryption. 5950 [Ben Laurie] 5951 5952 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 5953 signed attributes when verifying signatures (this would break them), 5954 the detached data encoding was wrong and public keys obtained using 5955 X509_get_pubkey() weren't freed. 5956 [Steve Henson] 5957 5958 *) Add text documentation for the BUFFER functions. Also added a work around 5959 to a Win95 console bug. This was triggered by the password read stuff: the 5960 last character typed gets carried over to the next fread(). If you were 5961 generating a new cert request using 'req' for example then the last 5962 character of the passphrase would be CR which would then enter the first 5963 field as blank. 5964 [Steve Henson] 5965 5966 *) Added the new `Includes OpenSSL Cryptography Software' button as 5967 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 5968 button and can be used by applications based on OpenSSL to show the 5969 relationship to the OpenSSL project. 5970 [Ralf S. Engelschall] 5971 5972 *) Remove confusing variables in function signatures in files 5973 ssl/ssl_lib.c and ssl/ssl.h. 5974 [Lennart Bong <lob@kulthea.stacken.kth.se>] 5975 5976 *) Don't install bss_file.c under PREFIX/include/ 5977 [Lennart Bong <lob@kulthea.stacken.kth.se>] 5978 5979 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 5980 functions that return function pointers and has support for NT specific 5981 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 5982 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 5983 unsigned to signed types: this was killing the Win32 compile. 5984 [Steve Henson] 5985 5986 *) Add new certificate file to stack functions, 5987 SSL_add_dir_cert_subjects_to_stack() and 5988 SSL_add_file_cert_subjects_to_stack(). These largely supplant 5989 SSL_load_client_CA_file(), and can be used to add multiple certs easily 5990 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 5991 This means that Apache-SSL and similar packages don't have to mess around 5992 to add as many CAs as they want to the preferred list. 5993 [Ben Laurie] 5994 5995 *) Experiment with doxygen documentation. Currently only partially applied to 5996 ssl/ssl_lib.c. 5997 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 5998 openssl.doxy as the configuration file. 5999 [Ben Laurie] 6000 6001 *) Get rid of remaining C++-style comments which strict C compilers hate. 6002 [Ralf S. Engelschall, pointed out by Carlos Amengual] 6003 6004 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 6005 compiled in by default: it has problems with large keys. 6006 [Steve Henson] 6007 6008 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 6009 DH private keys and/or callback functions which directly correspond to 6010 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 6011 is needed for applications which have to configure certificates on a 6012 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 6013 (e.g. s_server). 6014 For the RSA certificate situation is makes no difference, but 6015 for the DSA certificate situation this fixes the "no shared cipher" 6016 problem where the OpenSSL cipher selection procedure failed because the 6017 temporary keys were not overtaken from the context and the API provided 6018 no way to reconfigure them. 6019 The new functions now let applications reconfigure the stuff and they 6020 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 6021 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 6022 non-public-API function ssl_cert_instantiate() is used as a helper 6023 function and also to reduce code redundancy inside ssl_rsa.c. 6024 [Ralf S. Engelschall] 6025 6026 *) Move s_server -dcert and -dkey options out of the undocumented feature 6027 area because they are useful for the DSA situation and should be 6028 recognized by the users. 6029 [Ralf S. Engelschall] 6030 6031 *) Fix the cipher decision scheme for export ciphers: the export bits are 6032 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 6033 SSL_EXP_MASK. So, the original variable has to be used instead of the 6034 already masked variable. 6035 [Richard Levitte <levitte@stacken.kth.se>] 6036 6037 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 6038 [Richard Levitte <levitte@stacken.kth.se>] 6039 6040 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 6041 from `int' to `unsigned int' because it's a length and initialized by 6042 EVP_DigestFinal() which expects an `unsigned int *'. 