CHANGES revision 127128
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 6 7 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 8 by using the Codenomicon TLS Test Tool (CAN-2004-0079) 9 [Joe Orton, Steve Henson] 10 11 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 12 (CAN-2004-0112) 13 [Joe Orton, Steve Henson] 14 15 *) Make it possible to have multiple active certificates with the same 16 subject in the CA index file. This is done only if the keyword 17 'unique_subject' is set to 'no' in the main CA section (default 18 if 'CA_default') of the configuration file. The value is saved 19 with the database itself in a separate index attribute file, 20 named like the index file with '.attr' appended to the name. 21 [Richard Levitte] 22 23 *) X509 verify fixes. Disable broken certificate workarounds when 24 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 25 keyUsage extension present. Don't accept CRLs with unhandled critical 26 extensions: since verify currently doesn't process CRL extensions this 27 rejects a CRL with *any* critical extensions. Add new verify error codes 28 for these cases. 29 [Steve Henson] 30 31 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 32 A clarification of RFC2560 will require the use of OCTET STRINGs and 33 some implementations cannot handle the current raw format. Since OpenSSL 34 copies and compares OCSP nonces as opaque blobs without any attempt at 35 parsing them this should not create any compatibility issues. 36 [Steve Henson] 37 38 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 39 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 40 this HMAC (and other) operations are several times slower than OpenSSL 41 < 0.9.7. 42 [Steve Henson] 43 44 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 45 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 46 47 *) Use the correct content when signing type "other". 48 [Steve Henson] 49 50 Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 51 52 *) Fix various bugs revealed by running the NISCC test suite: 53 54 Stop out of bounds reads in the ASN1 code when presented with 55 invalid tags (CAN-2003-0543 and CAN-2003-0544). 56 57 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545). 58 59 If verify callback ignores invalid public key errors don't try to check 60 certificate signature with the NULL public key. 61 62 [Steve Henson] 63 64 *) New -ignore_err option in ocsp application to stop the server 65 exiting on the first error in a request. 66 [Steve Henson] 67 68 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 69 if the server requested one: as stated in TLS 1.0 and SSL 3.0 70 specifications. 71 [Steve Henson] 72 73 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 74 extra data after the compression methods not only for TLS 1.0 75 but also for SSL 3.0 (as required by the specification). 76 [Bodo Moeller; problem pointed out by Matthias Loepfe] 77 78 *) Change X509_certificate_type() to mark the key as exported/exportable 79 when it's 512 *bits* long, not 512 bytes. 80 [Richard Levitte] 81 82 *) Change AES_cbc_encrypt() so it outputs exact multiple of 83 blocks during encryption. 84 [Richard Levitte] 85 86 *) Various fixes to base64 BIO and non blocking I/O. On write 87 flushes were not handled properly if the BIO retried. On read 88 data was not being buffered properly and had various logic bugs. 89 This also affects blocking I/O when the data being decoded is a 90 certain size. 91 [Steve Henson] 92 93 *) Various S/MIME bugfixes and compatibility changes: 94 output correct application/pkcs7 MIME type if 95 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 96 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 97 of files as .eml work). Correctly handle very long lines in MIME 98 parser. 99 [Steve Henson] 100 101 Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 102 103 *) Countermeasure against the Klima-Pokorny-Rosa extension of 104 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 105 a protocol version number mismatch like a decryption error 106 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 107 [Bodo Moeller] 108 109 *) Turn on RSA blinding by default in the default implementation 110 to avoid a timing attack. Applications that don't want it can call 111 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 112 They would be ill-advised to do so in most cases. 113 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 114 115 *) Change RSA blinding code so that it works when the PRNG is not 116 seeded (in this case, the secret RSA exponent is abused as 117 an unpredictable seed -- if it is not unpredictable, there 118 is no point in blinding anyway). Make RSA blinding thread-safe 119 by remembering the creator's thread ID in rsa->blinding and 120 having all other threads use local one-time blinding factors 121 (this requires more computation than sharing rsa->blinding, but 122 avoids excessive locking; and if an RSA object is not shared 123 between threads, blinding will still be very fast). 124 [Bodo Moeller] 125 126 *) Fixed a typo bug that would cause ENGINE_set_default() to set an 127 ENGINE as defaults for all supported algorithms irrespective of 128 the 'flags' parameter. 'flags' is now honoured, so applications 129 should make sure they are passing it correctly. 130 [Geoff Thorpe] 131 132 *) Target "mingw" now allows native Windows code to be generated in 133 the Cygwin environment as well as with the MinGW compiler. 134 [Ulf Moeller] 135 136 Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 137 138 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 139 via timing by performing a MAC computation even if incorrrect 140 block cipher padding has been found. This is a countermeasure 141 against active attacks where the attacker has to distinguish 142 between bad padding and a MAC verification error. (CAN-2003-0078) 143 144 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 145 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 146 Martin Vuagnoux (EPFL, Ilion)] 147 148 *) Make the no-err option work as intended. The intention with no-err 149 is not to have the whole error stack handling routines removed from 150 libcrypto, it's only intended to remove all the function name and 151 reason texts, thereby removing some of the footprint that may not 152 be interesting if those errors aren't displayed anyway. 153 154 NOTE: it's still possible for any application or module to have it's 155 own set of error texts inserted. The routines are there, just not 156 used by default when no-err is given. 157 [Richard Levitte] 158 159 *) Add support for FreeBSD on IA64. 160 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 161 162 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 163 Kerberos function mit_des_cbc_cksum(). Before this change, 164 the value returned by DES_cbc_cksum() was like the one from 165 mit_des_cbc_cksum(), except the bytes were swapped. 166 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 167 168 *) Allow an application to disable the automatic SSL chain building. 169 Before this a rather primitive chain build was always performed in 170 ssl3_output_cert_chain(): an application had no way to send the 171 correct chain if the automatic operation produced an incorrect result. 172 173 Now the chain builder is disabled if either: 174 175 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 176 177 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 178 179 The reasoning behind this is that an application would not want the 180 auto chain building to take place if extra chain certificates are 181 present and it might also want a means of sending no additional 182 certificates (for example the chain has two certificates and the 183 root is omitted). 184 [Steve Henson] 185 186 *) Add the possibility to build without the ENGINE framework. 187 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 188 189 *) Under Win32 gmtime() can return NULL: check return value in 190 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 191 [Steve Henson] 192 193 *) DSA routines: under certain error conditions uninitialized BN objects 194 could be freed. Solution: make sure initialization is performed early 195 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 196 Nils Larsch <nla@trustcenter.de> via PR#459) 197 [Lutz Jaenicke] 198 199 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 200 checked on reconnect on the client side, therefore session resumption 201 could still fail with a "ssl session id is different" error. This 202 behaviour is masked when SSL_OP_ALL is used due to 203 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 204 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 205 followup to PR #377. 206 [Lutz Jaenicke] 207 208 *) IA-32 assembler support enhancements: unified ELF targets, support 209 for SCO/Caldera platforms, fix for Cygwin shared build. 210 [Andy Polyakov] 211 212 *) Add support for FreeBSD on sparc64. As a consequence, support for 213 FreeBSD on non-x86 processors is separate from x86 processors on 214 the config script, much like the NetBSD support. 215 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 216 217 Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 218 219 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 220 OpenSSL 0.9.7.] 221 222 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 223 code (06) was taken as the first octet of the session ID and the last 224 octet was ignored consequently. As a result SSLv2 client side session 225 caching could not have worked due to the session ID mismatch between 226 client and server. 227 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 228 PR #377. 229 [Lutz Jaenicke] 230 231 *) Change the declaration of needed Kerberos libraries to use EX_LIBS 232 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 233 removed entirely. 234 [Richard Levitte] 235 236 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 237 seems that in spite of existing for more than a year, many application 238 author have done nothing to provide the necessary callbacks, which 239 means that this particular engine will not work properly anywhere. 240 This is a very unfortunate situation which forces us, in the name 241 of usability, to give the hw_ncipher.c a static lock, which is part 242 of libcrypto. 243 NOTE: This is for the 0.9.7 series ONLY. This hack will never 244 appear in 0.9.8 or later. We EXPECT application authors to have 245 dealt properly with this when 0.9.8 is released (unless we actually 246 make such changes in the libcrypto locking code that changes will 247 have to be made anyway). 248 [Richard Levitte] 249 250 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 251 octets have been read, EOF or an error occurs. Without this change 252 some truncated ASN1 structures will not produce an error. 253 [Steve Henson] 254 255 *) Disable Heimdal support, since it hasn't been fully implemented. 256 Still give the possibility to force the use of Heimdal, but with 257 warnings and a request that patches get sent to openssl-dev. 258 [Richard Levitte] 259 260 *) Add the VC-CE target, introduce the WINCE sysname, and add 261 INSTALL.WCE and appropriate conditionals to make it build. 262 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 263 264 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 265 cygssl-x.y.z.dll, where x, y and z are the major, minor and 266 edit numbers of the version. 267 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 268 269 *) Introduce safe string copy and catenation functions 270 (BUF_strlcpy() and BUF_strlcat()). 271 [Ben Laurie (CHATS) and Richard Levitte] 272 273 *) Avoid using fixed-size buffers for one-line DNs. 274 [Ben Laurie (CHATS)] 275 276 *) Add BUF_MEM_grow_clean() to avoid information leakage when 277 resizing buffers containing secrets, and use where appropriate. 278 [Ben Laurie (CHATS)] 279 280 *) Avoid using fixed size buffers for configuration file location. 281 [Ben Laurie (CHATS)] 282 283 *) Avoid filename truncation for various CA files. 284 [Ben Laurie (CHATS)] 285 286 *) Use sizeof in preference to magic numbers. 287 [Ben Laurie (CHATS)] 288 289 *) Avoid filename truncation in cert requests. 290 [Ben Laurie (CHATS)] 291 292 *) Add assertions to check for (supposedly impossible) buffer 293 overflows. 294 [Ben Laurie (CHATS)] 295 296 *) Don't cache truncated DNS entries in the local cache (this could 297 potentially lead to a spoofing attack). 298 [Ben Laurie (CHATS)] 299 300 *) Fix various buffers to be large enough for hex/decimal 301 representations in a platform independent manner. 302 [Ben Laurie (CHATS)] 303 304 *) Add CRYPTO_realloc_clean() to avoid information leakage when 305 resizing buffers containing secrets, and use where appropriate. 306 [Ben Laurie (CHATS)] 307 308 *) Add BIO_indent() to avoid much slightly worrying code to do 309 indents. 310 [Ben Laurie (CHATS)] 311 312 *) Convert sprintf()/BIO_puts() to BIO_printf(). 313 [Ben Laurie (CHATS)] 314 315 *) buffer_gets() could terminate with the buffer only half 316 full. Fixed. 317 [Ben Laurie (CHATS)] 318 319 *) Add assertions to prevent user-supplied crypto functions from 320 overflowing internal buffers by having large block sizes, etc. 321 [Ben Laurie (CHATS)] 322 323 *) New OPENSSL_assert() macro (similar to assert(), but enabled 324 unconditionally). 325 [Ben Laurie (CHATS)] 326 327 *) Eliminate unused copy of key in RC4. 328 [Ben Laurie (CHATS)] 329 330 *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 331 [Ben Laurie (CHATS)] 332 333 *) Fix off-by-one error in EGD path. 334 [Ben Laurie (CHATS)] 335 336 *) If RANDFILE path is too long, ignore instead of truncating. 337 [Ben Laurie (CHATS)] 338 339 *) Eliminate unused and incorrectly sized X.509 structure 340 CBCParameter. 341 [Ben Laurie (CHATS)] 342 343 *) Eliminate unused and dangerous function knumber(). 344 [Ben Laurie (CHATS)] 345 346 *) Eliminate unused and dangerous structure, KSSL_ERR. 347 [Ben Laurie (CHATS)] 348 349 *) Protect against overlong session ID context length in an encoded 350 session object. Since these are local, this does not appear to be 351 exploitable. 352 [Ben Laurie (CHATS)] 353 354 *) Change from security patch (see 0.9.6e below) that did not affect 355 the 0.9.6 release series: 356 357 Remote buffer overflow in SSL3 protocol - an attacker could 358 supply an oversized master key in Kerberos-enabled versions. 359 (CAN-2002-0657) 360 [Ben Laurie (CHATS)] 361 362 *) Change the SSL kerb5 codes to match RFC 2712. 363 [Richard Levitte] 364 365 *) Make -nameopt work fully for req and add -reqopt switch. 366 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 367 368 *) The "block size" for block ciphers in CFB and OFB mode should be 1. 369 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 370 371 *) Make sure tests can be performed even if the corresponding algorithms 372 have been removed entirely. This was also the last step to make 373 OpenSSL compilable with DJGPP under all reasonable conditions. 374 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 375 376 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 377 to allow version independent disabling of normally unselected ciphers, 378 which may be activated as a side-effect of selecting a single cipher. 379 380 (E.g., cipher list string "RSA" enables ciphersuites that are left 381 out of "ALL" because they do not provide symmetric encryption. 382 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 383 [Lutz Jaenicke, Bodo Moeller] 384 385 *) Add appropriate support for separate platform-dependent build 386 directories. The recommended way to make a platform-dependent 387 build directory is the following (tested on Linux), maybe with 388 some local tweaks: 389 390 # Place yourself outside of the OpenSSL source tree. In 391 # this example, the environment variable OPENSSL_SOURCE 392 # is assumed to contain the absolute OpenSSL source directory. 393 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 394 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 395 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 396 mkdir -p `dirname $F` 397 ln -s $OPENSSL_SOURCE/$F $F 398 done 399 400 To be absolutely sure not to disturb the source tree, a "make clean" 401 is a good thing. If it isn't successfull, don't worry about it, 402 it probably means the source directory is very clean. 403 [Richard Levitte] 404 405 *) Make sure any ENGINE control commands make local copies of string 406 pointers passed to them whenever necessary. Otherwise it is possible 407 the caller may have overwritten (or deallocated) the original string 408 data when a later ENGINE operation tries to use the stored values. 409 [G�tz Babin-Ebell <babinebell@trustcenter.de>] 410 411 *) Improve diagnostics in file reading and command-line digests. 412 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 413 414 *) Add AES modes CFB and OFB to the object database. Correct an 415 error in AES-CFB decryption. 416 [Richard Levitte] 417 418 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 419 allows existing EVP_CIPHER_CTX structures to be reused after 420 calling EVP_*Final(). This behaviour is used by encryption 421 BIOs and some applications. This has the side effect that 422 applications must explicitly clean up cipher contexts with 423 EVP_CIPHER_CTX_cleanup() or they will leak memory. 424 [Steve Henson] 425 426 *) Check the values of dna and dnb in bn_mul_recursive before calling 427 bn_mul_comba (a non zero value means the a or b arrays do not contain 428 n2 elements) and fallback to bn_mul_normal if either is not zero. 429 [Steve Henson] 430 431 *) Fix escaping of non-ASCII characters when using the -subj option 432 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 433 [Lutz Jaenicke] 434 435 *) Make object definitions compliant to LDAP (RFC2256): SN is the short 436 form for "surname", serialNumber has no short form. 437 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 438 therefore remove "mail" short name for "internet 7". 439 The OID for unique identifiers in X509 certificates is 440 x500UniqueIdentifier, not uniqueIdentifier. 441 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 442 [Lutz Jaenicke] 443 444 *) Add an "init" command to the ENGINE config module and auto initialize 445 ENGINEs. Without any "init" command the ENGINE will be initialized 446 after all ctrl commands have been executed on it. If init=1 the 447 ENGINE is initailized at that point (ctrls before that point are run 448 on the uninitialized ENGINE and after on the initialized one). If 449 init=0 then the ENGINE will not be iniatialized at all. 450 [Steve Henson] 451 452 *) Fix the 'app_verify_callback' interface so that the user-defined 453 argument is actually passed to the callback: In the 454 SSL_CTX_set_cert_verify_callback() prototype, the callback 455 declaration has been changed from 456 int (*cb)() 457 into 458 int (*cb)(X509_STORE_CTX *,void *); 459 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 460 i=s->ctx->app_verify_callback(&ctx) 461 has been changed into 462 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 463 464 To update applications using SSL_CTX_set_cert_verify_callback(), 465 a dummy argument can be added to their callback functions. 466 [D. K. Smetters <smetters@parc.xerox.com>] 467 468 *) Added the '4758cca' ENGINE to support IBM 4758 cards. 469 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 470 471 *) Add and OPENSSL_LOAD_CONF define which will cause 472 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 473 This allows older applications to transparently support certain 474 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 475 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 476 load the config file and OPENSSL_add_all_algorithms_conf() which will 477 always load it have also been added. 478 [Steve Henson] 479 480 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 481 Adjust NIDs and EVP layer. 482 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 483 484 *) Config modules support in openssl utility. 485 486 Most commands now load modules from the config file, 487 though in a few (such as version) this isn't done 488 because it couldn't be used for anything. 489 490 In the case of ca and req the config file used is 491 the same as the utility itself: that is the -config 492 command line option can be used to specify an 493 alternative file. 494 [Steve Henson] 495 496 *) Move default behaviour from OPENSSL_config(). If appname is NULL 497 use "openssl_conf" if filename is NULL use default openssl config file. 498 [Steve Henson] 499 500 *) Add an argument to OPENSSL_config() to allow the use of an alternative 501 config section name. Add a new flag to tolerate a missing config file 502 and move code to CONF_modules_load_file(). 503 [Steve Henson] 504 505 *) Support for crypto accelerator cards from Accelerated Encryption 506 Processing, www.aep.ie. (Use engine 'aep') 507 The support was copied from 0.9.6c [engine] and adapted/corrected 508 to work with the new engine framework. 509 [AEP Inc. and Richard Levitte] 510 511 *) Support for SureWare crypto accelerator cards from Baltimore 512 Technologies. (Use engine 'sureware') 513 The support was copied from 0.9.6c [engine] and adapted 514 to work with the new engine framework. 515 [Richard Levitte] 516 517 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 518 make the newer ENGINE framework commands for the CHIL engine work. 519 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 520 521 *) Make it possible to produce shared libraries on ReliantUNIX. 522 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 523 524 *) Add the configuration target debug-linux-ppro. 525 Make 'openssl rsa' use the general key loading routines 526 implemented in apps.c, and make those routines able to 527 handle the key format FORMAT_NETSCAPE and the variant 528 FORMAT_IISSGC. 529 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 530 531 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 532 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 533 534 *) Add -keyform to rsautl, and document -engine. 535 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 536 537 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 538 BIO_R_NO_SUCH_FILE error code rather than the generic 539 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 540 [Ben Laurie] 541 542 *) Add new functions 543 ERR_peek_last_error 544 ERR_peek_last_error_line 545 ERR_peek_last_error_line_data. 546 These are similar to 547 ERR_peek_error 548 ERR_peek_error_line 549 ERR_peek_error_line_data, 550 but report on the latest error recorded rather than the first one 551 still in the error queue. 552 [Ben Laurie, Bodo Moeller] 553 554 *) default_algorithms option in ENGINE config module. This allows things 555 like: 556 default_algorithms = ALL 557 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 558 [Steve Henson] 559 560 *) Prelminary ENGINE config module. 561 [Steve Henson] 562 563 *) New experimental application configuration code. 564 [Steve Henson] 565 566 *) Change the AES code to follow the same name structure as all other 567 symmetric ciphers, and behave the same way. Move everything to 568 the directory crypto/aes, thereby obsoleting crypto/rijndael. 569 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 570 571 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 572 [Ben Laurie and Theo de Raadt] 573 574 *) Add option to output public keys in req command. 575 [Massimiliano Pala madwolf@openca.org] 576 577 *) Use wNAFs in EC_POINTs_mul() for improved efficiency 578 (up to about 10% better than before for P-192 and P-224). 579 [Bodo Moeller] 580 581 *) New functions/macros 582 583 SSL_CTX_set_msg_callback(ctx, cb) 584 SSL_CTX_set_msg_callback_arg(ctx, arg) 585 SSL_set_msg_callback(ssl, cb) 586 SSL_set_msg_callback_arg(ssl, arg) 587 588 to request calling a callback function 589 590 void cb(int write_p, int version, int content_type, 591 const void *buf, size_t len, SSL *ssl, void *arg) 592 593 whenever a protocol message has been completely received 594 (write_p == 0) or sent (write_p == 1). Here 'version' is the 595 protocol version according to which the SSL library interprets 596 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 597 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 598 the content type as defined in the SSL 3.0/TLS 1.0 protocol 599 specification (change_cipher_spec(20), alert(21), handshake(22)). 600 'buf' and 'len' point to the actual message, 'ssl' to the 601 SSL object, and 'arg' is the application-defined value set by 602 SSL[_CTX]_set_msg_callback_arg(). 603 604 'openssl s_client' and 'openssl s_server' have new '-msg' options 605 to enable a callback that displays all protocol messages. 606 [Bodo Moeller] 607 608 *) Change the shared library support so shared libraries are built as 609 soon as the corresponding static library is finished, and thereby get 610 openssl and the test programs linked against the shared library. 611 This still only happens when the keyword "shard" has been given to 612 the configuration scripts. 613 614 NOTE: shared library support is still an experimental thing, and 615 backward binary compatibility is still not guaranteed. 616 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 617 618 *) Add support for Subject Information Access extension. 619 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 620 621 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 622 additional bytes when new memory had to be allocated, not just 623 when reusing an existing buffer. 624 [Bodo Moeller] 625 626 *) New command line and configuration option 'utf8' for the req command. 627 This allows field values to be specified as UTF8 strings. 628 [Steve Henson] 629 630 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 631 runs for the former and machine-readable output for the latter. 632 [Ben Laurie] 633 634 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 635 of the e-mail address in the DN (i.e., it will go into a certificate 636 extension only). The new configuration file option 'email_in_dn = no' 637 has the same effect. 638 [Massimiliano Pala madwolf@openca.org] 639 640 *) Change all functions with names starting with des_ to be starting 641 with DES_ instead. Add wrappers that are compatible with libdes, 642 but are named _ossl_old_des_*. Finally, add macros that map the 643 des_* symbols to the corresponding _ossl_old_des_* if libdes 644 compatibility is desired. If OpenSSL 0.9.6c compatibility is 645 desired, the des_* symbols will be mapped to DES_*, with one 646 exception. 647 648 Since we provide two compatibility mappings, the user needs to 649 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 650 compatibility is desired. The default (i.e., when that macro 651 isn't defined) is OpenSSL 0.9.6c compatibility. 652 653 There are also macros that enable and disable the support of old 654 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 655 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 656 are defined, the default will apply: to support the old des routines. 657 658 In either case, one must include openssl/des.h to get the correct 659 definitions. Do not try to just include openssl/des_old.h, that 660 won't work. 661 662 NOTE: This is a major break of an old API into a new one. Software 663 authors are encouraged to switch to the DES_ style functions. Some 664 time in the future, des_old.h and the libdes compatibility functions 665 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 666 default), and then completely removed. 667 [Richard Levitte] 668 669 *) Test for certificates which contain unsupported critical extensions. 670 If such a certificate is found during a verify operation it is 671 rejected by default: this behaviour can be overridden by either 672 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 673 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 674 X509_supported_extension() has also been added which returns 1 if a 675 particular extension is supported. 676 [Steve Henson] 677 678 *) Modify the behaviour of EVP cipher functions in similar way to digests 679 to retain compatibility with existing code. 680 [Steve Henson] 681 682 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 683 compatibility with existing code. In particular the 'ctx' parameter does 684 not have to be to be initialized before the call to EVP_DigestInit() and 685 it is tidied up after a call to EVP_DigestFinal(). New function 686 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 687 EVP_MD_CTX_copy() changed to not require the destination to be 688 initialized valid and new function EVP_MD_CTX_copy_ex() added which 689 requires the destination to be valid. 690 691 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 692 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 693 [Steve Henson] 694 695 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 696 so that complete 'Handshake' protocol structures are kept in memory 697 instead of overwriting 'msg_type' and 'length' with 'body' data. 698 [Bodo Moeller] 699 700 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 701 [Massimo Santin via Richard Levitte] 702 703 *) Major restructuring to the underlying ENGINE code. This includes 704 reduction of linker bloat, separation of pure "ENGINE" manipulation 705 (initialisation, etc) from functionality dealing with implementations 706 of specific crypto iterfaces. This change also introduces integrated 707 support for symmetric ciphers and digest implementations - so ENGINEs 708 can now accelerate these by providing EVP_CIPHER and EVP_MD 709 implementations of their own. This is detailed in crypto/engine/README 710 as it couldn't be adequately described here. However, there are a few 711 API changes worth noting - some RSA, DSA, DH, and RAND functions that 712 were changed in the original introduction of ENGINE code have now 713 reverted back - the hooking from this code to ENGINE is now a good 714 deal more passive and at run-time, operations deal directly with 715 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 716 dereferencing through an ENGINE pointer any more. Also, the ENGINE 717 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 718 they were not being used by the framework as there is no concept of a 719 BIGNUM_METHOD and they could not be generalised to the new 720 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 721 ENGINE_cpy() has been removed as it cannot be consistently defined in 722 the new code. 723 [Geoff Thorpe] 724 725 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 726 [Steve Henson] 727 728 *) Change mkdef.pl to sort symbols that get the same entry number, 729 and make sure the automatically generated functions ERR_load_* 730 become part of libeay.num as well. 731 [Richard Levitte] 732 733 *) New function SSL_renegotiate_pending(). This returns true once 734 renegotiation has been requested (either SSL_renegotiate() call 735 or HelloRequest/ClientHello receveived from the peer) and becomes 736 false once a handshake has been completed. 737 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 738 sends a HelloRequest, but does not ensure that a handshake takes 739 place. SSL_renegotiate_pending() is useful for checking if the 740 client has followed the request.) 741 [Bodo Moeller] 742 743 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 744 By default, clients may request session resumption even during 745 renegotiation (if session ID contexts permit); with this option, 746 session resumption is possible only in the first handshake. 747 748 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 749 more bits available for options that should not be part of 750 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 751 [Bodo Moeller] 752 753 *) Add some demos for certificate and certificate request creation. 754 [Steve Henson] 755 756 *) Make maximum certificate chain size accepted from the peer application 757 settable (SSL*_get/set_max_cert_list()), as proposed by 758 "Douglas E. Engert" <deengert@anl.gov>. 759 [Lutz Jaenicke] 760 761 *) Add support for shared libraries for Unixware-7 762 (Boyd Lynn Gerber <gerberb@zenez.com>). 763 [Lutz Jaenicke] 764 765 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 766 be done prior to destruction. Use this to unload error strings from 767 ENGINEs that load their own error strings. NB: This adds two new API 768 functions to "get" and "set" this destroy handler in an ENGINE. 769 [Geoff Thorpe] 770 771 *) Alter all existing ENGINE implementations (except "openssl" and 772 "openbsd") to dynamically instantiate their own error strings. This 773 makes them more flexible to be built both as statically-linked ENGINEs 774 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 775 Also, add stub code to each that makes building them as self-contained 776 shared-libraries easier (see README.ENGINE). 777 [Geoff Thorpe] 778 779 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 780 implementations into applications that are completely implemented in 781 self-contained shared-libraries. The "dynamic" ENGINE exposes control 782 commands that can be used to configure what shared-library to load and 783 to control aspects of the way it is handled. Also, made an update to 784 the README.ENGINE file that brings its information up-to-date and 785 provides some information and instructions on the "dynamic" ENGINE 786 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 787 [Geoff Thorpe] 788 789 *) Make it possible to unload ranges of ERR strings with a new 790 "ERR_unload_strings" function. 791 [Geoff Thorpe] 792 793 *) Add a copy() function to EVP_MD. 794 [Ben Laurie] 795 796 *) Make EVP_MD routines take a context pointer instead of just the 797 md_data void pointer. 798 [Ben Laurie] 799 800 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 801 that the digest can only process a single chunk of data 802 (typically because it is provided by a piece of 803 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 804 is only going to provide a single chunk of data, and hence the 805 framework needn't accumulate the data for oneshot drivers. 806 [Ben Laurie] 807 808 *) As with "ERR", make it possible to replace the underlying "ex_data" 809 functions. This change also alters the storage and management of global 810 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 811 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 812 index counters. The API functions that use this state have been changed 813 to take a "class_index" rather than pointers to the class's local STACK 814 and counter, and there is now an API function to dynamically create new 815 classes. This centralisation allows us to (a) plug a lot of the 816 thread-safety problems that existed, and (b) makes it possible to clean 817 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 818 such data would previously have always leaked in application code and 819 workarounds were in place to make the memory debugging turn a blind eye 820 to it. Application code that doesn't use this new function will still 821 leak as before, but their memory debugging output will announce it now 822 rather than letting it slide. 