CHANGES revision 79998
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 6 7 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 8 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 9 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 10 PRNG state recovery was possible based on the output of 11 one PRNG request appropriately sized to gain knowledge on 12 'md' followed by enough consecutive 1-byte PRNG requests 13 to traverse all of 'state'. 14 15 1. When updating 'md_local' (the current thread's copy of 'md') 16 during PRNG output generation, hash all of the previous 17 'md_local' value, not just the half used for PRNG output. 18 19 2. Make the number of bytes from 'state' included into the hash 20 independent from the number of PRNG bytes requested. 21 22 The first measure alone would be sufficient to avoid 23 Markku-Juhani's attack. (Actually it had never occurred 24 to me that the half of 'md_local' used for chaining was the 25 half from which PRNG output bytes were taken -- I had always 26 assumed that the secret half would be used.) The second 27 measure makes sure that additional data from 'state' is never 28 mixed into 'md_local' in small portions; this heuristically 29 further strengthens the PRNG. 30 [Bodo Moeller] 31 32 *) Fix crypto/bn/asm/mips3.s. 33 [Andy Polyakov] 34 35 *) When only the key is given to "enc", the IV is undefined. Print out 36 an error message in this case. 37 [Lutz Jaenicke] 38 39 *) Handle special case when X509_NAME is empty in X509 printing routines. 40 [Steve Henson] 41 42 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 43 positive and less than q. 44 [Bodo Moeller] 45 46 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 47 used: it isn't thread safe and the add_lock_callback should handle 48 that itself. 49 [Paul Rose <Paul.Rose@bridge.com>] 50 51 *) Verify that incoming data obeys the block size in 52 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 53 [Bodo Moeller] 54 55 *) Fix OAEP check. 56 [Ulf M�ller, Bodo M�ller] 57 58 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 59 RSA encryption was accidentily removed in s3_srvr.c in OpenSSL 0.9.5 60 when fixing the server behaviour for backwards-compatible 'client 61 hello' messages. (Note that the attack is impractical against 62 SSL 3.0 and TLS 1.0 anyway because length and version checking 63 means that the probability of guessing a valid ciphertext is 64 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 65 paper.) 66 67 Before 0.9.5, the countermeasure (hide the error by generating a 68 random 'decryption result') did not work properly because 69 ERR_clear_error() was missing, meaning that SSL_get_error() would 70 detect the supposedly ignored error. 71 72 Both problems are now fixed. 73 [Bodo Moeller] 74 75 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 76 (previously it was 1024). 77 [Bodo Moeller] 78 79 *) Fix for compatibility mode trust settings: ignore trust settings 80 unless some valid trust or reject settings are present. 81 [Steve Henson] 82 83 *) Fix for blowfish EVP: its a variable length cipher. 84 [Steve Henson] 85 86 *) Fix various bugs related to DSA S/MIME verification. Handle missing 87 parameters in DSA public key structures and return an error in the 88 DSA routines if parameters are absent. 89 [Steve Henson] 90 91 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 92 in the current directory if neither $RANDFILE nor $HOME was set. 93 RAND_file_name() in 0.9.6a returned NULL in this case. This has 94 caused some confusion to Windows users who haven't defined $HOME. 95 Thus RAND_file_name() is changed again: e_os.h can define a 96 DEFAULT_HOME, which will be used if $HOME is not set. 97 For Windows, we use "C:"; on other platforms, we still require 98 environment variables. 99 100 *) Move 'if (!initialized) RAND_poll()' into regions protected by 101 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 102 having multiple threads call RAND_poll() concurrently. 103 [Bodo Moeller] 104 105 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 106 combination of a flag and a thread ID variable. 107 Otherwise while one thread is in ssleay_rand_bytes (which sets the 108 flag), *other* threads can enter ssleay_add_bytes without obeying 109 the CRYPTO_LOCK_RAND lock (and may even illegaly release the lock 110 that they do not hold after the first thread unsets add_do_not_lock). 111 [Bodo Moeller] 112 113 *) Change bctest again: '-x' expressions are not available in all 114 versions of 'test'. 115 [Bodo Moeller] 116 117 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 118 119 *) Fix a couple of memory leaks in PKCS7_dataDecode() 120 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 121 122 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 123 the default extension for executables, if any. Also, make the perl 124 scripts that use symlink() to test if it really exists and use "cp" 125 if it doesn't. All this made OpenSSL compilable and installable in 126 CygWin. 127 [Richard Levitte] 128 129 *) Fix for asn1_GetSequence() for indefinite length constructed data. 130 If SEQUENCE is length is indefinite just set c->slen to the total 131 amount of data available. 132 [Steve Henson, reported by shige@FreeBSD.org] 133 [This change does not apply to 0.9.7.] 134 135 *) Change bctest to avoid here-documents inside command substitution 136 (workaround for FreeBSD /bin/sh bug). 137 For compatibility with Ultrix, avoid shell functions (introduced 138 in the bctest version that searches along $PATH). 139 [Bodo Moeller] 140 141 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 142 with des_encrypt() defined on some operating systems, like Solaris 143 and UnixWare. 144 [Richard Levitte] 145 146 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 147 On the Importance of Eliminating Errors in Cryptographic 148 Computations, J. Cryptology 14 (2001) 2, 101-119, 149 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 150 [Ulf Moeller] 151 152 *) MIPS assembler BIGNUM division bug fix. 153 [Andy Polyakov] 154 155 *) Disabled incorrect Alpha assembler code. 156 [Richard Levitte] 157 158 *) Fix PKCS#7 decode routines so they correctly update the length 159 after reading an EOC for the EXPLICIT tag. 160 [Steve Henson] 161 [This change does not apply to 0.9.7.] 162 163 *) Fix bug in PKCS#12 key generation routines. This was triggered 164 if a 3DES key was generated with a 0 initial byte. Include 165 PKCS12_BROKEN_KEYGEN compilation option to retain the old 166 (but broken) behaviour. 167 [Steve Henson] 168 169 *) Enhance bctest to search for a working bc along $PATH and print 170 it when found. 171 [Tim Rice <tim@multitalents.net> via Richard Levitte] 172 173 *) Fix memory leaks in err.c: free err_data string if necessary; 174 don't write to the wrong index in ERR_set_error_data. 175 [Bodo Moeller] 176 177 *) Implement ssl23_peek (analogous to ssl23_read), which previously 178 did not exist. 179 [Bodo Moeller] 180 181 *) Replace rdtsc with _emit statements for VC++ version 5. 182 [Jeremy Cooper <jeremy@baymoo.org>] 183 184 *) Make it possible to reuse SSLv2 sessions. 185 [Richard Levitte] 186 187 *) In copy_email() check for >= 0 as a return value for 188 X509_NAME_get_index_by_NID() since 0 is a valid index. 189 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 190 191 *) Avoid coredump with unsupported or invalid public keys by checking if 192 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 193 PKCS7_verify() fails with non detached data. 194 [Steve Henson] 195 196 *) Don't use getenv in library functions when run as setuid/setgid. 197 New function OPENSSL_issetugid(). 198 [Ulf Moeller] 199 200 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 201 due to incorrect handling of multi-threading: 202 203 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 204 205 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 206 207 3. Count how many times MemCheck_off() has been called so that 208 nested use can be treated correctly. This also avoids 209 inband-signalling in the previous code (which relied on the 210 assumption that thread ID 0 is impossible). 211 [Bodo Moeller] 212 213 *) Add "-rand" option also to s_client and s_server. 214 [Lutz Jaenicke] 215 216 *) Fix CPU detection on Irix 6.x. 217 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 218 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 219 220 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 221 was empty. 222 [Steve Henson] 223 [This change does not apply to 0.9.7.] 224 225 *) Use the cached encoding of an X509_NAME structure rather than 226 copying it. This is apparently the reason for the libsafe "errors" 227 but the code is actually correct. 228 [Steve Henson] 229 230 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 231 Bleichenbacher's DSA attack. 232 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 233 to be set and top=0 forces the highest bit to be set; top=-1 is new 234 and leaves the highest bit random. 235 [Ulf Moeller, Bodo Moeller] 236 237 *) In the NCONF_...-based implementations for CONF_... queries 238 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 239 a temporary CONF structure with the data component set to NULL 240 (which gives segmentation faults in lh_retrieve). 241 Instead, use NULL for the CONF pointer in CONF_get_string and 242 CONF_get_number (which may use environment variables) and directly 243 return NULL from CONF_get_section. 244 [Bodo Moeller] 245 246 *) Fix potential buffer overrun for EBCDIC. 247 [Ulf Moeller] 248 249 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 250 keyUsage if basicConstraints absent for a CA. 251 [Steve Henson] 252 253 *) Make SMIME_write_PKCS7() write mail header values with a format that 254 is more generally accepted (no spaces before the semicolon), since 255 some programs can't parse those values properly otherwise. Also make 256 sure BIO's that break lines after each write do not create invalid 257 headers. 258 [Richard Levitte] 259 260 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 261 macros previously used would not encode an empty SEQUENCE OF 262 and break the signature. 263 [Steve Henson] 264 [This change does not apply to 0.9.7.] 265 266 *) Zero the premaster secret after deriving the master secret in 267 DH ciphersuites. 268 [Steve Henson] 269 270 *) Add some EVP_add_digest_alias registrations (as found in 271 OpenSSL_add_all_digests()) to SSL_library_init() 272 aka OpenSSL_add_ssl_algorithms(). This provides improved 273 compatibility with peers using X.509 certificates 274 with unconventional AlgorithmIdentifier OIDs. 275 [Bodo Moeller] 276 277 *) Fix for Irix with NO_ASM. 278 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 279 280 *) ./config script fixes. 281 [Ulf Moeller, Richard Levitte] 282 283 *) Fix 'openssl passwd -1'. 284 [Bodo Moeller] 285 286 *) Change PKCS12_key_gen_asc() so it can cope with non null 287 terminated strings whose length is passed in the passlen 288 parameter, for example from PEM callbacks. This was done 289 by adding an extra length parameter to asc2uni(). 290 [Steve Henson, reported by <oddissey@samsung.co.kr>] 291 292 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 293 call failed, free the DSA structure. 294 [Bodo Moeller] 295 296 *) Fix to uni2asc() to cope with zero length Unicode strings. 297 These are present in some PKCS#12 files. 298 [Steve Henson] 299 300 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 301 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 302 when writing a 32767 byte record. 303 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 304 305 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 306 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 307 308 (RSA objects have a reference count access to which is protected 309 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 310 so they are meant to be shared between threads.) 311 [Bodo Moeller, Geoff Thorpe; original patch submitted by 312 "Reddie, Steven" <Steven.Reddie@ca.com>] 313 314 *) Fix a deadlock in CRYPTO_mem_leaks(). 315 [Bodo Moeller] 316 317 *) Use better test patterns in bntest. 318 [Ulf M�ller] 319 320 *) rand_win.c fix for Borland C. 321 [Ulf M�ller] 322 323 *) BN_rshift bugfix for n == 0. 324 [Bodo Moeller] 325 326 *) Add a 'bctest' script that checks for some known 'bc' bugs 327 so that 'make test' does not abort just because 'bc' is broken. 328 [Bodo Moeller] 329 330 *) Store verify_result within SSL_SESSION also for client side to 331 avoid potential security hole. (Re-used sessions on the client side 332 always resulted in verify_result==X509_V_OK, not using the original 333 result of the server certificate verification.) 334 [Lutz Jaenicke] 335 336 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 337 SSL3_RT_APPLICATION_DATA, return 0. 338 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 339 [Bodo Moeller] 340 341 *) Fix SSL_peek: 342 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 343 releases, have been re-implemented by renaming the previous 344 implementations of ssl2_read and ssl3_read to ssl2_read_internal 345 and ssl3_read_internal, respectively, and adding 'peek' parameters 346 to them. The new ssl[23]_{read,peek} functions are calls to 347 ssl[23]_read_internal with the 'peek' flag set appropriately. 348 A 'peek' parameter has also been added to ssl3_read_bytes, which 349 does the actual work for ssl3_read_internal. 350 [Bodo Moeller] 351 352 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 353 the method-specific "init()" handler. Also clean up ex_data after 354 calling the method-specific "finish()" handler. Previously, this was 355 happening the other way round. 356 [Geoff Thorpe] 357 358 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 359 The previous value, 12, was not always sufficient for BN_mod_exp(). 360 [Bodo Moeller] 361 362 *) Make sure that shared libraries get the internal name engine with 363 the full version number and not just 0. This should mark the 364 shared libraries as not backward compatible. Of course, this should 365 be changed again when we can guarantee backward binary compatibility. 366 [Richard Levitte] 367 368 *) Fix typo in get_cert_by_subject() in by_dir.c 369 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 370 371 *) Rework the system to generate shared libraries: 372 373 - Make note of the expected extension for the shared libraries and 374 if there is a need for symbolic links from for example libcrypto.so.0 375 to libcrypto.so.0.9.7. There is extended info in Configure for 376 that. 377 378 - Make as few rebuilds of the shared libraries as possible. 379 380 - Still avoid linking the OpenSSL programs with the shared libraries. 381 382 - When installing, install the shared libraries separately from the 383 static ones. 384 [Richard Levitte] 385 386 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 387 388 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 389 and not in SSL_clear because the latter is also used by the 390 accept/connect functions; previously, the settings made by 391 SSL_set_read_ahead would be lost during the handshake. 392 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 393 394 *) Correct util/mkdef.pl to be selective about disabled algorithms. 395 Previously, it would create entries for disableed algorithms no 396 matter what. 397 [Richard Levitte] 398 399 *) Added several new manual pages for SSL_* function. 400 [Lutz Jaenicke] 401 402 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 403 404 *) In ssl23_get_client_hello, generate an error message when faced 405 with an initial SSL 3.0/TLS record that is too small to contain the 406 first two bytes of the ClientHello message, i.e. client_version. 407 (Note that this is a pathologic case that probably has never happened 408 in real life.) The previous approach was to use the version number 409 from the record header as a substitute; but our protocol choice 410 should not depend on that one because it is not authenticated 411 by the Finished messages. 412 [Bodo Moeller] 413 414 *) More robust randomness gathering functions for Windows. 415 [Jeffrey Altman <jaltman@columbia.edu>] 416 417 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 418 not set then we don't setup the error code for issuer check errors 419 to avoid possibly overwriting other errors which the callback does 420 handle. If an application does set the flag then we assume it knows 421 what it is doing and can handle the new informational codes 422 appropriately. 423 [Steve Henson] 424 425 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 426 a general "ANY" type, as such it should be able to decode anything 427 including tagged types. However it didn't check the class so it would 428 wrongly interpret tagged types in the same way as their universal 429 counterpart and unknown types were just rejected. Changed so that the 430 tagged and unknown types are handled in the same way as a SEQUENCE: 431 that is the encoding is stored intact. There is also a new type 432 "V_ASN1_OTHER" which is used when the class is not universal, in this 433 case we have no idea what the actual type is so we just lump them all 434 together. 435 [Steve Henson] 436 437 *) On VMS, stdout may very well lead to a file that is written to 438 in a record-oriented fashion. That means that every write() will 439 write a separate record, which will be read separately by the 440 programs trying to read from it. This can be very confusing. 441 442 The solution is to put a BIO filter in the way that will buffer 443 text until a linefeed is reached, and then write everything a 444 line at a time, so every record written will be an actual line, 445 not chunks of lines and not (usually doesn't happen, but I've 446 seen it once) several lines in one record. BIO_f_linebuffer() is 447 the answer. 448 449 Currently, it's a VMS-only method, because that's where it has 450 been tested well enough. 451 [Richard Levitte] 452 453 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 454 it can return incorrect results. 455 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 456 but it was in 0.9.6-beta[12].) 457 [Bodo Moeller] 458 459 *) Disable the check for content being present when verifying detached 460 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 461 include zero length content when signing messages. 462 [Steve Henson] 463 464 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 465 BIO_ctrl (for BIO pairs). 