CHANGES revision 72613
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 0.9.6 and 0.9.6a [xx XXX 2001] 6 7 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 8 Bleichenbacher's DSA attack. 9 [Ulf Moeller] 10 11 *) In the NCONF_...-based implementations for CONF_... queries 12 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 13 a temporary CONF structure with the data component set to NULL 14 (which gives segmentation faults in lh_retrieve). 15 Instead, use NULL for the CONF pointer in CONF_get_string and 16 CONF_get_number (which may use environment variables) and directly 17 return NULL from CONF_get_section. 18 [Bodo Moeller] 19 20 *) Fix potential buffer overrun for EBCDIC. 21 [Ulf Moeller] 22 23 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 24 keyUsage if basicConstraints absent for a CA. 25 [Steve Henson] 26 27 *) Make SMIME_write_PKCS7() write mail header values with a format that 28 is more generally accepted (no spaces before the semicolon), since 29 some programs can't parse those values properly otherwise. Also make 30 sure BIO's that break lines after each write do not create invalid 31 headers. 32 [Richard Levitte] 33 34 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 35 macros previously used would not encode an empty SEQUENCE OF 36 and break the signature. 37 [Steve Henson] 38 39 *) Zero the premaster secret after deriving the master secret in 40 DH ciphersuites. 41 [Steve Henson] 42 43 *) Add some EVP_add_digest_alias registrations (as found in 44 OpenSSL_add_all_digests()) to SSL_library_init() 45 aka OpenSSL_add_ssl_algorithms(). This provides improved 46 compatibility with peers using X.509 certificates 47 with unconventional AlgorithmIdentifier OIDs. 48 [Bodo Moeller] 49 50 *) Fix for Irix with NO_ASM. 51 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 52 53 *) ./config script fixes. 54 [Ulf Moeller, Richard Levitte] 55 56 *) Fix 'openssl passwd -1'. 57 [Bodo Moeller] 58 59 *) Change PKCS12_key_gen_asc() so it can cope with non null 60 terminated strings whose length is passed in the passlen 61 parameter, for example from PEM callbacks. This was done 62 by adding an extra length parameter to asc2uni(). 63 [Steve Henson, reported by <oddissey@samsung.co.kr>] 64 65 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 66 call failed, free the DSA structure. 67 [Bodo Moeller] 68 69 *) Fix to uni2asc() to cope with zero length Unicode strings. 70 These are present in some PKCS#12 files. 71 [Steve Henson] 72 73 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 74 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 75 when writing a 32767 byte record. 76 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 77 78 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 79 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 80 81 (RSA objects have a reference count access to which is protected 82 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 83 so they are meant to be shared between threads.) 84 [Bodo Moeller, Geoff Thorpe; original patch submitted by 85 "Reddie, Steven" <Steven.Reddie@ca.com>] 86 87 *) Fix a deadlock in CRYPTO_mem_leaks(). 88 [Bodo Moeller] 89 90 *) rand_win.c fix for Borland C. 91 [Ulf M�ller] 92 93 *) BN_rshift bugfix for n == 0. 94 [Bodo Moeller] 95 96 *) Store verify_result within SSL_SESSION also for client side to 97 avoid potential security hole. (Re-used sessions on the client side 98 always resulted in verify_result==X509_V_OK, not using the original 99 result of the server certificate verification.) 100 [Lutz Jaenicke] 101 102 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 103 SSL3_RT_APPLICATION_DATA, return 0. 104 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 105 [Bodo Moeller] 106 107 *) Fix SSL_peek: 108 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 109 releases, have been re-implemented by renaming the previous 110 implementations of ssl2_read and ssl3_read to ssl2_read_internal 111 and ssl3_read_internal, respectively, and adding 'peek' parameters 112 to them. The new ssl[23]_{read,peek} functions are calls to 113 ssl[23]_read_internal with the 'peek' flag set appropriately. 114 A 'peek' parameter has also been added to ssl3_read_bytes, which 115 does the actual work for ssl3_read_internal. 116 [Bodo Moeller] 117 118 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 119 The previous value, 12, was not always sufficient for BN_mod_exp(). 120 [Bodo Moeller] 121 122 *) Fix typo in get_cert_by_subject() in by_dir.c 123 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 124 125 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 126 127 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 128 and not in SSL_clear because the latter is also used by the 129 accept/connect functions; previously, the settings made by 130 SSL_set_read_ahead would be lost during the handshake. 131 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 132 133 *) Correct util/mkdef.pl to be selective about disabled algorithms. 134 Previously, it would create entries for disableed algorithms no 135 matter what. 136 [Richard Levitte] 137 138 *) Added several new manual pages for SSL_* function. 139 [Lutz Jaenicke] 140 141 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 142 143 *) In ssl23_get_client_hello, generate an error message when faced 144 with an initial SSL 3.0/TLS record that is too small to contain the 145 first two bytes of the ClientHello message, i.e. client_version. 146 (Note that this is a pathologic case that probably has never happened 147 in real life.) The previous approach was to use the version number 148 from the record header as a substitute; but our protocol choice 149 should not depend on that one because it is not authenticated 150 by the Finished messages. 151 [Bodo Moeller] 152 153 *) More robust randomness gathering functions for Windows. 154 [Jeffrey Altman <jaltman@columbia.edu>] 155 156 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 157 not set then we don't setup the error code for issuer check errors 158 to avoid possibly overwriting other errors which the callback does 159 handle. If an application does set the flag then we assume it knows 160 what it is doing and can handle the new informational codes 161 appropriately. 162 [Steve Henson] 163 164 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 165 a general "ANY" type, as such it should be able to decode anything 166 including tagged types. However it didn't check the class so it would 167 wrongly interpret tagged types in the same way as their universal 168 counterpart and unknown types were just rejected. Changed so that the 169 tagged and unknown types are handled in the same way as a SEQUENCE: 170 that is the encoding is stored intact. There is also a new type 171 "V_ASN1_OTHER" which is used when the class is not universal, in this 172 case we have no idea what the actual type is so we just lump them all 173 together. 174 [Steve Henson] 175 176 *) On VMS, stdout may very well lead to a file that is written to 177 in a record-oriented fashion. That means that every write() will 178 write a separate record, which will be read separately by the 179 programs trying to read from it. This can be very confusing. 180 181 The solution is to put a BIO filter in the way that will buffer 182 text until a linefeed is reached, and then write everything a 183 line at a time, so every record written will be an actual line, 184 not chunks of lines and not (usually doesn't happen, but I've 185 seen it once) several lines in one record. BIO_f_linebuffer() is 186 the answer. 187 188 Currently, it's a VMS-only method, because that's where it has 189 been tested well enough. 190 [Richard Levitte] 191 192 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 193 it can return incorrect results. 194 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 195 but it was in 0.9.6-beta[12].) 196 [Bodo Moeller] 197 198 *) Disable the check for content being present when verifying detached 199 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 200 include zero length content when signing messages. 201 [Steve Henson] 202 203 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 204 BIO_ctrl (for BIO pairs). 205 [Bodo M�ller] 206 207 *) Add DSO method for VMS. 208 [Richard Levitte] 209 210 *) Bug fix: Montgomery multiplication could produce results with the 211 wrong sign. 212 [Ulf M�ller] 213 214 *) Add RPM specification openssl.spec and modify it to build three 215 packages. The default package contains applications, application 216 documentation and run-time libraries. The devel package contains 217 include files, static libraries and function documentation. The 218 doc package contains the contents of the doc directory. The original 219 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 220 [Richard Levitte] 221 222 *) Add a large number of documentation files for many SSL routines. 223 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 224 225 *) Add a configuration entry for Sony News 4. 226 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 227 228 *) Don't set the two most significant bits to one when generating a 229 random number < q in the DSA library. 230 [Ulf M�ller] 231 232 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 233 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 234 the underlying transport is blocking) if a handshake took place. 235 (The default behaviour is needed by applications such as s_client 236 and s_server that use select() to determine when to use SSL_read; 237 but for applications that know in advance when to expect data, it 238 just makes things more complicated.) 239 [Bodo Moeller] 240 241 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 242 from EGD. 243 [Ben Laurie] 244 245 *) Add a few more EBCDIC conditionals that make `req' and `x509' 246 work better on such systems. 247 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 248 249 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 250 Update PKCS12_parse() so it copies the friendlyName and the 251 keyid to the certificates aux info. 252 [Steve Henson] 253 254 *) Fix bug in PKCS7_verify() which caused an infinite loop 255 if there was more than one signature. 256 [Sven Uszpelkat <su@celocom.de>] 257 258 *) Major change in util/mkdef.pl to include extra information 259 about each symbol, as well as presentig variables as well 260 as functions. This change means that there's n more need 261 to rebuild the .num files when some algorithms are excluded. 262 [Richard Levitte] 263 264 *) Allow the verify time to be set by an application, 265 rather than always using the current time. 266 [Steve Henson] 267 268 *) Phase 2 verify code reorganisation. The certificate 269 verify code now looks up an issuer certificate by a 270 number of criteria: subject name, authority key id 271 and key usage. It also verifies self signed certificates 272 by the same criteria. The main comparison function is 273 X509_check_issued() which performs these checks. 274 275 Lot of changes were necessary in order to support this 276 without completely rewriting the lookup code. 277 278 Authority and subject key identifier are now cached. 279 280 The LHASH 'certs' is X509_STORE has now been replaced 281 by a STACK_OF(X509_OBJECT). This is mainly because an 282 LHASH can't store or retrieve multiple objects with 283 the same hash value. 284 285 As a result various functions (which were all internal 286 use only) have changed to handle the new X509_STORE 287 structure. This will break anything that messed round 288 with X509_STORE internally. 289 290 The functions X509_STORE_add_cert() now checks for an 291 exact match, rather than just subject name. 292 293 The X509_STORE API doesn't directly support the retrieval 294 of multiple certificates matching a given criteria, however 295 this can be worked round by performing a lookup first 296 (which will fill the cache with candidate certificates) 297 and then examining the cache for matches. This is probably 298 the best we can do without throwing out X509_LOOKUP 299 entirely (maybe later...). 300 301 The X509_VERIFY_CTX structure has been enhanced considerably. 302 303 All certificate lookup operations now go via a get_issuer() 304 callback. Although this currently uses an X509_STORE it 305 can be replaced by custom lookups. This is a simple way 306 to bypass the X509_STORE hackery necessary to make this 307 work and makes it possible to use more efficient techniques 308 in future. A very simple version which uses a simple 309 STACK for its trusted certificate store is also provided 310 using X509_STORE_CTX_trusted_stack(). 311 312 The verify_cb() and verify() callbacks now have equivalents 313 in the X509_STORE_CTX structure. 314 315 X509_STORE_CTX also has a 'flags' field which can be used 316 to customise the verify behaviour. 317 [Steve Henson] 318 319 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 320 excludes S/MIME capabilities. 321 [Steve Henson] 322 323 *) When a certificate request is read in keep a copy of the 324 original encoding of the signed data and use it when outputing 325 again. Signatures then use the original encoding rather than 326 a decoded, encoded version which may cause problems if the 327 request is improperly encoded. 328 [Steve Henson] 329 330 *) For consistency with other BIO_puts implementations, call 331 buffer_write(b, ...) directly in buffer_puts instead of calling 332 BIO_write(b, ...). 333 334 In BIO_puts, increment b->num_write as in BIO_write. 335 [Peter.Sylvester@EdelWeb.fr] 336 337 *) Fix BN_mul_word for the case where the word is 0. (We have to use 338 BN_zero, we may not return a BIGNUM with an array consisting of 339 words set to zero.) 340 [Bodo Moeller] 341 342 *) Avoid calling abort() from within the library when problems are 343 detected, except if preprocessor symbols have been defined 344 (such as REF_CHECK, BN_DEBUG etc.). 345 [Bodo Moeller] 346 347 *) New openssl application 'rsautl'. This utility can be 348 used for low level RSA operations. DER public key 349 BIO/fp routines also added. 350 [Steve Henson] 351 352 *) New Configure entry and patches for compiling on QNX 4. 353 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 354 355 *) A demo state-machine implementation was sponsored by 356 Nuron (http://www.nuron.com/) and is now available in 357 demos/state_machine. 358 [Ben Laurie] 359 360 *) New options added to the 'dgst' utility for signature 361 generation and verification. 362 [Steve Henson] 363 364 *) Unrecognized PKCS#7 content types are now handled via a 365 catch all ASN1_TYPE structure. This allows unsupported 366 types to be stored as a "blob" and an application can 367 encode and decode it manually. 368 [Steve Henson] 369 370 *) Fix various signed/unsigned issues to make a_strex.c 371 compile under VC++. 372 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 373 374 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 375 length if passed a buffer. ASN1_INTEGER_to_BN failed 376 if passed a NULL BN and its argument was negative. 377 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 378 379 *) Modification to PKCS#7 encoding routines to output definite 380 length encoding. Since currently the whole structures are in 381 memory there's not real point in using indefinite length 382 constructed encoding. However if OpenSSL is compiled with 383 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 384 [Steve Henson] 385 386 *) Added BIO_vprintf() and BIO_vsnprintf(). 387 [Richard Levitte] 388 389 *) Added more prefixes to parse for in the the strings written 390 through a logging bio, to cover all the levels that are available 391 through syslog. The prefixes are now: 392 393 PANIC, EMERG, EMR => LOG_EMERG 394 ALERT, ALR => LOG_ALERT 395 CRIT, CRI => LOG_CRIT 396 ERROR, ERR => LOG_ERR 397 WARNING, WARN, WAR => LOG_WARNING 398 NOTICE, NOTE, NOT => LOG_NOTICE 399 INFO, INF => LOG_INFO 400 DEBUG, DBG => LOG_DEBUG 401 402 and as before, if none of those prefixes are present at the 403 beginning of the string, LOG_ERR is chosen. 404 405 On Win32, the LOG_* levels are mapped according to this: 406 407 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 408 LOG_WARNING => EVENTLOG_WARNING_TYPE 409 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 410 411 [Richard Levitte] 412 413 *) Made it possible to reconfigure with just the configuration 414 argument "reconf" or "reconfigure". The command line arguments 415 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 416 and are retrieved from there when reconfiguring. 417 [Richard Levitte] 418 419 *) MD4 implemented. 420 [Assar Westerlund <assar@sics.se>, Richard Levitte] 421 422 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 423 [Richard Levitte] 424 425 *) The obj_dat.pl script was messing up the sorting of object 426 names. The reason was that it compared the quoted version 427 of strings as a result "OCSP" > "OCSP Signing" because 428 " > SPACE. Changed script to store unquoted versions of 429 names and add quotes on output. It was also omitting some 430 names from the lookup table if they were given a default 431 value (that is if SN is missing it is given the same 432 value as LN and vice versa), these are now added on the 433 grounds that if an object has a name we should be able to 434 look it up. Finally added warning output when duplicate 435 short or long names are found. 436 [Steve Henson] 437 438 *) Changes needed for Tandem NSK. 439 [Scott Uroff <scott@xypro.com>] 440 441 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 442 RSA_padding_check_SSLv23(), special padding was never detected 443 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 444 version rollback attacks was not effective. 445 446 In s23_clnt.c, don't use special rollback-attack detection padding 447 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 448 client; similarly, in s23_srvr.c, don't do the rollback check if 449 SSL 2.0 is the only protocol enabled in the server. 450 [Bodo Moeller] 451 452 *) Make it possible to get hexdumps of unprintable data with 'openssl 453 asn1parse'. By implication, the functions ASN1_parse_dump() and 454 BIO_dump_indent() are added. 455 [Richard Levitte] 456 457 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 458 these print out strings and name structures based on various 459 flags including RFC2253 support and proper handling of 460 multibyte characters. Added options to the 'x509' utility 461 to allow the various flags to be set. 462 [Steve Henson] 463 464 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 465 Also change the functions X509_cmp_current_time() and 466 X509_gmtime_adj() work with an ASN1_TIME structure, 467 this will enable certificates using GeneralizedTime in validity 468 dates to be checked. 469 [Steve Henson] 470 471 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 472 negative public key encodings) on by default, 473 NO_NEG_PUBKEY_BUG can be set to disable it. 474 [Steve Henson] 475 476 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 477 content octets. An i2c_ASN1_OBJECT is unnecessary because 478 the encoding can be trivially obtained from the structure. 479 [Steve Henson] 480 481 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 482 not read locks (CRYPTO_r_[un]lock). 