CHANGES revision 120631
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 6 7 *) Fix various bugs revealed by running the NISCC test suite: 8 9 Stop out of bounds reads in the ASN1 code when presented with 10 invalid tags (CAN-2003-0543 and CAN-2003-0544). 11 12 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545). 13 14 If verify callback ignores invalid public key errors don't try to check 15 certificate signature with the NULL public key. 16 17 [Steve Henson] 18 19 *) New -ignore_err option in ocsp application to stop the server 20 exiting on the first error in a request. 21 [Steve Henson] 22 23 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 24 if the server requested one: as stated in TLS 1.0 and SSL 3.0 25 specifications. 26 [Steve Henson] 27 28 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 29 extra data after the compression methods not only for TLS 1.0 30 but also for SSL 3.0 (as required by the specification). 31 [Bodo Moeller; problem pointed out by Matthias Loepfe] 32 33 *) Change X509_certificate_type() to mark the key as exported/exportable 34 when it's 512 *bits* long, not 512 bytes. 35 [Richard Levitte] 36 37 *) Change AES_cbc_encrypt() so it outputs exact multiple of 38 blocks during encryption. 39 [Richard Levitte] 40 41 *) Various fixes to base64 BIO and non blocking I/O. On write 42 flushes were not handled properly if the BIO retried. On read 43 data was not being buffered properly and had various logic bugs. 44 This also affects blocking I/O when the data being decoded is a 45 certain size. 46 [Steve Henson] 47 48 *) Various S/MIME bugfixes and compatibility changes: 49 output correct application/pkcs7 MIME type if 50 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 51 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 52 of files as .eml work). Correctly handle very long lines in MIME 53 parser. 54 [Steve Henson] 55 56 Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 57 58 *) Countermeasure against the Klima-Pokorny-Rosa extension of 59 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 60 a protocol version number mismatch like a decryption error 61 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 62 [Bodo Moeller] 63 64 *) Turn on RSA blinding by default in the default implementation 65 to avoid a timing attack. Applications that don't want it can call 66 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 67 They would be ill-advised to do so in most cases. 68 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 69 70 *) Change RSA blinding code so that it works when the PRNG is not 71 seeded (in this case, the secret RSA exponent is abused as 72 an unpredictable seed -- if it is not unpredictable, there 73 is no point in blinding anyway). Make RSA blinding thread-safe 74 by remembering the creator's thread ID in rsa->blinding and 75 having all other threads use local one-time blinding factors 76 (this requires more computation than sharing rsa->blinding, but 77 avoids excessive locking; and if an RSA object is not shared 78 between threads, blinding will still be very fast). 79 [Bodo Moeller] 80 81 *) Fixed a typo bug that would cause ENGINE_set_default() to set an 82 ENGINE as defaults for all supported algorithms irrespective of 83 the 'flags' parameter. 'flags' is now honoured, so applications 84 should make sure they are passing it correctly. 85 [Geoff Thorpe] 86 87 *) Target "mingw" now allows native Windows code to be generated in 88 the Cygwin environment as well as with the MinGW compiler. 89 [Ulf Moeller] 90 91 Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 92 93 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 94 via timing by performing a MAC computation even if incorrrect 95 block cipher padding has been found. This is a countermeasure 96 against active attacks where the attacker has to distinguish 97 between bad padding and a MAC verification error. (CAN-2003-0078) 98 99 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 100 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 101 Martin Vuagnoux (EPFL, Ilion)] 102 103 *) Make the no-err option work as intended. The intention with no-err 104 is not to have the whole error stack handling routines removed from 105 libcrypto, it's only intended to remove all the function name and 106 reason texts, thereby removing some of the footprint that may not 107 be interesting if those errors aren't displayed anyway. 108 109 NOTE: it's still possible for any application or module to have it's 110 own set of error texts inserted. The routines are there, just not 111 used by default when no-err is given. 112 [Richard Levitte] 113 114 *) Add support for FreeBSD on IA64. 115 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 116 117 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 118 Kerberos function mit_des_cbc_cksum(). Before this change, 119 the value returned by DES_cbc_cksum() was like the one from 120 mit_des_cbc_cksum(), except the bytes were swapped. 121 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 122 123 *) Allow an application to disable the automatic SSL chain building. 124 Before this a rather primitive chain build was always performed in 125 ssl3_output_cert_chain(): an application had no way to send the 126 correct chain if the automatic operation produced an incorrect result. 127 128 Now the chain builder is disabled if either: 129 130 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 131 132 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 133 134 The reasoning behind this is that an application would not want the 135 auto chain building to take place if extra chain certificates are 136 present and it might also want a means of sending no additional 137 certificates (for example the chain has two certificates and the 138 root is omitted). 139 [Steve Henson] 140 141 *) Add the possibility to build without the ENGINE framework. 142 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 143 144 *) Under Win32 gmtime() can return NULL: check return value in 145 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 146 [Steve Henson] 147 148 *) DSA routines: under certain error conditions uninitialized BN objects 149 could be freed. Solution: make sure initialization is performed early 150 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 151 Nils Larsch <nla@trustcenter.de> via PR#459) 152 [Lutz Jaenicke] 153 154 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 155 checked on reconnect on the client side, therefore session resumption 156 could still fail with a "ssl session id is different" error. This 157 behaviour is masked when SSL_OP_ALL is used due to 158 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 159 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 160 followup to PR #377. 161 [Lutz Jaenicke] 162 163 *) IA-32 assembler support enhancements: unified ELF targets, support 164 for SCO/Caldera platforms, fix for Cygwin shared build. 165 [Andy Polyakov] 166 167 *) Add support for FreeBSD on sparc64. As a consequence, support for 168 FreeBSD on non-x86 processors is separate from x86 processors on 169 the config script, much like the NetBSD support. 170 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 171 172 Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 173 174 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 175 OpenSSL 0.9.7.] 176 177 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 178 code (06) was taken as the first octet of the session ID and the last 179 octet was ignored consequently. As a result SSLv2 client side session 180 caching could not have worked due to the session ID mismatch between 181 client and server. 182 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 183 PR #377. 184 [Lutz Jaenicke] 185 186 *) Change the declaration of needed Kerberos libraries to use EX_LIBS 187 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 188 removed entirely. 189 [Richard Levitte] 190 191 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 192 seems that in spite of existing for more than a year, many application 193 author have done nothing to provide the necessary callbacks, which 194 means that this particular engine will not work properly anywhere. 195 This is a very unfortunate situation which forces us, in the name 196 of usability, to give the hw_ncipher.c a static lock, which is part 197 of libcrypto. 198 NOTE: This is for the 0.9.7 series ONLY. This hack will never 199 appear in 0.9.8 or later. We EXPECT application authors to have 200 dealt properly with this when 0.9.8 is released (unless we actually 201 make such changes in the libcrypto locking code that changes will 202 have to be made anyway). 203 [Richard Levitte] 204 205 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 206 octets have been read, EOF or an error occurs. Without this change 207 some truncated ASN1 structures will not produce an error. 208 [Steve Henson] 209 210 *) Disable Heimdal support, since it hasn't been fully implemented. 211 Still give the possibility to force the use of Heimdal, but with 212 warnings and a request that patches get sent to openssl-dev. 213 [Richard Levitte] 214 215 *) Add the VC-CE target, introduce the WINCE sysname, and add 216 INSTALL.WCE and appropriate conditionals to make it build. 217 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 218 219 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 220 cygssl-x.y.z.dll, where x, y and z are the major, minor and 221 edit numbers of the version. 222 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 223 224 *) Introduce safe string copy and catenation functions 225 (BUF_strlcpy() and BUF_strlcat()). 226 [Ben Laurie (CHATS) and Richard Levitte] 227 228 *) Avoid using fixed-size buffers for one-line DNs. 229 [Ben Laurie (CHATS)] 230 231 *) Add BUF_MEM_grow_clean() to avoid information leakage when 232 resizing buffers containing secrets, and use where appropriate. 233 [Ben Laurie (CHATS)] 234 235 *) Avoid using fixed size buffers for configuration file location. 236 [Ben Laurie (CHATS)] 237 238 *) Avoid filename truncation for various CA files. 239 [Ben Laurie (CHATS)] 240 241 *) Use sizeof in preference to magic numbers. 242 [Ben Laurie (CHATS)] 243 244 *) Avoid filename truncation in cert requests. 245 [Ben Laurie (CHATS)] 246 247 *) Add assertions to check for (supposedly impossible) buffer 248 overflows. 249 [Ben Laurie (CHATS)] 250 251 *) Don't cache truncated DNS entries in the local cache (this could 252 potentially lead to a spoofing attack). 253 [Ben Laurie (CHATS)] 254 255 *) Fix various buffers to be large enough for hex/decimal 256 representations in a platform independent manner. 257 [Ben Laurie (CHATS)] 258 259 *) Add CRYPTO_realloc_clean() to avoid information leakage when 260 resizing buffers containing secrets, and use where appropriate. 261 [Ben Laurie (CHATS)] 262 263 *) Add BIO_indent() to avoid much slightly worrying code to do 264 indents. 265 [Ben Laurie (CHATS)] 266 267 *) Convert sprintf()/BIO_puts() to BIO_printf(). 268 [Ben Laurie (CHATS)] 269 270 *) buffer_gets() could terminate with the buffer only half 271 full. Fixed. 272 [Ben Laurie (CHATS)] 273 274 *) Add assertions to prevent user-supplied crypto functions from 275 overflowing internal buffers by having large block sizes, etc. 276 [Ben Laurie (CHATS)] 277 278 *) New OPENSSL_assert() macro (similar to assert(), but enabled 279 unconditionally). 280 [Ben Laurie (CHATS)] 281 282 *) Eliminate unused copy of key in RC4. 283 [Ben Laurie (CHATS)] 284 285 *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 286 [Ben Laurie (CHATS)] 287 288 *) Fix off-by-one error in EGD path. 289 [Ben Laurie (CHATS)] 290 291 *) If RANDFILE path is too long, ignore instead of truncating. 292 [Ben Laurie (CHATS)] 293 294 *) Eliminate unused and incorrectly sized X.509 structure 295 CBCParameter. 296 [Ben Laurie (CHATS)] 297 298 *) Eliminate unused and dangerous function knumber(). 299 [Ben Laurie (CHATS)] 300 301 *) Eliminate unused and dangerous structure, KSSL_ERR. 302 [Ben Laurie (CHATS)] 303 304 *) Protect against overlong session ID context length in an encoded 305 session object. Since these are local, this does not appear to be 306 exploitable. 307 [Ben Laurie (CHATS)] 308 309 *) Change from security patch (see 0.9.6e below) that did not affect 310 the 0.9.6 release series: 311 312 Remote buffer overflow in SSL3 protocol - an attacker could 313 supply an oversized master key in Kerberos-enabled versions. 314 (CAN-2002-0657) 315 [Ben Laurie (CHATS)] 316 317 *) Change the SSL kerb5 codes to match RFC 2712. 318 [Richard Levitte] 319 320 *) Make -nameopt work fully for req and add -reqopt switch. 321 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 322 323 *) The "block size" for block ciphers in CFB and OFB mode should be 1. 324 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 325 326 *) Make sure tests can be performed even if the corresponding algorithms 327 have been removed entirely. This was also the last step to make 328 OpenSSL compilable with DJGPP under all reasonable conditions. 329 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 330 331 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 332 to allow version independent disabling of normally unselected ciphers, 333 which may be activated as a side-effect of selecting a single cipher. 334 335 (E.g., cipher list string "RSA" enables ciphersuites that are left 336 out of "ALL" because they do not provide symmetric encryption. 337 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 338 [Lutz Jaenicke, Bodo Moeller] 339 340 *) Add appropriate support for separate platform-dependent build 341 directories. The recommended way to make a platform-dependent 342 build directory is the following (tested on Linux), maybe with 343 some local tweaks: 344 345 # Place yourself outside of the OpenSSL source tree. In 346 # this example, the environment variable OPENSSL_SOURCE 347 # is assumed to contain the absolute OpenSSL source directory. 348 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 349 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 350 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 351 mkdir -p `dirname $F` 352 ln -s $OPENSSL_SOURCE/$F $F 353 done 354 355 To be absolutely sure not to disturb the source tree, a "make clean" 356 is a good thing. If it isn't successfull, don't worry about it, 357 it probably means the source directory is very clean. 358 [Richard Levitte] 359 360 *) Make sure any ENGINE control commands make local copies of string 361 pointers passed to them whenever necessary. Otherwise it is possible 362 the caller may have overwritten (or deallocated) the original string 363 data when a later ENGINE operation tries to use the stored values. 364 [G�tz Babin-Ebell <babinebell@trustcenter.de>] 365 366 *) Improve diagnostics in file reading and command-line digests. 367 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 368 369 *) Add AES modes CFB and OFB to the object database. Correct an 370 error in AES-CFB decryption. 371 [Richard Levitte] 372 373 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 374 allows existing EVP_CIPHER_CTX structures to be reused after 375 calling EVP_*Final(). This behaviour is used by encryption 376 BIOs and some applications. This has the side effect that 377 applications must explicitly clean up cipher contexts with 378 EVP_CIPHER_CTX_cleanup() or they will leak memory. 379 [Steve Henson] 380 381 *) Check the values of dna and dnb in bn_mul_recursive before calling 382 bn_mul_comba (a non zero value means the a or b arrays do not contain 383 n2 elements) and fallback to bn_mul_normal if either is not zero. 384 [Steve Henson] 385 386 *) Fix escaping of non-ASCII characters when using the -subj option 387 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 388 [Lutz Jaenicke] 389 390 *) Make object definitions compliant to LDAP (RFC2256): SN is the short 391 form for "surname", serialNumber has no short form. 392 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 393 therefore remove "mail" short name for "internet 7". 394 The OID for unique identifiers in X509 certificates is 395 x500UniqueIdentifier, not uniqueIdentifier. 396 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 397 [Lutz Jaenicke] 398 399 *) Add an "init" command to the ENGINE config module and auto initialize 400 ENGINEs. Without any "init" command the ENGINE will be initialized 401 after all ctrl commands have been executed on it. If init=1 the 402 ENGINE is initailized at that point (ctrls before that point are run 403 on the uninitialized ENGINE and after on the initialized one). If 404 init=0 then the ENGINE will not be iniatialized at all. 405 [Steve Henson] 406 407 *) Fix the 'app_verify_callback' interface so that the user-defined 408 argument is actually passed to the callback: In the 409 SSL_CTX_set_cert_verify_callback() prototype, the callback 410 declaration has been changed from 411 int (*cb)() 412 into 413 int (*cb)(X509_STORE_CTX *,void *); 414 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 415 i=s->ctx->app_verify_callback(&ctx) 416 has been changed into 417 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 418 419 To update applications using SSL_CTX_set_cert_verify_callback(), 420 a dummy argument can be added to their callback functions. 421 [D. K. Smetters <smetters@parc.xerox.com>] 422 423 *) Added the '4758cca' ENGINE to support IBM 4758 cards. 424 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 425 426 *) Add and OPENSSL_LOAD_CONF define which will cause 427 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 428 This allows older applications to transparently support certain 429 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 430 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 431 load the config file and OPENSSL_add_all_algorithms_conf() which will 432 always load it have also been added. 433 [Steve Henson] 434 435 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 436 Adjust NIDs and EVP layer. 437 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 438 439 *) Config modules support in openssl utility. 440 441 Most commands now load modules from the config file, 442 though in a few (such as version) this isn't done 443 because it couldn't be used for anything. 444 445 In the case of ca and req the config file used is 446 the same as the utility itself: that is the -config 447 command line option can be used to specify an 448 alternative file. 449 [Steve Henson] 450 451 *) Move default behaviour from OPENSSL_config(). If appname is NULL 452 use "openssl_conf" if filename is NULL use default openssl config file. 453 [Steve Henson] 454 455 *) Add an argument to OPENSSL_config() to allow the use of an alternative 456 config section name. Add a new flag to tolerate a missing config file 457 and move code to CONF_modules_load_file(). 458 [Steve Henson] 459 460 *) Support for crypto accelerator cards from Accelerated Encryption 461 Processing, www.aep.ie. (Use engine 'aep') 462 The support was copied from 0.9.6c [engine] and adapted/corrected 463 to work with the new engine framework. 464 [AEP Inc. and Richard Levitte] 465 466 *) Support for SureWare crypto accelerator cards from Baltimore 467 Technologies. (Use engine 'sureware') 468 The support was copied from 0.9.6c [engine] and adapted 469 to work with the new engine framework. 470 [Richard Levitte] 471 472 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 473 make the newer ENGINE framework commands for the CHIL engine work. 474 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 475 476 *) Make it possible to produce shared libraries on ReliantUNIX. 477 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 478 479 *) Add the configuration target debug-linux-ppro. 480 Make 'openssl rsa' use the general key loading routines 481 implemented in apps.c, and make those routines able to 482 handle the key format FORMAT_NETSCAPE and the variant 483 FORMAT_IISSGC. 484 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 485 486 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 487 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 488 489 *) Add -keyform to rsautl, and document -engine. 490 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 491 492 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 493 BIO_R_NO_SUCH_FILE error code rather than the generic 494 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 495 [Ben Laurie] 496 497 *) Add new functions 498 ERR_peek_last_error 499 ERR_peek_last_error_line 500 ERR_peek_last_error_line_data. 501 These are similar to 502 ERR_peek_error 503 ERR_peek_error_line 504 ERR_peek_error_line_data, 505 but report on the latest error recorded rather than the first one 506 still in the error queue. 507 [Ben Laurie, Bodo Moeller] 508 509 *) default_algorithms option in ENGINE config module. This allows things 510 like: 511 default_algorithms = ALL 512 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 513 [Steve Henson] 514 515 *) Prelminary ENGINE config module. 516 [Steve Henson] 517 518 *) New experimental application configuration code. 519 [Steve Henson] 520 521 *) Change the AES code to follow the same name structure as all other 522 symmetric ciphers, and behave the same way. Move everything to 523 the directory crypto/aes, thereby obsoleting crypto/rijndael. 524 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 525 526 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 527 [Ben Laurie and Theo de Raadt] 528 529 *) Add option to output public keys in req command. 530 [Massimiliano Pala madwolf@openca.org] 531 532 *) Use wNAFs in EC_POINTs_mul() for improved efficiency 533 (up to about 10% better than before for P-192 and P-224). 534 [Bodo Moeller] 535 536 *) New functions/macros 537 538 SSL_CTX_set_msg_callback(ctx, cb) 539 SSL_CTX_set_msg_callback_arg(ctx, arg) 540 SSL_set_msg_callback(ssl, cb) 541 SSL_set_msg_callback_arg(ssl, arg) 542 543 to request calling a callback function 544 545 void cb(int write_p, int version, int content_type, 546 const void *buf, size_t len, SSL *ssl, void *arg) 547 548 whenever a protocol message has been completely received 549 (write_p == 0) or sent (write_p == 1). Here 'version' is the 550 protocol version according to which the SSL library interprets 551 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 552 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 553 the content type as defined in the SSL 3.0/TLS 1.0 protocol 554 specification (change_cipher_spec(20), alert(21), handshake(22)). 555 'buf' and 'len' point to the actual message, 'ssl' to the 556 SSL object, and 'arg' is the application-defined value set by 557 SSL[_CTX]_set_msg_callback_arg(). 558 559 'openssl s_client' and 'openssl s_server' have new '-msg' options 560 to enable a callback that displays all protocol messages. 561 [Bodo Moeller] 562 563 *) Change the shared library support so shared libraries are built as 564 soon as the corresponding static library is finished, and thereby get 565 openssl and the test programs linked against the shared library. 566 This still only happens when the keyword "shard" has been given to 567 the configuration scripts. 568 569 NOTE: shared library support is still an experimental thing, and 570 backward binary compatibility is still not guaranteed. 571 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 572 573 *) Add support for Subject Information Access extension. 574 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 575 576 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 577 additional bytes when new memory had to be allocated, not just 578 when reusing an existing buffer. 579 [Bodo Moeller] 580 581 *) New command line and configuration option 'utf8' for the req command. 582 This allows field values to be specified as UTF8 strings. 583 [Steve Henson] 584 585 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 586 runs for the former and machine-readable output for the latter. 587 [Ben Laurie] 588 589 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 590 of the e-mail address in the DN (i.e., it will go into a certificate 591 extension only). The new configuration file option 'email_in_dn = no' 592 has the same effect. 593 [Massimiliano Pala madwolf@openca.org] 594 595 *) Change all functions with names starting with des_ to be starting 596 with DES_ instead. Add wrappers that are compatible with libdes, 597 but are named _ossl_old_des_*. Finally, add macros that map the 598 des_* symbols to the corresponding _ossl_old_des_* if libdes 599 compatibility is desired. If OpenSSL 0.9.6c compatibility is 600 desired, the des_* symbols will be mapped to DES_*, with one 601 exception. 602 603 Since we provide two compatibility mappings, the user needs to 604 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 605 compatibility is desired. The default (i.e., when that macro 606 isn't defined) is OpenSSL 0.9.6c compatibility. 607 608 There are also macros that enable and disable the support of old 609 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 610 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 611 are defined, the default will apply: to support the old des routines. 612 613 In either case, one must include openssl/des.h to get the correct 614 definitions. Do not try to just include openssl/des_old.h, that 615 won't work. 616 617 NOTE: This is a major break of an old API into a new one. Software 618 authors are encouraged to switch to the DES_ style functions. Some 619 time in the future, des_old.h and the libdes compatibility functions 620 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 621 default), and then completely removed. 622 [Richard Levitte] 623 624 *) Test for certificates which contain unsupported critical extensions. 625 If such a certificate is found during a verify operation it is 626 rejected by default: this behaviour can be overridden by either 627 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 628 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 629 X509_supported_extension() has also been added which returns 1 if a 630 particular extension is supported. 631 [Steve Henson] 632 633 *) Modify the behaviour of EVP cipher functions in similar way to digests 634 to retain compatibility with existing code. 635 [Steve Henson] 636 637 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 638 compatibility with existing code. In particular the 'ctx' parameter does 639 not have to be to be initialized before the call to EVP_DigestInit() and 640 it is tidied up after a call to EVP_DigestFinal(). New function 641 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 642 EVP_MD_CTX_copy() changed to not require the destination to be 643 initialized valid and new function EVP_MD_CTX_copy_ex() added which 644 requires the destination to be valid. 645 646 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 647 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 648 [Steve Henson] 649 650 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 651 so that complete 'Handshake' protocol structures are kept in memory 652 instead of overwriting 'msg_type' and 'length' with 'body' data. 653 [Bodo Moeller] 654 655 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 656 [Massimo Santin via Richard Levitte] 657 658 *) Major restructuring to the underlying ENGINE code. This includes 659 reduction of linker bloat, separation of pure "ENGINE" manipulation 660 (initialisation, etc) from functionality dealing with implementations 661 of specific crypto iterfaces. This change also introduces integrated 662 support for symmetric ciphers and digest implementations - so ENGINEs 663 can now accelerate these by providing EVP_CIPHER and EVP_MD 664 implementations of their own. This is detailed in crypto/engine/README 665 as it couldn't be adequately described here. However, there are a few 666 API changes worth noting - some RSA, DSA, DH, and RAND functions that 667 were changed in the original introduction of ENGINE code have now 668 reverted back - the hooking from this code to ENGINE is now a good 669 deal more passive and at run-time, operations deal directly with 670 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 671 dereferencing through an ENGINE pointer any more. Also, the ENGINE 672 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 673 they were not being used by the framework as there is no concept of a 674 BIGNUM_METHOD and they could not be generalised to the new 675 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 676 ENGINE_cpy() has been removed as it cannot be consistently defined in 677 the new code. 678 [Geoff Thorpe] 679 680 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 681 [Steve Henson] 682 683 *) Change mkdef.pl to sort symbols that get the same entry number, 684 and make sure the automatically generated functions ERR_load_* 685 become part of libeay.num as well. 686 [Richard Levitte] 687 688 *) New function SSL_renegotiate_pending(). This returns true once 689 renegotiation has been requested (either SSL_renegotiate() call 690 or HelloRequest/ClientHello receveived from the peer) and becomes 691 false once a handshake has been completed. 692 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 693 sends a HelloRequest, but does not ensure that a handshake takes 694 place. SSL_renegotiate_pending() is useful for checking if the 695 client has followed the request.) 696 [Bodo Moeller] 697 698 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 699 By default, clients may request session resumption even during 700 renegotiation (if session ID contexts permit); with this option, 701 session resumption is possible only in the first handshake. 702 703 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 704 more bits available for options that should not be part of 705 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 706 [Bodo Moeller] 707 708 *) Add some demos for certificate and certificate request creation. 709 [Steve Henson] 710 711 *) Make maximum certificate chain size accepted from the peer application 712 settable (SSL*_get/set_max_cert_list()), as proposed by 713 "Douglas E. Engert" <deengert@anl.gov>. 714 [Lutz Jaenicke] 715 716 *) Add support for shared libraries for Unixware-7 717 (Boyd Lynn Gerber <gerberb@zenez.com>). 718 [Lutz Jaenicke] 719 720 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 721 be done prior to destruction. Use this to unload error strings from 722 ENGINEs that load their own error strings. NB: This adds two new API 723 functions to "get" and "set" this destroy handler in an ENGINE. 724 [Geoff Thorpe] 725 726 *) Alter all existing ENGINE implementations (except "openssl" and 727 "openbsd") to dynamically instantiate their own error strings. This 728 makes them more flexible to be built both as statically-linked ENGINEs 729 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 730 Also, add stub code to each that makes building them as self-contained 731 shared-libraries easier (see README.ENGINE). 732 [Geoff Thorpe] 733 734 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 735 implementations into applications that are completely implemented in 736 self-contained shared-libraries. The "dynamic" ENGINE exposes control 737 commands that can be used to configure what shared-library to load and 738 to control aspects of the way it is handled. Also, made an update to 739 the README.ENGINE file that brings its information up-to-date and 740 provides some information and instructions on the "dynamic" ENGINE 741 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 742 [Geoff Thorpe] 743 744 *) Make it possible to unload ranges of ERR strings with a new 745 "ERR_unload_strings" function. 746 [Geoff Thorpe] 747 748 *) Add a copy() function to EVP_MD. 749 [Ben Laurie] 750 751 *) Make EVP_MD routines take a context pointer instead of just the 752 md_data void pointer. 753 [Ben Laurie] 754 755 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 756 that the digest can only process a single chunk of data 757 (typically because it is provided by a piece of 758 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 759 is only going to provide a single chunk of data, and hence the 760 framework needn't accumulate the data for oneshot drivers. 761 [Ben Laurie] 762 763 *) As with "ERR", make it possible to replace the underlying "ex_data" 764 functions. This change also alters the storage and management of global 765 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 766 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 767 index counters. The API functions that use this state have been changed 768 to take a "class_index" rather than pointers to the class's local STACK 769 and counter, and there is now an API function to dynamically create new 770 classes. This centralisation allows us to (a) plug a lot of the 771 thread-safety problems that existed, and (b) makes it possible to clean 772 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 773 such data would previously have always leaked in application code and 774 workarounds were in place to make the memory debugging turn a blind eye 775 to it. Application code that doesn't use this new function will still 776 leak as before, but their memory debugging output will announce it now 777 rather than letting it slide. 778 779 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 780 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 781 has a return value to indicate success or failure. 782 [Geoff Thorpe] 783 784 *) Make it possible to replace the underlying "ERR" functions such that the 785 global state (2 LHASH tables and 2 locks) is only used by the "default" 786 implementation. This change also adds two functions to "get" and "set" 787 the implementation prior to it being automatically set the first time 788 any other ERR function takes place. Ie. an application can call "get", 789 pass the return value to a module it has just loaded, and that module 790 can call its own "set" function using that value. This means the 791 module's "ERR" operations will use (and modify) the error state in the 792 application and not in its own statically linked copy of OpenSSL code. 793 [Geoff Thorpe] 794 795 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 796 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 797 the operation, and provides a more encapsulated way for external code 798 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 799 to use these functions rather than manually incrementing the counts. 800 801 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 802 [Geoff Thorpe] 803 804 *) Add EVP test program. 805 [Ben Laurie] 806 807 *) Add symmetric cipher support to ENGINE. Expect the API to change! 