466 [Bodo M�ller] 467 468 *) Add DSO method for VMS. 469 [Richard Levitte] 470 471 *) Bug fix: Montgomery multiplication could produce results with the 472 wrong sign. 473 [Ulf M�ller] 474 475 *) Add RPM specification openssl.spec and modify it to build three 476 packages. The default package contains applications, application 477 documentation and run-time libraries. The devel package contains 478 include files, static libraries and function documentation. The 479 doc package contains the contents of the doc directory. The original 480 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 481 [Richard Levitte] 482 483 *) Add a large number of documentation files for many SSL routines. 484 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 485 486 *) Add a configuration entry for Sony News 4. 487 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 488 489 *) Don't set the two most significant bits to one when generating a 490 random number < q in the DSA library. 491 [Ulf M�ller] 492 493 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 494 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 495 the underlying transport is blocking) if a handshake took place. 496 (The default behaviour is needed by applications such as s_client 497 and s_server that use select() to determine when to use SSL_read; 498 but for applications that know in advance when to expect data, it 499 just makes things more complicated.) 500 [Bodo Moeller] 501 502 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 503 from EGD. 504 [Ben Laurie] 505 506 *) Add a few more EBCDIC conditionals that make `req' and `x509' 507 work better on such systems. 508 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 509 510 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 511 Update PKCS12_parse() so it copies the friendlyName and the 512 keyid to the certificates aux info. 513 [Steve Henson] 514 515 *) Fix bug in PKCS7_verify() which caused an infinite loop 516 if there was more than one signature. 517 [Sven Uszpelkat <su@celocom.de>] 518 519 *) Major change in util/mkdef.pl to include extra information 520 about each symbol, as well as presentig variables as well 521 as functions. This change means that there's n more need 522 to rebuild the .num files when some algorithms are excluded. 523 [Richard Levitte] 524 525 *) Allow the verify time to be set by an application, 526 rather than always using the current time. 527 [Steve Henson] 528 529 *) Phase 2 verify code reorganisation. The certificate 530 verify code now looks up an issuer certificate by a 531 number of criteria: subject name, authority key id 532 and key usage. It also verifies self signed certificates 533 by the same criteria. The main comparison function is 534 X509_check_issued() which performs these checks. 535 536 Lot of changes were necessary in order to support this 537 without completely rewriting the lookup code. 538 539 Authority and subject key identifier are now cached. 540 541 The LHASH 'certs' is X509_STORE has now been replaced 542 by a STACK_OF(X509_OBJECT). This is mainly because an 543 LHASH can't store or retrieve multiple objects with 544 the same hash value. 545 546 As a result various functions (which were all internal 547 use only) have changed to handle the new X509_STORE 548 structure. This will break anything that messed round 549 with X509_STORE internally. 550 551 The functions X509_STORE_add_cert() now checks for an 552 exact match, rather than just subject name. 553 554 The X509_STORE API doesn't directly support the retrieval 555 of multiple certificates matching a given criteria, however 556 this can be worked round by performing a lookup first 557 (which will fill the cache with candidate certificates) 558 and then examining the cache for matches. This is probably 559 the best we can do without throwing out X509_LOOKUP 560 entirely (maybe later...). 561 562 The X509_VERIFY_CTX structure has been enhanced considerably. 563 564 All certificate lookup operations now go via a get_issuer() 565 callback. Although this currently uses an X509_STORE it 566 can be replaced by custom lookups. This is a simple way 567 to bypass the X509_STORE hackery necessary to make this 568 work and makes it possible to use more efficient techniques 569 in future. A very simple version which uses a simple 570 STACK for its trusted certificate store is also provided 571 using X509_STORE_CTX_trusted_stack(). 572 573 The verify_cb() and verify() callbacks now have equivalents 574 in the X509_STORE_CTX structure. 575 576 X509_STORE_CTX also has a 'flags' field which can be used 577 to customise the verify behaviour. 578 [Steve Henson] 579 580 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 581 excludes S/MIME capabilities. 582 [Steve Henson] 583 584 *) When a certificate request is read in keep a copy of the 585 original encoding of the signed data and use it when outputing 586 again. Signatures then use the original encoding rather than 587 a decoded, encoded version which may cause problems if the 588 request is improperly encoded. 589 [Steve Henson] 590 591 *) For consistency with other BIO_puts implementations, call 592 buffer_write(b, ...) directly in buffer_puts instead of calling 593 BIO_write(b, ...). 594 595 In BIO_puts, increment b->num_write as in BIO_write. 596 [Peter.Sylvester@EdelWeb.fr] 597 598 *) Fix BN_mul_word for the case where the word is 0. (We have to use 599 BN_zero, we may not return a BIGNUM with an array consisting of 600 words set to zero.) 601 [Bodo Moeller] 602 603 *) Avoid calling abort() from within the library when problems are 604 detected, except if preprocessor symbols have been defined 605 (such as REF_CHECK, BN_DEBUG etc.). 606 [Bodo Moeller] 607 608 *) New openssl application 'rsautl'. This utility can be 609 used for low level RSA operations. DER public key 610 BIO/fp routines also added. 611 [Steve Henson] 612 613 *) New Configure entry and patches for compiling on QNX 4. 614 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 615 616 *) A demo state-machine implementation was sponsored by 617 Nuron (http://www.nuron.com/) and is now available in 618 demos/state_machine. 619 [Ben Laurie] 620 621 *) New options added to the 'dgst' utility for signature 622 generation and verification. 623 [Steve Henson] 624 625 *) Unrecognized PKCS#7 content types are now handled via a 626 catch all ASN1_TYPE structure. This allows unsupported 627 types to be stored as a "blob" and an application can 628 encode and decode it manually. 629 [Steve Henson] 630 631 *) Fix various signed/unsigned issues to make a_strex.c 632 compile under VC++. 633 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 634 635 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 636 length if passed a buffer. ASN1_INTEGER_to_BN failed 637 if passed a NULL BN and its argument was negative. 638 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 639 640 *) Modification to PKCS#7 encoding routines to output definite 641 length encoding. Since currently the whole structures are in 642 memory there's not real point in using indefinite length 643 constructed encoding. However if OpenSSL is compiled with 644 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 645 [Steve Henson] 646 647 *) Added BIO_vprintf() and BIO_vsnprintf(). 648 [Richard Levitte] 649 650 *) Added more prefixes to parse for in the the strings written 651 through a logging bio, to cover all the levels that are available 652 through syslog. The prefixes are now: 653 654 PANIC, EMERG, EMR => LOG_EMERG 655 ALERT, ALR => LOG_ALERT 656 CRIT, CRI => LOG_CRIT 657 ERROR, ERR => LOG_ERR 658 WARNING, WARN, WAR => LOG_WARNING 659 NOTICE, NOTE, NOT => LOG_NOTICE 660 INFO, INF => LOG_INFO 661 DEBUG, DBG => LOG_DEBUG 662 663 and as before, if none of those prefixes are present at the 664 beginning of the string, LOG_ERR is chosen. 665 666 On Win32, the LOG_* levels are mapped according to this: 667 668 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 669 LOG_WARNING => EVENTLOG_WARNING_TYPE 670 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 671 672 [Richard Levitte] 673 674 *) Made it possible to reconfigure with just the configuration 675 argument "reconf" or "reconfigure". The command line arguments 676 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 677 and are retrieved from there when reconfiguring. 678 [Richard Levitte] 679 680 *) MD4 implemented. 681 [Assar Westerlund <assar@sics.se>, Richard Levitte] 682 683 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 684 [Richard Levitte] 685 686 *) The obj_dat.pl script was messing up the sorting of object 687 names. The reason was that it compared the quoted version 688 of strings as a result "OCSP" > "OCSP Signing" because 689 " > SPACE. Changed script to store unquoted versions of 690 names and add quotes on output. It was also omitting some 691 names from the lookup table if they were given a default 692 value (that is if SN is missing it is given the same 693 value as LN and vice versa), these are now added on the 694 grounds that if an object has a name we should be able to 695 look it up. Finally added warning output when duplicate 696 short or long names are found. 697 [Steve Henson] 698 699 *) Changes needed for Tandem NSK. 700 [Scott Uroff <scott@xypro.com>] 701 702 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 703 RSA_padding_check_SSLv23(), special padding was never detected 704 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 705 version rollback attacks was not effective. 706 707 In s23_clnt.c, don't use special rollback-attack detection padding 708 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 709 client; similarly, in s23_srvr.c, don't do the rollback check if 710 SSL 2.0 is the only protocol enabled in the server. 711 [Bodo Moeller] 712 713 *) Make it possible to get hexdumps of unprintable data with 'openssl 714 asn1parse'. By implication, the functions ASN1_parse_dump() and 715 BIO_dump_indent() are added. 716 [Richard Levitte] 717 718 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 719 these print out strings and name structures based on various 720 flags including RFC2253 support and proper handling of 721 multibyte characters. Added options to the 'x509' utility 722 to allow the various flags to be set. 723 [Steve Henson] 724 725 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 726 Also change the functions X509_cmp_current_time() and 727 X509_gmtime_adj() work with an ASN1_TIME structure, 728 this will enable certificates using GeneralizedTime in validity 729 dates to be checked. 730 [Steve Henson] 731 732 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 733 negative public key encodings) on by default, 734 NO_NEG_PUBKEY_BUG can be set to disable it. 735 [Steve Henson] 736 737 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 738 content octets. An i2c_ASN1_OBJECT is unnecessary because 739 the encoding can be trivially obtained from the structure. 740 [Steve Henson] 741 742 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 743 not read locks (CRYPTO_r_[un]lock). 744 [Bodo Moeller] 745 746 *) A first attempt at creating official support for shared 747 libraries through configuration. I've kept it so the 748 default is static libraries only, and the OpenSSL programs 749 are always statically linked for now, but there are 750 preparations for dynamic linking in place. 751 This has been tested on Linux and True64. 752 [Richard Levitte] 753 754 *) Randomness polling function for Win9x, as described in: 755 Peter Gutmann, Software Generation of Practically Strong 756 Random Numbers. 757 [Ulf M�ller] 758 759 *) Fix so PRNG is seeded in req if using an already existing 760 DSA key. 761 [Steve Henson] 762 763 *) New options to smime application. -inform and -outform 764 allow alternative formats for the S/MIME message including 765 PEM and DER. The -content option allows the content to be 766 specified separately. This should allow things like Netscape 767 form signing output easier to verify. 768 [Steve Henson] 769 770 *) Fix the ASN1 encoding of tags using the 'long form'. 771 [Steve Henson] 772 773 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 774 STRING types. These convert content octets to and from the 775 underlying type. The actual tag and length octets are 776 already assumed to have been read in and checked. These 777 are needed because all other string types have virtually 778 identical handling apart from the tag. By having versions 779 of the ASN1 functions that just operate on content octets 780 IMPLICIT tagging can be handled properly. It also allows 781 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 782 and ASN1_INTEGER are identical apart from the tag. 783 [Steve Henson] 784 785 *) Change the handling of OID objects as follows: 786 787 - New object identifiers are inserted in objects.txt, following 788 the syntax given in objects.README. 789 - objects.pl is used to process obj_mac.num and create a new 790 obj_mac.h. 791 - obj_dat.pl is used to create a new obj_dat.h, using the data in 792 obj_mac.h. 793 794 This is currently kind of a hack, and the perl code in objects.pl 795 isn't very elegant, but it works as I intended. The simplest way 796 to check that it worked correctly is to look in obj_dat.h and 797 check the array nid_objs and make sure the objects haven't moved 798 around (this is important!). Additions are OK, as well as 799 consistent name changes. 800 [Richard Levitte] 801 802 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 803 [Bodo Moeller] 804 805 *) Addition of the command line parameter '-rand file' to 'openssl req'. 806 The given file adds to whatever has already been seeded into the 807 random pool through the RANDFILE configuration file option or 808 environment variable, or the default random state file. 809 [Richard Levitte] 810 811 *) mkstack.pl now sorts each macro group into lexical order. 812 Previously the output order depended on the order the files 813 appeared in the directory, resulting in needless rewriting 814 of safestack.h . 815 [Steve Henson] 816 817 *) Patches to make OpenSSL compile under Win32 again. Mostly 818 work arounds for the VC++ problem that it treats func() as 819 func(void). Also stripped out the parts of mkdef.pl that 820 added extra typesafe functions: these no longer exist. 821 [Steve Henson] 822 823 *) Reorganisation of the stack code. The macros are now all 824 collected in safestack.h . Each macro is defined in terms of 825 a "stack macro" of the form SKM_<name>(type, a, b). The 826 DEBUG_SAFESTACK is now handled in terms of function casts, 827 this has the advantage of retaining type safety without the 828 use of additional functions. If DEBUG_SAFESTACK is not defined 829 then the non typesafe macros are used instead. Also modified the 830 mkstack.pl script to handle the new form. Needs testing to see 831 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 832 the default if no major problems. Similar behaviour for ASN1_SET_OF 833 and PKCS12_STACK_OF. 834 [Steve Henson] 835 836 *) When some versions of IIS use the 'NET' form of private key the 837 key derivation algorithm is different. Normally MD5(password) is 838 used as a 128 bit RC4 key. In the modified case 839 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 840 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 841 as the old Netscape_RSA functions except they have an additional 842 'sgckey' parameter which uses the modified algorithm. Also added 843 an -sgckey command line option to the rsa utility. Thanks to 844 Adrian Peck <bertie@ncipher.com> for posting details of the modified 845 algorithm to openssl-dev. 846 [Steve Henson] 847 848 *) The evp_local.h macros were using 'c.##kname' which resulted in 849 invalid expansion on some systems (SCO 5.0.5 for example). 850 Corrected to 'c.kname'. 851 [Phillip Porch <root@theporch.com>] 852 853 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 854 a STACK of email addresses from a certificate or request, these look 855 in the subject name and the subject alternative name extensions and 856 omit any duplicate addresses. 857 [Steve Henson] 858 859 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 860 This makes DSA verification about 2 % faster. 861 [Bodo Moeller] 862 863 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 864 (meaning that now 2^5 values will be precomputed, which is only 4 KB 865 plus overhead for 1024 bit moduli). 866 This makes exponentiations about 0.5 % faster for 1024 bit 867 exponents (as measured by "openssl speed rsa2048"). 868 [Bodo Moeller] 869 870 *) Rename memory handling macros to avoid conflicts with other 871 software: 872 Malloc => OPENSSL_malloc 873 Malloc_locked => OPENSSL_malloc_locked 874 Realloc => OPENSSL_realloc 875 Free => OPENSSL_free 876 [Richard Levitte] 877 878 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 879 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 880 [Bodo Moeller] 881 882 *) CygWin32 support. 883 [John Jarvie <jjarvie@newsguy.com>] 884 885 *) The type-safe stack code has been rejigged. It is now only compiled 886 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 887 by default all type-specific stack functions are "#define"d back to 888 standard stack functions. This results in more streamlined output 889 but retains the type-safety checking possibilities of the original 890 approach. 891 [Geoff Thorpe] 892 893 *) The STACK code has been cleaned up, and certain type declarations 894 that didn't make a lot of sense have been brought in line. This has 895 also involved a cleanup of sorts in safestack.h to more correctly 896 map type-safe stack functions onto their plain stack counterparts. 897 This work has also resulted in a variety of "const"ifications of 898 lots of the code, especially "_cmp" operations which should normally 899 be prototyped with "const" parameters anyway. 900 [Geoff Thorpe] 901 902 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 903 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 904 (The PRNG state consists of two parts, the large pool 'state' and 'md', 905 where all of 'md' is used each time the PRNG is used, but 'state' 906 is used only indexed by a cyclic counter. As entropy may not be 907 well distributed from the beginning, 'md' is important as a 908 chaining variable. However, the output function chains only half 909 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 910 all of 'md', and seeding with STATE_SIZE dummy bytes will result 911 in all of 'state' being rewritten, with the new values depending 912 on virtually all of 'md'. This overcomes the 80 bit limitation.) 913 [Bodo Moeller] 914 915 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 916 the handshake is continued after ssl_verify_cert_chain(); 917 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 918 can lead to 'unexplainable' connection aborts later. 919 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 920 921 *) Major EVP API cipher revision. 