823 824 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 825 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 826 has a return value to indicate success or failure. 827 [Geoff Thorpe] 828 829 *) Make it possible to replace the underlying "ERR" functions such that the 830 global state (2 LHASH tables and 2 locks) is only used by the "default" 831 implementation. This change also adds two functions to "get" and "set" 832 the implementation prior to it being automatically set the first time 833 any other ERR function takes place. Ie. an application can call "get", 834 pass the return value to a module it has just loaded, and that module 835 can call its own "set" function using that value. This means the 836 module's "ERR" operations will use (and modify) the error state in the 837 application and not in its own statically linked copy of OpenSSL code. 838 [Geoff Thorpe] 839 840 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 841 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 842 the operation, and provides a more encapsulated way for external code 843 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 844 to use these functions rather than manually incrementing the counts. 845 846 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 847 [Geoff Thorpe] 848 849 *) Add EVP test program. 850 [Ben Laurie] 851 852 *) Add symmetric cipher support to ENGINE. Expect the API to change! 853 [Ben Laurie] 854 855 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 856 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 857 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 858 These allow a CRL to be built without having to access X509_CRL fields 859 directly. Modify 'ca' application to use new functions. 860 [Steve Henson] 861 862 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 863 bug workarounds. Rollback attack detection is a security feature. 864 The problem will only arise on OpenSSL servers when TLSv1 is not 865 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 866 Software authors not wanting to support TLSv1 will have special reasons 867 for their choice and can explicitly enable this option. 868 [Bodo Moeller, Lutz Jaenicke] 869 870 *) Rationalise EVP so it can be extended: don't include a union of 871 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 872 (similar to those existing for EVP_CIPHER_CTX). 873 Usage example: 874 875 EVP_MD_CTX md; 876 877 EVP_MD_CTX_init(&md); /* new function call */ 878 EVP_DigestInit(&md, EVP_sha1()); 879 EVP_DigestUpdate(&md, in, len); 880 EVP_DigestFinal(&md, out, NULL); 881 EVP_MD_CTX_cleanup(&md); /* new function call */ 882 883 [Ben Laurie] 884 885 *) Make DES key schedule conform to the usual scheme, as well as 886 correcting its structure. This means that calls to DES functions 887 now have to pass a pointer to a des_key_schedule instead of a 888 plain des_key_schedule (which was actually always a pointer 889 anyway): E.g., 890 891 des_key_schedule ks; 892 893 des_set_key_checked(..., &ks); 894 des_ncbc_encrypt(..., &ks, ...); 895 896 (Note that a later change renames 'des_...' into 'DES_...'.) 897 [Ben Laurie] 898 899 *) Initial reduction of linker bloat: the use of some functions, such as 900 PEM causes large amounts of unused functions to be linked in due to 901 poor organisation. For example pem_all.c contains every PEM function 902 which has a knock on effect of linking in large amounts of (unused) 903 ASN1 code. Grouping together similar functions and splitting unrelated 904 functions prevents this. 905 [Steve Henson] 906 907 *) Cleanup of EVP macros. 908 [Ben Laurie] 909 910 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 911 correct _ecb suffix. 912 [Ben Laurie] 913 914 *) Add initial OCSP responder support to ocsp application. The 915 revocation information is handled using the text based index 916 use by the ca application. The responder can either handle 917 requests generated internally, supplied in files (for example 918 via a CGI script) or using an internal minimal server. 919 [Steve Henson] 920 921 *) Add configuration choices to get zlib compression for TLS. 922 [Richard Levitte] 923 924 *) Changes to Kerberos SSL for RFC 2712 compliance: 925 1. Implemented real KerberosWrapper, instead of just using 926 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 927 2. Implemented optional authenticator field of KerberosWrapper. 928 929 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 930 and authenticator structs; see crypto/krb5/. 931 932 Generalized Kerberos calls to support multiple Kerberos libraries. 933 [Vern Staats <staatsvr@asc.hpc.mil>, 934 Jeffrey Altman <jaltman@columbia.edu> 935 via Richard Levitte] 936 937 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 938 already does with RSA. testdsa.h now has 'priv_key/pub_key' 939 values for each of the key sizes rather than having just 940 parameters (and 'speed' generating keys each time). 941 [Geoff Thorpe] 942 943 *) Speed up EVP routines. 944 Before: 945encrypt 946type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 947des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 948des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 949des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 950decrypt 951des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 952des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 953des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 954 After: 955encrypt 956des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 957decrypt 958des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 959 [Ben Laurie] 960 961 *) Added the OS2-EMX target. 962 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 963 964 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 965 to support NCONF routines in extension code. New function CONF_set_nconf() 966 to allow functions which take an NCONF to also handle the old LHASH 967 structure: this means that the old CONF compatible routines can be 968 retained (in particular wrt extensions) without having to duplicate the 969 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 970 [Steve Henson] 971 972 *) Enhance the general user interface with mechanisms for inner control 973 and with possibilities to have yes/no kind of prompts. 974 [Richard Levitte] 975 976 *) Change all calls to low level digest routines in the library and 977 applications to use EVP. Add missing calls to HMAC_cleanup() and 978 don't assume HMAC_CTX can be copied using memcpy(). 979 [Verdon Walker <VWalker@novell.com>, Steve Henson] 980 981 *) Add the possibility to control engines through control names but with 982 arbitrary arguments instead of just a string. 983 Change the key loaders to take a UI_METHOD instead of a callback 984 function pointer. NOTE: this breaks binary compatibility with earlier 985 versions of OpenSSL [engine]. 986 Adapt the nCipher code for these new conditions and add a card insertion 987 callback. 988 [Richard Levitte] 989 990 *) Enhance the general user interface with mechanisms to better support 991 dialog box interfaces, application-defined prompts, the possibility 992 to use defaults (for example default passwords from somewhere else) 993 and interrupts/cancellations. 994 [Richard Levitte] 995 996 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 997 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 998 [Steve Henson] 999 1000 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 1001 tidy up some unnecessarily weird code in 'sk_new()'). 1002 [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 1003 1004 *) Change the key loading routines for ENGINEs to use the same kind 1005 callback (pem_password_cb) as all other routines that need this 1006 kind of callback. 1007 [Richard Levitte] 1008 1009 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 1010 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 1011 than this minimum value is recommended. 1012 [Lutz Jaenicke] 1013 1014 *) New random seeder for OpenVMS, using the system process statistics 1015 that are easily reachable. 1016 [Richard Levitte] 1017 1018 *) Windows apparently can't transparently handle global 1019 variables defined in DLLs. Initialisations such as: 1020 1021 const ASN1_ITEM *it = &ASN1_INTEGER_it; 1022 1023 wont compile. This is used by the any applications that need to 1024 declare their own ASN1 modules. This was fixed by adding the option 1025 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 1026 needed for static libraries under Win32. 1027 [Steve Henson] 1028 1029 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 1030 setting of purpose and trust fields. New X509_STORE trust and 1031 purpose functions and tidy up setting in other SSL functions. 1032 [Steve Henson] 1033 1034 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 1035 structure. These are inherited by X509_STORE_CTX when it is 1036 initialised. This allows various defaults to be set in the 1037 X509_STORE structure (such as flags for CRL checking and custom 1038 purpose or trust settings) for functions which only use X509_STORE_CTX 1039 internally such as S/MIME. 1040 1041 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 1042 trust settings if they are not set in X509_STORE. This allows X509_STORE 1043 purposes and trust (in S/MIME for example) to override any set by default. 1044 1045 Add command line options for CRL checking to smime, s_client and s_server 1046 applications. 1047 [Steve Henson] 1048 1049 *) Initial CRL based revocation checking. If the CRL checking flag(s) 1050 are set then the CRL is looked up in the X509_STORE structure and 1051 its validity and signature checked, then if the certificate is found 1052 in the CRL the verify fails with a revoked error. 1053 1054 Various new CRL related callbacks added to X509_STORE_CTX structure. 1055 1056 Command line options added to 'verify' application to support this. 1057 1058 This needs some additional work, such as being able to handle multiple 1059 CRLs with different times, extension based lookup (rather than just 1060 by subject name) and ultimately more complete V2 CRL extension 1061 handling. 1062 [Steve Henson] 1063 1064 *) Add a general user interface API (crypto/ui/). This is designed 1065 to replace things like des_read_password and friends (backward 1066 compatibility functions using this new API are provided). 1067 The purpose is to remove prompting functions from the DES code 1068 section as well as provide for prompting through dialog boxes in 1069 a window system and the like. 1070 [Richard Levitte] 1071 1072 *) Add "ex_data" support to ENGINE so implementations can add state at a 1073 per-structure level rather than having to store it globally. 1074 [Geoff] 1075 1076 *) Make it possible for ENGINE structures to be copied when retrieved by 1077 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 1078 This causes the "original" ENGINE structure to act like a template, 1079 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 1080 operational state can be localised to each ENGINE structure, despite the 1081 fact they all share the same "methods". New ENGINE structures returned in 1082 this case have no functional references and the return value is the single 1083 structural reference. This matches the single structural reference returned 1084 by ENGINE_by_id() normally, when it is incremented on the pre-existing 1085 ENGINE structure. 1086 [Geoff] 1087 1088 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 1089 needs to match any other type at all we need to manually clear the 1090 tag cache. 1091 [Steve Henson] 1092 1093 *) Changes to the "openssl engine" utility to include; 1094 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 1095 about an ENGINE's available control commands. 1096 - executing control commands from command line arguments using the 1097 '-pre' and '-post' switches. '-post' is only used if '-t' is 1098 specified and the ENGINE is successfully initialised. The syntax for 1099 the individual commands are colon-separated, for example; 1100 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 1101 [Geoff] 1102 1103 *) New dynamic control command support for ENGINEs. ENGINEs can now 1104 declare their own commands (numbers), names (strings), descriptions, 1105 and input types for run-time discovery by calling applications. A 1106 subset of these commands are implicitly classed as "executable" 1107 depending on their input type, and only these can be invoked through 1108 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 1109 can be based on user input, config files, etc). The distinction is 1110 that "executable" commands cannot return anything other than a boolean 1111 result and can only support numeric or string input, whereas some 1112 discoverable commands may only be for direct use through 1113 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 1114 pointers, or other custom uses. The "executable" commands are to 1115 support parameterisations of ENGINE behaviour that can be 1116 unambiguously defined by ENGINEs and used consistently across any 1117 OpenSSL-based application. Commands have been added to all the 1118 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 1119 control over shared-library paths without source code alterations. 1120 [Geoff] 1121 1122 *) Changed all ENGINE implementations to dynamically allocate their 1123 ENGINEs rather than declaring them statically. Apart from this being 1124 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 1125 this also allows the implementations to compile without using the 1126 internal engine_int.h header. 1127 [Geoff] 1128 1129 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 1130 'const' value. Any code that should be able to modify a RAND_METHOD 1131 should already have non-const pointers to it (ie. they should only 1132 modify their own ones). 1133 [Geoff] 1134 1135 *) Made a variety of little tweaks to the ENGINE code. 1136 - "atalla" and "ubsec" string definitions were moved from header files 1137 to C code. "nuron" string definitions were placed in variables 1138 rather than hard-coded - allowing parameterisation of these values 1139 later on via ctrl() commands. 1140 - Removed unused "#if 0"'d code. 1141 - Fixed engine list iteration code so it uses ENGINE_free() to release 1142 structural references. 1143 - Constified the RAND_METHOD element of ENGINE structures. 1144 - Constified various get/set functions as appropriate and added 1145 missing functions (including a catch-all ENGINE_cpy that duplicates 1146 all ENGINE values onto a new ENGINE except reference counts/state). 1147 - Removed NULL parameter checks in get/set functions. Setting a method 1148 or function to NULL is a way of cancelling out a previously set 1149 value. Passing a NULL ENGINE parameter is just plain stupid anyway 1150 and doesn't justify the extra error symbols and code. 1151 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 1152 flags from engine_int.h to engine.h. 1153 - Changed prototypes for ENGINE handler functions (init(), finish(), 1154 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 1155 [Geoff] 1156 1157 *) Implement binary inversion algorithm for BN_mod_inverse in addition 1158 to the algorithm using long division. The binary algorithm can be 1159 used only if the modulus is odd. On 32-bit systems, it is faster 1160 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 1161 roughly 5-15% for 256-bit moduli), so we use it only for moduli 1162 up to 450 bits. In 64-bit environments, the binary algorithm 1163 appears to be advantageous for much longer moduli; here we use it 1164 for moduli up to 2048 bits. 1165 [Bodo Moeller] 1166 1167 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 1168 could not support the combine flag in choice fields. 1169 [Steve Henson] 1170 1171 *) Add a 'copy_extensions' option to the 'ca' utility. This copies 1172 extensions from a certificate request to the certificate. 1173 [Steve Henson] 1174 1175 *) Allow multiple 'certopt' and 'nameopt' options to be separated 1176 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 1177 file: this allows the display of the certificate about to be 1178 signed to be customised, to allow certain fields to be included 1179 or excluded and extension details. The old system didn't display 1180 multicharacter strings properly, omitted fields not in the policy 1181 and couldn't display additional details such as extensions. 1182 [Steve Henson] 1183 1184 *) Function EC_POINTs_mul for multiple scalar multiplication 1185 of an arbitrary number of elliptic curve points 1186 \sum scalars[i]*points[i], 1187 optionally including the generator defined for the EC_GROUP: 1188 scalar*generator + \sum scalars[i]*points[i]. 1189 1190 EC_POINT_mul is a simple wrapper function for the typical case 1191 that the point list has just one item (besides the optional 1192 generator). 1193 [Bodo Moeller] 1194 1195 *) First EC_METHODs for curves over GF(p): 1196 1197 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 1198 operations and provides various method functions that can also 1199 operate with faster implementations of modular arithmetic. 1200 1201 EC_GFp_mont_method() reuses most functions that are part of 1202 EC_GFp_simple_method, but uses Montgomery arithmetic. 1203 1204 [Bodo Moeller; point addition and point doubling 1205 implementation directly derived from source code provided by 1206 Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 1207 1208 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 1209 crypto/ec/ec_lib.c): 1210 1211 Curves are EC_GROUP objects (with an optional group generator) 1212 based on EC_METHODs that are built into the library. 1213 1214 Points are EC_POINT objects based on EC_GROUP objects. 1215 1216 Most of the framework would be able to handle curves over arbitrary 1217 finite fields, but as there are no obvious types for fields other 1218 than GF(p), some functions are limited to that for now. 1219 [Bodo Moeller] 1220 1221 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 1222 that the file contains a complete HTTP response. 1223 [Richard Levitte] 1224 1225 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 1226 change the def and num file printf format specifier from "%-40sXXX" 1227 to "%-39s XXX". The latter will always guarantee a space after the 1228 field while the former will cause them to run together if the field 1229 is 40 of more characters long. 1230 [Steve Henson] 1231 1232 *) Constify the cipher and digest 'method' functions and structures 1233 and modify related functions to take constant EVP_MD and EVP_CIPHER 1234 pointers. 1235 [Steve Henson] 1236 1237 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 1238 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 1239 [Bodo Moeller] 1240 1241 *) Modify EVP_Digest*() routines so they now return values. Although the 1242 internal software routines can never fail additional hardware versions 1243 might. 1244 [Steve Henson] 1245 1246 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 1247 1248 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 1249 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 1250 1251 ASN1 error codes 1252 ERR_R_NESTED_ASN1_ERROR 1253 ... 1254 ERR_R_MISSING_ASN1_EOS 1255 were 4 .. 9, conflicting with 1256 ERR_LIB_RSA (= ERR_R_RSA_LIB) 1257 ... 1258 ERR_LIB_PEM (= ERR_R_PEM_LIB). 1259 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 1260 1261 Add new error code 'ERR_R_INTERNAL_ERROR'. 1262 [Bodo Moeller] 1263 1264 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 1265 suffices. 1266 [Bodo Moeller] 1267 1268 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 1269 sets the subject name for a new request or supersedes the 1270 subject name in a given request. Formats that can be parsed are 1271 'CN=Some Name, OU=myOU, C=IT' 1272 and 1273 'CN=Some Name/OU=myOU/C=IT'. 1274 1275 Add options '-batch' and '-verbose' to 'openssl req'. 1276 [Massimiliano Pala <madwolf@hackmasters.net>] 1277 1278 *) Introduce the possibility to access global variables through 1279 functions on platform were that's the best way to handle exporting 1280 global variables in shared libraries. To enable this functionality, 1281 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 1282 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 1283 is normally done by Configure or something similar). 1284 1285 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 1286 in the source file (foo.c) like this: 1287 1288 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 1289 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 1290 1291 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 1292 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 1293 1294 OPENSSL_DECLARE_GLOBAL(int,foo); 1295 #define foo OPENSSL_GLOBAL_REF(foo) 1296 OPENSSL_DECLARE_GLOBAL(double,bar); 1297 #define bar OPENSSL_GLOBAL_REF(bar) 1298 1299 The #defines are very important, and therefore so is including the 1300 header file everywhere where the defined globals are used. 1301 1302 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 1303 of ASN.1 items, but that structure is a bit different. 1304 1305 The largest change is in util/mkdef.pl which has been enhanced with 1306 better and easier to understand logic to choose which symbols should 1307 go into the Windows .def files as well as a number of fixes and code 1308 cleanup (among others, algorithm keywords are now sorted 1309 lexicographically to avoid constant rewrites). 1310 [Richard Levitte] 1311 1312 *) In BN_div() keep a copy of the sign of 'num' before writing the 1313 result to 'rm' because if rm==num the value will be overwritten 1314 and produce the wrong result if 'num' is negative: this caused 1315 problems with BN_mod() and BN_nnmod(). 1316 [Steve Henson] 1317 1318 *) Function OCSP_request_verify(). This checks the signature on an 1319 OCSP request and verifies the signer certificate. The signer 1320 certificate is just checked for a generic purpose and OCSP request 1321 trust settings. 1322 [Steve Henson] 1323 1324 *) Add OCSP_check_validity() function to check the validity of OCSP 1325 responses. OCSP responses are prepared in real time and may only 1326 be a few seconds old. Simply checking that the current time lies 1327 between thisUpdate and nextUpdate max reject otherwise valid responses 1328 caused by either OCSP responder or client clock inaccuracy. Instead 1329 we allow thisUpdate and nextUpdate to fall within a certain period of 1330 the current time. The age of the response can also optionally be 1331 checked. Two new options -validity_period and -status_age added to 1332 ocsp utility. 1333 [Steve Henson] 1334 1335 *) If signature or public key algorithm is unrecognized print out its 1336 OID rather that just UNKNOWN. 1337 [Steve Henson] 1338 1339 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 1340 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 1341 ID to be generated from the issuer certificate alone which can then be 1342 passed to OCSP_id_issuer_cmp(). 1343 [Steve Henson] 1344 1345 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 1346 ASN1 modules to export functions returning ASN1_ITEM pointers 1347 instead of the ASN1_ITEM structures themselves. This adds several 1348 new macros which allow the underlying ASN1 function/structure to 1349 be accessed transparently. As a result code should not use ASN1_ITEM 1350 references directly (such as &X509_it) but instead use the relevant 1351 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 1352 use of the new ASN1 code on platforms where exporting structures 1353 is problematical (for example in shared libraries) but exporting 1354 functions returning pointers to structures is not. 1355 [Steve Henson] 1356 1357 *) Add support for overriding the generation of SSL/TLS session IDs. 1358 These callbacks can be registered either in an SSL_CTX or per SSL. 1359 The purpose of this is to allow applications to control, if they wish, 1360 the arbitrary values chosen for use as session IDs, particularly as it 1361 can be useful for session caching in multiple-server environments. A 1362 command-line switch for testing this (and any client code that wishes 1363 to use such a feature) has been added to "s_server". 1364 [Geoff Thorpe, Lutz Jaenicke] 1365 1366 *) Modify mkdef.pl to recognise and parse preprocessor conditionals 1367 of the form '#if defined(...) || defined(...) || ...' and 1368 '#if !defined(...) && !defined(...) && ...'. This also avoids 1369 the growing number of special cases it was previously handling. 1370 [Richard Levitte] 1371 1372 *) Make all configuration macros available for application by making 1373 sure they are available in opensslconf.h, by giving them names starting 1374 with "OPENSSL_" to avoid conflicts with other packages and by making 1375 sure e_os2.h will cover all platform-specific cases together with 1376 opensslconf.h. 1377 Additionally, it is now possible to define configuration/platform- 1378 specific names (called "system identities"). In the C code, these 1379 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 1380 macro with the name beginning with "OPENSSL_SYS_", which is determined 1381 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 1382 what is available. 1383 [Richard Levitte] 1384 1385 *) New option -set_serial to 'req' and 'x509' this allows the serial 1386 number to use to be specified on the command line. Previously self 1387 signed certificates were hard coded with serial number 0 and the 1388 CA options of 'x509' had to use a serial number in a file which was 1389 auto incremented. 1390 [Steve Henson] 1391 1392 *) New options to 'ca' utility to support V2 CRL entry extensions. 1393 Currently CRL reason, invalidity date and hold instruction are 1394 supported. Add new CRL extensions to V3 code and some new objects. 1395 [Steve Henson] 1396 1397 *) New function EVP_CIPHER_CTX_set_padding() this is used to 1398 disable standard block padding (aka PKCS#5 padding) in the EVP 1399 API, which was previously mandatory. This means that the data is 1400 not padded in any way and so the total length much be a multiple 1401 of the block size, otherwise an error occurs. 1402 [Steve Henson] 1403 1404 *) Initial (incomplete) OCSP SSL support. 1405 [Steve Henson] 1406 1407 *) New function OCSP_parse_url(). This splits up a URL into its host, 1408 port and path components: primarily to parse OCSP URLs. New -url 1409 option to ocsp utility. 1410 [Steve Henson] 1411 1412 *) New nonce behavior. The return value of OCSP_check_nonce() now 1413 reflects the various checks performed. Applications can decide 1414 whether to tolerate certain situations such as an absent nonce 1415 in a response when one was present in a request: the ocsp application 1416 just prints out a warning. New function OCSP_add1_basic_nonce() 1417 this is to allow responders to include a nonce in a response even if 1418 the request is nonce-less. 1419 [Steve Henson] 1420 1421 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 1422 skipped when using openssl x509 multiple times on a single input file, 1423 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 1424 [Bodo Moeller] 1425 1426 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 1427 set string type: to handle setting ASN1_TIME structures. Fix ca 1428 utility to correctly initialize revocation date of CRLs. 1429 [Steve Henson] 1430 1431 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 1432 the clients preferred ciphersuites and rather use its own preferences. 1433 Should help to work around M$ SGC (Server Gated Cryptography) bug in 1434 Internet Explorer by ensuring unchanged hash method during stepup. 1435 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 1436 [Lutz Jaenicke] 1437 1438 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 1439 to aes and add a new 'exist' option to print out symbols that don't 1440 appear to exist. 1441 [Steve Henson] 1442 1443 *) Additional options to ocsp utility to allow flags to be set and 1444 additional certificates supplied. 1445 [Steve Henson] 1446 1447 *) Add the option -VAfile to 'openssl ocsp', so the user can give the 1448 OCSP client a number of certificate to only verify the response 1449 signature against. 1450 [Richard Levitte] 1451 1452 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 1453 handle the new API. Currently only ECB, CBC modes supported. Add new 1454 AES OIDs. 1455 1456 Add TLS AES ciphersuites as described in RFC3268, "Advanced 1457 Encryption Standard (AES) Ciphersuites for Transport Layer 1458 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 1459 not enabled by default and were not part of the "ALL" ciphersuite 1460 alias because they were not yet official; they could be 1461 explicitly requested by specifying the "AESdraft" ciphersuite 1462 group alias. In the final release of OpenSSL 0.9.7, the group 1463 alias is called "AES" and is part of "ALL".) 1464 [Ben Laurie, Steve Henson, Bodo Moeller] 1465 1466 *) New function OCSP_copy_nonce() to copy nonce value (if present) from 1467 request to response. 1468 [Steve Henson] 1469 1470 *) Functions for OCSP responders. OCSP_request_onereq_count(), 1471 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 1472 extract information from a certificate request. OCSP_response_create() 1473 creates a response and optionally adds a basic response structure. 1474 OCSP_basic_add1_status() adds a complete single response to a basic 1475 response and returns the OCSP_SINGLERESP structure just added (to allow 1476 extensions to be included for example). OCSP_basic_add1_cert() adds a 1477 certificate to a basic response and OCSP_basic_sign() signs a basic 1478 response with various flags. New helper functions ASN1_TIME_check() 1479 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 1480 (converts ASN1_TIME to GeneralizedTime). 1481 [Steve Henson] 1482 1483 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 1484 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 1485 structure from a certificate. X509_pubkey_digest() digests the public_key 1486 contents: this is used in various key identifiers. 1487 [Steve Henson] 1488 1489 *) Make sk_sort() tolerate a NULL argument. 1490 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 1491 1492 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 1493 passed by the function are trusted implicitly. If any of them signed the 1494 response then it is assumed to be valid and is not verified. 1495 [Steve Henson] 1496 1497 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 1498 to data. This was previously part of the PKCS7 ASN1 code. This 1499 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 1500 [Steve Henson, reported by Kenneth R. Robinette 1501 <support@securenetterm.com>] 1502 1503 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 1504 routines: without these tracing memory leaks is very painful. 1505 Fix leaks in PKCS12 and PKCS7 routines. 1506 [Steve Henson] 1507 1508 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 1509 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 1510 effectively meant GeneralizedTime would never be used. Now it 1511 is initialised to -1 but X509_time_adj() now has to check the value 1512 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 1513 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 1514 [Steve Henson, reported by Kenneth R. Robinette 1515 <support@securenetterm.com>] 1516 1517 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 1518 result in a zero length in the ASN1_INTEGER structure which was 1519 not consistent with the structure when d2i_ASN1_INTEGER() was used 1520 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 1521 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 1522 where it did not print out a minus for negative ASN1_INTEGER. 1523 [Steve Henson] 1524 1525 *) Add summary printout to ocsp utility. The various functions which 1526 convert status values to strings have been renamed to: 1527 OCSP_response_status_str(), OCSP_cert_status_str() and 1528 OCSP_crl_reason_str() and are no longer static. New options 1529 to verify nonce values and to disable verification. OCSP response 1530 printout format cleaned up. 1531 [Steve Henson] 1532 1533 *) Add additional OCSP certificate checks. These are those specified 1534 in RFC2560. This consists of two separate checks: the CA of the 1535 certificate being checked must either be the OCSP signer certificate 1536 or the issuer of the OCSP signer certificate. In the latter case the 1537 OCSP signer certificate must contain the OCSP signing extended key 1538 usage. This check is performed by attempting to match the OCSP 1539 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 1540 in the OCSP_CERTID structures of the response. 1541 [Steve Henson] 1542 1543 *) Initial OCSP certificate verification added to OCSP_basic_verify() 1544 and related routines. This uses the standard OpenSSL certificate 1545 verify routines to perform initial checks (just CA validity) and 1546 to obtain the certificate chain. Then additional checks will be 1547 performed on the chain. Currently the root CA is checked to see 1548 if it is explicitly trusted for OCSP signing. This is used to set 1549 a root CA as a global signing root: that is any certificate that 1550 chains to that CA is an acceptable OCSP signing certificate. 1551 [Steve Henson] 1552 1553 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 1554 extensions from a separate configuration file. 1555 As when reading extensions from the main configuration file, 1556 the '-extensions ...' option may be used for specifying the 1557 section to use. 1558 [Massimiliano Pala <madwolf@comune.modena.it>] 1559 1560 *) New OCSP utility. Allows OCSP requests to be generated or 1561 read. The request can be sent to a responder and the output 1562 parsed, outputed or printed in text form. Not complete yet: 1563 still needs to check the OCSP response validity. 1564 [Steve Henson] 1565 1566 *) New subcommands for 'openssl ca': 1567 'openssl ca -status <serial>' prints the status of the cert with 1568 the given serial number (according to the index file). 1569 'openssl ca -updatedb' updates the expiry status of certificates 1570 in the index file. 1571 [Massimiliano Pala <madwolf@comune.modena.it>] 1572 1573 *) New '-newreq-nodes' command option to CA.pl. This is like 1574 '-newreq', but calls 'openssl req' with the '-nodes' option 1575 so that the resulting key is not encrypted. 1576 [Damien Miller <djm@mindrot.org>] 1577 1578 *) New configuration for the GNU Hurd. 1579 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 1580 1581 *) Initial code to implement OCSP basic response verify. This 1582 is currently incomplete. Currently just finds the signer's 1583 certificate and verifies the signature on the response. 