808 [Ben Laurie] 809 810 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 811 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 812 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 813 These allow a CRL to be built without having to access X509_CRL fields 814 directly. Modify 'ca' application to use new functions. 815 [Steve Henson] 816 817 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 818 bug workarounds. Rollback attack detection is a security feature. 819 The problem will only arise on OpenSSL servers when TLSv1 is not 820 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 821 Software authors not wanting to support TLSv1 will have special reasons 822 for their choice and can explicitly enable this option. 823 [Bodo Moeller, Lutz Jaenicke] 824 825 *) Rationalise EVP so it can be extended: don't include a union of 826 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 827 (similar to those existing for EVP_CIPHER_CTX). 828 Usage example: 829 830 EVP_MD_CTX md; 831 832 EVP_MD_CTX_init(&md); /* new function call */ 833 EVP_DigestInit(&md, EVP_sha1()); 834 EVP_DigestUpdate(&md, in, len); 835 EVP_DigestFinal(&md, out, NULL); 836 EVP_MD_CTX_cleanup(&md); /* new function call */ 837 838 [Ben Laurie] 839 840 *) Make DES key schedule conform to the usual scheme, as well as 841 correcting its structure. This means that calls to DES functions 842 now have to pass a pointer to a des_key_schedule instead of a 843 plain des_key_schedule (which was actually always a pointer 844 anyway): E.g., 845 846 des_key_schedule ks; 847 848 des_set_key_checked(..., &ks); 849 des_ncbc_encrypt(..., &ks, ...); 850 851 (Note that a later change renames 'des_...' into 'DES_...'.) 852 [Ben Laurie] 853 854 *) Initial reduction of linker bloat: the use of some functions, such as 855 PEM causes large amounts of unused functions to be linked in due to 856 poor organisation. For example pem_all.c contains every PEM function 857 which has a knock on effect of linking in large amounts of (unused) 858 ASN1 code. Grouping together similar functions and splitting unrelated 859 functions prevents this. 860 [Steve Henson] 861 862 *) Cleanup of EVP macros. 863 [Ben Laurie] 864 865 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 866 correct _ecb suffix. 867 [Ben Laurie] 868 869 *) Add initial OCSP responder support to ocsp application. The 870 revocation information is handled using the text based index 871 use by the ca application. The responder can either handle 872 requests generated internally, supplied in files (for example 873 via a CGI script) or using an internal minimal server. 874 [Steve Henson] 875 876 *) Add configuration choices to get zlib compression for TLS. 877 [Richard Levitte] 878 879 *) Changes to Kerberos SSL for RFC 2712 compliance: 880 1. Implemented real KerberosWrapper, instead of just using 881 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 882 2. Implemented optional authenticator field of KerberosWrapper. 883 884 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 885 and authenticator structs; see crypto/krb5/. 886 887 Generalized Kerberos calls to support multiple Kerberos libraries. 888 [Vern Staats <staatsvr@asc.hpc.mil>, 889 Jeffrey Altman <jaltman@columbia.edu> 890 via Richard Levitte] 891 892 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 893 already does with RSA. testdsa.h now has 'priv_key/pub_key' 894 values for each of the key sizes rather than having just 895 parameters (and 'speed' generating keys each time). 896 [Geoff Thorpe] 897 898 *) Speed up EVP routines. 899 Before: 900encrypt 901type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 902des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 903des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 904des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 905decrypt 906des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 907des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 908des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 909 After: 910encrypt 911des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 912decrypt 913des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 914 [Ben Laurie] 915 916 *) Added the OS2-EMX target. 917 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 918 919 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 920 to support NCONF routines in extension code. New function CONF_set_nconf() 921 to allow functions which take an NCONF to also handle the old LHASH 922 structure: this means that the old CONF compatible routines can be 923 retained (in particular wrt extensions) without having to duplicate the 924 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 925 [Steve Henson] 926 927 *) Enhance the general user interface with mechanisms for inner control 928 and with possibilities to have yes/no kind of prompts. 929 [Richard Levitte] 930 931 *) Change all calls to low level digest routines in the library and 932 applications to use EVP. Add missing calls to HMAC_cleanup() and 933 don't assume HMAC_CTX can be copied using memcpy(). 934 [Verdon Walker <VWalker@novell.com>, Steve Henson] 935 936 *) Add the possibility to control engines through control names but with 937 arbitrary arguments instead of just a string. 938 Change the key loaders to take a UI_METHOD instead of a callback 939 function pointer. NOTE: this breaks binary compatibility with earlier 940 versions of OpenSSL [engine]. 941 Adapt the nCipher code for these new conditions and add a card insertion 942 callback. 943 [Richard Levitte] 944 945 *) Enhance the general user interface with mechanisms to better support 946 dialog box interfaces, application-defined prompts, the possibility 947 to use defaults (for example default passwords from somewhere else) 948 and interrupts/cancellations. 949 [Richard Levitte] 950 951 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 952 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 953 [Steve Henson] 954 955 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 956 tidy up some unnecessarily weird code in 'sk_new()'). 957 [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 958 959 *) Change the key loading routines for ENGINEs to use the same kind 960 callback (pem_password_cb) as all other routines that need this 961 kind of callback. 962 [Richard Levitte] 963 964 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 965 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 966 than this minimum value is recommended. 967 [Lutz Jaenicke] 968 969 *) New random seeder for OpenVMS, using the system process statistics 970 that are easily reachable. 971 [Richard Levitte] 972 973 *) Windows apparently can't transparently handle global 974 variables defined in DLLs. Initialisations such as: 975 976 const ASN1_ITEM *it = &ASN1_INTEGER_it; 977 978 wont compile. This is used by the any applications that need to 979 declare their own ASN1 modules. This was fixed by adding the option 980 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 981 needed for static libraries under Win32. 982 [Steve Henson] 983 984 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 985 setting of purpose and trust fields. New X509_STORE trust and 986 purpose functions and tidy up setting in other SSL functions. 987 [Steve Henson] 988 989 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 990 structure. These are inherited by X509_STORE_CTX when it is 991 initialised. This allows various defaults to be set in the 992 X509_STORE structure (such as flags for CRL checking and custom 993 purpose or trust settings) for functions which only use X509_STORE_CTX 994 internally such as S/MIME. 995 996 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 997 trust settings if they are not set in X509_STORE. This allows X509_STORE 998 purposes and trust (in S/MIME for example) to override any set by default. 999 1000 Add command line options for CRL checking to smime, s_client and s_server 1001 applications. 1002 [Steve Henson] 1003 1004 *) Initial CRL based revocation checking. If the CRL checking flag(s) 1005 are set then the CRL is looked up in the X509_STORE structure and 1006 its validity and signature checked, then if the certificate is found 1007 in the CRL the verify fails with a revoked error. 1008 1009 Various new CRL related callbacks added to X509_STORE_CTX structure. 1010 1011 Command line options added to 'verify' application to support this. 1012 1013 This needs some additional work, such as being able to handle multiple 1014 CRLs with different times, extension based lookup (rather than just 1015 by subject name) and ultimately more complete V2 CRL extension 1016 handling. 1017 [Steve Henson] 1018 1019 *) Add a general user interface API (crypto/ui/). This is designed 1020 to replace things like des_read_password and friends (backward 1021 compatibility functions using this new API are provided). 1022 The purpose is to remove prompting functions from the DES code 1023 section as well as provide for prompting through dialog boxes in 1024 a window system and the like. 1025 [Richard Levitte] 1026 1027 *) Add "ex_data" support to ENGINE so implementations can add state at a 1028 per-structure level rather than having to store it globally. 1029 [Geoff] 1030 1031 *) Make it possible for ENGINE structures to be copied when retrieved by 1032 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 1033 This causes the "original" ENGINE structure to act like a template, 1034 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 1035 operational state can be localised to each ENGINE structure, despite the 1036 fact they all share the same "methods". New ENGINE structures returned in 1037 this case have no functional references and the return value is the single 1038 structural reference. This matches the single structural reference returned 1039 by ENGINE_by_id() normally, when it is incremented on the pre-existing 1040 ENGINE structure. 1041 [Geoff] 1042 1043 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 1044 needs to match any other type at all we need to manually clear the 1045 tag cache. 1046 [Steve Henson] 1047 1048 *) Changes to the "openssl engine" utility to include; 1049 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 1050 about an ENGINE's available control commands. 1051 - executing control commands from command line arguments using the 1052 '-pre' and '-post' switches. '-post' is only used if '-t' is 1053 specified and the ENGINE is successfully initialised. The syntax for 1054 the individual commands are colon-separated, for example; 1055 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 1056 [Geoff] 1057 1058 *) New dynamic control command support for ENGINEs. ENGINEs can now 1059 declare their own commands (numbers), names (strings), descriptions, 1060 and input types for run-time discovery by calling applications. A 1061 subset of these commands are implicitly classed as "executable" 1062 depending on their input type, and only these can be invoked through 1063 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 1064 can be based on user input, config files, etc). The distinction is 1065 that "executable" commands cannot return anything other than a boolean 1066 result and can only support numeric or string input, whereas some 1067 discoverable commands may only be for direct use through 1068 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 1069 pointers, or other custom uses. The "executable" commands are to 1070 support parameterisations of ENGINE behaviour that can be 1071 unambiguously defined by ENGINEs and used consistently across any 1072 OpenSSL-based application. Commands have been added to all the 1073 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 1074 control over shared-library paths without source code alterations. 1075 [Geoff] 1076 1077 *) Changed all ENGINE implementations to dynamically allocate their 1078 ENGINEs rather than declaring them statically. Apart from this being 1079 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 1080 this also allows the implementations to compile without using the 1081 internal engine_int.h header. 1082 [Geoff] 1083 1084 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 1085 'const' value. Any code that should be able to modify a RAND_METHOD 1086 should already have non-const pointers to it (ie. they should only 1087 modify their own ones). 1088 [Geoff] 1089 1090 *) Made a variety of little tweaks to the ENGINE code. 1091 - "atalla" and "ubsec" string definitions were moved from header files 1092 to C code. "nuron" string definitions were placed in variables 1093 rather than hard-coded - allowing parameterisation of these values 1094 later on via ctrl() commands. 1095 - Removed unused "#if 0"'d code. 1096 - Fixed engine list iteration code so it uses ENGINE_free() to release 1097 structural references. 1098 - Constified the RAND_METHOD element of ENGINE structures. 1099 - Constified various get/set functions as appropriate and added 1100 missing functions (including a catch-all ENGINE_cpy that duplicates 1101 all ENGINE values onto a new ENGINE except reference counts/state). 1102 - Removed NULL parameter checks in get/set functions. Setting a method 1103 or function to NULL is a way of cancelling out a previously set 1104 value. Passing a NULL ENGINE parameter is just plain stupid anyway 1105 and doesn't justify the extra error symbols and code. 1106 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 1107 flags from engine_int.h to engine.h. 1108 - Changed prototypes for ENGINE handler functions (init(), finish(), 1109 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 1110 [Geoff] 1111 1112 *) Implement binary inversion algorithm for BN_mod_inverse in addition 1113 to the algorithm using long division. The binary algorithm can be 1114 used only if the modulus is odd. On 32-bit systems, it is faster 1115 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 1116 roughly 5-15% for 256-bit moduli), so we use it only for moduli 1117 up to 450 bits. In 64-bit environments, the binary algorithm 1118 appears to be advantageous for much longer moduli; here we use it 1119 for moduli up to 2048 bits. 1120 [Bodo Moeller] 1121 1122 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 1123 could not support the combine flag in choice fields. 1124 [Steve Henson] 1125 1126 *) Add a 'copy_extensions' option to the 'ca' utility. This copies 1127 extensions from a certificate request to the certificate. 1128 [Steve Henson] 1129 1130 *) Allow multiple 'certopt' and 'nameopt' options to be separated 1131 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 1132 file: this allows the display of the certificate about to be 1133 signed to be customised, to allow certain fields to be included 1134 or excluded and extension details. The old system didn't display 1135 multicharacter strings properly, omitted fields not in the policy 1136 and couldn't display additional details such as extensions. 1137 [Steve Henson] 1138 1139 *) Function EC_POINTs_mul for multiple scalar multiplication 1140 of an arbitrary number of elliptic curve points 1141 \sum scalars[i]*points[i], 1142 optionally including the generator defined for the EC_GROUP: 1143 scalar*generator + \sum scalars[i]*points[i]. 1144 1145 EC_POINT_mul is a simple wrapper function for the typical case 1146 that the point list has just one item (besides the optional 1147 generator). 1148 [Bodo Moeller] 1149 1150 *) First EC_METHODs for curves over GF(p): 1151 1152 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 1153 operations and provides various method functions that can also 1154 operate with faster implementations of modular arithmetic. 1155 1156 EC_GFp_mont_method() reuses most functions that are part of 1157 EC_GFp_simple_method, but uses Montgomery arithmetic. 1158 1159 [Bodo Moeller; point addition and point doubling 1160 implementation directly derived from source code provided by 1161 Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 1162 1163 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 1164 crypto/ec/ec_lib.c): 1165 1166 Curves are EC_GROUP objects (with an optional group generator) 1167 based on EC_METHODs that are built into the library. 1168 1169 Points are EC_POINT objects based on EC_GROUP objects. 1170 1171 Most of the framework would be able to handle curves over arbitrary 1172 finite fields, but as there are no obvious types for fields other 1173 than GF(p), some functions are limited to that for now. 1174 [Bodo Moeller] 1175 1176 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 1177 that the file contains a complete HTTP response. 1178 [Richard Levitte] 1179 1180 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 1181 change the def and num file printf format specifier from "%-40sXXX" 1182 to "%-39s XXX". The latter will always guarantee a space after the 1183 field while the former will cause them to run together if the field 1184 is 40 of more characters long. 1185 [Steve Henson] 1186 1187 *) Constify the cipher and digest 'method' functions and structures 1188 and modify related functions to take constant EVP_MD and EVP_CIPHER 1189 pointers. 1190 [Steve Henson] 1191 1192 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 1193 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 1194 [Bodo Moeller] 1195 1196 *) Modify EVP_Digest*() routines so they now return values. Although the 1197 internal software routines can never fail additional hardware versions 1198 might. 1199 [Steve Henson] 1200 1201 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 1202 1203 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 1204 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 1205 1206 ASN1 error codes 1207 ERR_R_NESTED_ASN1_ERROR 1208 ... 1209 ERR_R_MISSING_ASN1_EOS 1210 were 4 .. 9, conflicting with 1211 ERR_LIB_RSA (= ERR_R_RSA_LIB) 1212 ... 1213 ERR_LIB_PEM (= ERR_R_PEM_LIB). 1214 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 1215 1216 Add new error code 'ERR_R_INTERNAL_ERROR'. 1217 [Bodo Moeller] 1218 1219 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 1220 suffices. 1221 [Bodo Moeller] 1222 1223 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 1224 sets the subject name for a new request or supersedes the 1225 subject name in a given request. Formats that can be parsed are 1226 'CN=Some Name, OU=myOU, C=IT' 1227 and 1228 'CN=Some Name/OU=myOU/C=IT'. 1229 1230 Add options '-batch' and '-verbose' to 'openssl req'. 1231 [Massimiliano Pala <madwolf@hackmasters.net>] 1232 1233 *) Introduce the possibility to access global variables through 1234 functions on platform were that's the best way to handle exporting 1235 global variables in shared libraries. To enable this functionality, 1236 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 1237 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 1238 is normally done by Configure or something similar). 1239 1240 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 1241 in the source file (foo.c) like this: 1242 1243 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 1244 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 1245 1246 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 1247 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 1248 1249 OPENSSL_DECLARE_GLOBAL(int,foo); 1250 #define foo OPENSSL_GLOBAL_REF(foo) 1251 OPENSSL_DECLARE_GLOBAL(double,bar); 1252 #define bar OPENSSL_GLOBAL_REF(bar) 1253 1254 The #defines are very important, and therefore so is including the 1255 header file everywhere where the defined globals are used. 1256 1257 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 1258 of ASN.1 items, but that structure is a bit different. 1259 1260 The largest change is in util/mkdef.pl which has been enhanced with 1261 better and easier to understand logic to choose which symbols should 1262 go into the Windows .def files as well as a number of fixes and code 1263 cleanup (among others, algorithm keywords are now sorted 1264 lexicographically to avoid constant rewrites). 1265 [Richard Levitte] 1266 1267 *) In BN_div() keep a copy of the sign of 'num' before writing the 1268 result to 'rm' because if rm==num the value will be overwritten 1269 and produce the wrong result if 'num' is negative: this caused 1270 problems with BN_mod() and BN_nnmod(). 1271 [Steve Henson] 1272 1273 *) Function OCSP_request_verify(). This checks the signature on an 1274 OCSP request and verifies the signer certificate. The signer 1275 certificate is just checked for a generic purpose and OCSP request 1276 trust settings. 1277 [Steve Henson] 1278 1279 *) Add OCSP_check_validity() function to check the validity of OCSP 1280 responses. OCSP responses are prepared in real time and may only 1281 be a few seconds old. Simply checking that the current time lies 1282 between thisUpdate and nextUpdate max reject otherwise valid responses 1283 caused by either OCSP responder or client clock inaccuracy. Instead 1284 we allow thisUpdate and nextUpdate to fall within a certain period of 1285 the current time. The age of the response can also optionally be 1286 checked. Two new options -validity_period and -status_age added to 1287 ocsp utility. 1288 [Steve Henson] 1289 1290 *) If signature or public key algorithm is unrecognized print out its 1291 OID rather that just UNKNOWN. 1292 [Steve Henson] 1293 1294 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 1295 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 1296 ID to be generated from the issuer certificate alone which can then be 1297 passed to OCSP_id_issuer_cmp(). 1298 [Steve Henson] 1299 1300 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 1301 ASN1 modules to export functions returning ASN1_ITEM pointers 1302 instead of the ASN1_ITEM structures themselves. This adds several 1303 new macros which allow the underlying ASN1 function/structure to 1304 be accessed transparently. As a result code should not use ASN1_ITEM 1305 references directly (such as &X509_it) but instead use the relevant 1306 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 1307 use of the new ASN1 code on platforms where exporting structures 1308 is problematical (for example in shared libraries) but exporting 1309 functions returning pointers to structures is not. 1310 [Steve Henson] 1311 1312 *) Add support for overriding the generation of SSL/TLS session IDs. 1313 These callbacks can be registered either in an SSL_CTX or per SSL. 1314 The purpose of this is to allow applications to control, if they wish, 1315 the arbitrary values chosen for use as session IDs, particularly as it 1316 can be useful for session caching in multiple-server environments. A 1317 command-line switch for testing this (and any client code that wishes 1318 to use such a feature) has been added to "s_server". 1319 [Geoff Thorpe, Lutz Jaenicke] 1320 1321 *) Modify mkdef.pl to recognise and parse preprocessor conditionals 1322 of the form '#if defined(...) || defined(...) || ...' and 1323 '#if !defined(...) && !defined(...) && ...'. This also avoids 1324 the growing number of special cases it was previously handling. 1325 [Richard Levitte] 1326 1327 *) Make all configuration macros available for application by making 1328 sure they are available in opensslconf.h, by giving them names starting 1329 with "OPENSSL_" to avoid conflicts with other packages and by making 1330 sure e_os2.h will cover all platform-specific cases together with 1331 opensslconf.h. 1332 Additionally, it is now possible to define configuration/platform- 1333 specific names (called "system identities"). In the C code, these 1334 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 1335 macro with the name beginning with "OPENSSL_SYS_", which is determined 1336 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 1337 what is available. 1338 [Richard Levitte] 1339 1340 *) New option -set_serial to 'req' and 'x509' this allows the serial 1341 number to use to be specified on the command line. Previously self 1342 signed certificates were hard coded with serial number 0 and the 1343 CA options of 'x509' had to use a serial number in a file which was 1344 auto incremented. 1345 [Steve Henson] 1346 1347 *) New options to 'ca' utility to support V2 CRL entry extensions. 1348 Currently CRL reason, invalidity date and hold instruction are 1349 supported. Add new CRL extensions to V3 code and some new objects. 1350 [Steve Henson] 1351 1352 *) New function EVP_CIPHER_CTX_set_padding() this is used to 1353 disable standard block padding (aka PKCS#5 padding) in the EVP 1354 API, which was previously mandatory. This means that the data is 1355 not padded in any way and so the total length much be a multiple 1356 of the block size, otherwise an error occurs. 1357 [Steve Henson] 1358 1359 *) Initial (incomplete) OCSP SSL support. 1360 [Steve Henson] 1361 1362 *) New function OCSP_parse_url(). This splits up a URL into its host, 1363 port and path components: primarily to parse OCSP URLs. New -url 1364 option to ocsp utility. 1365 [Steve Henson] 1366 1367 *) New nonce behavior. The return value of OCSP_check_nonce() now 1368 reflects the various checks performed. Applications can decide 1369 whether to tolerate certain situations such as an absent nonce 1370 in a response when one was present in a request: the ocsp application 1371 just prints out a warning. New function OCSP_add1_basic_nonce() 1372 this is to allow responders to include a nonce in a response even if 1373 the request is nonce-less. 1374 [Steve Henson] 1375 1376 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 1377 skipped when using openssl x509 multiple times on a single input file, 1378 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 1379 [Bodo Moeller] 1380 1381 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 1382 set string type: to handle setting ASN1_TIME structures. Fix ca 1383 utility to correctly initialize revocation date of CRLs. 1384 [Steve Henson] 1385 1386 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 1387 the clients preferred ciphersuites and rather use its own preferences. 1388 Should help to work around M$ SGC (Server Gated Cryptography) bug in 1389 Internet Explorer by ensuring unchanged hash method during stepup. 1390 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 1391 [Lutz Jaenicke] 1392 1393 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 1394 to aes and add a new 'exist' option to print out symbols that don't 1395 appear to exist. 1396 [Steve Henson] 1397 1398 *) Additional options to ocsp utility to allow flags to be set and 1399 additional certificates supplied. 1400 [Steve Henson] 1401 1402 *) Add the option -VAfile to 'openssl ocsp', so the user can give the 1403 OCSP client a number of certificate to only verify the response 1404 signature against. 1405 [Richard Levitte] 1406 1407 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 1408 handle the new API. Currently only ECB, CBC modes supported. Add new 1409 AES OIDs. 1410 1411 Add TLS AES ciphersuites as described in RFC3268, "Advanced 1412 Encryption Standard (AES) Ciphersuites for Transport Layer 1413 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 1414 not enabled by default and were not part of the "ALL" ciphersuite 1415 alias because they were not yet official; they could be 1416 explicitly requested by specifying the "AESdraft" ciphersuite 1417 group alias. In the final release of OpenSSL 0.9.7, the group 1418 alias is called "AES" and is part of "ALL".) 1419 [Ben Laurie, Steve Henson, Bodo Moeller] 1420 1421 *) New function OCSP_copy_nonce() to copy nonce value (if present) from 1422 request to response. 1423 [Steve Henson] 1424 1425 *) Functions for OCSP responders. OCSP_request_onereq_count(), 1426 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 1427 extract information from a certificate request. OCSP_response_create() 1428 creates a response and optionally adds a basic response structure. 1429 OCSP_basic_add1_status() adds a complete single response to a basic 1430 response and returns the OCSP_SINGLERESP structure just added (to allow 1431 extensions to be included for example). OCSP_basic_add1_cert() adds a 1432 certificate to a basic response and OCSP_basic_sign() signs a basic 1433 response with various flags. New helper functions ASN1_TIME_check() 1434 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 1435 (converts ASN1_TIME to GeneralizedTime). 1436 [Steve Henson] 1437 1438 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 1439 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 1440 structure from a certificate. X509_pubkey_digest() digests the public_key 1441 contents: this is used in various key identifiers. 1442 [Steve Henson] 1443 1444 *) Make sk_sort() tolerate a NULL argument. 1445 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 1446 1447 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 1448 passed by the function are trusted implicitly. If any of them signed the 1449 response then it is assumed to be valid and is not verified. 1450 [Steve Henson] 1451 1452 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 1453 to data. This was previously part of the PKCS7 ASN1 code. This 1454 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 1455 [Steve Henson, reported by Kenneth R. Robinette 1456 <support@securenetterm.com>] 1457 1458 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 1459 routines: without these tracing memory leaks is very painful. 1460 Fix leaks in PKCS12 and PKCS7 routines. 1461 [Steve Henson] 1462 1463 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 1464 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 1465 effectively meant GeneralizedTime would never be used. Now it 1466 is initialised to -1 but X509_time_adj() now has to check the value 1467 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 1468 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 1469 [Steve Henson, reported by Kenneth R. Robinette 1470 <support@securenetterm.com>] 1471 1472 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 1473 result in a zero length in the ASN1_INTEGER structure which was 1474 not consistent with the structure when d2i_ASN1_INTEGER() was used 1475 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 1476 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 1477 where it did not print out a minus for negative ASN1_INTEGER. 1478 [Steve Henson] 1479 1480 *) Add summary printout to ocsp utility. The various functions which 1481 convert status values to strings have been renamed to: 1482 OCSP_response_status_str(), OCSP_cert_status_str() and 1483 OCSP_crl_reason_str() and are no longer static. New options 1484 to verify nonce values and to disable verification. OCSP response 1485 printout format cleaned up. 1486 [Steve Henson] 1487 1488 *) Add additional OCSP certificate checks. These are those specified 1489 in RFC2560. This consists of two separate checks: the CA of the 1490 certificate being checked must either be the OCSP signer certificate 1491 or the issuer of the OCSP signer certificate. In the latter case the 1492 OCSP signer certificate must contain the OCSP signing extended key 1493 usage. This check is performed by attempting to match the OCSP 1494 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 1495 in the OCSP_CERTID structures of the response. 1496 [Steve Henson] 1497 1498 *) Initial OCSP certificate verification added to OCSP_basic_verify() 1499 and related routines. This uses the standard OpenSSL certificate 1500 verify routines to perform initial checks (just CA validity) and 1501 to obtain the certificate chain. Then additional checks will be 1502 performed on the chain. Currently the root CA is checked to see 1503 if it is explicitly trusted for OCSP signing. This is used to set 1504 a root CA as a global signing root: that is any certificate that 1505 chains to that CA is an acceptable OCSP signing certificate. 1506 [Steve Henson] 1507 1508 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 1509 extensions from a separate configuration file. 1510 As when reading extensions from the main configuration file, 1511 the '-extensions ...' option may be used for specifying the 1512 section to use. 1513 [Massimiliano Pala <madwolf@comune.modena.it>] 1514 1515 *) New OCSP utility. Allows OCSP requests to be generated or 1516 read. The request can be sent to a responder and the output 1517 parsed, outputed or printed in text form. Not complete yet: 1518 still needs to check the OCSP response validity. 1519 [Steve Henson] 1520 1521 *) New subcommands for 'openssl ca': 1522 'openssl ca -status <serial>' prints the status of the cert with 1523 the given serial number (according to the index file). 1524 'openssl ca -updatedb' updates the expiry status of certificates 1525 in the index file. 1526 [Massimiliano Pala <madwolf@comune.modena.it>] 1527 1528 *) New '-newreq-nodes' command option to CA.pl. This is like 1529 '-newreq', but calls 'openssl req' with the '-nodes' option 1530 so that the resulting key is not encrypted. 1531 [Damien Miller <djm@mindrot.org>] 1532 1533 *) New configuration for the GNU Hurd. 1534 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 1535 1536 *) Initial code to implement OCSP basic response verify. This 1537 is currently incomplete. Currently just finds the signer's 1538 certificate and verifies the signature on the response. 1539 [Steve Henson] 1540 1541 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 1542 value of OPENSSLDIR. This is available via the new '-d' option 1543 to 'openssl version', and is also included in 'openssl version -a'. 1544 [Bodo Moeller] 1545 1546 *) Allowing defining memory allocation callbacks that will be given 1547 file name and line number information in additional arguments 1548 (a const char* and an int). The basic functionality remains, as 1549 well as the original possibility to just replace malloc(), 1550 realloc() and free() by functions that do not know about these 1551 additional arguments. To register and find out the current 1552 settings for extended allocation functions, the following 1553 functions are provided: 1554 1555 CRYPTO_set_mem_ex_functions 1556 CRYPTO_set_locked_mem_ex_functions 1557 CRYPTO_get_mem_ex_functions 1558 CRYPTO_get_locked_mem_ex_functions 1559 1560 These work the same way as CRYPTO_set_mem_functions and friends. 1561 CRYPTO_get_[locked_]mem_functions now writes 0 where such an 1562 extended allocation function is enabled. 1563 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 1564 a conventional allocation function is enabled. 1565 [Richard Levitte, Bodo Moeller] 1566 1567 *) Finish off removing the remaining LHASH function pointer casts. 