922 Add hooks for extra EVP features. This allows various cipher 923 parameters to be set in the EVP interface. Support added for variable 924 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 925 setting of RC2 and RC5 parameters. 926 927 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 928 ciphers. 929 930 Remove lots of duplicated code from the EVP library. For example *every* 931 cipher init() function handles the 'iv' in the same way according to the 932 cipher mode. They also all do nothing if the 'key' parameter is NULL and 933 for CFB and OFB modes they zero ctx->num. 934 935 New functionality allows removal of S/MIME code RC2 hack. 936 937 Most of the routines have the same form and so can be declared in terms 938 of macros. 939 940 By shifting this to the top level EVP_CipherInit() it can be removed from 941 all individual ciphers. If the cipher wants to handle IVs or keys 942 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 943 flags. 944 945 Change lots of functions like EVP_EncryptUpdate() to now return a 946 value: although software versions of the algorithms cannot fail 947 any installed hardware versions can. 948 [Steve Henson] 949 950 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 951 this option is set, tolerate broken clients that send the negotiated 952 protocol version number instead of the requested protocol version 953 number. 954 [Bodo Moeller] 955 956 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 957 i.e. non-zero for export ciphersuites, zero otherwise. 958 Previous versions had this flag inverted, inconsistent with 959 rsa_tmp_cb (..._TMP_RSA_CB). 960 [Bodo Moeller; problem reported by Amit Chopra] 961 962 *) Add missing DSA library text string. Work around for some IIS 963 key files with invalid SEQUENCE encoding. 964 [Steve Henson] 965 966 *) Add a document (doc/standards.txt) that list all kinds of standards 967 and so on that are implemented in OpenSSL. 968 [Richard Levitte] 969 970 *) Enhance c_rehash script. Old version would mishandle certificates 971 with the same subject name hash and wouldn't handle CRLs at all. 972 Added -fingerprint option to crl utility, to support new c_rehash 973 features. 974 [Steve Henson] 975 976 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 977 [Ulf M�ller] 978 979 *) Fix for SSL server purpose checking. Server checking was 980 rejecting certificates which had extended key usage present 981 but no ssl client purpose. 982 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 983 984 *) Make PKCS#12 code work with no password. The PKCS#12 spec 985 is a little unclear about how a blank password is handled. 986 Since the password in encoded as a BMPString with terminating 987 double NULL a zero length password would end up as just the 988 double NULL. However no password at all is different and is 989 handled differently in the PKCS#12 key generation code. NS 990 treats a blank password as zero length. MSIE treats it as no 991 password on export: but it will try both on import. We now do 992 the same: PKCS12_parse() tries zero length and no password if 993 the password is set to "" or NULL (NULL is now a valid password: 994 it wasn't before) as does the pkcs12 application. 995 [Steve Henson] 996 997 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 998 perror when PEM_read_bio_X509_REQ fails, the error message must 999 be obtained from the error queue. 1000 [Bodo Moeller] 1001 1002 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 1003 it in ERR_remove_state if appropriate, and change ERR_get_state 1004 accordingly to avoid race conditions (this is necessary because 1005 thread_hash is no longer constant once set). 1006 [Bodo Moeller] 1007 1008 *) Bugfix for linux-elf makefile.one. 1009 [Ulf M�ller] 1010 1011 *) RSA_get_default_method() will now cause a default 1012 RSA_METHOD to be chosen if one doesn't exist already. 1013 Previously this was only set during a call to RSA_new() 1014 or RSA_new_method(NULL) meaning it was possible for 1015 RSA_get_default_method() to return NULL. 1016 [Geoff Thorpe] 1017 1018 *) Added native name translation to the existing DSO code 1019 that will convert (if the flag to do so is set) filenames 1020 that are sufficiently small and have no path information 1021 into a canonical native form. Eg. "blah" converted to 1022 "libblah.so" or "blah.dll" etc. 1023 [Geoff Thorpe] 1024 1025 *) New function ERR_error_string_n(e, buf, len) which is like 1026 ERR_error_string(e, buf), but writes at most 'len' bytes 1027 including the 0 terminator. For ERR_error_string_n, 'buf' 1028 may not be NULL. 1029 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 1030 1031 *) CONF library reworked to become more general. A new CONF 1032 configuration file reader "class" is implemented as well as a 1033 new functions (NCONF_*, for "New CONF") to handle it. The now 1034 old CONF_* functions are still there, but are reimplemented to 1035 work in terms of the new functions. Also, a set of functions 1036 to handle the internal storage of the configuration data is 1037 provided to make it easier to write new configuration file 1038 reader "classes" (I can definitely see something reading a 1039 configuration file in XML format, for example), called _CONF_*, 1040 or "the configuration storage API"... 1041 1042 The new configuration file reading functions are: 1043 1044 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 1045 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 1046 1047 NCONF_default, NCONF_WIN32 1048 1049 NCONF_dump_fp, NCONF_dump_bio 1050 1051 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 1052 NCONF_new creates a new CONF object. This works in the same way 1053 as other interfaces in OpenSSL, like the BIO interface. 1054 NCONF_dump_* dump the internal storage of the configuration file, 1055 which is useful for debugging. All other functions take the same 1056 arguments as the old CONF_* functions wth the exception of the 1057 first that must be a `CONF *' instead of a `LHASH *'. 1058 1059 To make it easer to use the new classes with the old CONF_* functions, 1060 the function CONF_set_default_method is provided. 1061 [Richard Levitte] 1062 1063 *) Add '-tls1' option to 'openssl ciphers', which was already 1064 mentioned in the documentation but had not been implemented. 1065 (This option is not yet really useful because even the additional 1066 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 1067 [Bodo Moeller] 1068 1069 *) Initial DSO code added into libcrypto for letting OpenSSL (and 1070 OpenSSL-based applications) load shared libraries and bind to 1071 them in a portable way. 1072 [Geoff Thorpe, with contributions from Richard Levitte] 1073 1074 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 1075 1076 *) Make sure _lrotl and _lrotr are only used with MSVC. 1077 1078 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 1079 (the default implementation of RAND_status). 1080 1081 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 1082 to '-clrext' (= clear extensions), as intended and documented. 1083 [Bodo Moeller; inconsistency pointed out by Michael Attili 1084 <attili@amaxo.com>] 1085 1086 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 1087 was larger than the MD block size. 1088 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 1089 1090 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 1091 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 1092 using the passed key: if the passed key was a private key the result 1093 of X509_print(), for example, would be to print out all the private key 1094 components. 1095 [Steve Henson] 1096 1097 *) des_quad_cksum() byte order bug fix. 1098 [Ulf M�ller, using the problem description in krb4-0.9.7, where 1099 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 1100 1101 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 1102 discouraged. 1103 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 1104 1105 *) For easily testing in shell scripts whether some command 1106 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 1107 returns with exit code 0 iff no command of the given name is available. 1108 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 1109 the output goes to stdout and nothing is printed to stderr. 1110 Additional arguments are always ignored. 1111 1112 Since for each cipher there is a command of the same name, 1113 the 'no-cipher' compilation switches can be tested this way. 1114 1115 ('openssl no-XXX' is not able to detect pseudo-commands such 1116 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 1117 [Bodo Moeller] 1118 1119 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 1120 [Bodo Moeller] 1121 1122 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 1123 is set; it will be thrown away anyway because each handshake creates 1124 its own key. 1125 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 1126 to parameters -- in previous versions (since OpenSSL 0.9.3) the 1127 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 1128 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 1129 [Bodo Moeller] 1130 1131 *) New s_client option -ign_eof: EOF at stdin is ignored, and 1132 'Q' and 'R' lose their special meanings (quit/renegotiate). 1133 This is part of what -quiet does; unlike -quiet, -ign_eof 1134 does not suppress any output. 1135 [Richard Levitte] 1136 1137 *) Add compatibility options to the purpose and trust code. The 1138 purpose X509_PURPOSE_ANY is "any purpose" which automatically 1139 accepts a certificate or CA, this was the previous behaviour, 1140 with all the associated security issues. 1141 1142 X509_TRUST_COMPAT is the old trust behaviour: only and 1143 automatically trust self signed roots in certificate store. A 1144 new trust setting X509_TRUST_DEFAULT is used to specify that 1145 a purpose has no associated trust setting and it should instead 1146 use the value in the default purpose. 1147 [Steve Henson] 1148 1149 *) Fix the PKCS#8 DSA private key code so it decodes keys again 1150 and fix a memory leak. 1151 [Steve Henson] 1152 1153 *) In util/mkerr.pl (which implements 'make errors'), preserve 1154 reason strings from the previous version of the .c file, as 1155 the default to have only downcase letters (and digits) in 1156 automatically generated reasons codes is not always appropriate. 1157 [Bodo Moeller] 1158 1159 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 1160 using strerror. Previously, ERR_reason_error_string() returned 1161 library names as reason strings for SYSerr; but SYSerr is a special 1162 case where small numbers are errno values, not library numbers. 1163 [Bodo Moeller] 1164 1165 *) Add '-dsaparam' option to 'openssl dhparam' application. This 1166 converts DSA parameters into DH parameters. (When creating parameters, 1167 DSA_generate_parameters is used.) 1168 [Bodo Moeller] 1169 1170 *) Include 'length' (recommended exponent length) in C code generated 1171 by 'openssl dhparam -C'. 1172 [Bodo Moeller] 1173 1174 *) The second argument to set_label in perlasm was already being used 1175 so couldn't be used as a "file scope" flag. Moved to third argument 1176 which was free. 1177 [Steve Henson] 1178 1179 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 1180 instead of RAND_bytes for encryption IVs and salts. 1181 [Bodo Moeller] 1182 1183 *) Include RAND_status() into RAND_METHOD instead of implementing 1184 it only for md_rand.c Otherwise replacing the PRNG by calling 1185 RAND_set_rand_method would be impossible. 1186 [Bodo Moeller] 1187 1188 *) Don't let DSA_generate_key() enter an infinite loop if the random 1189 number generation fails. 1190 [Bodo Moeller] 1191 1192 *) New 'rand' application for creating pseudo-random output. 1193 [Bodo Moeller] 1194 1195 *) Added configuration support for Linux/IA64 1196 [Rolf Haberrecker <rolf@suse.de>] 1197 1198 *) Assembler module support for Mingw32. 1199 [Ulf M�ller] 1200 1201 *) Shared library support for HPUX (in shlib/). 1202 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 1203 1204 *) Shared library support for Solaris gcc. 1205 [Lutz Behnke <behnke@trustcenter.de>] 1206 1207 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 1208 1209 *) PKCS7_encrypt() was adding text MIME headers twice because they 1210 were added manually and by SMIME_crlf_copy(). 1211 [Steve Henson] 1212 1213 *) In bntest.c don't call BN_rand with zero bits argument. 1214 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 1215 1216 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 1217 case was implemented. This caused BN_div_recp() to fail occasionally. 1218 [Ulf M�ller] 1219 1220 *) Add an optional second argument to the set_label() in the perl 1221 assembly language builder. If this argument exists and is set 1222 to 1 it signals that the assembler should use a symbol whose 1223 scope is the entire file, not just the current function. This 1224 is needed with MASM which uses the format label:: for this scope. 1225 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 1226 1227 *) Change the ASN1 types so they are typedefs by default. Before 1228 almost all types were #define'd to ASN1_STRING which was causing 1229 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 1230 for example. 1231 [Steve Henson] 1232 1233 *) Change names of new functions to the new get1/get0 naming 1234 convention: After 'get1', the caller owns a reference count 1235 and has to call ..._free; 'get0' returns a pointer to some 1236 data structure without incrementing reference counters. 1237 (Some of the existing 'get' functions increment a reference 1238 counter, some don't.) 1239 Similarly, 'set1' and 'add1' functions increase reference 1240 counters or duplicate objects. 1241 [Steve Henson] 1242 1243 *) Allow for the possibility of temp RSA key generation failure: 1244 the code used to assume it always worked and crashed on failure. 1245 [Steve Henson] 1246 1247 *) Fix potential buffer overrun problem in BIO_printf(). 1248 [Ulf M�ller, using public domain code by Patrick Powell; problem 1249 pointed out by David Sacerdote <das33@cornell.edu>] 1250 1251 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 1252 RAND_egd() and RAND_status(). In the command line application, 1253 the EGD socket can be specified like a seed file using RANDFILE 1254 or -rand. 1255 [Ulf M�ller] 1256 1257 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 1258 Some CAs (e.g. Verisign) distribute certificates in this form. 1259 [Steve Henson] 1260 1261 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 1262 list to exclude them. This means that no special compilation option 1263 is needed to use anonymous DH: it just needs to be included in the 1264 cipher list. 1265 [Steve Henson] 1266 1267 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 1268 EVP_MD_type. The old functionality is available in a new macro called 1269 EVP_MD_md(). Change code that uses it and update docs. 1270 [Steve Henson] 1271 1272 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 1273 where the 'void *' argument is replaced by a function pointer argument. 1274 Previously 'void *' was abused to point to functions, which works on 1275 many platforms, but is not correct. As these functions are usually 1276 called by macros defined in OpenSSL header files, most source code 1277 should work without changes. 1278 [Richard Levitte] 1279 1280 *) <openssl/opensslconf.h> (which is created by Configure) now contains 1281 sections with information on -D... compiler switches used for 1282 compiling the library so that applications can see them. To enable 1283 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 1284 must be defined. E.g., 1285 #define OPENSSL_ALGORITHM_DEFINES 1286 #include <openssl/opensslconf.h> 1287 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 1288 [Richard Levitte, Ulf and Bodo M�ller] 1289 1290 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 1291 record layer. 1292 [Bodo Moeller] 1293 1294 *) Change the 'other' type in certificate aux info to a STACK_OF 1295 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 1296 the required ASN1 format: arbitrary types determined by an OID. 1297 [Steve Henson] 1298 1299 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 1300 argument to 'req'. This is not because the function is newer or 1301 better than others it just uses the work 'NEW' in the certificate 1302 request header lines. Some software needs this. 1303 [Steve Henson] 1304 1305 *) Reorganise password command line arguments: now passwords can be 1306 obtained from various sources. Delete the PEM_cb function and make 1307 it the default behaviour: i.e. if the callback is NULL and the 1308 usrdata argument is not NULL interpret it as a null terminated pass 1309 phrase. If usrdata and the callback are NULL then the pass phrase 1310 is prompted for as usual. 1311 [Steve Henson] 1312 1313 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 1314 the support is automatically enabled. The resulting binaries will 1315 autodetect the card and use it if present. 1316 [Ben Laurie and Compaq Inc.] 1317 1318 *) Work around for Netscape hang bug. This sends certificate request 1319 and server done in one record. Since this is perfectly legal in the 1320 SSL/TLS protocol it isn't a "bug" option and is on by default. See 1321 the bugs/SSLv3 entry for more info. 1322 [Steve Henson] 1323 1324 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 1325 [Andy Polyakov] 1326 1327 *) Add -rand argument to smime and pkcs12 applications and read/write 1328 of seed file. 1329 [Steve Henson] 1330 1331 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 1332 [Bodo Moeller] 1333 1334 *) Add command line password options to the remaining applications. 1335 [Steve Henson] 1336 1337 *) Bug fix for BN_div_recp() for numerators with an even number of 1338 bits. 1339 [Ulf M�ller] 1340 1341 *) More tests in bntest.c, and changed test_bn output. 1342 [Ulf M�ller] 1343 1344 *) ./config recognizes MacOS X now. 1345 [Andy Polyakov] 1346 1347 *) Bug fix for BN_div() when the first words of num and divsor are 1348 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 1349 [Ulf M�ller] 1350 1351 *) Add support for various broken PKCS#8 formats, and command line 1352 options to produce them. 1353 [Steve Henson] 1354 1355 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 1356 get temporary BIGNUMs from a BN_CTX. 1357 [Ulf M�ller] 1358 1359 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 1360 for p == 0. 