1584 [Steve Henson] 1585 1586 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 1587 value of OPENSSLDIR. This is available via the new '-d' option 1588 to 'openssl version', and is also included in 'openssl version -a'. 1589 [Bodo Moeller] 1590 1591 *) Allowing defining memory allocation callbacks that will be given 1592 file name and line number information in additional arguments 1593 (a const char* and an int). The basic functionality remains, as 1594 well as the original possibility to just replace malloc(), 1595 realloc() and free() by functions that do not know about these 1596 additional arguments. To register and find out the current 1597 settings for extended allocation functions, the following 1598 functions are provided: 1599 1600 CRYPTO_set_mem_ex_functions 1601 CRYPTO_set_locked_mem_ex_functions 1602 CRYPTO_get_mem_ex_functions 1603 CRYPTO_get_locked_mem_ex_functions 1604 1605 These work the same way as CRYPTO_set_mem_functions and friends. 1606 CRYPTO_get_[locked_]mem_functions now writes 0 where such an 1607 extended allocation function is enabled. 1608 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 1609 a conventional allocation function is enabled. 1610 [Richard Levitte, Bodo Moeller] 1611 1612 *) Finish off removing the remaining LHASH function pointer casts. 1613 There should no longer be any prototype-casting required when using 1614 the LHASH abstraction, and any casts that remain are "bugs". See 1615 the callback types and macros at the head of lhash.h for details 1616 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 1617 [Geoff Thorpe] 1618 1619 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 1620 If /dev/[u]random devices are not available or do not return enough 1621 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 1622 be queried. 1623 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 1624 /etc/entropy will be queried once each in this sequence, quering stops 1625 when enough entropy was collected without querying more sockets. 1626 [Lutz Jaenicke] 1627 1628 *) Change the Unix RAND_poll() variant to be able to poll several 1629 random devices, as specified by DEVRANDOM, until a sufficient amount 1630 of data has been collected. We spend at most 10 ms on each file 1631 (select timeout) and read in non-blocking mode. DEVRANDOM now 1632 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 1633 (previously it was just the string "/dev/urandom"), so on typical 1634 platforms the 10 ms delay will never occur. 1635 Also separate out the Unix variant to its own file, rand_unix.c. 1636 For VMS, there's a currently-empty rand_vms.c. 1637 [Richard Levitte] 1638 1639 *) Move OCSP client related routines to ocsp_cl.c. These 1640 provide utility functions which an application needing 1641 to issue a request to an OCSP responder and analyse the 1642 response will typically need: as opposed to those which an 1643 OCSP responder itself would need which will be added later. 1644 1645 OCSP_request_sign() signs an OCSP request with an API similar 1646 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 1647 response. OCSP_response_get1_basic() extracts basic response 1648 from response. OCSP_resp_find_status(): finds and extracts status 1649 information from an OCSP_CERTID structure (which will be created 1650 when the request structure is built). These are built from lower 1651 level functions which work on OCSP_SINGLERESP structures but 1652 wont normally be used unless the application wishes to examine 1653 extensions in the OCSP response for example. 1654 1655 Replace nonce routines with a pair of functions. 1656 OCSP_request_add1_nonce() adds a nonce value and optionally 1657 generates a random value. OCSP_check_nonce() checks the 1658 validity of the nonce in an OCSP response. 1659 [Steve Henson] 1660 1661 *) Change function OCSP_request_add() to OCSP_request_add0_id(). 1662 This doesn't copy the supplied OCSP_CERTID and avoids the 1663 need to free up the newly created id. Change return type 1664 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 1665 This can then be used to add extensions to the request. 1666 Deleted OCSP_request_new(), since most of its functionality 1667 is now in OCSP_REQUEST_new() (and the case insensitive name 1668 clash) apart from the ability to set the request name which 1669 will be added elsewhere. 1670 [Steve Henson] 1671 1672 *) Update OCSP API. Remove obsolete extensions argument from 1673 various functions. Extensions are now handled using the new 1674 OCSP extension code. New simple OCSP HTTP function which 1675 can be used to send requests and parse the response. 1676 [Steve Henson] 1677 1678 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 1679 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 1680 uses the special reorder version of SET OF to sort the attributes 1681 and reorder them to match the encoded order. This resolves a long 1682 standing problem: a verify on a PKCS7 structure just after signing 1683 it used to fail because the attribute order did not match the 1684 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 1685 it uses the received order. This is necessary to tolerate some broken 1686 software that does not order SET OF. This is handled by encoding 1687 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 1688 to produce the required SET OF. 1689 [Steve Henson] 1690 1691 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 1692 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 1693 files to get correct declarations of the ASN.1 item variables. 1694 [Richard Levitte] 1695 1696 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 1697 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 1698 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 1699 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 1700 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 1701 ASN1_ITEM and no wrapper functions. 1702 [Steve Henson] 1703 1704 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 1705 replace the old function pointer based I/O routines. Change most of 1706 the *_d2i_bio() and *_d2i_fp() functions to use these. 1707 [Steve Henson] 1708 1709 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 1710 lines, recognice more "algorithms" that can be deselected, and make 1711 it complain about algorithm deselection that isn't recognised. 1712 [Richard Levitte] 1713 1714 *) New ASN1 functions to handle dup, sign, verify, digest, pack and 1715 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 1716 to use new functions. Add NO_ASN1_OLD which can be set to remove 1717 some old style ASN1 functions: this can be used to determine if old 1718 code will still work when these eventually go away. 1719 [Steve Henson] 1720 1721 *) New extension functions for OCSP structures, these follow the 1722 same conventions as certificates and CRLs. 1723 [Steve Henson] 1724 1725 *) New function X509V3_add1_i2d(). This automatically encodes and 1726 adds an extension. Its behaviour can be customised with various 1727 flags to append, replace or delete. Various wrappers added for 1728 certifcates and CRLs. 1729 [Steve Henson] 1730 1731 *) Fix to avoid calling the underlying ASN1 print routine when 1732 an extension cannot be parsed. Correct a typo in the 1733 OCSP_SERVICELOC extension. Tidy up print OCSP format. 1734 [Steve Henson] 1735 1736 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 1737 entries for variables. 1738 [Steve Henson] 1739 1740 *) Add functionality to apps/openssl.c for detecting locking 1741 problems: As the program is single-threaded, all we have 1742 to do is register a locking callback using an array for 1743 storing which locks are currently held by the program. 1744 [Bodo Moeller] 1745 1746 *) Use a lock around the call to CRYPTO_get_ex_new_index() in 1747 SSL_get_ex_data_X509_STORE_idx(), which is used in 1748 ssl_verify_cert_chain() and thus can be called at any time 1749 during TLS/SSL handshakes so that thread-safety is essential. 1750 Unfortunately, the ex_data design is not at all suited 1751 for multi-threaded use, so it probably should be abolished. 1752 [Bodo Moeller] 1753 1754 *) Added Broadcom "ubsec" ENGINE to OpenSSL. 1755 [Broadcom, tweaked and integrated by Geoff Thorpe] 1756 1757 *) Move common extension printing code to new function 1758 X509V3_print_extensions(). Reorganise OCSP print routines and 1759 implement some needed OCSP ASN1 functions. Add OCSP extensions. 1760 [Steve Henson] 1761 1762 *) New function X509_signature_print() to remove duplication in some 1763 print routines. 1764 [Steve Henson] 1765 1766 *) Add a special meaning when SET OF and SEQUENCE OF flags are both 1767 set (this was treated exactly the same as SET OF previously). This 1768 is used to reorder the STACK representing the structure to match the 1769 encoding. This will be used to get round a problem where a PKCS7 1770 structure which was signed could not be verified because the STACK 1771 order did not reflect the encoded order. 1772 [Steve Henson] 1773 1774 *) Reimplement the OCSP ASN1 module using the new code. 1775 [Steve Henson] 1776 1777 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 1778 for its ASN1 operations. The old style function pointers still exist 1779 for now but they will eventually go away. 1780 [Steve Henson] 1781 1782 *) Merge in replacement ASN1 code from the ASN1 branch. This almost 1783 completely replaces the old ASN1 functionality with a table driven 1784 encoder and decoder which interprets an ASN1_ITEM structure describing 1785 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 1786 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 1787 has also been converted to the new form. 1788 [Steve Henson] 1789 1790 *) Change BN_mod_exp_recp so that negative moduli are tolerated 1791 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 1792 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 1793 for negative moduli. 1794 [Bodo Moeller] 1795 1796 *) Fix BN_uadd and BN_usub: Always return non-negative results instead 1797 of not touching the result's sign bit. 1798 [Bodo Moeller] 1799 1800 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 1801 set. 1802 [Bodo Moeller] 1803 1804 *) Changed the LHASH code to use prototypes for callbacks, and created 1805 macros to declare and implement thin (optionally static) functions 1806 that provide type-safety and avoid function pointer casting for the 1807 type-specific callbacks. 1808 [Geoff Thorpe] 1809 1810 *) Added Kerberos Cipher Suites to be used with TLS, as written in 1811 RFC 2712. 1812 [Veers Staats <staatsvr@asc.hpc.mil>, 1813 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 1814 1815 *) Reformat the FAQ so the different questions and answers can be divided 1816 in sections depending on the subject. 1817 [Richard Levitte] 1818 1819 *) Have the zlib compression code load ZLIB.DLL dynamically under 1820 Windows. 1821 [Richard Levitte] 1822 1823 *) New function BN_mod_sqrt for computing square roots modulo a prime 1824 (using the probabilistic Tonelli-Shanks algorithm unless 1825 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 1826 be handled deterministically). 1827 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1828 1829 *) Make BN_mod_inverse faster by explicitly handling small quotients 1830 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 1831 512 bits], about 30% for larger ones [1024 or 2048 bits].) 1832 [Bodo Moeller] 1833 1834 *) New function BN_kronecker. 1835 [Bodo Moeller] 1836 1837 *) Fix BN_gcd so that it works on negative inputs; the result is 1838 positive unless both parameters are zero. 1839 Previously something reasonably close to an infinite loop was 1840 possible because numbers could be growing instead of shrinking 1841 in the implementation of Euclid's algorithm. 1842 [Bodo Moeller] 1843 1844 *) Fix BN_is_word() and BN_is_one() macros to take into account the 1845 sign of the number in question. 1846 1847 Fix BN_is_word(a,w) to work correctly for w == 0. 1848 1849 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 1850 because its test if the absolute value of 'a' equals 'w'. 1851 Note that BN_abs_is_word does *not* handle w == 0 reliably; 1852 it exists mostly for use in the implementations of BN_is_zero(), 1853 BN_is_one(), and BN_is_word(). 1854 [Bodo Moeller] 1855 1856 *) New function BN_swap. 1857 [Bodo Moeller] 1858 1859 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 1860 the exponentiation functions are more likely to produce reasonable 1861 results on negative inputs. 1862 [Bodo Moeller] 1863 1864 *) Change BN_mod_mul so that the result is always non-negative. 1865 Previously, it could be negative if one of the factors was negative; 1866 I don't think anyone really wanted that behaviour. 1867 [Bodo Moeller] 1868 1869 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 1870 (except for exponentiation, which stays in crypto/bn/bn_exp.c, 1871 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 1872 and add new functions: 1873 1874 BN_nnmod 1875 BN_mod_sqr 1876 BN_mod_add 1877 BN_mod_add_quick 1878 BN_mod_sub 1879 BN_mod_sub_quick 1880 BN_mod_lshift1 1881 BN_mod_lshift1_quick 1882 BN_mod_lshift 1883 BN_mod_lshift_quick 1884 1885 These functions always generate non-negative results. 1886 1887 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 1888 such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 1889 1890 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 1891 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 1892 be reduced modulo m. 1893 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 1894 1895#if 0 1896 The following entry accidentily appeared in the CHANGES file 1897 distributed with OpenSSL 0.9.7. The modifications described in 1898 it do *not* apply to OpenSSL 0.9.7. 1899 1900 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 1901 was actually never needed) and in BN_mul(). The removal in BN_mul() 1902 required a small change in bn_mul_part_recursive() and the addition 1903 of the functions bn_cmp_part_words(), bn_sub_part_words() and 1904 bn_add_part_words(), which do the same thing as bn_cmp_words(), 1905 bn_sub_words() and bn_add_words() except they take arrays with 1906 differing sizes. 1907 [Richard Levitte] 1908#endif 1909 1910 *) In 'openssl passwd', verify passwords read from the terminal 1911 unless the '-salt' option is used (which usually means that 1912 verification would just waste user's time since the resulting 1913 hash is going to be compared with some given password hash) 1914 or the new '-noverify' option is used. 1915 1916 This is an incompatible change, but it does not affect 1917 non-interactive use of 'openssl passwd' (passwords on the command 1918 line, '-stdin' option, '-in ...' option) and thus should not 1919 cause any problems. 1920 [Bodo Moeller] 1921 1922 *) Remove all references to RSAref, since there's no more need for it. 1923 [Richard Levitte] 1924 1925 *) Make DSO load along a path given through an environment variable 1926 (SHLIB_PATH) with shl_load(). 1927 [Richard Levitte] 1928 1929 *) Constify the ENGINE code as a result of BIGNUM constification. 1930 Also constify the RSA code and most things related to it. In a 1931 few places, most notable in the depth of the ASN.1 code, ugly 1932 casts back to non-const were required (to be solved at a later 1933 time) 1934 [Richard Levitte] 1935 1936 *) Make it so the openssl application has all engines loaded by default. 1937 [Richard Levitte] 1938 1939 *) Constify the BIGNUM routines a little more. 1940 [Richard Levitte] 1941 1942 *) Add the following functions: 1943 1944 ENGINE_load_cswift() 1945 ENGINE_load_chil() 1946 ENGINE_load_atalla() 1947 ENGINE_load_nuron() 1948 ENGINE_load_builtin_engines() 1949 1950 That way, an application can itself choose if external engines that 1951 are built-in in OpenSSL shall ever be used or not. The benefit is 1952 that applications won't have to be linked with libdl or other dso 1953 libraries unless it's really needed. 1954 1955 Changed 'openssl engine' to load all engines on demand. 1956 Changed the engine header files to avoid the duplication of some 1957 declarations (they differed!). 1958 [Richard Levitte] 1959 1960 *) 'openssl engine' can now list capabilities. 1961 [Richard Levitte] 1962 1963 *) Better error reporting in 'openssl engine'. 1964 [Richard Levitte] 1965 1966 *) Never call load_dh_param(NULL) in s_server. 1967 [Bodo Moeller] 1968 1969 *) Add engine application. It can currently list engines by name and 1970 identity, and test if they are actually available. 1971 [Richard Levitte] 1972 1973 *) Improve RPM specification file by forcing symbolic linking and making 1974 sure the installed documentation is also owned by root.root. 1975 [Damien Miller <djm@mindrot.org>] 1976 1977 *) Give the OpenSSL applications more possibilities to make use of 1978 keys (public as well as private) handled by engines. 1979 [Richard Levitte] 1980 1981 *) Add OCSP code that comes from CertCo. 1982 [Richard Levitte] 1983 1984 *) Add VMS support for the Rijndael code. 1985 [Richard Levitte] 1986 1987 *) Added untested support for Nuron crypto accelerator. 1988 [Ben Laurie] 1989 1990 *) Add support for external cryptographic devices. This code was 1991 previously distributed separately as the "engine" branch. 1992 [Geoff Thorpe, Richard Levitte] 1993 1994 *) Rework the filename-translation in the DSO code. It is now possible to 1995 have far greater control over how a "name" is turned into a filename 1996 depending on the operating environment and any oddities about the 1997 different shared library filenames on each system. 1998 [Geoff Thorpe] 1999 2000 *) Support threads on FreeBSD-elf in Configure. 2001 [Richard Levitte] 2002 2003 *) Fix for SHA1 assembly problem with MASM: it produces 2004 warnings about corrupt line number information when assembling 2005 with debugging information. This is caused by the overlapping 2006 of two sections. 2007 [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 2008 2009 *) NCONF changes. 2010 NCONF_get_number() has no error checking at all. As a replacement, 2011 NCONF_get_number_e() is defined (_e for "error checking") and is 2012 promoted strongly. The old NCONF_get_number is kept around for 2013 binary backward compatibility. 2014 Make it possible for methods to load from something other than a BIO, 2015 by providing a function pointer that is given a name instead of a BIO. 2016 For example, this could be used to load configuration data from an 2017 LDAP server. 2018 [Richard Levitte] 2019 2020 *) Fix for non blocking accept BIOs. Added new I/O special reason 2021 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 2022 with non blocking I/O was not possible because no retry code was 2023 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 2024 this case. 2025 [Steve Henson] 2026 2027 *) Added the beginnings of Rijndael support. 2028 [Ben Laurie] 2029 2030 *) Fix for bug in DirectoryString mask setting. Add support for 2031 X509_NAME_print_ex() in 'req' and X509_print_ex() function 2032 to allow certificate printing to more controllable, additional 2033 'certopt' option to 'x509' to allow new printing options to be 2034 set. 2035 [Steve Henson] 2036 2037 *) Clean old EAY MD5 hack from e_os.h. 2038 [Richard Levitte] 2039 2040 Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 2041 2042 *) Fix various bugs revealed by running the NISCC test suite: 2043 2044 Stop out of bounds reads in the ASN1 code when presented with 2045 invalid tags (CAN-2003-0543 and CAN-2003-0544). 2046 2047 If verify callback ignores invalid public key errors don't try to check 2048 certificate signature with the NULL public key. 2049 2050 [Steve Henson] 2051 2052 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 2053 if the server requested one: as stated in TLS 1.0 and SSL 3.0 2054 specifications. 2055 [Steve Henson] 2056 2057 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 2058 extra data after the compression methods not only for TLS 1.0 2059 but also for SSL 3.0 (as required by the specification). 2060 [Bodo Moeller; problem pointed out by Matthias Loepfe] 2061 2062 *) Change X509_certificate_type() to mark the key as exported/exportable 2063 when it's 512 *bits* long, not 512 bytes. 2064 [Richard Levitte] 2065 2066 Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 2067 2068 *) Countermeasure against the Klima-Pokorny-Rosa extension of 2069 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 2070 a protocol version number mismatch like a decryption error 2071 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 2072 [Bodo Moeller] 2073 2074 *) Turn on RSA blinding by default in the default implementation 2075 to avoid a timing attack. Applications that don't want it can call 2076 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 2077 They would be ill-advised to do so in most cases. 2078 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 2079 2080 *) Change RSA blinding code so that it works when the PRNG is not 2081 seeded (in this case, the secret RSA exponent is abused as 2082 an unpredictable seed -- if it is not unpredictable, there 2083 is no point in blinding anyway). Make RSA blinding thread-safe 2084 by remembering the creator's thread ID in rsa->blinding and 2085 having all other threads use local one-time blinding factors 2086 (this requires more computation than sharing rsa->blinding, but 2087 avoids excessive locking; and if an RSA object is not shared 2088 between threads, blinding will still be very fast). 2089 [Bodo Moeller] 2090 2091 Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 2092 2093 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 2094 via timing by performing a MAC computation even if incorrrect 2095 block cipher padding has been found. This is a countermeasure 2096 against active attacks where the attacker has to distinguish 2097 between bad padding and a MAC verification error. (CAN-2003-0078) 2098 2099 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 2100 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 2101 Martin Vuagnoux (EPFL, Ilion)] 2102 2103 Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 2104 2105 *) New function OPENSSL_cleanse(), which is used to cleanse a section of 2106 memory from it's contents. This is done with a counter that will 2107 place alternating values in each byte. This can be used to solve 2108 two issues: 1) the removal of calls to memset() by highly optimizing 2109 compilers, and 2) cleansing with other values than 0, since those can 2110 be read through on certain media, for example a swap space on disk. 2111 [Geoff Thorpe] 2112 2113 *) Bugfix: client side session caching did not work with external caching, 2114 because the session->cipher setting was not restored when reloading 2115 from the external cache. This problem was masked, when 2116 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 2117 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 2118 [Lutz Jaenicke] 2119 2120 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 2121 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 2122 [Zeev Lieber <zeev-l@yahoo.com>] 2123 2124 *) Undo an undocumented change introduced in 0.9.6e which caused 2125 repeated calls to OpenSSL_add_all_ciphers() and 2126 OpenSSL_add_all_digests() to be ignored, even after calling 2127 EVP_cleanup(). 2128 [Richard Levitte] 2129 2130 *) Change the default configuration reader to deal with last line not 2131 being properly terminated. 2132 [Richard Levitte] 2133 2134 *) Change X509_NAME_cmp() so it applies the special rules on handling 2135 DN values that are of type PrintableString, as well as RDNs of type 2136 emailAddress where the value has the type ia5String. 2137 [stefank@valicert.com via Richard Levitte] 2138 2139 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 2140 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 2141 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 2142 the bitwise-OR of the two for use by the majority of applications 2143 wanting this behaviour, and update the docs. The documented 2144 behaviour and actual behaviour were inconsistent and had been 2145 changing anyway, so this is more a bug-fix than a behavioural 2146 change. 2147 [Geoff Thorpe, diagnosed by Nadav Har'El] 2148 2149 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 2150 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 2151 [Bodo Moeller] 2152 2153 *) Fix initialization code race conditions in 2154 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 2155 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 2156 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 2157 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 2158 ssl2_get_cipher_by_char(), 2159 ssl3_get_cipher_by_char(). 2160 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 2161 2162 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 2163 the cached sessions are flushed, as the remove_cb() might use ex_data 2164 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 2165 (see [openssl.org #212]). 2166 [Geoff Thorpe, Lutz Jaenicke] 2167 2168 *) Fix typo in OBJ_txt2obj which incorrectly passed the content 2169 length, instead of the encoding length to d2i_ASN1_OBJECT. 2170 [Steve Henson] 2171 2172 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 2173 2174 *) [In 0.9.6g-engine release:] 2175 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 2176 [Lynn Gazis <lgazis@rainbow.com>] 2177 2178 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 2179 2180 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 2181 and get fix the header length calculation. 2182 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 2183 Alon Kantor <alonk@checkpoint.com> (and others), 2184 Steve Henson] 2185 2186 *) Use proper error handling instead of 'assertions' in buffer 2187 overflow checks added in 0.9.6e. This prevents DoS (the 2188 assertions could call abort()). 2189 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 2190 2191 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 2192 2193 *) Add various sanity checks to asn1_get_length() to reject 2194 the ASN1 length bytes if they exceed sizeof(long), will appear 2195 negative or the content length exceeds the length of the 2196 supplied buffer. 2197 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2198 2199 *) Fix cipher selection routines: ciphers without encryption had no flags 2200 for the cipher strength set and where therefore not handled correctly 2201 by the selection routines (PR #130). 2202 [Lutz Jaenicke] 2203 2204 *) Fix EVP_dsa_sha macro. 2205 [Nils Larsch] 2206 2207 *) New option 2208 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 2209 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 2210 that was added in OpenSSL 0.9.6d. 2211 2212 As the countermeasure turned out to be incompatible with some 2213 broken SSL implementations, the new option is part of SSL_OP_ALL. 2214 SSL_OP_ALL is usually employed when compatibility with weird SSL 2215 implementations is desired (e.g. '-bugs' option to 's_client' and 2216 's_server'), so the new option is automatically set in many 2217 applications. 2218 [Bodo Moeller] 2219 2220 *) Changes in security patch: 2221 2222 Changes marked "(CHATS)" were sponsored by the Defense Advanced 2223 Research Projects Agency (DARPA) and Air Force Research Laboratory, 2224 Air Force Materiel Command, USAF, under agreement number 2225 F30602-01-2-0537. 2226 2227 *) Add various sanity checks to asn1_get_length() to reject 2228 the ASN1 length bytes if they exceed sizeof(long), will appear 2229 negative or the content length exceeds the length of the 2230 supplied buffer. (CAN-2002-0659) 2231 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 2232 2233 *) Assertions for various potential buffer overflows, not known to 2234 happen in practice. 2235 [Ben Laurie (CHATS)] 2236 2237 *) Various temporary buffers to hold ASCII versions of integers were 2238 too small for 64 bit platforms. (CAN-2002-0655) 2239 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 2240 2241 *) Remote buffer overflow in SSL3 protocol - an attacker could 2242 supply an oversized session ID to a client. (CAN-2002-0656) 2243 [Ben Laurie (CHATS)] 2244 2245 *) Remote buffer overflow in SSL2 protocol - an attacker could 2246 supply an oversized client master key. (CAN-2002-0656) 2247 [Ben Laurie (CHATS)] 2248 2249 Changes between 0.9.6c and 0.9.6d [9 May 2002] 2250 2251 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 2252 encoded as NULL) with id-dsa-with-sha1. 2253 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 2254 2255 *) Check various X509_...() return values in apps/req.c. 2256 [Nils Larsch <nla@trustcenter.de>] 2257 2258 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 2259 an end-of-file condition would erronously be flagged, when the CRLF 2260 was just at the end of a processed block. The bug was discovered when 2261 processing data through a buffering memory BIO handing the data to a 2262 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 2263 <ptsekov@syntrex.com> and Nedelcho Stanev. 2264 [Lutz Jaenicke] 2265 2266 *) Implement a countermeasure against a vulnerability recently found 2267 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 2268 before application data chunks to avoid the use of known IVs 2269 with data potentially chosen by the attacker. 2270 [Bodo Moeller] 2271 2272 *) Fix length checks in ssl3_get_client_hello(). 2273 [Bodo Moeller] 2274 2275 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 2276 to prevent ssl3_read_internal() from incorrectly assuming that 2277 ssl3_read_bytes() found application data while handshake 2278 processing was enabled when in fact s->s3->in_read_app_data was 2279 merely automatically cleared during the initial handshake. 2280 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 2281 2282 *) Fix object definitions for Private and Enterprise: they were not 2283 recognized in their shortname (=lowercase) representation. Extend 2284 obj_dat.pl to issue an error when using undefined keywords instead 2285 of silently ignoring the problem (Svenning Sorensen 2286 <sss@sss.dnsalias.net>). 2287 [Lutz Jaenicke] 2288 2289 *) Fix DH_generate_parameters() so that it works for 'non-standard' 2290 generators, i.e. generators other than 2 and 5. (Previously, the 2291 code did not properly initialise the 'add' and 'rem' values to 2292 BN_generate_prime().) 2293 2294 In the new general case, we do not insist that 'generator' is 2295 actually a primitive root: This requirement is rather pointless; 2296 a generator of the order-q subgroup is just as good, if not 2297 better. 2298 [Bodo Moeller] 2299 2300 *) Map new X509 verification errors to alerts. Discovered and submitted by 2301 Tom Wu <tom@arcot.com>. 2302 [Lutz Jaenicke] 2303 2304 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 2305 returning non-zero before the data has been completely received 2306 when using non-blocking I/O. 2307 [Bodo Moeller; problem pointed out by John Hughes] 2308 2309 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 2310 [Ben Laurie, Lutz Jaenicke] 2311 2312 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 2313 Yoram Zahavi <YoramZ@gilian.com>). 2314 [Lutz Jaenicke] 2315 2316 *) Add information about CygWin 1.3 and on, and preserve proper 2317 configuration for the versions before that. 2318 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 2319 2320 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 2321 check whether we deal with a copy of a session and do not delete from 2322 the cache in this case. Problem reported by "Izhar Shoshani Levi" 2323 <izhar@checkpoint.com>. 2324 [Lutz Jaenicke] 2325 2326 *) Do not store session data into the internal session cache, if it 2327 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 2328 flag is set). Proposed by Aslam <aslam@funk.com>. 2329 [Lutz Jaenicke] 2330 2331 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 2332 value is 0. 2333 [Richard Levitte] 2334 2335 *) [In 0.9.6d-engine release:] 2336 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 2337 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 2338 2339 *) Add the configuration target linux-s390x. 2340 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 2341 2342 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 2343 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 2344 variable as an indication that a ClientHello message has been 2345 received. As the flag value will be lost between multiple 2346 invocations of ssl3_accept when using non-blocking I/O, the 2347 function may not be aware that a handshake has actually taken 2348 place, thus preventing a new session from being added to the 2349 session cache. 2350 2351 To avoid this problem, we now set s->new_session to 2 instead of 2352 using a local variable. 2353 [Lutz Jaenicke, Bodo Moeller] 2354 2355 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 2356 if the SSL_R_LENGTH_MISMATCH error is detected. 2357 [Geoff Thorpe, Bodo Moeller] 2358 2359 *) New 'shared_ldflag' column in Configure platform table. 2360 [Richard Levitte] 2361 2362 *) Fix EVP_CIPHER_mode macro. 2363 ["Dan S. Camper" <dan@bti.net>] 2364 2365 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 2366 type, we must throw them away by setting rr->length to 0. 2367 [D P Chang <dpc@qualys.com>] 2368 2369 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 2370 2371 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 2372 <Dominikus.Scherkl@biodata.com>. (The previous implementation 2373 worked incorrectly for those cases where range = 10..._2 and 2374 3*range is two bits longer than range.) 2375 [Bodo Moeller] 2376 2377 *) Only add signing time to PKCS7 structures if it is not already 2378 present. 2379 [Steve Henson] 2380 2381 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 2382 OBJ_ld_ce should be OBJ_id_ce. 2383 Also some ip-pda OIDs in crypto/objects/objects.txt were 2384 incorrect (cf. RFC 3039). 2385 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 2386 2387 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 2388 returns early because it has nothing to do. 2389 [Andy Schneider <andy.schneider@bjss.co.uk>] 2390 2391 *) [In 0.9.6c-engine release:] 2392 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 2393 [Andy Schneider <andy.schneider@bjss.co.uk>] 2394 2395 *) [In 0.9.6c-engine release:] 2396 Add support for Cryptographic Appliance's keyserver technology. 2397 (Use engine 'keyclient') 2398 [Cryptographic Appliances and Geoff Thorpe] 2399 2400 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 2401 is called via tools/c89.sh because arguments have to be 2402 rearranged (all '-L' options must appear before the first object 2403 modules). 2404 [Richard Shapiro <rshapiro@abinitio.com>] 2405 2406 *) [In 0.9.6c-engine release:] 2407 Add support for Broadcom crypto accelerator cards, backported 2408 from 0.9.7. 