6043 [Richard Levitte <levitte@stacken.kth.se>] 6044 6045 *) Don't hard-code path to Perl interpreter on shebang line of Configure 6046 script. Instead use the usual Shell->Perl transition trick. 6047 [Ralf S. Engelschall] 6048 6049 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 6050 (in addition to RSA certificates) to match the behaviour of `openssl dsa 6051 -noout -modulus' as it's already the case for `openssl rsa -noout 6052 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 6053 currently the public key is printed (a decision which was already done by 6054 `openssl dsa -modulus' in the past) which serves a similar purpose. 6055 Additionally the NO_RSA no longer completely removes the whole -modulus 6056 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 6057 now, too. 6058 [Ralf S. Engelschall] 6059 6060 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 6061 BIO. See the source (crypto/evp/bio_ok.c) for more info. 6062 [Arne Ansper <arne@ats.cyber.ee>] 6063 6064 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 6065 to be added. Now both 'req' and 'ca' can use new objects defined in the 6066 config file. 6067 [Steve Henson] 6068 6069 *) Add cool BIO that does syslog (or event log on NT). 6070 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 6071 6072 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 6073 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 6074 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 6075 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 6076 [Ben Laurie] 6077 6078 *) Add preliminary config info for new extension code. 6079 [Steve Henson] 6080 6081 *) Make RSA_NO_PADDING really use no padding. 6082 [Ulf Moeller <ulf@fitug.de>] 6083 6084 *) Generate errors when private/public key check is done. 6085 [Ben Laurie] 6086 6087 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 6088 for some CRL extensions and new objects added. 6089 [Steve Henson] 6090 6091 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 6092 key usage extension and fuller support for authority key id. 6093 [Steve Henson] 6094 6095 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 6096 padding method for RSA, which is recommended for new applications in PKCS 6097 #1 v2.0 (RFC 2437, October 1998). 6098 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 6099 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 6100 against Bleichbacher's attack on RSA. 6101 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 6102 Ben Laurie] 6103 6104 *) Updates to the new SSL compression code 6105 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 6106 6107 *) Fix so that the version number in the master secret, when passed 6108 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 6109 (because the server will not accept higher), that the version number 6110 is 0x03,0x01, not 0x03,0x00 6111 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 6112 6113 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 6114 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 6115 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 6116 [Steve Henson] 6117 6118 *) Support for RAW extensions where an arbitrary extension can be 6119 created by including its DER encoding. See apps/openssl.cnf for 6120 an example. 6121 [Steve Henson] 6122 6123 *) Make sure latest Perl versions don't interpret some generated C array 6124 code as Perl array code in the crypto/err/err_genc.pl script. 6125 [Lars Weber <3weber@informatik.uni-hamburg.de>] 6126 6127 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 6128 not many people have the assembler. Various Win32 compilation fixes and 6129 update to the INSTALL.W32 file with (hopefully) more accurate Win32 6130 build instructions. 6131 [Steve Henson] 6132 6133 *) Modify configure script 'Configure' to automatically create crypto/date.h 6134 file under Win32 and also build pem.h from pem.org. New script 6135 util/mkfiles.pl to create the MINFO file on environments that can't do a 6136 'make files': perl util/mkfiles.pl >MINFO should work. 6137 [Steve Henson] 6138 6139 *) Major rework of DES function declarations, in the pursuit of correctness 6140 and purity. As a result, many evil casts evaporated, and some weirdness, 6141 too. You may find this causes warnings in your code. Zapping your evil 6142 casts will probably fix them. Mostly. 6143 [Ben Laurie] 6144 6145 *) Fix for a typo in asn1.h. Bug fix to object creation script 6146 obj_dat.pl. It considered a zero in an object definition to mean 6147 "end of object": none of the objects in objects.h have any zeros 6148 so it wasn't spotted. 