1568 There should no longer be any prototype-casting required when using 1569 the LHASH abstraction, and any casts that remain are "bugs". See 1570 the callback types and macros at the head of lhash.h for details 1571 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 1572 [Geoff Thorpe] 1573 1574 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 1575 If /dev/[u]random devices are not available or do not return enough 1576 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 1577 be queried. 1578 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 1579 /etc/entropy will be queried once each in this sequence, quering stops 1580 when enough entropy was collected without querying more sockets. 1581 [Lutz Jaenicke] 1582 1583 *) Change the Unix RAND_poll() variant to be able to poll several 1584 random devices, as specified by DEVRANDOM, until a sufficient amount 1585 of data has been collected. We spend at most 10 ms on each file 1586 (select timeout) and read in non-blocking mode. DEVRANDOM now 1587 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 1588 (previously it was just the string "/dev/urandom"), so on typical 1589 platforms the 10 ms delay will never occur. 1590 Also separate out the Unix variant to its own file, rand_unix.c. 1591 For VMS, there's a currently-empty rand_vms.c. 1592 [Richard Levitte] 1593 1594 *) Move OCSP client related routines to ocsp_cl.c. These 1595 provide utility functions which an application needing 1596 to issue a request to an OCSP responder and analyse the 1597 response will typically need: as opposed to those which an 1598 OCSP responder itself would need which will be added later. 1599 1600 OCSP_request_sign() signs an OCSP request with an API similar 1601 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 1602 response. OCSP_response_get1_basic() extracts basic response 1603 from response. OCSP_resp_find_status(): finds and extracts status 1604 information from an OCSP_CERTID structure (which will be created 1605 when the request structure is built). These are built from lower 1606 level functions which work on OCSP_SINGLERESP structures but 1607 wont normally be used unless the application wishes to examine 1608 extensions in the OCSP response for example. 1609 1610 Replace nonce routines with a pair of functions. 1611 OCSP_request_add1_nonce() adds a nonce value and optionally 1612 generates a random value. OCSP_check_nonce() checks the 1613 validity of the nonce in an OCSP response. 1614 [Steve Henson] 1615 1616 *) Change function OCSP_request_add() to OCSP_request_add0_id(). 1617 This doesn't copy the supplied OCSP_CERTID and avoids the 1618 need to free up the newly created id. Change return type 1619 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 1620 This can then be used to add extensions to the request. 1621 Deleted OCSP_request_new(), since most of its functionality 1622 is now in OCSP_REQUEST_new() (and the case insensitive name 1623 clash) apart from the ability to set the request name which 1624 will be added elsewhere. 1625 [Steve Henson] 1626 1627 *) Update OCSP API. Remove obsolete extensions argument from 1628 various functions. Extensions are now handled using the new 1629 OCSP extension code. New simple OCSP HTTP function which 1630 can be used to send requests and parse the response. 1631 [Steve Henson] 1632 1633 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 1634 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 1635 uses the special reorder version of SET OF to sort the attributes 1636 and reorder them to match the encoded order. This resolves a long 1637 standing problem: a verify on a PKCS7 structure just after signing 1638 it used to fail because the attribute order did not match the 1639 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 1640 it uses the received order. This is necessary to tolerate some broken 1641 software that does not order SET OF. This is handled by encoding 1642 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 1643 to produce the required SET OF. 1644 [Steve Henson] 1645 1646 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 1647 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 1648 files to get correct declarations of the ASN.1 item variables. 1649 [Richard Levitte] 1650 1651 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 1652 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 1653 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 1654 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 1655 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 1656 ASN1_ITEM and no wrapper functions. 1657 [Steve Henson] 1658 1659 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 1660 replace the old function pointer based I/O routines. Change most of 1661 the *_d2i_bio() and *_d2i_fp() functions to use these. 1662 [Steve Henson] 1663 1664 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 1665 lines, recognice more "algorithms" that can be deselected, and make 1666 it complain about algorithm deselection that isn't recognised. 1667 [Richard Levitte] 1668 1669 *) New ASN1 functions to handle dup, sign, verify, digest, pack and 1670 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 1671 to use new functions. Add NO_ASN1_OLD which can be set to remove 1672 some old style ASN1 functions: this can be used to determine if old 1673 code will still work when these eventually go away. 1674 [Steve Henson] 1675 1676 *) New extension functions for OCSP structures, these follow the 1677 same conventions as certificates and CRLs. 1678 [Steve Henson] 1679 1680 *) New function X509V3_add1_i2d(). This automatically encodes and 1681 adds an extension. Its behaviour can be customised with various 1682 flags to append, replace or delete. Various wrappers added for 1683 certifcates and CRLs. 1684 [Steve Henson] 1685 1686 *) Fix to avoid calling the underlying ASN1 print routine when 1687 an extension cannot be parsed. Correct a typo in the 1688 OCSP_SERVICELOC extension. Tidy up print OCSP format. 1689 [Steve Henson] 1690 1691 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 1692 entries for variables. 1693 [Steve Henson] 1694 1695 *) Add functionality to apps/openssl.c for detecting locking 1696 problems: As the program is single-threaded, all we have 1697 to do is register a locking callback using an array for 1698 storing which locks are currently held by the program. 1699 [Bodo Moeller] 1700 1701 *) Use a lock around the call to CRYPTO_get_ex_new_index() in 1702 SSL_get_ex_data_X509_STORE_idx(), which is used in 1703 ssl_verify_cert_chain() and thus can be called at any time 1704 during TLS/SSL handshakes so that thread-safety is essential. 1705 Unfortunately, the ex_data design is not at all suited 1706 for multi-threaded use, so it probably should be abolished. 1707 [Bodo Moeller] 1708 1709 *) Added Broadcom "ubsec" ENGINE to OpenSSL. 1710 [Broadcom, tweaked and integrated by Geoff Thorpe] 1711 1712 *) Move common extension printing code to new function 1713 X509V3_print_extensions(). Reorganise OCSP print routines and 1714 implement some needed OCSP ASN1 functions. Add OCSP extensions. 1715 [Steve Henson] 1716 1717 *) New function X509_signature_print() to remove duplication in some 1718 print routines. 1719 [Steve Henson] 1720 1721 *) Add a special meaning when SET OF and SEQUENCE OF flags are both 1722 set (this was treated exactly the same as SET OF previously). This 1723 is used to reorder the STACK representing the structure to match the 1724 encoding. This will be used to get round a problem where a PKCS7 1725 structure which was signed could not be verified because the STACK 1726 order did not reflect the encoded order. 1727 [Steve Henson] 1728 1729 *) Reimplement the OCSP ASN1 module using the new code. 1730 [Steve Henson] 1731 1732 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 1733 for its ASN1 operations. The old style function pointers still exist 1734 for now but they will eventually go away. 1735 [Steve Henson] 1736 1737 *) Merge in replacement ASN1 code from the ASN1 branch. This almost 1738 completely replaces the old ASN1 functionality with a table driven 1739 encoder and decoder which interprets an ASN1_ITEM structure describing 1740 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 1741 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 1742 has also been converted to the new form. 1743 [Steve Henson] 1744 1745 *) Change BN_mod_exp_recp so that negative moduli are tolerated 1746 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 1747 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 1748 for negative moduli. 1749 [Bodo Moeller] 1750 1751 *) Fix BN_uadd and BN_usub: Always return non-negative results instead 1752 of not touching the result's sign bit. 1753 [Bodo Moeller] 1754 1755 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 1756 set. 1757 [Bodo Moeller] 1758 1759 *) Changed the LHASH code to use prototypes for callbacks, and created 1760 macros to declare and implement thin (optionally static) functions 1761 that provide type-safety and avoid function pointer casting for the 1762 type-specific callbacks. 1763 [Geoff Thorpe] 1764 1765 *) Added Kerberos Cipher Suites to be used with TLS, as written in 1766 RFC 2712. 1767 [Veers Staats <staatsvr@asc.hpc.mil>, 1768 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 1769 1770 *) Reformat the FAQ so the different questions and answers can be divided 1771 in sections depending on the subject. 1772 [Richard Levitte] 1773 1774 *) Have the zlib compression code load ZLIB.DLL dynamically under 1775 Windows. 1776 [Richard Levitte] 1777 1778 *) New function BN_mod_sqrt for computing square roots modulo a prime 1779 (using the probabilistic Tonelli-Shanks algorithm unless 1780 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 1781 be handled deterministically). 1782 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1783 1784 *) Make BN_mod_inverse faster by explicitly handling small quotients 1785 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 1786 512 bits], about 30% for larger ones [1024 or 2048 bits].) 1787 [Bodo Moeller] 1788 1789 *) New function BN_kronecker. 1790 [Bodo Moeller] 1791 1792 *) Fix BN_gcd so that it works on negative inputs; the result is 1793 positive unless both parameters are zero. 1794 Previously something reasonably close to an infinite loop was 1795 possible because numbers could be growing instead of shrinking 1796 in the implementation of Euclid's algorithm. 1797 [Bodo Moeller] 1798 1799 *) Fix BN_is_word() and BN_is_one() macros to take into account the 1800 sign of the number in question. 1801 1802 Fix BN_is_word(a,w) to work correctly for w == 0. 1803 1804 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 1805 because its test if the absolute value of 'a' equals 'w'. 1806 Note that BN_abs_is_word does *not* handle w == 0 reliably; 1807 it exists mostly for use in the implementations of BN_is_zero(), 1808 BN_is_one(), and BN_is_word(). 1809 [Bodo Moeller] 1810 1811 *) New function BN_swap. 1812 [Bodo Moeller] 1813 1814 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 1815 the exponentiation functions are more likely to produce reasonable 1816 results on negative inputs. 1817 [Bodo Moeller] 1818 1819 *) Change BN_mod_mul so that the result is always non-negative. 1820 Previously, it could be negative if one of the factors was negative; 1821 I don't think anyone really wanted that behaviour. 1822 [Bodo Moeller] 1823 1824 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 1825 (except for exponentiation, which stays in crypto/bn/bn_exp.c, 1826 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 1827 and add new functions: 1828 1829 BN_nnmod 1830 BN_mod_sqr 1831 BN_mod_add 1832 BN_mod_add_quick 1833 BN_mod_sub 1834 BN_mod_sub_quick 1835 BN_mod_lshift1 1836 BN_mod_lshift1_quick 1837 BN_mod_lshift 1838 BN_mod_lshift_quick 1839 1840 These functions always generate non-negative results. 1841 1842 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 1843 such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 1844 1845 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 1846 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 1847 be reduced modulo m. 1848 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1849 1850#if 0 1851 The following entry accidentily appeared in the CHANGES file 1852 distributed with OpenSSL 0.9.7. The modifications described in 1853 it do *not* apply to OpenSSL 0.9.7. 1854 1855 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1856 was actually never needed) and in BN_mul(). The removal in BN_mul() 1857 required a small change in bn_mul_part_recursive() and the addition 1858 of the functions bn_cmp_part_words(), bn_sub_part_words() and 1859 bn_add_part_words(), which do the same thing as bn_cmp_words(), 1860 bn_sub_words() and bn_add_words() except they take arrays with 1861 differing sizes. 1862 [Richard Levitte] 1863#endif 1864 1865 *) In 'openssl passwd', verify passwords read from the terminal 1866 unless the '-salt' option is used (which usually means that 1867 verification would just waste user's time since the resulting 1868 hash is going to be compared with some given password hash) 1869 or the new '-noverify' option is used. 1870 1871 This is an incompatible change, but it does not affect 1872 non-interactive use of 'openssl passwd' (passwords on the command 1873 line, '-stdin' option, '-in ...' option) and thus should not 1874 cause any problems. 1875 [Bodo Moeller] 1876 1877 *) Remove all references to RSAref, since there's no more need for it. 1878 [Richard Levitte] 1879 1880 *) Make DSO load along a path given through an environment variable 1881 (SHLIB_PATH) with shl_load(). 1882 [Richard Levitte] 1883 1884 *) Constify the ENGINE code as a result of BIGNUM constification. 1885 Also constify the RSA code and most things related to it. In a 1886 few places, most notable in the depth of the ASN.1 code, ugly 1887 casts back to non-const were required (to be solved at a later 1888 time) 1889 [Richard Levitte] 1890 1891 *) Make it so the openssl application has all engines loaded by default. 1892 [Richard Levitte] 1893 1894 *) Constify the BIGNUM routines a little more. 1895 [Richard Levitte] 1896 1897 *) Add the following functions: 1898 1899 ENGINE_load_cswift() 1900 ENGINE_load_chil() 1901 ENGINE_load_atalla() 1902 ENGINE_load_nuron() 1903 ENGINE_load_builtin_engines() 1904 1905 That way, an application can itself choose if external engines that 1906 are built-in in OpenSSL shall ever be used or not. The benefit is 1907 that applications won't have to be linked with libdl or other dso 1908 libraries unless it's really needed. 1909 1910 Changed 'openssl engine' to load all engines on demand. 1911 Changed the engine header files to avoid the duplication of some 1912 declarations (they differed!). 1913 [Richard Levitte] 1914 1915 *) 'openssl engine' can now list capabilities. 1916 [Richard Levitte] 1917 1918 *) Better error reporting in 'openssl engine'. 1919 [Richard Levitte] 1920 1921 *) Never call load_dh_param(NULL) in s_server. 1922 [Bodo Moeller] 1923 1924 *) Add engine application. It can currently list engines by name and 1925 identity, and test if they are actually available. 1926 [Richard Levitte] 1927 1928 *) Improve RPM specification file by forcing symbolic linking and making 1929 sure the installed documentation is also owned by root.root. 1930 [Damien Miller <djm@mindrot.org>] 1931 1932 *) Give the OpenSSL applications more possibilities to make use of 1933 keys (public as well as private) handled by engines. 1934 [Richard Levitte] 1935 1936 *) Add OCSP code that comes from CertCo. 1937 [Richard Levitte] 1938 1939 *) Add VMS support for the Rijndael code. 1940 [Richard Levitte] 1941 1942 *) Added untested support for Nuron crypto accelerator. 1943 [Ben Laurie] 1944 1945 *) Add support for external cryptographic devices. This code was 1946 previously distributed separately as the "engine" branch. 1947 [Geoff Thorpe, Richard Levitte] 1948 1949 *) Rework the filename-translation in the DSO code. It is now possible to 1950 have far greater control over how a "name" is turned into a filename 1951 depending on the operating environment and any oddities about the 1952 different shared library filenames on each system. 1953 [Geoff Thorpe] 1954 1955 *) Support threads on FreeBSD-elf in Configure. 1956 [Richard Levitte] 1957 1958 *) Fix for SHA1 assembly problem with MASM: it produces 1959 warnings about corrupt line number information when assembling 1960 with debugging information. This is caused by the overlapping 1961 of two sections. 1962 [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 1963 1964 *) NCONF changes. 1965 NCONF_get_number() has no error checking at all. As a replacement, 1966 NCONF_get_number_e() is defined (_e for "error checking") and is 1967 promoted strongly. The old NCONF_get_number is kept around for 1968 binary backward compatibility. 1969 Make it possible for methods to load from something other than a BIO, 1970 by providing a function pointer that is given a name instead of a BIO. 1971 For example, this could be used to load configuration data from an 1972 LDAP server. 1973 [Richard Levitte] 1974 1975 *) Fix for non blocking accept BIOs. Added new I/O special reason 1976 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 1977 with non blocking I/O was not possible because no retry code was 1978 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 1979 this case. 1980 [Steve Henson] 1981 1982 *) Added the beginnings of Rijndael support. 1983 [Ben Laurie] 1984 1985 *) Fix for bug in DirectoryString mask setting. Add support for 1986 X509_NAME_print_ex() in 'req' and X509_print_ex() function 1987 to allow certificate printing to more controllable, additional 1988 'certopt' option to 'x509' to allow new printing options to be 1989 set. 1990 [Steve Henson] 1991 1992 *) Clean old EAY MD5 hack from e_os.h. 1993 [Richard Levitte] 1994 1995 Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 1996 1997 *) Fix various bugs revealed by running the NISCC test suite: 1998 1999 Stop out of bounds reads in the ASN1 code when presented with 2000 invalid tags (CAN-2003-0543 and CAN-2003-0544). 2001 2002 If verify callback ignores invalid public key errors don't try to check 2003 certificate signature with the NULL public key. 2004 2005 [Steve Henson] 2006 2007 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 2008 if the server requested one: as stated in TLS 1.0 and SSL 3.0 2009 specifications. 2010 [Steve Henson] 2011 2012 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 2013 extra data after the compression methods not only for TLS 1.0 2014 but also for SSL 3.0 (as required by the specification). 2015 [Bodo Moeller; problem pointed out by Matthias Loepfe] 2016 2017 *) Change X509_certificate_type() to mark the key as exported/exportable 2018 when it's 512 *bits* long, not 512 bytes. 2019 [Richard Levitte] 2020 2021 Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 2022 2023 *) Countermeasure against the Klima-Pokorny-Rosa extension of 2024 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 2025 a protocol version number mismatch like a decryption error 2026 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 2027 [Bodo Moeller] 2028 2029 *) Turn on RSA blinding by default in the default implementation 2030 to avoid a timing attack. Applications that don't want it can call 2031 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 2032 They would be ill-advised to do so in most cases. 2033 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 2034 2035 *) Change RSA blinding code so that it works when the PRNG is not 2036 seeded (in this case, the secret RSA exponent is abused as 2037 an unpredictable seed -- if it is not unpredictable, there 2038 is no point in blinding anyway). Make RSA blinding thread-safe 2039 by remembering the creator's thread ID in rsa->blinding and 2040 having all other threads use local one-time blinding factors 2041 (this requires more computation than sharing rsa->blinding, but 2042 avoids excessive locking; and if an RSA object is not shared 2043 between threads, blinding will still be very fast). 2044 [Bodo Moeller] 2045 2046 Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 2047 2048 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 2049 via timing by performing a MAC computation even if incorrrect 2050 block cipher padding has been found. This is a countermeasure 2051 against active attacks where the attacker has to distinguish 2052 between bad padding and a MAC verification error. (CAN-2003-0078) 2053 2054 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 2055 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 2056 Martin Vuagnoux (EPFL, Ilion)] 2057 2058 Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 2059 2060 *) New function OPENSSL_cleanse(), which is used to cleanse a section of 2061 memory from it's contents. This is done with a counter that will 2062 place alternating values in each byte. This can be used to solve 2063 two issues: 1) the removal of calls to memset() by highly optimizing 2064 compilers, and 2) cleansing with other values than 0, since those can 2065 be read through on certain media, for example a swap space on disk. 2066 [Geoff Thorpe] 2067 2068 *) Bugfix: client side session caching did not work with external caching, 2069 because the session->cipher setting was not restored when reloading 2070 from the external cache. This problem was masked, when 2071 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 2072 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 2073 [Lutz Jaenicke] 2074 2075 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 2076 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 2077 [Zeev Lieber <zeev-l@yahoo.com>] 2078 2079 *) Undo an undocumented change introduced in 0.9.6e which caused 2080 repeated calls to OpenSSL_add_all_ciphers() and 2081 OpenSSL_add_all_digests() to be ignored, even after calling 2082 EVP_cleanup(). 2083 [Richard Levitte] 2084 2085 *) Change the default configuration reader to deal with last line not 2086 being properly terminated. 2087 [Richard Levitte] 2088 2089 *) Change X509_NAME_cmp() so it applies the special rules on handling 2090 DN values that are of type PrintableString, as well as RDNs of type 2091 emailAddress where the value has the type ia5String. 2092 [stefank@valicert.com via Richard Levitte] 2093 2094 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 2095 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 2096 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 2097 the bitwise-OR of the two for use by the majority of applications 2098 wanting this behaviour, and update the docs. The documented 2099 behaviour and actual behaviour were inconsistent and had been 2100 changing anyway, so this is more a bug-fix than a behavioural 2101 change. 2102 [Geoff Thorpe, diagnosed by Nadav Har'El] 2103 2104 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 2105 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 2106 [Bodo Moeller] 2107 2108 *) Fix initialization code race conditions in 2109 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 2110 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 2111 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 2112 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 2113 ssl2_get_cipher_by_char(), 2114 ssl3_get_cipher_by_char(). 2115 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 2116 2117 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 2118 the cached sessions are flushed, as the remove_cb() might use ex_data 2119 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 2120 (see [openssl.org #212]). 2121 [Geoff Thorpe, Lutz Jaenicke] 2122 2123 *) Fix typo in OBJ_txt2obj which incorrectly passed the content 2124 length, instead of the encoding length to d2i_ASN1_OBJECT. 2125 [Steve Henson] 2126 2127 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 2128 2129 *) [In 0.9.6g-engine release:] 2130 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 2131 [Lynn Gazis <lgazis@rainbow.com>] 2132 2133 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 2134 2135 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 2136 and get fix the header length calculation. 2137 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 2138 Alon Kantor <alonk@checkpoint.com> (and others), 2139 Steve Henson] 2140 2141 *) Use proper error handling instead of 'assertions' in buffer 2142 overflow checks added in 0.9.6e. This prevents DoS (the 2143 assertions could call abort()). 2144 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 2145 2146 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 2147 2148 *) Add various sanity checks to asn1_get_length() to reject 2149 the ASN1 length bytes if they exceed sizeof(long), will appear 2150 negative or the content length exceeds the length of the 2151 supplied buffer. 2152 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2153 2154 *) Fix cipher selection routines: ciphers without encryption had no flags 2155 for the cipher strength set and where therefore not handled correctly 2156 by the selection routines (PR #130). 2157 [Lutz Jaenicke] 2158 2159 *) Fix EVP_dsa_sha macro. 2160 [Nils Larsch] 2161 2162 *) New option 2163 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 2164 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 2165 that was added in OpenSSL 0.9.6d. 2166 2167 As the countermeasure turned out to be incompatible with some 2168 broken SSL implementations, the new option is part of SSL_OP_ALL. 2169 SSL_OP_ALL is usually employed when compatibility with weird SSL 2170 implementations is desired (e.g. '-bugs' option to 's_client' and 2171 's_server'), so the new option is automatically set in many 2172 applications. 2173 [Bodo Moeller] 2174 2175 *) Changes in security patch: 2176 2177 Changes marked "(CHATS)" were sponsored by the Defense Advanced 2178 Research Projects Agency (DARPA) and Air Force Research Laboratory, 2179 Air Force Materiel Command, USAF, under agreement number 2180 F30602-01-2-0537. 2181 2182 *) Add various sanity checks to asn1_get_length() to reject 2183 the ASN1 length bytes if they exceed sizeof(long), will appear 2184 negative or the content length exceeds the length of the 2185 supplied buffer. (CAN-2002-0659) 2186 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2187 2188 *) Assertions for various potential buffer overflows, not known to 2189 happen in practice. 2190 [Ben Laurie (CHATS)] 2191 2192 *) Various temporary buffers to hold ASCII versions of integers were 2193 too small for 64 bit platforms. (CAN-2002-0655) 2194 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 2195 2196 *) Remote buffer overflow in SSL3 protocol - an attacker could 2197 supply an oversized session ID to a client. (CAN-2002-0656) 2198 [Ben Laurie (CHATS)] 2199 2200 *) Remote buffer overflow in SSL2 protocol - an attacker could 2201 supply an oversized client master key. (CAN-2002-0656) 2202 [Ben Laurie (CHATS)] 2203 2204 Changes between 0.9.6c and 0.9.6d [9 May 2002] 2205 2206 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 2207 encoded as NULL) with id-dsa-with-sha1. 2208 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 2209 2210 *) Check various X509_...() return values in apps/req.c. 2211 [Nils Larsch <nla@trustcenter.de>] 2212 2213 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 2214 an end-of-file condition would erronously be flagged, when the CRLF 2215 was just at the end of a processed block. The bug was discovered when 2216 processing data through a buffering memory BIO handing the data to a 2217 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 2218 <ptsekov@syntrex.com> and Nedelcho Stanev. 2219 [Lutz Jaenicke] 2220 2221 *) Implement a countermeasure against a vulnerability recently found 2222 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 2223 before application data chunks to avoid the use of known IVs 2224 with data potentially chosen by the attacker. 2225 [Bodo Moeller] 2226 2227 *) Fix length checks in ssl3_get_client_hello(). 2228 [Bodo Moeller] 2229 2230 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 2231 to prevent ssl3_read_internal() from incorrectly assuming that 2232 ssl3_read_bytes() found application data while handshake 2233 processing was enabled when in fact s->s3->in_read_app_data was 2234 merely automatically cleared during the initial handshake. 2235 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 2236 2237 *) Fix object definitions for Private and Enterprise: they were not 2238 recognized in their shortname (=lowercase) representation. Extend 2239 obj_dat.pl to issue an error when using undefined keywords instead 2240 of silently ignoring the problem (Svenning Sorensen 2241 <sss@sss.dnsalias.net>). 2242 [Lutz Jaenicke] 2243 2244 *) Fix DH_generate_parameters() so that it works for 'non-standard' 2245 generators, i.e. generators other than 2 and 5. (Previously, the 2246 code did not properly initialise the 'add' and 'rem' values to 2247 BN_generate_prime().) 2248 2249 In the new general case, we do not insist that 'generator' is 2250 actually a primitive root: This requirement is rather pointless; 2251 a generator of the order-q subgroup is just as good, if not 2252 better. 2253 [Bodo Moeller] 2254 2255 *) Map new X509 verification errors to alerts. Discovered and submitted by 2256 Tom Wu <tom@arcot.com>. 2257 [Lutz Jaenicke] 2258 2259 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 2260 returning non-zero before the data has been completely received 2261 when using non-blocking I/O. 2262 [Bodo Moeller; problem pointed out by John Hughes] 2263 2264 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 2265 [Ben Laurie, Lutz Jaenicke] 2266 2267 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 2268 Yoram Zahavi <YoramZ@gilian.com>). 2269 [Lutz Jaenicke] 2270 2271 *) Add information about CygWin 1.3 and on, and preserve proper 2272 configuration for the versions before that. 2273 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2274 2275 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 2276 check whether we deal with a copy of a session and do not delete from 2277 the cache in this case. Problem reported by "Izhar Shoshani Levi" 2278 <izhar@checkpoint.com>. 2279 [Lutz Jaenicke] 2280 2281 *) Do not store session data into the internal session cache, if it 2282 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 2283 flag is set). Proposed by Aslam <aslam@funk.com>. 2284 [Lutz Jaenicke] 2285 2286 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 2287 value is 0. 2288 [Richard Levitte] 2289 2290 *) [In 0.9.6d-engine release:] 2291 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2292 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2293 2294 *) Add the configuration target linux-s390x. 2295 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 2296 2297 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 2298 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 2299 variable as an indication that a ClientHello message has been 2300 received. As the flag value will be lost between multiple 2301 invocations of ssl3_accept when using non-blocking I/O, the 2302 function may not be aware that a handshake has actually taken 2303 place, thus preventing a new session from being added to the 2304 session cache. 2305 2306 To avoid this problem, we now set s->new_session to 2 instead of 2307 using a local variable. 2308 [Lutz Jaenicke, Bodo Moeller] 2309 2310 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 2311 if the SSL_R_LENGTH_MISMATCH error is detected. 2312 [Geoff Thorpe, Bodo Moeller] 2313 2314 *) New 'shared_ldflag' column in Configure platform table. 2315 [Richard Levitte] 2316 2317 *) Fix EVP_CIPHER_mode macro. 2318 ["Dan S. Camper" <dan@bti.net>] 2319 2320 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 2321 type, we must throw them away by setting rr->length to 0. 2322 [D P Chang <dpc@qualys.com>] 2323 2324 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 2325 2326 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 2327 <Dominikus.Scherkl@biodata.com>. (The previous implementation 2328 worked incorrectly for those cases where range = 10..._2 and 2329 3*range is two bits longer than range.) 2330 [Bodo Moeller] 2331 2332 *) Only add signing time to PKCS7 structures if it is not already 2333 present. 2334 [Steve Henson] 2335 2336 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 2337 OBJ_ld_ce should be OBJ_id_ce. 2338 Also some ip-pda OIDs in crypto/objects/objects.txt were 2339 incorrect (cf. RFC 3039). 2340 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 2341 2342 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 2343 returns early because it has nothing to do. 2344 [Andy Schneider <andy.schneider@bjss.co.uk>] 2345 2346 *) [In 0.9.6c-engine release:] 2347 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 2348 [Andy Schneider <andy.schneider@bjss.co.uk>] 2349 2350 *) [In 0.9.6c-engine release:] 2351 Add support for Cryptographic Appliance's keyserver technology. 2352 (Use engine 'keyclient') 2353 [Cryptographic Appliances and Geoff Thorpe] 2354 2355 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 2356 is called via tools/c89.sh because arguments have to be 2357 rearranged (all '-L' options must appear before the first object 2358 modules). 2359 [Richard Shapiro <rshapiro@abinitio.com>] 2360 2361 *) [In 0.9.6c-engine release:] 2362 Add support for Broadcom crypto accelerator cards, backported 2363 from 0.9.7. 2364 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 2365 2366 *) [In 0.9.6c-engine release:] 2367 Add support for SureWare crypto accelerator cards from 2368 Baltimore Technologies. (Use engine 'sureware') 2369 [Baltimore Technologies and Mark Cox] 2370 2371 *) [In 0.9.6c-engine release:] 2372 Add support for crypto accelerator cards from Accelerated 2373 Encryption Processing, www.aep.ie. (Use engine 'aep') 2374 [AEP Inc. and Mark Cox] 2375 2376 *) Add a configuration entry for gcc on UnixWare. 2377 [Gary Benson <gbenson@redhat.com>] 2378 2379 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 2380 messages are stored in a single piece (fixed-length part and 2381 variable-length part combined) and fix various bugs found on the way. 2382 [Bodo Moeller] 2383 2384 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 2385 instead. BIO_gethostbyname() does not know what timeouts are 2386 appropriate, so entries would stay in cache even when they have 2387 become invalid. 