2409 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 2410 2411 *) [In 0.9.6c-engine release:] 2412 Add support for SureWare crypto accelerator cards from 2413 Baltimore Technologies. (Use engine 'sureware') 2414 [Baltimore Technologies and Mark Cox] 2415 2416 *) [In 0.9.6c-engine release:] 2417 Add support for crypto accelerator cards from Accelerated 2418 Encryption Processing, www.aep.ie. (Use engine 'aep') 2419 [AEP Inc. and Mark Cox] 2420 2421 *) Add a configuration entry for gcc on UnixWare. 2422 [Gary Benson <gbenson@redhat.com>] 2423 2424 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 2425 messages are stored in a single piece (fixed-length part and 2426 variable-length part combined) and fix various bugs found on the way. 2427 [Bodo Moeller] 2428 2429 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 2430 instead. BIO_gethostbyname() does not know what timeouts are 2431 appropriate, so entries would stay in cache even when they have 2432 become invalid. 2433 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 2434 2435 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 2436 faced with a pathologically small ClientHello fragment that does 2437 not contain client_version: Instead of aborting with an error, 2438 simply choose the highest available protocol version (i.e., 2439 TLS 1.0 unless it is disabled). In practice, ClientHello 2440 messages are never sent like this, but this change gives us 2441 strictly correct behaviour at least for TLS. 2442 [Bodo Moeller] 2443 2444 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 2445 never resets s->method to s->ctx->method when called from within 2446 one of the SSL handshake functions. 2447 [Bodo Moeller; problem pointed out by Niko Baric] 2448 2449 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 2450 (sent using the client's version number) if client_version is 2451 smaller than the protocol version in use. Also change 2452 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 2453 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 2454 the client will at least see that alert. 2455 [Bodo Moeller] 2456 2457 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 2458 correctly. 2459 [Bodo Moeller] 2460 2461 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 2462 client receives HelloRequest while in a handshake. 2463 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 2464 2465 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 2466 should end in 'break', not 'goto end' which circuments various 2467 cleanups done in state SSL_ST_OK. But session related stuff 2468 must be disabled for SSL_ST_OK in the case that we just sent a 2469 HelloRequest. 2470 2471 Also avoid some overhead by not calling ssl_init_wbio_buffer() 2472 before just sending a HelloRequest. 2473 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 2474 2475 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 2476 reveal whether illegal block cipher padding was found or a MAC 2477 verification error occured. (Neither SSLerr() codes nor alerts 2478 are directly visible to potential attackers, but the information 2479 may leak via logfiles.) 2480 2481 Similar changes are not required for the SSL 2.0 implementation 2482 because the number of padding bytes is sent in clear for SSL 2.0, 2483 and the extra bytes are just ignored. However ssl/s2_pkt.c 2484 failed to verify that the purported number of padding bytes is in 2485 the legal range. 2486 [Bodo Moeller] 2487 2488 *) Add OpenUNIX-8 support including shared libraries 2489 (Boyd Lynn Gerber <gerberb@zenez.com>). 2490 [Lutz Jaenicke] 2491 2492 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 2493 'wristwatch attack' using huge encoding parameters (cf. 2494 James H. Manger's CRYPTO 2001 paper). Note that the 2495 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 2496 encoding parameters and hence was not vulnerable. 2497 [Bodo Moeller] 2498 2499 *) BN_sqr() bug fix. 2500 [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 2501 2502 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 2503 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 2504 followed by modular reduction. 2505 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 2506 2507 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 2508 equivalent based on BN_pseudo_rand() instead of BN_rand(). 2509 [Bodo Moeller] 2510 2511 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 2512 This function was broken, as the check for a new client hello message 2513 to handle SGC did not allow these large messages. 2514 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 2515 [Lutz Jaenicke] 2516 2517 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 2518 [Lutz Jaenicke] 2519 2520 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 2521 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 2522 [Lutz Jaenicke] 2523 2524 *) Rework the configuration and shared library support for Tru64 Unix. 2525 The configuration part makes use of modern compiler features and 2526 still retains old compiler behavior for those that run older versions 2527 of the OS. The shared library support part includes a variant that 2528 uses the RPATH feature, and is available through the special 2529 configuration target "alpha-cc-rpath", which will never be selected 2530 automatically. 2531 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 2532 2533 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 2534 with the same message size as in ssl3_get_certificate_request(). 2535 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 2536 messages might inadvertently be reject as too long. 2537 [Petr Lampa <lampa@fee.vutbr.cz>] 2538 2539 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 2540 [Andy Polyakov] 2541 2542 *) Modified SSL library such that the verify_callback that has been set 2543 specificly for an SSL object with SSL_set_verify() is actually being 2544 used. Before the change, a verify_callback set with this function was 2545 ignored and the verify_callback() set in the SSL_CTX at the time of 2546 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 2547 to allow the necessary settings. 2548 [Lutz Jaenicke] 2549 2550 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 2551 explicitly to NULL, as at least on Solaris 8 this seems not always to be 2552 done automatically (in contradiction to the requirements of the C 2553 standard). This made problems when used from OpenSSH. 2554 [Lutz Jaenicke] 2555 2556 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 2557 dh->length and always used 2558 2559 BN_rand_range(priv_key, dh->p). 2560 2561 BN_rand_range() is not necessary for Diffie-Hellman, and this 2562 specific range makes Diffie-Hellman unnecessarily inefficient if 2563 dh->length (recommended exponent length) is much smaller than the 2564 length of dh->p. We could use BN_rand_range() if the order of 2565 the subgroup was stored in the DH structure, but we only have 2566 dh->length. 2567 2568 So switch back to 2569 2570 BN_rand(priv_key, l, ...) 2571 2572 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 2573 otherwise. 2574 [Bodo Moeller] 2575 2576 *) In 2577 2578 RSA_eay_public_encrypt 2579 RSA_eay_private_decrypt 2580 RSA_eay_private_encrypt (signing) 2581 RSA_eay_public_decrypt (signature verification) 2582 2583 (default implementations for RSA_public_encrypt, 2584 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 2585 always reject numbers >= n. 2586 [Bodo Moeller] 2587 2588 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 2589 to synchronize access to 'locking_thread'. This is necessary on 2590 systems where access to 'locking_thread' (an 'unsigned long' 2591 variable) is not atomic. 2592 [Bodo Moeller] 2593 2594 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 2595 *before* setting the 'crypto_lock_rand' flag. The previous code had 2596 a race condition if 0 is a valid thread ID. 2597 [Travis Vitek <vitek@roguewave.com>] 2598 2599 *) Add support for shared libraries under Irix. 2600 [Albert Chin-A-Young <china@thewrittenword.com>] 2601 2602 *) Add configuration option to build on Linux on both big-endian and 2603 little-endian MIPS. 2604 [Ralf Baechle <ralf@uni-koblenz.de>] 2605 2606 *) Add the possibility to create shared libraries on HP-UX. 2607 [Richard Levitte] 2608 2609 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 2610 2611 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 2612 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 2613 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 2614 PRNG state recovery was possible based on the output of 2615 one PRNG request appropriately sized to gain knowledge on 2616 'md' followed by enough consecutive 1-byte PRNG requests 2617 to traverse all of 'state'. 2618 2619 1. When updating 'md_local' (the current thread's copy of 'md') 2620 during PRNG output generation, hash all of the previous 2621 'md_local' value, not just the half used for PRNG output. 2622 2623 2. Make the number of bytes from 'state' included into the hash 2624 independent from the number of PRNG bytes requested. 2625 2626 The first measure alone would be sufficient to avoid 2627 Markku-Juhani's attack. (Actually it had never occurred 2628 to me that the half of 'md_local' used for chaining was the 2629 half from which PRNG output bytes were taken -- I had always 2630 assumed that the secret half would be used.) The second 2631 measure makes sure that additional data from 'state' is never 2632 mixed into 'md_local' in small portions; this heuristically 2633 further strengthens the PRNG. 2634 [Bodo Moeller] 2635 2636 *) Fix crypto/bn/asm/mips3.s. 2637 [Andy Polyakov] 2638 2639 *) When only the key is given to "enc", the IV is undefined. Print out 2640 an error message in this case. 2641 [Lutz Jaenicke] 2642 2643 *) Handle special case when X509_NAME is empty in X509 printing routines. 2644 [Steve Henson] 2645 2646 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 2647 positive and less than q. 2648 [Bodo Moeller] 2649 2650 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 2651 used: it isn't thread safe and the add_lock_callback should handle 2652 that itself. 2653 [Paul Rose <Paul.Rose@bridge.com>] 2654 2655 *) Verify that incoming data obeys the block size in 2656 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 2657 [Bodo Moeller] 2658 2659 *) Fix OAEP check. 2660 [Ulf M�ller, Bodo M�ller] 2661 2662 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 2663 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 2664 when fixing the server behaviour for backwards-compatible 'client 2665 hello' messages. (Note that the attack is impractical against 2666 SSL 3.0 and TLS 1.0 anyway because length and version checking 2667 means that the probability of guessing a valid ciphertext is 2668 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 2669 paper.) 2670 2671 Before 0.9.5, the countermeasure (hide the error by generating a 2672 random 'decryption result') did not work properly because 2673 ERR_clear_error() was missing, meaning that SSL_get_error() would 2674 detect the supposedly ignored error. 2675 2676 Both problems are now fixed. 2677 [Bodo Moeller] 2678 2679 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 2680 (previously it was 1024). 2681 [Bodo Moeller] 2682 2683 *) Fix for compatibility mode trust settings: ignore trust settings 2684 unless some valid trust or reject settings are present. 2685 [Steve Henson] 2686 2687 *) Fix for blowfish EVP: its a variable length cipher. 2688 [Steve Henson] 2689 2690 *) Fix various bugs related to DSA S/MIME verification. Handle missing 2691 parameters in DSA public key structures and return an error in the 2692 DSA routines if parameters are absent. 2693 [Steve Henson] 2694 2695 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 2696 in the current directory if neither $RANDFILE nor $HOME was set. 2697 RAND_file_name() in 0.9.6a returned NULL in this case. This has 2698 caused some confusion to Windows users who haven't defined $HOME. 2699 Thus RAND_file_name() is changed again: e_os.h can define a 2700 DEFAULT_HOME, which will be used if $HOME is not set. 2701 For Windows, we use "C:"; on other platforms, we still require 2702 environment variables. 2703 2704 *) Move 'if (!initialized) RAND_poll()' into regions protected by 2705 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 2706 having multiple threads call RAND_poll() concurrently. 2707 [Bodo Moeller] 2708 2709 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 2710 combination of a flag and a thread ID variable. 2711 Otherwise while one thread is in ssleay_rand_bytes (which sets the 2712 flag), *other* threads can enter ssleay_add_bytes without obeying 2713 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 2714 that they do not hold after the first thread unsets add_do_not_lock). 2715 [Bodo Moeller] 2716 2717 *) Change bctest again: '-x' expressions are not available in all 2718 versions of 'test'. 2719 [Bodo Moeller] 2720 2721 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 2722 2723 *) Fix a couple of memory leaks in PKCS7_dataDecode() 2724 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 2725 2726 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 2727 the default extension for executables, if any. Also, make the perl 2728 scripts that use symlink() to test if it really exists and use "cp" 2729 if it doesn't. All this made OpenSSL compilable and installable in 2730 CygWin. 2731 [Richard Levitte] 2732 2733 *) Fix for asn1_GetSequence() for indefinite length constructed data. 2734 If SEQUENCE is length is indefinite just set c->slen to the total 2735 amount of data available. 2736 [Steve Henson, reported by shige@FreeBSD.org] 2737 [This change does not apply to 0.9.7.] 2738 2739 *) Change bctest to avoid here-documents inside command substitution 2740 (workaround for FreeBSD /bin/sh bug). 2741 For compatibility with Ultrix, avoid shell functions (introduced 2742 in the bctest version that searches along $PATH). 2743 [Bodo Moeller] 2744 2745 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 2746 with des_encrypt() defined on some operating systems, like Solaris 2747 and UnixWare. 2748 [Richard Levitte] 2749 2750 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 2751 On the Importance of Eliminating Errors in Cryptographic 2752 Computations, J. Cryptology 14 (2001) 2, 101-119, 2753 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 2754 [Ulf Moeller] 2755 2756 *) MIPS assembler BIGNUM division bug fix. 2757 [Andy Polyakov] 2758 2759 *) Disabled incorrect Alpha assembler code. 2760 [Richard Levitte] 2761 2762 *) Fix PKCS#7 decode routines so they correctly update the length 2763 after reading an EOC for the EXPLICIT tag. 2764 [Steve Henson] 2765 [This change does not apply to 0.9.7.] 2766 2767 *) Fix bug in PKCS#12 key generation routines. This was triggered 2768 if a 3DES key was generated with a 0 initial byte. Include 2769 PKCS12_BROKEN_KEYGEN compilation option to retain the old 2770 (but broken) behaviour. 2771 [Steve Henson] 2772 2773 *) Enhance bctest to search for a working bc along $PATH and print 2774 it when found. 2775 [Tim Rice <tim@multitalents.net> via Richard Levitte] 2776 2777 *) Fix memory leaks in err.c: free err_data string if necessary; 2778 don't write to the wrong index in ERR_set_error_data. 2779 [Bodo Moeller] 2780 2781 *) Implement ssl23_peek (analogous to ssl23_read), which previously 2782 did not exist. 2783 [Bodo Moeller] 2784 2785 *) Replace rdtsc with _emit statements for VC++ version 5. 2786 [Jeremy Cooper <jeremy@baymoo.org>] 2787 2788 *) Make it possible to reuse SSLv2 sessions. 2789 [Richard Levitte] 2790 2791 *) In copy_email() check for >= 0 as a return value for 2792 X509_NAME_get_index_by_NID() since 0 is a valid index. 2793 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 2794 2795 *) Avoid coredump with unsupported or invalid public keys by checking if 2796 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 2797 PKCS7_verify() fails with non detached data. 2798 [Steve Henson] 2799 2800 *) Don't use getenv in library functions when run as setuid/setgid. 2801 New function OPENSSL_issetugid(). 2802 [Ulf Moeller] 2803 2804 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 2805 due to incorrect handling of multi-threading: 2806 2807 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 2808 2809 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 2810 2811 3. Count how many times MemCheck_off() has been called so that 2812 nested use can be treated correctly. This also avoids 2813 inband-signalling in the previous code (which relied on the 2814 assumption that thread ID 0 is impossible). 2815 [Bodo Moeller] 2816 2817 *) Add "-rand" option also to s_client and s_server. 2818 [Lutz Jaenicke] 2819 2820 *) Fix CPU detection on Irix 6.x. 2821 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 2822 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 2823 2824 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 2825 was empty. 2826 [Steve Henson] 2827 [This change does not apply to 0.9.7.] 2828 2829 *) Use the cached encoding of an X509_NAME structure rather than 2830 copying it. This is apparently the reason for the libsafe "errors" 2831 but the code is actually correct. 2832 [Steve Henson] 2833 2834 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 2835 Bleichenbacher's DSA attack. 2836 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 2837 to be set and top=0 forces the highest bit to be set; top=-1 is new 2838 and leaves the highest bit random. 2839 [Ulf Moeller, Bodo Moeller] 2840 2841 *) In the NCONF_...-based implementations for CONF_... queries 2842 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 2843 a temporary CONF structure with the data component set to NULL 2844 (which gives segmentation faults in lh_retrieve). 2845 Instead, use NULL for the CONF pointer in CONF_get_string and 2846 CONF_get_number (which may use environment variables) and directly 2847 return NULL from CONF_get_section. 2848 [Bodo Moeller] 2849 2850 *) Fix potential buffer overrun for EBCDIC. 2851 [Ulf Moeller] 2852 2853 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 2854 keyUsage if basicConstraints absent for a CA. 2855 [Steve Henson] 2856 2857 *) Make SMIME_write_PKCS7() write mail header values with a format that 2858 is more generally accepted (no spaces before the semicolon), since 2859 some programs can't parse those values properly otherwise. Also make 2860 sure BIO's that break lines after each write do not create invalid 2861 headers. 2862 [Richard Levitte] 2863 2864 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 2865 macros previously used would not encode an empty SEQUENCE OF 2866 and break the signature. 2867 [Steve Henson] 2868 [This change does not apply to 0.9.7.] 2869 2870 *) Zero the premaster secret after deriving the master secret in 2871 DH ciphersuites. 2872 [Steve Henson] 2873 2874 *) Add some EVP_add_digest_alias registrations (as found in 2875 OpenSSL_add_all_digests()) to SSL_library_init() 2876 aka OpenSSL_add_ssl_algorithms(). This provides improved 2877 compatibility with peers using X.509 certificates 2878 with unconventional AlgorithmIdentifier OIDs. 2879 [Bodo Moeller] 2880 2881 *) Fix for Irix with NO_ASM. 2882 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 2883 2884 *) ./config script fixes. 2885 [Ulf Moeller, Richard Levitte] 2886 2887 *) Fix 'openssl passwd -1'. 2888 [Bodo Moeller] 2889 2890 *) Change PKCS12_key_gen_asc() so it can cope with non null 2891 terminated strings whose length is passed in the passlen 2892 parameter, for example from PEM callbacks. This was done 2893 by adding an extra length parameter to asc2uni(). 2894 [Steve Henson, reported by <oddissey@samsung.co.kr>] 2895 2896 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 2897 call failed, free the DSA structure. 2898 [Bodo Moeller] 2899 2900 *) Fix to uni2asc() to cope with zero length Unicode strings. 2901 These are present in some PKCS#12 files. 2902 [Steve Henson] 2903 2904 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 2905 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 2906 when writing a 32767 byte record. 2907 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 2908 2909 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 2910 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 2911 2912 (RSA objects have a reference count access to which is protected 2913 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 2914 so they are meant to be shared between threads.) 2915 [Bodo Moeller, Geoff Thorpe; original patch submitted by 2916 "Reddie, Steven" <Steven.Reddie@ca.com>] 2917 2918 *) Fix a deadlock in CRYPTO_mem_leaks(). 2919 [Bodo Moeller] 2920 2921 *) Use better test patterns in bntest. 2922 [Ulf M�ller] 2923 2924 *) rand_win.c fix for Borland C. 2925 [Ulf M�ller] 2926 2927 *) BN_rshift bugfix for n == 0. 2928 [Bodo Moeller] 2929 2930 *) Add a 'bctest' script that checks for some known 'bc' bugs 2931 so that 'make test' does not abort just because 'bc' is broken. 2932 [Bodo Moeller] 2933 2934 *) Store verify_result within SSL_SESSION also for client side to 2935 avoid potential security hole. (Re-used sessions on the client side 2936 always resulted in verify_result==X509_V_OK, not using the original 2937 result of the server certificate verification.) 2938 [Lutz Jaenicke] 2939 2940 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 2941 SSL3_RT_APPLICATION_DATA, return 0. 2942 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 2943 [Bodo Moeller] 2944 2945 *) Fix SSL_peek: 2946 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 2947 releases, have been re-implemented by renaming the previous 2948 implementations of ssl2_read and ssl3_read to ssl2_read_internal 2949 and ssl3_read_internal, respectively, and adding 'peek' parameters 2950 to them. The new ssl[23]_{read,peek} functions are calls to 2951 ssl[23]_read_internal with the 'peek' flag set appropriately. 2952 A 'peek' parameter has also been added to ssl3_read_bytes, which 2953 does the actual work for ssl3_read_internal. 2954 [Bodo Moeller] 2955 2956 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 2957 the method-specific "init()" handler. Also clean up ex_data after 2958 calling the method-specific "finish()" handler. Previously, this was 2959 happening the other way round. 2960 [Geoff Thorpe] 2961 2962 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 2963 The previous value, 12, was not always sufficient for BN_mod_exp(). 2964 [Bodo Moeller] 2965 2966 *) Make sure that shared libraries get the internal name engine with 2967 the full version number and not just 0. This should mark the 2968 shared libraries as not backward compatible. Of course, this should 2969 be changed again when we can guarantee backward binary compatibility. 2970 [Richard Levitte] 2971 2972 *) Fix typo in get_cert_by_subject() in by_dir.c 2973 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 2974 2975 *) Rework the system to generate shared libraries: 2976 2977 - Make note of the expected extension for the shared libraries and 2978 if there is a need for symbolic links from for example libcrypto.so.0 2979 to libcrypto.so.0.9.7. There is extended info in Configure for 2980 that. 2981 2982 - Make as few rebuilds of the shared libraries as possible. 2983 2984 - Still avoid linking the OpenSSL programs with the shared libraries. 2985 2986 - When installing, install the shared libraries separately from the 2987 static ones. 2988 [Richard Levitte] 2989 2990 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 2991 2992 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 2993 and not in SSL_clear because the latter is also used by the 2994 accept/connect functions; previously, the settings made by 2995 SSL_set_read_ahead would be lost during the handshake. 2996 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 2997 2998 *) Correct util/mkdef.pl to be selective about disabled algorithms. 2999 Previously, it would create entries for disableed algorithms no 3000 matter what. 3001 [Richard Levitte] 3002 3003 *) Added several new manual pages for SSL_* function. 3004 [Lutz Jaenicke] 3005 3006 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 3007 3008 *) In ssl23_get_client_hello, generate an error message when faced 3009 with an initial SSL 3.0/TLS record that is too small to contain the 3010 first two bytes of the ClientHello message, i.e. client_version. 3011 (Note that this is a pathologic case that probably has never happened 3012 in real life.) The previous approach was to use the version number 3013 from the record header as a substitute; but our protocol choice 3014 should not depend on that one because it is not authenticated 3015 by the Finished messages. 3016 [Bodo Moeller] 3017 3018 *) More robust randomness gathering functions for Windows. 3019 [Jeffrey Altman <jaltman@columbia.edu>] 3020 3021 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 3022 not set then we don't setup the error code for issuer check errors 3023 to avoid possibly overwriting other errors which the callback does 3024 handle. If an application does set the flag then we assume it knows 3025 what it is doing and can handle the new informational codes 3026 appropriately. 3027 [Steve Henson] 3028 3029 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 3030 a general "ANY" type, as such it should be able to decode anything 3031 including tagged types. However it didn't check the class so it would 3032 wrongly interpret tagged types in the same way as their universal 3033 counterpart and unknown types were just rejected. Changed so that the 3034 tagged and unknown types are handled in the same way as a SEQUENCE: 3035 that is the encoding is stored intact. There is also a new type 3036 "V_ASN1_OTHER" which is used when the class is not universal, in this 3037 case we have no idea what the actual type is so we just lump them all 3038 together. 3039 [Steve Henson] 3040 3041 *) On VMS, stdout may very well lead to a file that is written to 3042 in a record-oriented fashion. That means that every write() will 3043 write a separate record, which will be read separately by the 3044 programs trying to read from it. This can be very confusing. 3045 3046 The solution is to put a BIO filter in the way that will buffer 3047 text until a linefeed is reached, and then write everything a 3048 line at a time, so every record written will be an actual line, 3049 not chunks of lines and not (usually doesn't happen, but I've 3050 seen it once) several lines in one record. BIO_f_linebuffer() is 3051 the answer. 3052 3053 Currently, it's a VMS-only method, because that's where it has 3054 been tested well enough. 3055 [Richard Levitte] 3056 3057 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 3058 it can return incorrect results. 3059 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 3060 but it was in 0.9.6-beta[12].) 3061 [Bodo Moeller] 3062 3063 *) Disable the check for content being present when verifying detached 3064 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 3065 include zero length content when signing messages. 3066 [Steve Henson] 3067 3068 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 3069 BIO_ctrl (for BIO pairs). 3070 [Bodo M�ller] 3071 3072 *) Add DSO method for VMS. 3073 [Richard Levitte] 3074 3075 *) Bug fix: Montgomery multiplication could produce results with the 3076 wrong sign. 3077 [Ulf M�ller] 3078 3079 *) Add RPM specification openssl.spec and modify it to build three 3080 packages. The default package contains applications, application 3081 documentation and run-time libraries. The devel package contains 3082 include files, static libraries and function documentation. The 3083 doc package contains the contents of the doc directory. The original 3084 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 3085 [Richard Levitte] 3086 3087 *) Add a large number of documentation files for many SSL routines. 3088 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 3089 3090 *) Add a configuration entry for Sony News 4. 3091 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 3092 3093 *) Don't set the two most significant bits to one when generating a 3094 random number < q in the DSA library. 3095 [Ulf M�ller] 3096 3097 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 3098 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 3099 the underlying transport is blocking) if a handshake took place. 3100 (The default behaviour is needed by applications such as s_client 3101 and s_server that use select() to determine when to use SSL_read; 3102 but for applications that know in advance when to expect data, it 3103 just makes things more complicated.) 3104 [Bodo Moeller] 3105 3106 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 3107 from EGD. 3108 [Ben Laurie] 3109 3110 *) Add a few more EBCDIC conditionals that make `req' and `x509' 3111 work better on such systems. 3112 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 3113 3114 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 3115 Update PKCS12_parse() so it copies the friendlyName and the 3116 keyid to the certificates aux info. 3117 [Steve Henson] 3118 3119 *) Fix bug in PKCS7_verify() which caused an infinite loop 3120 if there was more than one signature. 3121 [Sven Uszpelkat <su@celocom.de>] 3122 3123 *) Major change in util/mkdef.pl to include extra information 3124 about each symbol, as well as presentig variables as well 3125 as functions. This change means that there's n more need 3126 to rebuild the .num files when some algorithms are excluded. 3127 [Richard Levitte] 3128 3129 *) Allow the verify time to be set by an application, 3130 rather than always using the current time. 3131 [Steve Henson] 3132 3133 *) Phase 2 verify code reorganisation. The certificate 3134 verify code now looks up an issuer certificate by a 3135 number of criteria: subject name, authority key id 3136 and key usage. It also verifies self signed certificates 3137 by the same criteria. The main comparison function is 3138 X509_check_issued() which performs these checks. 3139 3140 Lot of changes were necessary in order to support this 3141 without completely rewriting the lookup code. 3142 3143 Authority and subject key identifier are now cached. 3144 3145 The LHASH 'certs' is X509_STORE has now been replaced 3146 by a STACK_OF(X509_OBJECT). This is mainly because an 3147 LHASH can't store or retrieve multiple objects with 3148 the same hash value. 3149 3150 As a result various functions (which were all internal 3151 use only) have changed to handle the new X509_STORE 3152 structure. This will break anything that messed round 3153 with X509_STORE internally. 3154 3155 The functions X509_STORE_add_cert() now checks for an 3156 exact match, rather than just subject name. 3157 3158 The X509_STORE API doesn't directly support the retrieval 3159 of multiple certificates matching a given criteria, however 3160 this can be worked round by performing a lookup first 3161 (which will fill the cache with candidate certificates) 3162 and then examining the cache for matches. This is probably 3163 the best we can do without throwing out X509_LOOKUP 3164 entirely (maybe later...). 3165 3166 The X509_VERIFY_CTX structure has been enhanced considerably. 3167 3168 All certificate lookup operations now go via a get_issuer() 3169 callback. Although this currently uses an X509_STORE it 3170 can be replaced by custom lookups. This is a simple way 3171 to bypass the X509_STORE hackery necessary to make this 3172 work and makes it possible to use more efficient techniques 3173 in future. A very simple version which uses a simple 3174 STACK for its trusted certificate store is also provided 3175 using X509_STORE_CTX_trusted_stack(). 3176 3177 The verify_cb() and verify() callbacks now have equivalents 3178 in the X509_STORE_CTX structure. 3179 3180 X509_STORE_CTX also has a 'flags' field which can be used 3181 to customise the verify behaviour. 3182 [Steve Henson] 3183 3184 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 3185 excludes S/MIME capabilities. 3186 [Steve Henson] 3187 3188 *) When a certificate request is read in keep a copy of the 3189 original encoding of the signed data and use it when outputing 3190 again. Signatures then use the original encoding rather than 3191 a decoded, encoded version which may cause problems if the 3192 request is improperly encoded. 3193 [Steve Henson] 3194 3195 *) For consistency with other BIO_puts implementations, call 3196 buffer_write(b, ...) directly in buffer_puts instead of calling 3197 BIO_write(b, ...). 3198 3199 In BIO_puts, increment b->num_write as in BIO_write. 3200 [Peter.Sylvester@EdelWeb.fr] 3201 3202 *) Fix BN_mul_word for the case where the word is 0. (We have to use 3203 BN_zero, we may not return a BIGNUM with an array consisting of 3204 words set to zero.) 3205 [Bodo Moeller] 3206 3207 *) Avoid calling abort() from within the library when problems are 3208 detected, except if preprocessor symbols have been defined 3209 (such as REF_CHECK, BN_DEBUG etc.). 3210 [Bodo Moeller] 3211 3212 *) New openssl application 'rsautl'. This utility can be 3213 used for low level RSA operations. DER public key 3214 BIO/fp routines also added. 3215 [Steve Henson] 3216 3217 *) New Configure entry and patches for compiling on QNX 4. 3218 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 3219 3220 *) A demo state-machine implementation was sponsored by 3221 Nuron (http://www.nuron.com/) and is now available in 3222 demos/state_machine. 3223 [Ben Laurie] 3224 3225 *) New options added to the 'dgst' utility for signature 3226 generation and verification. 3227 [Steve Henson] 3228 3229 *) Unrecognized PKCS#7 content types are now handled via a 3230 catch all ASN1_TYPE structure. This allows unsupported 3231 types to be stored as a "blob" and an application can 3232 encode and decode it manually. 3233 [Steve Henson] 3234 3235 *) Fix various signed/unsigned issues to make a_strex.c 3236 compile under VC++. 3237 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 3238 3239 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 3240 length if passed a buffer. ASN1_INTEGER_to_BN failed 3241 if passed a NULL BN and its argument was negative. 3242 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 3243 3244 *) Modification to PKCS#7 encoding routines to output definite 3245 length encoding. Since currently the whole structures are in 3246 memory there's not real point in using indefinite length 3247 constructed encoding. However if OpenSSL is compiled with 3248 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 3249 [Steve Henson] 3250 3251 *) Added BIO_vprintf() and BIO_vsnprintf(). 3252 [Richard Levitte] 3253 3254 *) Added more prefixes to parse for in the the strings written 3255 through a logging bio, to cover all the levels that are available 3256 through syslog. The prefixes are now: 3257 3258 PANIC, EMERG, EMR => LOG_EMERG 3259 ALERT, ALR => LOG_ALERT 3260 CRIT, CRI => LOG_CRIT 3261 ERROR, ERR => LOG_ERR 3262 WARNING, WARN, WAR => LOG_WARNING 3263 NOTICE, NOTE, NOT => LOG_NOTICE 3264 INFO, INF => LOG_INFO 3265 DEBUG, DBG => LOG_DEBUG 3266 3267 and as before, if none of those prefixes are present at the 3268 beginning of the string, LOG_ERR is chosen. 