483 [Bodo Moeller] 484 485 *) A first attempt at creating official support for shared 486 libraries through configuration. I've kept it so the 487 default is static libraries only, and the OpenSSL programs 488 are always statically linked for now, but there are 489 preparations for dynamic linking in place. 490 This has been tested on Linux and True64. 491 [Richard Levitte] 492 493 *) Randomness polling function for Win9x, as described in: 494 Peter Gutmann, Software Generation of Practically Strong 495 Random Numbers. 496 [Ulf M�ller] 497 498 *) Fix so PRNG is seeded in req if using an already existing 499 DSA key. 500 [Steve Henson] 501 502 *) New options to smime application. -inform and -outform 503 allow alternative formats for the S/MIME message including 504 PEM and DER. The -content option allows the content to be 505 specified separately. This should allow things like Netscape 506 form signing output easier to verify. 507 [Steve Henson] 508 509 *) Fix the ASN1 encoding of tags using the 'long form'. 510 [Steve Henson] 511 512 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 513 STRING types. These convert content octets to and from the 514 underlying type. The actual tag and length octets are 515 already assumed to have been read in and checked. These 516 are needed because all other string types have virtually 517 identical handling apart from the tag. By having versions 518 of the ASN1 functions that just operate on content octets 519 IMPLICIT tagging can be handled properly. It also allows 520 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 521 and ASN1_INTEGER are identical apart from the tag. 522 [Steve Henson] 523 524 *) Change the handling of OID objects as follows: 525 526 - New object identifiers are inserted in objects.txt, following 527 the syntax given in objects.README. 528 - objects.pl is used to process obj_mac.num and create a new 529 obj_mac.h. 530 - obj_dat.pl is used to create a new obj_dat.h, using the data in 531 obj_mac.h. 532 533 This is currently kind of a hack, and the perl code in objects.pl 534 isn't very elegant, but it works as I intended. The simplest way 535 to check that it worked correctly is to look in obj_dat.h and 536 check the array nid_objs and make sure the objects haven't moved 537 around (this is important!). Additions are OK, as well as 538 consistent name changes. 539 [Richard Levitte] 540 541 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 542 [Bodo Moeller] 543 544 *) Addition of the command line parameter '-rand file' to 'openssl req'. 545 The given file adds to whatever has already been seeded into the 546 random pool through the RANDFILE configuration file option or 547 environment variable, or the default random state file. 548 [Richard Levitte] 549 550 *) mkstack.pl now sorts each macro group into lexical order. 551 Previously the output order depended on the order the files 552 appeared in the directory, resulting in needless rewriting 553 of safestack.h . 554 [Steve Henson] 555 556 *) Patches to make OpenSSL compile under Win32 again. Mostly 557 work arounds for the VC++ problem that it treats func() as 558 func(void). Also stripped out the parts of mkdef.pl that 559 added extra typesafe functions: these no longer exist. 560 [Steve Henson] 561 562 *) Reorganisation of the stack code. The macros are now all 563 collected in safestack.h . Each macro is defined in terms of 564 a "stack macro" of the form SKM_<name>(type, a, b). The 565 DEBUG_SAFESTACK is now handled in terms of function casts, 566 this has the advantage of retaining type safety without the 567 use of additional functions. If DEBUG_SAFESTACK is not defined 568 then the non typesafe macros are used instead. Also modified the 569 mkstack.pl script to handle the new form. Needs testing to see 570 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 571 the default if no major problems. Similar behaviour for ASN1_SET_OF 572 and PKCS12_STACK_OF. 573 [Steve Henson] 574 575 *) When some versions of IIS use the 'NET' form of private key the 576 key derivation algorithm is different. Normally MD5(password) is 577 used as a 128 bit RC4 key. In the modified case 578 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 579 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 580 as the old Netscape_RSA functions except they have an additional 581 'sgckey' parameter which uses the modified algorithm. Also added 582 an -sgckey command line option to the rsa utility. Thanks to 583 Adrian Peck <bertie@ncipher.com> for posting details of the modified 584 algorithm to openssl-dev. 585 [Steve Henson] 586 587 *) The evp_local.h macros were using 'c.##kname' which resulted in 588 invalid expansion on some systems (SCO 5.0.5 for example). 589 Corrected to 'c.kname'. 590 [Phillip Porch <root@theporch.com>] 591 592 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 593 a STACK of email addresses from a certificate or request, these look 594 in the subject name and the subject alternative name extensions and 595 omit any duplicate addresses. 596 [Steve Henson] 597 598 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 599 This makes DSA verification about 2 % faster. 600 [Bodo Moeller] 601 602 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 603 (meaning that now 2^5 values will be precomputed, which is only 4 KB 604 plus overhead for 1024 bit moduli). 605 This makes exponentiations about 0.5 % faster for 1024 bit 606 exponents (as measured by "openssl speed rsa2048"). 607 [Bodo Moeller] 608 609 *) Rename memory handling macros to avoid conflicts with other 610 software: 611 Malloc => OPENSSL_malloc 612 Malloc_locked => OPENSSL_malloc_locked 613 Realloc => OPENSSL_realloc 614 Free => OPENSSL_free 615 [Richard Levitte] 616 617 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 618 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 619 [Bodo Moeller] 620 621 *) CygWin32 support. 622 [John Jarvie <jjarvie@newsguy.com>] 623 624 *) The type-safe stack code has been rejigged. It is now only compiled 625 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 626 by default all type-specific stack functions are "#define"d back to 627 standard stack functions. This results in more streamlined output 628 but retains the type-safety checking possibilities of the original 629 approach. 630 [Geoff Thorpe] 631 632 *) The STACK code has been cleaned up, and certain type declarations 633 that didn't make a lot of sense have been brought in line. This has 634 also involved a cleanup of sorts in safestack.h to more correctly 635 map type-safe stack functions onto their plain stack counterparts. 636 This work has also resulted in a variety of "const"ifications of 637 lots of the code, especially "_cmp" operations which should normally 638 be prototyped with "const" parameters anyway. 639 [Geoff Thorpe] 640 641 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 642 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 643 (The PRNG state consists of two parts, the large pool 'state' and 'md', 644 where all of 'md' is used each time the PRNG is used, but 'state' 645 is used only indexed by a cyclic counter. As entropy may not be 646 well distributed from the beginning, 'md' is important as a 647 chaining variable. However, the output function chains only half 648 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 649 all of 'md', and seeding with STATE_SIZE dummy bytes will result 650 in all of 'state' being rewritten, with the new values depending 651 on virtually all of 'md'. This overcomes the 80 bit limitation.) 652 [Bodo Moeller] 653 654 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 655 the handshake is continued after ssl_verify_cert_chain(); 656 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 657 can lead to 'unexplainable' connection aborts later. 658 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 659 660 *) Major EVP API cipher revision. 661 Add hooks for extra EVP features. This allows various cipher 662 parameters to be set in the EVP interface. Support added for variable 663 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 664 setting of RC2 and RC5 parameters. 665 666 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 667 ciphers. 668 669 Remove lots of duplicated code from the EVP library. For example *every* 670 cipher init() function handles the 'iv' in the same way according to the 671 cipher mode. They also all do nothing if the 'key' parameter is NULL and 672 for CFB and OFB modes they zero ctx->num. 673 674 New functionality allows removal of S/MIME code RC2 hack. 675 676 Most of the routines have the same form and so can be declared in terms 677 of macros. 678 679 By shifting this to the top level EVP_CipherInit() it can be removed from 680 all individual ciphers. If the cipher wants to handle IVs or keys 681 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 682 flags. 683 684 Change lots of functions like EVP_EncryptUpdate() to now return a 685 value: although software versions of the algorithms cannot fail 686 any installed hardware versions can. 687 [Steve Henson] 688 689 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 690 this option is set, tolerate broken clients that send the negotiated 691 protocol version number instead of the requested protocol version 692 number. 693 [Bodo Moeller] 694 695 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 696 i.e. non-zero for export ciphersuites, zero otherwise. 697 Previous versions had this flag inverted, inconsistent with 698 rsa_tmp_cb (..._TMP_RSA_CB). 699 [Bodo Moeller; problem reported by Amit Chopra] 700 701 *) Add missing DSA library text string. Work around for some IIS 702 key files with invalid SEQUENCE encoding. 703 [Steve Henson] 704 705 *) Add a document (doc/standards.txt) that list all kinds of standards 706 and so on that are implemented in OpenSSL. 707 [Richard Levitte] 708 709 *) Enhance c_rehash script. Old version would mishandle certificates 710 with the same subject name hash and wouldn't handle CRLs at all. 711 Added -fingerprint option to crl utility, to support new c_rehash 712 features. 713 [Steve Henson] 714 715 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 716 [Ulf M�ller] 717 718 *) Fix for SSL server purpose checking. Server checking was 719 rejecting certificates which had extended key usage present 720 but no ssl client purpose. 721 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 722 723 *) Make PKCS#12 code work with no password. The PKCS#12 spec 724 is a little unclear about how a blank password is handled. 725 Since the password in encoded as a BMPString with terminating 726 double NULL a zero length password would end up as just the 727 double NULL. However no password at all is different and is 728 handled differently in the PKCS#12 key generation code. NS 729 treats a blank password as zero length. MSIE treats it as no 730 password on export: but it will try both on import. We now do 731 the same: PKCS12_parse() tries zero length and no password if 732 the password is set to "" or NULL (NULL is now a valid password: 733 it wasn't before) as does the pkcs12 application. 734 [Steve Henson] 735 736 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 737 perror when PEM_read_bio_X509_REQ fails, the error message must 738 be obtained from the error queue. 739 [Bodo Moeller] 740 741 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 742 it in ERR_remove_state if appropriate, and change ERR_get_state 743 accordingly to avoid race conditions (this is necessary because 744 thread_hash is no longer constant once set). 745 [Bodo Moeller] 746 747 *) Bugfix for linux-elf makefile.one. 748 [Ulf M�ller] 749 750 *) RSA_get_default_method() will now cause a default 751 RSA_METHOD to be chosen if one doesn't exist already. 752 Previously this was only set during a call to RSA_new() 753 or RSA_new_method(NULL) meaning it was possible for 754 RSA_get_default_method() to return NULL. 755 [Geoff Thorpe] 756 757 *) Added native name translation to the existing DSO code 758 that will convert (if the flag to do so is set) filenames 759 that are sufficiently small and have no path information 760 into a canonical native form. Eg. "blah" converted to 761 "libblah.so" or "blah.dll" etc. 762 [Geoff Thorpe] 763 764 *) New function ERR_error_string_n(e, buf, len) which is like 765 ERR_error_string(e, buf), but writes at most 'len' bytes 766 including the 0 terminator. For ERR_error_string_n, 'buf' 767 may not be NULL. 768 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 769 770 *) CONF library reworked to become more general. A new CONF 771 configuration file reader "class" is implemented as well as a 772 new functions (NCONF_*, for "New CONF") to handle it. The now 773 old CONF_* functions are still there, but are reimplemented to 774 work in terms of the new functions. Also, a set of functions 775 to handle the internal storage of the configuration data is 776 provided to make it easier to write new configuration file 777 reader "classes" (I can definitely see something reading a 778 configuration file in XML format, for example), called _CONF_*, 779 or "the configuration storage API"... 780 781 The new configuration file reading functions are: 782 783 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 784 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 785 786 NCONF_default, NCONF_WIN32 787 788 NCONF_dump_fp, NCONF_dump_bio 789 790 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 791 NCONF_new creates a new CONF object. This works in the same way 792 as other interfaces in OpenSSL, like the BIO interface. 793 NCONF_dump_* dump the internal storage of the configuration file, 794 which is useful for debugging. All other functions take the same 795 arguments as the old CONF_* functions wth the exception of the 796 first that must be a `CONF *' instead of a `LHASH *'. 797 798 To make it easer to use the new classes with the old CONF_* functions, 799 the function CONF_set_default_method is provided. 800 [Richard Levitte] 801 802 *) Add '-tls1' option to 'openssl ciphers', which was already 803 mentioned in the documentation but had not been implemented. 804 (This option is not yet really useful because even the additional 805 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 806 [Bodo Moeller] 807 808 *) Initial DSO code added into libcrypto for letting OpenSSL (and 809 OpenSSL-based applications) load shared libraries and bind to 810 them in a portable way. 811 [Geoff Thorpe, with contributions from Richard Levitte] 812 813 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 814 815 *) Make sure _lrotl and _lrotr are only used with MSVC. 816 817 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 818 (the default implementation of RAND_status). 819 820 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 821 to '-clrext' (= clear extensions), as intended and documented. 822 [Bodo Moeller; inconsistency pointed out by Michael Attili 823 <attili@amaxo.com>] 824 825 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 826 was larger than the MD block size. 827 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 828 829 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 830 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 831 using the passed key: if the passed key was a private key the result 832 of X509_print(), for example, would be to print out all the private key 833 components. 834 [Steve Henson] 835 836 *) des_quad_cksum() byte order bug fix. 837 [Ulf M�ller, using the problem description in krb4-0.9.7, where 838 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 839 840 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 841 discouraged. 842 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 843 844 *) For easily testing in shell scripts whether some command 845 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 846 returns with exit code 0 iff no command of the given name is available. 847 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 848 the output goes to stdout and nothing is printed to stderr. 849 Additional arguments are always ignored. 850 851 Since for each cipher there is a command of the same name, 852 the 'no-cipher' compilation switches can be tested this way. 853 854 ('openssl no-XXX' is not able to detect pseudo-commands such 855 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 856 [Bodo Moeller] 857 858 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 859 [Bodo Moeller] 860 861 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 862 is set; it will be thrown away anyway because each handshake creates 863 its own key. 864 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 865 to parameters -- in previous versions (since OpenSSL 0.9.3) the 866 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 867 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 868 [Bodo Moeller] 869 870 *) New s_client option -ign_eof: EOF at stdin is ignored, and 871 'Q' and 'R' lose their special meanings (quit/renegotiate). 872 This is part of what -quiet does; unlike -quiet, -ign_eof 873 does not suppress any output. 874 [Richard Levitte] 875 876 *) Add compatibility options to the purpose and trust code. The 877 purpose X509_PURPOSE_ANY is "any purpose" which automatically 878 accepts a certificate or CA, this was the previous behaviour, 879 with all the associated security issues. 880 881 X509_TRUST_COMPAT is the old trust behaviour: only and 882 automatically trust self signed roots in certificate store. A 883 new trust setting X509_TRUST_DEFAULT is used to specify that 884 a purpose has no associated trust setting and it should instead 885 use the value in the default purpose. 886 [Steve Henson] 887 888 *) Fix the PKCS#8 DSA private key code so it decodes keys again 889 and fix a memory leak. 890 [Steve Henson] 891 892 *) In util/mkerr.pl (which implements 'make errors'), preserve 893 reason strings from the previous version of the .c file, as 894 the default to have only downcase letters (and digits) in 895 automatically generated reasons codes is not always appropriate. 896 [Bodo Moeller] 897 898 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 899 using strerror. Previously, ERR_reason_error_string() returned 900 library names as reason strings for SYSerr; but SYSerr is a special 901 case where small numbers are errno values, not library numbers. 902 [Bodo Moeller] 903 904 *) Add '-dsaparam' option to 'openssl dhparam' application. This 905 converts DSA parameters into DH parameters. (When creating parameters, 906 DSA_generate_parameters is used.) 907 [Bodo Moeller] 908 909 *) Include 'length' (recommended exponent length) in C code generated 910 by 'openssl dhparam -C'. 911 [Bodo Moeller] 912 913 *) The second argument to set_label in perlasm was already being used 914 so couldn't be used as a "file scope" flag. Moved to third argument 915 which was free. 916 [Steve Henson] 917 918 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 919 instead of RAND_bytes for encryption IVs and salts. 920 [Bodo Moeller] 921 922 *) Include RAND_status() into RAND_METHOD instead of implementing 923 it only for md_rand.c Otherwise replacing the PRNG by calling 924 RAND_set_rand_method would be impossible. 925 [Bodo Moeller] 926 927 *) Don't let DSA_generate_key() enter an infinite loop if the random 928 number generation fails. 929 [Bodo Moeller] 930 931 *) New 'rand' application for creating pseudo-random output. 932 [Bodo Moeller] 933 934 *) Added configuration support for Linux/IA64 935 [Rolf Haberrecker <rolf@suse.de>] 936 937 *) Assembler module support for Mingw32. 