2388 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 2389 2390 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 2391 faced with a pathologically small ClientHello fragment that does 2392 not contain client_version: Instead of aborting with an error, 2393 simply choose the highest available protocol version (i.e., 2394 TLS 1.0 unless it is disabled). In practice, ClientHello 2395 messages are never sent like this, but this change gives us 2396 strictly correct behaviour at least for TLS. 2397 [Bodo Moeller] 2398 2399 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 2400 never resets s->method to s->ctx->method when called from within 2401 one of the SSL handshake functions. 2402 [Bodo Moeller; problem pointed out by Niko Baric] 2403 2404 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 2405 (sent using the client's version number) if client_version is 2406 smaller than the protocol version in use. Also change 2407 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 2408 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 2409 the client will at least see that alert. 2410 [Bodo Moeller] 2411 2412 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 2413 correctly. 2414 [Bodo Moeller] 2415 2416 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 2417 client receives HelloRequest while in a handshake. 2418 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 2419 2420 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 2421 should end in 'break', not 'goto end' which circuments various 2422 cleanups done in state SSL_ST_OK. But session related stuff 2423 must be disabled for SSL_ST_OK in the case that we just sent a 2424 HelloRequest. 2425 2426 Also avoid some overhead by not calling ssl_init_wbio_buffer() 2427 before just sending a HelloRequest. 2428 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 2429 2430 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 2431 reveal whether illegal block cipher padding was found or a MAC 2432 verification error occured. (Neither SSLerr() codes nor alerts 2433 are directly visible to potential attackers, but the information 2434 may leak via logfiles.) 2435 2436 Similar changes are not required for the SSL 2.0 implementation 2437 because the number of padding bytes is sent in clear for SSL 2.0, 2438 and the extra bytes are just ignored. However ssl/s2_pkt.c 2439 failed to verify that the purported number of padding bytes is in 2440 the legal range. 2441 [Bodo Moeller] 2442 2443 *) Add OpenUNIX-8 support including shared libraries 2444 (Boyd Lynn Gerber <gerberb@zenez.com>). 2445 [Lutz Jaenicke] 2446 2447 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 2448 'wristwatch attack' using huge encoding parameters (cf. 2449 James H. Manger's CRYPTO 2001 paper). Note that the 2450 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 2451 encoding parameters and hence was not vulnerable. 2452 [Bodo Moeller] 2453 2454 *) BN_sqr() bug fix. 2455 [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 2456 2457 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 2458 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 2459 followed by modular reduction. 2460 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 2461 2462 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 2463 equivalent based on BN_pseudo_rand() instead of BN_rand(). 2464 [Bodo Moeller] 2465 2466 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 2467 This function was broken, as the check for a new client hello message 2468 to handle SGC did not allow these large messages. 2469 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 2470 [Lutz Jaenicke] 2471 2472 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 2473 [Lutz Jaenicke] 2474 2475 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 2476 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 2477 [Lutz Jaenicke] 2478 2479 *) Rework the configuration and shared library support for Tru64 Unix. 2480 The configuration part makes use of modern compiler features and 2481 still retains old compiler behavior for those that run older versions 2482 of the OS. The shared library support part includes a variant that 2483 uses the RPATH feature, and is available through the special 2484 configuration target "alpha-cc-rpath", which will never be selected 2485 automatically. 2486 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 2487 2488 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 2489 with the same message size as in ssl3_get_certificate_request(). 2490 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 2491 messages might inadvertently be reject as too long. 2492 [Petr Lampa <lampa@fee.vutbr.cz>] 2493 2494 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 2495 [Andy Polyakov] 2496 2497 *) Modified SSL library such that the verify_callback that has been set 2498 specificly for an SSL object with SSL_set_verify() is actually being 2499 used. Before the change, a verify_callback set with this function was 2500 ignored and the verify_callback() set in the SSL_CTX at the time of 2501 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 2502 to allow the necessary settings. 2503 [Lutz Jaenicke] 2504 2505 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 2506 explicitly to NULL, as at least on Solaris 8 this seems not always to be 2507 done automatically (in contradiction to the requirements of the C 2508 standard). This made problems when used from OpenSSH. 2509 [Lutz Jaenicke] 2510 2511 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 2512 dh->length and always used 2513 2514 BN_rand_range(priv_key, dh->p). 2515 2516 BN_rand_range() is not necessary for Diffie-Hellman, and this 2517 specific range makes Diffie-Hellman unnecessarily inefficient if 2518 dh->length (recommended exponent length) is much smaller than the 2519 length of dh->p. We could use BN_rand_range() if the order of 2520 the subgroup was stored in the DH structure, but we only have 2521 dh->length. 2522 2523 So switch back to 2524 2525 BN_rand(priv_key, l, ...) 2526 2527 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 2528 otherwise. 2529 [Bodo Moeller] 2530 2531 *) In 2532 2533 RSA_eay_public_encrypt 2534 RSA_eay_private_decrypt 2535 RSA_eay_private_encrypt (signing) 2536 RSA_eay_public_decrypt (signature verification) 2537 2538 (default implementations for RSA_public_encrypt, 2539 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 2540 always reject numbers >= n. 2541 [Bodo Moeller] 2542 2543 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 2544 to synchronize access to 'locking_thread'. This is necessary on 2545 systems where access to 'locking_thread' (an 'unsigned long' 2546 variable) is not atomic. 2547 [Bodo Moeller] 2548 2549 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 2550 *before* setting the 'crypto_lock_rand' flag. The previous code had 2551 a race condition if 0 is a valid thread ID. 2552 [Travis Vitek <vitek@roguewave.com>] 2553 2554 *) Add support for shared libraries under Irix. 2555 [Albert Chin-A-Young <china@thewrittenword.com>] 2556 2557 *) Add configuration option to build on Linux on both big-endian and 2558 little-endian MIPS. 2559 [Ralf Baechle <ralf@uni-koblenz.de>] 2560 2561 *) Add the possibility to create shared libraries on HP-UX. 2562 [Richard Levitte] 2563 2564 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 2565 2566 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 2567 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 2568 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 2569 PRNG state recovery was possible based on the output of 2570 one PRNG request appropriately sized to gain knowledge on 2571 'md' followed by enough consecutive 1-byte PRNG requests 2572 to traverse all of 'state'. 2573 2574 1. When updating 'md_local' (the current thread's copy of 'md') 2575 during PRNG output generation, hash all of the previous 2576 'md_local' value, not just the half used for PRNG output. 2577 2578 2. Make the number of bytes from 'state' included into the hash 2579 independent from the number of PRNG bytes requested. 2580 2581 The first measure alone would be sufficient to avoid 2582 Markku-Juhani's attack. (Actually it had never occurred 2583 to me that the half of 'md_local' used for chaining was the 2584 half from which PRNG output bytes were taken -- I had always 2585 assumed that the secret half would be used.) The second 2586 measure makes sure that additional data from 'state' is never 2587 mixed into 'md_local' in small portions; this heuristically 2588 further strengthens the PRNG. 2589 [Bodo Moeller] 2590 2591 *) Fix crypto/bn/asm/mips3.s. 2592 [Andy Polyakov] 2593 2594 *) When only the key is given to "enc", the IV is undefined. Print out 2595 an error message in this case. 2596 [Lutz Jaenicke] 2597 2598 *) Handle special case when X509_NAME is empty in X509 printing routines. 2599 [Steve Henson] 2600 2601 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 2602 positive and less than q. 2603 [Bodo Moeller] 2604 2605 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 2606 used: it isn't thread safe and the add_lock_callback should handle 2607 that itself. 2608 [Paul Rose <Paul.Rose@bridge.com>] 2609 2610 *) Verify that incoming data obeys the block size in 2611 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 2612 [Bodo Moeller] 2613 2614 *) Fix OAEP check. 2615 [Ulf M�ller, Bodo M�ller] 2616 2617 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 2618 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 2619 when fixing the server behaviour for backwards-compatible 'client 2620 hello' messages. (Note that the attack is impractical against 2621 SSL 3.0 and TLS 1.0 anyway because length and version checking 2622 means that the probability of guessing a valid ciphertext is 2623 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 2624 paper.) 2625 2626 Before 0.9.5, the countermeasure (hide the error by generating a 2627 random 'decryption result') did not work properly because 2628 ERR_clear_error() was missing, meaning that SSL_get_error() would 2629 detect the supposedly ignored error. 2630 2631 Both problems are now fixed. 2632 [Bodo Moeller] 2633 2634 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 2635 (previously it was 1024). 2636 [Bodo Moeller] 2637 2638 *) Fix for compatibility mode trust settings: ignore trust settings 2639 unless some valid trust or reject settings are present. 2640 [Steve Henson] 2641 2642 *) Fix for blowfish EVP: its a variable length cipher. 2643 [Steve Henson] 2644 2645 *) Fix various bugs related to DSA S/MIME verification. Handle missing 2646 parameters in DSA public key structures and return an error in the 2647 DSA routines if parameters are absent. 2648 [Steve Henson] 2649 2650 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 2651 in the current directory if neither $RANDFILE nor $HOME was set. 2652 RAND_file_name() in 0.9.6a returned NULL in this case. This has 2653 caused some confusion to Windows users who haven't defined $HOME. 2654 Thus RAND_file_name() is changed again: e_os.h can define a 2655 DEFAULT_HOME, which will be used if $HOME is not set. 2656 For Windows, we use "C:"; on other platforms, we still require 2657 environment variables. 2658 2659 *) Move 'if (!initialized) RAND_poll()' into regions protected by 2660 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 2661 having multiple threads call RAND_poll() concurrently. 2662 [Bodo Moeller] 2663 2664 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 2665 combination of a flag and a thread ID variable. 2666 Otherwise while one thread is in ssleay_rand_bytes (which sets the 2667 flag), *other* threads can enter ssleay_add_bytes without obeying 2668 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 2669 that they do not hold after the first thread unsets add_do_not_lock). 2670 [Bodo Moeller] 2671 2672 *) Change bctest again: '-x' expressions are not available in all 2673 versions of 'test'. 2674 [Bodo Moeller] 2675 2676 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 2677 2678 *) Fix a couple of memory leaks in PKCS7_dataDecode() 2679 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 2680 2681 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 2682 the default extension for executables, if any. Also, make the perl 2683 scripts that use symlink() to test if it really exists and use "cp" 2684 if it doesn't. All this made OpenSSL compilable and installable in 2685 CygWin. 2686 [Richard Levitte] 2687 2688 *) Fix for asn1_GetSequence() for indefinite length constructed data. 2689 If SEQUENCE is length is indefinite just set c->slen to the total 2690 amount of data available. 2691 [Steve Henson, reported by shige@FreeBSD.org] 2692 [This change does not apply to 0.9.7.] 2693 2694 *) Change bctest to avoid here-documents inside command substitution 2695 (workaround for FreeBSD /bin/sh bug). 2696 For compatibility with Ultrix, avoid shell functions (introduced 2697 in the bctest version that searches along $PATH). 2698 [Bodo Moeller] 2699 2700 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 2701 with des_encrypt() defined on some operating systems, like Solaris 2702 and UnixWare. 2703 [Richard Levitte] 2704 2705 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 2706 On the Importance of Eliminating Errors in Cryptographic 2707 Computations, J. Cryptology 14 (2001) 2, 101-119, 2708 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 2709 [Ulf Moeller] 2710 2711 *) MIPS assembler BIGNUM division bug fix. 2712 [Andy Polyakov] 2713 2714 *) Disabled incorrect Alpha assembler code. 2715 [Richard Levitte] 2716 2717 *) Fix PKCS#7 decode routines so they correctly update the length 2718 after reading an EOC for the EXPLICIT tag. 2719 [Steve Henson] 2720 [This change does not apply to 0.9.7.] 2721 2722 *) Fix bug in PKCS#12 key generation routines. This was triggered 2723 if a 3DES key was generated with a 0 initial byte. Include 2724 PKCS12_BROKEN_KEYGEN compilation option to retain the old 2725 (but broken) behaviour. 2726 [Steve Henson] 2727 2728 *) Enhance bctest to search for a working bc along $PATH and print 2729 it when found. 2730 [Tim Rice <tim@multitalents.net> via Richard Levitte] 2731 2732 *) Fix memory leaks in err.c: free err_data string if necessary; 2733 don't write to the wrong index in ERR_set_error_data. 2734 [Bodo Moeller] 2735 2736 *) Implement ssl23_peek (analogous to ssl23_read), which previously 2737 did not exist. 2738 [Bodo Moeller] 2739 2740 *) Replace rdtsc with _emit statements for VC++ version 5. 2741 [Jeremy Cooper <jeremy@baymoo.org>] 2742 2743 *) Make it possible to reuse SSLv2 sessions. 2744 [Richard Levitte] 2745 2746 *) In copy_email() check for >= 0 as a return value for 2747 X509_NAME_get_index_by_NID() since 0 is a valid index. 2748 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 2749 2750 *) Avoid coredump with unsupported or invalid public keys by checking if 2751 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 2752 PKCS7_verify() fails with non detached data. 2753 [Steve Henson] 2754 2755 *) Don't use getenv in library functions when run as setuid/setgid. 2756 New function OPENSSL_issetugid(). 2757 [Ulf Moeller] 2758 2759 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 2760 due to incorrect handling of multi-threading: 2761 2762 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 2763 2764 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 2765 2766 3. Count how many times MemCheck_off() has been called so that 2767 nested use can be treated correctly. This also avoids 2768 inband-signalling in the previous code (which relied on the 2769 assumption that thread ID 0 is impossible). 2770 [Bodo Moeller] 2771 2772 *) Add "-rand" option also to s_client and s_server. 2773 [Lutz Jaenicke] 2774 2775 *) Fix CPU detection on Irix 6.x. 2776 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 2777 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 2778 2779 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 2780 was empty. 2781 [Steve Henson] 2782 [This change does not apply to 0.9.7.] 2783 2784 *) Use the cached encoding of an X509_NAME structure rather than 2785 copying it. This is apparently the reason for the libsafe "errors" 2786 but the code is actually correct. 2787 [Steve Henson] 2788 2789 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 2790 Bleichenbacher's DSA attack. 2791 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 2792 to be set and top=0 forces the highest bit to be set; top=-1 is new 2793 and leaves the highest bit random. 2794 [Ulf Moeller, Bodo Moeller] 2795 2796 *) In the NCONF_...-based implementations for CONF_... queries 2797 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 2798 a temporary CONF structure with the data component set to NULL 2799 (which gives segmentation faults in lh_retrieve). 2800 Instead, use NULL for the CONF pointer in CONF_get_string and 2801 CONF_get_number (which may use environment variables) and directly 2802 return NULL from CONF_get_section. 2803 [Bodo Moeller] 2804 2805 *) Fix potential buffer overrun for EBCDIC. 2806 [Ulf Moeller] 2807 2808 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 2809 keyUsage if basicConstraints absent for a CA. 2810 [Steve Henson] 2811 2812 *) Make SMIME_write_PKCS7() write mail header values with a format that 2813 is more generally accepted (no spaces before the semicolon), since 2814 some programs can't parse those values properly otherwise. Also make 2815 sure BIO's that break lines after each write do not create invalid 2816 headers. 2817 [Richard Levitte] 2818 2819 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 2820 macros previously used would not encode an empty SEQUENCE OF 2821 and break the signature. 2822 [Steve Henson] 2823 [This change does not apply to 0.9.7.] 2824 2825 *) Zero the premaster secret after deriving the master secret in 2826 DH ciphersuites. 2827 [Steve Henson] 2828 2829 *) Add some EVP_add_digest_alias registrations (as found in 2830 OpenSSL_add_all_digests()) to SSL_library_init() 2831 aka OpenSSL_add_ssl_algorithms(). This provides improved 2832 compatibility with peers using X.509 certificates 2833 with unconventional AlgorithmIdentifier OIDs. 2834 [Bodo Moeller] 2835 2836 *) Fix for Irix with NO_ASM. 2837 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 2838 2839 *) ./config script fixes. 2840 [Ulf Moeller, Richard Levitte] 2841 2842 *) Fix 'openssl passwd -1'. 2843 [Bodo Moeller] 2844 2845 *) Change PKCS12_key_gen_asc() so it can cope with non null 2846 terminated strings whose length is passed in the passlen 2847 parameter, for example from PEM callbacks. This was done 2848 by adding an extra length parameter to asc2uni(). 2849 [Steve Henson, reported by <oddissey@samsung.co.kr>] 2850 2851 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 2852 call failed, free the DSA structure. 2853 [Bodo Moeller] 2854 2855 *) Fix to uni2asc() to cope with zero length Unicode strings. 2856 These are present in some PKCS#12 files. 2857 [Steve Henson] 2858 2859 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 2860 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 2861 when writing a 32767 byte record. 2862 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 2863 2864 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 2865 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 2866 2867 (RSA objects have a reference count access to which is protected 2868 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 2869 so they are meant to be shared between threads.) 2870 [Bodo Moeller, Geoff Thorpe; original patch submitted by 2871 "Reddie, Steven" <Steven.Reddie@ca.com>] 2872 2873 *) Fix a deadlock in CRYPTO_mem_leaks(). 2874 [Bodo Moeller] 2875 2876 *) Use better test patterns in bntest. 2877 [Ulf M�ller] 2878 2879 *) rand_win.c fix for Borland C. 2880 [Ulf M�ller] 2881 2882 *) BN_rshift bugfix for n == 0. 2883 [Bodo Moeller] 2884 2885 *) Add a 'bctest' script that checks for some known 'bc' bugs 2886 so that 'make test' does not abort just because 'bc' is broken. 2887 [Bodo Moeller] 2888 2889 *) Store verify_result within SSL_SESSION also for client side to 2890 avoid potential security hole. (Re-used sessions on the client side 2891 always resulted in verify_result==X509_V_OK, not using the original 2892 result of the server certificate verification.) 2893 [Lutz Jaenicke] 2894 2895 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 2896 SSL3_RT_APPLICATION_DATA, return 0. 2897 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 2898 [Bodo Moeller] 2899 2900 *) Fix SSL_peek: 2901 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 2902 releases, have been re-implemented by renaming the previous 2903 implementations of ssl2_read and ssl3_read to ssl2_read_internal 2904 and ssl3_read_internal, respectively, and adding 'peek' parameters 2905 to them. The new ssl[23]_{read,peek} functions are calls to 2906 ssl[23]_read_internal with the 'peek' flag set appropriately. 2907 A 'peek' parameter has also been added to ssl3_read_bytes, which 2908 does the actual work for ssl3_read_internal. 2909 [Bodo Moeller] 2910 2911 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 2912 the method-specific "init()" handler. Also clean up ex_data after 2913 calling the method-specific "finish()" handler. Previously, this was 2914 happening the other way round. 2915 [Geoff Thorpe] 2916 2917 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 2918 The previous value, 12, was not always sufficient for BN_mod_exp(). 2919 [Bodo Moeller] 2920 2921 *) Make sure that shared libraries get the internal name engine with 2922 the full version number and not just 0. This should mark the 2923 shared libraries as not backward compatible. Of course, this should 2924 be changed again when we can guarantee backward binary compatibility. 2925 [Richard Levitte] 2926 2927 *) Fix typo in get_cert_by_subject() in by_dir.c 2928 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 2929 2930 *) Rework the system to generate shared libraries: 2931 2932 - Make note of the expected extension for the shared libraries and 2933 if there is a need for symbolic links from for example libcrypto.so.0 2934 to libcrypto.so.0.9.7. There is extended info in Configure for 2935 that. 2936 2937 - Make as few rebuilds of the shared libraries as possible. 2938 2939 - Still avoid linking the OpenSSL programs with the shared libraries. 2940 2941 - When installing, install the shared libraries separately from the 2942 static ones. 2943 [Richard Levitte] 2944 2945 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 2946 2947 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 2948 and not in SSL_clear because the latter is also used by the 2949 accept/connect functions; previously, the settings made by 2950 SSL_set_read_ahead would be lost during the handshake. 2951 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 2952 2953 *) Correct util/mkdef.pl to be selective about disabled algorithms. 2954 Previously, it would create entries for disableed algorithms no 2955 matter what. 2956 [Richard Levitte] 2957 2958 *) Added several new manual pages for SSL_* function. 2959 [Lutz Jaenicke] 2960 2961 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 2962 2963 *) In ssl23_get_client_hello, generate an error message when faced 2964 with an initial SSL 3.0/TLS record that is too small to contain the 2965 first two bytes of the ClientHello message, i.e. client_version. 2966 (Note that this is a pathologic case that probably has never happened 2967 in real life.) The previous approach was to use the version number 2968 from the record header as a substitute; but our protocol choice 2969 should not depend on that one because it is not authenticated 2970 by the Finished messages. 2971 [Bodo Moeller] 2972 2973 *) More robust randomness gathering functions for Windows. 2974 [Jeffrey Altman <jaltman@columbia.edu>] 2975 2976 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 2977 not set then we don't setup the error code for issuer check errors 2978 to avoid possibly overwriting other errors which the callback does 2979 handle. If an application does set the flag then we assume it knows 2980 what it is doing and can handle the new informational codes 2981 appropriately. 2982 [Steve Henson] 2983 2984 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 2985 a general "ANY" type, as such it should be able to decode anything 2986 including tagged types. However it didn't check the class so it would 2987 wrongly interpret tagged types in the same way as their universal 2988 counterpart and unknown types were just rejected. Changed so that the 2989 tagged and unknown types are handled in the same way as a SEQUENCE: 2990 that is the encoding is stored intact. There is also a new type 2991 "V_ASN1_OTHER" which is used when the class is not universal, in this 2992 case we have no idea what the actual type is so we just lump them all 2993 together. 2994 [Steve Henson] 2995 2996 *) On VMS, stdout may very well lead to a file that is written to 2997 in a record-oriented fashion. That means that every write() will 2998 write a separate record, which will be read separately by the 2999 programs trying to read from it. This can be very confusing. 3000 3001 The solution is to put a BIO filter in the way that will buffer 3002 text until a linefeed is reached, and then write everything a 3003 line at a time, so every record written will be an actual line, 3004 not chunks of lines and not (usually doesn't happen, but I've 3005 seen it once) several lines in one record. BIO_f_linebuffer() is 3006 the answer. 3007 3008 Currently, it's a VMS-only method, because that's where it has 3009 been tested well enough. 3010 [Richard Levitte] 3011 3012 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 3013 it can return incorrect results. 3014 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 3015 but it was in 0.9.6-beta[12].) 3016 [Bodo Moeller] 3017 3018 *) Disable the check for content being present when verifying detached 3019 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 3020 include zero length content when signing messages. 3021 [Steve Henson] 3022 3023 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 3024 BIO_ctrl (for BIO pairs). 3025 [Bodo M�ller] 3026 3027 *) Add DSO method for VMS. 3028 [Richard Levitte] 3029 3030 *) Bug fix: Montgomery multiplication could produce results with the 3031 wrong sign. 3032 [Ulf M�ller] 3033 3034 *) Add RPM specification openssl.spec and modify it to build three 3035 packages. The default package contains applications, application 3036 documentation and run-time libraries. The devel package contains 3037 include files, static libraries and function documentation. The 3038 doc package contains the contents of the doc directory. The original 3039 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 3040 [Richard Levitte] 3041 3042 *) Add a large number of documentation files for many SSL routines. 3043 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 3044 3045 *) Add a configuration entry for Sony News 4. 3046 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 3047 3048 *) Don't set the two most significant bits to one when generating a 3049 random number < q in the DSA library. 3050 [Ulf M�ller] 3051 3052 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 3053 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 3054 the underlying transport is blocking) if a handshake took place. 3055 (The default behaviour is needed by applications such as s_client 3056 and s_server that use select() to determine when to use SSL_read; 3057 but for applications that know in advance when to expect data, it 3058 just makes things more complicated.) 3059 [Bodo Moeller] 3060 3061 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 3062 from EGD. 3063 [Ben Laurie] 3064 3065 *) Add a few more EBCDIC conditionals that make `req' and `x509' 3066 work better on such systems. 3067 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 3068 3069 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 3070 Update PKCS12_parse() so it copies the friendlyName and the 3071 keyid to the certificates aux info. 3072 [Steve Henson] 3073 3074 *) Fix bug in PKCS7_verify() which caused an infinite loop 3075 if there was more than one signature. 3076 [Sven Uszpelkat <su@celocom.de>] 3077 3078 *) Major change in util/mkdef.pl to include extra information 3079 about each symbol, as well as presentig variables as well 3080 as functions. This change means that there's n more need 3081 to rebuild the .num files when some algorithms are excluded. 3082 [Richard Levitte] 3083 3084 *) Allow the verify time to be set by an application, 3085 rather than always using the current time. 3086 [Steve Henson] 3087 3088 *) Phase 2 verify code reorganisation. The certificate 3089 verify code now looks up an issuer certificate by a 3090 number of criteria: subject name, authority key id 3091 and key usage. It also verifies self signed certificates 3092 by the same criteria. The main comparison function is 3093 X509_check_issued() which performs these checks. 3094 3095 Lot of changes were necessary in order to support this 3096 without completely rewriting the lookup code. 3097 3098 Authority and subject key identifier are now cached. 3099 3100 The LHASH 'certs' is X509_STORE has now been replaced 3101 by a STACK_OF(X509_OBJECT). This is mainly because an 3102 LHASH can't store or retrieve multiple objects with 3103 the same hash value. 3104 3105 As a result various functions (which were all internal 3106 use only) have changed to handle the new X509_STORE 3107 structure. This will break anything that messed round 3108 with X509_STORE internally. 3109 3110 The functions X509_STORE_add_cert() now checks for an 3111 exact match, rather than just subject name. 3112 3113 The X509_STORE API doesn't directly support the retrieval 3114 of multiple certificates matching a given criteria, however 3115 this can be worked round by performing a lookup first 3116 (which will fill the cache with candidate certificates) 3117 and then examining the cache for matches. This is probably 3118 the best we can do without throwing out X509_LOOKUP 3119 entirely (maybe later...). 3120 3121 The X509_VERIFY_CTX structure has been enhanced considerably. 3122 3123 All certificate lookup operations now go via a get_issuer() 3124 callback. Although this currently uses an X509_STORE it 3125 can be replaced by custom lookups. This is a simple way 3126 to bypass the X509_STORE hackery necessary to make this 3127 work and makes it possible to use more efficient techniques 3128 in future. A very simple version which uses a simple 3129 STACK for its trusted certificate store is also provided 3130 using X509_STORE_CTX_trusted_stack(). 3131 3132 The verify_cb() and verify() callbacks now have equivalents 3133 in the X509_STORE_CTX structure. 3134 3135 X509_STORE_CTX also has a 'flags' field which can be used 3136 to customise the verify behaviour. 3137 [Steve Henson] 3138 3139 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 3140 excludes S/MIME capabilities. 3141 [Steve Henson] 3142 3143 *) When a certificate request is read in keep a copy of the 3144 original encoding of the signed data and use it when outputing 3145 again. Signatures then use the original encoding rather than 3146 a decoded, encoded version which may cause problems if the 3147 request is improperly encoded. 3148 [Steve Henson] 3149 3150 *) For consistency with other BIO_puts implementations, call 3151 buffer_write(b, ...) directly in buffer_puts instead of calling 3152 BIO_write(b, ...). 3153 3154 In BIO_puts, increment b->num_write as in BIO_write. 3155 [Peter.Sylvester@EdelWeb.fr] 3156 3157 *) Fix BN_mul_word for the case where the word is 0. (We have to use 3158 BN_zero, we may not return a BIGNUM with an array consisting of 3159 words set to zero.) 3160 [Bodo Moeller] 3161 3162 *) Avoid calling abort() from within the library when problems are 3163 detected, except if preprocessor symbols have been defined 3164 (such as REF_CHECK, BN_DEBUG etc.). 3165 [Bodo Moeller] 3166 3167 *) New openssl application 'rsautl'. This utility can be 3168 used for low level RSA operations. DER public key 3169 BIO/fp routines also added. 3170 [Steve Henson] 3171 3172 *) New Configure entry and patches for compiling on QNX 4. 3173 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 3174 3175 *) A demo state-machine implementation was sponsored by 3176 Nuron (http://www.nuron.com/) and is now available in 3177 demos/state_machine. 3178 [Ben Laurie] 3179 3180 *) New options added to the 'dgst' utility for signature 3181 generation and verification. 3182 [Steve Henson] 3183 3184 *) Unrecognized PKCS#7 content types are now handled via a 3185 catch all ASN1_TYPE structure. This allows unsupported 3186 types to be stored as a "blob" and an application can 3187 encode and decode it manually. 3188 [Steve Henson] 3189 3190 *) Fix various signed/unsigned issues to make a_strex.c 3191 compile under VC++. 3192 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 3193 3194 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 3195 length if passed a buffer. ASN1_INTEGER_to_BN failed 3196 if passed a NULL BN and its argument was negative. 3197 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 3198 3199 *) Modification to PKCS#7 encoding routines to output definite 3200 length encoding. Since currently the whole structures are in 3201 memory there's not real point in using indefinite length 3202 constructed encoding. However if OpenSSL is compiled with 3203 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 3204 [Steve Henson] 3205 3206 *) Added BIO_vprintf() and BIO_vsnprintf(). 3207 [Richard Levitte] 3208 3209 *) Added more prefixes to parse for in the the strings written 3210 through a logging bio, to cover all the levels that are available 3211 through syslog. The prefixes are now: 3212 3213 PANIC, EMERG, EMR => LOG_EMERG 3214 ALERT, ALR => LOG_ALERT 3215 CRIT, CRI => LOG_CRIT 3216 ERROR, ERR => LOG_ERR 3217 WARNING, WARN, WAR => LOG_WARNING 3218 NOTICE, NOTE, NOT => LOG_NOTICE 3219 INFO, INF => LOG_INFO 3220 DEBUG, DBG => LOG_DEBUG 3221 3222 and as before, if none of those prefixes are present at the 3223 beginning of the string, LOG_ERR is chosen. 3224 3225 On Win32, the LOG_* levels are mapped according to this: 3226 3227 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 3228 LOG_WARNING => EVENTLOG_WARNING_TYPE 3229 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 3230 3231 [Richard Levitte] 3232 3233 *) Made it possible to reconfigure with just the configuration 3234 argument "reconf" or "reconfigure". The command line arguments 3235 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 3236 and are retrieved from there when reconfiguring. 3237 [Richard Levitte] 3238 3239 *) MD4 implemented. 