3269 3270 On Win32, the LOG_* levels are mapped according to this: 3271 3272 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 3273 LOG_WARNING => EVENTLOG_WARNING_TYPE 3274 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 3275 3276 [Richard Levitte] 3277 3278 *) Made it possible to reconfigure with just the configuration 3279 argument "reconf" or "reconfigure". The command line arguments 3280 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 3281 and are retrieved from there when reconfiguring. 3282 [Richard Levitte] 3283 3284 *) MD4 implemented. 3285 [Assar Westerlund <assar@sics.se>, Richard Levitte] 3286 3287 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 3288 [Richard Levitte] 3289 3290 *) The obj_dat.pl script was messing up the sorting of object 3291 names. The reason was that it compared the quoted version 3292 of strings as a result "OCSP" > "OCSP Signing" because 3293 " > SPACE. Changed script to store unquoted versions of 3294 names and add quotes on output. It was also omitting some 3295 names from the lookup table if they were given a default 3296 value (that is if SN is missing it is given the same 3297 value as LN and vice versa), these are now added on the 3298 grounds that if an object has a name we should be able to 3299 look it up. Finally added warning output when duplicate 3300 short or long names are found. 3301 [Steve Henson] 3302 3303 *) Changes needed for Tandem NSK. 3304 [Scott Uroff <scott@xypro.com>] 3305 3306 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 3307 RSA_padding_check_SSLv23(), special padding was never detected 3308 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 3309 version rollback attacks was not effective. 3310 3311 In s23_clnt.c, don't use special rollback-attack detection padding 3312 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 3313 client; similarly, in s23_srvr.c, don't do the rollback check if 3314 SSL 2.0 is the only protocol enabled in the server. 3315 [Bodo Moeller] 3316 3317 *) Make it possible to get hexdumps of unprintable data with 'openssl 3318 asn1parse'. By implication, the functions ASN1_parse_dump() and 3319 BIO_dump_indent() are added. 3320 [Richard Levitte] 3321 3322 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 3323 these print out strings and name structures based on various 3324 flags including RFC2253 support and proper handling of 3325 multibyte characters. Added options to the 'x509' utility 3326 to allow the various flags to be set. 3327 [Steve Henson] 3328 3329 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 3330 Also change the functions X509_cmp_current_time() and 3331 X509_gmtime_adj() work with an ASN1_TIME structure, 3332 this will enable certificates using GeneralizedTime in validity 3333 dates to be checked. 3334 [Steve Henson] 3335 3336 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 3337 negative public key encodings) on by default, 3338 NO_NEG_PUBKEY_BUG can be set to disable it. 3339 [Steve Henson] 3340 3341 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 3342 content octets. An i2c_ASN1_OBJECT is unnecessary because 3343 the encoding can be trivially obtained from the structure. 3344 [Steve Henson] 3345 3346 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 3347 not read locks (CRYPTO_r_[un]lock). 3348 [Bodo Moeller] 3349 3350 *) A first attempt at creating official support for shared 3351 libraries through configuration. I've kept it so the 3352 default is static libraries only, and the OpenSSL programs 3353 are always statically linked for now, but there are 3354 preparations for dynamic linking in place. 3355 This has been tested on Linux and Tru64. 3356 [Richard Levitte] 3357 3358 *) Randomness polling function for Win9x, as described in: 3359 Peter Gutmann, Software Generation of Practically Strong 3360 Random Numbers. 3361 [Ulf M�ller] 3362 3363 *) Fix so PRNG is seeded in req if using an already existing 3364 DSA key. 3365 [Steve Henson] 3366 3367 *) New options to smime application. -inform and -outform 3368 allow alternative formats for the S/MIME message including 3369 PEM and DER. The -content option allows the content to be 3370 specified separately. This should allow things like Netscape 3371 form signing output easier to verify. 3372 [Steve Henson] 3373 3374 *) Fix the ASN1 encoding of tags using the 'long form'. 3375 [Steve Henson] 3376 3377 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 3378 STRING types. These convert content octets to and from the 3379 underlying type. The actual tag and length octets are 3380 already assumed to have been read in and checked. These 3381 are needed because all other string types have virtually 3382 identical handling apart from the tag. By having versions 3383 of the ASN1 functions that just operate on content octets 3384 IMPLICIT tagging can be handled properly. It also allows 3385 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 3386 and ASN1_INTEGER are identical apart from the tag. 3387 [Steve Henson] 3388 3389 *) Change the handling of OID objects as follows: 3390 3391 - New object identifiers are inserted in objects.txt, following 3392 the syntax given in objects.README. 3393 - objects.pl is used to process obj_mac.num and create a new 3394 obj_mac.h. 3395 - obj_dat.pl is used to create a new obj_dat.h, using the data in 3396 obj_mac.h. 3397 3398 This is currently kind of a hack, and the perl code in objects.pl 3399 isn't very elegant, but it works as I intended. The simplest way 3400 to check that it worked correctly is to look in obj_dat.h and 3401 check the array nid_objs and make sure the objects haven't moved 3402 around (this is important!). Additions are OK, as well as 3403 consistent name changes. 3404 [Richard Levitte] 3405 3406 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 3407 [Bodo Moeller] 3408 3409 *) Addition of the command line parameter '-rand file' to 'openssl req'. 3410 The given file adds to whatever has already been seeded into the 3411 random pool through the RANDFILE configuration file option or 3412 environment variable, or the default random state file. 3413 [Richard Levitte] 3414 3415 *) mkstack.pl now sorts each macro group into lexical order. 3416 Previously the output order depended on the order the files 3417 appeared in the directory, resulting in needless rewriting 3418 of safestack.h . 3419 [Steve Henson] 3420 3421 *) Patches to make OpenSSL compile under Win32 again. Mostly 3422 work arounds for the VC++ problem that it treats func() as 3423 func(void). Also stripped out the parts of mkdef.pl that 3424 added extra typesafe functions: these no longer exist. 3425 [Steve Henson] 3426 3427 *) Reorganisation of the stack code. The macros are now all 3428 collected in safestack.h . Each macro is defined in terms of 3429 a "stack macro" of the form SKM_<name>(type, a, b). The 3430 DEBUG_SAFESTACK is now handled in terms of function casts, 3431 this has the advantage of retaining type safety without the 3432 use of additional functions. If DEBUG_SAFESTACK is not defined 3433 then the non typesafe macros are used instead. Also modified the 3434 mkstack.pl script to handle the new form. Needs testing to see 3435 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 3436 the default if no major problems. Similar behaviour for ASN1_SET_OF 3437 and PKCS12_STACK_OF. 3438 [Steve Henson] 3439 3440 *) When some versions of IIS use the 'NET' form of private key the 3441 key derivation algorithm is different. Normally MD5(password) is 3442 used as a 128 bit RC4 key. In the modified case 3443 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 3444 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 3445 as the old Netscape_RSA functions except they have an additional 3446 'sgckey' parameter which uses the modified algorithm. Also added 3447 an -sgckey command line option to the rsa utility. Thanks to 3448 Adrian Peck <bertie@ncipher.com> for posting details of the modified 3449 algorithm to openssl-dev. 3450 [Steve Henson] 3451 3452 *) The evp_local.h macros were using 'c.##kname' which resulted in 3453 invalid expansion on some systems (SCO 5.0.5 for example). 3454 Corrected to 'c.kname'. 3455 [Phillip Porch <root@theporch.com>] 3456 3457 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 3458 a STACK of email addresses from a certificate or request, these look 3459 in the subject name and the subject alternative name extensions and 3460 omit any duplicate addresses. 3461 [Steve Henson] 3462 3463 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 3464 This makes DSA verification about 2 % faster. 3465 [Bodo Moeller] 3466 3467 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 3468 (meaning that now 2^5 values will be precomputed, which is only 4 KB 3469 plus overhead for 1024 bit moduli). 3470 This makes exponentiations about 0.5 % faster for 1024 bit 3471 exponents (as measured by "openssl speed rsa2048"). 3472 [Bodo Moeller] 3473 3474 *) Rename memory handling macros to avoid conflicts with other 3475 software: 3476 Malloc => OPENSSL_malloc 3477 Malloc_locked => OPENSSL_malloc_locked 3478 Realloc => OPENSSL_realloc 3479 Free => OPENSSL_free 3480 [Richard Levitte] 3481 3482 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 3483 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 3484 [Bodo Moeller] 3485 3486 *) CygWin32 support. 3487 [John Jarvie <jjarvie@newsguy.com>] 3488 3489 *) The type-safe stack code has been rejigged. It is now only compiled 3490 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 3491 by default all type-specific stack functions are "#define"d back to 3492 standard stack functions. This results in more streamlined output 3493 but retains the type-safety checking possibilities of the original 3494 approach. 3495 [Geoff Thorpe] 3496 3497 *) The STACK code has been cleaned up, and certain type declarations 3498 that didn't make a lot of sense have been brought in line. This has 3499 also involved a cleanup of sorts in safestack.h to more correctly 3500 map type-safe stack functions onto their plain stack counterparts. 3501 This work has also resulted in a variety of "const"ifications of 3502 lots of the code, especially "_cmp" operations which should normally 3503 be prototyped with "const" parameters anyway. 3504 [Geoff Thorpe] 3505 3506 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 3507 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 3508 (The PRNG state consists of two parts, the large pool 'state' and 'md', 3509 where all of 'md' is used each time the PRNG is used, but 'state' 3510 is used only indexed by a cyclic counter. As entropy may not be 3511 well distributed from the beginning, 'md' is important as a 3512 chaining variable. However, the output function chains only half 3513 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 3514 all of 'md', and seeding with STATE_SIZE dummy bytes will result 3515 in all of 'state' being rewritten, with the new values depending 3516 on virtually all of 'md'. This overcomes the 80 bit limitation.) 3517 [Bodo Moeller] 3518 3519 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 3520 the handshake is continued after ssl_verify_cert_chain(); 3521 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 3522 can lead to 'unexplainable' connection aborts later. 3523 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 3524 3525 *) Major EVP API cipher revision. 3526 Add hooks for extra EVP features. This allows various cipher 3527 parameters to be set in the EVP interface. Support added for variable 3528 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 3529 setting of RC2 and RC5 parameters. 3530 3531 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 3532 ciphers. 3533 3534 Remove lots of duplicated code from the EVP library. For example *every* 3535 cipher init() function handles the 'iv' in the same way according to the 3536 cipher mode. They also all do nothing if the 'key' parameter is NULL and 3537 for CFB and OFB modes they zero ctx->num. 3538 3539 New functionality allows removal of S/MIME code RC2 hack. 3540 3541 Most of the routines have the same form and so can be declared in terms 3542 of macros. 3543 3544 By shifting this to the top level EVP_CipherInit() it can be removed from 3545 all individual ciphers. If the cipher wants to handle IVs or keys 3546 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 3547 flags. 3548 3549 Change lots of functions like EVP_EncryptUpdate() to now return a 3550 value: although software versions of the algorithms cannot fail 3551 any installed hardware versions can. 3552 [Steve Henson] 3553 3554 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 3555 this option is set, tolerate broken clients that send the negotiated 3556 protocol version number instead of the requested protocol version 3557 number. 3558 [Bodo Moeller] 3559 3560 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 3561 i.e. non-zero for export ciphersuites, zero otherwise. 3562 Previous versions had this flag inverted, inconsistent with 3563 rsa_tmp_cb (..._TMP_RSA_CB). 3564 [Bodo Moeller; problem reported by Amit Chopra] 3565 3566 *) Add missing DSA library text string. Work around for some IIS 3567 key files with invalid SEQUENCE encoding. 3568 [Steve Henson] 3569 3570 *) Add a document (doc/standards.txt) that list all kinds of standards 3571 and so on that are implemented in OpenSSL. 3572 [Richard Levitte] 3573 3574 *) Enhance c_rehash script. Old version would mishandle certificates 3575 with the same subject name hash and wouldn't handle CRLs at all. 3576 Added -fingerprint option to crl utility, to support new c_rehash 3577 features. 3578 [Steve Henson] 3579 3580 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 3581 [Ulf M�ller] 3582 3583 *) Fix for SSL server purpose checking. Server checking was 3584 rejecting certificates which had extended key usage present 3585 but no ssl client purpose. 3586 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 3587 3588 *) Make PKCS#12 code work with no password. The PKCS#12 spec 3589 is a little unclear about how a blank password is handled. 3590 Since the password in encoded as a BMPString with terminating 3591 double NULL a zero length password would end up as just the 3592 double NULL. However no password at all is different and is 3593 handled differently in the PKCS#12 key generation code. NS 3594 treats a blank password as zero length. MSIE treats it as no 3595 password on export: but it will try both on import. We now do 3596 the same: PKCS12_parse() tries zero length and no password if 3597 the password is set to "" or NULL (NULL is now a valid password: 3598 it wasn't before) as does the pkcs12 application. 3599 [Steve Henson] 3600 3601 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 3602 perror when PEM_read_bio_X509_REQ fails, the error message must 3603 be obtained from the error queue. 3604 [Bodo Moeller] 3605 3606 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 3607 it in ERR_remove_state if appropriate, and change ERR_get_state 3608 accordingly to avoid race conditions (this is necessary because 3609 thread_hash is no longer constant once set). 3610 [Bodo Moeller] 3611 3612 *) Bugfix for linux-elf makefile.one. 3613 [Ulf M�ller] 3614 3615 *) RSA_get_default_method() will now cause a default 3616 RSA_METHOD to be chosen if one doesn't exist already. 3617 Previously this was only set during a call to RSA_new() 3618 or RSA_new_method(NULL) meaning it was possible for 3619 RSA_get_default_method() to return NULL. 3620 [Geoff Thorpe] 3621 3622 *) Added native name translation to the existing DSO code 3623 that will convert (if the flag to do so is set) filenames 3624 that are sufficiently small and have no path information 3625 into a canonical native form. Eg. "blah" converted to 3626 "libblah.so" or "blah.dll" etc. 3627 [Geoff Thorpe] 3628 3629 *) New function ERR_error_string_n(e, buf, len) which is like 3630 ERR_error_string(e, buf), but writes at most 'len' bytes 3631 including the 0 terminator. For ERR_error_string_n, 'buf' 3632 may not be NULL. 3633 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 3634 3635 *) CONF library reworked to become more general. A new CONF 3636 configuration file reader "class" is implemented as well as a 3637 new functions (NCONF_*, for "New CONF") to handle it. The now 3638 old CONF_* functions are still there, but are reimplemented to 3639 work in terms of the new functions. Also, a set of functions 3640 to handle the internal storage of the configuration data is 3641 provided to make it easier to write new configuration file 3642 reader "classes" (I can definitely see something reading a 3643 configuration file in XML format, for example), called _CONF_*, 3644 or "the configuration storage API"... 3645 3646 The new configuration file reading functions are: 3647 3648 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 3649 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 3650 3651 NCONF_default, NCONF_WIN32 3652 3653 NCONF_dump_fp, NCONF_dump_bio 3654 3655 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 3656 NCONF_new creates a new CONF object. This works in the same way 3657 as other interfaces in OpenSSL, like the BIO interface. 3658 NCONF_dump_* dump the internal storage of the configuration file, 3659 which is useful for debugging. All other functions take the same 3660 arguments as the old CONF_* functions wth the exception of the 3661 first that must be a `CONF *' instead of a `LHASH *'. 3662 3663 To make it easer to use the new classes with the old CONF_* functions, 3664 the function CONF_set_default_method is provided. 3665 [Richard Levitte] 3666 3667 *) Add '-tls1' option to 'openssl ciphers', which was already 3668 mentioned in the documentation but had not been implemented. 3669 (This option is not yet really useful because even the additional 3670 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 3671 [Bodo Moeller] 3672 3673 *) Initial DSO code added into libcrypto for letting OpenSSL (and 3674 OpenSSL-based applications) load shared libraries and bind to 3675 them in a portable way. 3676 [Geoff Thorpe, with contributions from Richard Levitte] 3677 3678 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 3679 3680 *) Make sure _lrotl and _lrotr are only used with MSVC. 3681 3682 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 3683 (the default implementation of RAND_status). 3684 3685 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 3686 to '-clrext' (= clear extensions), as intended and documented. 3687 [Bodo Moeller; inconsistency pointed out by Michael Attili 3688 <attili@amaxo.com>] 3689 3690 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 3691 was larger than the MD block size. 3692 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 3693 3694 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 3695 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 3696 using the passed key: if the passed key was a private key the result 3697 of X509_print(), for example, would be to print out all the private key 3698 components. 3699 [Steve Henson] 3700 3701 *) des_quad_cksum() byte order bug fix. 3702 [Ulf M�ller, using the problem description in krb4-0.9.7, where 3703 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 3704 3705 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 3706 discouraged. 3707 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 3708 3709 *) For easily testing in shell scripts whether some command 3710 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 3711 returns with exit code 0 iff no command of the given name is available. 3712 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 3713 the output goes to stdout and nothing is printed to stderr. 3714 Additional arguments are always ignored. 3715 3716 Since for each cipher there is a command of the same name, 3717 the 'no-cipher' compilation switches can be tested this way. 3718 3719 ('openssl no-XXX' is not able to detect pseudo-commands such 3720 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 3721 [Bodo Moeller] 3722 3723 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 3724 [Bodo Moeller] 3725 3726 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 3727 is set; it will be thrown away anyway because each handshake creates 3728 its own key. 3729 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 3730 to parameters -- in previous versions (since OpenSSL 0.9.3) the 3731 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 3732 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 3733 [Bodo Moeller] 3734 3735 *) New s_client option -ign_eof: EOF at stdin is ignored, and 3736 'Q' and 'R' lose their special meanings (quit/renegotiate). 3737 This is part of what -quiet does; unlike -quiet, -ign_eof 3738 does not suppress any output. 3739 [Richard Levitte] 3740 3741 *) Add compatibility options to the purpose and trust code. The 3742 purpose X509_PURPOSE_ANY is "any purpose" which automatically 3743 accepts a certificate or CA, this was the previous behaviour, 3744 with all the associated security issues. 3745 3746 X509_TRUST_COMPAT is the old trust behaviour: only and 3747 automatically trust self signed roots in certificate store. A 3748 new trust setting X509_TRUST_DEFAULT is used to specify that 3749 a purpose has no associated trust setting and it should instead 3750 use the value in the default purpose. 3751 [Steve Henson] 3752 3753 *) Fix the PKCS#8 DSA private key code so it decodes keys again 3754 and fix a memory leak. 3755 [Steve Henson] 3756 3757 *) In util/mkerr.pl (which implements 'make errors'), preserve 3758 reason strings from the previous version of the .c file, as 3759 the default to have only downcase letters (and digits) in 3760 automatically generated reasons codes is not always appropriate. 3761 [Bodo Moeller] 3762 3763 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 3764 using strerror. Previously, ERR_reason_error_string() returned 3765 library names as reason strings for SYSerr; but SYSerr is a special 3766 case where small numbers are errno values, not library numbers. 3767 [Bodo Moeller] 3768 3769 *) Add '-dsaparam' option to 'openssl dhparam' application. This 3770 converts DSA parameters into DH parameters. (When creating parameters, 3771 DSA_generate_parameters is used.) 3772 [Bodo Moeller] 3773 3774 *) Include 'length' (recommended exponent length) in C code generated 3775 by 'openssl dhparam -C'. 3776 [Bodo Moeller] 3777 3778 *) The second argument to set_label in perlasm was already being used 3779 so couldn't be used as a "file scope" flag. Moved to third argument 3780 which was free. 3781 [Steve Henson] 3782 3783 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 3784 instead of RAND_bytes for encryption IVs and salts. 3785 [Bodo Moeller] 3786 3787 *) Include RAND_status() into RAND_METHOD instead of implementing 3788 it only for md_rand.c Otherwise replacing the PRNG by calling 3789 RAND_set_rand_method would be impossible. 3790 [Bodo Moeller] 3791 3792 *) Don't let DSA_generate_key() enter an infinite loop if the random 3793 number generation fails. 3794 [Bodo Moeller] 3795 3796 *) New 'rand' application for creating pseudo-random output. 3797 [Bodo Moeller] 3798 3799 *) Added configuration support for Linux/IA64 3800 [Rolf Haberrecker <rolf@suse.de>] 3801 3802 *) Assembler module support for Mingw32. 3803 [Ulf M�ller] 3804 3805 *) Shared library support for HPUX (in shlib/). 3806 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 3807 3808 *) Shared library support for Solaris gcc. 3809 [Lutz Behnke <behnke@trustcenter.de>] 3810 3811 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 3812 3813 *) PKCS7_encrypt() was adding text MIME headers twice because they 3814 were added manually and by SMIME_crlf_copy(). 3815 [Steve Henson] 3816 3817 *) In bntest.c don't call BN_rand with zero bits argument. 3818 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 3819 3820 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 3821 case was implemented. This caused BN_div_recp() to fail occasionally. 3822 [Ulf M�ller] 3823 3824 *) Add an optional second argument to the set_label() in the perl 3825 assembly language builder. If this argument exists and is set 3826 to 1 it signals that the assembler should use a symbol whose 3827 scope is the entire file, not just the current function. This 3828 is needed with MASM which uses the format label:: for this scope. 3829 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 3830 3831 *) Change the ASN1 types so they are typedefs by default. Before 3832 almost all types were #define'd to ASN1_STRING which was causing 3833 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 3834 for example. 3835 [Steve Henson] 3836 3837 *) Change names of new functions to the new get1/get0 naming 3838 convention: After 'get1', the caller owns a reference count 3839 and has to call ..._free; 'get0' returns a pointer to some 3840 data structure without incrementing reference counters. 3841 (Some of the existing 'get' functions increment a reference 3842 counter, some don't.) 3843 Similarly, 'set1' and 'add1' functions increase reference 3844 counters or duplicate objects. 3845 [Steve Henson] 3846 3847 *) Allow for the possibility of temp RSA key generation failure: 3848 the code used to assume it always worked and crashed on failure. 3849 [Steve Henson] 3850 3851 *) Fix potential buffer overrun problem in BIO_printf(). 3852 [Ulf M�ller, using public domain code by Patrick Powell; problem 3853 pointed out by David Sacerdote <das33@cornell.edu>] 3854 3855 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 3856 RAND_egd() and RAND_status(). In the command line application, 3857 the EGD socket can be specified like a seed file using RANDFILE 3858 or -rand. 3859 [Ulf M�ller] 3860 3861 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 3862 Some CAs (e.g. Verisign) distribute certificates in this form. 3863 [Steve Henson] 3864 3865 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 3866 list to exclude them. This means that no special compilation option 3867 is needed to use anonymous DH: it just needs to be included in the 3868 cipher list. 3869 [Steve Henson] 3870 3871 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 3872 EVP_MD_type. The old functionality is available in a new macro called 3873 EVP_MD_md(). Change code that uses it and update docs. 3874 [Steve Henson] 3875 3876 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 3877 where the 'void *' argument is replaced by a function pointer argument. 3878 Previously 'void *' was abused to point to functions, which works on 3879 many platforms, but is not correct. As these functions are usually 3880 called by macros defined in OpenSSL header files, most source code 3881 should work without changes. 3882 [Richard Levitte] 3883 3884 *) <openssl/opensslconf.h> (which is created by Configure) now contains 3885 sections with information on -D... compiler switches used for 3886 compiling the library so that applications can see them. To enable 3887 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 3888 must be defined. E.g., 3889 #define OPENSSL_ALGORITHM_DEFINES 3890 #include <openssl/opensslconf.h> 3891 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 3892 [Richard Levitte, Ulf and Bodo M�ller] 3893 3894 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 3895 record layer. 3896 [Bodo Moeller] 3897 3898 *) Change the 'other' type in certificate aux info to a STACK_OF 3899 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 3900 the required ASN1 format: arbitrary types determined by an OID. 3901 [Steve Henson] 3902 3903 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 3904 argument to 'req'. This is not because the function is newer or 3905 better than others it just uses the work 'NEW' in the certificate 3906 request header lines. Some software needs this. 3907 [Steve Henson] 3908 3909 *) Reorganise password command line arguments: now passwords can be 3910 obtained from various sources. Delete the PEM_cb function and make 3911 it the default behaviour: i.e. if the callback is NULL and the 3912 usrdata argument is not NULL interpret it as a null terminated pass 3913 phrase. If usrdata and the callback are NULL then the pass phrase 3914 is prompted for as usual. 3915 [Steve Henson] 3916 3917 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 3918 the support is automatically enabled. The resulting binaries will 3919 autodetect the card and use it if present. 3920 [Ben Laurie and Compaq Inc.] 3921 3922 *) Work around for Netscape hang bug. This sends certificate request 3923 and server done in one record. Since this is perfectly legal in the 3924 SSL/TLS protocol it isn't a "bug" option and is on by default. See 3925 the bugs/SSLv3 entry for more info. 3926 [Steve Henson] 3927 3928 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 3929 [Andy Polyakov] 3930 3931 *) Add -rand argument to smime and pkcs12 applications and read/write 3932 of seed file. 3933 [Steve Henson] 3934 3935 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 3936 [Bodo Moeller] 3937 3938 *) Add command line password options to the remaining applications. 3939 [Steve Henson] 3940 3941 *) Bug fix for BN_div_recp() for numerators with an even number of 3942 bits. 3943 [Ulf M�ller] 3944 3945 *) More tests in bntest.c, and changed test_bn output. 3946 [Ulf M�ller] 3947 3948 *) ./config recognizes MacOS X now. 3949 [Andy Polyakov] 3950 3951 *) Bug fix for BN_div() when the first words of num and divsor are 3952 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 3953 [Ulf M�ller] 3954 3955 *) Add support for various broken PKCS#8 formats, and command line 3956 options to produce them. 3957 [Steve Henson] 3958 3959 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 3960 get temporary BIGNUMs from a BN_CTX. 3961 [Ulf M�ller] 3962 3963 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 3964 for p == 0. 3965 [Ulf M�ller] 3966 3967 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 3968 include a #define from the old name to the new. The original intent 3969 was that statically linked binaries could for example just call 3970 SSLeay_add_all_ciphers() to just add ciphers to the table and not 3971 link with digests. This never worked becayse SSLeay_add_all_digests() 3972 and SSLeay_add_all_ciphers() were in the same source file so calling 3973 one would link with the other. They are now in separate source files. 3974 [Steve Henson] 3975 3976 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 3977 [Steve Henson] 3978 3979 *) Use a less unusual form of the Miller-Rabin primality test (it used 3980 a binary algorithm for exponentiation integrated into the Miller-Rabin 3981 loop, our standard modexp algorithms are faster). 3982 [Bodo Moeller] 3983 3984 *) Support for the EBCDIC character set completed. 3985 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 3986 3987 *) Source code cleanups: use const where appropriate, eliminate casts, 3988 use void * instead of char * in lhash. 3989 [Ulf M�ller] 3990 3991 *) Bugfix: ssl3_send_server_key_exchange was not restartable 3992 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 3993 this the server could overwrite ephemeral keys that the client 3994 has already seen). 3995 [Bodo Moeller] 3996 3997 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 3998 using 50 iterations of the Rabin-Miller test. 3999 4000 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 4001 iterations of the Rabin-Miller test as required by the appendix 4002 to FIPS PUB 186[-1]) instead of DSA_is_prime. 4003 As BN_is_prime_fasttest includes trial division, DSA parameter 4004 generation becomes much faster. 4005 4006 This implies a change for the callback functions in DSA_is_prime 4007 and DSA_generate_parameters: The callback function is called once 4008 for each positive witness in the Rabin-Miller test, not just 4009 occasionally in the inner loop; and the parameters to the 4010 callback function now provide an iteration count for the outer 4011 loop rather than for the current invocation of the inner loop. 4012 DSA_generate_parameters additionally can call the callback 4013 function with an 'iteration count' of -1, meaning that a 4014 candidate has passed the trial division test (when q is generated 4015 from an application-provided seed, trial division is skipped). 4016 [Bodo Moeller] 4017 4018 *) New function BN_is_prime_fasttest that optionally does trial 4019 division before starting the Rabin-Miller test and has 4020 an additional BN_CTX * argument (whereas BN_is_prime always 4021 has to allocate at least one BN_CTX). 4022 'callback(1, -1, cb_arg)' is called when a number has passed the 4023 trial division stage. 4024 [Bodo Moeller] 4025 4026 *) Fix for bug in CRL encoding. The validity dates weren't being handled 4027 as ASN1_TIME. 4028 [Steve Henson] 4029 4030 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 4031 [Steve Henson] 4032 4033 *) New function BN_pseudo_rand(). 4034 [Ulf M�ller] 4035 4036 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 4037 bignum version of BN_from_montgomery() with the working code from 4038 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 4039 the comments. 4040 [Ulf M�ller] 4041 4042 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 4043 made it impossible to use the same SSL_SESSION data structure in 4044 SSL2 clients in multiple threads. 4045 [Bodo Moeller] 4046 4047 *) The return value of RAND_load_file() no longer counts bytes obtained 4048 by stat(). RAND_load_file(..., -1) is new and uses the complete file 4049 to seed the PRNG (previously an explicit byte count was required). 4050 [Ulf M�ller, Bodo M�ller] 4051 4052 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 4053 used (char *) instead of (void *) and had casts all over the place. 4054 [Steve Henson] 4055 4056 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 4057 [Ulf M�ller] 4058 4059 *) Retain source code compatibility for BN_prime_checks macro: 4060 BN_is_prime(..., BN_prime_checks, ...) now uses 4061 BN_prime_checks_for_size to determine the appropriate number of 4062 Rabin-Miller iterations. 4063 [Ulf M�ller] 4064 4065 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 4066 DH_CHECK_P_NOT_SAFE_PRIME. 4067 (Check if this is true? OpenPGP calls them "strong".) 4068 [Ulf M�ller] 4069 4070 *) Merge the functionality of "dh" and "gendh" programs into a new program 4071 "dhparam". The old programs are retained for now but will handle DH keys 4072 (instead of parameters) in future. 4073 [Steve Henson] 4074 4075 *) Make the ciphers, s_server and s_client programs check the return values 4076 when a new cipher list is set. 4077 [Steve Henson] 4078 4079 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 4080 ciphers. Before when the 56bit ciphers were enabled the sorting was 4081 wrong. 