938 [Ulf M�ller] 939 940 *) Shared library support for HPUX (in shlib/). 941 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 942 943 *) Shared library support for Solaris gcc. 944 [Lutz Behnke <behnke@trustcenter.de>] 945 946 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 947 948 *) PKCS7_encrypt() was adding text MIME headers twice because they 949 were added manually and by SMIME_crlf_copy(). 950 [Steve Henson] 951 952 *) In bntest.c don't call BN_rand with zero bits argument. 953 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 954 955 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 956 case was implemented. This caused BN_div_recp() to fail occasionally. 957 [Ulf M�ller] 958 959 *) Add an optional second argument to the set_label() in the perl 960 assembly language builder. If this argument exists and is set 961 to 1 it signals that the assembler should use a symbol whose 962 scope is the entire file, not just the current function. This 963 is needed with MASM which uses the format label:: for this scope. 964 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 965 966 *) Change the ASN1 types so they are typedefs by default. Before 967 almost all types were #define'd to ASN1_STRING which was causing 968 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 969 for example. 970 [Steve Henson] 971 972 *) Change names of new functions to the new get1/get0 naming 973 convention: After 'get1', the caller owns a reference count 974 and has to call ..._free; 'get0' returns a pointer to some 975 data structure without incrementing reference counters. 976 (Some of the existing 'get' functions increment a reference 977 counter, some don't.) 978 Similarly, 'set1' and 'add1' functions increase reference 979 counters or duplicate objects. 980 [Steve Henson] 981 982 *) Allow for the possibility of temp RSA key generation failure: 983 the code used to assume it always worked and crashed on failure. 984 [Steve Henson] 985 986 *) Fix potential buffer overrun problem in BIO_printf(). 987 [Ulf M�ller, using public domain code by Patrick Powell; problem 988 pointed out by David Sacerdote <das33@cornell.edu>] 989 990 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 991 RAND_egd() and RAND_status(). In the command line application, 992 the EGD socket can be specified like a seed file using RANDFILE 993 or -rand. 994 [Ulf M�ller] 995 996 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 997 Some CAs (e.g. Verisign) distribute certificates in this form. 998 [Steve Henson] 999 1000 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 1001 list to exclude them. This means that no special compilation option 1002 is needed to use anonymous DH: it just needs to be included in the 1003 cipher list. 1004 [Steve Henson] 1005 1006 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 1007 EVP_MD_type. The old functionality is available in a new macro called 1008 EVP_MD_md(). Change code that uses it and update docs. 1009 [Steve Henson] 1010 1011 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 1012 where the 'void *' argument is replaced by a function pointer argument. 1013 Previously 'void *' was abused to point to functions, which works on 1014 many platforms, but is not correct. As these functions are usually 1015 called by macros defined in OpenSSL header files, most source code 1016 should work without changes. 1017 [Richard Levitte] 1018 1019 *) <openssl/opensslconf.h> (which is created by Configure) now contains 1020 sections with information on -D... compiler switches used for 1021 compiling the library so that applications can see them. To enable 1022 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 1023 must be defined. E.g., 1024 #define OPENSSL_ALGORITHM_DEFINES 1025 #include <openssl/opensslconf.h> 1026 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 1027 [Richard Levitte, Ulf and Bodo M�ller] 1028 1029 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 1030 record layer. 1031 [Bodo Moeller] 1032 1033 *) Change the 'other' type in certificate aux info to a STACK_OF 1034 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 1035 the required ASN1 format: arbitrary types determined by an OID. 1036 [Steve Henson] 1037 1038 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 1039 argument to 'req'. This is not because the function is newer or 1040 better than others it just uses the work 'NEW' in the certificate 1041 request header lines. Some software needs this. 1042 [Steve Henson] 1043 1044 *) Reorganise password command line arguments: now passwords can be 1045 obtained from various sources. Delete the PEM_cb function and make 1046 it the default behaviour: i.e. if the callback is NULL and the 1047 usrdata argument is not NULL interpret it as a null terminated pass 1048 phrase. If usrdata and the callback are NULL then the pass phrase 1049 is prompted for as usual. 1050 [Steve Henson] 1051 1052 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 1053 the support is automatically enabled. The resulting binaries will 1054 autodetect the card and use it if present. 1055 [Ben Laurie and Compaq Inc.] 1056 1057 *) Work around for Netscape hang bug. This sends certificate request 1058 and server done in one record. Since this is perfectly legal in the 1059 SSL/TLS protocol it isn't a "bug" option and is on by default. See 1060 the bugs/SSLv3 entry for more info. 1061 [Steve Henson] 1062 1063 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 1064 [Andy Polyakov] 1065 1066 *) Add -rand argument to smime and pkcs12 applications and read/write 1067 of seed file. 1068 [Steve Henson] 1069 1070 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 1071 [Bodo Moeller] 1072 1073 *) Add command line password options to the remaining applications. 1074 [Steve Henson] 1075 1076 *) Bug fix for BN_div_recp() for numerators with an even number of 1077 bits. 1078 [Ulf M�ller] 1079 1080 *) More tests in bntest.c, and changed test_bn output. 1081 [Ulf M�ller] 1082 1083 *) ./config recognizes MacOS X now. 1084 [Andy Polyakov] 1085 1086 *) Bug fix for BN_div() when the first words of num and divsor are 1087 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 1088 [Ulf M�ller] 1089 1090 *) Add support for various broken PKCS#8 formats, and command line 1091 options to produce them. 1092 [Steve Henson] 1093 1094 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 1095 get temporary BIGNUMs from a BN_CTX. 1096 [Ulf M�ller] 1097 1098 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 1099 for p == 0. 1100 [Ulf M�ller] 1101 1102 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 1103 include a #define from the old name to the new. The original intent 1104 was that statically linked binaries could for example just call 1105 SSLeay_add_all_ciphers() to just add ciphers to the table and not 1106 link with digests. This never worked becayse SSLeay_add_all_digests() 1107 and SSLeay_add_all_ciphers() were in the same source file so calling 1108 one would link with the other. They are now in separate source files. 1109 [Steve Henson] 1110 1111 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 1112 [Steve Henson] 1113 1114 *) Use a less unusual form of the Miller-Rabin primality test (it used 1115 a binary algorithm for exponentiation integrated into the Miller-Rabin 1116 loop, our standard modexp algorithms are faster). 1117 [Bodo Moeller] 1118 1119 *) Support for the EBCDIC character set completed. 1120 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 1121 1122 *) Source code cleanups: use const where appropriate, eliminate casts, 1123 use void * instead of char * in lhash. 1124 [Ulf M�ller] 1125 1126 *) Bugfix: ssl3_send_server_key_exchange was not restartable 1127 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 1128 this the server could overwrite ephemeral keys that the client 1129 has already seen). 1130 [Bodo Moeller] 1131 1132 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 1133 using 50 iterations of the Rabin-Miller test. 1134 1135 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 1136 iterations of the Rabin-Miller test as required by the appendix 1137 to FIPS PUB 186[-1]) instead of DSA_is_prime. 1138 As BN_is_prime_fasttest includes trial division, DSA parameter 1139 generation becomes much faster. 1140 1141 This implies a change for the callback functions in DSA_is_prime 1142 and DSA_generate_parameters: The callback function is called once 1143 for each positive witness in the Rabin-Miller test, not just 1144 occasionally in the inner loop; and the parameters to the 1145 callback function now provide an iteration count for the outer 1146 loop rather than for the current invocation of the inner loop. 1147 DSA_generate_parameters additionally can call the callback 1148 function with an 'iteration count' of -1, meaning that a 1149 candidate has passed the trial division test (when q is generated 1150 from an application-provided seed, trial division is skipped). 1151 [Bodo Moeller] 1152 1153 *) New function BN_is_prime_fasttest that optionally does trial 1154 division before starting the Rabin-Miller test and has 1155 an additional BN_CTX * argument (whereas BN_is_prime always 1156 has to allocate at least one BN_CTX). 1157 'callback(1, -1, cb_arg)' is called when a number has passed the 1158 trial division stage. 1159 [Bodo Moeller] 1160 1161 *) Fix for bug in CRL encoding. The validity dates weren't being handled 1162 as ASN1_TIME. 1163 [Steve Henson] 1164 1165 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 1166 [Steve Henson] 1167 1168 *) New function BN_pseudo_rand(). 1169 [Ulf M�ller] 1170 1171 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 1172 bignum version of BN_from_montgomery() with the working code from 1173 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 1174 the comments. 1175 [Ulf M�ller] 1176 1177 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 1178 made it impossible to use the same SSL_SESSION data structure in 1179 SSL2 clients in multiple threads. 1180 [Bodo Moeller] 1181 1182 *) The return value of RAND_load_file() no longer counts bytes obtained 1183 by stat(). RAND_load_file(..., -1) is new and uses the complete file 1184 to seed the PRNG (previously an explicit byte count was required). 1185 [Ulf M�ller, Bodo M�ller] 1186 1187 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 1188 used (char *) instead of (void *) and had casts all over the place. 1189 [Steve Henson] 1190 1191 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 1192 [Ulf M�ller] 1193 1194 *) Retain source code compatibility for BN_prime_checks macro: 1195 BN_is_prime(..., BN_prime_checks, ...) now uses 1196 BN_prime_checks_for_size to determine the appropriate number of 1197 Rabin-Miller iterations. 1198 [Ulf M�ller] 1199 1200 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 1201 DH_CHECK_P_NOT_SAFE_PRIME. 1202 (Check if this is true? OpenPGP calls them "strong".) 1203 [Ulf M�ller] 1204 1205 *) Merge the functionality of "dh" and "gendh" programs into a new program 1206 "dhparam". The old programs are retained for now but will handle DH keys 1207 (instead of parameters) in future. 1208 [Steve Henson] 1209 1210 *) Make the ciphers, s_server and s_client programs check the return values 1211 when a new cipher list is set. 1212 [Steve Henson] 1213 1214 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 1215 ciphers. Before when the 56bit ciphers were enabled the sorting was 1216 wrong. 1217 1218 The syntax for the cipher sorting has been extended to support sorting by 1219 cipher-strength (using the strength_bits hard coded in the tables). 1220 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 1221 1222 Fix a bug in the cipher-command parser: when supplying a cipher command 1223 string with an "undefined" symbol (neither command nor alphanumeric 1224 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 1225 an error is flagged. 1226 1227 Due to the strength-sorting extension, the code of the 1228 ssl_create_cipher_list() function was completely rearranged. I hope that 1229 the readability was also increased :-) 1230 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 1231 1232 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 1233 for the first serial number and places 2 in the serial number file. This 1234 avoids problems when the root CA is created with serial number zero and 1235 the first user certificate has the same issuer name and serial number 1236 as the root CA. 1237 [Steve Henson] 1238 1239 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 1240 the new code. Add documentation for this stuff. 1241 [Steve Henson] 1242 1243 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 1244 X509_*() to X509at_*() on the grounds that they don't handle X509 1245 structures and behave in an analagous way to the X509v3 functions: 1246 they shouldn't be called directly but wrapper functions should be used 1247 instead. 1248 1249 So we also now have some wrapper functions that call the X509at functions 1250 when passed certificate requests. (TO DO: similar things can be done with 1251 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 1252 things. Some of these need some d2i or i2d and print functionality 1253 because they handle more complex structures.) 1254 [Steve Henson] 1255 1256 *) Add missing #ifndefs that caused missing symbols when building libssl 1257 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 1258 NO_RSA in ssl/s2*.c. 1259 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 1260 1261 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 1262 has a return value which indicates the quality of the random data 1263 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 1264 error queue. New function RAND_pseudo_bytes() generates output that is 1265 guaranteed to be unique but not unpredictable. RAND_add is like 1266 RAND_seed, but takes an extra argument for an entropy estimate 1267 (RAND_seed always assumes full entropy). 1268 [Ulf M�ller] 1269 1270 *) Do more iterations of Rabin-Miller probable prime test (specifically, 1271 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 1272 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 1273 in crypto/bn/bn_prime.c for the complete table). This guarantees a 1274 false-positive rate of at most 2^-80 for random input. 1275 [Bodo Moeller] 1276 1277 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 1278 [Bodo Moeller] 1279 1280 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 1281 in the 0.9.5 release), this returns the chain 1282 from an X509_CTX structure with a dup of the stack and all 1283 the X509 reference counts upped: so the stack will exist 1284 after X509_CTX_cleanup() has been called. Modify pkcs12.c 1285 to use this. 1286 1287 Also make SSL_SESSION_print() print out the verify return 1288 code. 1289 [Steve Henson] 1290 1291 *) Add manpage for the pkcs12 command. Also change the default 1292 behaviour so MAC iteration counts are used unless the new 1293 -nomaciter option is used. This improves file security and 1294 only older versions of MSIE (4.0 for example) need it. 1295 [Steve Henson] 1296 1297 *) Honor the no-xxx Configure options when creating .DEF files. 1298 [Ulf M�ller] 1299 1300 *) Add PKCS#10 attributes to field table: challengePassword, 1301 unstructuredName and unstructuredAddress. These are taken from 1302 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 1303 international characters are used. 1304 1305 More changes to X509_ATTRIBUTE code: allow the setting of types 1306 based on strings. Remove the 'loc' parameter when adding 1307 attributes because these will be a SET OF encoding which is sorted 1308 in ASN1 order. 1309 [Steve Henson] 1310 1311 *) Initial changes to the 'req' utility to allow request generation 1312 automation. This will allow an application to just generate a template 1313 file containing all the field values and have req construct the 1314 request. 1315 1316 Initial support for X509_ATTRIBUTE handling. Stacks of these are 1317 used all over the place including certificate requests and PKCS#7 1318 structures. They are currently handled manually where necessary with 1319 some primitive wrappers for PKCS#7. The new functions behave in a 1320 manner analogous to the X509 extension functions: they allow 1321 attributes to be looked up by NID and added. 1322 1323 Later something similar to the X509V3 code would be desirable to 1324 automatically handle the encoding, decoding and printing of the 1325 more complex types. The string types like challengePassword can 1326 be handled by the string table functions. 1327 1328 Also modified the multi byte string table handling. Now there is 1329 a 'global mask' which masks out certain types. The table itself 1330 can use the flag STABLE_NO_MASK to ignore the mask setting: this 1331 is useful when for example there is only one permissible type 1332 (as in countryName) and using the mask might result in no valid 1333 types at all. 1334 [Steve Henson] 1335 1336 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 1337 SSL_get_peer_finished to allow applications to obtain the latest 1338 Finished messages sent to the peer or expected from the peer, 1339 respectively. (SSL_get_peer_finished is usually the Finished message 1340 actually received from the peer, otherwise the protocol will be aborted.) 1341 1342 As the Finished message are message digests of the complete handshake 1343 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 1344 be used for external authentication procedures when the authentication 1345 provided by SSL/TLS is not desired or is not enough. 1346 [Bodo Moeller] 1347 1348 *) Enhanced support for Alpha Linux is added. Now ./config checks if 1349 the host supports BWX extension and if Compaq C is present on the 1350 $PATH. Just exploiting of the BWX extension results in 20-30% 1351 performance kick for some algorithms, e.g. DES and RC4 to mention 1352 a couple. Compaq C in turn generates ~20% faster code for MD5 and 1353 SHA1. 1354 [Andy Polyakov] 1355 1356 *) Add support for MS "fast SGC". This is arguably a violation of the 1357 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 1358 weak crypto and after checking the certificate is SGC a second one 1359 with strong crypto. MS SGC stops the first handshake after receiving 1360 the server certificate message and sends a second client hello. Since 1361 a server will typically do all the time consuming operations before 1362 expecting any further messages from the client (server key exchange 1363 is the most expensive) there is little difference between the two. 1364 1365 To get OpenSSL to support MS SGC we have to permit a second client 1366 hello message after we have sent server done. In addition we have to 1367 reset the MAC if we do get this second client hello. 1368 [Steve Henson] 1369 1370 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 1371 if a DER encoded private key is RSA or DSA traditional format. Changed 1372 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 1373 format DER encoded private key. Newer code should use PKCS#8 format which 1374 has the key type encoded in the ASN1 structure. Added DER private key 1375 support to pkcs8 application. 