3240 [Assar Westerlund <assar@sics.se>, Richard Levitte] 3241 3242 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 3243 [Richard Levitte] 3244 3245 *) The obj_dat.pl script was messing up the sorting of object 3246 names. The reason was that it compared the quoted version 3247 of strings as a result "OCSP" > "OCSP Signing" because 3248 " > SPACE. Changed script to store unquoted versions of 3249 names and add quotes on output. It was also omitting some 3250 names from the lookup table if they were given a default 3251 value (that is if SN is missing it is given the same 3252 value as LN and vice versa), these are now added on the 3253 grounds that if an object has a name we should be able to 3254 look it up. Finally added warning output when duplicate 3255 short or long names are found. 3256 [Steve Henson] 3257 3258 *) Changes needed for Tandem NSK. 3259 [Scott Uroff <scott@xypro.com>] 3260 3261 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 3262 RSA_padding_check_SSLv23(), special padding was never detected 3263 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 3264 version rollback attacks was not effective. 3265 3266 In s23_clnt.c, don't use special rollback-attack detection padding 3267 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 3268 client; similarly, in s23_srvr.c, don't do the rollback check if 3269 SSL 2.0 is the only protocol enabled in the server. 3270 [Bodo Moeller] 3271 3272 *) Make it possible to get hexdumps of unprintable data with 'openssl 3273 asn1parse'. By implication, the functions ASN1_parse_dump() and 3274 BIO_dump_indent() are added. 3275 [Richard Levitte] 3276 3277 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 3278 these print out strings and name structures based on various 3279 flags including RFC2253 support and proper handling of 3280 multibyte characters. Added options to the 'x509' utility 3281 to allow the various flags to be set. 3282 [Steve Henson] 3283 3284 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 3285 Also change the functions X509_cmp_current_time() and 3286 X509_gmtime_adj() work with an ASN1_TIME structure, 3287 this will enable certificates using GeneralizedTime in validity 3288 dates to be checked. 3289 [Steve Henson] 3290 3291 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 3292 negative public key encodings) on by default, 3293 NO_NEG_PUBKEY_BUG can be set to disable it. 3294 [Steve Henson] 3295 3296 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 3297 content octets. An i2c_ASN1_OBJECT is unnecessary because 3298 the encoding can be trivially obtained from the structure. 3299 [Steve Henson] 3300 3301 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 3302 not read locks (CRYPTO_r_[un]lock). 3303 [Bodo Moeller] 3304 3305 *) A first attempt at creating official support for shared 3306 libraries through configuration. I've kept it so the 3307 default is static libraries only, and the OpenSSL programs 3308 are always statically linked for now, but there are 3309 preparations for dynamic linking in place. 3310 This has been tested on Linux and Tru64. 3311 [Richard Levitte] 3312 3313 *) Randomness polling function for Win9x, as described in: 3314 Peter Gutmann, Software Generation of Practically Strong 3315 Random Numbers. 3316 [Ulf M�ller] 3317 3318 *) Fix so PRNG is seeded in req if using an already existing 3319 DSA key. 3320 [Steve Henson] 3321 3322 *) New options to smime application. -inform and -outform 3323 allow alternative formats for the S/MIME message including 3324 PEM and DER. The -content option allows the content to be 3325 specified separately. This should allow things like Netscape 3326 form signing output easier to verify. 3327 [Steve Henson] 3328 3329 *) Fix the ASN1 encoding of tags using the 'long form'. 3330 [Steve Henson] 3331 3332 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 3333 STRING types. These convert content octets to and from the 3334 underlying type. The actual tag and length octets are 3335 already assumed to have been read in and checked. These 3336 are needed because all other string types have virtually 3337 identical handling apart from the tag. By having versions 3338 of the ASN1 functions that just operate on content octets 3339 IMPLICIT tagging can be handled properly. It also allows 3340 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 3341 and ASN1_INTEGER are identical apart from the tag. 3342 [Steve Henson] 3343 3344 *) Change the handling of OID objects as follows: 3345 3346 - New object identifiers are inserted in objects.txt, following 3347 the syntax given in objects.README. 3348 - objects.pl is used to process obj_mac.num and create a new 3349 obj_mac.h. 3350 - obj_dat.pl is used to create a new obj_dat.h, using the data in 3351 obj_mac.h. 3352 3353 This is currently kind of a hack, and the perl code in objects.pl 3354 isn't very elegant, but it works as I intended. The simplest way 3355 to check that it worked correctly is to look in obj_dat.h and 3356 check the array nid_objs and make sure the objects haven't moved 3357 around (this is important!). Additions are OK, as well as 3358 consistent name changes. 3359 [Richard Levitte] 3360 3361 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 3362 [Bodo Moeller] 3363 3364 *) Addition of the command line parameter '-rand file' to 'openssl req'. 3365 The given file adds to whatever has already been seeded into the 3366 random pool through the RANDFILE configuration file option or 3367 environment variable, or the default random state file. 3368 [Richard Levitte] 3369 3370 *) mkstack.pl now sorts each macro group into lexical order. 3371 Previously the output order depended on the order the files 3372 appeared in the directory, resulting in needless rewriting 3373 of safestack.h . 3374 [Steve Henson] 3375 3376 *) Patches to make OpenSSL compile under Win32 again. Mostly 3377 work arounds for the VC++ problem that it treats func() as 3378 func(void). Also stripped out the parts of mkdef.pl that 3379 added extra typesafe functions: these no longer exist. 3380 [Steve Henson] 3381 3382 *) Reorganisation of the stack code. The macros are now all 3383 collected in safestack.h . Each macro is defined in terms of 3384 a "stack macro" of the form SKM_<name>(type, a, b). The 3385 DEBUG_SAFESTACK is now handled in terms of function casts, 3386 this has the advantage of retaining type safety without the 3387 use of additional functions. If DEBUG_SAFESTACK is not defined 3388 then the non typesafe macros are used instead. Also modified the 3389 mkstack.pl script to handle the new form. Needs testing to see 3390 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 3391 the default if no major problems. Similar behaviour for ASN1_SET_OF 3392 and PKCS12_STACK_OF. 3393 [Steve Henson] 3394 3395 *) When some versions of IIS use the 'NET' form of private key the 3396 key derivation algorithm is different. Normally MD5(password) is 3397 used as a 128 bit RC4 key. In the modified case 3398 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 3399 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 3400 as the old Netscape_RSA functions except they have an additional 3401 'sgckey' parameter which uses the modified algorithm. Also added 3402 an -sgckey command line option to the rsa utility. Thanks to 3403 Adrian Peck <bertie@ncipher.com> for posting details of the modified 3404 algorithm to openssl-dev. 3405 [Steve Henson] 3406 3407 *) The evp_local.h macros were using 'c.##kname' which resulted in 3408 invalid expansion on some systems (SCO 5.0.5 for example). 3409 Corrected to 'c.kname'. 3410 [Phillip Porch <root@theporch.com>] 3411 3412 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 3413 a STACK of email addresses from a certificate or request, these look 3414 in the subject name and the subject alternative name extensions and 3415 omit any duplicate addresses. 3416 [Steve Henson] 3417 3418 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 3419 This makes DSA verification about 2 % faster. 3420 [Bodo Moeller] 3421 3422 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 3423 (meaning that now 2^5 values will be precomputed, which is only 4 KB 3424 plus overhead for 1024 bit moduli). 3425 This makes exponentiations about 0.5 % faster for 1024 bit 3426 exponents (as measured by "openssl speed rsa2048"). 3427 [Bodo Moeller] 3428 3429 *) Rename memory handling macros to avoid conflicts with other 3430 software: 3431 Malloc => OPENSSL_malloc 3432 Malloc_locked => OPENSSL_malloc_locked 3433 Realloc => OPENSSL_realloc 3434 Free => OPENSSL_free 3435 [Richard Levitte] 3436 3437 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 3438 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 3439 [Bodo Moeller] 3440 3441 *) CygWin32 support. 3442 [John Jarvie <jjarvie@newsguy.com>] 3443 3444 *) The type-safe stack code has been rejigged. It is now only compiled 3445 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 3446 by default all type-specific stack functions are "#define"d back to 3447 standard stack functions. This results in more streamlined output 3448 but retains the type-safety checking possibilities of the original 3449 approach. 3450 [Geoff Thorpe] 3451 3452 *) The STACK code has been cleaned up, and certain type declarations 3453 that didn't make a lot of sense have been brought in line. This has 3454 also involved a cleanup of sorts in safestack.h to more correctly 3455 map type-safe stack functions onto their plain stack counterparts. 3456 This work has also resulted in a variety of "const"ifications of 3457 lots of the code, especially "_cmp" operations which should normally 3458 be prototyped with "const" parameters anyway. 3459 [Geoff Thorpe] 3460 3461 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 3462 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 3463 (The PRNG state consists of two parts, the large pool 'state' and 'md', 3464 where all of 'md' is used each time the PRNG is used, but 'state' 3465 is used only indexed by a cyclic counter. As entropy may not be 3466 well distributed from the beginning, 'md' is important as a 3467 chaining variable. However, the output function chains only half 3468 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 3469 all of 'md', and seeding with STATE_SIZE dummy bytes will result 3470 in all of 'state' being rewritten, with the new values depending 3471 on virtually all of 'md'. This overcomes the 80 bit limitation.) 3472 [Bodo Moeller] 3473 3474 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 3475 the handshake is continued after ssl_verify_cert_chain(); 3476 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 3477 can lead to 'unexplainable' connection aborts later. 3478 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 3479 3480 *) Major EVP API cipher revision. 3481 Add hooks for extra EVP features. This allows various cipher 3482 parameters to be set in the EVP interface. Support added for variable 3483 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 3484 setting of RC2 and RC5 parameters. 3485 3486 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 3487 ciphers. 3488 3489 Remove lots of duplicated code from the EVP library. For example *every* 3490 cipher init() function handles the 'iv' in the same way according to the 3491 cipher mode. They also all do nothing if the 'key' parameter is NULL and 3492 for CFB and OFB modes they zero ctx->num. 3493 3494 New functionality allows removal of S/MIME code RC2 hack. 3495 3496 Most of the routines have the same form and so can be declared in terms 3497 of macros. 3498 3499 By shifting this to the top level EVP_CipherInit() it can be removed from 3500 all individual ciphers. If the cipher wants to handle IVs or keys 3501 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 3502 flags. 3503 3504 Change lots of functions like EVP_EncryptUpdate() to now return a 3505 value: although software versions of the algorithms cannot fail 3506 any installed hardware versions can. 3507 [Steve Henson] 3508 3509 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 3510 this option is set, tolerate broken clients that send the negotiated 3511 protocol version number instead of the requested protocol version 3512 number. 3513 [Bodo Moeller] 3514 3515 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 3516 i.e. non-zero for export ciphersuites, zero otherwise. 3517 Previous versions had this flag inverted, inconsistent with 3518 rsa_tmp_cb (..._TMP_RSA_CB). 3519 [Bodo Moeller; problem reported by Amit Chopra] 3520 3521 *) Add missing DSA library text string. Work around for some IIS 3522 key files with invalid SEQUENCE encoding. 3523 [Steve Henson] 3524 3525 *) Add a document (doc/standards.txt) that list all kinds of standards 3526 and so on that are implemented in OpenSSL. 3527 [Richard Levitte] 3528 3529 *) Enhance c_rehash script. Old version would mishandle certificates 3530 with the same subject name hash and wouldn't handle CRLs at all. 3531 Added -fingerprint option to crl utility, to support new c_rehash 3532 features. 3533 [Steve Henson] 3534 3535 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 3536 [Ulf M�ller] 3537 3538 *) Fix for SSL server purpose checking. Server checking was 3539 rejecting certificates which had extended key usage present 3540 but no ssl client purpose. 3541 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 3542 3543 *) Make PKCS#12 code work with no password. The PKCS#12 spec 3544 is a little unclear about how a blank password is handled. 3545 Since the password in encoded as a BMPString with terminating 3546 double NULL a zero length password would end up as just the 3547 double NULL. However no password at all is different and is 3548 handled differently in the PKCS#12 key generation code. NS 3549 treats a blank password as zero length. MSIE treats it as no 3550 password on export: but it will try both on import. We now do 3551 the same: PKCS12_parse() tries zero length and no password if 3552 the password is set to "" or NULL (NULL is now a valid password: 3553 it wasn't before) as does the pkcs12 application. 3554 [Steve Henson] 3555 3556 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 3557 perror when PEM_read_bio_X509_REQ fails, the error message must 3558 be obtained from the error queue. 3559 [Bodo Moeller] 3560 3561 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 3562 it in ERR_remove_state if appropriate, and change ERR_get_state 3563 accordingly to avoid race conditions (this is necessary because 3564 thread_hash is no longer constant once set). 3565 [Bodo Moeller] 3566 3567 *) Bugfix for linux-elf makefile.one. 3568 [Ulf M�ller] 3569 3570 *) RSA_get_default_method() will now cause a default 3571 RSA_METHOD to be chosen if one doesn't exist already. 3572 Previously this was only set during a call to RSA_new() 3573 or RSA_new_method(NULL) meaning it was possible for 3574 RSA_get_default_method() to return NULL. 3575 [Geoff Thorpe] 3576 3577 *) Added native name translation to the existing DSO code 3578 that will convert (if the flag to do so is set) filenames 3579 that are sufficiently small and have no path information 3580 into a canonical native form. Eg. "blah" converted to 3581 "libblah.so" or "blah.dll" etc. 3582 [Geoff Thorpe] 3583 3584 *) New function ERR_error_string_n(e, buf, len) which is like 3585 ERR_error_string(e, buf), but writes at most 'len' bytes 3586 including the 0 terminator. For ERR_error_string_n, 'buf' 3587 may not be NULL. 3588 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 3589 3590 *) CONF library reworked to become more general. A new CONF 3591 configuration file reader "class" is implemented as well as a 3592 new functions (NCONF_*, for "New CONF") to handle it. The now 3593 old CONF_* functions are still there, but are reimplemented to 3594 work in terms of the new functions. Also, a set of functions 3595 to handle the internal storage of the configuration data is 3596 provided to make it easier to write new configuration file 3597 reader "classes" (I can definitely see something reading a 3598 configuration file in XML format, for example), called _CONF_*, 3599 or "the configuration storage API"... 3600 3601 The new configuration file reading functions are: 3602 3603 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 3604 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 3605 3606 NCONF_default, NCONF_WIN32 3607 3608 NCONF_dump_fp, NCONF_dump_bio 3609 3610 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 3611 NCONF_new creates a new CONF object. This works in the same way 3612 as other interfaces in OpenSSL, like the BIO interface. 3613 NCONF_dump_* dump the internal storage of the configuration file, 3614 which is useful for debugging. All other functions take the same 3615 arguments as the old CONF_* functions wth the exception of the 3616 first that must be a `CONF *' instead of a `LHASH *'. 3617 3618 To make it easer to use the new classes with the old CONF_* functions, 3619 the function CONF_set_default_method is provided. 3620 [Richard Levitte] 3621 3622 *) Add '-tls1' option to 'openssl ciphers', which was already 3623 mentioned in the documentation but had not been implemented. 3624 (This option is not yet really useful because even the additional 3625 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 3626 [Bodo Moeller] 3627 3628 *) Initial DSO code added into libcrypto for letting OpenSSL (and 3629 OpenSSL-based applications) load shared libraries and bind to 3630 them in a portable way. 3631 [Geoff Thorpe, with contributions from Richard Levitte] 3632 3633 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 3634 3635 *) Make sure _lrotl and _lrotr are only used with MSVC. 3636 3637 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 3638 (the default implementation of RAND_status). 3639 3640 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 3641 to '-clrext' (= clear extensions), as intended and documented. 3642 [Bodo Moeller; inconsistency pointed out by Michael Attili 3643 <attili@amaxo.com>] 3644 3645 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 3646 was larger than the MD block size. 3647 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 3648 3649 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 3650 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 3651 using the passed key: if the passed key was a private key the result 3652 of X509_print(), for example, would be to print out all the private key 3653 components. 3654 [Steve Henson] 3655 3656 *) des_quad_cksum() byte order bug fix. 3657 [Ulf M�ller, using the problem description in krb4-0.9.7, where 3658 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 3659 3660 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 3661 discouraged. 3662 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 3663 3664 *) For easily testing in shell scripts whether some command 3665 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 3666 returns with exit code 0 iff no command of the given name is available. 3667 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 3668 the output goes to stdout and nothing is printed to stderr. 3669 Additional arguments are always ignored. 3670 3671 Since for each cipher there is a command of the same name, 3672 the 'no-cipher' compilation switches can be tested this way. 3673 3674 ('openssl no-XXX' is not able to detect pseudo-commands such 3675 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 3676 [Bodo Moeller] 3677 3678 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 3679 [Bodo Moeller] 3680 3681 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 3682 is set; it will be thrown away anyway because each handshake creates 3683 its own key. 3684 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 3685 to parameters -- in previous versions (since OpenSSL 0.9.3) the 3686 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 3687 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 3688 [Bodo Moeller] 3689 3690 *) New s_client option -ign_eof: EOF at stdin is ignored, and 3691 'Q' and 'R' lose their special meanings (quit/renegotiate). 3692 This is part of what -quiet does; unlike -quiet, -ign_eof 3693 does not suppress any output. 3694 [Richard Levitte] 3695 3696 *) Add compatibility options to the purpose and trust code. The 3697 purpose X509_PURPOSE_ANY is "any purpose" which automatically 3698 accepts a certificate or CA, this was the previous behaviour, 3699 with all the associated security issues. 3700 3701 X509_TRUST_COMPAT is the old trust behaviour: only and 3702 automatically trust self signed roots in certificate store. A 3703 new trust setting X509_TRUST_DEFAULT is used to specify that 3704 a purpose has no associated trust setting and it should instead 3705 use the value in the default purpose. 3706 [Steve Henson] 3707 3708 *) Fix the PKCS#8 DSA private key code so it decodes keys again 3709 and fix a memory leak. 3710 [Steve Henson] 3711 3712 *) In util/mkerr.pl (which implements 'make errors'), preserve 3713 reason strings from the previous version of the .c file, as 3714 the default to have only downcase letters (and digits) in 3715 automatically generated reasons codes is not always appropriate. 3716 [Bodo Moeller] 3717 3718 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 3719 using strerror. Previously, ERR_reason_error_string() returned 3720 library names as reason strings for SYSerr; but SYSerr is a special 3721 case where small numbers are errno values, not library numbers. 3722 [Bodo Moeller] 3723 3724 *) Add '-dsaparam' option to 'openssl dhparam' application. This 3725 converts DSA parameters into DH parameters. (When creating parameters, 3726 DSA_generate_parameters is used.) 3727 [Bodo Moeller] 3728 3729 *) Include 'length' (recommended exponent length) in C code generated 3730 by 'openssl dhparam -C'. 3731 [Bodo Moeller] 3732 3733 *) The second argument to set_label in perlasm was already being used 3734 so couldn't be used as a "file scope" flag. Moved to third argument 3735 which was free. 3736 [Steve Henson] 3737 3738 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 3739 instead of RAND_bytes for encryption IVs and salts. 3740 [Bodo Moeller] 3741 3742 *) Include RAND_status() into RAND_METHOD instead of implementing 3743 it only for md_rand.c Otherwise replacing the PRNG by calling 3744 RAND_set_rand_method would be impossible. 3745 [Bodo Moeller] 3746 3747 *) Don't let DSA_generate_key() enter an infinite loop if the random 3748 number generation fails. 3749 [Bodo Moeller] 3750 3751 *) New 'rand' application for creating pseudo-random output. 3752 [Bodo Moeller] 3753 3754 *) Added configuration support for Linux/IA64 3755 [Rolf Haberrecker <rolf@suse.de>] 3756 3757 *) Assembler module support for Mingw32. 3758 [Ulf M�ller] 3759 3760 *) Shared library support for HPUX (in shlib/). 3761 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 3762 3763 *) Shared library support for Solaris gcc. 3764 [Lutz Behnke <behnke@trustcenter.de>] 3765 3766 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 3767 3768 *) PKCS7_encrypt() was adding text MIME headers twice because they 3769 were added manually and by SMIME_crlf_copy(). 3770 [Steve Henson] 3771 3772 *) In bntest.c don't call BN_rand with zero bits argument. 3773 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 3774 3775 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 3776 case was implemented. This caused BN_div_recp() to fail occasionally. 3777 [Ulf M�ller] 3778 3779 *) Add an optional second argument to the set_label() in the perl 3780 assembly language builder. If this argument exists and is set 3781 to 1 it signals that the assembler should use a symbol whose 3782 scope is the entire file, not just the current function. This 3783 is needed with MASM which uses the format label:: for this scope. 3784 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 3785 3786 *) Change the ASN1 types so they are typedefs by default. Before 3787 almost all types were #define'd to ASN1_STRING which was causing 3788 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 3789 for example. 3790 [Steve Henson] 3791 3792 *) Change names of new functions to the new get1/get0 naming 3793 convention: After 'get1', the caller owns a reference count 3794 and has to call ..._free; 'get0' returns a pointer to some 3795 data structure without incrementing reference counters. 3796 (Some of the existing 'get' functions increment a reference 3797 counter, some don't.) 3798 Similarly, 'set1' and 'add1' functions increase reference 3799 counters or duplicate objects. 3800 [Steve Henson] 3801 3802 *) Allow for the possibility of temp RSA key generation failure: 3803 the code used to assume it always worked and crashed on failure. 3804 [Steve Henson] 3805 3806 *) Fix potential buffer overrun problem in BIO_printf(). 3807 [Ulf M�ller, using public domain code by Patrick Powell; problem 3808 pointed out by David Sacerdote <das33@cornell.edu>] 3809 3810 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 3811 RAND_egd() and RAND_status(). In the command line application, 3812 the EGD socket can be specified like a seed file using RANDFILE 3813 or -rand. 3814 [Ulf M�ller] 3815 3816 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 3817 Some CAs (e.g. Verisign) distribute certificates in this form. 3818 [Steve Henson] 3819 3820 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 3821 list to exclude them. This means that no special compilation option 3822 is needed to use anonymous DH: it just needs to be included in the 3823 cipher list. 3824 [Steve Henson] 3825 3826 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 3827 EVP_MD_type. The old functionality is available in a new macro called 3828 EVP_MD_md(). Change code that uses it and update docs. 3829 [Steve Henson] 3830 3831 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 3832 where the 'void *' argument is replaced by a function pointer argument. 3833 Previously 'void *' was abused to point to functions, which works on 3834 many platforms, but is not correct. As these functions are usually 3835 called by macros defined in OpenSSL header files, most source code 3836 should work without changes. 3837 [Richard Levitte] 3838 3839 *) <openssl/opensslconf.h> (which is created by Configure) now contains 3840 sections with information on -D... compiler switches used for 3841 compiling the library so that applications can see them. To enable 3842 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 3843 must be defined. E.g., 3844 #define OPENSSL_ALGORITHM_DEFINES 3845 #include <openssl/opensslconf.h> 3846 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 3847 [Richard Levitte, Ulf and Bodo M�ller] 3848 3849 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 3850 record layer. 3851 [Bodo Moeller] 3852 3853 *) Change the 'other' type in certificate aux info to a STACK_OF 3854 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 3855 the required ASN1 format: arbitrary types determined by an OID. 3856 [Steve Henson] 3857 3858 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 3859 argument to 'req'. This is not because the function is newer or 3860 better than others it just uses the work 'NEW' in the certificate 3861 request header lines. Some software needs this. 3862 [Steve Henson] 3863 3864 *) Reorganise password command line arguments: now passwords can be 3865 obtained from various sources. Delete the PEM_cb function and make 3866 it the default behaviour: i.e. if the callback is NULL and the 3867 usrdata argument is not NULL interpret it as a null terminated pass 3868 phrase. If usrdata and the callback are NULL then the pass phrase 3869 is prompted for as usual. 3870 [Steve Henson] 3871 3872 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 3873 the support is automatically enabled. The resulting binaries will 3874 autodetect the card and use it if present. 3875 [Ben Laurie and Compaq Inc.] 3876 3877 *) Work around for Netscape hang bug. This sends certificate request 3878 and server done in one record. Since this is perfectly legal in the 3879 SSL/TLS protocol it isn't a "bug" option and is on by default. See 3880 the bugs/SSLv3 entry for more info. 3881 [Steve Henson] 3882 3883 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 3884 [Andy Polyakov] 3885 3886 *) Add -rand argument to smime and pkcs12 applications and read/write 3887 of seed file. 3888 [Steve Henson] 3889 3890 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 3891 [Bodo Moeller] 3892 3893 *) Add command line password options to the remaining applications. 3894 [Steve Henson] 3895 3896 *) Bug fix for BN_div_recp() for numerators with an even number of 3897 bits. 3898 [Ulf M�ller] 3899 3900 *) More tests in bntest.c, and changed test_bn output. 3901 [Ulf M�ller] 3902 3903 *) ./config recognizes MacOS X now. 3904 [Andy Polyakov] 3905 3906 *) Bug fix for BN_div() when the first words of num and divsor are 3907 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 3908 [Ulf M�ller] 3909 3910 *) Add support for various broken PKCS#8 formats, and command line 3911 options to produce them. 3912 [Steve Henson] 3913 3914 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 3915 get temporary BIGNUMs from a BN_CTX. 3916 [Ulf M�ller] 3917 3918 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 3919 for p == 0. 3920 [Ulf M�ller] 3921 3922 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 3923 include a #define from the old name to the new. The original intent 3924 was that statically linked binaries could for example just call 3925 SSLeay_add_all_ciphers() to just add ciphers to the table and not 3926 link with digests. This never worked becayse SSLeay_add_all_digests() 3927 and SSLeay_add_all_ciphers() were in the same source file so calling 3928 one would link with the other. They are now in separate source files. 3929 [Steve Henson] 3930 3931 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 3932 [Steve Henson] 3933 3934 *) Use a less unusual form of the Miller-Rabin primality test (it used 3935 a binary algorithm for exponentiation integrated into the Miller-Rabin 3936 loop, our standard modexp algorithms are faster). 3937 [Bodo Moeller] 3938 3939 *) Support for the EBCDIC character set completed. 3940 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 3941 3942 *) Source code cleanups: use const where appropriate, eliminate casts, 3943 use void * instead of char * in lhash. 3944 [Ulf M�ller] 3945 3946 *) Bugfix: ssl3_send_server_key_exchange was not restartable 3947 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 3948 this the server could overwrite ephemeral keys that the client 3949 has already seen). 3950 [Bodo Moeller] 3951 3952 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 3953 using 50 iterations of the Rabin-Miller test. 3954 3955 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 3956 iterations of the Rabin-Miller test as required by the appendix 3957 to FIPS PUB 186[-1]) instead of DSA_is_prime. 3958 As BN_is_prime_fasttest includes trial division, DSA parameter 3959 generation becomes much faster. 3960 3961 This implies a change for the callback functions in DSA_is_prime 3962 and DSA_generate_parameters: The callback function is called once 3963 for each positive witness in the Rabin-Miller test, not just 3964 occasionally in the inner loop; and the parameters to the 3965 callback function now provide an iteration count for the outer 3966 loop rather than for the current invocation of the inner loop. 3967 DSA_generate_parameters additionally can call the callback 3968 function with an 'iteration count' of -1, meaning that a 3969 candidate has passed the trial division test (when q is generated 3970 from an application-provided seed, trial division is skipped). 3971 [Bodo Moeller] 3972 3973 *) New function BN_is_prime_fasttest that optionally does trial 3974 division before starting the Rabin-Miller test and has 3975 an additional BN_CTX * argument (whereas BN_is_prime always 3976 has to allocate at least one BN_CTX). 3977 'callback(1, -1, cb_arg)' is called when a number has passed the 3978 trial division stage. 3979 [Bodo Moeller] 3980 3981 *) Fix for bug in CRL encoding. The validity dates weren't being handled 3982 as ASN1_TIME. 3983 [Steve Henson] 3984 3985 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 3986 [Steve Henson] 3987 3988 *) New function BN_pseudo_rand(). 3989 [Ulf M�ller] 3990 3991 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 3992 bignum version of BN_from_montgomery() with the working code from 3993 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 3994 the comments. 3995 [Ulf M�ller] 3996 3997 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 3998 made it impossible to use the same SSL_SESSION data structure in 3999 SSL2 clients in multiple threads. 4000 [Bodo Moeller] 4001 4002 *) The return value of RAND_load_file() no longer counts bytes obtained 4003 by stat(). RAND_load_file(..., -1) is new and uses the complete file 4004 to seed the PRNG (previously an explicit byte count was required). 4005 [Ulf M�ller, Bodo M�ller] 4006 4007 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 4008 used (char *) instead of (void *) and had casts all over the place. 4009 [Steve Henson] 4010 4011 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 4012 [Ulf M�ller] 4013 4014 *) Retain source code compatibility for BN_prime_checks macro: 4015 BN_is_prime(..., BN_prime_checks, ...) now uses 4016 BN_prime_checks_for_size to determine the appropriate number of 4017 Rabin-Miller iterations. 4018 [Ulf M�ller] 4019 4020 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 4021 DH_CHECK_P_NOT_SAFE_PRIME. 4022 (Check if this is true? OpenPGP calls them "strong".) 4023 [Ulf M�ller] 4024 4025 *) Merge the functionality of "dh" and "gendh" programs into a new program 4026 "dhparam". The old programs are retained for now but will handle DH keys 4027 (instead of parameters) in future. 4028 [Steve Henson] 4029 4030 *) Make the ciphers, s_server and s_client programs check the return values 4031 when a new cipher list is set. 4032 [Steve Henson] 4033 4034 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 4035 ciphers. Before when the 56bit ciphers were enabled the sorting was 4036 wrong. 4037 4038 The syntax for the cipher sorting has been extended to support sorting by 4039 cipher-strength (using the strength_bits hard coded in the tables). 4040 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 4041 4042 Fix a bug in the cipher-command parser: when supplying a cipher command 4043 string with an "undefined" symbol (neither command nor alphanumeric 4044 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 4045 an error is flagged. 