6149 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 6150 6151 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 6152 Masking (CBCM). In the absence of test vectors, the best I have been able 6153 to do is check that the decrypt undoes the encrypt, so far. Send me test 6154 vectors if you have them. 6155 [Ben Laurie] 6156 6157 *) Correct calculation of key length for export ciphers (too much space was 6158 allocated for null ciphers). This has not been tested! 6159 [Ben Laurie] 6160 6161 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 6162 message is now correct (it understands "crypto" and "ssl" on its 6163 command line). There is also now an "update" option. This will update 6164 the util/ssleay.num and util/libeay.num files with any new functions. 6165 If you do a: 6166 perl util/mkdef.pl crypto ssl update 6167 it will update them. 6168 [Steve Henson] 6169 6170 *) Overhauled the Perl interface (perl/*): 6171 - ported BN stuff to OpenSSL's different BN library 6172 - made the perl/ source tree CVS-aware 6173 - renamed the package from SSLeay to OpenSSL (the files still contain 6174 their history because I've copied them in the repository) 6175 - removed obsolete files (the test scripts will be replaced 6176 by better Test::Harness variants in the future) 6177 [Ralf S. Engelschall] 6178 6179 *) First cut for a very conservative source tree cleanup: 6180 1. merge various obsolete readme texts into doc/ssleay.txt 6181 where we collect the old documents and readme texts. 6182 2. remove the first part of files where I'm already sure that we no 6183 longer need them because of three reasons: either they are just temporary 6184 files which were left by Eric or they are preserved original files where 6185 I've verified that the diff is also available in the CVS via "cvs diff 6186 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 6187 the crypto/md/ stuff). 6188 [Ralf S. Engelschall] 6189 6190 *) More extension code. Incomplete support for subject and issuer alt 6191 name, issuer and authority key id. Change the i2v function parameters 6192 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 6193 what that's for :-) Fix to ASN1 macro which messed up 6194 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 6195 [Steve Henson] 6196 6197 *) Preliminary support for ENUMERATED type. This is largely copied from the 6198 INTEGER code. 6199 [Steve Henson] 6200 6201 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 6202 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 6203 6204 *) Make sure `make rehash' target really finds the `openssl' program. 6205 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 6206 6207 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 6208 like to hear about it if this slows down other processors. 6209 [Ben Laurie] 6210 6211 *) Add CygWin32 platform information to Configure script. 6212 [Alan Batie <batie@aahz.jf.intel.com>] 6213 6214 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 6215 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 6216 6217 *) New program nseq to manipulate netscape certificate sequences 6218 [Steve Henson] 6219 6220 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 6221 few typos. 6222 [Steve Henson] 6223 6224 *) Fixes to BN code. Previously the default was to define BN_RECURSION 6225 but the BN code had some problems that would cause failures when 6226 doing certificate verification and some other functions. 6227 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 6228 6229 *) Add ASN1 and PEM code to support netscape certificate sequences. 6230 [Steve Henson] 6231 6232 *) Add ASN1 and PEM code to support netscape certificate sequences. 6233 [Steve Henson] 6234 6235 *) Add several PKIX and private extended key usage OIDs. 6236 [Steve Henson] 6237 6238 *) Modify the 'ca' program to handle the new extension code. Modify 6239 openssl.cnf for new extension format, add comments. 6240 [Steve Henson] 6241 6242 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 6243 and add a sample to openssl.cnf so req -x509 now adds appropriate 6244 CA extensions. 6245 [Steve Henson] 6246 6247 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 6248 error code, add initial support to X509_print() and x509 application. 