1376 [Steve Henson] 1377 1378 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 1379 ciphersuites has been selected (as required by the SSL 3/TLS 1 1380 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 1381 is set, we interpret this as a request to violate the specification 1382 (the worst that can happen is a handshake failure, and 'correct' 1383 behaviour would result in a handshake failure anyway). 1384 [Bodo Moeller] 1385 1386 *) In SSL_CTX_add_session, take into account that there might be multiple 1387 SSL_SESSION structures with the same session ID (e.g. when two threads 1388 concurrently obtain them from an external cache). 1389 The internal cache can handle only one SSL_SESSION with a given ID, 1390 so if there's a conflict, we now throw out the old one to achieve 1391 consistency. 1392 [Bodo Moeller] 1393 1394 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 1395 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 1396 some routines that use cipher OIDs: some ciphers do not have OIDs 1397 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 1398 example. 1399 [Steve Henson] 1400 1401 *) Simplify the trust setting structure and code. Now we just have 1402 two sequences of OIDs for trusted and rejected settings. These will 1403 typically have values the same as the extended key usage extension 1404 and any application specific purposes. 1405 1406 The trust checking code now has a default behaviour: it will just 1407 check for an object with the same NID as the passed id. Functions can 1408 be provided to override either the default behaviour or the behaviour 1409 for a given id. SSL client, server and email already have functions 1410 in place for compatibility: they check the NID and also return "trusted" 1411 if the certificate is self signed. 1412 [Steve Henson] 1413 1414 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 1415 traditional format into an EVP_PKEY structure. 1416 [Steve Henson] 1417 1418 *) Add a password callback function PEM_cb() which either prompts for 1419 a password if usr_data is NULL or otherwise assumes it is a null 1420 terminated password. Allow passwords to be passed on command line 1421 environment or config files in a few more utilities. 1422 [Steve Henson] 1423 1424 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 1425 keys. Add some short names for PKCS#8 PBE algorithms and allow them 1426 to be specified on the command line for the pkcs8 and pkcs12 utilities. 1427 Update documentation. 1428 [Steve Henson] 1429 1430 *) Support for ASN1 "NULL" type. This could be handled before by using 1431 ASN1_TYPE but there wasn't any function that would try to read a NULL 1432 and produce an error if it couldn't. For compatibility we also have 1433 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 1434 don't allocate anything because they don't need to. 1435 [Steve Henson] 1436 1437 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 1438 for details. 1439 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 1440 1441 *) Rebuild of the memory allocation routines used by OpenSSL code and 1442 possibly others as well. The purpose is to make an interface that 1443 provide hooks so anyone can build a separate set of allocation and 1444 deallocation routines to be used by OpenSSL, for example memory 1445 pool implementations, or something else, which was previously hard 1446 since Malloc(), Realloc() and Free() were defined as macros having 1447 the values malloc, realloc and free, respectively (except for Win32 1448 compilations). The same is provided for memory debugging code. 1449 OpenSSL already comes with functionality to find memory leaks, but 1450 this gives people a chance to debug other memory problems. 1451 1452 With these changes, a new set of functions and macros have appeared: 1453 1454 CRYPTO_set_mem_debug_functions() [F] 1455 CRYPTO_get_mem_debug_functions() [F] 1456 CRYPTO_dbg_set_options() [F] 1457 CRYPTO_dbg_get_options() [F] 1458 CRYPTO_malloc_debug_init() [M] 1459 1460 The memory debug functions are NULL by default, unless the library 1461 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 1462 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 1463 gives the standard debugging functions that come with OpenSSL) or 1464 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 1465 provided by the library user) must be used. When the standard 1466 debugging functions are used, CRYPTO_dbg_set_options can be used to 1467 request additional information: 1468 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 1469 the CRYPTO_MDEBUG_xxx macro when compiling the library. 1470 1471 Also, things like CRYPTO_set_mem_functions will always give the 1472 expected result (the new set of functions is used for allocation 1473 and deallocation) at all times, regardless of platform and compiler 1474 options. 1475 1476 To finish it up, some functions that were never use in any other 1477 way than through macros have a new API and new semantic: 1478 1479 CRYPTO_dbg_malloc() 1480 CRYPTO_dbg_realloc() 1481 CRYPTO_dbg_free() 1482 1483 All macros of value have retained their old syntax. 1484 [Richard Levitte and Bodo Moeller] 1485 1486 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 1487 ordering of SMIMECapabilities wasn't in "strength order" and there 1488 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 1489 algorithm. 1490 [Steve Henson] 1491 1492 *) Some ASN1 types with illegal zero length encoding (INTEGER, 1493 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 1494 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 1495 1496 *) Merge in my S/MIME library for OpenSSL. This provides a simple 1497 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 1498 functionality to handle multipart/signed properly) and a utility 1499 called 'smime' to call all this stuff. This is based on code I 1500 originally wrote for Celo who have kindly allowed it to be 1501 included in OpenSSL. 1502 [Steve Henson] 1503 1504 *) Add variants des_set_key_checked and des_set_key_unchecked of 1505 des_set_key (aka des_key_sched). Global variable des_check_key 1506 decides which of these is called by des_set_key; this way 1507 des_check_key behaves as it always did, but applications and 1508 the library itself, which was buggy for des_check_key == 1, 1509 have a cleaner way to pick the version they need. 1510 [Bodo Moeller] 1511 1512 *) New function PKCS12_newpass() which changes the password of a 1513 PKCS12 structure. 1514 [Steve Henson] 1515 1516 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 1517 dynamic mix. In both cases the ids can be used as an index into the 1518 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 1519 functions so they accept a list of the field values and the 1520 application doesn't need to directly manipulate the X509_TRUST 1521 structure. 1522 [Steve Henson] 1523 1524 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 1525 need initialising. 1526 [Steve Henson] 1527 1528 *) Modify the way the V3 extension code looks up extensions. This now 1529 works in a similar way to the object code: we have some "standard" 1530 extensions in a static table which is searched with OBJ_bsearch() 1531 and the application can add dynamic ones if needed. The file 1532 crypto/x509v3/ext_dat.h now has the info: this file needs to be 1533 updated whenever a new extension is added to the core code and kept 1534 in ext_nid order. There is a simple program 'tabtest.c' which checks 1535 this. New extensions are not added too often so this file can readily 1536 be maintained manually. 1537 1538 There are two big advantages in doing things this way. The extensions 1539 can be looked up immediately and no longer need to be "added" using 1540 X509V3_add_standard_extensions(): this function now does nothing. 1541 [Side note: I get *lots* of email saying the extension code doesn't 1542 work because people forget to call this function] 1543 Also no dynamic allocation is done unless new extensions are added: 1544 so if we don't add custom extensions there is no need to call 1545 X509V3_EXT_cleanup(). 1546 [Steve Henson] 1547 1548 *) Modify enc utility's salting as follows: make salting the default. Add a 1549 magic header, so unsalted files fail gracefully instead of just decrypting 1550 to garbage. This is because not salting is a big security hole, so people 1551 should be discouraged from doing it. 1552 [Ben Laurie] 1553 1554 *) Fixes and enhancements to the 'x509' utility. It allowed a message 1555 digest to be passed on the command line but it only used this 1556 parameter when signing a certificate. Modified so all relevant 1557 operations are affected by the digest parameter including the 1558 -fingerprint and -x509toreq options. Also -x509toreq choked if a 1559 DSA key was used because it didn't fix the digest. 1560 [Steve Henson] 1561 1562 *) Initial certificate chain verify code. Currently tests the untrusted 1563 certificates for consistency with the verify purpose (which is set 1564 when the X509_STORE_CTX structure is set up) and checks the pathlength. 1565 1566 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 1567 this is because it will reject chains with invalid extensions whereas 1568 every previous version of OpenSSL and SSLeay made no checks at all. 1569 1570 Trust code: checks the root CA for the relevant trust settings. Trust 1571 settings have an initial value consistent with the verify purpose: e.g. 1572 if the verify purpose is for SSL client use it expects the CA to be 1573 trusted for SSL client use. However the default value can be changed to 1574 permit custom trust settings: one example of this would be to only trust 1575 certificates from a specific "secure" set of CAs. 1576 1577 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 1578 which should be used for version portability: especially since the 1579 verify structure is likely to change more often now. 1580 1581 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 1582 to set them. If not set then assume SSL clients will verify SSL servers 1583 and vice versa. 1584 1585 Two new options to the verify program: -untrusted allows a set of 1586 untrusted certificates to be passed in and -purpose which sets the 1587 intended purpose of the certificate. If a purpose is set then the 1588 new chain verify code is used to check extension consistency. 1589 [Steve Henson] 1590 1591 *) Support for the authority information access extension. 1592 [Steve Henson] 1593 1594 *) Modify RSA and DSA PEM read routines to transparently handle 1595 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 1596 public keys in a format compatible with certificate 1597 SubjectPublicKeyInfo structures. Unfortunately there were already 1598 functions called *_PublicKey_* which used various odd formats so 1599 these are retained for compatibility: however the DSA variants were 1600 never in a public release so they have been deleted. Changed dsa/rsa 1601 utilities to handle the new format: note no releases ever handled public 1602 keys so we should be OK. 1603 1604 The primary motivation for this change is to avoid the same fiasco 1605 that dogs private keys: there are several incompatible private key 1606 formats some of which are standard and some OpenSSL specific and 1607 require various evil hacks to allow partial transparent handling and 1608 even then it doesn't work with DER formats. Given the option anything 1609 other than PKCS#8 should be dumped: but the other formats have to 1610 stay in the name of compatibility. 1611 1612 With public keys and the benefit of hindsight one standard format 1613 is used which works with EVP_PKEY, RSA or DSA structures: though 1614 it clearly returns an error if you try to read the wrong kind of key. 1615 1616 Added a -pubkey option to the 'x509' utility to output the public key. 1617 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 1618 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 1619 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 1620 that do the same as the EVP_PKEY_assign_*() except they up the 1621 reference count of the added key (they don't "swallow" the 1622 supplied key). 1623 [Steve Henson] 1624 1625 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 1626 CRLs would fail if the file contained no certificates or no CRLs: 1627 added a new function to read in both types and return the number 1628 read: this means that if none are read it will be an error. The 1629 DER versions of the certificate and CRL reader would always fail 1630 because it isn't possible to mix certificates and CRLs in DER format 1631 without choking one or the other routine. Changed this to just read 1632 a certificate: this is the best we can do. Also modified the code 1633 in apps/verify.c to take notice of return codes: it was previously 1634 attempting to read in certificates from NULL pointers and ignoring 1635 any errors: this is one reason why the cert and CRL reader seemed 1636 to work. It doesn't check return codes from the default certificate 1637 routines: these may well fail if the certificates aren't installed. 1638 [Steve Henson] 1639 1640 *) Code to support otherName option in GeneralName. 1641 [Steve Henson] 1642 1643 *) First update to verify code. Change the verify utility 1644 so it warns if it is passed a self signed certificate: 1645 for consistency with the normal behaviour. X509_verify 1646 has been modified to it will now verify a self signed 1647 certificate if *exactly* the same certificate appears 1648 in the store: it was previously impossible to trust a 1649 single self signed certificate. This means that: 1650 openssl verify ss.pem 1651 now gives a warning about a self signed certificate but 1652 openssl verify -CAfile ss.pem ss.pem 1653 is OK. 1654 [Steve Henson] 1655 1656 *) For servers, store verify_result in SSL_SESSION data structure 1657 (and add it to external session representation). 1658 This is needed when client certificate verifications fails, 1659 but an application-provided verification callback (set by 1660 SSL_CTX_set_cert_verify_callback) allows accepting the session 1661 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 1662 but returns 1): When the session is reused, we have to set 1663 ssl->verify_result to the appropriate error code to avoid 1664 security holes. 1665 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 1666 1667 *) Fix a bug in the new PKCS#7 code: it didn't consider the 1668 case in PKCS7_dataInit() where the signed PKCS7 structure 1669 didn't contain any existing data because it was being created. 1670 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 1671 1672 *) Add a salt to the key derivation routines in enc.c. This 1673 forms the first 8 bytes of the encrypted file. Also add a 1674 -S option to allow a salt to be input on the command line. 1675 [Steve Henson] 1676 1677 *) New function X509_cmp(). Oddly enough there wasn't a function 1678 to compare two certificates. We do this by working out the SHA1 1679 hash and comparing that. X509_cmp() will be needed by the trust 1680 code. 1681 [Steve Henson] 1682 1683 *) SSL_get1_session() is like SSL_get_session(), but increments 1684 the reference count in the SSL_SESSION returned. 1685 [Geoff Thorpe <geoff@eu.c2.net>] 1686 1687 *) Fix for 'req': it was adding a null to request attributes. 1688 Also change the X509_LOOKUP and X509_INFO code to handle 1689 certificate auxiliary information. 1690 [Steve Henson] 1691 1692 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 1693 the 'enc' command. 1694 [Steve Henson] 1695 1696 *) Add the possibility to add extra information to the memory leak 1697 detecting output, to form tracebacks, showing from where each 1698 allocation was originated: CRYPTO_push_info("constant string") adds 1699 the string plus current file name and line number to a per-thread 1700 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 1701 is like calling CYRPTO_pop_info() until the stack is empty. 1702 Also updated memory leak detection code to be multi-thread-safe. 1703 [Richard Levitte] 1704 1705 *) Add options -text and -noout to pkcs7 utility and delete the 1706 encryption options which never did anything. Update docs. 1707 [Steve Henson] 1708 1709 *) Add options to some of the utilities to allow the pass phrase 1710 to be included on either the command line (not recommended on 1711 OSes like Unix) or read from the environment. Update the 1712 manpages and fix a few bugs. 1713 [Steve Henson] 1714 1715 *) Add a few manpages for some of the openssl commands. 1716 [Steve Henson] 1717 1718 *) Fix the -revoke option in ca. It was freeing up memory twice, 1719 leaking and not finding already revoked certificates. 1720 [Steve Henson] 1721 1722 *) Extensive changes to support certificate auxiliary information. 1723 This involves the use of X509_CERT_AUX structure and X509_AUX 1724 functions. An X509_AUX function such as PEM_read_X509_AUX() 1725 can still read in a certificate file in the usual way but it 1726 will also read in any additional "auxiliary information". By 1727 doing things this way a fair degree of compatibility can be 1728 retained: existing certificates can have this information added 1729 using the new 'x509' options. 1730 1731 Current auxiliary information includes an "alias" and some trust 1732 settings. The trust settings will ultimately be used in enhanced 1733 certificate chain verification routines: currently a certificate 1734 can only be trusted if it is self signed and then it is trusted 1735 for all purposes. 1736 [Steve Henson] 1737 1738 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 1739 The problem was that one of the replacement routines had not been working 1740 since SSLeay releases. For now the offending routine has been replaced 1741 with non-optimised assembler. Even so, this now gives around 95% 1742 performance improvement for 1024 bit RSA signs. 1743 [Mark Cox] 1744 1745 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 1746 handling. Most clients have the effective key size in bits equal to 1747 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 1748 A few however don't do this and instead use the size of the decrypted key 1749 to determine the RC2 key length and the AlgorithmIdentifier to determine 1750 the effective key length. In this case the effective key length can still 1751 be 40 bits but the key length can be 168 bits for example. This is fixed 1752 by manually forcing an RC2 key into the EVP_PKEY structure because the 1753 EVP code can't currently handle unusual RC2 key sizes: it always assumes 1754 the key length and effective key length are equal. 1755 [Steve Henson] 1756 1757 *) Add a bunch of functions that should simplify the creation of 1758 X509_NAME structures. Now you should be able to do: 1759 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 1760 and have it automatically work out the correct field type and fill in 1761 the structures. The more adventurous can try: 1762 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 1763 and it will (hopefully) work out the correct multibyte encoding. 1764 [Steve Henson] 1765 1766 *) Change the 'req' utility to use the new field handling and multibyte 1767 copy routines. Before the DN field creation was handled in an ad hoc 1768 way in req, ca, and x509 which was rather broken and didn't support 1769 BMPStrings or UTF8Strings. Since some software doesn't implement 1770 BMPStrings or UTF8Strings yet, they can be enabled using the config file 1771 using the dirstring_type option. See the new comment in the default 1772 openssl.cnf for more info. 1773 [Steve Henson] 1774 1775 *) Make crypto/rand/md_rand.c more robust: 1776 - Assure unique random numbers after fork(). 