4046 4047 Due to the strength-sorting extension, the code of the 4048 ssl_create_cipher_list() function was completely rearranged. I hope that 4049 the readability was also increased :-) 4050 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 4051 4052 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 4053 for the first serial number and places 2 in the serial number file. This 4054 avoids problems when the root CA is created with serial number zero and 4055 the first user certificate has the same issuer name and serial number 4056 as the root CA. 4057 [Steve Henson] 4058 4059 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 4060 the new code. Add documentation for this stuff. 4061 [Steve Henson] 4062 4063 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 4064 X509_*() to X509at_*() on the grounds that they don't handle X509 4065 structures and behave in an analagous way to the X509v3 functions: 4066 they shouldn't be called directly but wrapper functions should be used 4067 instead. 4068 4069 So we also now have some wrapper functions that call the X509at functions 4070 when passed certificate requests. (TO DO: similar things can be done with 4071 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 4072 things. Some of these need some d2i or i2d and print functionality 4073 because they handle more complex structures.) 4074 [Steve Henson] 4075 4076 *) Add missing #ifndefs that caused missing symbols when building libssl 4077 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 4078 NO_RSA in ssl/s2*.c. 4079 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 4080 4081 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 4082 has a return value which indicates the quality of the random data 4083 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 4084 error queue. New function RAND_pseudo_bytes() generates output that is 4085 guaranteed to be unique but not unpredictable. RAND_add is like 4086 RAND_seed, but takes an extra argument for an entropy estimate 4087 (RAND_seed always assumes full entropy). 4088 [Ulf M�ller] 4089 4090 *) Do more iterations of Rabin-Miller probable prime test (specifically, 4091 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 4092 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 4093 in crypto/bn/bn_prime.c for the complete table). This guarantees a 4094 false-positive rate of at most 2^-80 for random input. 4095 [Bodo Moeller] 4096 4097 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 4098 [Bodo Moeller] 4099 4100 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 4101 in the 0.9.5 release), this returns the chain 4102 from an X509_CTX structure with a dup of the stack and all 4103 the X509 reference counts upped: so the stack will exist 4104 after X509_CTX_cleanup() has been called. Modify pkcs12.c 4105 to use this. 4106 4107 Also make SSL_SESSION_print() print out the verify return 4108 code. 4109 [Steve Henson] 4110 4111 *) Add manpage for the pkcs12 command. Also change the default 4112 behaviour so MAC iteration counts are used unless the new 4113 -nomaciter option is used. This improves file security and 4114 only older versions of MSIE (4.0 for example) need it. 4115 [Steve Henson] 4116 4117 *) Honor the no-xxx Configure options when creating .DEF files. 4118 [Ulf M�ller] 4119 4120 *) Add PKCS#10 attributes to field table: challengePassword, 4121 unstructuredName and unstructuredAddress. These are taken from 4122 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 4123 international characters are used. 4124 4125 More changes to X509_ATTRIBUTE code: allow the setting of types 4126 based on strings. Remove the 'loc' parameter when adding 4127 attributes because these will be a SET OF encoding which is sorted 4128 in ASN1 order. 4129 [Steve Henson] 4130 4131 *) Initial changes to the 'req' utility to allow request generation 4132 automation. This will allow an application to just generate a template 4133 file containing all the field values and have req construct the 4134 request. 4135 4136 Initial support for X509_ATTRIBUTE handling. Stacks of these are 4137 used all over the place including certificate requests and PKCS#7 4138 structures. They are currently handled manually where necessary with 4139 some primitive wrappers for PKCS#7. The new functions behave in a 4140 manner analogous to the X509 extension functions: they allow 4141 attributes to be looked up by NID and added. 4142 4143 Later something similar to the X509V3 code would be desirable to 4144 automatically handle the encoding, decoding and printing of the 4145 more complex types. The string types like challengePassword can 4146 be handled by the string table functions. 4147 4148 Also modified the multi byte string table handling. Now there is 4149 a 'global mask' which masks out certain types. The table itself 4150 can use the flag STABLE_NO_MASK to ignore the mask setting: this 4151 is useful when for example there is only one permissible type 4152 (as in countryName) and using the mask might result in no valid 4153 types at all. 4154 [Steve Henson] 4155 4156 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 4157 SSL_get_peer_finished to allow applications to obtain the latest 4158 Finished messages sent to the peer or expected from the peer, 4159 respectively. (SSL_get_peer_finished is usually the Finished message 4160 actually received from the peer, otherwise the protocol will be aborted.) 4161 4162 As the Finished message are message digests of the complete handshake 4163 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 4164 be used for external authentication procedures when the authentication 4165 provided by SSL/TLS is not desired or is not enough. 4166 [Bodo Moeller] 4167 4168 *) Enhanced support for Alpha Linux is added. Now ./config checks if 4169 the host supports BWX extension and if Compaq C is present on the 4170 $PATH. Just exploiting of the BWX extension results in 20-30% 4171 performance kick for some algorithms, e.g. DES and RC4 to mention 4172 a couple. Compaq C in turn generates ~20% faster code for MD5 and 4173 SHA1. 4174 [Andy Polyakov] 4175 4176 *) Add support for MS "fast SGC". This is arguably a violation of the 4177 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 4178 weak crypto and after checking the certificate is SGC a second one 4179 with strong crypto. MS SGC stops the first handshake after receiving 4180 the server certificate message and sends a second client hello. Since 4181 a server will typically do all the time consuming operations before 4182 expecting any further messages from the client (server key exchange 4183 is the most expensive) there is little difference between the two. 4184 4185 To get OpenSSL to support MS SGC we have to permit a second client 4186 hello message after we have sent server done. In addition we have to 4187 reset the MAC if we do get this second client hello. 4188 [Steve Henson] 4189 4190 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 4191 if a DER encoded private key is RSA or DSA traditional format. Changed 4192 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 4193 format DER encoded private key. Newer code should use PKCS#8 format which 4194 has the key type encoded in the ASN1 structure. Added DER private key 4195 support to pkcs8 application. 4196 [Steve Henson] 4197 4198 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 4199 ciphersuites has been selected (as required by the SSL 3/TLS 1 4200 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 4201 is set, we interpret this as a request to violate the specification 4202 (the worst that can happen is a handshake failure, and 'correct' 4203 behaviour would result in a handshake failure anyway). 4204 [Bodo Moeller] 4205 4206 *) In SSL_CTX_add_session, take into account that there might be multiple 4207 SSL_SESSION structures with the same session ID (e.g. when two threads 4208 concurrently obtain them from an external cache). 4209 The internal cache can handle only one SSL_SESSION with a given ID, 4210 so if there's a conflict, we now throw out the old one to achieve 4211 consistency. 4212 [Bodo Moeller] 4213 4214 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 4215 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 4216 some routines that use cipher OIDs: some ciphers do not have OIDs 4217 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 4218 example. 4219 [Steve Henson] 4220 4221 *) Simplify the trust setting structure and code. Now we just have 4222 two sequences of OIDs for trusted and rejected settings. These will 4223 typically have values the same as the extended key usage extension 4224 and any application specific purposes. 4225 4226 The trust checking code now has a default behaviour: it will just 4227 check for an object with the same NID as the passed id. Functions can 4228 be provided to override either the default behaviour or the behaviour 4229 for a given id. SSL client, server and email already have functions 4230 in place for compatibility: they check the NID and also return "trusted" 4231 if the certificate is self signed. 4232 [Steve Henson] 4233 4234 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 4235 traditional format into an EVP_PKEY structure. 4236 [Steve Henson] 4237 4238 *) Add a password callback function PEM_cb() which either prompts for 4239 a password if usr_data is NULL or otherwise assumes it is a null 4240 terminated password. Allow passwords to be passed on command line 4241 environment or config files in a few more utilities. 4242 [Steve Henson] 4243 4244 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 4245 keys. Add some short names for PKCS#8 PBE algorithms and allow them 4246 to be specified on the command line for the pkcs8 and pkcs12 utilities. 4247 Update documentation. 4248 [Steve Henson] 4249 4250 *) Support for ASN1 "NULL" type. This could be handled before by using 4251 ASN1_TYPE but there wasn't any function that would try to read a NULL 4252 and produce an error if it couldn't. For compatibility we also have 4253 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 4254 don't allocate anything because they don't need to. 4255 [Steve Henson] 4256 4257 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 4258 for details. 4259 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 4260 4261 *) Rebuild of the memory allocation routines used by OpenSSL code and 4262 possibly others as well. The purpose is to make an interface that 4263 provide hooks so anyone can build a separate set of allocation and 4264 deallocation routines to be used by OpenSSL, for example memory 4265 pool implementations, or something else, which was previously hard 4266 since Malloc(), Realloc() and Free() were defined as macros having 4267 the values malloc, realloc and free, respectively (except for Win32 4268 compilations). The same is provided for memory debugging code. 4269 OpenSSL already comes with functionality to find memory leaks, but 4270 this gives people a chance to debug other memory problems. 4271 4272 With these changes, a new set of functions and macros have appeared: 4273 4274 CRYPTO_set_mem_debug_functions() [F] 4275 CRYPTO_get_mem_debug_functions() [F] 4276 CRYPTO_dbg_set_options() [F] 4277 CRYPTO_dbg_get_options() [F] 4278 CRYPTO_malloc_debug_init() [M] 4279 4280 The memory debug functions are NULL by default, unless the library 4281 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 4282 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 4283 gives the standard debugging functions that come with OpenSSL) or 4284 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 4285 provided by the library user) must be used. When the standard 4286 debugging functions are used, CRYPTO_dbg_set_options can be used to 4287 request additional information: 4288 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 4289 the CRYPTO_MDEBUG_xxx macro when compiling the library. 4290 4291 Also, things like CRYPTO_set_mem_functions will always give the 4292 expected result (the new set of functions is used for allocation 4293 and deallocation) at all times, regardless of platform and compiler 4294 options. 4295 4296 To finish it up, some functions that were never use in any other 4297 way than through macros have a new API and new semantic: 4298 4299 CRYPTO_dbg_malloc() 4300 CRYPTO_dbg_realloc() 4301 CRYPTO_dbg_free() 4302 4303 All macros of value have retained their old syntax. 4304 [Richard Levitte and Bodo Moeller] 4305 4306 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 4307 ordering of SMIMECapabilities wasn't in "strength order" and there 4308 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 4309 algorithm. 4310 [Steve Henson] 4311 4312 *) Some ASN1 types with illegal zero length encoding (INTEGER, 4313 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 4314 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 4315 4316 *) Merge in my S/MIME library for OpenSSL. This provides a simple 4317 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 4318 functionality to handle multipart/signed properly) and a utility 4319 called 'smime' to call all this stuff. This is based on code I 4320 originally wrote for Celo who have kindly allowed it to be 4321 included in OpenSSL. 4322 [Steve Henson] 4323 4324 *) Add variants des_set_key_checked and des_set_key_unchecked of 4325 des_set_key (aka des_key_sched). Global variable des_check_key 4326 decides which of these is called by des_set_key; this way 4327 des_check_key behaves as it always did, but applications and 4328 the library itself, which was buggy for des_check_key == 1, 4329 have a cleaner way to pick the version they need. 4330 [Bodo Moeller] 4331 4332 *) New function PKCS12_newpass() which changes the password of a 4333 PKCS12 structure. 4334 [Steve Henson] 4335 4336 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 4337 dynamic mix. In both cases the ids can be used as an index into the 4338 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 4339 functions so they accept a list of the field values and the 4340 application doesn't need to directly manipulate the X509_TRUST 4341 structure. 4342 [Steve Henson] 4343 4344 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 4345 need initialising. 4346 [Steve Henson] 4347 4348 *) Modify the way the V3 extension code looks up extensions. This now 4349 works in a similar way to the object code: we have some "standard" 4350 extensions in a static table which is searched with OBJ_bsearch() 4351 and the application can add dynamic ones if needed. The file 4352 crypto/x509v3/ext_dat.h now has the info: this file needs to be 4353 updated whenever a new extension is added to the core code and kept 4354 in ext_nid order. There is a simple program 'tabtest.c' which checks 4355 this. New extensions are not added too often so this file can readily 4356 be maintained manually. 4357 4358 There are two big advantages in doing things this way. The extensions 4359 can be looked up immediately and no longer need to be "added" using 4360 X509V3_add_standard_extensions(): this function now does nothing. 4361 [Side note: I get *lots* of email saying the extension code doesn't 4362 work because people forget to call this function] 4363 Also no dynamic allocation is done unless new extensions are added: 4364 so if we don't add custom extensions there is no need to call 4365 X509V3_EXT_cleanup(). 4366 [Steve Henson] 4367 4368 *) Modify enc utility's salting as follows: make salting the default. Add a 4369 magic header, so unsalted files fail gracefully instead of just decrypting 4370 to garbage. This is because not salting is a big security hole, so people 4371 should be discouraged from doing it. 4372 [Ben Laurie] 4373 4374 *) Fixes and enhancements to the 'x509' utility. It allowed a message 4375 digest to be passed on the command line but it only used this 4376 parameter when signing a certificate. Modified so all relevant 4377 operations are affected by the digest parameter including the 4378 -fingerprint and -x509toreq options. Also -x509toreq choked if a 4379 DSA key was used because it didn't fix the digest. 4380 [Steve Henson] 4381 4382 *) Initial certificate chain verify code. Currently tests the untrusted 4383 certificates for consistency with the verify purpose (which is set 4384 when the X509_STORE_CTX structure is set up) and checks the pathlength. 4385 4386 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 4387 this is because it will reject chains with invalid extensions whereas 4388 every previous version of OpenSSL and SSLeay made no checks at all. 4389 4390 Trust code: checks the root CA for the relevant trust settings. Trust 4391 settings have an initial value consistent with the verify purpose: e.g. 4392 if the verify purpose is for SSL client use it expects the CA to be 4393 trusted for SSL client use. However the default value can be changed to 4394 permit custom trust settings: one example of this would be to only trust 4395 certificates from a specific "secure" set of CAs. 4396 4397 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 4398 which should be used for version portability: especially since the 4399 verify structure is likely to change more often now. 4400 4401 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 4402 to set them. If not set then assume SSL clients will verify SSL servers 4403 and vice versa. 4404 4405 Two new options to the verify program: -untrusted allows a set of 4406 untrusted certificates to be passed in and -purpose which sets the 4407 intended purpose of the certificate. If a purpose is set then the 4408 new chain verify code is used to check extension consistency. 4409 [Steve Henson] 4410 4411 *) Support for the authority information access extension. 4412 [Steve Henson] 4413 4414 *) Modify RSA and DSA PEM read routines to transparently handle 4415 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 4416 public keys in a format compatible with certificate 4417 SubjectPublicKeyInfo structures. Unfortunately there were already 4418 functions called *_PublicKey_* which used various odd formats so 4419 these are retained for compatibility: however the DSA variants were 4420 never in a public release so they have been deleted. Changed dsa/rsa 4421 utilities to handle the new format: note no releases ever handled public 4422 keys so we should be OK. 4423 4424 The primary motivation for this change is to avoid the same fiasco 4425 that dogs private keys: there are several incompatible private key 4426 formats some of which are standard and some OpenSSL specific and 4427 require various evil hacks to allow partial transparent handling and 4428 even then it doesn't work with DER formats. Given the option anything 4429 other than PKCS#8 should be dumped: but the other formats have to 4430 stay in the name of compatibility. 4431 4432 With public keys and the benefit of hindsight one standard format 4433 is used which works with EVP_PKEY, RSA or DSA structures: though 4434 it clearly returns an error if you try to read the wrong kind of key. 4435 4436 Added a -pubkey option to the 'x509' utility to output the public key. 4437 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 4438 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 4439 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 4440 that do the same as the EVP_PKEY_assign_*() except they up the 4441 reference count of the added key (they don't "swallow" the 4442 supplied key). 4443 [Steve Henson] 4444 4445 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 4446 CRLs would fail if the file contained no certificates or no CRLs: 4447 added a new function to read in both types and return the number 4448 read: this means that if none are read it will be an error. The 4449 DER versions of the certificate and CRL reader would always fail 4450 because it isn't possible to mix certificates and CRLs in DER format 4451 without choking one or the other routine. Changed this to just read 4452 a certificate: this is the best we can do. Also modified the code 4453 in apps/verify.c to take notice of return codes: it was previously 4454 attempting to read in certificates from NULL pointers and ignoring 4455 any errors: this is one reason why the cert and CRL reader seemed 4456 to work. It doesn't check return codes from the default certificate 4457 routines: these may well fail if the certificates aren't installed. 4458 [Steve Henson] 4459 4460 *) Code to support otherName option in GeneralName. 4461 [Steve Henson] 4462 4463 *) First update to verify code. Change the verify utility 4464 so it warns if it is passed a self signed certificate: 4465 for consistency with the normal behaviour. X509_verify 4466 has been modified to it will now verify a self signed 4467 certificate if *exactly* the same certificate appears 4468 in the store: it was previously impossible to trust a 4469 single self signed certificate. This means that: 4470 openssl verify ss.pem 4471 now gives a warning about a self signed certificate but 4472 openssl verify -CAfile ss.pem ss.pem 4473 is OK. 4474 [Steve Henson] 4475 4476 *) For servers, store verify_result in SSL_SESSION data structure 4477 (and add it to external session representation). 4478 This is needed when client certificate verifications fails, 4479 but an application-provided verification callback (set by 4480 SSL_CTX_set_cert_verify_callback) allows accepting the session 4481 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 4482 but returns 1): When the session is reused, we have to set 4483 ssl->verify_result to the appropriate error code to avoid 4484 security holes. 4485 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 4486 4487 *) Fix a bug in the new PKCS#7 code: it didn't consider the 4488 case in PKCS7_dataInit() where the signed PKCS7 structure 4489 didn't contain any existing data because it was being created. 4490 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 4491 4492 *) Add a salt to the key derivation routines in enc.c. This 4493 forms the first 8 bytes of the encrypted file. Also add a 4494 -S option to allow a salt to be input on the command line. 4495 [Steve Henson] 4496 4497 *) New function X509_cmp(). Oddly enough there wasn't a function 4498 to compare two certificates. We do this by working out the SHA1 4499 hash and comparing that. X509_cmp() will be needed by the trust 4500 code. 4501 [Steve Henson] 4502 4503 *) SSL_get1_session() is like SSL_get_session(), but increments 4504 the reference count in the SSL_SESSION returned. 4505 [Geoff Thorpe <geoff@eu.c2.net>] 4506 4507 *) Fix for 'req': it was adding a null to request attributes. 4508 Also change the X509_LOOKUP and X509_INFO code to handle 4509 certificate auxiliary information. 4510 [Steve Henson] 4511 4512 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 4513 the 'enc' command. 4514 [Steve Henson] 4515 4516 *) Add the possibility to add extra information to the memory leak 4517 detecting output, to form tracebacks, showing from where each 4518 allocation was originated: CRYPTO_push_info("constant string") adds 4519 the string plus current file name and line number to a per-thread 4520 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 4521 is like calling CYRPTO_pop_info() until the stack is empty. 4522 Also updated memory leak detection code to be multi-thread-safe. 4523 [Richard Levitte] 4524 4525 *) Add options -text and -noout to pkcs7 utility and delete the 4526 encryption options which never did anything. Update docs. 4527 [Steve Henson] 4528 4529 *) Add options to some of the utilities to allow the pass phrase 4530 to be included on either the command line (not recommended on 4531 OSes like Unix) or read from the environment. Update the 4532 manpages and fix a few bugs. 4533 [Steve Henson] 4534 4535 *) Add a few manpages for some of the openssl commands. 4536 [Steve Henson] 4537 4538 *) Fix the -revoke option in ca. It was freeing up memory twice, 4539 leaking and not finding already revoked certificates. 4540 [Steve Henson] 4541 4542 *) Extensive changes to support certificate auxiliary information. 4543 This involves the use of X509_CERT_AUX structure and X509_AUX 4544 functions. An X509_AUX function such as PEM_read_X509_AUX() 4545 can still read in a certificate file in the usual way but it 4546 will also read in any additional "auxiliary information". By 4547 doing things this way a fair degree of compatibility can be 4548 retained: existing certificates can have this information added 4549 using the new 'x509' options. 4550 4551 Current auxiliary information includes an "alias" and some trust 4552 settings. The trust settings will ultimately be used in enhanced 4553 certificate chain verification routines: currently a certificate 4554 can only be trusted if it is self signed and then it is trusted 4555 for all purposes. 4556 [Steve Henson] 4557 4558 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 4559 The problem was that one of the replacement routines had not been working 4560 since SSLeay releases. For now the offending routine has been replaced 4561 with non-optimised assembler. Even so, this now gives around 95% 4562 performance improvement for 1024 bit RSA signs. 4563 [Mark Cox] 4564 4565 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 4566 handling. Most clients have the effective key size in bits equal to 4567 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 4568 A few however don't do this and instead use the size of the decrypted key 4569 to determine the RC2 key length and the AlgorithmIdentifier to determine 4570 the effective key length. In this case the effective key length can still 4571 be 40 bits but the key length can be 168 bits for example. This is fixed 4572 by manually forcing an RC2 key into the EVP_PKEY structure because the 4573 EVP code can't currently handle unusual RC2 key sizes: it always assumes 4574 the key length and effective key length are equal. 4575 [Steve Henson] 4576 4577 *) Add a bunch of functions that should simplify the creation of 4578 X509_NAME structures. Now you should be able to do: 4579 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 4580 and have it automatically work out the correct field type and fill in 4581 the structures. The more adventurous can try: 4582 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 4583 and it will (hopefully) work out the correct multibyte encoding. 4584 [Steve Henson] 4585 4586 *) Change the 'req' utility to use the new field handling and multibyte 4587 copy routines. Before the DN field creation was handled in an ad hoc 4588 way in req, ca, and x509 which was rather broken and didn't support 4589 BMPStrings or UTF8Strings. Since some software doesn't implement 4590 BMPStrings or UTF8Strings yet, they can be enabled using the config file 4591 using the dirstring_type option. See the new comment in the default 4592 openssl.cnf for more info. 4593 [Steve Henson] 4594 4595 *) Make crypto/rand/md_rand.c more robust: 4596 - Assure unique random numbers after fork(). 4597 - Make sure that concurrent threads access the global counter and 4598 md serializably so that we never lose entropy in them 4599 or use exactly the same state in multiple threads. 4600 Access to the large state is not always serializable because 4601 the additional locking could be a performance killer, and 4602 md should be large enough anyway. 4603 [Bodo Moeller] 4604 4605 *) New file apps/app_rand.c with commonly needed functionality 4606 for handling the random seed file. 4607 4608 Use the random seed file in some applications that previously did not: 4609 ca, 4610 dsaparam -genkey (which also ignored its '-rand' option), 4611 s_client, 4612 s_server, 4613 x509 (when signing). 4614 Except on systems with /dev/urandom, it is crucial to have a random 4615 seed file at least for key creation, DSA signing, and for DH exchanges; 4616 for RSA signatures we could do without one. 4617 4618 gendh and gendsa (unlike genrsa) used to read only the first byte 4619 of each file listed in the '-rand' option. The function as previously 4620 found in genrsa is now in app_rand.c and is used by all programs 4621 that support '-rand'. 4622 [Bodo Moeller] 4623 4624 *) In RAND_write_file, use mode 0600 for creating files; 4625 don't just chmod when it may be too late. 4626 [Bodo Moeller] 4627 4628 *) Report an error from X509_STORE_load_locations 4629 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 4630 [Bill Perry] 4631 4632 *) New function ASN1_mbstring_copy() this copies a string in either 4633 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 4634 into an ASN1_STRING type. A mask of permissible types is passed 4635 and it chooses the "minimal" type to use or an error if not type 4636 is suitable. 4637 [Steve Henson] 4638 4639 *) Add function equivalents to the various macros in asn1.h. The old 4640 macros are retained with an M_ prefix. Code inside the library can 4641 use the M_ macros. External code (including the openssl utility) 4642 should *NOT* in order to be "shared library friendly". 4643 [Steve Henson] 4644 4645 *) Add various functions that can check a certificate's extensions 4646 to see if it usable for various purposes such as SSL client, 4647 server or S/MIME and CAs of these types. This is currently 4648 VERY EXPERIMENTAL but will ultimately be used for certificate chain 4649 verification. Also added a -purpose flag to x509 utility to 4650 print out all the purposes. 4651 [Steve Henson] 4652 4653 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 4654 functions. 4655 [Steve Henson] 4656 4657 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 4658 for, obtain and decode and extension and obtain its critical flag. 4659 This allows all the necessary extension code to be handled in a 4660 single function call. 4661 [Steve Henson] 4662 4663 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 4664 platforms. See crypto/rc4/rc4_enc.c for further details. 4665 [Andy Polyakov] 4666 4667 *) New -noout option to asn1parse. This causes no output to be produced 4668 its main use is when combined with -strparse and -out to extract data 4669 from a file (which may not be in ASN.1 format). 4670 [Steve Henson] 4671 4672 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 4673 when producing the local key id. 4674 [Richard Levitte <levitte@stacken.kth.se>] 4675 4676 *) New option -dhparam in s_server. This allows a DH parameter file to be 4677 stated explicitly. If it is not stated then it tries the first server 4678 certificate file. The previous behaviour hard coded the filename 4679 "server.pem". 4680 [Steve Henson] 4681 4682 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 4683 a public key to be input or output. For example: 4684 openssl rsa -in key.pem -pubout -out pubkey.pem 4685 Also added necessary DSA public key functions to handle this. 4686 [Steve Henson] 4687 4688 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 4689 in the message. This was handled by allowing 4690 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 4691 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 4692 4693 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 4694 to the end of the strings whereas this didn't. This would cause problems 4695 if strings read with d2i_ASN1_bytes() were later modified. 4696 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 4697 4698 *) Fix for base64 decode bug. When a base64 bio reads only one line of 4699 data and it contains EOF it will end up returning an error. This is 4700 caused by input 46 bytes long. The cause is due to the way base64 4701 BIOs find the start of base64 encoded data. They do this by trying a 4702 trial decode on each line until they find one that works. When they 4703 do a flag is set and it starts again knowing it can pass all the 4704 data directly through the decoder. Unfortunately it doesn't reset 4705 the context it uses. This means that if EOF is reached an attempt 4706 is made to pass two EOFs through the context and this causes the 4707 resulting error. This can also cause other problems as well. As is 4708 usual with these problems it takes *ages* to find and the fix is 4709 trivial: move one line. 4710 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 4711 4712 *) Ugly workaround to get s_client and s_server working under Windows. The 4713 old code wouldn't work because it needed to select() on sockets and the 4714 tty (for keypresses and to see if data could be written). Win32 only 4715 supports select() on sockets so we select() with a 1s timeout on the 4716 sockets and then see if any characters are waiting to be read, if none 4717 are present then we retry, we also assume we can always write data to 4718 the tty. This isn't nice because the code then blocks until we've 4719 received a complete line of data and it is effectively polling the 4720 keyboard at 1s intervals: however it's quite a bit better than not 4721 working at all :-) A dedicated Windows application might handle this 4722 with an event loop for example. 4723 [Steve Henson] 4724 4725 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 4726 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 4727 will be called when RSA_sign() and RSA_verify() are used. This is useful 4728 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 4729 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 4730 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 4731 This necessitated the support of an extra signature type NID_md5_sha1 4732 for SSL signatures and modifications to the SSL library to use it instead 4733 of calling RSA_public_decrypt() and RSA_private_encrypt(). 4734 [Steve Henson] 4735 4736 *) Add new -verify -CAfile and -CApath options to the crl program, these 4737 will lookup a CRL issuers certificate and verify the signature in a 4738 similar way to the verify program. Tidy up the crl program so it 4739 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 4740 less strict. It will now permit CRL extensions even if it is not 4741 a V2 CRL: this will allow it to tolerate some broken CRLs. 4742 [Steve Henson] 4743 4744 *) Initialize all non-automatic variables each time one of the openssl 4745 sub-programs is started (this is necessary as they may be started 4746 multiple times from the "OpenSSL>" prompt). 4747 [Lennart Bang, Bodo Moeller] 4748 4749 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 4750 removing all other RSA functionality (this is what NO_RSA does). This 4751 is so (for example) those in the US can disable those operations covered 4752 by the RSA patent while allowing storage and parsing of RSA keys and RSA 4753 key generation. 4754 [Steve Henson] 4755 4756 *) Non-copying interface to BIO pairs. 4757 (still largely untested) 4758 [Bodo Moeller] 4759 4760 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 4761 ASCII string. This was handled independently in various places before. 