1361 [Ulf M�ller] 1362 1363 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 1364 include a #define from the old name to the new. The original intent 1365 was that statically linked binaries could for example just call 1366 SSLeay_add_all_ciphers() to just add ciphers to the table and not 1367 link with digests. This never worked becayse SSLeay_add_all_digests() 1368 and SSLeay_add_all_ciphers() were in the same source file so calling 1369 one would link with the other. They are now in separate source files. 1370 [Steve Henson] 1371 1372 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 1373 [Steve Henson] 1374 1375 *) Use a less unusual form of the Miller-Rabin primality test (it used 1376 a binary algorithm for exponentiation integrated into the Miller-Rabin 1377 loop, our standard modexp algorithms are faster). 1378 [Bodo Moeller] 1379 1380 *) Support for the EBCDIC character set completed. 1381 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 1382 1383 *) Source code cleanups: use const where appropriate, eliminate casts, 1384 use void * instead of char * in lhash. 1385 [Ulf M�ller] 1386 1387 *) Bugfix: ssl3_send_server_key_exchange was not restartable 1388 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 1389 this the server could overwrite ephemeral keys that the client 1390 has already seen). 1391 [Bodo Moeller] 1392 1393 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 1394 using 50 iterations of the Rabin-Miller test. 1395 1396 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 1397 iterations of the Rabin-Miller test as required by the appendix 1398 to FIPS PUB 186[-1]) instead of DSA_is_prime. 1399 As BN_is_prime_fasttest includes trial division, DSA parameter 1400 generation becomes much faster. 1401 1402 This implies a change for the callback functions in DSA_is_prime 1403 and DSA_generate_parameters: The callback function is called once 1404 for each positive witness in the Rabin-Miller test, not just 1405 occasionally in the inner loop; and the parameters to the 1406 callback function now provide an iteration count for the outer 1407 loop rather than for the current invocation of the inner loop. 1408 DSA_generate_parameters additionally can call the callback 1409 function with an 'iteration count' of -1, meaning that a 1410 candidate has passed the trial division test (when q is generated 1411 from an application-provided seed, trial division is skipped). 1412 [Bodo Moeller] 1413 1414 *) New function BN_is_prime_fasttest that optionally does trial 1415 division before starting the Rabin-Miller test and has 1416 an additional BN_CTX * argument (whereas BN_is_prime always 1417 has to allocate at least one BN_CTX). 1418 'callback(1, -1, cb_arg)' is called when a number has passed the 1419 trial division stage. 1420 [Bodo Moeller] 1421 1422 *) Fix for bug in CRL encoding. The validity dates weren't being handled 1423 as ASN1_TIME. 1424 [Steve Henson] 1425 1426 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 1427 [Steve Henson] 1428 1429 *) New function BN_pseudo_rand(). 1430 [Ulf M�ller] 1431 1432 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 1433 bignum version of BN_from_montgomery() with the working code from 1434 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 1435 the comments. 1436 [Ulf M�ller] 1437 1438 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 1439 made it impossible to use the same SSL_SESSION data structure in 1440 SSL2 clients in multiple threads. 1441 [Bodo Moeller] 1442 1443 *) The return value of RAND_load_file() no longer counts bytes obtained 1444 by stat(). RAND_load_file(..., -1) is new and uses the complete file 1445 to seed the PRNG (previously an explicit byte count was required). 1446 [Ulf M�ller, Bodo M�ller] 1447 1448 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 1449 used (char *) instead of (void *) and had casts all over the place. 1450 [Steve Henson] 1451 1452 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 1453 [Ulf M�ller] 1454 1455 *) Retain source code compatibility for BN_prime_checks macro: 1456 BN_is_prime(..., BN_prime_checks, ...) now uses 1457 BN_prime_checks_for_size to determine the appropriate number of 1458 Rabin-Miller iterations. 1459 [Ulf M�ller] 1460 1461 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 1462 DH_CHECK_P_NOT_SAFE_PRIME. 1463 (Check if this is true? OpenPGP calls them "strong".) 1464 [Ulf M�ller] 1465 1466 *) Merge the functionality of "dh" and "gendh" programs into a new program 1467 "dhparam". The old programs are retained for now but will handle DH keys 1468 (instead of parameters) in future. 1469 [Steve Henson] 1470 1471 *) Make the ciphers, s_server and s_client programs check the return values 1472 when a new cipher list is set. 1473 [Steve Henson] 1474 1475 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 1476 ciphers. Before when the 56bit ciphers were enabled the sorting was 1477 wrong. 1478 1479 The syntax for the cipher sorting has been extended to support sorting by 1480 cipher-strength (using the strength_bits hard coded in the tables). 1481 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 1482 1483 Fix a bug in the cipher-command parser: when supplying a cipher command 1484 string with an "undefined" symbol (neither command nor alphanumeric 1485 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 1486 an error is flagged. 1487 1488 Due to the strength-sorting extension, the code of the 1489 ssl_create_cipher_list() function was completely rearranged. I hope that 1490 the readability was also increased :-) 1491 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 1492 1493 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 1494 for the first serial number and places 2 in the serial number file. This 1495 avoids problems when the root CA is created with serial number zero and 1496 the first user certificate has the same issuer name and serial number 1497 as the root CA. 1498 [Steve Henson] 1499 1500 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 1501 the new code. Add documentation for this stuff. 1502 [Steve Henson] 1503 1504 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 1505 X509_*() to X509at_*() on the grounds that they don't handle X509 1506 structures and behave in an analagous way to the X509v3 functions: 1507 they shouldn't be called directly but wrapper functions should be used 1508 instead. 1509 1510 So we also now have some wrapper functions that call the X509at functions 1511 when passed certificate requests. (TO DO: similar things can be done with 1512 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 1513 things. Some of these need some d2i or i2d and print functionality 1514 because they handle more complex structures.) 1515 [Steve Henson] 1516 1517 *) Add missing #ifndefs that caused missing symbols when building libssl 1518 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 1519 NO_RSA in ssl/s2*.c. 1520 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 1521 1522 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 1523 has a return value which indicates the quality of the random data 1524 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 1525 error queue. New function RAND_pseudo_bytes() generates output that is 1526 guaranteed to be unique but not unpredictable. RAND_add is like 1527 RAND_seed, but takes an extra argument for an entropy estimate 1528 (RAND_seed always assumes full entropy). 1529 [Ulf M�ller] 1530 1531 *) Do more iterations of Rabin-Miller probable prime test (specifically, 1532 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 1533 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 1534 in crypto/bn/bn_prime.c for the complete table). This guarantees a 1535 false-positive rate of at most 2^-80 for random input. 1536 [Bodo Moeller] 1537 1538 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 1539 [Bodo Moeller] 1540 1541 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 1542 in the 0.9.5 release), this returns the chain 1543 from an X509_CTX structure with a dup of the stack and all 1544 the X509 reference counts upped: so the stack will exist 1545 after X509_CTX_cleanup() has been called. Modify pkcs12.c 1546 to use this. 1547 1548 Also make SSL_SESSION_print() print out the verify return 1549 code. 1550 [Steve Henson] 1551 1552 *) Add manpage for the pkcs12 command. Also change the default 1553 behaviour so MAC iteration counts are used unless the new 1554 -nomaciter option is used. This improves file security and 1555 only older versions of MSIE (4.0 for example) need it. 1556 [Steve Henson] 1557 1558 *) Honor the no-xxx Configure options when creating .DEF files. 1559 [Ulf M�ller] 1560 1561 *) Add PKCS#10 attributes to field table: challengePassword, 1562 unstructuredName and unstructuredAddress. These are taken from 1563 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 1564 international characters are used. 1565 1566 More changes to X509_ATTRIBUTE code: allow the setting of types 1567 based on strings. Remove the 'loc' parameter when adding 1568 attributes because these will be a SET OF encoding which is sorted 1569 in ASN1 order. 1570 [Steve Henson] 1571 1572 *) Initial changes to the 'req' utility to allow request generation 1573 automation. This will allow an application to just generate a template 1574 file containing all the field values and have req construct the 1575 request. 1576 1577 Initial support for X509_ATTRIBUTE handling. Stacks of these are 1578 used all over the place including certificate requests and PKCS#7 1579 structures. They are currently handled manually where necessary with 1580 some primitive wrappers for PKCS#7. The new functions behave in a 1581 manner analogous to the X509 extension functions: they allow 1582 attributes to be looked up by NID and added. 1583 1584 Later something similar to the X509V3 code would be desirable to 1585 automatically handle the encoding, decoding and printing of the 1586 more complex types. The string types like challengePassword can 1587 be handled by the string table functions. 1588 1589 Also modified the multi byte string table handling. Now there is 1590 a 'global mask' which masks out certain types. The table itself 1591 can use the flag STABLE_NO_MASK to ignore the mask setting: this 1592 is useful when for example there is only one permissible type 1593 (as in countryName) and using the mask might result in no valid 1594 types at all. 1595 [Steve Henson] 1596 1597 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 1598 SSL_get_peer_finished to allow applications to obtain the latest 1599 Finished messages sent to the peer or expected from the peer, 1600 respectively. (SSL_get_peer_finished is usually the Finished message 1601 actually received from the peer, otherwise the protocol will be aborted.) 1602 1603 As the Finished message are message digests of the complete handshake 1604 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 1605 be used for external authentication procedures when the authentication 1606 provided by SSL/TLS is not desired or is not enough. 1607 [Bodo Moeller] 1608 1609 *) Enhanced support for Alpha Linux is added. Now ./config checks if 1610 the host supports BWX extension and if Compaq C is present on the 1611 $PATH. Just exploiting of the BWX extension results in 20-30% 1612 performance kick for some algorithms, e.g. DES and RC4 to mention 1613 a couple. Compaq C in turn generates ~20% faster code for MD5 and 1614 SHA1. 1615 [Andy Polyakov] 1616 1617 *) Add support for MS "fast SGC". This is arguably a violation of the 1618 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 1619 weak crypto and after checking the certificate is SGC a second one 1620 with strong crypto. MS SGC stops the first handshake after receiving 1621 the server certificate message and sends a second client hello. Since 1622 a server will typically do all the time consuming operations before 1623 expecting any further messages from the client (server key exchange 1624 is the most expensive) there is little difference between the two. 1625 1626 To get OpenSSL to support MS SGC we have to permit a second client 1627 hello message after we have sent server done. In addition we have to 1628 reset the MAC if we do get this second client hello. 1629 [Steve Henson] 1630 1631 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 1632 if a DER encoded private key is RSA or DSA traditional format. Changed 1633 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 1634 format DER encoded private key. Newer code should use PKCS#8 format which 1635 has the key type encoded in the ASN1 structure. Added DER private key 1636 support to pkcs8 application. 1637 [Steve Henson] 1638 1639 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 1640 ciphersuites has been selected (as required by the SSL 3/TLS 1 1641 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 1642 is set, we interpret this as a request to violate the specification 1643 (the worst that can happen is a handshake failure, and 'correct' 1644 behaviour would result in a handshake failure anyway). 1645 [Bodo Moeller] 1646 1647 *) In SSL_CTX_add_session, take into account that there might be multiple 1648 SSL_SESSION structures with the same session ID (e.g. when two threads 1649 concurrently obtain them from an external cache). 1650 The internal cache can handle only one SSL_SESSION with a given ID, 1651 so if there's a conflict, we now throw out the old one to achieve 1652 consistency. 1653 [Bodo Moeller] 1654 1655 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 1656 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 1657 some routines that use cipher OIDs: some ciphers do not have OIDs 1658 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 1659 example. 1660 [Steve Henson] 1661 1662 *) Simplify the trust setting structure and code. Now we just have 1663 two sequences of OIDs for trusted and rejected settings. These will 1664 typically have values the same as the extended key usage extension 1665 and any application specific purposes. 1666 1667 The trust checking code now has a default behaviour: it will just 1668 check for an object with the same NID as the passed id. Functions can 1669 be provided to override either the default behaviour or the behaviour 1670 for a given id. SSL client, server and email already have functions 1671 in place for compatibility: they check the NID and also return "trusted" 1672 if the certificate is self signed. 1673 [Steve Henson] 1674 1675 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 1676 traditional format into an EVP_PKEY structure. 1677 [Steve Henson] 1678 1679 *) Add a password callback function PEM_cb() which either prompts for 1680 a password if usr_data is NULL or otherwise assumes it is a null 1681 terminated password. Allow passwords to be passed on command line 1682 environment or config files in a few more utilities. 1683 [Steve Henson] 1684 1685 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 1686 keys. Add some short names for PKCS#8 PBE algorithms and allow them 1687 to be specified on the command line for the pkcs8 and pkcs12 utilities. 1688 Update documentation. 1689 [Steve Henson] 1690 1691 *) Support for ASN1 "NULL" type. This could be handled before by using 1692 ASN1_TYPE but there wasn't any function that would try to read a NULL 1693 and produce an error if it couldn't. For compatibility we also have 1694 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 1695 don't allocate anything because they don't need to. 1696 [Steve Henson] 1697 1698 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 1699 for details. 1700 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 1701 1702 *) Rebuild of the memory allocation routines used by OpenSSL code and 1703 possibly others as well. The purpose is to make an interface that 1704 provide hooks so anyone can build a separate set of allocation and 1705 deallocation routines to be used by OpenSSL, for example memory 1706 pool implementations, or something else, which was previously hard 1707 since Malloc(), Realloc() and Free() were defined as macros having 1708 the values malloc, realloc and free, respectively (except for Win32 1709 compilations). The same is provided for memory debugging code. 1710 OpenSSL already comes with functionality to find memory leaks, but 1711 this gives people a chance to debug other memory problems. 1712 1713 With these changes, a new set of functions and macros have appeared: 1714 1715 CRYPTO_set_mem_debug_functions() [F] 1716 CRYPTO_get_mem_debug_functions() [F] 1717 CRYPTO_dbg_set_options() [F] 1718 CRYPTO_dbg_get_options() [F] 1719 CRYPTO_malloc_debug_init() [M] 1720 1721 The memory debug functions are NULL by default, unless the library 1722 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 1723 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 1724 gives the standard debugging functions that come with OpenSSL) or 1725 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 1726 provided by the library user) must be used. When the standard 1727 debugging functions are used, CRYPTO_dbg_set_options can be used to 1728 request additional information: 1729 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 1730 the CRYPTO_MDEBUG_xxx macro when compiling the library. 1731 1732 Also, things like CRYPTO_set_mem_functions will always give the 1733 expected result (the new set of functions is used for allocation 1734 and deallocation) at all times, regardless of platform and compiler 1735 options. 1736 1737 To finish it up, some functions that were never use in any other 1738 way than through macros have a new API and new semantic: 1739 1740 CRYPTO_dbg_malloc() 1741 CRYPTO_dbg_realloc() 1742 CRYPTO_dbg_free() 1743 1744 All macros of value have retained their old syntax. 1745 [Richard Levitte and Bodo Moeller] 1746 1747 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 1748 ordering of SMIMECapabilities wasn't in "strength order" and there 1749 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 1750 algorithm. 1751 [Steve Henson] 1752 1753 *) Some ASN1 types with illegal zero length encoding (INTEGER, 1754 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 1755 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 1756 1757 *) Merge in my S/MIME library for OpenSSL. This provides a simple 1758 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 1759 functionality to handle multipart/signed properly) and a utility 1760 called 'smime' to call all this stuff. This is based on code I 1761 originally wrote for Celo who have kindly allowed it to be 1762 included in OpenSSL. 1763 [Steve Henson] 1764 1765 *) Add variants des_set_key_checked and des_set_key_unchecked of 1766 des_set_key (aka des_key_sched). Global variable des_check_key 1767 decides which of these is called by des_set_key; this way 1768 des_check_key behaves as it always did, but applications and 1769 the library itself, which was buggy for des_check_key == 1, 1770 have a cleaner way to pick the version they need. 1771 [Bodo Moeller] 1772 1773 *) New function PKCS12_newpass() which changes the password of a 1774 PKCS12 structure. 