1777 - Make sure that concurrent threads access the global counter and 1778 md serializably so that we never lose entropy in them 1779 or use exactly the same state in multiple threads. 1780 Access to the large state is not always serializable because 1781 the additional locking could be a performance killer, and 1782 md should be large enough anyway. 1783 [Bodo Moeller] 1784 1785 *) New file apps/app_rand.c with commonly needed functionality 1786 for handling the random seed file. 1787 1788 Use the random seed file in some applications that previously did not: 1789 ca, 1790 dsaparam -genkey (which also ignored its '-rand' option), 1791 s_client, 1792 s_server, 1793 x509 (when signing). 1794 Except on systems with /dev/urandom, it is crucial to have a random 1795 seed file at least for key creation, DSA signing, and for DH exchanges; 1796 for RSA signatures we could do without one. 1797 1798 gendh and gendsa (unlike genrsa) used to read only the first byte 1799 of each file listed in the '-rand' option. The function as previously 1800 found in genrsa is now in app_rand.c and is used by all programs 1801 that support '-rand'. 1802 [Bodo Moeller] 1803 1804 *) In RAND_write_file, use mode 0600 for creating files; 1805 don't just chmod when it may be too late. 1806 [Bodo Moeller] 1807 1808 *) Report an error from X509_STORE_load_locations 1809 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 1810 [Bill Perry] 1811 1812 *) New function ASN1_mbstring_copy() this copies a string in either 1813 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 1814 into an ASN1_STRING type. A mask of permissible types is passed 1815 and it chooses the "minimal" type to use or an error if not type 1816 is suitable. 1817 [Steve Henson] 1818 1819 *) Add function equivalents to the various macros in asn1.h. The old 1820 macros are retained with an M_ prefix. Code inside the library can 1821 use the M_ macros. External code (including the openssl utility) 1822 should *NOT* in order to be "shared library friendly". 1823 [Steve Henson] 1824 1825 *) Add various functions that can check a certificate's extensions 1826 to see if it usable for various purposes such as SSL client, 1827 server or S/MIME and CAs of these types. This is currently 1828 VERY EXPERIMENTAL but will ultimately be used for certificate chain 1829 verification. Also added a -purpose flag to x509 utility to 1830 print out all the purposes. 1831 [Steve Henson] 1832 1833 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 1834 functions. 1835 [Steve Henson] 1836 1837 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 1838 for, obtain and decode and extension and obtain its critical flag. 1839 This allows all the necessary extension code to be handled in a 1840 single function call. 1841 [Steve Henson] 1842 1843 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 1844 platforms. See crypto/rc4/rc4_enc.c for further details. 1845 [Andy Polyakov] 1846 1847 *) New -noout option to asn1parse. This causes no output to be produced 1848 its main use is when combined with -strparse and -out to extract data 1849 from a file (which may not be in ASN.1 format). 1850 [Steve Henson] 1851 1852 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 1853 when producing the local key id. 1854 [Richard Levitte <levitte@stacken.kth.se>] 1855 1856 *) New option -dhparam in s_server. This allows a DH parameter file to be 1857 stated explicitly. If it is not stated then it tries the first server 1858 certificate file. The previous behaviour hard coded the filename 1859 "server.pem". 1860 [Steve Henson] 1861 1862 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 1863 a public key to be input or output. For example: 1864 openssl rsa -in key.pem -pubout -out pubkey.pem 1865 Also added necessary DSA public key functions to handle this. 1866 [Steve Henson] 1867 1868 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 1869 in the message. This was handled by allowing 1870 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 1871 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 1872 1873 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 1874 to the end of the strings whereas this didn't. This would cause problems 1875 if strings read with d2i_ASN1_bytes() were later modified. 1876 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 1877 1878 *) Fix for base64 decode bug. When a base64 bio reads only one line of 1879 data and it contains EOF it will end up returning an error. This is 1880 caused by input 46 bytes long. The cause is due to the way base64 1881 BIOs find the start of base64 encoded data. They do this by trying a 1882 trial decode on each line until they find one that works. When they 1883 do a flag is set and it starts again knowing it can pass all the 1884 data directly through the decoder. Unfortunately it doesn't reset 1885 the context it uses. This means that if EOF is reached an attempt 1886 is made to pass two EOFs through the context and this causes the 1887 resulting error. This can also cause other problems as well. As is 1888 usual with these problems it takes *ages* to find and the fix is 1889 trivial: move one line. 1890 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 1891 1892 *) Ugly workaround to get s_client and s_server working under Windows. The 1893 old code wouldn't work because it needed to select() on sockets and the 1894 tty (for keypresses and to see if data could be written). Win32 only 1895 supports select() on sockets so we select() with a 1s timeout on the 1896 sockets and then see if any characters are waiting to be read, if none 1897 are present then we retry, we also assume we can always write data to 1898 the tty. This isn't nice because the code then blocks until we've 1899 received a complete line of data and it is effectively polling the 1900 keyboard at 1s intervals: however it's quite a bit better than not 1901 working at all :-) A dedicated Windows application might handle this 1902 with an event loop for example. 1903 [Steve Henson] 1904 1905 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 1906 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 1907 will be called when RSA_sign() and RSA_verify() are used. This is useful 1908 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 1909 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 1910 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 1911 This necessitated the support of an extra signature type NID_md5_sha1 1912 for SSL signatures and modifications to the SSL library to use it instead 1913 of calling RSA_public_decrypt() and RSA_private_encrypt(). 1914 [Steve Henson] 1915 1916 *) Add new -verify -CAfile and -CApath options to the crl program, these 1917 will lookup a CRL issuers certificate and verify the signature in a 1918 similar way to the verify program. Tidy up the crl program so it 1919 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 1920 less strict. It will now permit CRL extensions even if it is not 1921 a V2 CRL: this will allow it to tolerate some broken CRLs. 1922 [Steve Henson] 1923 1924 *) Initialize all non-automatic variables each time one of the openssl 1925 sub-programs is started (this is necessary as they may be started 1926 multiple times from the "OpenSSL>" prompt). 1927 [Lennart Bang, Bodo Moeller] 1928 1929 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 1930 removing all other RSA functionality (this is what NO_RSA does). This 1931 is so (for example) those in the US can disable those operations covered 1932 by the RSA patent while allowing storage and parsing of RSA keys and RSA 1933 key generation. 1934 [Steve Henson] 1935 1936 *) Non-copying interface to BIO pairs. 1937 (still largely untested) 1938 [Bodo Moeller] 1939 1940 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 1941 ASCII string. This was handled independently in various places before. 1942 [Steve Henson] 1943 1944 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 1945 UTF8 strings a character at a time. 1946 [Steve Henson] 1947 1948 *) Use client_version from client hello to select the protocol 1949 (s23_srvr.c) and for RSA client key exchange verification 1950 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 1951 [Bodo Moeller] 1952 1953 *) Add various utility functions to handle SPKACs, these were previously 1954 handled by poking round in the structure internals. Added new function 1955 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 1956 print, verify and generate SPKACs. Based on an original idea from 1957 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 1958 [Steve Henson] 1959 1960 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 1961 [Andy Polyakov] 1962 1963 *) Allow the config file extension section to be overwritten on the 1964 command line. Based on an original idea from Massimiliano Pala 1965 <madwolf@comune.modena.it>. The new option is called -extensions 1966 and can be applied to ca, req and x509. Also -reqexts to override 1967 the request extensions in req and -crlexts to override the crl extensions 1968 in ca. 1969 [Steve Henson] 1970 1971 *) Add new feature to the SPKAC handling in ca. Now you can include 1972 the same field multiple times by preceding it by "XXXX." for example: 1973 1.OU="Unit name 1" 1974 2.OU="Unit name 2" 1975 this is the same syntax as used in the req config file. 1976 [Steve Henson] 1977 1978 *) Allow certificate extensions to be added to certificate requests. These 1979 are specified in a 'req_extensions' option of the req section of the 1980 config file. They can be printed out with the -text option to req but 1981 are otherwise ignored at present. 1982 [Steve Henson] 1983 1984 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 1985 data read consists of only the final block it would not decrypted because 1986 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 1987 A misplaced 'break' also meant the decrypted final block might not be 1988 copied until the next read. 1989 [Steve Henson] 1990 1991 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 1992 a few extra parameters to the DH structure: these will be useful if 1993 for example we want the value of 'q' or implement X9.42 DH. 1994 [Steve Henson] 1995 1996 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 1997 provides hooks that allow the default DSA functions or functions on a 1998 "per key" basis to be replaced. This allows hardware acceleration and 1999 hardware key storage to be handled without major modification to the 2000 library. Also added low level modexp hooks and CRYPTO_EX structure and 2001 associated functions. 2002 [Steve Henson] 2003 2004 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 2005 as "read only": it can't be written to and the buffer it points to will 2006 not be freed. Reading from a read only BIO is much more efficient than 2007 a normal memory BIO. This was added because there are several times when 2008 an area of memory needs to be read from a BIO. The previous method was 2009 to create a memory BIO and write the data to it, this results in two 2010 copies of the data and an O(n^2) reading algorithm. There is a new 2011 function BIO_new_mem_buf() which creates a read only memory BIO from 2012 an area of memory. Also modified the PKCS#7 routines to use read only 2013 memory BIOs. 2014 [Steve Henson] 2015 2016 *) Bugfix: ssl23_get_client_hello did not work properly when called in 2017 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 2018 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 2019 but a retry condition occured while trying to read the rest. 2020 [Bodo Moeller] 2021 2022 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 2023 NID_pkcs7_encrypted by default: this was wrong since this should almost 2024 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 2025 the encrypted data type: this is a more sensible place to put it and it 2026 allows the PKCS#12 code to be tidied up that duplicated this 2027 functionality. 2028 [Steve Henson] 2029 2030 *) Changed obj_dat.pl script so it takes its input and output files on 2031 the command line. This should avoid shell escape redirection problems 2032 under Win32. 2033 [Steve Henson] 2034 2035 *) Initial support for certificate extension requests, these are included 2036 in things like Xenroll certificate requests. Included functions to allow 2037 extensions to be obtained and added. 2038 [Steve Henson] 2039 2040 *) -crlf option to s_client and s_server for sending newlines as 2041 CRLF (as required by many protocols). 2042 [Bodo Moeller] 2043 2044 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 2045 2046 *) Install libRSAglue.a when OpenSSL is built with RSAref. 2047 [Ralf S. Engelschall] 2048 2049 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 2050 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 2051 2052 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 2053 program. 2054 [Steve Henson] 2055 2056 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 2057 DH parameters/keys (q is lost during that conversion, but the resulting 2058 DH parameters contain its length). 2059 2060 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 2061 much faster than DH_generate_parameters (which creates parameters 2062 where p = 2*q + 1), and also the smaller q makes DH computations 2063 much more efficient (160-bit exponentiation instead of 1024-bit 2064 exponentiation); so this provides a convenient way to support DHE 2065 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 2066 utter importance to use 2067 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 2068 or 2069 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 2070 when such DH parameters are used, because otherwise small subgroup 2071 attacks may become possible! 2072 [Bodo Moeller] 2073 2074 *) Avoid memory leak in i2d_DHparams. 2075 [Bodo Moeller] 2076 2077 *) Allow the -k option to be used more than once in the enc program: 2078 this allows the same encrypted message to be read by multiple recipients. 2079 [Steve Henson] 2080 2081 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 2082 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 2083 it will always use the numerical form of the OID, even if it has a short 2084 or long name. 2085 [Steve Henson] 2086 2087 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 2088 method only got called if p,q,dmp1,dmq1,iqmp components were present, 2089 otherwise bn_mod_exp was called. In the case of hardware keys for example 2090 no private key components need be present and it might store extra data 2091 in the RSA structure, which cannot be accessed from bn_mod_exp. 2092 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 2093 private key operations. 2094 [Steve Henson] 2095 2096 *) Added support for SPARC Linux. 2097 [Andy Polyakov] 2098 2099 *) pem_password_cb function type incompatibly changed from 2100 typedef int pem_password_cb(char *buf, int size, int rwflag); 2101 to 2102 ....(char *buf, int size, int rwflag, void *userdata); 2103 so that applications can pass data to their callbacks: 2104 The PEM[_ASN1]_{read,write}... functions and macros now take an 2105 additional void * argument, which is just handed through whenever 2106 the password callback is called. 2107 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 2108 2109 New function SSL_CTX_set_default_passwd_cb_userdata. 2110 2111 Compatibility note: As many C implementations push function arguments 2112 onto the stack in reverse order, the new library version is likely to 2113 interoperate with programs that have been compiled with the old 2114 pem_password_cb definition (PEM_whatever takes some data that 2115 happens to be on the stack as its last argument, and the callback 2116 just ignores this garbage); but there is no guarantee whatsoever that 2117 this will work. 2118 2119 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 2120 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 2121 problems not only on Windows, but also on some Unix platforms. 2122 To avoid problematic command lines, these definitions are now in an 2123 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 2124 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 2125 [Bodo Moeller] 2126 2127 *) MIPS III/IV assembler module is reimplemented. 2128 [Andy Polyakov] 2129 2130 *) More DES library cleanups: remove references to srand/rand and 2131 delete an unused file. 2132 [Ulf M�ller] 2133 2134 *) Add support for the the free Netwide assembler (NASM) under Win32, 2135 since not many people have MASM (ml) and it can be hard to obtain. 2136 This is currently experimental but it seems to work OK and pass all 2137 the tests. Check out INSTALL.W32 for info. 2138 [Steve Henson] 2139 2140 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 2141 without temporary keys kept an extra copy of the server key, 2142 and connections with temporary keys did not free everything in case 2143 of an error. 2144 [Bodo Moeller] 2145 2146 *) New function RSA_check_key and new openssl rsa option -check 2147 for verifying the consistency of RSA keys. 2148 [Ulf Moeller, Bodo Moeller] 2149 2150 *) Various changes to make Win32 compile work: 2151 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 2152 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 2153 comparison" warnings. 2154 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 2155 [Steve Henson] 2156 2157 *) Add a debugging option to PKCS#5 v2 key generation function: when 2158 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 2159 derived keys are printed to stderr. 2160 [Steve Henson] 2161 2162 *) Copy the flags in ASN1_STRING_dup(). 2163 [Roman E. Pavlov <pre@mo.msk.ru>] 2164 2165 *) The x509 application mishandled signing requests containing DSA 2166 keys when the signing key was also DSA and the parameters didn't match. 2167 2168 It was supposed to omit the parameters when they matched the signing key: 2169 the verifying software was then supposed to automatically use the CA's 2170 parameters if they were absent from the end user certificate. 2171 2172 Omitting parameters is no longer recommended. The test was also 2173 the wrong way round! This was probably due to unusual behaviour in 2174 EVP_cmp_parameters() which returns 1 if the parameters match. 2175 This meant that parameters were omitted when they *didn't* match and 2176 the certificate was useless. Certificates signed with 'ca' didn't have 2177 this bug. 2178 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 2179 2180 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 2181 The interface is as follows: 2182 Applications can use 2183 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 2184 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 2185 "off" is now the default. 2186 The library internally uses 2187 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 2188 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 2189 to disable memory-checking temporarily. 2190 2191 Some inconsistent states that previously were possible (and were 2192 even the default) are now avoided. 2193 2194 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 2195 with each memory chunk allocated; this is occasionally more helpful 2196 than just having a counter. 