4762 [Steve Henson] 4763 4764 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 4765 UTF8 strings a character at a time. 4766 [Steve Henson] 4767 4768 *) Use client_version from client hello to select the protocol 4769 (s23_srvr.c) and for RSA client key exchange verification 4770 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 4771 [Bodo Moeller] 4772 4773 *) Add various utility functions to handle SPKACs, these were previously 4774 handled by poking round in the structure internals. Added new function 4775 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 4776 print, verify and generate SPKACs. Based on an original idea from 4777 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 4778 [Steve Henson] 4779 4780 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 4781 [Andy Polyakov] 4782 4783 *) Allow the config file extension section to be overwritten on the 4784 command line. Based on an original idea from Massimiliano Pala 4785 <madwolf@comune.modena.it>. The new option is called -extensions 4786 and can be applied to ca, req and x509. Also -reqexts to override 4787 the request extensions in req and -crlexts to override the crl extensions 4788 in ca. 4789 [Steve Henson] 4790 4791 *) Add new feature to the SPKAC handling in ca. Now you can include 4792 the same field multiple times by preceding it by "XXXX." for example: 4793 1.OU="Unit name 1" 4794 2.OU="Unit name 2" 4795 this is the same syntax as used in the req config file. 4796 [Steve Henson] 4797 4798 *) Allow certificate extensions to be added to certificate requests. These 4799 are specified in a 'req_extensions' option of the req section of the 4800 config file. They can be printed out with the -text option to req but 4801 are otherwise ignored at present. 4802 [Steve Henson] 4803 4804 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 4805 data read consists of only the final block it would not decrypted because 4806 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 4807 A misplaced 'break' also meant the decrypted final block might not be 4808 copied until the next read. 4809 [Steve Henson] 4810 4811 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 4812 a few extra parameters to the DH structure: these will be useful if 4813 for example we want the value of 'q' or implement X9.42 DH. 4814 [Steve Henson] 4815 4816 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 4817 provides hooks that allow the default DSA functions or functions on a 4818 "per key" basis to be replaced. This allows hardware acceleration and 4819 hardware key storage to be handled without major modification to the 4820 library. Also added low level modexp hooks and CRYPTO_EX structure and 4821 associated functions. 4822 [Steve Henson] 4823 4824 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 4825 as "read only": it can't be written to and the buffer it points to will 4826 not be freed. Reading from a read only BIO is much more efficient than 4827 a normal memory BIO. This was added because there are several times when 4828 an area of memory needs to be read from a BIO. The previous method was 4829 to create a memory BIO and write the data to it, this results in two 4830 copies of the data and an O(n^2) reading algorithm. There is a new 4831 function BIO_new_mem_buf() which creates a read only memory BIO from 4832 an area of memory. Also modified the PKCS#7 routines to use read only 4833 memory BIOs. 4834 [Steve Henson] 4835 4836 *) Bugfix: ssl23_get_client_hello did not work properly when called in 4837 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 4838 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 4839 but a retry condition occured while trying to read the rest. 4840 [Bodo Moeller] 4841 4842 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 4843 NID_pkcs7_encrypted by default: this was wrong since this should almost 4844 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 4845 the encrypted data type: this is a more sensible place to put it and it 4846 allows the PKCS#12 code to be tidied up that duplicated this 4847 functionality. 4848 [Steve Henson] 4849 4850 *) Changed obj_dat.pl script so it takes its input and output files on 4851 the command line. This should avoid shell escape redirection problems 4852 under Win32. 4853 [Steve Henson] 4854 4855 *) Initial support for certificate extension requests, these are included 4856 in things like Xenroll certificate requests. Included functions to allow 4857 extensions to be obtained and added. 4858 [Steve Henson] 4859 4860 *) -crlf option to s_client and s_server for sending newlines as 4861 CRLF (as required by many protocols). 4862 [Bodo Moeller] 4863 4864 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 4865 4866 *) Install libRSAglue.a when OpenSSL is built with RSAref. 4867 [Ralf S. Engelschall] 4868 4869 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 4870 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 4871 4872 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 4873 program. 4874 [Steve Henson] 4875 4876 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 4877 DH parameters/keys (q is lost during that conversion, but the resulting 4878 DH parameters contain its length). 4879 4880 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 4881 much faster than DH_generate_parameters (which creates parameters 4882 where p = 2*q + 1), and also the smaller q makes DH computations 4883 much more efficient (160-bit exponentiation instead of 1024-bit 4884 exponentiation); so this provides a convenient way to support DHE 4885 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 4886 utter importance to use 4887 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 4888 or 4889 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 4890 when such DH parameters are used, because otherwise small subgroup 4891 attacks may become possible! 4892 [Bodo Moeller] 4893 4894 *) Avoid memory leak in i2d_DHparams. 4895 [Bodo Moeller] 4896 4897 *) Allow the -k option to be used more than once in the enc program: 4898 this allows the same encrypted message to be read by multiple recipients. 4899 [Steve Henson] 4900 4901 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 4902 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 4903 it will always use the numerical form of the OID, even if it has a short 4904 or long name. 4905 [Steve Henson] 4906 4907 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 4908 method only got called if p,q,dmp1,dmq1,iqmp components were present, 4909 otherwise bn_mod_exp was called. In the case of hardware keys for example 4910 no private key components need be present and it might store extra data 4911 in the RSA structure, which cannot be accessed from bn_mod_exp. 4912 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 4913 private key operations. 4914 [Steve Henson] 4915 4916 *) Added support for SPARC Linux. 4917 [Andy Polyakov] 4918 4919 *) pem_password_cb function type incompatibly changed from 4920 typedef int pem_password_cb(char *buf, int size, int rwflag); 4921 to 4922 ....(char *buf, int size, int rwflag, void *userdata); 4923 so that applications can pass data to their callbacks: 4924 The PEM[_ASN1]_{read,write}... functions and macros now take an 4925 additional void * argument, which is just handed through whenever 4926 the password callback is called. 4927 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 4928 4929 New function SSL_CTX_set_default_passwd_cb_userdata. 4930 4931 Compatibility note: As many C implementations push function arguments 4932 onto the stack in reverse order, the new library version is likely to 4933 interoperate with programs that have been compiled with the old 4934 pem_password_cb definition (PEM_whatever takes some data that 4935 happens to be on the stack as its last argument, and the callback 4936 just ignores this garbage); but there is no guarantee whatsoever that 4937 this will work. 4938 4939 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 4940 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 4941 problems not only on Windows, but also on some Unix platforms. 4942 To avoid problematic command lines, these definitions are now in an 4943 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 4944 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 4945 [Bodo Moeller] 4946 4947 *) MIPS III/IV assembler module is reimplemented. 4948 [Andy Polyakov] 4949 4950 *) More DES library cleanups: remove references to srand/rand and 4951 delete an unused file. 4952 [Ulf M�ller] 4953 4954 *) Add support for the the free Netwide assembler (NASM) under Win32, 4955 since not many people have MASM (ml) and it can be hard to obtain. 4956 This is currently experimental but it seems to work OK and pass all 4957 the tests. Check out INSTALL.W32 for info. 4958 [Steve Henson] 4959 4960 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 4961 without temporary keys kept an extra copy of the server key, 4962 and connections with temporary keys did not free everything in case 4963 of an error. 4964 [Bodo Moeller] 4965 4966 *) New function RSA_check_key and new openssl rsa option -check 4967 for verifying the consistency of RSA keys. 4968 [Ulf Moeller, Bodo Moeller] 4969 4970 *) Various changes to make Win32 compile work: 4971 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 4972 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 4973 comparison" warnings. 4974 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 4975 [Steve Henson] 4976 4977 *) Add a debugging option to PKCS#5 v2 key generation function: when 4978 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 4979 derived keys are printed to stderr. 4980 [Steve Henson] 4981 4982 *) Copy the flags in ASN1_STRING_dup(). 4983 [Roman E. Pavlov <pre@mo.msk.ru>] 4984 4985 *) The x509 application mishandled signing requests containing DSA 4986 keys when the signing key was also DSA and the parameters didn't match. 4987 4988 It was supposed to omit the parameters when they matched the signing key: 4989 the verifying software was then supposed to automatically use the CA's 4990 parameters if they were absent from the end user certificate. 4991 4992 Omitting parameters is no longer recommended. The test was also 4993 the wrong way round! This was probably due to unusual behaviour in 4994 EVP_cmp_parameters() which returns 1 if the parameters match. 4995 This meant that parameters were omitted when they *didn't* match and 4996 the certificate was useless. Certificates signed with 'ca' didn't have 4997 this bug. 4998 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 4999 5000 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 5001 The interface is as follows: 5002 Applications can use 5003 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 5004 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 5005 "off" is now the default. 5006 The library internally uses 5007 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 5008 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 5009 to disable memory-checking temporarily. 5010 5011 Some inconsistent states that previously were possible (and were 5012 even the default) are now avoided. 5013 5014 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 5015 with each memory chunk allocated; this is occasionally more helpful 5016 than just having a counter. 5017 5018 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 5019 5020 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 5021 extensions. 5022 [Bodo Moeller] 5023 5024 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 5025 which largely parallels "options", but is for changing API behaviour, 5026 whereas "options" are about protocol behaviour. 5027 Initial "mode" flags are: 5028 5029 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 5030 a single record has been written. 5031 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 5032 retries use the same buffer location. 5033 (But all of the contents must be 5034 copied!) 5035 [Bodo Moeller] 5036 5037 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 5038 worked. 5039 5040 *) Fix problems with no-hmac etc. 5041 [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 5042 5043 *) New functions RSA_get_default_method(), RSA_set_method() and 5044 RSA_get_method(). These allows replacement of RSA_METHODs without having 5045 to mess around with the internals of an RSA structure. 5046 [Steve Henson] 5047 5048 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 5049 Also really enable memory leak checks in openssl.c and in some 5050 test programs. 5051 [Chad C. Mulligan, Bodo Moeller] 5052 5053 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 5054 up the length of negative integers. This has now been simplified to just 5055 store the length when it is first determined and use it later, rather 5056 than trying to keep track of where data is copied and updating it to 5057 point to the end. 5058 [Steve Henson, reported by Brien Wheeler 5059 <bwheeler@authentica-security.com>] 5060 5061 *) Add a new function PKCS7_signatureVerify. This allows the verification 5062 of a PKCS#7 signature but with the signing certificate passed to the 5063 function itself. This contrasts with PKCS7_dataVerify which assumes the 5064 certificate is present in the PKCS#7 structure. This isn't always the 5065 case: certificates can be omitted from a PKCS#7 structure and be 5066 distributed by "out of band" means (such as a certificate database). 5067 [Steve Henson] 5068 5069 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 5070 function prototypes in pem.h, also change util/mkdef.pl to add the 5071 necessary function names. 5072 [Steve Henson] 5073 5074 *) mk1mf.pl (used by Windows builds) did not properly read the 5075 options set by Configure in the top level Makefile, and Configure 5076 was not even able to write more than one option correctly. 5077 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 5078 [Bodo Moeller] 5079 5080 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 5081 file to be loaded from a BIO or FILE pointer. The BIO version will 5082 for example allow memory BIOs to contain config info. 5083 [Steve Henson] 5084 5085 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 5086 Whoever hopes to achieve shared-library compatibility across versions 5087 must use this, not the compile-time macro. 5088 (Exercise 0.9.4: Which is the minimum library version required by 5089 such programs?) 5090 Note: All this applies only to multi-threaded programs, others don't 5091 need locks. 5092 [Bodo Moeller] 5093 5094 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 5095 through a BIO pair triggered the default case, i.e. 5096 SSLerr(...,SSL_R_UNKNOWN_STATE). 5097 [Bodo Moeller] 5098 5099 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 5100 can use the SSL library even if none of the specific BIOs is 5101 appropriate. 5102 [Bodo Moeller] 5103 5104 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 5105 for the encoded length. 5106 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 5107 5108 *) Add initial documentation of the X509V3 functions. 5109 [Steve Henson] 5110 5111 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 5112 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 5113 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 5114 secure PKCS#8 private key format with a high iteration count. 5115 [Steve Henson] 5116 5117 *) Fix determination of Perl interpreter: A perl or perl5 5118 _directory_ in $PATH was also accepted as the interpreter. 5119 [Ralf S. Engelschall] 5120 5121 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 5122 wrong with it but it was very old and did things like calling 5123 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 5124 unusual formatting. 5125 [Steve Henson] 5126 5127 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 5128 to use the new extension code. 5129 [Steve Henson] 5130 5131 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 5132 with macros. This should make it easier to change their form, add extra 5133 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 5134 constant. 5135 [Steve Henson] 5136 5137 *) Add to configuration table a new entry that can specify an alternative 5138 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 5139 according to Mark Crispin <MRC@Panda.COM>. 5140 [Bodo Moeller] 5141 5142#if 0 5143 *) DES CBC did not update the IV. Weird. 5144 [Ben Laurie] 5145#else 5146 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 5147 Changing the behaviour of the former might break existing programs -- 5148 where IV updating is needed, des_ncbc_encrypt can be used. 5149#endif 5150 5151 *) When bntest is run from "make test" it drives bc to check its 5152 calculations, as well as internally checking them. If an internal check 5153 fails, it needs to cause bc to give a non-zero result or make test carries 5154 on without noticing the failure. Fixed. 5155 [Ben Laurie] 5156 5157 *) DES library cleanups. 5158 [Ulf M�ller] 5159 5160 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 5161 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 5162 ciphers. NOTE: although the key derivation function has been verified 5163 against some published test vectors it has not been extensively tested 5164 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 5165 of v2.0. 5166 [Steve Henson] 5167 5168 *) Instead of "mkdir -p", which is not fully portable, use new 5169 Perl script "util/mkdir-p.pl". 5170 [Bodo Moeller] 5171 5172 *) Rewrite the way password based encryption (PBE) is handled. It used to 5173 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 5174 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 5175 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 5176 the 'parameter' field of the AlgorithmIdentifier is passed to the 5177 underlying key generation function so it must do its own ASN1 parsing. 5178 This has also changed the EVP_PBE_CipherInit() function which now has a 5179 'parameter' argument instead of literal salt and iteration count values 5180 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 5181 [Steve Henson] 5182 5183 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 5184 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 5185 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 5186 KEY" because this clashed with PKCS#8 unencrypted string. Since this 5187 value was just used as a "magic string" and not used directly its 5188 value doesn't matter. 5189 [Steve Henson] 5190 5191 *) Introduce some semblance of const correctness to BN. Shame C doesn't 5192 support mutable. 5193 [Ben Laurie] 5194 5195 *) "linux-sparc64" configuration (ultrapenguin). 5196 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 5197 "linux-sparc" configuration. 5198 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 5199 5200 *) config now generates no-xxx options for missing ciphers. 5201 [Ulf M�ller] 5202 5203 *) Support the EBCDIC character set (work in progress). 5204 File ebcdic.c not yet included because it has a different license. 5205 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 5206 5207 *) Support BS2000/OSD-POSIX. 5208 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 5209 5210 *) Make callbacks for key generation use void * instead of char *. 5211 [Ben Laurie] 5212 5213 *) Make S/MIME samples compile (not yet tested). 5214 [Ben Laurie] 5215 5216 *) Additional typesafe stacks. 5217 [Ben Laurie] 5218 5219 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 5220 [Bodo Moeller] 5221 5222 5223 Changes between 0.9.3 and 0.9.3a [29 May 1999] 5224 5225 *) New configuration variant "sco5-gcc". 5226 5227 *) Updated some demos. 5228 [Sean O Riordain, Wade Scholine] 5229 5230 *) Add missing BIO_free at exit of pkcs12 application. 5231 [Wu Zhigang] 5232 5233 *) Fix memory leak in conf.c. 5234 [Steve Henson] 5235 5236 *) Updates for Win32 to assembler version of MD5. 5237 [Steve Henson] 5238 5239 *) Set #! path to perl in apps/der_chop to where we found it 5240 instead of using a fixed path. 5241 [Bodo Moeller] 5242 5243 *) SHA library changes for irix64-mips4-cc. 5244 [Andy Polyakov] 5245 5246 *) Improvements for VMS support. 5247 [Richard Levitte] 5248 5249 5250 Changes between 0.9.2b and 0.9.3 [24 May 1999] 5251 5252 *) Bignum library bug fix. IRIX 6 passes "make test" now! 5253 This also avoids the problems with SC4.2 and unpatched SC5. 5254 [Andy Polyakov <appro@fy.chalmers.se>] 5255 5256 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 5257 These are required because of the typesafe stack would otherwise break 5258 existing code. If old code used a structure member which used to be STACK 5259 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 5260 sk_num or sk_value it would produce an error because the num, data members 5261 are not present in STACK_OF. Now it just produces a warning. sk_set 5262 replaces the old method of assigning a value to sk_value 5263 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 5264 that does this will no longer work (and should use sk_set instead) but 5265 this could be regarded as a "questionable" behaviour anyway. 5266 [Steve Henson] 5267 5268 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 5269 correctly handle encrypted S/MIME data. 5270 [Steve Henson] 5271 5272 *) Change type of various DES function arguments from des_cblock 5273 (which means, in function argument declarations, pointer to char) 5274 to des_cblock * (meaning pointer to array with 8 char elements), 5275 which allows the compiler to do more typechecking; it was like 5276 that back in SSLeay, but with lots of ugly casts. 5277 5278 Introduce new type const_des_cblock. 5279 [Bodo Moeller] 5280 5281 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 5282 problems: find RecipientInfo structure that matches recipient certificate 5283 and initialise the ASN1 structures properly based on passed cipher. 5284 [Steve Henson] 5285 5286 *) Belatedly make the BN tests actually check the results. 5287 [Ben Laurie] 5288 5289 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 5290 to and from BNs: it was completely broken. New compilation option 5291 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 5292 key elements as negative integers. 5293 [Steve Henson] 5294 5295 *) Reorganize and speed up MD5. 5296 [Andy Polyakov <appro@fy.chalmers.se>] 5297 5298 *) VMS support. 5299 [Richard Levitte <richard@levitte.org>] 5300 5301 *) New option -out to asn1parse to allow the parsed structure to be 5302 output to a file. This is most useful when combined with the -strparse 5303 option to examine the output of things like OCTET STRINGS. 5304 [Steve Henson] 5305 5306 *) Make SSL library a little more fool-proof by not requiring any longer 5307 that SSL_set_{accept,connect}_state be called before 5308 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 5309 in many applications because usually everything *appeared* to work as 5310 intended anyway -- now it really works as intended). 5311 [Bodo Moeller] 5312 5313 *) Move openssl.cnf out of lib/. 5314 [Ulf M�ller] 5315 5316 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 5317 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 5318 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 5319 [Ralf S. Engelschall] 5320 5321 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 5322 handle PKCS#7 enveloped data properly. 5323 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 5324 5325 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 5326 copying pointers. The cert_st handling is changed by this in 5327 various ways (and thus what used to be known as ctx->default_cert 5328 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 5329 any longer when s->cert does not give us what we need). 5330 ssl_cert_instantiate becomes obsolete by this change. 5331 As soon as we've got the new code right (possibly it already is?), 5332 we have solved a couple of bugs of the earlier code where s->cert 5333 was used as if it could not have been shared with other SSL structures. 5334 5335 Note that using the SSL API in certain dirty ways now will result 5336 in different behaviour than observed with earlier library versions: 5337 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 5338 does not influence s as it used to. 5339 5340 In order to clean up things more thoroughly, inside SSL_SESSION 5341 we don't use CERT any longer, but a new structure SESS_CERT 5342 that holds per-session data (if available); currently, this is 5343 the peer's certificate chain and, for clients, the server's certificate 5344 and temporary key. CERT holds only those values that can have 5345 meaningful defaults in an SSL_CTX. 5346 [Bodo Moeller] 5347 5348 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 5349 from the internal representation. Various PKCS#7 fixes: remove some 5350 evil casts and set the enc_dig_alg field properly based on the signing 5351 key type. 5352 [Steve Henson] 5353 5354 *) Allow PKCS#12 password to be set from the command line or the 5355 environment. Let 'ca' get its config file name from the environment 5356 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 5357 and 'x509'). 5358 [Steve Henson] 5359 5360 *) Allow certificate policies extension to use an IA5STRING for the 5361 organization field. This is contrary to the PKIX definition but 5362 VeriSign uses it and IE5 only recognises this form. Document 'x509' 5363 extension option. 5364 [Steve Henson] 5365 5366 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 5367 without disallowing inline assembler and the like for non-pedantic builds. 5368 [Ben Laurie] 5369 5370 *) Support Borland C++ builder. 5371 [Janez Jere <jj@void.si>, modified by Ulf M�ller] 5372 5373 *) Support Mingw32. 5374 [Ulf M�ller] 5375 5376 *) SHA-1 cleanups and performance enhancements. 5377 [Andy Polyakov <appro@fy.chalmers.se>] 5378 5379 *) Sparc v8plus assembler for the bignum library. 5380 [Andy Polyakov <appro@fy.chalmers.se>] 5381 5382 *) Accept any -xxx and +xxx compiler options in Configure. 5383 [Ulf M�ller] 5384 5385 *) Update HPUX configuration. 5386 [Anonymous] 5387 5388 *) Add missing sk_<type>_unshift() function to safestack.h 5389 [Ralf S. Engelschall] 5390 5391 *) New function SSL_CTX_use_certificate_chain_file that sets the 5392 "extra_cert"s in addition to the certificate. (This makes sense 5393 only for "PEM" format files, as chains as a whole are not 5394 DER-encoded.) 5395 [Bodo Moeller] 5396 5397 *) Support verify_depth from the SSL API. 5398 x509_vfy.c had what can be considered an off-by-one-error: 5399 Its depth (which was not part of the external interface) 5400 was actually counting the number of certificates in a chain; 5401 now it really counts the depth. 5402 [Bodo Moeller] 5403 5404 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 5405 instead of X509err, which often resulted in confusing error 5406 messages since the error codes are not globally unique 5407 (e.g. an alleged error in ssl3_accept when a certificate 5408 didn't match the private key). 5409 5410 *) New function SSL_CTX_set_session_id_context that allows to set a default 5411 value (so that you don't need SSL_set_session_id_context for each 5412 connection using the SSL_CTX). 5413 [Bodo Moeller] 5414 5415 *) OAEP decoding bug fix. 5416 [Ulf M�ller] 5417 5418 *) Support INSTALL_PREFIX for package builders, as proposed by 5419 David Harris. 5420 [Bodo Moeller] 5421 5422 *) New Configure options "threads" and "no-threads". For systems 5423 where the proper compiler options are known (currently Solaris 5424 and Linux), "threads" is the default. 5425 [Bodo Moeller] 5426 5427 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 5428 [Bodo Moeller] 5429 5430 *) Install various scripts to $(OPENSSLDIR)/misc, not to 5431 $(INSTALLTOP)/bin -- they shouldn't clutter directories 5432 such as /usr/local/bin. 5433 [Bodo Moeller] 5434 5435 *) "make linux-shared" to build shared libraries. 5436 [Niels Poppe <niels@netbox.org>] 5437 5438 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 5439 [Ulf M�ller] 5440 5441 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 5442 extension adding in x509 utility. 5443 [Steve Henson] 5444 5445 *) Remove NOPROTO sections and error code comments. 5446 [Ulf M�ller] 5447 5448 *) Partial rewrite of the DEF file generator to now parse the ANSI 5449 prototypes. 5450 [Steve Henson] 5451 5452 *) New Configure options --prefix=DIR and --openssldir=DIR. 5453 [Ulf M�ller] 5454 5455 *) Complete rewrite of the error code script(s). It is all now handled 5456 by one script at the top level which handles error code gathering, 5457 header rewriting and C source file generation. It should be much better 5458 than the old method: it now uses a modified version of Ulf's parser to 5459 read the ANSI prototypes in all header files (thus the old K&R definitions 5460 aren't needed for error creation any more) and do a better job of 5461 translating function codes into names. The old 'ASN1 error code imbedded 5462 in a comment' is no longer necessary and it doesn't use .err files which 5463 have now been deleted. Also the error code call doesn't have to appear all 5464 on one line (which resulted in some large lines...). 5465 [Steve Henson] 5466 5467 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 5468 [Bodo Moeller] 5469 5470 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 5471 0 (which usually indicates a closed connection), but continue reading. 5472 [Bodo Moeller] 5473 5474 *) Fix some race conditions. 5475 [Bodo Moeller] 5476 5477 *) Add support for CRL distribution points extension. Add Certificate 5478 Policies and CRL distribution points documentation. 5479 [Steve Henson] 5480 5481 *) Move the autogenerated header file parts to crypto/opensslconf.h. 5482 [Ulf M�ller] 5483 5484 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 5485 8 of keying material. Merlin has also confirmed interop with this fix 5486 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 5487 [Merlin Hughes <merlin@baltimore.ie>] 5488 5489 *) Fix lots of warnings. 5490 [Richard Levitte <levitte@stacken.kth.se>] 5491 5492 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 5493 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 5494 [Richard Levitte <levitte@stacken.kth.se>] 5495 5496 *) Fix problems with sizeof(long) == 8. 5497 [Andy Polyakov <appro@fy.chalmers.se>] 5498 5499 *) Change functions to ANSI C. 5500 [Ulf M�ller] 5501 5502 *) Fix typos in error codes. 5503 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 5504 5505 *) Remove defunct assembler files from Configure. 5506 [Ulf M�ller] 5507 5508 *) SPARC v8 assembler BIGNUM implementation. 5509 [Andy Polyakov <appro@fy.chalmers.se>] 5510 5511 *) Support for Certificate Policies extension: both print and set. 5512 Various additions to support the r2i method this uses. 5513 [Steve Henson] 5514 5515 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 5516 return a const string when you are expecting an allocated buffer. 5517 [Ben Laurie] 5518 5519 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 5520 types DirectoryString and DisplayText. 5521 [Steve Henson] 5522 5523 *) Add code to allow r2i extensions to access the configuration database, 5524 add an LHASH database driver and add several ctx helper functions. 5525 [Steve Henson] 5526 5527 *) Fix an evil bug in bn_expand2() which caused various BN functions to 5528 fail when they extended the size of a BIGNUM. 5529 [Steve Henson] 5530 5531 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 5532 support typesafe stack. 5533 [Steve Henson] 5534 5535 *) Fix typo in SSL_[gs]et_options(). 5536 [Nils Frostberg <nils@medcom.se>] 5537 5538 *) Delete various functions and files that belonged to the (now obsolete) 5539 old X509V3 handling code. 5540 [Steve Henson] 5541 5542 *) New Configure option "rsaref". 5543 [Ulf M�ller] 5544 5545 *) Don't auto-generate pem.h. 5546 [Bodo Moeller] 5547 5548 *) Introduce type-safe ASN.1 SETs. 5549 [Ben Laurie] 5550 5551 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 5552 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 5553 5554 *) Introduce type-safe STACKs. This will almost certainly break lots of code 5555 that links with OpenSSL (well at least cause lots of warnings), but fear 5556 not: the conversion is trivial, and it eliminates loads of evil casts. A 5557 few STACKed things have been converted already. Feel free to convert more. 5558 In the fullness of time, I'll do away with the STACK type altogether. 5559 [Ben Laurie] 5560 5561 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 5562 specified in <certfile> by updating the entry in the index.txt file. 5563 This way one no longer has to edit the index.txt file manually for 5564 revoking a certificate. The -revoke option does the gory details now. 5565 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 5566 5567 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 5568 `-text' option at all and this way the `-noout -text' combination was 5569 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 5570 [Ralf S. Engelschall] 5571 5572 *) Make sure a corresponding plain text error message exists for the 5573 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 5574 verify callback function determined that a certificate was revoked. 5575 [Ralf S. Engelschall] 5576 5577 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 5578 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 5579 all available cipers including rc5, which was forgotten until now. 5580 In order to let the testing shell script know which algorithms 5581 are available, a new (up to now undocumented) command 5582 "openssl list-cipher-commands" is used. 5583 [Bodo Moeller] 5584 5585 *) Bugfix: s_client occasionally would sleep in select() when 5586 it should have checked SSL_pending() first. 5587 [Bodo Moeller] 5588 5589 *) New functions DSA_do_sign and DSA_do_verify to provide access to 5590 the raw DSA values prior to ASN.1 encoding. 5591 [Ulf M�ller] 5592 5593 *) Tweaks to Configure 5594 [Niels Poppe <niels@netbox.org>] 5595 5596 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 5597 yet... 5598 [Steve Henson] 5599 5600 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 5601 [Ulf M�ller] 5602 5603 *) New config option to avoid instructions that are illegal on the 80386. 5604 The default code is faster, but requires at least a 486. 5605 [Ulf M�ller] 5606 5607 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 5608 SSL2_SERVER_VERSION (not used at all) macros, which are now the 5609 same as SSL2_VERSION anyway. 5610 [Bodo Moeller] 5611 5612 *) New "-showcerts" option for s_client. 5613 [Bodo Moeller] 5614 5615 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 5616 application. Various cleanups and fixes. 