4082 4083 The syntax for the cipher sorting has been extended to support sorting by 4084 cipher-strength (using the strength_bits hard coded in the tables). 4085 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 4086 4087 Fix a bug in the cipher-command parser: when supplying a cipher command 4088 string with an "undefined" symbol (neither command nor alphanumeric 4089 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 4090 an error is flagged. 4091 4092 Due to the strength-sorting extension, the code of the 4093 ssl_create_cipher_list() function was completely rearranged. I hope that 4094 the readability was also increased :-) 4095 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 4096 4097 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 4098 for the first serial number and places 2 in the serial number file. This 4099 avoids problems when the root CA is created with serial number zero and 4100 the first user certificate has the same issuer name and serial number 4101 as the root CA. 4102 [Steve Henson] 4103 4104 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 4105 the new code. Add documentation for this stuff. 4106 [Steve Henson] 4107 4108 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 4109 X509_*() to X509at_*() on the grounds that they don't handle X509 4110 structures and behave in an analagous way to the X509v3 functions: 4111 they shouldn't be called directly but wrapper functions should be used 4112 instead. 4113 4114 So we also now have some wrapper functions that call the X509at functions 4115 when passed certificate requests. (TO DO: similar things can be done with 4116 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 4117 things. Some of these need some d2i or i2d and print functionality 4118 because they handle more complex structures.) 4119 [Steve Henson] 4120 4121 *) Add missing #ifndefs that caused missing symbols when building libssl 4122 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 4123 NO_RSA in ssl/s2*.c. 4124 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 4125 4126 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 4127 has a return value which indicates the quality of the random data 4128 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 4129 error queue. New function RAND_pseudo_bytes() generates output that is 4130 guaranteed to be unique but not unpredictable. RAND_add is like 4131 RAND_seed, but takes an extra argument for an entropy estimate 4132 (RAND_seed always assumes full entropy). 4133 [Ulf M�ller] 4134 4135 *) Do more iterations of Rabin-Miller probable prime test (specifically, 4136 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 4137 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 4138 in crypto/bn/bn_prime.c for the complete table). This guarantees a 4139 false-positive rate of at most 2^-80 for random input. 4140 [Bodo Moeller] 4141 4142 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 4143 [Bodo Moeller] 4144 4145 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 4146 in the 0.9.5 release), this returns the chain 4147 from an X509_CTX structure with a dup of the stack and all 4148 the X509 reference counts upped: so the stack will exist 4149 after X509_CTX_cleanup() has been called. Modify pkcs12.c 4150 to use this. 4151 4152 Also make SSL_SESSION_print() print out the verify return 4153 code. 4154 [Steve Henson] 4155 4156 *) Add manpage for the pkcs12 command. Also change the default 4157 behaviour so MAC iteration counts are used unless the new 4158 -nomaciter option is used. This improves file security and 4159 only older versions of MSIE (4.0 for example) need it. 4160 [Steve Henson] 4161 4162 *) Honor the no-xxx Configure options when creating .DEF files. 4163 [Ulf M�ller] 4164 4165 *) Add PKCS#10 attributes to field table: challengePassword, 4166 unstructuredName and unstructuredAddress. These are taken from 4167 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 4168 international characters are used. 4169 4170 More changes to X509_ATTRIBUTE code: allow the setting of types 4171 based on strings. Remove the 'loc' parameter when adding 4172 attributes because these will be a SET OF encoding which is sorted 4173 in ASN1 order. 4174 [Steve Henson] 4175 4176 *) Initial changes to the 'req' utility to allow request generation 4177 automation. This will allow an application to just generate a template 4178 file containing all the field values and have req construct the 4179 request. 4180 4181 Initial support for X509_ATTRIBUTE handling. Stacks of these are 4182 used all over the place including certificate requests and PKCS#7 4183 structures. They are currently handled manually where necessary with 4184 some primitive wrappers for PKCS#7. The new functions behave in a 4185 manner analogous to the X509 extension functions: they allow 4186 attributes to be looked up by NID and added. 4187 4188 Later something similar to the X509V3 code would be desirable to 4189 automatically handle the encoding, decoding and printing of the 4190 more complex types. The string types like challengePassword can 4191 be handled by the string table functions. 4192 4193 Also modified the multi byte string table handling. Now there is 4194 a 'global mask' which masks out certain types. The table itself 4195 can use the flag STABLE_NO_MASK to ignore the mask setting: this 4196 is useful when for example there is only one permissible type 4197 (as in countryName) and using the mask might result in no valid 4198 types at all. 4199 [Steve Henson] 4200 4201 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 4202 SSL_get_peer_finished to allow applications to obtain the latest 4203 Finished messages sent to the peer or expected from the peer, 4204 respectively. (SSL_get_peer_finished is usually the Finished message 4205 actually received from the peer, otherwise the protocol will be aborted.) 4206 4207 As the Finished message are message digests of the complete handshake 4208 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 4209 be used for external authentication procedures when the authentication 4210 provided by SSL/TLS is not desired or is not enough. 4211 [Bodo Moeller] 4212 4213 *) Enhanced support for Alpha Linux is added. Now ./config checks if 4214 the host supports BWX extension and if Compaq C is present on the 4215 $PATH. Just exploiting of the BWX extension results in 20-30% 4216 performance kick for some algorithms, e.g. DES and RC4 to mention 4217 a couple. Compaq C in turn generates ~20% faster code for MD5 and 4218 SHA1. 4219 [Andy Polyakov] 4220 4221 *) Add support for MS "fast SGC". This is arguably a violation of the 4222 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 4223 weak crypto and after checking the certificate is SGC a second one 4224 with strong crypto. MS SGC stops the first handshake after receiving 4225 the server certificate message and sends a second client hello. Since 4226 a server will typically do all the time consuming operations before 4227 expecting any further messages from the client (server key exchange 4228 is the most expensive) there is little difference between the two. 4229 4230 To get OpenSSL to support MS SGC we have to permit a second client 4231 hello message after we have sent server done. In addition we have to 4232 reset the MAC if we do get this second client hello. 4233 [Steve Henson] 4234 4235 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 4236 if a DER encoded private key is RSA or DSA traditional format. Changed 4237 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 4238 format DER encoded private key. Newer code should use PKCS#8 format which 4239 has the key type encoded in the ASN1 structure. Added DER private key 4240 support to pkcs8 application. 4241 [Steve Henson] 4242 4243 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 4244 ciphersuites has been selected (as required by the SSL 3/TLS 1 4245 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 4246 is set, we interpret this as a request to violate the specification 4247 (the worst that can happen is a handshake failure, and 'correct' 4248 behaviour would result in a handshake failure anyway). 4249 [Bodo Moeller] 4250 4251 *) In SSL_CTX_add_session, take into account that there might be multiple 4252 SSL_SESSION structures with the same session ID (e.g. when two threads 4253 concurrently obtain them from an external cache). 4254 The internal cache can handle only one SSL_SESSION with a given ID, 4255 so if there's a conflict, we now throw out the old one to achieve 4256 consistency. 4257 [Bodo Moeller] 4258 4259 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 4260 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 4261 some routines that use cipher OIDs: some ciphers do not have OIDs 4262 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 4263 example. 4264 [Steve Henson] 4265 4266 *) Simplify the trust setting structure and code. Now we just have 4267 two sequences of OIDs for trusted and rejected settings. These will 4268 typically have values the same as the extended key usage extension 4269 and any application specific purposes. 4270 4271 The trust checking code now has a default behaviour: it will just 4272 check for an object with the same NID as the passed id. Functions can 4273 be provided to override either the default behaviour or the behaviour 4274 for a given id. SSL client, server and email already have functions 4275 in place for compatibility: they check the NID and also return "trusted" 4276 if the certificate is self signed. 4277 [Steve Henson] 4278 4279 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 4280 traditional format into an EVP_PKEY structure. 4281 [Steve Henson] 4282 4283 *) Add a password callback function PEM_cb() which either prompts for 4284 a password if usr_data is NULL or otherwise assumes it is a null 4285 terminated password. Allow passwords to be passed on command line 4286 environment or config files in a few more utilities. 4287 [Steve Henson] 4288 4289 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 4290 keys. Add some short names for PKCS#8 PBE algorithms and allow them 4291 to be specified on the command line for the pkcs8 and pkcs12 utilities. 4292 Update documentation. 4293 [Steve Henson] 4294 4295 *) Support for ASN1 "NULL" type. This could be handled before by using 4296 ASN1_TYPE but there wasn't any function that would try to read a NULL 4297 and produce an error if it couldn't. For compatibility we also have 4298 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 4299 don't allocate anything because they don't need to. 4300 [Steve Henson] 4301 4302 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 4303 for details. 4304 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 4305 4306 *) Rebuild of the memory allocation routines used by OpenSSL code and 4307 possibly others as well. The purpose is to make an interface that 4308 provide hooks so anyone can build a separate set of allocation and 4309 deallocation routines to be used by OpenSSL, for example memory 4310 pool implementations, or something else, which was previously hard 4311 since Malloc(), Realloc() and Free() were defined as macros having 4312 the values malloc, realloc and free, respectively (except for Win32 4313 compilations). The same is provided for memory debugging code. 4314 OpenSSL already comes with functionality to find memory leaks, but 4315 this gives people a chance to debug other memory problems. 4316 4317 With these changes, a new set of functions and macros have appeared: 4318 4319 CRYPTO_set_mem_debug_functions() [F] 4320 CRYPTO_get_mem_debug_functions() [F] 4321 CRYPTO_dbg_set_options() [F] 4322 CRYPTO_dbg_get_options() [F] 4323 CRYPTO_malloc_debug_init() [M] 4324 4325 The memory debug functions are NULL by default, unless the library 4326 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 4327 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 4328 gives the standard debugging functions that come with OpenSSL) or 4329 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 4330 provided by the library user) must be used. When the standard 4331 debugging functions are used, CRYPTO_dbg_set_options can be used to 4332 request additional information: 4333 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 4334 the CRYPTO_MDEBUG_xxx macro when compiling the library. 4335 4336 Also, things like CRYPTO_set_mem_functions will always give the 4337 expected result (the new set of functions is used for allocation 4338 and deallocation) at all times, regardless of platform and compiler 4339 options. 4340 4341 To finish it up, some functions that were never use in any other 4342 way than through macros have a new API and new semantic: 4343 4344 CRYPTO_dbg_malloc() 4345 CRYPTO_dbg_realloc() 4346 CRYPTO_dbg_free() 4347 4348 All macros of value have retained their old syntax. 4349 [Richard Levitte and Bodo Moeller] 4350 4351 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 4352 ordering of SMIMECapabilities wasn't in "strength order" and there 4353 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 4354 algorithm. 4355 [Steve Henson] 4356 4357 *) Some ASN1 types with illegal zero length encoding (INTEGER, 4358 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 4359 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 4360 4361 *) Merge in my S/MIME library for OpenSSL. This provides a simple 4362 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 4363 functionality to handle multipart/signed properly) and a utility 4364 called 'smime' to call all this stuff. This is based on code I 4365 originally wrote for Celo who have kindly allowed it to be 4366 included in OpenSSL. 4367 [Steve Henson] 4368 4369 *) Add variants des_set_key_checked and des_set_key_unchecked of 4370 des_set_key (aka des_key_sched). Global variable des_check_key 4371 decides which of these is called by des_set_key; this way 4372 des_check_key behaves as it always did, but applications and 4373 the library itself, which was buggy for des_check_key == 1, 4374 have a cleaner way to pick the version they need. 4375 [Bodo Moeller] 4376 4377 *) New function PKCS12_newpass() which changes the password of a 4378 PKCS12 structure. 4379 [Steve Henson] 4380 4381 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 4382 dynamic mix. In both cases the ids can be used as an index into the 4383 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 4384 functions so they accept a list of the field values and the 4385 application doesn't need to directly manipulate the X509_TRUST 4386 structure. 4387 [Steve Henson] 4388 4389 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 4390 need initialising. 4391 [Steve Henson] 4392 4393 *) Modify the way the V3 extension code looks up extensions. This now 4394 works in a similar way to the object code: we have some "standard" 4395 extensions in a static table which is searched with OBJ_bsearch() 4396 and the application can add dynamic ones if needed. The file 4397 crypto/x509v3/ext_dat.h now has the info: this file needs to be 4398 updated whenever a new extension is added to the core code and kept 4399 in ext_nid order. There is a simple program 'tabtest.c' which checks 4400 this. New extensions are not added too often so this file can readily 4401 be maintained manually. 4402 4403 There are two big advantages in doing things this way. The extensions 4404 can be looked up immediately and no longer need to be "added" using 4405 X509V3_add_standard_extensions(): this function now does nothing. 4406 [Side note: I get *lots* of email saying the extension code doesn't 4407 work because people forget to call this function] 4408 Also no dynamic allocation is done unless new extensions are added: 4409 so if we don't add custom extensions there is no need to call 4410 X509V3_EXT_cleanup(). 4411 [Steve Henson] 4412 4413 *) Modify enc utility's salting as follows: make salting the default. Add a 4414 magic header, so unsalted files fail gracefully instead of just decrypting 4415 to garbage. This is because not salting is a big security hole, so people 4416 should be discouraged from doing it. 4417 [Ben Laurie] 4418 4419 *) Fixes and enhancements to the 'x509' utility. It allowed a message 4420 digest to be passed on the command line but it only used this 4421 parameter when signing a certificate. Modified so all relevant 4422 operations are affected by the digest parameter including the 4423 -fingerprint and -x509toreq options. Also -x509toreq choked if a 4424 DSA key was used because it didn't fix the digest. 4425 [Steve Henson] 4426 4427 *) Initial certificate chain verify code. Currently tests the untrusted 4428 certificates for consistency with the verify purpose (which is set 4429 when the X509_STORE_CTX structure is set up) and checks the pathlength. 4430 4431 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 4432 this is because it will reject chains with invalid extensions whereas 4433 every previous version of OpenSSL and SSLeay made no checks at all. 4434 4435 Trust code: checks the root CA for the relevant trust settings. Trust 4436 settings have an initial value consistent with the verify purpose: e.g. 4437 if the verify purpose is for SSL client use it expects the CA to be 4438 trusted for SSL client use. However the default value can be changed to 4439 permit custom trust settings: one example of this would be to only trust 4440 certificates from a specific "secure" set of CAs. 4441 4442 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 4443 which should be used for version portability: especially since the 4444 verify structure is likely to change more often now. 4445 4446 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 4447 to set them. If not set then assume SSL clients will verify SSL servers 4448 and vice versa. 4449 4450 Two new options to the verify program: -untrusted allows a set of 4451 untrusted certificates to be passed in and -purpose which sets the 4452 intended purpose of the certificate. If a purpose is set then the 4453 new chain verify code is used to check extension consistency. 4454 [Steve Henson] 4455 4456 *) Support for the authority information access extension. 4457 [Steve Henson] 4458 4459 *) Modify RSA and DSA PEM read routines to transparently handle 4460 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 4461 public keys in a format compatible with certificate 4462 SubjectPublicKeyInfo structures. Unfortunately there were already 4463 functions called *_PublicKey_* which used various odd formats so 4464 these are retained for compatibility: however the DSA variants were 4465 never in a public release so they have been deleted. Changed dsa/rsa 4466 utilities to handle the new format: note no releases ever handled public 4467 keys so we should be OK. 4468 4469 The primary motivation for this change is to avoid the same fiasco 4470 that dogs private keys: there are several incompatible private key 4471 formats some of which are standard and some OpenSSL specific and 4472 require various evil hacks to allow partial transparent handling and 4473 even then it doesn't work with DER formats. Given the option anything 4474 other than PKCS#8 should be dumped: but the other formats have to 4475 stay in the name of compatibility. 4476 4477 With public keys and the benefit of hindsight one standard format 4478 is used which works with EVP_PKEY, RSA or DSA structures: though 4479 it clearly returns an error if you try to read the wrong kind of key. 4480 4481 Added a -pubkey option to the 'x509' utility to output the public key. 4482 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 4483 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 4484 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 4485 that do the same as the EVP_PKEY_assign_*() except they up the 4486 reference count of the added key (they don't "swallow" the 4487 supplied key). 4488 [Steve Henson] 4489 4490 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 4491 CRLs would fail if the file contained no certificates or no CRLs: 4492 added a new function to read in both types and return the number 4493 read: this means that if none are read it will be an error. The 4494 DER versions of the certificate and CRL reader would always fail 4495 because it isn't possible to mix certificates and CRLs in DER format 4496 without choking one or the other routine. Changed this to just read 4497 a certificate: this is the best we can do. Also modified the code 4498 in apps/verify.c to take notice of return codes: it was previously 4499 attempting to read in certificates from NULL pointers and ignoring 4500 any errors: this is one reason why the cert and CRL reader seemed 4501 to work. It doesn't check return codes from the default certificate 4502 routines: these may well fail if the certificates aren't installed. 4503 [Steve Henson] 4504 4505 *) Code to support otherName option in GeneralName. 4506 [Steve Henson] 4507 4508 *) First update to verify code. Change the verify utility 4509 so it warns if it is passed a self signed certificate: 4510 for consistency with the normal behaviour. X509_verify 4511 has been modified to it will now verify a self signed 4512 certificate if *exactly* the same certificate appears 4513 in the store: it was previously impossible to trust a 4514 single self signed certificate. This means that: 4515 openssl verify ss.pem 4516 now gives a warning about a self signed certificate but 4517 openssl verify -CAfile ss.pem ss.pem 4518 is OK. 4519 [Steve Henson] 4520 4521 *) For servers, store verify_result in SSL_SESSION data structure 4522 (and add it to external session representation). 4523 This is needed when client certificate verifications fails, 4524 but an application-provided verification callback (set by 4525 SSL_CTX_set_cert_verify_callback) allows accepting the session 4526 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 4527 but returns 1): When the session is reused, we have to set 4528 ssl->verify_result to the appropriate error code to avoid 4529 security holes. 4530 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 4531 4532 *) Fix a bug in the new PKCS#7 code: it didn't consider the 4533 case in PKCS7_dataInit() where the signed PKCS7 structure 4534 didn't contain any existing data because it was being created. 4535 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 4536 4537 *) Add a salt to the key derivation routines in enc.c. This 4538 forms the first 8 bytes of the encrypted file. Also add a 4539 -S option to allow a salt to be input on the command line. 4540 [Steve Henson] 4541 4542 *) New function X509_cmp(). Oddly enough there wasn't a function 4543 to compare two certificates. We do this by working out the SHA1 4544 hash and comparing that. X509_cmp() will be needed by the trust 4545 code. 4546 [Steve Henson] 4547 4548 *) SSL_get1_session() is like SSL_get_session(), but increments 4549 the reference count in the SSL_SESSION returned. 4550 [Geoff Thorpe <geoff@eu.c2.net>] 4551 4552 *) Fix for 'req': it was adding a null to request attributes. 4553 Also change the X509_LOOKUP and X509_INFO code to handle 4554 certificate auxiliary information. 4555 [Steve Henson] 4556 4557 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 4558 the 'enc' command. 4559 [Steve Henson] 4560 4561 *) Add the possibility to add extra information to the memory leak 4562 detecting output, to form tracebacks, showing from where each 4563 allocation was originated: CRYPTO_push_info("constant string") adds 4564 the string plus current file name and line number to a per-thread 4565 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 4566 is like calling CYRPTO_pop_info() until the stack is empty. 4567 Also updated memory leak detection code to be multi-thread-safe. 4568 [Richard Levitte] 4569 4570 *) Add options -text and -noout to pkcs7 utility and delete the 4571 encryption options which never did anything. Update docs. 4572 [Steve Henson] 4573 4574 *) Add options to some of the utilities to allow the pass phrase 4575 to be included on either the command line (not recommended on 4576 OSes like Unix) or read from the environment. Update the 4577 manpages and fix a few bugs. 4578 [Steve Henson] 4579 4580 *) Add a few manpages for some of the openssl commands. 4581 [Steve Henson] 4582 4583 *) Fix the -revoke option in ca. It was freeing up memory twice, 4584 leaking and not finding already revoked certificates. 4585 [Steve Henson] 4586 4587 *) Extensive changes to support certificate auxiliary information. 4588 This involves the use of X509_CERT_AUX structure and X509_AUX 4589 functions. An X509_AUX function such as PEM_read_X509_AUX() 4590 can still read in a certificate file in the usual way but it 4591 will also read in any additional "auxiliary information". By 4592 doing things this way a fair degree of compatibility can be 4593 retained: existing certificates can have this information added 4594 using the new 'x509' options. 4595 4596 Current auxiliary information includes an "alias" and some trust 4597 settings. The trust settings will ultimately be used in enhanced 4598 certificate chain verification routines: currently a certificate 4599 can only be trusted if it is self signed and then it is trusted 4600 for all purposes. 4601 [Steve Henson] 4602 4603 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 4604 The problem was that one of the replacement routines had not been working 4605 since SSLeay releases. For now the offending routine has been replaced 4606 with non-optimised assembler. Even so, this now gives around 95% 4607 performance improvement for 1024 bit RSA signs. 4608 [Mark Cox] 4609 4610 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 4611 handling. Most clients have the effective key size in bits equal to 4612 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 4613 A few however don't do this and instead use the size of the decrypted key 4614 to determine the RC2 key length and the AlgorithmIdentifier to determine 4615 the effective key length. In this case the effective key length can still 4616 be 40 bits but the key length can be 168 bits for example. This is fixed 4617 by manually forcing an RC2 key into the EVP_PKEY structure because the 4618 EVP code can't currently handle unusual RC2 key sizes: it always assumes 4619 the key length and effective key length are equal. 4620 [Steve Henson] 4621 4622 *) Add a bunch of functions that should simplify the creation of 4623 X509_NAME structures. Now you should be able to do: 4624 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 4625 and have it automatically work out the correct field type and fill in 4626 the structures. The more adventurous can try: 4627 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 4628 and it will (hopefully) work out the correct multibyte encoding. 4629 [Steve Henson] 4630 4631 *) Change the 'req' utility to use the new field handling and multibyte 4632 copy routines. Before the DN field creation was handled in an ad hoc 4633 way in req, ca, and x509 which was rather broken and didn't support 4634 BMPStrings or UTF8Strings. Since some software doesn't implement 4635 BMPStrings or UTF8Strings yet, they can be enabled using the config file 4636 using the dirstring_type option. See the new comment in the default 4637 openssl.cnf for more info. 4638 [Steve Henson] 4639 4640 *) Make crypto/rand/md_rand.c more robust: 4641 - Assure unique random numbers after fork(). 4642 - Make sure that concurrent threads access the global counter and 4643 md serializably so that we never lose entropy in them 4644 or use exactly the same state in multiple threads. 4645 Access to the large state is not always serializable because 4646 the additional locking could be a performance killer, and 4647 md should be large enough anyway. 4648 [Bodo Moeller] 4649 4650 *) New file apps/app_rand.c with commonly needed functionality 4651 for handling the random seed file. 4652 4653 Use the random seed file in some applications that previously did not: 4654 ca, 4655 dsaparam -genkey (which also ignored its '-rand' option), 4656 s_client, 4657 s_server, 4658 x509 (when signing). 4659 Except on systems with /dev/urandom, it is crucial to have a random 4660 seed file at least for key creation, DSA signing, and for DH exchanges; 4661 for RSA signatures we could do without one. 4662 4663 gendh and gendsa (unlike genrsa) used to read only the first byte 4664 of each file listed in the '-rand' option. The function as previously 4665 found in genrsa is now in app_rand.c and is used by all programs 4666 that support '-rand'. 4667 [Bodo Moeller] 4668 4669 *) In RAND_write_file, use mode 0600 for creating files; 4670 don't just chmod when it may be too late. 4671 [Bodo Moeller] 4672 4673 *) Report an error from X509_STORE_load_locations 4674 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 4675 [Bill Perry] 4676 4677 *) New function ASN1_mbstring_copy() this copies a string in either 4678 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 4679 into an ASN1_STRING type. A mask of permissible types is passed 4680 and it chooses the "minimal" type to use or an error if not type 4681 is suitable. 4682 [Steve Henson] 4683 4684 *) Add function equivalents to the various macros in asn1.h. The old 4685 macros are retained with an M_ prefix. Code inside the library can 4686 use the M_ macros. External code (including the openssl utility) 4687 should *NOT* in order to be "shared library friendly". 4688 [Steve Henson] 4689 4690 *) Add various functions that can check a certificate's extensions 4691 to see if it usable for various purposes such as SSL client, 4692 server or S/MIME and CAs of these types. This is currently 4693 VERY EXPERIMENTAL but will ultimately be used for certificate chain 4694 verification. Also added a -purpose flag to x509 utility to 4695 print out all the purposes. 4696 [Steve Henson] 4697 4698 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 4699 functions. 4700 [Steve Henson] 4701 4702 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 4703 for, obtain and decode and extension and obtain its critical flag. 4704 This allows all the necessary extension code to be handled in a 4705 single function call. 4706 [Steve Henson] 4707 4708 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 4709 platforms. See crypto/rc4/rc4_enc.c for further details. 4710 [Andy Polyakov] 4711 4712 *) New -noout option to asn1parse. This causes no output to be produced 4713 its main use is when combined with -strparse and -out to extract data 4714 from a file (which may not be in ASN.1 format). 4715 [Steve Henson] 4716 4717 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 4718 when producing the local key id. 4719 [Richard Levitte <levitte@stacken.kth.se>] 4720 4721 *) New option -dhparam in s_server. This allows a DH parameter file to be 4722 stated explicitly. If it is not stated then it tries the first server 4723 certificate file. The previous behaviour hard coded the filename 4724 "server.pem". 4725 [Steve Henson] 4726 4727 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 4728 a public key to be input or output. For example: 4729 openssl rsa -in key.pem -pubout -out pubkey.pem 4730 Also added necessary DSA public key functions to handle this. 4731 [Steve Henson] 4732 4733 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 4734 in the message. This was handled by allowing 4735 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 4736 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 4737 4738 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 4739 to the end of the strings whereas this didn't. This would cause problems 4740 if strings read with d2i_ASN1_bytes() were later modified. 4741 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 4742 4743 *) Fix for base64 decode bug. When a base64 bio reads only one line of 4744 data and it contains EOF it will end up returning an error. This is 4745 caused by input 46 bytes long. The cause is due to the way base64 4746 BIOs find the start of base64 encoded data. They do this by trying a 4747 trial decode on each line until they find one that works. When they 4748 do a flag is set and it starts again knowing it can pass all the 4749 data directly through the decoder. Unfortunately it doesn't reset 4750 the context it uses. This means that if EOF is reached an attempt 4751 is made to pass two EOFs through the context and this causes the 4752 resulting error. This can also cause other problems as well. As is 4753 usual with these problems it takes *ages* to find and the fix is 4754 trivial: move one line. 4755 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 4756 4757 *) Ugly workaround to get s_client and s_server working under Windows. The 4758 old code wouldn't work because it needed to select() on sockets and the 4759 tty (for keypresses and to see if data could be written). Win32 only 4760 supports select() on sockets so we select() with a 1s timeout on the 4761 sockets and then see if any characters are waiting to be read, if none 4762 are present then we retry, we also assume we can always write data to 4763 the tty. This isn't nice because the code then blocks until we've 4764 received a complete line of data and it is effectively polling the 4765 keyboard at 1s intervals: however it's quite a bit better than not 4766 working at all :-) A dedicated Windows application might handle this 4767 with an event loop for example. 4768 [Steve Henson] 4769 4770 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 4771 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 4772 will be called when RSA_sign() and RSA_verify() are used. This is useful 4773 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 4774 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 4775 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 4776 This necessitated the support of an extra signature type NID_md5_sha1 4777 for SSL signatures and modifications to the SSL library to use it instead 4778 of calling RSA_public_decrypt() and RSA_private_encrypt(). 4779 [Steve Henson] 4780 4781 *) Add new -verify -CAfile and -CApath options to the crl program, these 4782 will lookup a CRL issuers certificate and verify the signature in a 4783 similar way to the verify program. Tidy up the crl program so it 4784 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 4785 less strict. It will now permit CRL extensions even if it is not 4786 a V2 CRL: this will allow it to tolerate some broken CRLs. 4787 [Steve Henson] 4788 4789 *) Initialize all non-automatic variables each time one of the openssl 4790 sub-programs is started (this is necessary as they may be started 4791 multiple times from the "OpenSSL>" prompt). 4792 [Lennart Bang, Bodo Moeller] 4793 4794 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 4795 removing all other RSA functionality (this is what NO_RSA does). This 4796 is so (for example) those in the US can disable those operations covered 4797 by the RSA patent while allowing storage and parsing of RSA keys and RSA 4798 key generation. 