2197 2198 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 2199 2200 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 2201 extensions. 2202 [Bodo Moeller] 2203 2204 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 2205 which largely parallels "options", but is for changing API behaviour, 2206 whereas "options" are about protocol behaviour. 2207 Initial "mode" flags are: 2208 2209 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 2210 a single record has been written. 2211 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 2212 retries use the same buffer location. 2213 (But all of the contents must be 2214 copied!) 2215 [Bodo Moeller] 2216 2217 *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode 2218 worked. 2219 2220 *) Fix problems with no-hmac etc. 2221 [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 2222 2223 *) New functions RSA_get_default_method(), RSA_set_method() and 2224 RSA_get_method(). These allows replacement of RSA_METHODs without having 2225 to mess around with the internals of an RSA structure. 2226 [Steve Henson] 2227 2228 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 2229 Also really enable memory leak checks in openssl.c and in some 2230 test programs. 2231 [Chad C. Mulligan, Bodo Moeller] 2232 2233 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 2234 up the length of negative integers. This has now been simplified to just 2235 store the length when it is first determined and use it later, rather 2236 than trying to keep track of where data is copied and updating it to 2237 point to the end. 2238 [Steve Henson, reported by Brien Wheeler 2239 <bwheeler@authentica-security.com>] 2240 2241 *) Add a new function PKCS7_signatureVerify. This allows the verification 2242 of a PKCS#7 signature but with the signing certificate passed to the 2243 function itself. This contrasts with PKCS7_dataVerify which assumes the 2244 certificate is present in the PKCS#7 structure. This isn't always the 2245 case: certificates can be omitted from a PKCS#7 structure and be 2246 distributed by "out of band" means (such as a certificate database). 2247 [Steve Henson] 2248 2249 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 2250 function prototypes in pem.h, also change util/mkdef.pl to add the 2251 necessary function names. 2252 [Steve Henson] 2253 2254 *) mk1mf.pl (used by Windows builds) did not properly read the 2255 options set by Configure in the top level Makefile, and Configure 2256 was not even able to write more than one option correctly. 2257 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 2258 [Bodo Moeller] 2259 2260 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 2261 file to be loaded from a BIO or FILE pointer. The BIO version will 2262 for example allow memory BIOs to contain config info. 2263 [Steve Henson] 2264 2265 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 2266 Whoever hopes to achieve shared-library compatibility across versions 2267 must use this, not the compile-time macro. 2268 (Exercise 0.9.4: Which is the minimum library version required by 2269 such programs?) 2270 Note: All this applies only to multi-threaded programs, others don't 2271 need locks. 2272 [Bodo Moeller] 2273 2274 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 2275 through a BIO pair triggered the default case, i.e. 2276 SSLerr(...,SSL_R_UNKNOWN_STATE). 2277 [Bodo Moeller] 2278 2279 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 2280 can use the SSL library even if none of the specific BIOs is 2281 appropriate. 2282 [Bodo Moeller] 2283 2284 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 2285 for the encoded length. 2286 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 2287 2288 *) Add initial documentation of the X509V3 functions. 2289 [Steve Henson] 2290 2291 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 2292 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 2293 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 2294 secure PKCS#8 private key format with a high iteration count. 2295 [Steve Henson] 2296 2297 *) Fix determination of Perl interpreter: A perl or perl5 2298 _directory_ in $PATH was also accepted as the interpreter. 2299 [Ralf S. Engelschall] 2300 2301 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 2302 wrong with it but it was very old and did things like calling 2303 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 2304 unusual formatting. 2305 [Steve Henson] 2306 2307 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 2308 to use the new extension code. 2309 [Steve Henson] 2310 2311 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 2312 with macros. This should make it easier to change their form, add extra 2313 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 2314 constant. 2315 [Steve Henson] 2316 2317 *) Add to configuration table a new entry that can specify an alternative 2318 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 2319 according to Mark Crispin <MRC@Panda.COM>. 2320 [Bodo Moeller] 2321 2322#if 0 2323 *) DES CBC did not update the IV. Weird. 2324 [Ben Laurie] 2325#else 2326 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 2327 Changing the behaviour of the former might break existing programs -- 2328 where IV updating is needed, des_ncbc_encrypt can be used. 2329#endif 2330 2331 *) When bntest is run from "make test" it drives bc to check its 2332 calculations, as well as internally checking them. If an internal check 2333 fails, it needs to cause bc to give a non-zero result or make test carries 2334 on without noticing the failure. Fixed. 2335 [Ben Laurie] 2336 2337 *) DES library cleanups. 2338 [Ulf M�ller] 2339 2340 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 2341 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 2342 ciphers. NOTE: although the key derivation function has been verified 2343 against some published test vectors it has not been extensively tested 2344 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 2345 of v2.0. 2346 [Steve Henson] 2347 2348 *) Instead of "mkdir -p", which is not fully portable, use new 2349 Perl script "util/mkdir-p.pl". 2350 [Bodo Moeller] 2351 2352 *) Rewrite the way password based encryption (PBE) is handled. It used to 2353 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 2354 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 2355 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 2356 the 'parameter' field of the AlgorithmIdentifier is passed to the 2357 underlying key generation function so it must do its own ASN1 parsing. 2358 This has also changed the EVP_PBE_CipherInit() function which now has a 2359 'parameter' argument instead of literal salt and iteration count values 2360 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 2361 [Steve Henson] 2362 2363 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 2364 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 2365 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 2366 KEY" because this clashed with PKCS#8 unencrypted string. Since this 2367 value was just used as a "magic string" and not used directly its 2368 value doesn't matter. 2369 [Steve Henson] 2370 2371 *) Introduce some semblance of const correctness to BN. Shame C doesn't 2372 support mutable. 2373 [Ben Laurie] 2374 2375 *) "linux-sparc64" configuration (ultrapenguin). 2376 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 2377 "linux-sparc" configuration. 2378 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 2379 2380 *) config now generates no-xxx options for missing ciphers. 2381 [Ulf M�ller] 2382 2383 *) Support the EBCDIC character set (work in progress). 2384 File ebcdic.c not yet included because it has a different license. 2385 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 2386 2387 *) Support BS2000/OSD-POSIX. 2388 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 2389 2390 *) Make callbacks for key generation use void * instead of char *. 2391 [Ben Laurie] 2392 2393 *) Make S/MIME samples compile (not yet tested). 2394 [Ben Laurie] 2395 2396 *) Additional typesafe stacks. 2397 [Ben Laurie] 2398 2399 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 2400 [Bodo Moeller] 2401 2402 2403 Changes between 0.9.3 and 0.9.3a [29 May 1999] 2404 2405 *) New configuration variant "sco5-gcc". 2406 2407 *) Updated some demos. 2408 [Sean O Riordain, Wade Scholine] 2409 2410 *) Add missing BIO_free at exit of pkcs12 application. 2411 [Wu Zhigang] 2412 2413 *) Fix memory leak in conf.c. 2414 [Steve Henson] 2415 2416 *) Updates for Win32 to assembler version of MD5. 2417 [Steve Henson] 2418 2419 *) Set #! path to perl in apps/der_chop to where we found it 2420 instead of using a fixed path. 2421 [Bodo Moeller] 2422 2423 *) SHA library changes for irix64-mips4-cc. 2424 [Andy Polyakov] 2425 2426 *) Improvements for VMS support. 2427 [Richard Levitte] 2428 2429 2430 Changes between 0.9.2b and 0.9.3 [24 May 1999] 2431 2432 *) Bignum library bug fix. IRIX 6 passes "make test" now! 2433 This also avoids the problems with SC4.2 and unpatched SC5. 2434 [Andy Polyakov <appro@fy.chalmers.se>] 2435 2436 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 2437 These are required because of the typesafe stack would otherwise break 2438 existing code. If old code used a structure member which used to be STACK 2439 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 2440 sk_num or sk_value it would produce an error because the num, data members 2441 are not present in STACK_OF. Now it just produces a warning. sk_set 2442 replaces the old method of assigning a value to sk_value 2443 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 2444 that does this will no longer work (and should use sk_set instead) but 2445 this could be regarded as a "questionable" behaviour anyway. 2446 [Steve Henson] 2447 2448 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 2449 correctly handle encrypted S/MIME data. 2450 [Steve Henson] 2451 2452 *) Change type of various DES function arguments from des_cblock 2453 (which means, in function argument declarations, pointer to char) 2454 to des_cblock * (meaning pointer to array with 8 char elements), 2455 which allows the compiler to do more typechecking; it was like 2456 that back in SSLeay, but with lots of ugly casts. 2457 2458 Introduce new type const_des_cblock. 2459 [Bodo Moeller] 2460 2461 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 2462 problems: find RecipientInfo structure that matches recipient certificate 2463 and initialise the ASN1 structures properly based on passed cipher. 2464 [Steve Henson] 2465 2466 *) Belatedly make the BN tests actually check the results. 2467 [Ben Laurie] 2468 2469 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 2470 to and from BNs: it was completely broken. New compilation option 2471 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 2472 key elements as negative integers. 2473 [Steve Henson] 2474 2475 *) Reorganize and speed up MD5. 2476 [Andy Polyakov <appro@fy.chalmers.se>] 2477 2478 *) VMS support. 2479 [Richard Levitte <richard@levitte.org>] 2480 2481 *) New option -out to asn1parse to allow the parsed structure to be 2482 output to a file. This is most useful when combined with the -strparse 2483 option to examine the output of things like OCTET STRINGS. 2484 [Steve Henson] 2485 2486 *) Make SSL library a little more fool-proof by not requiring any longer 2487 that SSL_set_{accept,connect}_state be called before 2488 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 2489 in many applications because usually everything *appeared* to work as 2490 intended anyway -- now it really works as intended). 2491 [Bodo Moeller] 2492 2493 *) Move openssl.cnf out of lib/. 2494 [Ulf M�ller] 2495 2496 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 2497 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 2498 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 2499 [Ralf S. Engelschall] 2500 2501 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 2502 handle PKCS#7 enveloped data properly. 2503 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 2504 2505 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 2506 copying pointers. The cert_st handling is changed by this in 2507 various ways (and thus what used to be known as ctx->default_cert 2508 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 2509 any longer when s->cert does not give us what we need). 2510 ssl_cert_instantiate becomes obsolete by this change. 2511 As soon as we've got the new code right (possibly it already is?), 2512 we have solved a couple of bugs of the earlier code where s->cert 2513 was used as if it could not have been shared with other SSL structures. 2514 2515 Note that using the SSL API in certain dirty ways now will result 2516 in different behaviour than observed with earlier library versions: 2517 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 2518 does not influence s as it used to. 2519 2520 In order to clean up things more thoroughly, inside SSL_SESSION 2521 we don't use CERT any longer, but a new structure SESS_CERT 2522 that holds per-session data (if available); currently, this is 2523 the peer's certificate chain and, for clients, the server's certificate 2524 and temporary key. CERT holds only those values that can have 2525 meaningful defaults in an SSL_CTX. 2526 [Bodo Moeller] 2527 2528 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 2529 from the internal representation. Various PKCS#7 fixes: remove some 2530 evil casts and set the enc_dig_alg field properly based on the signing 2531 key type. 2532 [Steve Henson] 2533 2534 *) Allow PKCS#12 password to be set from the command line or the 2535 environment. Let 'ca' get its config file name from the environment 2536 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 2537 and 'x509'). 2538 [Steve Henson] 2539 2540 *) Allow certificate policies extension to use an IA5STRING for the 2541 organization field. This is contrary to the PKIX definition but 2542 VeriSign uses it and IE5 only recognises this form. Document 'x509' 2543 extension option. 2544 [Steve Henson] 2545 2546 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 2547 without disallowing inline assembler and the like for non-pedantic builds. 2548 [Ben Laurie] 2549 2550 *) Support Borland C++ builder. 2551 [Janez Jere <jj@void.si>, modified by Ulf M�ller] 2552 2553 *) Support Mingw32. 2554 [Ulf M�ller] 2555 2556 *) SHA-1 cleanups and performance enhancements. 2557 [Andy Polyakov <appro@fy.chalmers.se>] 2558 2559 *) Sparc v8plus assembler for the bignum library. 2560 [Andy Polyakov <appro@fy.chalmers.se>] 2561 2562 *) Accept any -xxx and +xxx compiler options in Configure. 2563 [Ulf M�ller] 2564 2565 *) Update HPUX configuration. 2566 [Anonymous] 2567 2568 *) Add missing sk_<type>_unshift() function to safestack.h 2569 [Ralf S. Engelschall] 2570 2571 *) New function SSL_CTX_use_certificate_chain_file that sets the 2572 "extra_cert"s in addition to the certificate. (This makes sense 2573 only for "PEM" format files, as chains as a whole are not 2574 DER-encoded.) 2575 [Bodo Moeller] 2576 2577 *) Support verify_depth from the SSL API. 2578 x509_vfy.c had what can be considered an off-by-one-error: 2579 Its depth (which was not part of the external interface) 2580 was actually counting the number of certificates in a chain; 2581 now it really counts the depth. 2582 [Bodo Moeller] 2583 2584 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 2585 instead of X509err, which often resulted in confusing error 2586 messages since the error codes are not globally unique 2587 (e.g. an alleged error in ssl3_accept when a certificate 2588 didn't match the private key). 2589 2590 *) New function SSL_CTX_set_session_id_context that allows to set a default 2591 value (so that you don't need SSL_set_session_id_context for each 2592 connection using the SSL_CTX). 2593 [Bodo Moeller] 2594 2595 *) OAEP decoding bug fix. 2596 [Ulf M�ller] 2597 2598 *) Support INSTALL_PREFIX for package builders, as proposed by 2599 David Harris. 2600 [Bodo Moeller] 2601 2602 *) New Configure options "threads" and "no-threads". For systems 2603 where the proper compiler options are known (currently Solaris 2604 and Linux), "threads" is the default. 2605 [Bodo Moeller] 2606 2607 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 2608 [Bodo Moeller] 2609 2610 *) Install various scripts to $(OPENSSLDIR)/misc, not to 2611 $(INSTALLTOP)/bin -- they shouldn't clutter directories 2612 such as /usr/local/bin. 2613 [Bodo Moeller] 2614 2615 *) "make linux-shared" to build shared libraries. 2616 [Niels Poppe <niels@netbox.org>] 2617 2618 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 2619 [Ulf M�ller] 2620 2621 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 2622 extension adding in x509 utility. 2623 [Steve Henson] 2624 2625 *) Remove NOPROTO sections and error code comments. 2626 [Ulf M�ller] 2627 2628 *) Partial rewrite of the DEF file generator to now parse the ANSI 2629 prototypes. 2630 [Steve Henson] 2631 2632 *) New Configure options --prefix=DIR and --openssldir=DIR. 2633 [Ulf M�ller] 2634 2635 *) Complete rewrite of the error code script(s). It is all now handled 2636 by one script at the top level which handles error code gathering, 2637 header rewriting and C source file generation. It should be much better 2638 than the old method: it now uses a modified version of Ulf's parser to 2639 read the ANSI prototypes in all header files (thus the old K&R definitions 2640 aren't needed for error creation any more) and do a better job of 2641 translating function codes into names. The old 'ASN1 error code imbedded 2642 in a comment' is no longer necessary and it doesn't use .err files which 2643 have now been deleted. Also the error code call doesn't have to appear all 2644 on one line (which resulted in some large lines...). 2645 [Steve Henson] 2646 2647 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 2648 [Bodo Moeller] 2649 2650 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 2651 0 (which usually indicates a closed connection), but continue reading. 2652 [Bodo Moeller] 2653 2654 *) Fix some race conditions. 2655 [Bodo Moeller] 2656 2657 *) Add support for CRL distribution points extension. Add Certificate 2658 Policies and CRL distribution points documentation. 2659 [Steve Henson] 2660 2661 *) Move the autogenerated header file parts to crypto/opensslconf.h. 2662 [Ulf M�ller] 2663 2664 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 2665 8 of keying material. Merlin has also confirmed interop with this fix 2666 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 2667 [Merlin Hughes <merlin@baltimore.ie>] 2668 2669 *) Fix lots of warnings. 2670 [Richard Levitte <levitte@stacken.kth.se>] 2671 2672 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 2673 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 2674 [Richard Levitte <levitte@stacken.kth.se>] 2675 2676 *) Fix problems with sizeof(long) == 8. 2677 [Andy Polyakov <appro@fy.chalmers.se>] 2678 2679 *) Change functions to ANSI C. 2680 [Ulf M�ller] 2681 2682 *) Fix typos in error codes. 2683 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 2684 2685 *) Remove defunct assembler files from Configure. 2686 [Ulf M�ller] 2687 2688 *) SPARC v8 assembler BIGNUM implementation. 