5617 [Steve Henson] 5618 5619 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 5620 modify error routines to work internally. Add error codes and PBE init 5621 to library startup routines. 5622 [Steve Henson] 5623 5624 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 5625 packing functions to asn1 and evp. Changed function names and error 5626 codes along the way. 5627 [Steve Henson] 5628 5629 *) PKCS12 integration: and so it begins... First of several patches to 5630 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 5631 objects to objects.h 5632 [Steve Henson] 5633 5634 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 5635 and display support for Thawte strong extranet extension. 5636 [Steve Henson] 5637 5638 *) Add LinuxPPC support. 5639 [Jeff Dubrule <igor@pobox.org>] 5640 5641 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 5642 bn_div_words in alpha.s. 5643 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 5644 5645 *) Make sure the RSA OAEP test is skipped under -DRSAref because 5646 OAEP isn't supported when OpenSSL is built with RSAref. 5647 [Ulf Moeller <ulf@fitug.de>] 5648 5649 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 5650 so they no longer are missing under -DNOPROTO. 5651 [Soren S. Jorvang <soren@t.dk>] 5652 5653 5654 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 5655 5656 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 5657 doesn't work when the session is reused. Coming soon! 5658 [Ben Laurie] 5659 5660 *) Fix a security hole, that allows sessions to be reused in the wrong 5661 context thus bypassing client cert protection! All software that uses 5662 client certs and session caches in multiple contexts NEEDS PATCHING to 5663 allow session reuse! A fuller solution is in the works. 5664 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 5665 5666 *) Some more source tree cleanups (removed obsolete files 5667 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 5668 permission on "config" script to be executable) and a fix for the INSTALL 5669 document. 5670 [Ulf Moeller <ulf@fitug.de>] 5671 5672 *) Remove some legacy and erroneous uses of malloc, free instead of 5673 Malloc, Free. 5674 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 5675 5676 *) Make rsa_oaep_test return non-zero on error. 5677 [Ulf Moeller <ulf@fitug.de>] 5678 5679 *) Add support for native Solaris shared libraries. Configure 5680 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 5681 if someone would make that last step automatic. 5682 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 5683 5684 *) ctx_size was not built with the right compiler during "make links". Fixed. 5685 [Ben Laurie] 5686 5687 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 5688 except NULL ciphers". This means the default cipher list will no longer 5689 enable NULL ciphers. They need to be specifically enabled e.g. with 5690 the string "DEFAULT:eNULL". 5691 [Steve Henson] 5692 5693 *) Fix to RSA private encryption routines: if p < q then it would 5694 occasionally produce an invalid result. This will only happen with 5695 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 5696 [Steve Henson] 5697 5698 *) Be less restrictive and allow also `perl util/perlpath.pl 5699 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 5700 because this way one can also use an interpreter named `perl5' (which is 5701 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 5702 installed as `perl'). 5703 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 5704 5705 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 5706 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 5707 5708 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 5709 advapi32.lib to Win32 build and change the pem test comparision 5710 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 5711 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 5712 and crypto/des/ede_cbcm_enc.c. 5713 [Steve Henson] 5714 5715 *) DES quad checksum was broken on big-endian architectures. Fixed. 5716 [Ben Laurie] 5717 5718 *) Comment out two functions in bio.h that aren't implemented. Fix up the 5719 Win32 test batch file so it (might) work again. The Win32 test batch file 5720 is horrible: I feel ill.... 5721 [Steve Henson] 5722 5723 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 5724 in e_os.h. Audit of header files to check ANSI and non ANSI 5725 sections: 10 functions were absent from non ANSI section and not exported 5726 from Windows DLLs. Fixed up libeay.num for new functions. 5727 [Steve Henson] 5728 5729 *) Make `openssl version' output lines consistent. 5730 [Ralf S. Engelschall] 5731 5732 *) Fix Win32 symbol export lists for BIO functions: Added 5733 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 5734 to ms/libeay{16,32}.def. 5735 [Ralf S. Engelschall] 5736 5737 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 5738 fine under Unix and passes some trivial tests I've now added. But the 5739 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 5740 added to make sure no one expects that this stuff really works in the 5741 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 5742 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 5743 openssl_bio.xs. 5744 [Ralf S. Engelschall] 5745 5746 *) Fix the generation of two part addresses in perl. 5747 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 5748 5749 *) Add config entry for Linux on MIPS. 5750 [John Tobey <jtobey@channel1.com>] 5751 5752 *) Make links whenever Configure is run, unless we are on Windoze. 5753 [Ben Laurie] 5754 5755 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 5756 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 5757 in CRLs. 5758 [Steve Henson] 5759 5760 *) Add a useful kludge to allow package maintainers to specify compiler and 5761 other platforms details on the command line without having to patch the 5762 Configure script everytime: One now can use ``perl Configure 5763 <id>:<details>'', i.e. platform ids are allowed to have details appended 5764 to them (seperated by colons). This is treated as there would be a static 5765 pre-configured entry in Configure's %table under key <id> with value 5766 <details> and ``perl Configure <id>'' is called. So, when you want to 5767 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 5768 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 5769 now, which overrides the FreeBSD-elf entry on-the-fly. 5770 [Ralf S. Engelschall] 5771 5772 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 5773 [Ben Laurie] 5774 5775 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 5776 on the `perl Configure ...' command line. This way one can compile 5777 OpenSSL libraries with Position Independent Code (PIC) which is needed 5778 for linking it into DSOs. 5779 [Ralf S. Engelschall] 5780 5781 *) Remarkably, export ciphers were totally broken and no-one had noticed! 5782 Fixed. 5783 [Ben Laurie] 5784 5785 *) Cleaned up the LICENSE document: The official contact for any license 5786 questions now is the OpenSSL core team under openssl-core@openssl.org. 5787 And add a paragraph about the dual-license situation to make sure people 5788 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 5789 to the OpenSSL toolkit. 5790 [Ralf S. Engelschall] 5791 5792 *) General source tree makefile cleanups: Made `making xxx in yyy...' 5793 display consistent in the source tree and replaced `/bin/rm' by `rm'. 5794 Additonally cleaned up the `make links' target: Remove unnecessary 5795 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 5796 to speed processing and no longer clutter the display with confusing 5797 stuff. Instead only the actually done links are displayed. 5798 [Ralf S. Engelschall] 5799 5800 *) Permit null encryption ciphersuites, used for authentication only. It used 5801 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 5802 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 5803 encryption. 5804 [Ben Laurie] 5805 5806 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 5807 signed attributes when verifying signatures (this would break them), 5808 the detached data encoding was wrong and public keys obtained using 5809 X509_get_pubkey() weren't freed. 5810 [Steve Henson] 5811 5812 *) Add text documentation for the BUFFER functions. Also added a work around 5813 to a Win95 console bug. This was triggered by the password read stuff: the 5814 last character typed gets carried over to the next fread(). If you were 5815 generating a new cert request using 'req' for example then the last 5816 character of the passphrase would be CR which would then enter the first 5817 field as blank. 5818 [Steve Henson] 5819 5820 *) Added the new `Includes OpenSSL Cryptography Software' button as 5821 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 5822 button and can be used by applications based on OpenSSL to show the 5823 relationship to the OpenSSL project. 5824 [Ralf S. Engelschall] 5825 5826 *) Remove confusing variables in function signatures in files 5827 ssl/ssl_lib.c and ssl/ssl.h. 5828 [Lennart Bong <lob@kulthea.stacken.kth.se>] 5829 5830 *) Don't install bss_file.c under PREFIX/include/ 5831 [Lennart Bong <lob@kulthea.stacken.kth.se>] 5832 5833 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 5834 functions that return function pointers and has support for NT specific 5835 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 5836 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 5837 unsigned to signed types: this was killing the Win32 compile. 5838 [Steve Henson] 5839 5840 *) Add new certificate file to stack functions, 5841 SSL_add_dir_cert_subjects_to_stack() and 5842 SSL_add_file_cert_subjects_to_stack(). These largely supplant 5843 SSL_load_client_CA_file(), and can be used to add multiple certs easily 5844 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 5845 This means that Apache-SSL and similar packages don't have to mess around 5846 to add as many CAs as they want to the preferred list. 5847 [Ben Laurie] 5848 5849 *) Experiment with doxygen documentation. Currently only partially applied to 5850 ssl/ssl_lib.c. 5851 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 5852 openssl.doxy as the configuration file. 5853 [Ben Laurie] 5854 5855 *) Get rid of remaining C++-style comments which strict C compilers hate. 5856 [Ralf S. Engelschall, pointed out by Carlos Amengual] 5857 5858 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 5859 compiled in by default: it has problems with large keys. 5860 [Steve Henson] 5861 5862 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 5863 DH private keys and/or callback functions which directly correspond to 5864 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 5865 is needed for applications which have to configure certificates on a 5866 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 5867 (e.g. s_server). 5868 For the RSA certificate situation is makes no difference, but 5869 for the DSA certificate situation this fixes the "no shared cipher" 5870 problem where the OpenSSL cipher selection procedure failed because the 5871 temporary keys were not overtaken from the context and the API provided 5872 no way to reconfigure them. 5873 The new functions now let applications reconfigure the stuff and they 5874 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 5875 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 5876 non-public-API function ssl_cert_instantiate() is used as a helper 5877 function and also to reduce code redundancy inside ssl_rsa.c. 5878 [Ralf S. Engelschall] 5879 5880 *) Move s_server -dcert and -dkey options out of the undocumented feature 5881 area because they are useful for the DSA situation and should be 5882 recognized by the users. 5883 [Ralf S. Engelschall] 5884 5885 *) Fix the cipher decision scheme for export ciphers: the export bits are 5886 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 5887 SSL_EXP_MASK. So, the original variable has to be used instead of the 5888 already masked variable. 5889 [Richard Levitte <levitte@stacken.kth.se>] 5890 5891 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 5892 [Richard Levitte <levitte@stacken.kth.se>] 5893 5894 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 5895 from `int' to `unsigned int' because it's a length and initialized by 5896 EVP_DigestFinal() which expects an `unsigned int *'. 5897 [Richard Levitte <levitte@stacken.kth.se>] 5898 5899 *) Don't hard-code path to Perl interpreter on shebang line of Configure 5900 script. Instead use the usual Shell->Perl transition trick. 5901 [Ralf S. Engelschall] 5902 5903 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 5904 (in addition to RSA certificates) to match the behaviour of `openssl dsa 5905 -noout -modulus' as it's already the case for `openssl rsa -noout 5906 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 5907 currently the public key is printed (a decision which was already done by 5908 `openssl dsa -modulus' in the past) which serves a similar purpose. 5909 Additionally the NO_RSA no longer completely removes the whole -modulus 5910 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 5911 now, too. 5912 [Ralf S. Engelschall] 5913 5914 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 5915 BIO. See the source (crypto/evp/bio_ok.c) for more info. 5916 [Arne Ansper <arne@ats.cyber.ee>] 5917 5918 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 5919 to be added. Now both 'req' and 'ca' can use new objects defined in the 5920 config file. 5921 [Steve Henson] 5922 5923 *) Add cool BIO that does syslog (or event log on NT). 5924 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 5925 5926 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 5927 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 5928 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 5929 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 5930 [Ben Laurie] 5931 5932 *) Add preliminary config info for new extension code. 5933 [Steve Henson] 5934 5935 *) Make RSA_NO_PADDING really use no padding. 5936 [Ulf Moeller <ulf@fitug.de>] 5937 5938 *) Generate errors when private/public key check is done. 5939 [Ben Laurie] 5940 5941 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 5942 for some CRL extensions and new objects added. 5943 [Steve Henson] 5944 5945 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 5946 key usage extension and fuller support for authority key id. 5947 [Steve Henson] 5948 5949 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 5950 padding method for RSA, which is recommended for new applications in PKCS 5951 #1 v2.0 (RFC 2437, October 1998). 5952 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 5953 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 5954 against Bleichbacher's attack on RSA. 5955 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 5956 Ben Laurie] 5957 5958 *) Updates to the new SSL compression code 5959 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 5960 5961 *) Fix so that the version number in the master secret, when passed 5962 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 5963 (because the server will not accept higher), that the version number 5964 is 0x03,0x01, not 0x03,0x00 5965 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 5966 5967 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 5968 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 5969 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 5970 [Steve Henson] 5971 5972 *) Support for RAW extensions where an arbitrary extension can be 5973 created by including its DER encoding. See apps/openssl.cnf for 5974 an example. 5975 [Steve Henson] 5976 5977 *) Make sure latest Perl versions don't interpret some generated C array 5978 code as Perl array code in the crypto/err/err_genc.pl script. 5979 [Lars Weber <3weber@informatik.uni-hamburg.de>] 5980 5981 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 5982 not many people have the assembler. Various Win32 compilation fixes and 5983 update to the INSTALL.W32 file with (hopefully) more accurate Win32 5984 build instructions. 5985 [Steve Henson] 5986 5987 *) Modify configure script 'Configure' to automatically create crypto/date.h 5988 file under Win32 and also build pem.h from pem.org. New script 5989 util/mkfiles.pl to create the MINFO file on environments that can't do a 5990 'make files': perl util/mkfiles.pl >MINFO should work. 5991 [Steve Henson] 5992 5993 *) Major rework of DES function declarations, in the pursuit of correctness 5994 and purity. As a result, many evil casts evaporated, and some weirdness, 5995 too. You may find this causes warnings in your code. Zapping your evil 5996 casts will probably fix them. Mostly. 5997 [Ben Laurie] 5998 5999 *) Fix for a typo in asn1.h. Bug fix to object creation script 6000 obj_dat.pl. It considered a zero in an object definition to mean 6001 "end of object": none of the objects in objects.h have any zeros 6002 so it wasn't spotted. 6003 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 6004 6005 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 6006 Masking (CBCM). In the absence of test vectors, the best I have been able 6007 to do is check that the decrypt undoes the encrypt, so far. Send me test 6008 vectors if you have them. 6009 [Ben Laurie] 6010 6011 *) Correct calculation of key length for export ciphers (too much space was 6012 allocated for null ciphers). This has not been tested! 6013 [Ben Laurie] 6014 6015 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 6016 message is now correct (it understands "crypto" and "ssl" on its 6017 command line). There is also now an "update" option. This will update 6018 the util/ssleay.num and util/libeay.num files with any new functions. 6019 If you do a: 6020 perl util/mkdef.pl crypto ssl update 6021 it will update them. 6022 [Steve Henson] 6023 6024 *) Overhauled the Perl interface (perl/*): 6025 - ported BN stuff to OpenSSL's different BN library 6026 - made the perl/ source tree CVS-aware 6027 - renamed the package from SSLeay to OpenSSL (the files still contain 6028 their history because I've copied them in the repository) 6029 - removed obsolete files (the test scripts will be replaced 6030 by better Test::Harness variants in the future) 6031 [Ralf S. Engelschall] 6032 6033 *) First cut for a very conservative source tree cleanup: 6034 1. merge various obsolete readme texts into doc/ssleay.txt 6035 where we collect the old documents and readme texts. 6036 2. remove the first part of files where I'm already sure that we no 6037 longer need them because of three reasons: either they are just temporary 6038 files which were left by Eric or they are preserved original files where 6039 I've verified that the diff is also available in the CVS via "cvs diff 6040 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 6041 the crypto/md/ stuff). 6042 [Ralf S. Engelschall] 6043 6044 *) More extension code. Incomplete support for subject and issuer alt 6045 name, issuer and authority key id. Change the i2v function parameters 6046 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 6047 what that's for :-) Fix to ASN1 macro which messed up 6048 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 6049 [Steve Henson] 6050 6051 *) Preliminary support for ENUMERATED type. This is largely copied from the 6052 INTEGER code. 6053 [Steve Henson] 6054 6055 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 6056 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 6057 6058 *) Make sure `make rehash' target really finds the `openssl' program. 6059 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 6060 6061 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 6062 like to hear about it if this slows down other processors. 6063 [Ben Laurie] 6064 6065 *) Add CygWin32 platform information to Configure script. 6066 [Alan Batie <batie@aahz.jf.intel.com>] 6067 6068 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 6069 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 6070 6071 *) New program nseq to manipulate netscape certificate sequences 6072 [Steve Henson] 6073 6074 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 6075 few typos. 6076 [Steve Henson] 6077 6078 *) Fixes to BN code. Previously the default was to define BN_RECURSION 6079 but the BN code had some problems that would cause failures when 6080 doing certificate verification and some other functions. 6081 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 6082 6083 *) Add ASN1 and PEM code to support netscape certificate sequences. 6084 [Steve Henson] 6085 6086 *) Add ASN1 and PEM code to support netscape certificate sequences. 6087 [Steve Henson] 6088 6089 *) Add several PKIX and private extended key usage OIDs. 6090 [Steve Henson] 6091 6092 *) Modify the 'ca' program to handle the new extension code. Modify 6093 openssl.cnf for new extension format, add comments. 6094 [Steve Henson] 6095 6096 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 6097 and add a sample to openssl.cnf so req -x509 now adds appropriate 6098 CA extensions. 6099 [Steve Henson] 6100 6101 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 6102 error code, add initial support to X509_print() and x509 application. 6103 [Steve Henson] 6104 6105 *) Takes a deep breath and start addding X509 V3 extension support code. Add 6106 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 6107 stuff is currently isolated and isn't even compiled yet. 6108 [Steve Henson] 6109 6110 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 6111 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 6112 Removed the versions check from X509 routines when loading extensions: 6113 this allows certain broken certificates that don't set the version 6114 properly to be processed. 6115 [Steve Henson] 6116 6117 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 6118 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 6119 can still be regenerated with "make depend". 6120 [Ben Laurie] 6121 6122 *) Spelling mistake in C version of CAST-128. 6123 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 6124 6125 *) Changes to the error generation code. The perl script err-code.pl 6126 now reads in the old error codes and retains the old numbers, only 6127 adding new ones if necessary. It also only changes the .err files if new 6128 codes are added. The makefiles have been modified to only insert errors 6129 when needed (to avoid needlessly modifying header files). This is done 6130 by only inserting errors if the .err file is newer than the auto generated 6131 C file. To rebuild all the error codes from scratch (the old behaviour) 6132 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 6133 or delete all the .err files. 6134 [Steve Henson] 6135 6136 *) CAST-128 was incorrectly implemented for short keys. The C version has 6137 been fixed, but is untested. The assembler versions are also fixed, but 6138 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 6139 to regenerate it if needed. 6140 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 6141 Hagino <itojun@kame.net>] 6142 6143 *) File was opened incorrectly in randfile.c. 6144 [Ulf M�ller <ulf@fitug.de>] 6145 6146 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 6147 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 6148 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 6149 al: it's just almost always a UTCTime. Note this patch adds new error 6150 codes so do a "make errors" if there are problems. 6151 [Steve Henson] 6152 6153 *) Correct Linux 1 recognition in config. 6154 [Ulf M�ller <ulf@fitug.de>] 6155 6156 *) Remove pointless MD5 hash when using DSA keys in ca. 6157 [Anonymous <nobody@replay.com>] 6158 6159 *) Generate an error if given an empty string as a cert directory. Also 6160 generate an error if handed NULL (previously returned 0 to indicate an 6161 error, but didn't set one). 6162 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 6163 6164 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 6165 [Ben Laurie] 6166 6167 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 6168 parameters. This was causing a warning which killed off the Win32 compile. 6169 [Steve Henson] 6170 6171 *) Remove C++ style comments from crypto/bn/bn_local.h. 6172 [Neil Costigan <neil.costigan@celocom.com>] 6173 6174 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 6175 based on a text string, looking up short and long names and finally 6176 "dot" format. The "dot" format stuff didn't work. Added new function 6177 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 6178 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 6179 OID is not part of the table. 6180 [Steve Henson] 6181 6182 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 6183 X509_LOOKUP_by_alias(). 6184 [Ben Laurie] 6185 6186 *) Sort openssl functions by name. 6187 [Ben Laurie] 6188 6189 *) Get the gendsa program working (hopefully) and add it to app list. Remove 6190 encryption from sample DSA keys (in case anyone is interested the password 6191 was "1234"). 6192 [Steve Henson] 6193 6194 *) Make _all_ *_free functions accept a NULL pointer. 6195 [Frans Heymans <fheymans@isaserver.be>] 6196 6197 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 6198 NULL pointers. 6199 [Anonymous <nobody@replay.com>] 6200 6201 *) s_server should send the CAfile as acceptable CAs, not its own cert. 6202 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 6203 6204 *) Don't blow it for numeric -newkey arguments to apps/req. 6205 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 6206 6207 *) Temp key "for export" tests were wrong in s3_srvr.c. 6208 [Anonymous <nobody@replay.com>] 6209 6210 *) Add prototype for temp key callback functions 6211 SSL_CTX_set_tmp_{rsa,dh}_callback(). 6212 [Ben Laurie] 6213 6214 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 6215 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 6216 [Steve Henson] 6217 6218 *) X509_name_add_entry() freed the wrong thing after an error. 6219 [Arne Ansper <arne@ats.cyber.ee>] 6220 6221 *) rsa_eay.c would attempt to free a NULL context. 6222 [Arne Ansper <arne@ats.cyber.ee>] 6223 6224 *) BIO_s_socket() had a broken should_retry() on Windoze. 6225 [Arne Ansper <arne@ats.cyber.ee>] 6226 6227 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 6228 [Arne Ansper <arne@ats.cyber.ee>] 6229 6230 *) Make sure the already existing X509_STORE->depth variable is initialized 6231 in X509_STORE_new(), but document the fact that this variable is still 6232 unused in the certificate verification process. 6233 [Ralf S. Engelschall] 6234 6235 *) Fix the various library and apps files to free up pkeys obtained from 6236 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 6237 [Steve Henson] 6238 6239 *) Fix reference counting in X509_PUBKEY_get(). This makes 6240 demos/maurice/example2.c work, amongst others, probably. 6241 [Steve Henson and Ben Laurie] 6242 6243 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 6244 `openssl' and second, the shortcut symlinks for the `openssl <command>' 6245 are no longer created. This way we have a single and consistent command 6246 line interface `openssl <command>', similar to `cvs <command>'. 6247 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 6248 6249 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 6250 BIT STRING wrapper always have zero unused bits. 6251 [Steve Henson] 6252 6253 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 6254 [Steve Henson] 6255 6256 *) Make the top-level INSTALL documentation easier to understand. 6257 [Paul Sutton] 6258 6259 *) Makefiles updated to exit if an error occurs in a sub-directory 6260 make (including if user presses ^C) [Paul Sutton] 6261 6262 *) Make Montgomery context stuff explicit in RSA data structure. 6263 [Ben Laurie] 6264 6265 *) Fix build order of pem and err to allow for generated pem.h. 6266 [Ben Laurie] 6267 6268 *) Fix renumbering bug in X509_NAME_delete_entry(). 6269 [Ben Laurie] 6270 6271 *) Enhanced the err-ins.pl script so it makes the error library number 6272 global and can add a library name. This is needed for external ASN1 and 6273 other error libraries. 6274 [Steve Henson] 6275 6276 *) Fixed sk_insert which never worked properly. 6277 [Steve Henson] 6278 6279 *) Fix ASN1 macros so they can handle indefinite length construted 6280 EXPLICIT tags. Some non standard certificates use these: they can now 6281 be read in. 6282 [Steve Henson] 6283 6284 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 6285 into a single doc/ssleay.txt bundle. This way the information is still 6286 preserved but no longer messes up this directory. Now it's new room for 6287 the new set of documenation files. 6288 [Ralf S. Engelschall] 6289 6290 *) SETs were incorrectly DER encoded. This was a major pain, because they 6291 shared code with SEQUENCEs, which aren't coded the same. This means that 6292 almost everything to do with SETs or SEQUENCEs has either changed name or 6293 number of arguments. 6294 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 6295 6296 *) Fix test data to work with the above. 6297 [Ben Laurie] 6298 6299 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 6300 was already fixed by Eric for 0.9.1 it seems. 6301 [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 6302 6303 *) Autodetect FreeBSD3. 6304 [Ben Laurie] 6305 6306 *) Fix various bugs in Configure. This affects the following platforms: 6307 nextstep 6308 ncr-scde 6309 unixware-2.0 6310 unixware-2.0-pentium 6311 sco5-cc. 6312 [Ben Laurie] 6313 6314 *) Eliminate generated files from CVS. Reorder tests to regenerate files 6315 before they are needed. 6316 [Ben Laurie] 6317 6318 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 6319 [Ben Laurie] 6320 6321 6322 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 6323 6324 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 6325 changed SSLeay to OpenSSL in version strings. 6326 [Ralf S. Engelschall] 6327 6328 *) Some fixups to the top-level documents. 6329 [Paul Sutton] 6330 6331 *) Fixed the nasty bug where rsaref.h was not found under compile-time 6332 because the symlink to include/ was missing. 6333 [Ralf S. Engelschall] 6334 6335 *) Incorporated the popular no-RSA/DSA-only patches 6336 which allow to compile a RSA-free SSLeay. 6337 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 6338 6339 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 6340 when "ssleay" is still not found. 6341 [Ralf S. Engelschall] 6342 6343 *) Added more platforms to Configure: Cray T3E, HPUX 11, 6344 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 6345 6346 *) Updated the README file. 6347 [Ralf S. Engelschall] 6348 6349 *) Added various .cvsignore files in the CVS repository subdirs 6350 to make a "cvs update" really silent. 6351 [Ralf S. Engelschall] 6352 6353 *) Recompiled the error-definition header files and added 6354 missing symbols to the Win32 linker tables. 6355 [Ralf S. Engelschall] 6356 6357 *) Cleaned up the top-level documents; 6358 o new files: CHANGES and LICENSE 6359 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 6360 o merged COPYRIGHT into LICENSE 6361 o removed obsolete TODO file 6362 o renamed MICROSOFT to INSTALL.W32 6363 [Ralf S. Engelschall] 6364 6365 *) Removed dummy files from the 0.9.1b source tree: 6366 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 6367 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 6368 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 6369 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 6370 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 6371 [Ralf S. Engelschall] 6372 6373 *) Added various platform portability fixes. 6374 [Mark J. Cox] 6375 6376 *) The Genesis of the OpenSSL rpject: 6377 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 6378 Young and Tim J. Hudson created while they were working for C2Net until 6379 summer 1998. 6380 [The OpenSSL Project] 6381 6382 6383 Changes between 0.9.0b and 0.9.1b [not released] 6384 6385 *) Updated a few CA certificates under certs/ 6386 [Eric A. Young] 6387 6388 *) Changed some BIGNUM api stuff. 6389 [Eric A. Young] 6390 6391 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 6392 DGUX x86, Linux Alpha, etc. 6393 [Eric A. Young] 6394 6395 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 6396 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 6397 available). 6398 [Eric A. Young] 6399 6400 *) Add -strparse option to asn1pars program which parses nested 6401 binary structures 6402 [Dr Stephen Henson <shenson@bigfoot.com>] 6403 6404 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 6405 [Eric A. Young] 6406 6407 *) DSA fix for "ca" program. 6408 [Eric A. Young] 6409 6410 *) Added "-genkey" option to "dsaparam" program. 6411 [Eric A. Young] 6412 6413 *) Added RIPE MD160 (rmd160) message digest. 6414 [Eric A. Young] 6415 6416 *) Added -a (all) option to "ssleay version" command. 6417 [Eric A. Young] 6418 6419 *) Added PLATFORM define which is the id given to Configure. 6420 [Eric A. Young] 6421 6422 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 6423 [Eric A. Young] 6424 6425 *) Extended the ASN.1 parser routines. 6426 [Eric A. Young] 6427 6428 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 6429 [Eric A. Young] 6430 6431 *) Added a BN_CTX to the BN library. 6432 [Eric A. Young] 6433 6434 *) Fixed the weak key values in DES library 6435 [Eric A. Young] 6436 6437 *) Changed API in EVP library for cipher aliases. 6438 [Eric A. Young] 6439 6440 *) Added support for RC2/64bit cipher. 6441 [Eric A. Young] 6442 6443 *) Converted the lhash library to the crypto/mem.c functions. 6444 [Eric A. Young] 6445 6446 *) Added more recognized ASN.1 object ids. 6447 [Eric A. Young] 6448 6449 *) Added more RSA padding checks for SSL/TLS. 6450 [Eric A. Young] 6451 6452 *) Added BIO proxy/filter functionality. 6453 [Eric A. Young] 6454 6455 *) Added extra_certs to SSL_CTX which can be used 6456 send extra CA certificates to the client in the CA cert chain sending 6457 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 6458 [Eric A. Young] 6459 6460 *) Now Fortezza is denied in the authentication phase because 6461 this is key exchange mechanism is not supported by SSLeay at all. 6462 [Eric A. Young] 6463 6464 *) Additional PKCS1 checks. 6465 [Eric A. Young] 6466 6467 *) Support the string "TLSv1" for all TLS v1 ciphers. 6468 [Eric A. Young] 6469 6470 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 6471 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 6472 [Eric A. Young] 6473 6474 *) Fixed a few memory leaks. 6475 [Eric A. Young] 6476 6477 *) Fixed various code and comment typos. 6478 [Eric A. Young] 6479 6480 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 6481 bytes sent in the client random. 6482 [Edward Bishop <ebishop@spyglass.com>] 6483 6484