4799 [Steve Henson] 4800 4801 *) Non-copying interface to BIO pairs. 4802 (still largely untested) 4803 [Bodo Moeller] 4804 4805 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 4806 ASCII string. This was handled independently in various places before. 4807 [Steve Henson] 4808 4809 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 4810 UTF8 strings a character at a time. 4811 [Steve Henson] 4812 4813 *) Use client_version from client hello to select the protocol 4814 (s23_srvr.c) and for RSA client key exchange verification 4815 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 4816 [Bodo Moeller] 4817 4818 *) Add various utility functions to handle SPKACs, these were previously 4819 handled by poking round in the structure internals. Added new function 4820 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 4821 print, verify and generate SPKACs. Based on an original idea from 4822 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 4823 [Steve Henson] 4824 4825 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 4826 [Andy Polyakov] 4827 4828 *) Allow the config file extension section to be overwritten on the 4829 command line. Based on an original idea from Massimiliano Pala 4830 <madwolf@comune.modena.it>. The new option is called -extensions 4831 and can be applied to ca, req and x509. Also -reqexts to override 4832 the request extensions in req and -crlexts to override the crl extensions 4833 in ca. 4834 [Steve Henson] 4835 4836 *) Add new feature to the SPKAC handling in ca. Now you can include 4837 the same field multiple times by preceding it by "XXXX." for example: 4838 1.OU="Unit name 1" 4839 2.OU="Unit name 2" 4840 this is the same syntax as used in the req config file. 4841 [Steve Henson] 4842 4843 *) Allow certificate extensions to be added to certificate requests. These 4844 are specified in a 'req_extensions' option of the req section of the 4845 config file. They can be printed out with the -text option to req but 4846 are otherwise ignored at present. 4847 [Steve Henson] 4848 4849 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 4850 data read consists of only the final block it would not decrypted because 4851 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 4852 A misplaced 'break' also meant the decrypted final block might not be 4853 copied until the next read. 4854 [Steve Henson] 4855 4856 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 4857 a few extra parameters to the DH structure: these will be useful if 4858 for example we want the value of 'q' or implement X9.42 DH. 4859 [Steve Henson] 4860 4861 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 4862 provides hooks that allow the default DSA functions or functions on a 4863 "per key" basis to be replaced. This allows hardware acceleration and 4864 hardware key storage to be handled without major modification to the 4865 library. Also added low level modexp hooks and CRYPTO_EX structure and 4866 associated functions. 4867 [Steve Henson] 4868 4869 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 4870 as "read only": it can't be written to and the buffer it points to will 4871 not be freed. Reading from a read only BIO is much more efficient than 4872 a normal memory BIO. This was added because there are several times when 4873 an area of memory needs to be read from a BIO. The previous method was 4874 to create a memory BIO and write the data to it, this results in two 4875 copies of the data and an O(n^2) reading algorithm. There is a new 4876 function BIO_new_mem_buf() which creates a read only memory BIO from 4877 an area of memory. Also modified the PKCS#7 routines to use read only 4878 memory BIOs. 4879 [Steve Henson] 4880 4881 *) Bugfix: ssl23_get_client_hello did not work properly when called in 4882 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 4883 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 4884 but a retry condition occured while trying to read the rest. 4885 [Bodo Moeller] 4886 4887 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 4888 NID_pkcs7_encrypted by default: this was wrong since this should almost 4889 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 4890 the encrypted data type: this is a more sensible place to put it and it 4891 allows the PKCS#12 code to be tidied up that duplicated this 4892 functionality. 4893 [Steve Henson] 4894 4895 *) Changed obj_dat.pl script so it takes its input and output files on 4896 the command line. This should avoid shell escape redirection problems 4897 under Win32. 4898 [Steve Henson] 4899 4900 *) Initial support for certificate extension requests, these are included 4901 in things like Xenroll certificate requests. Included functions to allow 4902 extensions to be obtained and added. 4903 [Steve Henson] 4904 4905 *) -crlf option to s_client and s_server for sending newlines as 4906 CRLF (as required by many protocols). 4907 [Bodo Moeller] 4908 4909 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 4910 4911 *) Install libRSAglue.a when OpenSSL is built with RSAref. 4912 [Ralf S. Engelschall] 4913 4914 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 4915 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 4916 4917 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 4918 program. 4919 [Steve Henson] 4920 4921 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 4922 DH parameters/keys (q is lost during that conversion, but the resulting 4923 DH parameters contain its length). 4924 4925 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 4926 much faster than DH_generate_parameters (which creates parameters 4927 where p = 2*q + 1), and also the smaller q makes DH computations 4928 much more efficient (160-bit exponentiation instead of 1024-bit 4929 exponentiation); so this provides a convenient way to support DHE 4930 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 4931 utter importance to use 4932 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 4933 or 4934 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 4935 when such DH parameters are used, because otherwise small subgroup 4936 attacks may become possible! 4937 [Bodo Moeller] 4938 4939 *) Avoid memory leak in i2d_DHparams. 4940 [Bodo Moeller] 4941 4942 *) Allow the -k option to be used more than once in the enc program: 4943 this allows the same encrypted message to be read by multiple recipients. 4944 [Steve Henson] 4945 4946 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 4947 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 4948 it will always use the numerical form of the OID, even if it has a short 4949 or long name. 4950 [Steve Henson] 4951 4952 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 4953 method only got called if p,q,dmp1,dmq1,iqmp components were present, 4954 otherwise bn_mod_exp was called. In the case of hardware keys for example 4955 no private key components need be present and it might store extra data 4956 in the RSA structure, which cannot be accessed from bn_mod_exp. 4957 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 4958 private key operations. 4959 [Steve Henson] 4960 4961 *) Added support for SPARC Linux. 4962 [Andy Polyakov] 4963 4964 *) pem_password_cb function type incompatibly changed from 4965 typedef int pem_password_cb(char *buf, int size, int rwflag); 4966 to 4967 ....(char *buf, int size, int rwflag, void *userdata); 4968 so that applications can pass data to their callbacks: 4969 The PEM[_ASN1]_{read,write}... functions and macros now take an 4970 additional void * argument, which is just handed through whenever 4971 the password callback is called. 4972 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 4973 4974 New function SSL_CTX_set_default_passwd_cb_userdata. 4975 4976 Compatibility note: As many C implementations push function arguments 4977 onto the stack in reverse order, the new library version is likely to 4978 interoperate with programs that have been compiled with the old 4979 pem_password_cb definition (PEM_whatever takes some data that 4980 happens to be on the stack as its last argument, and the callback 4981 just ignores this garbage); but there is no guarantee whatsoever that 4982 this will work. 4983 4984 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 4985 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 4986 problems not only on Windows, but also on some Unix platforms. 4987 To avoid problematic command lines, these definitions are now in an 4988 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 4989 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 4990 [Bodo Moeller] 4991 4992 *) MIPS III/IV assembler module is reimplemented. 4993 [Andy Polyakov] 4994 4995 *) More DES library cleanups: remove references to srand/rand and 4996 delete an unused file. 4997 [Ulf M�ller] 4998 4999 *) Add support for the the free Netwide assembler (NASM) under Win32, 5000 since not many people have MASM (ml) and it can be hard to obtain. 5001 This is currently experimental but it seems to work OK and pass all 5002 the tests. Check out INSTALL.W32 for info. 5003 [Steve Henson] 5004 5005 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 5006 without temporary keys kept an extra copy of the server key, 5007 and connections with temporary keys did not free everything in case 5008 of an error. 5009 [Bodo Moeller] 5010 5011 *) New function RSA_check_key and new openssl rsa option -check 5012 for verifying the consistency of RSA keys. 5013 [Ulf Moeller, Bodo Moeller] 5014 5015 *) Various changes to make Win32 compile work: 5016 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 5017 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 5018 comparison" warnings. 5019 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 5020 [Steve Henson] 5021 5022 *) Add a debugging option to PKCS#5 v2 key generation function: when 5023 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 5024 derived keys are printed to stderr. 5025 [Steve Henson] 5026 5027 *) Copy the flags in ASN1_STRING_dup(). 5028 [Roman E. Pavlov <pre@mo.msk.ru>] 5029 5030 *) The x509 application mishandled signing requests containing DSA 5031 keys when the signing key was also DSA and the parameters didn't match. 5032 5033 It was supposed to omit the parameters when they matched the signing key: 5034 the verifying software was then supposed to automatically use the CA's 5035 parameters if they were absent from the end user certificate. 5036 5037 Omitting parameters is no longer recommended. The test was also 5038 the wrong way round! This was probably due to unusual behaviour in 5039 EVP_cmp_parameters() which returns 1 if the parameters match. 5040 This meant that parameters were omitted when they *didn't* match and 5041 the certificate was useless. Certificates signed with 'ca' didn't have 5042 this bug. 5043 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 5044 5045 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 5046 The interface is as follows: 5047 Applications can use 5048 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 5049 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 5050 "off" is now the default. 5051 The library internally uses 5052 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 5053 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 5054 to disable memory-checking temporarily. 5055 5056 Some inconsistent states that previously were possible (and were 5057 even the default) are now avoided. 5058 5059 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 5060 with each memory chunk allocated; this is occasionally more helpful 5061 than just having a counter. 5062 5063 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 5064 5065 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 5066 extensions. 5067 [Bodo Moeller] 5068 5069 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 5070 which largely parallels "options", but is for changing API behaviour, 5071 whereas "options" are about protocol behaviour. 5072 Initial "mode" flags are: 5073 5074 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 5075 a single record has been written. 5076 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 5077 retries use the same buffer location. 5078 (But all of the contents must be 5079 copied!) 5080 [Bodo Moeller] 5081 5082 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 5083 worked. 5084 5085 *) Fix problems with no-hmac etc. 5086 [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 5087 5088 *) New functions RSA_get_default_method(), RSA_set_method() and 5089 RSA_get_method(). These allows replacement of RSA_METHODs without having 5090 to mess around with the internals of an RSA structure. 5091 [Steve Henson] 5092 5093 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 5094 Also really enable memory leak checks in openssl.c and in some 5095 test programs. 5096 [Chad C. Mulligan, Bodo Moeller] 5097 5098 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 5099 up the length of negative integers. This has now been simplified to just 5100 store the length when it is first determined and use it later, rather 5101 than trying to keep track of where data is copied and updating it to 5102 point to the end. 5103 [Steve Henson, reported by Brien Wheeler 5104 <bwheeler@authentica-security.com>] 5105 5106 *) Add a new function PKCS7_signatureVerify. This allows the verification 5107 of a PKCS#7 signature but with the signing certificate passed to the 5108 function itself. This contrasts with PKCS7_dataVerify which assumes the 5109 certificate is present in the PKCS#7 structure. This isn't always the 5110 case: certificates can be omitted from a PKCS#7 structure and be 5111 distributed by "out of band" means (such as a certificate database). 5112 [Steve Henson] 5113 5114 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 5115 function prototypes in pem.h, also change util/mkdef.pl to add the 5116 necessary function names. 5117 [Steve Henson] 5118 5119 *) mk1mf.pl (used by Windows builds) did not properly read the 5120 options set by Configure in the top level Makefile, and Configure 5121 was not even able to write more than one option correctly. 5122 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 5123 [Bodo Moeller] 5124 5125 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 5126 file to be loaded from a BIO or FILE pointer. The BIO version will 5127 for example allow memory BIOs to contain config info. 5128 [Steve Henson] 5129 5130 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 5131 Whoever hopes to achieve shared-library compatibility across versions 5132 must use this, not the compile-time macro. 5133 (Exercise 0.9.4: Which is the minimum library version required by 5134 such programs?) 5135 Note: All this applies only to multi-threaded programs, others don't 5136 need locks. 5137 [Bodo Moeller] 5138 5139 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 5140 through a BIO pair triggered the default case, i.e. 5141 SSLerr(...,SSL_R_UNKNOWN_STATE). 5142 [Bodo Moeller] 5143 5144 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 5145 can use the SSL library even if none of the specific BIOs is 5146 appropriate. 5147 [Bodo Moeller] 5148 5149 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 5150 for the encoded length. 5151 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 5152 5153 *) Add initial documentation of the X509V3 functions. 5154 [Steve Henson] 5155 5156 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 5157 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 5158 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 5159 secure PKCS#8 private key format with a high iteration count. 5160 [Steve Henson] 5161 5162 *) Fix determination of Perl interpreter: A perl or perl5 5163 _directory_ in $PATH was also accepted as the interpreter. 5164 [Ralf S. Engelschall] 5165 5166 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 5167 wrong with it but it was very old and did things like calling 5168 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 5169 unusual formatting. 5170 [Steve Henson] 5171 5172 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 5173 to use the new extension code. 5174 [Steve Henson] 5175 5176 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 5177 with macros. This should make it easier to change their form, add extra 5178 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 5179 constant. 5180 [Steve Henson] 5181 5182 *) Add to configuration table a new entry that can specify an alternative 5183 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 5184 according to Mark Crispin <MRC@Panda.COM>. 5185 [Bodo Moeller] 5186 5187#if 0 5188 *) DES CBC did not update the IV. Weird. 5189 [Ben Laurie] 5190#else 5191 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 5192 Changing the behaviour of the former might break existing programs -- 5193 where IV updating is needed, des_ncbc_encrypt can be used. 5194#endif 5195 5196 *) When bntest is run from "make test" it drives bc to check its 5197 calculations, as well as internally checking them. If an internal check 5198 fails, it needs to cause bc to give a non-zero result or make test carries 5199 on without noticing the failure. Fixed. 5200 [Ben Laurie] 5201 5202 *) DES library cleanups. 5203 [Ulf M�ller] 5204 5205 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 5206 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 5207 ciphers. NOTE: although the key derivation function has been verified 5208 against some published test vectors it has not been extensively tested 5209 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 5210 of v2.0. 5211 [Steve Henson] 5212 5213 *) Instead of "mkdir -p", which is not fully portable, use new 5214 Perl script "util/mkdir-p.pl". 5215 [Bodo Moeller] 5216 5217 *) Rewrite the way password based encryption (PBE) is handled. It used to 5218 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 5219 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 5220 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 5221 the 'parameter' field of the AlgorithmIdentifier is passed to the 5222 underlying key generation function so it must do its own ASN1 parsing. 5223 This has also changed the EVP_PBE_CipherInit() function which now has a 5224 'parameter' argument instead of literal salt and iteration count values 5225 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 5226 [Steve Henson] 5227 5228 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 5229 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 5230 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 5231 KEY" because this clashed with PKCS#8 unencrypted string. Since this 5232 value was just used as a "magic string" and not used directly its 5233 value doesn't matter. 5234 [Steve Henson] 5235 5236 *) Introduce some semblance of const correctness to BN. Shame C doesn't 5237 support mutable. 5238 [Ben Laurie] 5239 5240 *) "linux-sparc64" configuration (ultrapenguin). 5241 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 5242 "linux-sparc" configuration. 5243 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 5244 5245 *) config now generates no-xxx options for missing ciphers. 5246 [Ulf M�ller] 5247 5248 *) Support the EBCDIC character set (work in progress). 5249 File ebcdic.c not yet included because it has a different license. 5250 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 5251 5252 *) Support BS2000/OSD-POSIX. 5253 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 5254 5255 *) Make callbacks for key generation use void * instead of char *. 5256 [Ben Laurie] 5257 5258 *) Make S/MIME samples compile (not yet tested). 5259 [Ben Laurie] 5260 5261 *) Additional typesafe stacks. 5262 [Ben Laurie] 5263 5264 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 5265 [Bodo Moeller] 5266 5267 5268 Changes between 0.9.3 and 0.9.3a [29 May 1999] 5269 5270 *) New configuration variant "sco5-gcc". 5271 5272 *) Updated some demos. 5273 [Sean O Riordain, Wade Scholine] 5274 5275 *) Add missing BIO_free at exit of pkcs12 application. 5276 [Wu Zhigang] 5277 5278 *) Fix memory leak in conf.c. 5279 [Steve Henson] 5280 5281 *) Updates for Win32 to assembler version of MD5. 5282 [Steve Henson] 5283 5284 *) Set #! path to perl in apps/der_chop to where we found it 5285 instead of using a fixed path. 5286 [Bodo Moeller] 5287 5288 *) SHA library changes for irix64-mips4-cc. 5289 [Andy Polyakov] 5290 5291 *) Improvements for VMS support. 5292 [Richard Levitte] 5293 5294 5295 Changes between 0.9.2b and 0.9.3 [24 May 1999] 5296 5297 *) Bignum library bug fix. IRIX 6 passes "make test" now! 5298 This also avoids the problems with SC4.2 and unpatched SC5. 5299 [Andy Polyakov <appro@fy.chalmers.se>] 5300 5301 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 5302 These are required because of the typesafe stack would otherwise break 5303 existing code. If old code used a structure member which used to be STACK 5304 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 5305 sk_num or sk_value it would produce an error because the num, data members 5306 are not present in STACK_OF. Now it just produces a warning. sk_set 5307 replaces the old method of assigning a value to sk_value 5308 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 5309 that does this will no longer work (and should use sk_set instead) but 5310 this could be regarded as a "questionable" behaviour anyway. 5311 [Steve Henson] 5312 5313 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 5314 correctly handle encrypted S/MIME data. 5315 [Steve Henson] 5316 5317 *) Change type of various DES function arguments from des_cblock 5318 (which means, in function argument declarations, pointer to char) 5319 to des_cblock * (meaning pointer to array with 8 char elements), 5320 which allows the compiler to do more typechecking; it was like 5321 that back in SSLeay, but with lots of ugly casts. 5322 5323 Introduce new type const_des_cblock. 5324 [Bodo Moeller] 5325 5326 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 5327 problems: find RecipientInfo structure that matches recipient certificate 5328 and initialise the ASN1 structures properly based on passed cipher. 5329 [Steve Henson] 5330 5331 *) Belatedly make the BN tests actually check the results. 5332 [Ben Laurie] 5333 5334 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 5335 to and from BNs: it was completely broken. New compilation option 5336 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 5337 key elements as negative integers. 5338 [Steve Henson] 5339 5340 *) Reorganize and speed up MD5. 5341 [Andy Polyakov <appro@fy.chalmers.se>] 5342 5343 *) VMS support. 5344 [Richard Levitte <richard@levitte.org>] 5345 5346 *) New option -out to asn1parse to allow the parsed structure to be 5347 output to a file. This is most useful when combined with the -strparse 5348 option to examine the output of things like OCTET STRINGS. 5349 [Steve Henson] 5350 5351 *) Make SSL library a little more fool-proof by not requiring any longer 5352 that SSL_set_{accept,connect}_state be called before 5353 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 5354 in many applications because usually everything *appeared* to work as 5355 intended anyway -- now it really works as intended). 5356 [Bodo Moeller] 5357 5358 *) Move openssl.cnf out of lib/. 5359 [Ulf M�ller] 5360 5361 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 5362 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 5363 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 5364 [Ralf S. Engelschall] 5365 5366 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 5367 handle PKCS#7 enveloped data properly. 5368 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 5369 5370 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 5371 copying pointers. The cert_st handling is changed by this in 5372 various ways (and thus what used to be known as ctx->default_cert 5373 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 5374 any longer when s->cert does not give us what we need). 5375 ssl_cert_instantiate becomes obsolete by this change. 5376 As soon as we've got the new code right (possibly it already is?), 5377 we have solved a couple of bugs of the earlier code where s->cert 5378 was used as if it could not have been shared with other SSL structures. 5379 5380 Note that using the SSL API in certain dirty ways now will result 5381 in different behaviour than observed with earlier library versions: 5382 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 5383 does not influence s as it used to. 5384 5385 In order to clean up things more thoroughly, inside SSL_SESSION 5386 we don't use CERT any longer, but a new structure SESS_CERT 5387 that holds per-session data (if available); currently, this is 5388 the peer's certificate chain and, for clients, the server's certificate 5389 and temporary key. CERT holds only those values that can have 5390 meaningful defaults in an SSL_CTX. 5391 [Bodo Moeller] 5392 5393 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 5394 from the internal representation. Various PKCS#7 fixes: remove some 5395 evil casts and set the enc_dig_alg field properly based on the signing 5396 key type. 5397 [Steve Henson] 5398 5399 *) Allow PKCS#12 password to be set from the command line or the 5400 environment. Let 'ca' get its config file name from the environment 5401 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 5402 and 'x509'). 5403 [Steve Henson] 5404 5405 *) Allow certificate policies extension to use an IA5STRING for the 5406 organization field. This is contrary to the PKIX definition but 5407 VeriSign uses it and IE5 only recognises this form. Document 'x509' 5408 extension option. 5409 [Steve Henson] 5410 5411 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 5412 without disallowing inline assembler and the like for non-pedantic builds. 5413 [Ben Laurie] 5414 5415 *) Support Borland C++ builder. 5416 [Janez Jere <jj@void.si>, modified by Ulf M�ller] 5417 5418 *) Support Mingw32. 5419 [Ulf M�ller] 5420 5421 *) SHA-1 cleanups and performance enhancements. 5422 [Andy Polyakov <appro@fy.chalmers.se>] 5423 5424 *) Sparc v8plus assembler for the bignum library. 5425 [Andy Polyakov <appro@fy.chalmers.se>] 5426 5427 *) Accept any -xxx and +xxx compiler options in Configure. 5428 [Ulf M�ller] 5429 5430 *) Update HPUX configuration. 5431 [Anonymous] 5432 5433 *) Add missing sk_<type>_unshift() function to safestack.h 5434 [Ralf S. Engelschall] 5435 5436 *) New function SSL_CTX_use_certificate_chain_file that sets the 5437 "extra_cert"s in addition to the certificate. (This makes sense 5438 only for "PEM" format files, as chains as a whole are not 5439 DER-encoded.) 5440 [Bodo Moeller] 5441 5442 *) Support verify_depth from the SSL API. 5443 x509_vfy.c had what can be considered an off-by-one-error: 5444 Its depth (which was not part of the external interface) 5445 was actually counting the number of certificates in a chain; 5446 now it really counts the depth. 5447 [Bodo Moeller] 5448 5449 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 5450 instead of X509err, which often resulted in confusing error 5451 messages since the error codes are not globally unique 5452 (e.g. an alleged error in ssl3_accept when a certificate 5453 didn't match the private key). 5454 5455 *) New function SSL_CTX_set_session_id_context that allows to set a default 5456 value (so that you don't need SSL_set_session_id_context for each 5457 connection using the SSL_CTX). 5458 [Bodo Moeller] 5459 5460 *) OAEP decoding bug fix. 5461 [Ulf M�ller] 5462 5463 *) Support INSTALL_PREFIX for package builders, as proposed by 5464 David Harris. 5465 [Bodo Moeller] 5466 5467 *) New Configure options "threads" and "no-threads". For systems 5468 where the proper compiler options are known (currently Solaris 5469 and Linux), "threads" is the default. 5470 [Bodo Moeller] 5471 5472 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 5473 [Bodo Moeller] 5474 5475 *) Install various scripts to $(OPENSSLDIR)/misc, not to 5476 $(INSTALLTOP)/bin -- they shouldn't clutter directories 5477 such as /usr/local/bin. 5478 [Bodo Moeller] 5479 5480 *) "make linux-shared" to build shared libraries. 5481 [Niels Poppe <niels@netbox.org>] 5482 5483 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 5484 [Ulf M�ller] 5485 5486 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 5487 extension adding in x509 utility. 5488 [Steve Henson] 5489 5490 *) Remove NOPROTO sections and error code comments. 5491 [Ulf M�ller] 5492 5493 *) Partial rewrite of the DEF file generator to now parse the ANSI 5494 prototypes. 5495 [Steve Henson] 5496 5497 *) New Configure options --prefix=DIR and --openssldir=DIR. 5498 [Ulf M�ller] 5499 5500 *) Complete rewrite of the error code script(s). It is all now handled 5501 by one script at the top level which handles error code gathering, 5502 header rewriting and C source file generation. It should be much better 5503 than the old method: it now uses a modified version of Ulf's parser to 5504 read the ANSI prototypes in all header files (thus the old K&R definitions 5505 aren't needed for error creation any more) and do a better job of 5506 translating function codes into names. The old 'ASN1 error code imbedded 5507 in a comment' is no longer necessary and it doesn't use .err files which 5508 have now been deleted. Also the error code call doesn't have to appear all 5509 on one line (which resulted in some large lines...). 5510 [Steve Henson] 5511 5512 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 5513 [Bodo Moeller] 5514 5515 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 5516 0 (which usually indicates a closed connection), but continue reading. 5517 [Bodo Moeller] 5518 5519 *) Fix some race conditions. 5520 [Bodo Moeller] 5521 5522 *) Add support for CRL distribution points extension. Add Certificate 5523 Policies and CRL distribution points documentation. 5524 [Steve Henson] 5525 5526 *) Move the autogenerated header file parts to crypto/opensslconf.h. 5527 [Ulf M�ller] 5528 5529 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 5530 8 of keying material. Merlin has also confirmed interop with this fix 5531 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 5532 [Merlin Hughes <merlin@baltimore.ie>] 5533 5534 *) Fix lots of warnings. 5535 [Richard Levitte <levitte@stacken.kth.se>] 5536 5537 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 5538 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 5539 [Richard Levitte <levitte@stacken.kth.se>] 5540 5541 *) Fix problems with sizeof(long) == 8. 5542 [Andy Polyakov <appro@fy.chalmers.se>] 5543 5544 *) Change functions to ANSI C. 5545 [Ulf M�ller] 5546 5547 *) Fix typos in error codes. 5548 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 5549 5550 *) Remove defunct assembler files from Configure. 5551 [Ulf M�ller] 5552 5553 *) SPARC v8 assembler BIGNUM implementation. 5554 [Andy Polyakov <appro@fy.chalmers.se>] 5555 5556 *) Support for Certificate Policies extension: both print and set. 5557 Various additions to support the r2i method this uses. 5558 [Steve Henson] 5559 5560 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 5561 return a const string when you are expecting an allocated buffer. 5562 [Ben Laurie] 5563 5564 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 5565 types DirectoryString and DisplayText. 5566 [Steve Henson] 5567 5568 *) Add code to allow r2i extensions to access the configuration database, 5569 add an LHASH database driver and add several ctx helper functions. 5570 [Steve Henson] 5571 5572 *) Fix an evil bug in bn_expand2() which caused various BN functions to 5573 fail when they extended the size of a BIGNUM. 5574 [Steve Henson] 5575 5576 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 5577 support typesafe stack. 5578 [Steve Henson] 5579 5580 *) Fix typo in SSL_[gs]et_options(). 5581 [Nils Frostberg <nils@medcom.se>] 5582 5583 *) Delete various functions and files that belonged to the (now obsolete) 5584 old X509V3 handling code. 5585 [Steve Henson] 5586 5587 *) New Configure option "rsaref". 5588 [Ulf M�ller] 5589 5590 *) Don't auto-generate pem.h. 5591 [Bodo Moeller] 5592 5593 *) Introduce type-safe ASN.1 SETs. 5594 [Ben Laurie] 5595 5596 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 5597 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 5598 5599 *) Introduce type-safe STACKs. This will almost certainly break lots of code 5600 that links with OpenSSL (well at least cause lots of warnings), but fear 5601 not: the conversion is trivial, and it eliminates loads of evil casts. A 5602 few STACKed things have been converted already. Feel free to convert more. 5603 In the fullness of time, I'll do away with the STACK type altogether. 5604 [Ben Laurie] 5605 5606 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 5607 specified in <certfile> by updating the entry in the index.txt file. 5608 This way one no longer has to edit the index.txt file manually for 5609 revoking a certificate. The -revoke option does the gory details now. 5610 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 5611 5612 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 5613 `-text' option at all and this way the `-noout -text' combination was 5614 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 5615 [Ralf S. Engelschall] 5616 5617 *) Make sure a corresponding plain text error message exists for the 5618 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 5619 verify callback function determined that a certificate was revoked. 5620 [Ralf S. Engelschall] 5621 5622 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 5623 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 5624 all available cipers including rc5, which was forgotten until now. 5625 In order to let the testing shell script know which algorithms 5626 are available, a new (up to now undocumented) command 5627 "openssl list-cipher-commands" is used. 5628 [Bodo Moeller] 5629 5630 *) Bugfix: s_client occasionally would sleep in select() when 5631 it should have checked SSL_pending() first. 5632 [Bodo Moeller] 5633 5634 *) New functions DSA_do_sign and DSA_do_verify to provide access to 5635 the raw DSA values prior to ASN.1 encoding. 5636 [Ulf M�ller] 5637 5638 *) Tweaks to Configure 5639 [Niels Poppe <niels@netbox.org>] 5640 5641 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 5642 yet... 5643 [Steve Henson] 5644 5645 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 5646 [Ulf M�ller] 5647 5648 *) New config option to avoid instructions that are illegal on the 80386. 5649 The default code is faster, but requires at least a 486. 5650 [Ulf M�ller] 5651 5652 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 5653 SSL2_SERVER_VERSION (not used at all) macros, which are now the 5654 same as SSL2_VERSION anyway. 