1775 [Steve Henson] 1776 1777 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 1778 dynamic mix. In both cases the ids can be used as an index into the 1779 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 1780 functions so they accept a list of the field values and the 1781 application doesn't need to directly manipulate the X509_TRUST 1782 structure. 1783 [Steve Henson] 1784 1785 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 1786 need initialising. 1787 [Steve Henson] 1788 1789 *) Modify the way the V3 extension code looks up extensions. This now 1790 works in a similar way to the object code: we have some "standard" 1791 extensions in a static table which is searched with OBJ_bsearch() 1792 and the application can add dynamic ones if needed. The file 1793 crypto/x509v3/ext_dat.h now has the info: this file needs to be 1794 updated whenever a new extension is added to the core code and kept 1795 in ext_nid order. There is a simple program 'tabtest.c' which checks 1796 this. New extensions are not added too often so this file can readily 1797 be maintained manually. 1798 1799 There are two big advantages in doing things this way. The extensions 1800 can be looked up immediately and no longer need to be "added" using 1801 X509V3_add_standard_extensions(): this function now does nothing. 1802 [Side note: I get *lots* of email saying the extension code doesn't 1803 work because people forget to call this function] 1804 Also no dynamic allocation is done unless new extensions are added: 1805 so if we don't add custom extensions there is no need to call 1806 X509V3_EXT_cleanup(). 1807 [Steve Henson] 1808 1809 *) Modify enc utility's salting as follows: make salting the default. Add a 1810 magic header, so unsalted files fail gracefully instead of just decrypting 1811 to garbage. This is because not salting is a big security hole, so people 1812 should be discouraged from doing it. 1813 [Ben Laurie] 1814 1815 *) Fixes and enhancements to the 'x509' utility. It allowed a message 1816 digest to be passed on the command line but it only used this 1817 parameter when signing a certificate. Modified so all relevant 1818 operations are affected by the digest parameter including the 1819 -fingerprint and -x509toreq options. Also -x509toreq choked if a 1820 DSA key was used because it didn't fix the digest. 1821 [Steve Henson] 1822 1823 *) Initial certificate chain verify code. Currently tests the untrusted 1824 certificates for consistency with the verify purpose (which is set 1825 when the X509_STORE_CTX structure is set up) and checks the pathlength. 1826 1827 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 1828 this is because it will reject chains with invalid extensions whereas 1829 every previous version of OpenSSL and SSLeay made no checks at all. 1830 1831 Trust code: checks the root CA for the relevant trust settings. Trust 1832 settings have an initial value consistent with the verify purpose: e.g. 1833 if the verify purpose is for SSL client use it expects the CA to be 1834 trusted for SSL client use. However the default value can be changed to 1835 permit custom trust settings: one example of this would be to only trust 1836 certificates from a specific "secure" set of CAs. 1837 1838 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 1839 which should be used for version portability: especially since the 1840 verify structure is likely to change more often now. 1841 1842 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 1843 to set them. If not set then assume SSL clients will verify SSL servers 1844 and vice versa. 1845 1846 Two new options to the verify program: -untrusted allows a set of 1847 untrusted certificates to be passed in and -purpose which sets the 1848 intended purpose of the certificate. If a purpose is set then the 1849 new chain verify code is used to check extension consistency. 1850 [Steve Henson] 1851 1852 *) Support for the authority information access extension. 1853 [Steve Henson] 1854 1855 *) Modify RSA and DSA PEM read routines to transparently handle 1856 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 1857 public keys in a format compatible with certificate 1858 SubjectPublicKeyInfo structures. Unfortunately there were already 1859 functions called *_PublicKey_* which used various odd formats so 1860 these are retained for compatibility: however the DSA variants were 1861 never in a public release so they have been deleted. Changed dsa/rsa 1862 utilities to handle the new format: note no releases ever handled public 1863 keys so we should be OK. 1864 1865 The primary motivation for this change is to avoid the same fiasco 1866 that dogs private keys: there are several incompatible private key 1867 formats some of which are standard and some OpenSSL specific and 1868 require various evil hacks to allow partial transparent handling and 1869 even then it doesn't work with DER formats. Given the option anything 1870 other than PKCS#8 should be dumped: but the other formats have to 1871 stay in the name of compatibility. 1872 1873 With public keys and the benefit of hindsight one standard format 1874 is used which works with EVP_PKEY, RSA or DSA structures: though 1875 it clearly returns an error if you try to read the wrong kind of key. 1876 1877 Added a -pubkey option to the 'x509' utility to output the public key. 1878 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 1879 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 1880 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 1881 that do the same as the EVP_PKEY_assign_*() except they up the 1882 reference count of the added key (they don't "swallow" the 1883 supplied key). 1884 [Steve Henson] 1885 1886 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 1887 CRLs would fail if the file contained no certificates or no CRLs: 1888 added a new function to read in both types and return the number 1889 read: this means that if none are read it will be an error. The 1890 DER versions of the certificate and CRL reader would always fail 1891 because it isn't possible to mix certificates and CRLs in DER format 1892 without choking one or the other routine. Changed this to just read 1893 a certificate: this is the best we can do. Also modified the code 1894 in apps/verify.c to take notice of return codes: it was previously 1895 attempting to read in certificates from NULL pointers and ignoring 1896 any errors: this is one reason why the cert and CRL reader seemed 1897 to work. It doesn't check return codes from the default certificate 1898 routines: these may well fail if the certificates aren't installed. 1899 [Steve Henson] 1900 1901 *) Code to support otherName option in GeneralName. 1902 [Steve Henson] 1903 1904 *) First update to verify code. Change the verify utility 1905 so it warns if it is passed a self signed certificate: 1906 for consistency with the normal behaviour. X509_verify 1907 has been modified to it will now verify a self signed 1908 certificate if *exactly* the same certificate appears 1909 in the store: it was previously impossible to trust a 1910 single self signed certificate. This means that: 1911 openssl verify ss.pem 1912 now gives a warning about a self signed certificate but 1913 openssl verify -CAfile ss.pem ss.pem 1914 is OK. 1915 [Steve Henson] 1916 1917 *) For servers, store verify_result in SSL_SESSION data structure 1918 (and add it to external session representation). 1919 This is needed when client certificate verifications fails, 1920 but an application-provided verification callback (set by 1921 SSL_CTX_set_cert_verify_callback) allows accepting the session 1922 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 1923 but returns 1): When the session is reused, we have to set 1924 ssl->verify_result to the appropriate error code to avoid 1925 security holes. 1926 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 1927 1928 *) Fix a bug in the new PKCS#7 code: it didn't consider the 1929 case in PKCS7_dataInit() where the signed PKCS7 structure 1930 didn't contain any existing data because it was being created. 1931 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 1932 1933 *) Add a salt to the key derivation routines in enc.c. This 1934 forms the first 8 bytes of the encrypted file. Also add a 1935 -S option to allow a salt to be input on the command line. 1936 [Steve Henson] 1937 1938 *) New function X509_cmp(). Oddly enough there wasn't a function 1939 to compare two certificates. We do this by working out the SHA1 1940 hash and comparing that. X509_cmp() will be needed by the trust 1941 code. 1942 [Steve Henson] 1943 1944 *) SSL_get1_session() is like SSL_get_session(), but increments 1945 the reference count in the SSL_SESSION returned. 1946 [Geoff Thorpe <geoff@eu.c2.net>] 1947 1948 *) Fix for 'req': it was adding a null to request attributes. 1949 Also change the X509_LOOKUP and X509_INFO code to handle 1950 certificate auxiliary information. 1951 [Steve Henson] 1952 1953 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 1954 the 'enc' command. 1955 [Steve Henson] 1956 1957 *) Add the possibility to add extra information to the memory leak 1958 detecting output, to form tracebacks, showing from where each 1959 allocation was originated: CRYPTO_push_info("constant string") adds 1960 the string plus current file name and line number to a per-thread 1961 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 1962 is like calling CYRPTO_pop_info() until the stack is empty. 1963 Also updated memory leak detection code to be multi-thread-safe. 1964 [Richard Levitte] 1965 1966 *) Add options -text and -noout to pkcs7 utility and delete the 1967 encryption options which never did anything. Update docs. 1968 [Steve Henson] 1969 1970 *) Add options to some of the utilities to allow the pass phrase 1971 to be included on either the command line (not recommended on 1972 OSes like Unix) or read from the environment. Update the 1973 manpages and fix a few bugs. 1974 [Steve Henson] 1975 1976 *) Add a few manpages for some of the openssl commands. 1977 [Steve Henson] 1978 1979 *) Fix the -revoke option in ca. It was freeing up memory twice, 1980 leaking and not finding already revoked certificates. 1981 [Steve Henson] 1982 1983 *) Extensive changes to support certificate auxiliary information. 1984 This involves the use of X509_CERT_AUX structure and X509_AUX 1985 functions. An X509_AUX function such as PEM_read_X509_AUX() 1986 can still read in a certificate file in the usual way but it 1987 will also read in any additional "auxiliary information". By 1988 doing things this way a fair degree of compatibility can be 1989 retained: existing certificates can have this information added 1990 using the new 'x509' options. 1991 1992 Current auxiliary information includes an "alias" and some trust 1993 settings. The trust settings will ultimately be used in enhanced 1994 certificate chain verification routines: currently a certificate 1995 can only be trusted if it is self signed and then it is trusted 1996 for all purposes. 1997 [Steve Henson] 1998 1999 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 2000 The problem was that one of the replacement routines had not been working 2001 since SSLeay releases. For now the offending routine has been replaced 2002 with non-optimised assembler. Even so, this now gives around 95% 2003 performance improvement for 1024 bit RSA signs. 2004 [Mark Cox] 2005 2006 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 2007 handling. Most clients have the effective key size in bits equal to 2008 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 2009 A few however don't do this and instead use the size of the decrypted key 2010 to determine the RC2 key length and the AlgorithmIdentifier to determine 2011 the effective key length. In this case the effective key length can still 2012 be 40 bits but the key length can be 168 bits for example. This is fixed 2013 by manually forcing an RC2 key into the EVP_PKEY structure because the 2014 EVP code can't currently handle unusual RC2 key sizes: it always assumes 2015 the key length and effective key length are equal. 2016 [Steve Henson] 2017 2018 *) Add a bunch of functions that should simplify the creation of 2019 X509_NAME structures. Now you should be able to do: 2020 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 2021 and have it automatically work out the correct field type and fill in 2022 the structures. The more adventurous can try: 2023 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 2024 and it will (hopefully) work out the correct multibyte encoding. 2025 [Steve Henson] 2026 2027 *) Change the 'req' utility to use the new field handling and multibyte 2028 copy routines. Before the DN field creation was handled in an ad hoc 2029 way in req, ca, and x509 which was rather broken and didn't support 2030 BMPStrings or UTF8Strings. Since some software doesn't implement 2031 BMPStrings or UTF8Strings yet, they can be enabled using the config file 2032 using the dirstring_type option. See the new comment in the default 2033 openssl.cnf for more info. 2034 [Steve Henson] 2035 2036 *) Make crypto/rand/md_rand.c more robust: 2037 - Assure unique random numbers after fork(). 2038 - Make sure that concurrent threads access the global counter and 2039 md serializably so that we never lose entropy in them 2040 or use exactly the same state in multiple threads. 2041 Access to the large state is not always serializable because 2042 the additional locking could be a performance killer, and 2043 md should be large enough anyway. 2044 [Bodo Moeller] 2045 2046 *) New file apps/app_rand.c with commonly needed functionality 2047 for handling the random seed file. 2048 2049 Use the random seed file in some applications that previously did not: 2050 ca, 2051 dsaparam -genkey (which also ignored its '-rand' option), 2052 s_client, 2053 s_server, 2054 x509 (when signing). 2055 Except on systems with /dev/urandom, it is crucial to have a random 2056 seed file at least for key creation, DSA signing, and for DH exchanges; 2057 for RSA signatures we could do without one. 2058 2059 gendh and gendsa (unlike genrsa) used to read only the first byte 2060 of each file listed in the '-rand' option. The function as previously 2061 found in genrsa is now in app_rand.c and is used by all programs 2062 that support '-rand'. 2063 [Bodo Moeller] 2064 2065 *) In RAND_write_file, use mode 0600 for creating files; 2066 don't just chmod when it may be too late. 2067 [Bodo Moeller] 2068 2069 *) Report an error from X509_STORE_load_locations 2070 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 2071 [Bill Perry] 2072 2073 *) New function ASN1_mbstring_copy() this copies a string in either 2074 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 2075 into an ASN1_STRING type. A mask of permissible types is passed 2076 and it chooses the "minimal" type to use or an error if not type 2077 is suitable. 2078 [Steve Henson] 2079 2080 *) Add function equivalents to the various macros in asn1.h. The old 2081 macros are retained with an M_ prefix. Code inside the library can 2082 use the M_ macros. External code (including the openssl utility) 2083 should *NOT* in order to be "shared library friendly". 2084 [Steve Henson] 2085 2086 *) Add various functions that can check a certificate's extensions 2087 to see if it usable for various purposes such as SSL client, 2088 server or S/MIME and CAs of these types. This is currently 2089 VERY EXPERIMENTAL but will ultimately be used for certificate chain 2090 verification. Also added a -purpose flag to x509 utility to 2091 print out all the purposes. 2092 [Steve Henson] 2093 2094 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 2095 functions. 2096 [Steve Henson] 2097 2098 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 2099 for, obtain and decode and extension and obtain its critical flag. 2100 This allows all the necessary extension code to be handled in a 2101 single function call. 2102 [Steve Henson] 2103 2104 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 2105 platforms. See crypto/rc4/rc4_enc.c for further details. 2106 [Andy Polyakov] 2107 2108 *) New -noout option to asn1parse. This causes no output to be produced 2109 its main use is when combined with -strparse and -out to extract data 2110 from a file (which may not be in ASN.1 format). 2111 [Steve Henson] 2112 2113 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 2114 when producing the local key id. 2115 [Richard Levitte <levitte@stacken.kth.se>] 2116 2117 *) New option -dhparam in s_server. This allows a DH parameter file to be 2118 stated explicitly. If it is not stated then it tries the first server 2119 certificate file. The previous behaviour hard coded the filename 2120 "server.pem". 2121 [Steve Henson] 2122 2123 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 2124 a public key to be input or output. For example: 2125 openssl rsa -in key.pem -pubout -out pubkey.pem 2126 Also added necessary DSA public key functions to handle this. 2127 [Steve Henson] 2128 2129 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 2130 in the message. This was handled by allowing 2131 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 2132 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 2133 2134 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 2135 to the end of the strings whereas this didn't. This would cause problems 2136 if strings read with d2i_ASN1_bytes() were later modified. 2137 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 2138 2139 *) Fix for base64 decode bug. When a base64 bio reads only one line of 2140 data and it contains EOF it will end up returning an error. This is 2141 caused by input 46 bytes long. The cause is due to the way base64 2142 BIOs find the start of base64 encoded data. They do this by trying a 2143 trial decode on each line until they find one that works. When they 2144 do a flag is set and it starts again knowing it can pass all the 2145 data directly through the decoder. Unfortunately it doesn't reset 2146 the context it uses. This means that if EOF is reached an attempt 2147 is made to pass two EOFs through the context and this causes the 2148 resulting error. This can also cause other problems as well. As is 2149 usual with these problems it takes *ages* to find and the fix is 2150 trivial: move one line. 2151 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 2152 2153 *) Ugly workaround to get s_client and s_server working under Windows. The 2154 old code wouldn't work because it needed to select() on sockets and the 2155 tty (for keypresses and to see if data could be written). Win32 only 2156 supports select() on sockets so we select() with a 1s timeout on the 2157 sockets and then see if any characters are waiting to be read, if none 2158 are present then we retry, we also assume we can always write data to 2159 the tty. This isn't nice because the code then blocks until we've 2160 received a complete line of data and it is effectively polling the 2161 keyboard at 1s intervals: however it's quite a bit better than not 2162 working at all :-) A dedicated Windows application might handle this 2163 with an event loop for example. 