6249 [Steve Henson] 6250 6251 *) Takes a deep breath and start addding X509 V3 extension support code. Add 6252 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 6253 stuff is currently isolated and isn't even compiled yet. 6254 [Steve Henson] 6255 6256 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 6257 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 6258 Removed the versions check from X509 routines when loading extensions: 6259 this allows certain broken certificates that don't set the version 6260 properly to be processed. 6261 [Steve Henson] 6262 6263 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 6264 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 6265 can still be regenerated with "make depend". 6266 [Ben Laurie] 6267 6268 *) Spelling mistake in C version of CAST-128. 6269 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 6270 6271 *) Changes to the error generation code. The perl script err-code.pl 6272 now reads in the old error codes and retains the old numbers, only 6273 adding new ones if necessary. It also only changes the .err files if new 6274 codes are added. The makefiles have been modified to only insert errors 6275 when needed (to avoid needlessly modifying header files). This is done 6276 by only inserting errors if the .err file is newer than the auto generated 6277 C file. To rebuild all the error codes from scratch (the old behaviour) 6278 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 6279 or delete all the .err files. 6280 [Steve Henson] 6281 6282 *) CAST-128 was incorrectly implemented for short keys. The C version has 6283 been fixed, but is untested. The assembler versions are also fixed, but 6284 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 6285 to regenerate it if needed. 6286 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 6287 Hagino <itojun@kame.net>] 6288 6289 *) File was opened incorrectly in randfile.c. 6290 [Ulf M�ller <ulf@fitug.de>] 6291 6292 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 6293 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 6294 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 6295 al: it's just almost always a UTCTime. Note this patch adds new error 6296 codes so do a "make errors" if there are problems. 6297 [Steve Henson] 6298 6299 *) Correct Linux 1 recognition in config. 6300 [Ulf M�ller <ulf@fitug.de>] 6301 6302 *) Remove pointless MD5 hash when using DSA keys in ca. 6303 [Anonymous <nobody@replay.com>] 6304 6305 *) Generate an error if given an empty string as a cert directory. Also 6306 generate an error if handed NULL (previously returned 0 to indicate an 6307 error, but didn't set one). 6308 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 6309 6310 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 6311 [Ben Laurie] 6312 6313 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 6314 parameters. This was causing a warning which killed off the Win32 compile. 6315 [Steve Henson] 6316 6317 *) Remove C++ style comments from crypto/bn/bn_local.h. 6318 [Neil Costigan <neil.costigan@celocom.com>] 6319 6320 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 6321 based on a text string, looking up short and long names and finally 6322 "dot" format. The "dot" format stuff didn't work. Added new function 6323 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 6324 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 6325 OID is not part of the table. 6326 [Steve Henson] 6327 6328 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 6329 X509_LOOKUP_by_alias(). 6330 [Ben Laurie] 6331 6332 *) Sort openssl functions by name. 6333 [Ben Laurie] 6334 6335 *) Get the gendsa program working (hopefully) and add it to app list. Remove 6336 encryption from sample DSA keys (in case anyone is interested the password 6337 was "1234"). 6338 [Steve Henson] 6339 6340 *) Make _all_ *_free functions accept a NULL pointer. 6341 [Frans Heymans <fheymans@isaserver.be>] 6342 6343 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 6344 NULL pointers. 6345 [Anonymous <nobody@replay.com>] 6346 6347 *) s_server should send the CAfile as acceptable CAs, not its own cert. 6348 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 6349 6350 *) Don't blow it for numeric -newkey arguments to apps/req. 6351 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 6352 6353 *) Temp key "for export" tests were wrong in s3_srvr.c. 6354 [Anonymous <nobody@replay.com>] 6355 6356 *) Add prototype for temp key callback functions 6357 SSL_CTX_set_tmp_{rsa,dh}_callback(). 