2689 [Andy Polyakov <appro@fy.chalmers.se>] 2690 2691 *) Support for Certificate Policies extension: both print and set. 2692 Various additions to support the r2i method this uses. 2693 [Steve Henson] 2694 2695 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 2696 return a const string when you are expecting an allocated buffer. 2697 [Ben Laurie] 2698 2699 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 2700 types DirectoryString and DisplayText. 2701 [Steve Henson] 2702 2703 *) Add code to allow r2i extensions to access the configuration database, 2704 add an LHASH database driver and add several ctx helper functions. 2705 [Steve Henson] 2706 2707 *) Fix an evil bug in bn_expand2() which caused various BN functions to 2708 fail when they extended the size of a BIGNUM. 2709 [Steve Henson] 2710 2711 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 2712 support typesafe stack. 2713 [Steve Henson] 2714 2715 *) Fix typo in SSL_[gs]et_options(). 2716 [Nils Frostberg <nils@medcom.se>] 2717 2718 *) Delete various functions and files that belonged to the (now obsolete) 2719 old X509V3 handling code. 2720 [Steve Henson] 2721 2722 *) New Configure option "rsaref". 2723 [Ulf M�ller] 2724 2725 *) Don't auto-generate pem.h. 2726 [Bodo Moeller] 2727 2728 *) Introduce type-safe ASN.1 SETs. 2729 [Ben Laurie] 2730 2731 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 2732 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 2733 2734 *) Introduce type-safe STACKs. This will almost certainly break lots of code 2735 that links with OpenSSL (well at least cause lots of warnings), but fear 2736 not: the conversion is trivial, and it eliminates loads of evil casts. A 2737 few STACKed things have been converted already. Feel free to convert more. 2738 In the fullness of time, I'll do away with the STACK type altogether. 2739 [Ben Laurie] 2740 2741 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 2742 specified in <certfile> by updating the entry in the index.txt file. 2743 This way one no longer has to edit the index.txt file manually for 2744 revoking a certificate. The -revoke option does the gory details now. 2745 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 2746 2747 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 2748 `-text' option at all and this way the `-noout -text' combination was 2749 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 2750 [Ralf S. Engelschall] 2751 2752 *) Make sure a corresponding plain text error message exists for the 2753 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 2754 verify callback function determined that a certificate was revoked. 2755 [Ralf S. Engelschall] 2756 2757 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 2758 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 2759 all available cipers including rc5, which was forgotten until now. 2760 In order to let the testing shell script know which algorithms 2761 are available, a new (up to now undocumented) command 2762 "openssl list-cipher-commands" is used. 2763 [Bodo Moeller] 2764 2765 *) Bugfix: s_client occasionally would sleep in select() when 2766 it should have checked SSL_pending() first. 2767 [Bodo Moeller] 2768 2769 *) New functions DSA_do_sign and DSA_do_verify to provide access to 2770 the raw DSA values prior to ASN.1 encoding. 2771 [Ulf M�ller] 2772 2773 *) Tweaks to Configure 2774 [Niels Poppe <niels@netbox.org>] 2775 2776 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 2777 yet... 2778 [Steve Henson] 2779 2780 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 2781 [Ulf M�ller] 2782 2783 *) New config option to avoid instructions that are illegal on the 80386. 2784 The default code is faster, but requires at least a 486. 2785 [Ulf M�ller] 2786 2787 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 2788 SSL2_SERVER_VERSION (not used at all) macros, which are now the 2789 same as SSL2_VERSION anyway. 2790 [Bodo Moeller] 2791 2792 *) New "-showcerts" option for s_client. 2793 [Bodo Moeller] 2794 2795 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 2796 application. Various cleanups and fixes. 2797 [Steve Henson] 2798 2799 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 2800 modify error routines to work internally. Add error codes and PBE init 2801 to library startup routines. 2802 [Steve Henson] 2803 2804 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 2805 packing functions to asn1 and evp. Changed function names and error 2806 codes along the way. 2807 [Steve Henson] 2808 2809 *) PKCS12 integration: and so it begins... First of several patches to 2810 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 2811 objects to objects.h 2812 [Steve Henson] 2813 2814 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 2815 and display support for Thawte strong extranet extension. 2816 [Steve Henson] 2817 2818 *) Add LinuxPPC support. 2819 [Jeff Dubrule <igor@pobox.org>] 2820 2821 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 2822 bn_div_words in alpha.s. 2823 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 2824 2825 *) Make sure the RSA OAEP test is skipped under -DRSAref because 2826 OAEP isn't supported when OpenSSL is built with RSAref. 2827 [Ulf Moeller <ulf@fitug.de>] 2828 2829 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 2830 so they no longer are missing under -DNOPROTO. 2831 [Soren S. Jorvang <soren@t.dk>] 2832 2833 2834 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 2835 2836 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 2837 doesn't work when the session is reused. Coming soon! 2838 [Ben Laurie] 2839 2840 *) Fix a security hole, that allows sessions to be reused in the wrong 2841 context thus bypassing client cert protection! All software that uses 2842 client certs and session caches in multiple contexts NEEDS PATCHING to 2843 allow session reuse! A fuller solution is in the works. 2844 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 2845 2846 *) Some more source tree cleanups (removed obsolete files 2847 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 2848 permission on "config" script to be executable) and a fix for the INSTALL 2849 document. 2850 [Ulf Moeller <ulf@fitug.de>] 2851 2852 *) Remove some legacy and erroneous uses of malloc, free instead of 2853 Malloc, Free. 2854 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 2855 2856 *) Make rsa_oaep_test return non-zero on error. 2857 [Ulf Moeller <ulf@fitug.de>] 2858 2859 *) Add support for native Solaris shared libraries. Configure 2860 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 2861 if someone would make that last step automatic. 2862 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 2863 2864 *) ctx_size was not built with the right compiler during "make links". Fixed. 2865 [Ben Laurie] 2866 2867 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 2868 except NULL ciphers". This means the default cipher list will no longer 2869 enable NULL ciphers. They need to be specifically enabled e.g. with 2870 the string "DEFAULT:eNULL". 2871 [Steve Henson] 2872 2873 *) Fix to RSA private encryption routines: if p < q then it would 2874 occasionally produce an invalid result. This will only happen with 2875 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 2876 [Steve Henson] 2877 2878 *) Be less restrictive and allow also `perl util/perlpath.pl 2879 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 2880 because this way one can also use an interpreter named `perl5' (which is 2881 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 2882 installed as `perl'). 2883 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 2884 2885 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 2886 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 2887 2888 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 2889 advapi32.lib to Win32 build and change the pem test comparision 2890 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 2891 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 2892 and crypto/des/ede_cbcm_enc.c. 2893 [Steve Henson] 2894 2895 *) DES quad checksum was broken on big-endian architectures. Fixed. 2896 [Ben Laurie] 2897 2898 *) Comment out two functions in bio.h that aren't implemented. Fix up the 2899 Win32 test batch file so it (might) work again. The Win32 test batch file 2900 is horrible: I feel ill.... 2901 [Steve Henson] 2902 2903 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 2904 in e_os.h. Audit of header files to check ANSI and non ANSI 2905 sections: 10 functions were absent from non ANSI section and not exported 2906 from Windows DLLs. Fixed up libeay.num for new functions. 2907 [Steve Henson] 2908 2909 *) Make `openssl version' output lines consistent. 2910 [Ralf S. Engelschall] 2911 2912 *) Fix Win32 symbol export lists for BIO functions: Added 2913 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 2914 to ms/libeay{16,32}.def. 2915 [Ralf S. Engelschall] 2916 2917 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 2918 fine under Unix and passes some trivial tests I've now added. But the 2919 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 2920 added to make sure no one expects that this stuff really works in the 2921 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 2922 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 2923 openssl_bio.xs. 2924 [Ralf S. Engelschall] 2925 2926 *) Fix the generation of two part addresses in perl. 2927 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 2928 2929 *) Add config entry for Linux on MIPS. 2930 [John Tobey <jtobey@channel1.com>] 2931 2932 *) Make links whenever Configure is run, unless we are on Windoze. 2933 [Ben Laurie] 2934 2935 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 2936 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 2937 in CRLs. 2938 [Steve Henson] 2939 2940 *) Add a useful kludge to allow package maintainers to specify compiler and 2941 other platforms details on the command line without having to patch the 2942 Configure script everytime: One now can use ``perl Configure 2943 <id>:<details>'', i.e. platform ids are allowed to have details appended 2944 to them (seperated by colons). This is treated as there would be a static 2945 pre-configured entry in Configure's %table under key <id> with value 2946 <details> and ``perl Configure <id>'' is called. So, when you want to 2947 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 2948 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 2949 now, which overrides the FreeBSD-elf entry on-the-fly. 2950 [Ralf S. Engelschall] 2951 2952 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 2953 [Ben Laurie] 2954 2955 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 2956 on the `perl Configure ...' command line. This way one can compile 2957 OpenSSL libraries with Position Independent Code (PIC) which is needed 2958 for linking it into DSOs. 2959 [Ralf S. Engelschall] 2960 2961 *) Remarkably, export ciphers were totally broken and no-one had noticed! 2962 Fixed. 2963 [Ben Laurie] 2964 2965 *) Cleaned up the LICENSE document: The official contact for any license 2966 questions now is the OpenSSL core team under openssl-core@openssl.org. 2967 And add a paragraph about the dual-license situation to make sure people 2968 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 2969 to the OpenSSL toolkit. 2970 [Ralf S. Engelschall] 2971 2972 *) General source tree makefile cleanups: Made `making xxx in yyy...' 2973 display consistent in the source tree and replaced `/bin/rm' by `rm'. 2974 Additonally cleaned up the `make links' target: Remove unnecessary 2975 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 2976 to speed processing and no longer clutter the display with confusing 2977 stuff. Instead only the actually done links are displayed. 2978 [Ralf S. Engelschall] 2979 2980 *) Permit null encryption ciphersuites, used for authentication only. It used 2981 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 2982 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 2983 encryption. 2984 [Ben Laurie] 2985 2986 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 2987 signed attributes when verifying signatures (this would break them), 2988 the detached data encoding was wrong and public keys obtained using 2989 X509_get_pubkey() weren't freed. 2990 [Steve Henson] 2991 2992 *) Add text documentation for the BUFFER functions. Also added a work around 2993 to a Win95 console bug. This was triggered by the password read stuff: the 2994 last character typed gets carried over to the next fread(). If you were 2995 generating a new cert request using 'req' for example then the last 2996 character of the passphrase would be CR which would then enter the first 2997 field as blank. 2998 [Steve Henson] 2999 3000 *) Added the new `Includes OpenSSL Cryptography Software' button as 3001 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 3002 button and can be used by applications based on OpenSSL to show the 3003 relationship to the OpenSSL project. 3004 [Ralf S. Engelschall] 3005 3006 *) Remove confusing variables in function signatures in files 3007 ssl/ssl_lib.c and ssl/ssl.h. 3008 [Lennart Bong <lob@kulthea.stacken.kth.se>] 3009 3010 *) Don't install bss_file.c under PREFIX/include/ 3011 [Lennart Bong <lob@kulthea.stacken.kth.se>] 3012 3013 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 3014 functions that return function pointers and has support for NT specific 3015 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 3016 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 3017 unsigned to signed types: this was killing the Win32 compile. 3018 [Steve Henson] 3019 3020 *) Add new certificate file to stack functions, 3021 SSL_add_dir_cert_subjects_to_stack() and 3022 SSL_add_file_cert_subjects_to_stack(). These largely supplant 3023 SSL_load_client_CA_file(), and can be used to add multiple certs easily 3024 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 3025 This means that Apache-SSL and similar packages don't have to mess around 3026 to add as many CAs as they want to the preferred list. 3027 [Ben Laurie] 3028 3029 *) Experiment with doxygen documentation. Currently only partially applied to 3030 ssl/ssl_lib.c. 3031 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 3032 openssl.doxy as the configuration file. 3033 [Ben Laurie] 3034 3035 *) Get rid of remaining C++-style comments which strict C compilers hate. 3036 [Ralf S. Engelschall, pointed out by Carlos Amengual] 3037 3038 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 3039 compiled in by default: it has problems with large keys. 3040 [Steve Henson] 3041 3042 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 3043 DH private keys and/or callback functions which directly correspond to 3044 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 3045 is needed for applications which have to configure certificates on a 3046 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 3047 (e.g. s_server). 3048 For the RSA certificate situation is makes no difference, but 3049 for the DSA certificate situation this fixes the "no shared cipher" 3050 problem where the OpenSSL cipher selection procedure failed because the 3051 temporary keys were not overtaken from the context and the API provided 3052 no way to reconfigure them. 3053 The new functions now let applications reconfigure the stuff and they 3054 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 3055 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 3056 non-public-API function ssl_cert_instantiate() is used as a helper 3057 function and also to reduce code redundancy inside ssl_rsa.c. 3058 [Ralf S. Engelschall] 3059 3060 *) Move s_server -dcert and -dkey options out of the undocumented feature 3061 area because they are useful for the DSA situation and should be 3062 recognized by the users. 3063 [Ralf S. Engelschall] 3064 3065 *) Fix the cipher decision scheme for export ciphers: the export bits are 3066 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 3067 SSL_EXP_MASK. So, the original variable has to be used instead of the 3068 already masked variable. 3069 [Richard Levitte <levitte@stacken.kth.se>] 3070 3071 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 3072 [Richard Levitte <levitte@stacken.kth.se>] 3073 3074 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 3075 from `int' to `unsigned int' because it's a length and initialized by 3076 EVP_DigestFinal() which expects an `unsigned int *'. 3077 [Richard Levitte <levitte@stacken.kth.se>] 3078 3079 *) Don't hard-code path to Perl interpreter on shebang line of Configure 3080 script. Instead use the usual Shell->Perl transition trick. 3081 [Ralf S. Engelschall] 3082 3083 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 3084 (in addition to RSA certificates) to match the behaviour of `openssl dsa 3085 -noout -modulus' as it's already the case for `openssl rsa -noout 3086 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 3087 currently the public key is printed (a decision which was already done by 3088 `openssl dsa -modulus' in the past) which serves a similar purpose. 3089 Additionally the NO_RSA no longer completely removes the whole -modulus 3090 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 3091 now, too. 3092 [Ralf S. Engelschall] 3093 3094 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 3095 BIO. See the source (crypto/evp/bio_ok.c) for more info. 3096 [Arne Ansper <arne@ats.cyber.ee>] 3097 3098 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 3099 to be added. Now both 'req' and 'ca' can use new objects defined in the 3100 config file. 3101 [Steve Henson] 3102 3103 *) Add cool BIO that does syslog (or event log on NT). 3104 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 3105 3106 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 3107 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 3108 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 3109 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 3110 [Ben Laurie] 3111 3112 *) Add preliminary config info for new extension code. 3113 [Steve Henson] 3114 3115 *) Make RSA_NO_PADDING really use no padding. 3116 [Ulf Moeller <ulf@fitug.de>] 3117 3118 *) Generate errors when private/public key check is done. 3119 [Ben Laurie] 3120 3121 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 3122 for some CRL extensions and new objects added. 3123 [Steve Henson] 3124 3125 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 3126 key usage extension and fuller support for authority key id. 3127 [Steve Henson] 3128 3129 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 3130 padding method for RSA, which is recommended for new applications in PKCS 3131 #1 v2.0 (RFC 2437, October 1998). 3132 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 3133 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 3134 against Bleichbacher's attack on RSA. 3135 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 3136 Ben Laurie] 3137 3138 *) Updates to the new SSL compression code 3139 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3140 3141 *) Fix so that the version number in the master secret, when passed 3142 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 3143 (because the server will not accept higher), that the version number 3144 is 0x03,0x01, not 0x03,0x00 3145 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3146 3147 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 3148 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 3149 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 3150 [Steve Henson] 3151 3152 *) Support for RAW extensions where an arbitrary extension can be 3153 created by including its DER encoding. See apps/openssl.cnf for 3154 an example. 