2164 [Steve Henson] 2165 2166 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 2167 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 2168 will be called when RSA_sign() and RSA_verify() are used. This is useful 2169 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 2170 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 2171 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 2172 This necessitated the support of an extra signature type NID_md5_sha1 2173 for SSL signatures and modifications to the SSL library to use it instead 2174 of calling RSA_public_decrypt() and RSA_private_encrypt(). 2175 [Steve Henson] 2176 2177 *) Add new -verify -CAfile and -CApath options to the crl program, these 2178 will lookup a CRL issuers certificate and verify the signature in a 2179 similar way to the verify program. Tidy up the crl program so it 2180 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 2181 less strict. It will now permit CRL extensions even if it is not 2182 a V2 CRL: this will allow it to tolerate some broken CRLs. 2183 [Steve Henson] 2184 2185 *) Initialize all non-automatic variables each time one of the openssl 2186 sub-programs is started (this is necessary as they may be started 2187 multiple times from the "OpenSSL>" prompt). 2188 [Lennart Bang, Bodo Moeller] 2189 2190 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 2191 removing all other RSA functionality (this is what NO_RSA does). This 2192 is so (for example) those in the US can disable those operations covered 2193 by the RSA patent while allowing storage and parsing of RSA keys and RSA 2194 key generation. 2195 [Steve Henson] 2196 2197 *) Non-copying interface to BIO pairs. 2198 (still largely untested) 2199 [Bodo Moeller] 2200 2201 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 2202 ASCII string. This was handled independently in various places before. 2203 [Steve Henson] 2204 2205 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 2206 UTF8 strings a character at a time. 2207 [Steve Henson] 2208 2209 *) Use client_version from client hello to select the protocol 2210 (s23_srvr.c) and for RSA client key exchange verification 2211 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 2212 [Bodo Moeller] 2213 2214 *) Add various utility functions to handle SPKACs, these were previously 2215 handled by poking round in the structure internals. Added new function 2216 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 2217 print, verify and generate SPKACs. Based on an original idea from 2218 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 2219 [Steve Henson] 2220 2221 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 2222 [Andy Polyakov] 2223 2224 *) Allow the config file extension section to be overwritten on the 2225 command line. Based on an original idea from Massimiliano Pala 2226 <madwolf@comune.modena.it>. The new option is called -extensions 2227 and can be applied to ca, req and x509. Also -reqexts to override 2228 the request extensions in req and -crlexts to override the crl extensions 2229 in ca. 2230 [Steve Henson] 2231 2232 *) Add new feature to the SPKAC handling in ca. Now you can include 2233 the same field multiple times by preceding it by "XXXX." for example: 2234 1.OU="Unit name 1" 2235 2.OU="Unit name 2" 2236 this is the same syntax as used in the req config file. 2237 [Steve Henson] 2238 2239 *) Allow certificate extensions to be added to certificate requests. These 2240 are specified in a 'req_extensions' option of the req section of the 2241 config file. They can be printed out with the -text option to req but 2242 are otherwise ignored at present. 2243 [Steve Henson] 2244 2245 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 2246 data read consists of only the final block it would not decrypted because 2247 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 2248 A misplaced 'break' also meant the decrypted final block might not be 2249 copied until the next read. 2250 [Steve Henson] 2251 2252 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 2253 a few extra parameters to the DH structure: these will be useful if 2254 for example we want the value of 'q' or implement X9.42 DH. 2255 [Steve Henson] 2256 2257 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 2258 provides hooks that allow the default DSA functions or functions on a 2259 "per key" basis to be replaced. This allows hardware acceleration and 2260 hardware key storage to be handled without major modification to the 2261 library. Also added low level modexp hooks and CRYPTO_EX structure and 2262 associated functions. 2263 [Steve Henson] 2264 2265 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 2266 as "read only": it can't be written to and the buffer it points to will 2267 not be freed. Reading from a read only BIO is much more efficient than 2268 a normal memory BIO. This was added because there are several times when 2269 an area of memory needs to be read from a BIO. The previous method was 2270 to create a memory BIO and write the data to it, this results in two 2271 copies of the data and an O(n^2) reading algorithm. There is a new 2272 function BIO_new_mem_buf() which creates a read only memory BIO from 2273 an area of memory. Also modified the PKCS#7 routines to use read only 2274 memory BIOs. 2275 [Steve Henson] 2276 2277 *) Bugfix: ssl23_get_client_hello did not work properly when called in 2278 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 2279 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 2280 but a retry condition occured while trying to read the rest. 2281 [Bodo Moeller] 2282 2283 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 2284 NID_pkcs7_encrypted by default: this was wrong since this should almost 2285 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 2286 the encrypted data type: this is a more sensible place to put it and it 2287 allows the PKCS#12 code to be tidied up that duplicated this 2288 functionality. 2289 [Steve Henson] 2290 2291 *) Changed obj_dat.pl script so it takes its input and output files on 2292 the command line. This should avoid shell escape redirection problems 2293 under Win32. 2294 [Steve Henson] 2295 2296 *) Initial support for certificate extension requests, these are included 2297 in things like Xenroll certificate requests. Included functions to allow 2298 extensions to be obtained and added. 2299 [Steve Henson] 2300 2301 *) -crlf option to s_client and s_server for sending newlines as 2302 CRLF (as required by many protocols). 2303 [Bodo Moeller] 2304 2305 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 2306 2307 *) Install libRSAglue.a when OpenSSL is built with RSAref. 2308 [Ralf S. Engelschall] 2309 2310 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 2311 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 2312 2313 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 2314 program. 2315 [Steve Henson] 2316 2317 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 2318 DH parameters/keys (q is lost during that conversion, but the resulting 2319 DH parameters contain its length). 2320 2321 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 2322 much faster than DH_generate_parameters (which creates parameters 2323 where p = 2*q + 1), and also the smaller q makes DH computations 2324 much more efficient (160-bit exponentiation instead of 1024-bit 2325 exponentiation); so this provides a convenient way to support DHE 2326 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 2327 utter importance to use 2328 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 2329 or 2330 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 2331 when such DH parameters are used, because otherwise small subgroup 2332 attacks may become possible! 2333 [Bodo Moeller] 2334 2335 *) Avoid memory leak in i2d_DHparams. 2336 [Bodo Moeller] 2337 2338 *) Allow the -k option to be used more than once in the enc program: 2339 this allows the same encrypted message to be read by multiple recipients. 2340 [Steve Henson] 2341 2342 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 2343 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 2344 it will always use the numerical form of the OID, even if it has a short 2345 or long name. 2346 [Steve Henson] 2347 2348 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 2349 method only got called if p,q,dmp1,dmq1,iqmp components were present, 2350 otherwise bn_mod_exp was called. In the case of hardware keys for example 2351 no private key components need be present and it might store extra data 2352 in the RSA structure, which cannot be accessed from bn_mod_exp. 2353 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 2354 private key operations. 2355 [Steve Henson] 2356 2357 *) Added support for SPARC Linux. 2358 [Andy Polyakov] 2359 2360 *) pem_password_cb function type incompatibly changed from 2361 typedef int pem_password_cb(char *buf, int size, int rwflag); 2362 to 2363 ....(char *buf, int size, int rwflag, void *userdata); 2364 so that applications can pass data to their callbacks: 2365 The PEM[_ASN1]_{read,write}... functions and macros now take an 2366 additional void * argument, which is just handed through whenever 2367 the password callback is called. 2368 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 2369 2370 New function SSL_CTX_set_default_passwd_cb_userdata. 2371 2372 Compatibility note: As many C implementations push function arguments 2373 onto the stack in reverse order, the new library version is likely to 2374 interoperate with programs that have been compiled with the old 2375 pem_password_cb definition (PEM_whatever takes some data that 2376 happens to be on the stack as its last argument, and the callback 2377 just ignores this garbage); but there is no guarantee whatsoever that 2378 this will work. 2379 2380 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 2381 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 2382 problems not only on Windows, but also on some Unix platforms. 2383 To avoid problematic command lines, these definitions are now in an 2384 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 2385 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 2386 [Bodo Moeller] 2387 2388 *) MIPS III/IV assembler module is reimplemented. 2389 [Andy Polyakov] 2390 2391 *) More DES library cleanups: remove references to srand/rand and 2392 delete an unused file. 2393 [Ulf M�ller] 2394 2395 *) Add support for the the free Netwide assembler (NASM) under Win32, 2396 since not many people have MASM (ml) and it can be hard to obtain. 2397 This is currently experimental but it seems to work OK and pass all 2398 the tests. Check out INSTALL.W32 for info. 2399 [Steve Henson] 2400 2401 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 2402 without temporary keys kept an extra copy of the server key, 2403 and connections with temporary keys did not free everything in case 2404 of an error. 2405 [Bodo Moeller] 2406 2407 *) New function RSA_check_key and new openssl rsa option -check 2408 for verifying the consistency of RSA keys. 2409 [Ulf Moeller, Bodo Moeller] 2410 2411 *) Various changes to make Win32 compile work: 2412 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 2413 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 2414 comparison" warnings. 2415 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 2416 [Steve Henson] 2417 2418 *) Add a debugging option to PKCS#5 v2 key generation function: when 2419 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 2420 derived keys are printed to stderr. 2421 [Steve Henson] 2422 2423 *) Copy the flags in ASN1_STRING_dup(). 2424 [Roman E. Pavlov <pre@mo.msk.ru>] 2425 2426 *) The x509 application mishandled signing requests containing DSA 2427 keys when the signing key was also DSA and the parameters didn't match. 2428 2429 It was supposed to omit the parameters when they matched the signing key: 2430 the verifying software was then supposed to automatically use the CA's 2431 parameters if they were absent from the end user certificate. 2432 2433 Omitting parameters is no longer recommended. The test was also 2434 the wrong way round! This was probably due to unusual behaviour in 2435 EVP_cmp_parameters() which returns 1 if the parameters match. 2436 This meant that parameters were omitted when they *didn't* match and 2437 the certificate was useless. Certificates signed with 'ca' didn't have 2438 this bug. 2439 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 2440 2441 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 2442 The interface is as follows: 2443 Applications can use 2444 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 2445 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 2446 "off" is now the default. 2447 The library internally uses 2448 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 2449 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 2450 to disable memory-checking temporarily. 2451 2452 Some inconsistent states that previously were possible (and were 2453 even the default) are now avoided. 2454 2455 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 2456 with each memory chunk allocated; this is occasionally more helpful 2457 than just having a counter. 2458 2459 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 2460 2461 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 2462 extensions. 2463 [Bodo Moeller] 2464 2465 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 2466 which largely parallels "options", but is for changing API behaviour, 2467 whereas "options" are about protocol behaviour. 2468 Initial "mode" flags are: 2469 2470 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 2471 a single record has been written. 2472 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 2473 retries use the same buffer location. 2474 (But all of the contents must be 2475 copied!) 2476 [Bodo Moeller] 2477 2478 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 2479 worked. 2480 2481 *) Fix problems with no-hmac etc. 2482 [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 2483 2484 *) New functions RSA_get_default_method(), RSA_set_method() and 2485 RSA_get_method(). These allows replacement of RSA_METHODs without having 2486 to mess around with the internals of an RSA structure. 2487 [Steve Henson] 2488 2489 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 2490 Also really enable memory leak checks in openssl.c and in some 2491 test programs. 2492 [Chad C. Mulligan, Bodo Moeller] 2493 2494 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 2495 up the length of negative integers. This has now been simplified to just 2496 store the length when it is first determined and use it later, rather 2497 than trying to keep track of where data is copied and updating it to 2498 point to the end. 2499 [Steve Henson, reported by Brien Wheeler 2500 <bwheeler@authentica-security.com>] 2501 2502 *) Add a new function PKCS7_signatureVerify. This allows the verification 2503 of a PKCS#7 signature but with the signing certificate passed to the 2504 function itself. This contrasts with PKCS7_dataVerify which assumes the 2505 certificate is present in the PKCS#7 structure. This isn't always the 2506 case: certificates can be omitted from a PKCS#7 structure and be 2507 distributed by "out of band" means (such as a certificate database). 2508 [Steve Henson] 2509 2510 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 2511 function prototypes in pem.h, also change util/mkdef.pl to add the 2512 necessary function names. 2513 [Steve Henson] 2514 2515 *) mk1mf.pl (used by Windows builds) did not properly read the 2516 options set by Configure in the top level Makefile, and Configure 2517 was not even able to write more than one option correctly. 2518 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 2519 [Bodo Moeller] 2520 2521 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 2522 file to be loaded from a BIO or FILE pointer. The BIO version will 2523 for example allow memory BIOs to contain config info. 2524 [Steve Henson] 2525 2526 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 2527 Whoever hopes to achieve shared-library compatibility across versions 2528 must use this, not the compile-time macro. 2529 (Exercise 0.9.4: Which is the minimum library version required by 2530 such programs?) 2531 Note: All this applies only to multi-threaded programs, others don't 2532 need locks. 2533 [Bodo Moeller] 2534 2535 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 2536 through a BIO pair triggered the default case, i.e. 2537 SSLerr(...,SSL_R_UNKNOWN_STATE). 2538 [Bodo Moeller] 2539 2540 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 2541 can use the SSL library even if none of the specific BIOs is 2542 appropriate. 2543 [Bodo Moeller] 2544 2545 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 2546 for the encoded length. 2547 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 2548 2549 *) Add initial documentation of the X509V3 functions. 2550 [Steve Henson] 2551 2552 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 2553 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 2554 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 2555 secure PKCS#8 private key format with a high iteration count. 2556 [Steve Henson] 2557 2558 *) Fix determination of Perl interpreter: A perl or perl5 2559 _directory_ in $PATH was also accepted as the interpreter. 2560 [Ralf S. Engelschall] 2561 2562 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 2563 wrong with it but it was very old and did things like calling 2564 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 2565 unusual formatting. 2566 [Steve Henson] 2567 2568 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 2569 to use the new extension code. 2570 [Steve Henson] 2571 2572 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 2573 with macros. This should make it easier to change their form, add extra 2574 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 2575 constant. 