6358 [Ben Laurie] 6359 6360 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 6361 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 6362 [Steve Henson] 6363 6364 *) X509_name_add_entry() freed the wrong thing after an error. 6365 [Arne Ansper <arne@ats.cyber.ee>] 6366 6367 *) rsa_eay.c would attempt to free a NULL context. 6368 [Arne Ansper <arne@ats.cyber.ee>] 6369 6370 *) BIO_s_socket() had a broken should_retry() on Windoze. 6371 [Arne Ansper <arne@ats.cyber.ee>] 6372 6373 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 6374 [Arne Ansper <arne@ats.cyber.ee>] 6375 6376 *) Make sure the already existing X509_STORE->depth variable is initialized 6377 in X509_STORE_new(), but document the fact that this variable is still 6378 unused in the certificate verification process. 6379 [Ralf S. Engelschall] 6380 6381 *) Fix the various library and apps files to free up pkeys obtained from 6382 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 6383 [Steve Henson] 6384 6385 *) Fix reference counting in X509_PUBKEY_get(). This makes 6386 demos/maurice/example2.c work, amongst others, probably. 6387 [Steve Henson and Ben Laurie] 6388 6389 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 6390 `openssl' and second, the shortcut symlinks for the `openssl <command>' 6391 are no longer created. This way we have a single and consistent command 6392 line interface `openssl <command>', similar to `cvs <command>'. 6393 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 6394 6395 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 6396 BIT STRING wrapper always have zero unused bits. 6397 [Steve Henson] 6398 6399 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 6400 [Steve Henson] 6401 6402 *) Make the top-level INSTALL documentation easier to understand. 6403 [Paul Sutton] 6404 6405 *) Makefiles updated to exit if an error occurs in a sub-directory 6406 make (including if user presses ^C) [Paul Sutton] 6407 6408 *) Make Montgomery context stuff explicit in RSA data structure. 6409 [Ben Laurie] 6410 6411 *) Fix build order of pem and err to allow for generated pem.h. 6412 [Ben Laurie] 6413 6414 *) Fix renumbering bug in X509_NAME_delete_entry(). 6415 [Ben Laurie] 6416 6417 *) Enhanced the err-ins.pl script so it makes the error library number 6418 global and can add a library name. This is needed for external ASN1 and 6419 other error libraries. 6420 [Steve Henson] 6421 6422 *) Fixed sk_insert which never worked properly. 6423 [Steve Henson] 6424 6425 *) Fix ASN1 macros so they can handle indefinite length construted 6426 EXPLICIT tags. Some non standard certificates use these: they can now 6427 be read in. 6428 [Steve Henson] 6429 6430 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 6431 into a single doc/ssleay.txt bundle. This way the information is still 6432 preserved but no longer messes up this directory. Now it's new room for 6433 the new set of documenation files. 6434 [Ralf S. Engelschall] 6435 6436 *) SETs were incorrectly DER encoded. This was a major pain, because they 6437 shared code with SEQUENCEs, which aren't coded the same. This means that 6438 almost everything to do with SETs or SEQUENCEs has either changed name or 6439 number of arguments. 6440 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 6441 6442 *) Fix test data to work with the above. 6443 [Ben Laurie] 6444 6445 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 6446 was already fixed by Eric for 0.9.1 it seems. 6447 [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 6448 6449 *) Autodetect FreeBSD3. 6450 [Ben Laurie] 6451 6452 *) Fix various bugs in Configure. This affects the following platforms: 6453 nextstep 6454 ncr-scde 6455 unixware-2.0 6456 unixware-2.0-pentium 6457 sco5-cc. 6458 [Ben Laurie] 6459 6460 *) Eliminate generated files from CVS. Reorder tests to regenerate files 6461 before they are needed. 6462 [Ben Laurie] 6463 6464 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 6465 [Ben Laurie] 6466 6467 6468 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 6469 6470 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 6471 changed SSLeay to OpenSSL in version strings. 6472 [Ralf S. Engelschall] 6473 6474 *) Some fixups to the top-level documents. 6475 [Paul Sutton] 6476 6477 *) Fixed the nasty bug where rsaref.h was not found under compile-time 6478 because the symlink to include/ was missing. 6479 [Ralf S. Engelschall] 6480 6481 *) Incorporated the popular no-RSA/DSA-only patches 6482 which allow to compile a RSA-free SSLeay. 