3155 [Steve Henson] 3156 3157 *) Make sure latest Perl versions don't interpret some generated C array 3158 code as Perl array code in the crypto/err/err_genc.pl script. 3159 [Lars Weber <3weber@informatik.uni-hamburg.de>] 3160 3161 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 3162 not many people have the assembler. Various Win32 compilation fixes and 3163 update to the INSTALL.W32 file with (hopefully) more accurate Win32 3164 build instructions. 3165 [Steve Henson] 3166 3167 *) Modify configure script 'Configure' to automatically create crypto/date.h 3168 file under Win32 and also build pem.h from pem.org. New script 3169 util/mkfiles.pl to create the MINFO file on environments that can't do a 3170 'make files': perl util/mkfiles.pl >MINFO should work. 3171 [Steve Henson] 3172 3173 *) Major rework of DES function declarations, in the pursuit of correctness 3174 and purity. As a result, many evil casts evaporated, and some weirdness, 3175 too. You may find this causes warnings in your code. Zapping your evil 3176 casts will probably fix them. Mostly. 3177 [Ben Laurie] 3178 3179 *) Fix for a typo in asn1.h. Bug fix to object creation script 3180 obj_dat.pl. It considered a zero in an object definition to mean 3181 "end of object": none of the objects in objects.h have any zeros 3182 so it wasn't spotted. 3183 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 3184 3185 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 3186 Masking (CBCM). In the absence of test vectors, the best I have been able 3187 to do is check that the decrypt undoes the encrypt, so far. Send me test 3188 vectors if you have them. 3189 [Ben Laurie] 3190 3191 *) Correct calculation of key length for export ciphers (too much space was 3192 allocated for null ciphers). This has not been tested! 3193 [Ben Laurie] 3194 3195 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 3196 message is now correct (it understands "crypto" and "ssl" on its 3197 command line). There is also now an "update" option. This will update 3198 the util/ssleay.num and util/libeay.num files with any new functions. 3199 If you do a: 3200 perl util/mkdef.pl crypto ssl update 3201 it will update them. 3202 [Steve Henson] 3203 3204 *) Overhauled the Perl interface (perl/*): 3205 - ported BN stuff to OpenSSL's different BN library 3206 - made the perl/ source tree CVS-aware 3207 - renamed the package from SSLeay to OpenSSL (the files still contain 3208 their history because I've copied them in the repository) 3209 - removed obsolete files (the test scripts will be replaced 3210 by better Test::Harness variants in the future) 3211 [Ralf S. Engelschall] 3212 3213 *) First cut for a very conservative source tree cleanup: 3214 1. merge various obsolete readme texts into doc/ssleay.txt 3215 where we collect the old documents and readme texts. 3216 2. remove the first part of files where I'm already sure that we no 3217 longer need them because of three reasons: either they are just temporary 3218 files which were left by Eric or they are preserved original files where 3219 I've verified that the diff is also available in the CVS via "cvs diff 3220 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 3221 the crypto/md/ stuff). 3222 [Ralf S. Engelschall] 3223 3224 *) More extension code. Incomplete support for subject and issuer alt 3225 name, issuer and authority key id. Change the i2v function parameters 3226 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 3227 what that's for :-) Fix to ASN1 macro which messed up 3228 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 3229 [Steve Henson] 3230 3231 *) Preliminary support for ENUMERATED type. This is largely copied from the 3232 INTEGER code. 3233 [Steve Henson] 3234 3235 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 3236 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3237 3238 *) Make sure `make rehash' target really finds the `openssl' program. 3239 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 3240 3241 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 3242 like to hear about it if this slows down other processors. 3243 [Ben Laurie] 3244 3245 *) Add CygWin32 platform information to Configure script. 3246 [Alan Batie <batie@aahz.jf.intel.com>] 3247 3248 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 3249 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 3250 3251 *) New program nseq to manipulate netscape certificate sequences 3252 [Steve Henson] 3253 3254 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 3255 few typos. 3256 [Steve Henson] 3257 3258 *) Fixes to BN code. Previously the default was to define BN_RECURSION 3259 but the BN code had some problems that would cause failures when 3260 doing certificate verification and some other functions. 3261 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3262 3263 *) Add ASN1 and PEM code to support netscape certificate sequences. 3264 [Steve Henson] 3265 3266 *) Add ASN1 and PEM code to support netscape certificate sequences. 3267 [Steve Henson] 3268 3269 *) Add several PKIX and private extended key usage OIDs. 3270 [Steve Henson] 3271 3272 *) Modify the 'ca' program to handle the new extension code. Modify 3273 openssl.cnf for new extension format, add comments. 3274 [Steve Henson] 3275 3276 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 3277 and add a sample to openssl.cnf so req -x509 now adds appropriate 3278 CA extensions. 3279 [Steve Henson] 3280 3281 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 3282 error code, add initial support to X509_print() and x509 application. 3283 [Steve Henson] 3284 3285 *) Takes a deep breath and start addding X509 V3 extension support code. Add 3286 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 3287 stuff is currently isolated and isn't even compiled yet. 3288 [Steve Henson] 3289 3290 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 3291 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 3292 Removed the versions check from X509 routines when loading extensions: 3293 this allows certain broken certificates that don't set the version 3294 properly to be processed. 3295 [Steve Henson] 3296 3297 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 3298 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 3299 can still be regenerated with "make depend". 3300 [Ben Laurie] 3301 3302 *) Spelling mistake in C version of CAST-128. 3303 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 3304 3305 *) Changes to the error generation code. The perl script err-code.pl 3306 now reads in the old error codes and retains the old numbers, only 3307 adding new ones if necessary. It also only changes the .err files if new 3308 codes are added. The makefiles have been modified to only insert errors 3309 when needed (to avoid needlessly modifying header files). This is done 3310 by only inserting errors if the .err file is newer than the auto generated 3311 C file. To rebuild all the error codes from scratch (the old behaviour) 3312 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 3313 or delete all the .err files. 3314 [Steve Henson] 3315 3316 *) CAST-128 was incorrectly implemented for short keys. The C version has 3317 been fixed, but is untested. The assembler versions are also fixed, but 3318 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 3319 to regenerate it if needed. 3320 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 3321 Hagino <itojun@kame.net>] 3322 3323 *) File was opened incorrectly in randfile.c. 3324 [Ulf M�ller <ulf@fitug.de>] 3325 3326 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 3327 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 3328 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 3329 al: it's just almost always a UTCTime. Note this patch adds new error 3330 codes so do a "make errors" if there are problems. 3331 [Steve Henson] 3332 3333 *) Correct Linux 1 recognition in config. 3334 [Ulf M�ller <ulf@fitug.de>] 3335 3336 *) Remove pointless MD5 hash when using DSA keys in ca. 3337 [Anonymous <nobody@replay.com>] 3338 3339 *) Generate an error if given an empty string as a cert directory. Also 3340 generate an error if handed NULL (previously returned 0 to indicate an 3341 error, but didn't set one). 3342 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 3343 3344 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 3345 [Ben Laurie] 3346 3347 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 3348 parameters. This was causing a warning which killed off the Win32 compile. 3349 [Steve Henson] 3350 3351 *) Remove C++ style comments from crypto/bn/bn_local.h. 3352 [Neil Costigan <neil.costigan@celocom.com>] 3353 3354 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 3355 based on a text string, looking up short and long names and finally 3356 "dot" format. The "dot" format stuff didn't work. Added new function 3357 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 3358 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 3359 OID is not part of the table. 3360 [Steve Henson] 3361 3362 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 3363 X509_LOOKUP_by_alias(). 3364 [Ben Laurie] 3365 3366 *) Sort openssl functions by name. 3367 [Ben Laurie] 3368 3369 *) Get the gendsa program working (hopefully) and add it to app list. Remove 3370 encryption from sample DSA keys (in case anyone is interested the password 3371 was "1234"). 3372 [Steve Henson] 3373 3374 *) Make _all_ *_free functions accept a NULL pointer. 3375 [Frans Heymans <fheymans@isaserver.be>] 3376 3377 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 3378 NULL pointers. 3379 [Anonymous <nobody@replay.com>] 3380 3381 *) s_server should send the CAfile as acceptable CAs, not its own cert. 3382 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 3383 3384 *) Don't blow it for numeric -newkey arguments to apps/req. 3385 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 3386 3387 *) Temp key "for export" tests were wrong in s3_srvr.c. 3388 [Anonymous <nobody@replay.com>] 3389 3390 *) Add prototype for temp key callback functions 3391 SSL_CTX_set_tmp_{rsa,dh}_callback(). 3392 [Ben Laurie] 3393 3394 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 3395 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 3396 [Steve Henson] 3397 3398 *) X509_name_add_entry() freed the wrong thing after an error. 3399 [Arne Ansper <arne@ats.cyber.ee>] 3400 3401 *) rsa_eay.c would attempt to free a NULL context. 3402 [Arne Ansper <arne@ats.cyber.ee>] 3403 3404 *) BIO_s_socket() had a broken should_retry() on Windoze. 3405 [Arne Ansper <arne@ats.cyber.ee>] 3406 3407 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 3408 [Arne Ansper <arne@ats.cyber.ee>] 3409 3410 *) Make sure the already existing X509_STORE->depth variable is initialized 3411 in X509_STORE_new(), but document the fact that this variable is still 3412 unused in the certificate verification process. 3413 [Ralf S. Engelschall] 3414 3415 *) Fix the various library and apps files to free up pkeys obtained from 3416 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 3417 [Steve Henson] 3418 3419 *) Fix reference counting in X509_PUBKEY_get(). This makes 3420 demos/maurice/example2.c work, amongst others, probably. 3421 [Steve Henson and Ben Laurie] 3422 3423 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 3424 `openssl' and second, the shortcut symlinks for the `openssl <command>' 3425 are no longer created. This way we have a single and consistent command 3426 line interface `openssl <command>', similar to `cvs <command>'. 3427 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 3428 3429 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 3430 BIT STRING wrapper always have zero unused bits. 3431 [Steve Henson] 3432 3433 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 3434 [Steve Henson] 3435 3436 *) Make the top-level INSTALL documentation easier to understand. 3437 [Paul Sutton] 3438 3439 *) Makefiles updated to exit if an error occurs in a sub-directory 3440 make (including if user presses ^C) [Paul Sutton] 3441 3442 *) Make Montgomery context stuff explicit in RSA data structure. 3443 [Ben Laurie] 3444 3445 *) Fix build order of pem and err to allow for generated pem.h. 3446 [Ben Laurie] 3447 3448 *) Fix renumbering bug in X509_NAME_delete_entry(). 3449 [Ben Laurie] 3450 3451 *) Enhanced the err-ins.pl script so it makes the error library number 3452 global and can add a library name. This is needed for external ASN1 and 3453 other error libraries. 3454 [Steve Henson] 3455 3456 *) Fixed sk_insert which never worked properly. 3457 [Steve Henson] 3458 3459 *) Fix ASN1 macros so they can handle indefinite length construted 3460 EXPLICIT tags. Some non standard certificates use these: they can now 3461 be read in. 3462 [Steve Henson] 3463 3464 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 3465 into a single doc/ssleay.txt bundle. This way the information is still 3466 preserved but no longer messes up this directory. Now it's new room for 3467 the new set of documenation files. 3468 [Ralf S. Engelschall] 3469 3470 *) SETs were incorrectly DER encoded. This was a major pain, because they 3471 shared code with SEQUENCEs, which aren't coded the same. This means that 3472 almost everything to do with SETs or SEQUENCEs has either changed name or 3473 number of arguments. 3474 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 3475 3476 *) Fix test data to work with the above. 3477 [Ben Laurie] 3478 3479 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 3480 was already fixed by Eric for 0.9.1 it seems. 3481 [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 3482 3483 *) Autodetect FreeBSD3. 3484 [Ben Laurie] 3485 3486 *) Fix various bugs in Configure. This affects the following platforms: 3487 nextstep 3488 ncr-scde 3489 unixware-2.0 3490 unixware-2.0-pentium 3491 sco5-cc. 3492 [Ben Laurie] 3493 3494 *) Eliminate generated files from CVS. Reorder tests to regenerate files 3495 before they are needed. 3496 [Ben Laurie] 3497 3498 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 3499 [Ben Laurie] 3500 3501 3502 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 3503 3504 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 3505 changed SSLeay to OpenSSL in version strings. 3506 [Ralf S. Engelschall] 3507 3508 *) Some fixups to the top-level documents. 3509 [Paul Sutton] 3510 3511 *) Fixed the nasty bug where rsaref.h was not found under compile-time 3512 because the symlink to include/ was missing. 3513 [Ralf S. Engelschall] 3514 3515 *) Incorporated the popular no-RSA/DSA-only patches 3516 which allow to compile a RSA-free SSLeay. 3517 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 3518 3519 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 3520 when "ssleay" is still not found. 3521 [Ralf S. Engelschall] 3522 3523 *) Added more platforms to Configure: Cray T3E, HPUX 11, 3524 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 3525 3526 *) Updated the README file. 3527 [Ralf S. Engelschall] 3528 3529 *) Added various .cvsignore files in the CVS repository subdirs 3530 to make a "cvs update" really silent. 3531 [Ralf S. Engelschall] 3532 3533 *) Recompiled the error-definition header files and added 3534 missing symbols to the Win32 linker tables. 3535 [Ralf S. Engelschall] 3536 3537 *) Cleaned up the top-level documents; 3538 o new files: CHANGES and LICENSE 3539 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 3540 o merged COPYRIGHT into LICENSE 3541 o removed obsolete TODO file 3542 o renamed MICROSOFT to INSTALL.W32 3543 [Ralf S. Engelschall] 3544 3545 *) Removed dummy files from the 0.9.1b source tree: 3546 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 3547 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 3548 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 3549 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 3550 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 3551 [Ralf S. Engelschall] 3552 3553 *) Added various platform portability fixes. 3554 [Mark J. Cox] 3555 3556 *) The Genesis of the OpenSSL rpject: 3557 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 3558 Young and Tim J. Hudson created while they were working for C2Net until 3559 summer 1998. 3560 [The OpenSSL Project] 3561 3562 3563 Changes between 0.9.0b and 0.9.1b [not released] 3564 3565 *) Updated a few CA certificates under certs/ 3566 [Eric A. Young] 3567 3568 *) Changed some BIGNUM api stuff. 3569 [Eric A. Young] 3570 3571 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 3572 DGUX x86, Linux Alpha, etc. 3573 [Eric A. Young] 3574 3575 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 3576 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 3577 available). 3578 [Eric A. Young] 3579 3580 *) Add -strparse option to asn1pars program which parses nested 3581 binary structures 3582 [Dr Stephen Henson <shenson@bigfoot.com>] 3583 3584 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 3585 [Eric A. Young] 3586 3587 *) DSA fix for "ca" program. 3588 [Eric A. Young] 3589 3590 *) Added "-genkey" option to "dsaparam" program. 3591 [Eric A. Young] 3592 3593 *) Added RIPE MD160 (rmd160) message digest. 3594 [Eric A. Young] 3595 3596 *) Added -a (all) option to "ssleay version" command. 3597 [Eric A. Young] 3598 3599 *) Added PLATFORM define which is the id given to Configure. 3600 [Eric A. Young] 3601 3602 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 3603 [Eric A. Young] 3604 3605 *) Extended the ASN.1 parser routines. 3606 [Eric A. Young] 3607 3608 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 3609 [Eric A. Young] 3610 3611 *) Added a BN_CTX to the BN library. 3612 [Eric A. Young] 3613 3614 *) Fixed the weak key values in DES library 3615 [Eric A. Young] 3616 3617 *) Changed API in EVP library for cipher aliases. 3618 [Eric A. Young] 3619 3620 *) Added support for RC2/64bit cipher. 3621 [Eric A. Young] 3622 3623 *) Converted the lhash library to the crypto/mem.c functions. 3624 [Eric A. Young] 3625 3626 *) Added more recognized ASN.1 object ids. 3627 [Eric A. Young] 3628 3629 *) Added more RSA padding checks for SSL/TLS. 3630 [Eric A. Young] 3631 3632 *) Added BIO proxy/filter functionality. 3633 [Eric A. Young] 3634 3635 *) Added extra_certs to SSL_CTX which can be used 3636 send extra CA certificates to the client in the CA cert chain sending 3637 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 3638 [Eric A. Young] 3639 3640 *) Now Fortezza is denied in the authentication phase because 3641 this is key exchange mechanism is not supported by SSLeay at all. 3642 [Eric A. Young] 3643 3644 *) Additional PKCS1 checks. 3645 [Eric A. Young] 3646 3647 *) Support the string "TLSv1" for all TLS v1 ciphers. 3648 [Eric A. Young] 3649 3650 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 3651 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 3652 [Eric A. Young] 3653 3654 *) Fixed a few memory leaks. 3655 [Eric A. Young] 3656 3657 *) Fixed various code and comment typos. 3658 [Eric A. Young] 3659 3660 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 3661 bytes sent in the client random. 3662 [Edward Bishop <ebishop@spyglass.com>] 3663 3664