2576 [Steve Henson] 2577 2578 *) Add to configuration table a new entry that can specify an alternative 2579 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 2580 according to Mark Crispin <MRC@Panda.COM>. 2581 [Bodo Moeller] 2582 2583#if 0 2584 *) DES CBC did not update the IV. Weird. 2585 [Ben Laurie] 2586#else 2587 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 2588 Changing the behaviour of the former might break existing programs -- 2589 where IV updating is needed, des_ncbc_encrypt can be used. 2590#endif 2591 2592 *) When bntest is run from "make test" it drives bc to check its 2593 calculations, as well as internally checking them. If an internal check 2594 fails, it needs to cause bc to give a non-zero result or make test carries 2595 on without noticing the failure. Fixed. 2596 [Ben Laurie] 2597 2598 *) DES library cleanups. 2599 [Ulf M�ller] 2600 2601 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 2602 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 2603 ciphers. NOTE: although the key derivation function has been verified 2604 against some published test vectors it has not been extensively tested 2605 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 2606 of v2.0. 2607 [Steve Henson] 2608 2609 *) Instead of "mkdir -p", which is not fully portable, use new 2610 Perl script "util/mkdir-p.pl". 2611 [Bodo Moeller] 2612 2613 *) Rewrite the way password based encryption (PBE) is handled. It used to 2614 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 2615 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 2616 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 2617 the 'parameter' field of the AlgorithmIdentifier is passed to the 2618 underlying key generation function so it must do its own ASN1 parsing. 2619 This has also changed the EVP_PBE_CipherInit() function which now has a 2620 'parameter' argument instead of literal salt and iteration count values 2621 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 2622 [Steve Henson] 2623 2624 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 2625 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 2626 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 2627 KEY" because this clashed with PKCS#8 unencrypted string. Since this 2628 value was just used as a "magic string" and not used directly its 2629 value doesn't matter. 2630 [Steve Henson] 2631 2632 *) Introduce some semblance of const correctness to BN. Shame C doesn't 2633 support mutable. 2634 [Ben Laurie] 2635 2636 *) "linux-sparc64" configuration (ultrapenguin). 2637 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 2638 "linux-sparc" configuration. 2639 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 2640 2641 *) config now generates no-xxx options for missing ciphers. 2642 [Ulf M�ller] 2643 2644 *) Support the EBCDIC character set (work in progress). 2645 File ebcdic.c not yet included because it has a different license. 2646 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 2647 2648 *) Support BS2000/OSD-POSIX. 2649 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 2650 2651 *) Make callbacks for key generation use void * instead of char *. 2652 [Ben Laurie] 2653 2654 *) Make S/MIME samples compile (not yet tested). 2655 [Ben Laurie] 2656 2657 *) Additional typesafe stacks. 2658 [Ben Laurie] 2659 2660 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 2661 [Bodo Moeller] 2662 2663 2664 Changes between 0.9.3 and 0.9.3a [29 May 1999] 2665 2666 *) New configuration variant "sco5-gcc". 2667 2668 *) Updated some demos. 2669 [Sean O Riordain, Wade Scholine] 2670 2671 *) Add missing BIO_free at exit of pkcs12 application. 2672 [Wu Zhigang] 2673 2674 *) Fix memory leak in conf.c. 2675 [Steve Henson] 2676 2677 *) Updates for Win32 to assembler version of MD5. 2678 [Steve Henson] 2679 2680 *) Set #! path to perl in apps/der_chop to where we found it 2681 instead of using a fixed path. 2682 [Bodo Moeller] 2683 2684 *) SHA library changes for irix64-mips4-cc. 2685 [Andy Polyakov] 2686 2687 *) Improvements for VMS support. 2688 [Richard Levitte] 2689 2690 2691 Changes between 0.9.2b and 0.9.3 [24 May 1999] 2692 2693 *) Bignum library bug fix. IRIX 6 passes "make test" now! 2694 This also avoids the problems with SC4.2 and unpatched SC5. 2695 [Andy Polyakov <appro@fy.chalmers.se>] 2696 2697 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 2698 These are required because of the typesafe stack would otherwise break 2699 existing code. If old code used a structure member which used to be STACK 2700 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 2701 sk_num or sk_value it would produce an error because the num, data members 2702 are not present in STACK_OF. Now it just produces a warning. sk_set 2703 replaces the old method of assigning a value to sk_value 2704 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 2705 that does this will no longer work (and should use sk_set instead) but 2706 this could be regarded as a "questionable" behaviour anyway. 2707 [Steve Henson] 2708 2709 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 2710 correctly handle encrypted S/MIME data. 2711 [Steve Henson] 2712 2713 *) Change type of various DES function arguments from des_cblock 2714 (which means, in function argument declarations, pointer to char) 2715 to des_cblock * (meaning pointer to array with 8 char elements), 2716 which allows the compiler to do more typechecking; it was like 2717 that back in SSLeay, but with lots of ugly casts. 2718 2719 Introduce new type const_des_cblock. 2720 [Bodo Moeller] 2721 2722 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 2723 problems: find RecipientInfo structure that matches recipient certificate 2724 and initialise the ASN1 structures properly based on passed cipher. 2725 [Steve Henson] 2726 2727 *) Belatedly make the BN tests actually check the results. 2728 [Ben Laurie] 2729 2730 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 2731 to and from BNs: it was completely broken. New compilation option 2732 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 2733 key elements as negative integers. 2734 [Steve Henson] 2735 2736 *) Reorganize and speed up MD5. 2737 [Andy Polyakov <appro@fy.chalmers.se>] 2738 2739 *) VMS support. 2740 [Richard Levitte <richard@levitte.org>] 2741 2742 *) New option -out to asn1parse to allow the parsed structure to be 2743 output to a file. This is most useful when combined with the -strparse 2744 option to examine the output of things like OCTET STRINGS. 2745 [Steve Henson] 2746 2747 *) Make SSL library a little more fool-proof by not requiring any longer 2748 that SSL_set_{accept,connect}_state be called before 2749 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 2750 in many applications because usually everything *appeared* to work as 2751 intended anyway -- now it really works as intended). 2752 [Bodo Moeller] 2753 2754 *) Move openssl.cnf out of lib/. 2755 [Ulf M�ller] 2756 2757 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 2758 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 2759 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 2760 [Ralf S. Engelschall] 2761 2762 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 2763 handle PKCS#7 enveloped data properly. 2764 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 2765 2766 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 2767 copying pointers. The cert_st handling is changed by this in 2768 various ways (and thus what used to be known as ctx->default_cert 2769 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 2770 any longer when s->cert does not give us what we need). 2771 ssl_cert_instantiate becomes obsolete by this change. 2772 As soon as we've got the new code right (possibly it already is?), 2773 we have solved a couple of bugs of the earlier code where s->cert 2774 was used as if it could not have been shared with other SSL structures. 2775 2776 Note that using the SSL API in certain dirty ways now will result 2777 in different behaviour than observed with earlier library versions: 2778 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 2779 does not influence s as it used to. 2780 2781 In order to clean up things more thoroughly, inside SSL_SESSION 2782 we don't use CERT any longer, but a new structure SESS_CERT 2783 that holds per-session data (if available); currently, this is 2784 the peer's certificate chain and, for clients, the server's certificate 2785 and temporary key. CERT holds only those values that can have 2786 meaningful defaults in an SSL_CTX. 2787 [Bodo Moeller] 2788 2789 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 2790 from the internal representation. Various PKCS#7 fixes: remove some 2791 evil casts and set the enc_dig_alg field properly based on the signing 2792 key type. 2793 [Steve Henson] 2794 2795 *) Allow PKCS#12 password to be set from the command line or the 2796 environment. Let 'ca' get its config file name from the environment 2797 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 2798 and 'x509'). 2799 [Steve Henson] 2800 2801 *) Allow certificate policies extension to use an IA5STRING for the 2802 organization field. This is contrary to the PKIX definition but 2803 VeriSign uses it and IE5 only recognises this form. Document 'x509' 2804 extension option. 2805 [Steve Henson] 2806 2807 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 2808 without disallowing inline assembler and the like for non-pedantic builds. 2809 [Ben Laurie] 2810 2811 *) Support Borland C++ builder. 2812 [Janez Jere <jj@void.si>, modified by Ulf M�ller] 2813 2814 *) Support Mingw32. 2815 [Ulf M�ller] 2816 2817 *) SHA-1 cleanups and performance enhancements. 2818 [Andy Polyakov <appro@fy.chalmers.se>] 2819 2820 *) Sparc v8plus assembler for the bignum library. 2821 [Andy Polyakov <appro@fy.chalmers.se>] 2822 2823 *) Accept any -xxx and +xxx compiler options in Configure. 2824 [Ulf M�ller] 2825 2826 *) Update HPUX configuration. 2827 [Anonymous] 2828 2829 *) Add missing sk_<type>_unshift() function to safestack.h 2830 [Ralf S. Engelschall] 2831 2832 *) New function SSL_CTX_use_certificate_chain_file that sets the 2833 "extra_cert"s in addition to the certificate. (This makes sense 2834 only for "PEM" format files, as chains as a whole are not 2835 DER-encoded.) 2836 [Bodo Moeller] 2837 2838 *) Support verify_depth from the SSL API. 2839 x509_vfy.c had what can be considered an off-by-one-error: 2840 Its depth (which was not part of the external interface) 2841 was actually counting the number of certificates in a chain; 2842 now it really counts the depth. 2843 [Bodo Moeller] 2844 2845 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 2846 instead of X509err, which often resulted in confusing error 2847 messages since the error codes are not globally unique 2848 (e.g. an alleged error in ssl3_accept when a certificate 2849 didn't match the private key). 2850 2851 *) New function SSL_CTX_set_session_id_context that allows to set a default 2852 value (so that you don't need SSL_set_session_id_context for each 2853 connection using the SSL_CTX). 2854 [Bodo Moeller] 2855 2856 *) OAEP decoding bug fix. 2857 [Ulf M�ller] 2858 2859 *) Support INSTALL_PREFIX for package builders, as proposed by 2860 David Harris. 2861 [Bodo Moeller] 2862 2863 *) New Configure options "threads" and "no-threads". For systems 2864 where the proper compiler options are known (currently Solaris 2865 and Linux), "threads" is the default. 2866 [Bodo Moeller] 2867 2868 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 2869 [Bodo Moeller] 2870 2871 *) Install various scripts to $(OPENSSLDIR)/misc, not to 2872 $(INSTALLTOP)/bin -- they shouldn't clutter directories 2873 such as /usr/local/bin. 2874 [Bodo Moeller] 2875 2876 *) "make linux-shared" to build shared libraries. 2877 [Niels Poppe <niels@netbox.org>] 2878 2879 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 2880 [Ulf M�ller] 2881 2882 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 2883 extension adding in x509 utility. 2884 [Steve Henson] 2885 2886 *) Remove NOPROTO sections and error code comments. 2887 [Ulf M�ller] 2888 2889 *) Partial rewrite of the DEF file generator to now parse the ANSI 2890 prototypes. 2891 [Steve Henson] 2892 2893 *) New Configure options --prefix=DIR and --openssldir=DIR. 2894 [Ulf M�ller] 2895 2896 *) Complete rewrite of the error code script(s). It is all now handled 2897 by one script at the top level which handles error code gathering, 2898 header rewriting and C source file generation. It should be much better 2899 than the old method: it now uses a modified version of Ulf's parser to 2900 read the ANSI prototypes in all header files (thus the old K&R definitions 2901 aren't needed for error creation any more) and do a better job of 2902 translating function codes into names. The old 'ASN1 error code imbedded 2903 in a comment' is no longer necessary and it doesn't use .err files which 2904 have now been deleted. Also the error code call doesn't have to appear all 2905 on one line (which resulted in some large lines...). 2906 [Steve Henson] 2907 2908 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 2909 [Bodo Moeller] 2910 2911 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 2912 0 (which usually indicates a closed connection), but continue reading. 2913 [Bodo Moeller] 2914 2915 *) Fix some race conditions. 2916 [Bodo Moeller] 2917 2918 *) Add support for CRL distribution points extension. Add Certificate 2919 Policies and CRL distribution points documentation. 2920 [Steve Henson] 2921 2922 *) Move the autogenerated header file parts to crypto/opensslconf.h. 2923 [Ulf M�ller] 2924 2925 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 2926 8 of keying material. Merlin has also confirmed interop with this fix 2927 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 2928 [Merlin Hughes <merlin@baltimore.ie>] 2929 2930 *) Fix lots of warnings. 2931 [Richard Levitte <levitte@stacken.kth.se>] 2932 2933 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 2934 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 2935 [Richard Levitte <levitte@stacken.kth.se>] 2936 2937 *) Fix problems with sizeof(long) == 8. 2938 [Andy Polyakov <appro@fy.chalmers.se>] 2939 2940 *) Change functions to ANSI C. 2941 [Ulf M�ller] 2942 2943 *) Fix typos in error codes. 2944 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 2945 2946 *) Remove defunct assembler files from Configure. 2947 [Ulf M�ller] 2948 2949 *) SPARC v8 assembler BIGNUM implementation. 2950 [Andy Polyakov <appro@fy.chalmers.se>] 2951 2952 *) Support for Certificate Policies extension: both print and set. 2953 Various additions to support the r2i method this uses. 2954 [Steve Henson] 2955 2956 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 2957 return a const string when you are expecting an allocated buffer. 2958 [Ben Laurie] 2959 2960 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 2961 types DirectoryString and DisplayText. 2962 [Steve Henson] 2963 2964 *) Add code to allow r2i extensions to access the configuration database, 2965 add an LHASH database driver and add several ctx helper functions. 2966 [Steve Henson] 2967 2968 *) Fix an evil bug in bn_expand2() which caused various BN functions to 2969 fail when they extended the size of a BIGNUM. 2970 [Steve Henson] 2971 2972 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 2973 support typesafe stack. 2974 [Steve Henson] 2975 2976 *) Fix typo in SSL_[gs]et_options(). 2977 [Nils Frostberg <nils@medcom.se>] 2978 2979 *) Delete various functions and files that belonged to the (now obsolete) 2980 old X509V3 handling code. 2981 [Steve Henson] 2982 2983 *) New Configure option "rsaref". 2984 [Ulf M�ller] 2985 2986 *) Don't auto-generate pem.h. 2987 [Bodo Moeller] 2988 2989 *) Introduce type-safe ASN.1 SETs. 2990 [Ben Laurie] 2991 2992 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 2993 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 2994 2995 *) Introduce type-safe STACKs. This will almost certainly break lots of code 2996 that links with OpenSSL (well at least cause lots of warnings), but fear 2997 not: the conversion is trivial, and it eliminates loads of evil casts. A 2998 few STACKed things have been converted already. Feel free to convert more. 2999 In the fullness of time, I'll do away with the STACK type altogether. 3000 [Ben Laurie] 3001 3002 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 3003 specified in <certfile> by updating the entry in the index.txt file. 3004 This way one no longer has to edit the index.txt file manually for 3005 revoking a certificate. The -revoke option does the gory details now. 3006 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 3007 3008 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 3009 `-text' option at all and this way the `-noout -text' combination was 3010 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 3011 [Ralf S. Engelschall] 3012 3013 *) Make sure a corresponding plain text error message exists for the 3014 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 3015 verify callback function determined that a certificate was revoked. 3016 [Ralf S. Engelschall] 3017 3018 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 3019 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 3020 all available cipers including rc5, which was forgotten until now. 3021 In order to let the testing shell script know which algorithms 3022 are available, a new (up to now undocumented) command 3023 "openssl list-cipher-commands" is used. 3024 [Bodo Moeller] 3025 3026 *) Bugfix: s_client occasionally would sleep in select() when 3027 it should have checked SSL_pending() first. 3028 [Bodo Moeller] 3029 3030 *) New functions DSA_do_sign and DSA_do_verify to provide access to 3031 the raw DSA values prior to ASN.1 encoding. 3032 [Ulf M�ller] 3033 3034 *) Tweaks to Configure 3035 [Niels Poppe <niels@netbox.org>] 3036 3037 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 3038 yet... 3039 [Steve Henson] 3040 3041 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 3042 [Ulf M�ller] 3043 3044 *) New config option to avoid instructions that are illegal on the 80386. 3045 The default code is faster, but requires at least a 486. 3046 [Ulf M�ller] 3047 3048 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 3049 SSL2_SERVER_VERSION (not used at all) macros, which are now the 3050 same as SSL2_VERSION anyway. 3051 [Bodo Moeller] 3052 3053 *) New "-showcerts" option for s_client. 3054 [Bodo Moeller] 3055 3056 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 3057 application. Various cleanups and fixes. 