6483 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 6484 6485 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 6486 when "ssleay" is still not found. 6487 [Ralf S. Engelschall] 6488 6489 *) Added more platforms to Configure: Cray T3E, HPUX 11, 6490 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 6491 6492 *) Updated the README file. 6493 [Ralf S. Engelschall] 6494 6495 *) Added various .cvsignore files in the CVS repository subdirs 6496 to make a "cvs update" really silent. 6497 [Ralf S. Engelschall] 6498 6499 *) Recompiled the error-definition header files and added 6500 missing symbols to the Win32 linker tables. 6501 [Ralf S. Engelschall] 6502 6503 *) Cleaned up the top-level documents; 6504 o new files: CHANGES and LICENSE 6505 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 6506 o merged COPYRIGHT into LICENSE 6507 o removed obsolete TODO file 6508 o renamed MICROSOFT to INSTALL.W32 6509 [Ralf S. Engelschall] 6510 6511 *) Removed dummy files from the 0.9.1b source tree: 6512 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 6513 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 6514 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 6515 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 6516 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 6517 [Ralf S. Engelschall] 6518 6519 *) Added various platform portability fixes. 6520 [Mark J. Cox] 6521 6522 *) The Genesis of the OpenSSL rpject: 6523 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 6524 Young and Tim J. Hudson created while they were working for C2Net until 6525 summer 1998. 6526 [The OpenSSL Project] 6527 6528 6529 Changes between 0.9.0b and 0.9.1b [not released] 6530 6531 *) Updated a few CA certificates under certs/ 6532 [Eric A. Young] 6533 6534 *) Changed some BIGNUM api stuff. 6535 [Eric A. Young] 6536 6537 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 6538 DGUX x86, Linux Alpha, etc. 6539 [Eric A. Young] 6540 6541 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 6542 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 6543 available). 6544 [Eric A. Young] 6545 6546 *) Add -strparse option to asn1pars program which parses nested 6547 binary structures 6548 [Dr Stephen Henson <shenson@bigfoot.com>] 6549 6550 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 6551 [Eric A. Young] 6552 6553 *) DSA fix for "ca" program. 6554 [Eric A. Young] 6555 6556 *) Added "-genkey" option to "dsaparam" program. 6557 [Eric A. Young] 6558 6559 *) Added RIPE MD160 (rmd160) message digest. 6560 [Eric A. Young] 6561 6562 *) Added -a (all) option to "ssleay version" command. 6563 [Eric A. Young] 6564 6565 *) Added PLATFORM define which is the id given to Configure. 6566 [Eric A. Young] 6567 6568 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 6569 [Eric A. Young] 6570 6571 *) Extended the ASN.1 parser routines. 6572 [Eric A. Young] 6573 6574 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 6575 [Eric A. Young] 6576 6577 *) Added a BN_CTX to the BN library. 6578 [Eric A. Young] 6579 6580 *) Fixed the weak key values in DES library 6581 [Eric A. Young] 6582 6583 *) Changed API in EVP library for cipher aliases. 6584 [Eric A. Young] 6585 6586 *) Added support for RC2/64bit cipher. 6587 [Eric A. Young] 6588 6589 *) Converted the lhash library to the crypto/mem.c functions. 6590 [Eric A. Young] 6591 6592 *) Added more recognized ASN.1 object ids. 6593 [Eric A. Young] 6594 6595 *) Added more RSA padding checks for SSL/TLS. 6596 [Eric A. Young] 6597 6598 *) Added BIO proxy/filter functionality. 6599 [Eric A. Young] 6600 6601 *) Added extra_certs to SSL_CTX which can be used 6602 send extra CA certificates to the client in the CA cert chain sending 6603 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 6604 [Eric A. Young] 6605 6606 *) Now Fortezza is denied in the authentication phase because 6607 this is key exchange mechanism is not supported by SSLeay at all. 6608 [Eric A. Young] 6609 6610 *) Additional PKCS1 checks. 6611 [Eric A. Young] 6612 6613 *) Support the string "TLSv1" for all TLS v1 ciphers. 6614 [Eric A. Young] 6615 6616 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 6617 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 6618 [Eric A. Young] 6619 6620 *) Fixed a few memory leaks. 6621 [Eric A. Young] 6622 6623 *) Fixed various code and comment typos. 6624 [Eric A. Young] 6625 6626 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 6627 bytes sent in the client random. 6628 [Edward Bishop <ebishop@spyglass.com>] 6629 6630