5655 [Bodo Moeller] 5656 5657 *) New "-showcerts" option for s_client. 5658 [Bodo Moeller] 5659 5660 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 5661 application. Various cleanups and fixes. 5662 [Steve Henson] 5663 5664 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 5665 modify error routines to work internally. Add error codes and PBE init 5666 to library startup routines. 5667 [Steve Henson] 5668 5669 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 5670 packing functions to asn1 and evp. Changed function names and error 5671 codes along the way. 5672 [Steve Henson] 5673 5674 *) PKCS12 integration: and so it begins... First of several patches to 5675 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 5676 objects to objects.h 5677 [Steve Henson] 5678 5679 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 5680 and display support for Thawte strong extranet extension. 5681 [Steve Henson] 5682 5683 *) Add LinuxPPC support. 5684 [Jeff Dubrule <igor@pobox.org>] 5685 5686 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 5687 bn_div_words in alpha.s. 5688 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 5689 5690 *) Make sure the RSA OAEP test is skipped under -DRSAref because 5691 OAEP isn't supported when OpenSSL is built with RSAref. 5692 [Ulf Moeller <ulf@fitug.de>] 5693 5694 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 5695 so they no longer are missing under -DNOPROTO. 5696 [Soren S. Jorvang <soren@t.dk>] 5697 5698 5699 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 5700 5701 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 5702 doesn't work when the session is reused. Coming soon! 5703 [Ben Laurie] 5704 5705 *) Fix a security hole, that allows sessions to be reused in the wrong 5706 context thus bypassing client cert protection! All software that uses 5707 client certs and session caches in multiple contexts NEEDS PATCHING to 5708 allow session reuse! A fuller solution is in the works. 5709 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 5710 5711 *) Some more source tree cleanups (removed obsolete files 5712 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 5713 permission on "config" script to be executable) and a fix for the INSTALL 5714 document. 5715 [Ulf Moeller <ulf@fitug.de>] 5716 5717 *) Remove some legacy and erroneous uses of malloc, free instead of 5718 Malloc, Free. 5719 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 5720 5721 *) Make rsa_oaep_test return non-zero on error. 5722 [Ulf Moeller <ulf@fitug.de>] 5723 5724 *) Add support for native Solaris shared libraries. Configure 5725 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 5726 if someone would make that last step automatic. 5727 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 5728 5729 *) ctx_size was not built with the right compiler during "make links". Fixed. 5730 [Ben Laurie] 5731 5732 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 5733 except NULL ciphers". This means the default cipher list will no longer 5734 enable NULL ciphers. They need to be specifically enabled e.g. with 5735 the string "DEFAULT:eNULL". 5736 [Steve Henson] 5737 5738 *) Fix to RSA private encryption routines: if p < q then it would 5739 occasionally produce an invalid result. This will only happen with 5740 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 5741 [Steve Henson] 5742 5743 *) Be less restrictive and allow also `perl util/perlpath.pl 5744 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 5745 because this way one can also use an interpreter named `perl5' (which is 5746 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 5747 installed as `perl'). 5748 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 5749 5750 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 5751 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 5752 5753 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 5754 advapi32.lib to Win32 build and change the pem test comparision 5755 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 5756 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 5757 and crypto/des/ede_cbcm_enc.c. 5758 [Steve Henson] 5759 5760 *) DES quad checksum was broken on big-endian architectures. Fixed. 5761 [Ben Laurie] 5762 5763 *) Comment out two functions in bio.h that aren't implemented. Fix up the 5764 Win32 test batch file so it (might) work again. The Win32 test batch file 5765 is horrible: I feel ill.... 5766 [Steve Henson] 5767 5768 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 5769 in e_os.h. Audit of header files to check ANSI and non ANSI 5770 sections: 10 functions were absent from non ANSI section and not exported 5771 from Windows DLLs. Fixed up libeay.num for new functions. 5772 [Steve Henson] 5773 5774 *) Make `openssl version' output lines consistent. 5775 [Ralf S. Engelschall] 5776 5777 *) Fix Win32 symbol export lists for BIO functions: Added 5778 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 5779 to ms/libeay{16,32}.def. 5780 [Ralf S. Engelschall] 5781 5782 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 5783 fine under Unix and passes some trivial tests I've now added. But the 5784 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 5785 added to make sure no one expects that this stuff really works in the 5786 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 5787 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 5788 openssl_bio.xs. 5789 [Ralf S. Engelschall] 5790 5791 *) Fix the generation of two part addresses in perl. 5792 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 5793 5794 *) Add config entry for Linux on MIPS. 5795 [John Tobey <jtobey@channel1.com>] 5796 5797 *) Make links whenever Configure is run, unless we are on Windoze. 5798 [Ben Laurie] 5799 5800 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 5801 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 5802 in CRLs. 5803 [Steve Henson] 5804 5805 *) Add a useful kludge to allow package maintainers to specify compiler and 5806 other platforms details on the command line without having to patch the 5807 Configure script everytime: One now can use ``perl Configure 5808 <id>:<details>'', i.e. platform ids are allowed to have details appended 5809 to them (seperated by colons). This is treated as there would be a static 5810 pre-configured entry in Configure's %table under key <id> with value 5811 <details> and ``perl Configure <id>'' is called. So, when you want to 5812 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 5813 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 5814 now, which overrides the FreeBSD-elf entry on-the-fly. 5815 [Ralf S. Engelschall] 5816 5817 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 5818 [Ben Laurie] 5819 5820 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 5821 on the `perl Configure ...' command line. This way one can compile 5822 OpenSSL libraries with Position Independent Code (PIC) which is needed 5823 for linking it into DSOs. 5824 [Ralf S. Engelschall] 5825 5826 *) Remarkably, export ciphers were totally broken and no-one had noticed! 5827 Fixed. 5828 [Ben Laurie] 5829 5830 *) Cleaned up the LICENSE document: The official contact for any license 5831 questions now is the OpenSSL core team under openssl-core@openssl.org. 5832 And add a paragraph about the dual-license situation to make sure people 5833 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 5834 to the OpenSSL toolkit. 5835 [Ralf S. Engelschall] 5836 5837 *) General source tree makefile cleanups: Made `making xxx in yyy...' 5838 display consistent in the source tree and replaced `/bin/rm' by `rm'. 5839 Additonally cleaned up the `make links' target: Remove unnecessary 5840 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 5841 to speed processing and no longer clutter the display with confusing 5842 stuff. Instead only the actually done links are displayed. 5843 [Ralf S. Engelschall] 5844 5845 *) Permit null encryption ciphersuites, used for authentication only. It used 5846 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 5847 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 5848 encryption. 5849 [Ben Laurie] 5850 5851 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 5852 signed attributes when verifying signatures (this would break them), 5853 the detached data encoding was wrong and public keys obtained using 5854 X509_get_pubkey() weren't freed. 5855 [Steve Henson] 5856 5857 *) Add text documentation for the BUFFER functions. Also added a work around 5858 to a Win95 console bug. This was triggered by the password read stuff: the 5859 last character typed gets carried over to the next fread(). If you were 5860 generating a new cert request using 'req' for example then the last 5861 character of the passphrase would be CR which would then enter the first 5862 field as blank. 5863 [Steve Henson] 5864 5865 *) Added the new `Includes OpenSSL Cryptography Software' button as 5866 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 5867 button and can be used by applications based on OpenSSL to show the 5868 relationship to the OpenSSL project. 5869 [Ralf S. Engelschall] 5870 5871 *) Remove confusing variables in function signatures in files 5872 ssl/ssl_lib.c and ssl/ssl.h. 5873 [Lennart Bong <lob@kulthea.stacken.kth.se>] 5874 5875 *) Don't install bss_file.c under PREFIX/include/ 5876 [Lennart Bong <lob@kulthea.stacken.kth.se>] 5877 5878 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 5879 functions that return function pointers and has support for NT specific 5880 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 5881 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 5882 unsigned to signed types: this was killing the Win32 compile. 5883 [Steve Henson] 5884 5885 *) Add new certificate file to stack functions, 5886 SSL_add_dir_cert_subjects_to_stack() and 5887 SSL_add_file_cert_subjects_to_stack(). These largely supplant 5888 SSL_load_client_CA_file(), and can be used to add multiple certs easily 5889 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 5890 This means that Apache-SSL and similar packages don't have to mess around 5891 to add as many CAs as they want to the preferred list. 5892 [Ben Laurie] 5893 5894 *) Experiment with doxygen documentation. Currently only partially applied to 5895 ssl/ssl_lib.c. 5896 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 5897 openssl.doxy as the configuration file. 5898 [Ben Laurie] 5899 5900 *) Get rid of remaining C++-style comments which strict C compilers hate. 5901 [Ralf S. Engelschall, pointed out by Carlos Amengual] 5902 5903 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 5904 compiled in by default: it has problems with large keys. 5905 [Steve Henson] 5906 5907 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 5908 DH private keys and/or callback functions which directly correspond to 5909 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 5910 is needed for applications which have to configure certificates on a 5911 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 5912 (e.g. s_server). 5913 For the RSA certificate situation is makes no difference, but 5914 for the DSA certificate situation this fixes the "no shared cipher" 5915 problem where the OpenSSL cipher selection procedure failed because the 5916 temporary keys were not overtaken from the context and the API provided 5917 no way to reconfigure them. 5918 The new functions now let applications reconfigure the stuff and they 5919 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 5920 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 5921 non-public-API function ssl_cert_instantiate() is used as a helper 5922 function and also to reduce code redundancy inside ssl_rsa.c. 5923 [Ralf S. Engelschall] 5924 5925 *) Move s_server -dcert and -dkey options out of the undocumented feature 5926 area because they are useful for the DSA situation and should be 5927 recognized by the users. 5928 [Ralf S. Engelschall] 5929 5930 *) Fix the cipher decision scheme for export ciphers: the export bits are 5931 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 5932 SSL_EXP_MASK. So, the original variable has to be used instead of the 5933 already masked variable. 5934 [Richard Levitte <levitte@stacken.kth.se>] 5935 5936 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 5937 [Richard Levitte <levitte@stacken.kth.se>] 5938 5939 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 5940 from `int' to `unsigned int' because it's a length and initialized by 5941 EVP_DigestFinal() which expects an `unsigned int *'. 5942 [Richard Levitte <levitte@stacken.kth.se>] 5943 5944 *) Don't hard-code path to Perl interpreter on shebang line of Configure 5945 script. Instead use the usual Shell->Perl transition trick. 5946 [Ralf S. Engelschall] 5947 5948 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 5949 (in addition to RSA certificates) to match the behaviour of `openssl dsa 5950 -noout -modulus' as it's already the case for `openssl rsa -noout 5951 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 5952 currently the public key is printed (a decision which was already done by 5953 `openssl dsa -modulus' in the past) which serves a similar purpose. 5954 Additionally the NO_RSA no longer completely removes the whole -modulus 5955 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 5956 now, too. 5957 [Ralf S. Engelschall] 5958 5959 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 5960 BIO. See the source (crypto/evp/bio_ok.c) for more info. 5961 [Arne Ansper <arne@ats.cyber.ee>] 5962 5963 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 5964 to be added. Now both 'req' and 'ca' can use new objects defined in the 5965 config file. 5966 [Steve Henson] 5967 5968 *) Add cool BIO that does syslog (or event log on NT). 5969 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 5970 5971 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 5972 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 5973 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 5974 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 5975 [Ben Laurie] 5976 5977 *) Add preliminary config info for new extension code. 5978 [Steve Henson] 5979 5980 *) Make RSA_NO_PADDING really use no padding. 5981 [Ulf Moeller <ulf@fitug.de>] 5982 5983 *) Generate errors when private/public key check is done. 5984 [Ben Laurie] 5985 5986 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 5987 for some CRL extensions and new objects added. 5988 [Steve Henson] 5989 5990 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 5991 key usage extension and fuller support for authority key id. 5992 [Steve Henson] 5993 5994 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 5995 padding method for RSA, which is recommended for new applications in PKCS 5996 #1 v2.0 (RFC 2437, October 1998). 5997 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 5998 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 5999 against Bleichbacher's attack on RSA. 6000 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 6001 Ben Laurie] 6002 6003 *) Updates to the new SSL compression code 6004 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 6005 6006 *) Fix so that the version number in the master secret, when passed 6007 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 6008 (because the server will not accept higher), that the version number 6009 is 0x03,0x01, not 0x03,0x00 6010 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 6011 6012 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 6013 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 6014 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 6015 [Steve Henson] 6016 6017 *) Support for RAW extensions where an arbitrary extension can be 6018 created by including its DER encoding. See apps/openssl.cnf for 6019 an example. 6020 [Steve Henson] 6021 6022 *) Make sure latest Perl versions don't interpret some generated C array 6023 code as Perl array code in the crypto/err/err_genc.pl script. 6024 [Lars Weber <3weber@informatik.uni-hamburg.de>] 6025 6026 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 6027 not many people have the assembler. Various Win32 compilation fixes and 6028 update to the INSTALL.W32 file with (hopefully) more accurate Win32 6029 build instructions. 6030 [Steve Henson] 6031 6032 *) Modify configure script 'Configure' to automatically create crypto/date.h 6033 file under Win32 and also build pem.h from pem.org. New script 6034 util/mkfiles.pl to create the MINFO file on environments that can't do a 6035 'make files': perl util/mkfiles.pl >MINFO should work. 6036 [Steve Henson] 6037 6038 *) Major rework of DES function declarations, in the pursuit of correctness 6039 and purity. As a result, many evil casts evaporated, and some weirdness, 6040 too. You may find this causes warnings in your code. Zapping your evil 6041 casts will probably fix them. Mostly. 6042 [Ben Laurie] 6043 6044 *) Fix for a typo in asn1.h. Bug fix to object creation script 6045 obj_dat.pl. It considered a zero in an object definition to mean 6046 "end of object": none of the objects in objects.h have any zeros 6047 so it wasn't spotted. 6048 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 6049 6050 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 6051 Masking (CBCM). In the absence of test vectors, the best I have been able 6052 to do is check that the decrypt undoes the encrypt, so far. Send me test 6053 vectors if you have them. 6054 [Ben Laurie] 6055 6056 *) Correct calculation of key length for export ciphers (too much space was 6057 allocated for null ciphers). This has not been tested! 6058 [Ben Laurie] 6059 6060 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 6061 message is now correct (it understands "crypto" and "ssl" on its 6062 command line). There is also now an "update" option. This will update 6063 the util/ssleay.num and util/libeay.num files with any new functions. 6064 If you do a: 6065 perl util/mkdef.pl crypto ssl update 6066 it will update them. 6067 [Steve Henson] 6068 6069 *) Overhauled the Perl interface (perl/*): 6070 - ported BN stuff to OpenSSL's different BN library 6071 - made the perl/ source tree CVS-aware 6072 - renamed the package from SSLeay to OpenSSL (the files still contain 6073 their history because I've copied them in the repository) 6074 - removed obsolete files (the test scripts will be replaced 6075 by better Test::Harness variants in the future) 6076 [Ralf S. Engelschall] 6077 6078 *) First cut for a very conservative source tree cleanup: 6079 1. merge various obsolete readme texts into doc/ssleay.txt 6080 where we collect the old documents and readme texts. 6081 2. remove the first part of files where I'm already sure that we no 6082 longer need them because of three reasons: either they are just temporary 6083 files which were left by Eric or they are preserved original files where 6084 I've verified that the diff is also available in the CVS via "cvs diff 6085 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 6086 the crypto/md/ stuff). 6087 [Ralf S. Engelschall] 6088 6089 *) More extension code. Incomplete support for subject and issuer alt 6090 name, issuer and authority key id. Change the i2v function parameters 6091 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 6092 what that's for :-) Fix to ASN1 macro which messed up 6093 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 6094 [Steve Henson] 6095 6096 *) Preliminary support for ENUMERATED type. This is largely copied from the 6097 INTEGER code. 6098 [Steve Henson] 6099 6100 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 6101 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 6102 6103 *) Make sure `make rehash' target really finds the `openssl' program. 6104 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 6105 6106 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 6107 like to hear about it if this slows down other processors. 6108 [Ben Laurie] 6109 6110 *) Add CygWin32 platform information to Configure script. 6111 [Alan Batie <batie@aahz.jf.intel.com>] 6112 6113 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 6114 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 6115 6116 *) New program nseq to manipulate netscape certificate sequences 6117 [Steve Henson] 6118 6119 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 6120 few typos. 6121 [Steve Henson] 6122 6123 *) Fixes to BN code. Previously the default was to define BN_RECURSION 6124 but the BN code had some problems that would cause failures when 6125 doing certificate verification and some other functions. 6126 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 6127 6128 *) Add ASN1 and PEM code to support netscape certificate sequences. 6129 [Steve Henson] 6130 6131 *) Add ASN1 and PEM code to support netscape certificate sequences. 6132 [Steve Henson] 6133 6134 *) Add several PKIX and private extended key usage OIDs. 6135 [Steve Henson] 6136 6137 *) Modify the 'ca' program to handle the new extension code. Modify 6138 openssl.cnf for new extension format, add comments. 6139 [Steve Henson] 6140 6141 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 6142 and add a sample to openssl.cnf so req -x509 now adds appropriate 6143 CA extensions. 6144 [Steve Henson] 6145 6146 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 6147 error code, add initial support to X509_print() and x509 application. 6148 [Steve Henson] 6149 6150 *) Takes a deep breath and start addding X509 V3 extension support code. Add 6151 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 6152 stuff is currently isolated and isn't even compiled yet. 6153 [Steve Henson] 6154 6155 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 6156 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 6157 Removed the versions check from X509 routines when loading extensions: 6158 this allows certain broken certificates that don't set the version 6159 properly to be processed. 6160 [Steve Henson] 6161 6162 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 6163 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 6164 can still be regenerated with "make depend". 6165 [Ben Laurie] 6166 6167 *) Spelling mistake in C version of CAST-128. 6168 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 6169 6170 *) Changes to the error generation code. The perl script err-code.pl 6171 now reads in the old error codes and retains the old numbers, only 6172 adding new ones if necessary. It also only changes the .err files if new 6173 codes are added. The makefiles have been modified to only insert errors 6174 when needed (to avoid needlessly modifying header files). This is done 6175 by only inserting errors if the .err file is newer than the auto generated 6176 C file. To rebuild all the error codes from scratch (the old behaviour) 6177 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 6178 or delete all the .err files. 6179 [Steve Henson] 6180 6181 *) CAST-128 was incorrectly implemented for short keys. The C version has 6182 been fixed, but is untested. The assembler versions are also fixed, but 6183 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 6184 to regenerate it if needed. 6185 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 6186 Hagino <itojun@kame.net>] 6187 6188 *) File was opened incorrectly in randfile.c. 6189 [Ulf M�ller <ulf@fitug.de>] 6190 6191 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 6192 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 6193 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 6194 al: it's just almost always a UTCTime. Note this patch adds new error 6195 codes so do a "make errors" if there are problems. 6196 [Steve Henson] 6197 6198 *) Correct Linux 1 recognition in config. 6199 [Ulf M�ller <ulf@fitug.de>] 6200 6201 *) Remove pointless MD5 hash when using DSA keys in ca. 6202 [Anonymous <nobody@replay.com>] 6203 6204 *) Generate an error if given an empty string as a cert directory. Also 6205 generate an error if handed NULL (previously returned 0 to indicate an 6206 error, but didn't set one). 6207 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 6208 6209 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 6210 [Ben Laurie] 6211 6212 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 6213 parameters. This was causing a warning which killed off the Win32 compile. 6214 [Steve Henson] 6215 6216 *) Remove C++ style comments from crypto/bn/bn_local.h. 6217 [Neil Costigan <neil.costigan@celocom.com>] 6218 6219 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 6220 based on a text string, looking up short and long names and finally 6221 "dot" format. The "dot" format stuff didn't work. Added new function 6222 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 6223 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 6224 OID is not part of the table. 6225 [Steve Henson] 6226 6227 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 6228 X509_LOOKUP_by_alias(). 6229 [Ben Laurie] 6230 6231 *) Sort openssl functions by name. 6232 [Ben Laurie] 6233 6234 *) Get the gendsa program working (hopefully) and add it to app list. Remove 6235 encryption from sample DSA keys (in case anyone is interested the password 6236 was "1234"). 6237 [Steve Henson] 6238 6239 *) Make _all_ *_free functions accept a NULL pointer. 6240 [Frans Heymans <fheymans@isaserver.be>] 6241 6242 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 6243 NULL pointers. 6244 [Anonymous <nobody@replay.com>] 6245 6246 *) s_server should send the CAfile as acceptable CAs, not its own cert. 6247 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 6248 6249 *) Don't blow it for numeric -newkey arguments to apps/req. 6250 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 6251 6252 *) Temp key "for export" tests were wrong in s3_srvr.c. 6253 [Anonymous <nobody@replay.com>] 6254 6255 *) Add prototype for temp key callback functions 6256 SSL_CTX_set_tmp_{rsa,dh}_callback(). 6257 [Ben Laurie] 6258 6259 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 6260 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 6261 [Steve Henson] 6262 6263 *) X509_name_add_entry() freed the wrong thing after an error. 6264 [Arne Ansper <arne@ats.cyber.ee>] 6265 6266 *) rsa_eay.c would attempt to free a NULL context. 6267 [Arne Ansper <arne@ats.cyber.ee>] 6268 6269 *) BIO_s_socket() had a broken should_retry() on Windoze. 6270 [Arne Ansper <arne@ats.cyber.ee>] 6271 6272 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 6273 [Arne Ansper <arne@ats.cyber.ee>] 6274 6275 *) Make sure the already existing X509_STORE->depth variable is initialized 6276 in X509_STORE_new(), but document the fact that this variable is still 6277 unused in the certificate verification process. 6278 [Ralf S. Engelschall] 6279 6280 *) Fix the various library and apps files to free up pkeys obtained from 6281 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 6282 [Steve Henson] 6283 6284 *) Fix reference counting in X509_PUBKEY_get(). This makes 6285 demos/maurice/example2.c work, amongst others, probably. 6286 [Steve Henson and Ben Laurie] 6287 6288 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 6289 `openssl' and second, the shortcut symlinks for the `openssl <command>' 6290 are no longer created. This way we have a single and consistent command 6291 line interface `openssl <command>', similar to `cvs <command>'. 6292 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 6293 6294 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 6295 BIT STRING wrapper always have zero unused bits. 6296 [Steve Henson] 6297 6298 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 6299 [Steve Henson] 6300 6301 *) Make the top-level INSTALL documentation easier to understand. 6302 [Paul Sutton] 6303 6304 *) Makefiles updated to exit if an error occurs in a sub-directory 6305 make (including if user presses ^C) [Paul Sutton] 6306 6307 *) Make Montgomery context stuff explicit in RSA data structure. 6308 [Ben Laurie] 6309 6310 *) Fix build order of pem and err to allow for generated pem.h. 6311 [Ben Laurie] 6312 6313 *) Fix renumbering bug in X509_NAME_delete_entry(). 6314 [Ben Laurie] 6315 6316 *) Enhanced the err-ins.pl script so it makes the error library number 6317 global and can add a library name. This is needed for external ASN1 and 6318 other error libraries. 6319 [Steve Henson] 6320 6321 *) Fixed sk_insert which never worked properly. 6322 [Steve Henson] 6323 6324 *) Fix ASN1 macros so they can handle indefinite length construted 6325 EXPLICIT tags. Some non standard certificates use these: they can now 6326 be read in. 6327 [Steve Henson] 6328 6329 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 6330 into a single doc/ssleay.txt bundle. This way the information is still 6331 preserved but no longer messes up this directory. Now it's new room for 6332 the new set of documenation files. 6333 [Ralf S. Engelschall] 6334 6335 *) SETs were incorrectly DER encoded. This was a major pain, because they 6336 shared code with SEQUENCEs, which aren't coded the same. This means that 6337 almost everything to do with SETs or SEQUENCEs has either changed name or 6338 number of arguments. 6339 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 6340 6341 *) Fix test data to work with the above. 6342 [Ben Laurie] 6343 6344 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 6345 was already fixed by Eric for 0.9.1 it seems. 6346 [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 6347 6348 *) Autodetect FreeBSD3. 6349 [Ben Laurie] 6350 6351 *) Fix various bugs in Configure. This affects the following platforms: 6352 nextstep 6353 ncr-scde 6354 unixware-2.0 6355 unixware-2.0-pentium 6356 sco5-cc. 6357 [Ben Laurie] 6358 6359 *) Eliminate generated files from CVS. Reorder tests to regenerate files 6360 before they are needed. 6361 [Ben Laurie] 6362 6363 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 6364 [Ben Laurie] 6365 6366 6367 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 6368 6369 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 6370 changed SSLeay to OpenSSL in version strings. 6371 [Ralf S. Engelschall] 6372 6373 *) Some fixups to the top-level documents. 6374 [Paul Sutton] 6375 6376 *) Fixed the nasty bug where rsaref.h was not found under compile-time 6377 because the symlink to include/ was missing. 6378 [Ralf S. Engelschall] 6379 6380 *) Incorporated the popular no-RSA/DSA-only patches 6381 which allow to compile a RSA-free SSLeay. 6382 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 6383 6384 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 6385 when "ssleay" is still not found. 6386 [Ralf S. Engelschall] 6387 6388 *) Added more platforms to Configure: Cray T3E, HPUX 11, 6389 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 6390 6391 *) Updated the README file. 6392 [Ralf S. Engelschall] 6393 6394 *) Added various .cvsignore files in the CVS repository subdirs 6395 to make a "cvs update" really silent. 6396 [Ralf S. Engelschall] 6397 6398 *) Recompiled the error-definition header files and added 6399 missing symbols to the Win32 linker tables. 6400 [Ralf S. Engelschall] 6401 6402 *) Cleaned up the top-level documents; 6403 o new files: CHANGES and LICENSE 6404 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 6405 o merged COPYRIGHT into LICENSE 6406 o removed obsolete TODO file 6407 o renamed MICROSOFT to INSTALL.W32 6408 [Ralf S. Engelschall] 6409 6410 *) Removed dummy files from the 0.9.1b source tree: 6411 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 6412 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 6413 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 6414 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 6415 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 6416 [Ralf S. Engelschall] 6417 6418 *) Added various platform portability fixes. 6419 [Mark J. Cox] 6420 6421 *) The Genesis of the OpenSSL rpject: 6422 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 6423 Young and Tim J. Hudson created while they were working for C2Net until 6424 summer 1998. 6425 [The OpenSSL Project] 6426 6427 6428 Changes between 0.9.0b and 0.9.1b [not released] 6429 6430 *) Updated a few CA certificates under certs/ 6431 [Eric A. Young] 6432 6433 *) Changed some BIGNUM api stuff. 6434 [Eric A. Young] 6435 6436 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 6437 DGUX x86, Linux Alpha, etc. 6438 [Eric A. Young] 6439 6440 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 6441 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 6442 available). 6443 [Eric A. Young] 6444 6445 *) Add -strparse option to asn1pars program which parses nested 6446 binary structures 6447 [Dr Stephen Henson <shenson@bigfoot.com>] 6448 6449 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 6450 [Eric A. Young] 6451 6452 *) DSA fix for "ca" program. 6453 [Eric A. Young] 6454 6455 *) Added "-genkey" option to "dsaparam" program. 6456 [Eric A. Young] 6457 6458 *) Added RIPE MD160 (rmd160) message digest. 6459 [Eric A. Young] 6460 6461 *) Added -a (all) option to "ssleay version" command. 6462 [Eric A. Young] 6463 6464 *) Added PLATFORM define which is the id given to Configure. 6465 [Eric A. Young] 6466 6467 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 6468 [Eric A. Young] 6469 6470 *) Extended the ASN.1 parser routines. 6471 [Eric A. Young] 6472 6473 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 6474 [Eric A. Young] 6475 6476 *) Added a BN_CTX to the BN library. 6477 [Eric A. Young] 6478 6479 *) Fixed the weak key values in DES library 6480 [Eric A. Young] 6481 6482 *) Changed API in EVP library for cipher aliases. 6483 [Eric A. Young] 6484 6485 *) Added support for RC2/64bit cipher. 6486 [Eric A. Young] 6487 6488 *) Converted the lhash library to the crypto/mem.c functions. 6489 [Eric A. Young] 6490 6491 *) Added more recognized ASN.1 object ids. 6492 [Eric A. Young] 6493 6494 *) Added more RSA padding checks for SSL/TLS. 6495 [Eric A. Young] 6496 6497 *) Added BIO proxy/filter functionality. 6498 [Eric A. Young] 6499 6500 *) Added extra_certs to SSL_CTX which can be used 6501 send extra CA certificates to the client in the CA cert chain sending 6502 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 6503 [Eric A. Young] 6504 6505 *) Now Fortezza is denied in the authentication phase because 6506 this is key exchange mechanism is not supported by SSLeay at all. 6507 [Eric A. Young] 6508 6509 *) Additional PKCS1 checks. 6510 [Eric A. Young] 6511 6512 *) Support the string "TLSv1" for all TLS v1 ciphers. 6513 [Eric A. Young] 6514 6515 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 6516 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 6517 [Eric A. Young] 6518 6519 *) Fixed a few memory leaks. 6520 [Eric A. Young] 6521 6522 *) Fixed various code and comment typos. 6523 [Eric A. Young] 6524 6525 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 6526 bytes sent in the client random. 6527 [Edward Bishop <ebishop@spyglass.com>] 6528 6529