3058 [Steve Henson] 3059 3060 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 3061 modify error routines to work internally. Add error codes and PBE init 3062 to library startup routines. 3063 [Steve Henson] 3064 3065 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 3066 packing functions to asn1 and evp. Changed function names and error 3067 codes along the way. 3068 [Steve Henson] 3069 3070 *) PKCS12 integration: and so it begins... First of several patches to 3071 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 3072 objects to objects.h 3073 [Steve Henson] 3074 3075 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 3076 and display support for Thawte strong extranet extension. 3077 [Steve Henson] 3078 3079 *) Add LinuxPPC support. 3080 [Jeff Dubrule <igor@pobox.org>] 3081 3082 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 3083 bn_div_words in alpha.s. 3084 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 3085 3086 *) Make sure the RSA OAEP test is skipped under -DRSAref because 3087 OAEP isn't supported when OpenSSL is built with RSAref. 3088 [Ulf Moeller <ulf@fitug.de>] 3089 3090 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 3091 so they no longer are missing under -DNOPROTO. 3092 [Soren S. Jorvang <soren@t.dk>] 3093 3094 3095 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 3096 3097 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 3098 doesn't work when the session is reused. Coming soon! 3099 [Ben Laurie] 3100 3101 *) Fix a security hole, that allows sessions to be reused in the wrong 3102 context thus bypassing client cert protection! All software that uses 3103 client certs and session caches in multiple contexts NEEDS PATCHING to 3104 allow session reuse! A fuller solution is in the works. 3105 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 3106 3107 *) Some more source tree cleanups (removed obsolete files 3108 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 3109 permission on "config" script to be executable) and a fix for the INSTALL 3110 document. 3111 [Ulf Moeller <ulf@fitug.de>] 3112 3113 *) Remove some legacy and erroneous uses of malloc, free instead of 3114 Malloc, Free. 3115 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 3116 3117 *) Make rsa_oaep_test return non-zero on error. 3118 [Ulf Moeller <ulf@fitug.de>] 3119 3120 *) Add support for native Solaris shared libraries. Configure 3121 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 3122 if someone would make that last step automatic. 3123 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 3124 3125 *) ctx_size was not built with the right compiler during "make links". Fixed. 3126 [Ben Laurie] 3127 3128 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 3129 except NULL ciphers". This means the default cipher list will no longer 3130 enable NULL ciphers. They need to be specifically enabled e.g. with 3131 the string "DEFAULT:eNULL". 3132 [Steve Henson] 3133 3134 *) Fix to RSA private encryption routines: if p < q then it would 3135 occasionally produce an invalid result. This will only happen with 3136 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 3137 [Steve Henson] 3138 3139 *) Be less restrictive and allow also `perl util/perlpath.pl 3140 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 3141 because this way one can also use an interpreter named `perl5' (which is 3142 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 3143 installed as `perl'). 3144 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 3145 3146 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 3147 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 3148 3149 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 3150 advapi32.lib to Win32 build and change the pem test comparision 3151 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 3152 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 3153 and crypto/des/ede_cbcm_enc.c. 3154 [Steve Henson] 3155 3156 *) DES quad checksum was broken on big-endian architectures. Fixed. 3157 [Ben Laurie] 3158 3159 *) Comment out two functions in bio.h that aren't implemented. Fix up the 3160 Win32 test batch file so it (might) work again. The Win32 test batch file 3161 is horrible: I feel ill.... 3162 [Steve Henson] 3163 3164 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 3165 in e_os.h. Audit of header files to check ANSI and non ANSI 3166 sections: 10 functions were absent from non ANSI section and not exported 3167 from Windows DLLs. Fixed up libeay.num for new functions. 3168 [Steve Henson] 3169 3170 *) Make `openssl version' output lines consistent. 3171 [Ralf S. Engelschall] 3172 3173 *) Fix Win32 symbol export lists for BIO functions: Added 3174 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 3175 to ms/libeay{16,32}.def. 3176 [Ralf S. Engelschall] 3177 3178 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 3179 fine under Unix and passes some trivial tests I've now added. But the 3180 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 3181 added to make sure no one expects that this stuff really works in the 3182 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 3183 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 3184 openssl_bio.xs. 3185 [Ralf S. Engelschall] 3186 3187 *) Fix the generation of two part addresses in perl. 3188 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 3189 3190 *) Add config entry for Linux on MIPS. 3191 [John Tobey <jtobey@channel1.com>] 3192 3193 *) Make links whenever Configure is run, unless we are on Windoze. 3194 [Ben Laurie] 3195 3196 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 3197 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 3198 in CRLs. 3199 [Steve Henson] 3200 3201 *) Add a useful kludge to allow package maintainers to specify compiler and 3202 other platforms details on the command line without having to patch the 3203 Configure script everytime: One now can use ``perl Configure 3204 <id>:<details>'', i.e. platform ids are allowed to have details appended 3205 to them (seperated by colons). This is treated as there would be a static 3206 pre-configured entry in Configure's %table under key <id> with value 3207 <details> and ``perl Configure <id>'' is called. So, when you want to 3208 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 3209 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 3210 now, which overrides the FreeBSD-elf entry on-the-fly. 3211 [Ralf S. Engelschall] 3212 3213 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 3214 [Ben Laurie] 3215 3216 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 3217 on the `perl Configure ...' command line. This way one can compile 3218 OpenSSL libraries with Position Independent Code (PIC) which is needed 3219 for linking it into DSOs. 3220 [Ralf S. Engelschall] 3221 3222 *) Remarkably, export ciphers were totally broken and no-one had noticed! 3223 Fixed. 3224 [Ben Laurie] 3225 3226 *) Cleaned up the LICENSE document: The official contact for any license 3227 questions now is the OpenSSL core team under openssl-core@openssl.org. 3228 And add a paragraph about the dual-license situation to make sure people 3229 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 3230 to the OpenSSL toolkit. 3231 [Ralf S. Engelschall] 3232 3233 *) General source tree makefile cleanups: Made `making xxx in yyy...' 3234 display consistent in the source tree and replaced `/bin/rm' by `rm'. 3235 Additonally cleaned up the `make links' target: Remove unnecessary 3236 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 3237 to speed processing and no longer clutter the display with confusing 3238 stuff. Instead only the actually done links are displayed. 3239 [Ralf S. Engelschall] 3240 3241 *) Permit null encryption ciphersuites, used for authentication only. It used 3242 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 3243 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 3244 encryption. 3245 [Ben Laurie] 3246 3247 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 3248 signed attributes when verifying signatures (this would break them), 3249 the detached data encoding was wrong and public keys obtained using 3250 X509_get_pubkey() weren't freed. 3251 [Steve Henson] 3252 3253 *) Add text documentation for the BUFFER functions. Also added a work around 3254 to a Win95 console bug. This was triggered by the password read stuff: the 3255 last character typed gets carried over to the next fread(). If you were 3256 generating a new cert request using 'req' for example then the last 3257 character of the passphrase would be CR which would then enter the first 3258 field as blank. 3259 [Steve Henson] 3260 3261 *) Added the new `Includes OpenSSL Cryptography Software' button as 3262 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 3263 button and can be used by applications based on OpenSSL to show the 3264 relationship to the OpenSSL project. 3265 [Ralf S. Engelschall] 3266 3267 *) Remove confusing variables in function signatures in files 3268 ssl/ssl_lib.c and ssl/ssl.h. 3269 [Lennart Bong <lob@kulthea.stacken.kth.se>] 3270 3271 *) Don't install bss_file.c under PREFIX/include/ 3272 [Lennart Bong <lob@kulthea.stacken.kth.se>] 3273 3274 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 3275 functions that return function pointers and has support for NT specific 3276 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 3277 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 3278 unsigned to signed types: this was killing the Win32 compile. 3279 [Steve Henson] 3280 3281 *) Add new certificate file to stack functions, 3282 SSL_add_dir_cert_subjects_to_stack() and 3283 SSL_add_file_cert_subjects_to_stack(). These largely supplant 3284 SSL_load_client_CA_file(), and can be used to add multiple certs easily 3285 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 3286 This means that Apache-SSL and similar packages don't have to mess around 3287 to add as many CAs as they want to the preferred list. 3288 [Ben Laurie] 3289 3290 *) Experiment with doxygen documentation. Currently only partially applied to 3291 ssl/ssl_lib.c. 3292 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 3293 openssl.doxy as the configuration file. 3294 [Ben Laurie] 3295 3296 *) Get rid of remaining C++-style comments which strict C compilers hate. 3297 [Ralf S. Engelschall, pointed out by Carlos Amengual] 3298 3299 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 3300 compiled in by default: it has problems with large keys. 3301 [Steve Henson] 3302 3303 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 3304 DH private keys and/or callback functions which directly correspond to 3305 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 3306 is needed for applications which have to configure certificates on a 3307 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 3308 (e.g. s_server). 3309 For the RSA certificate situation is makes no difference, but 3310 for the DSA certificate situation this fixes the "no shared cipher" 3311 problem where the OpenSSL cipher selection procedure failed because the 3312 temporary keys were not overtaken from the context and the API provided 3313 no way to reconfigure them. 3314 The new functions now let applications reconfigure the stuff and they 3315 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 3316 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 3317 non-public-API function ssl_cert_instantiate() is used as a helper 3318 function and also to reduce code redundancy inside ssl_rsa.c. 3319 [Ralf S. Engelschall] 3320 3321 *) Move s_server -dcert and -dkey options out of the undocumented feature 3322 area because they are useful for the DSA situation and should be 3323 recognized by the users. 3324 [Ralf S. Engelschall] 3325 3326 *) Fix the cipher decision scheme for export ciphers: the export bits are 3327 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 3328 SSL_EXP_MASK. So, the original variable has to be used instead of the 3329 already masked variable. 3330 [Richard Levitte <levitte@stacken.kth.se>] 3331 3332 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 3333 [Richard Levitte <levitte@stacken.kth.se>] 3334 3335 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 3336 from `int' to `unsigned int' because it's a length and initialized by 3337 EVP_DigestFinal() which expects an `unsigned int *'. 3338 [Richard Levitte <levitte@stacken.kth.se>] 3339 3340 *) Don't hard-code path to Perl interpreter on shebang line of Configure 3341 script. Instead use the usual Shell->Perl transition trick. 3342 [Ralf S. Engelschall] 3343 3344 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 3345 (in addition to RSA certificates) to match the behaviour of `openssl dsa 3346 -noout -modulus' as it's already the case for `openssl rsa -noout 3347 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 3348 currently the public key is printed (a decision which was already done by 3349 `openssl dsa -modulus' in the past) which serves a similar purpose. 3350 Additionally the NO_RSA no longer completely removes the whole -modulus 3351 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 3352 now, too. 3353 [Ralf S. Engelschall] 3354 3355 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 3356 BIO. See the source (crypto/evp/bio_ok.c) for more info. 3357 [Arne Ansper <arne@ats.cyber.ee>] 3358 3359 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 3360 to be added. Now both 'req' and 'ca' can use new objects defined in the 3361 config file. 3362 [Steve Henson] 3363 3364 *) Add cool BIO that does syslog (or event log on NT). 3365 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 3366 3367 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 3368 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 3369 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 3370 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 3371 [Ben Laurie] 3372 3373 *) Add preliminary config info for new extension code. 3374 [Steve Henson] 3375 3376 *) Make RSA_NO_PADDING really use no padding. 3377 [Ulf Moeller <ulf@fitug.de>] 3378 3379 *) Generate errors when private/public key check is done. 3380 [Ben Laurie] 3381 3382 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 3383 for some CRL extensions and new objects added. 3384 [Steve Henson] 3385 3386 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 3387 key usage extension and fuller support for authority key id. 3388 [Steve Henson] 3389 3390 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 3391 padding method for RSA, which is recommended for new applications in PKCS 3392 #1 v2.0 (RFC 2437, October 1998). 3393 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 3394 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 3395 against Bleichbacher's attack on RSA. 3396 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 3397 Ben Laurie] 3398 3399 *) Updates to the new SSL compression code 3400 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3401 3402 *) Fix so that the version number in the master secret, when passed 3403 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 3404 (because the server will not accept higher), that the version number 3405 is 0x03,0x01, not 0x03,0x00 3406 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3407 3408 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 3409 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 3410 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 3411 [Steve Henson] 3412 3413 *) Support for RAW extensions where an arbitrary extension can be 3414 created by including its DER encoding. See apps/openssl.cnf for 3415 an example. 3416 [Steve Henson] 3417 3418 *) Make sure latest Perl versions don't interpret some generated C array 3419 code as Perl array code in the crypto/err/err_genc.pl script. 3420 [Lars Weber <3weber@informatik.uni-hamburg.de>] 3421 3422 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 3423 not many people have the assembler. Various Win32 compilation fixes and 3424 update to the INSTALL.W32 file with (hopefully) more accurate Win32 3425 build instructions. 3426 [Steve Henson] 3427 3428 *) Modify configure script 'Configure' to automatically create crypto/date.h 3429 file under Win32 and also build pem.h from pem.org. New script 3430 util/mkfiles.pl to create the MINFO file on environments that can't do a 3431 'make files': perl util/mkfiles.pl >MINFO should work. 3432 [Steve Henson] 3433 3434 *) Major rework of DES function declarations, in the pursuit of correctness 3435 and purity. As a result, many evil casts evaporated, and some weirdness, 3436 too. You may find this causes warnings in your code. Zapping your evil 3437 casts will probably fix them. Mostly. 3438 [Ben Laurie] 3439 3440 *) Fix for a typo in asn1.h. Bug fix to object creation script 3441 obj_dat.pl. It considered a zero in an object definition to mean 3442 "end of object": none of the objects in objects.h have any zeros 3443 so it wasn't spotted. 3444 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 3445 3446 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 3447 Masking (CBCM). In the absence of test vectors, the best I have been able 3448 to do is check that the decrypt undoes the encrypt, so far. Send me test 3449 vectors if you have them. 3450 [Ben Laurie] 3451 3452 *) Correct calculation of key length for export ciphers (too much space was 3453 allocated for null ciphers). This has not been tested! 3454 [Ben Laurie] 3455 3456 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 3457 message is now correct (it understands "crypto" and "ssl" on its 3458 command line). There is also now an "update" option. This will update 3459 the util/ssleay.num and util/libeay.num files with any new functions. 3460 If you do a: 3461 perl util/mkdef.pl crypto ssl update 3462 it will update them. 3463 [Steve Henson] 3464 3465 *) Overhauled the Perl interface (perl/*): 3466 - ported BN stuff to OpenSSL's different BN library 3467 - made the perl/ source tree CVS-aware 3468 - renamed the package from SSLeay to OpenSSL (the files still contain 3469 their history because I've copied them in the repository) 3470 - removed obsolete files (the test scripts will be replaced 3471 by better Test::Harness variants in the future) 3472 [Ralf S. Engelschall] 3473 3474 *) First cut for a very conservative source tree cleanup: 3475 1. merge various obsolete readme texts into doc/ssleay.txt 3476 where we collect the old documents and readme texts. 3477 2. remove the first part of files where I'm already sure that we no 3478 longer need them because of three reasons: either they are just temporary 3479 files which were left by Eric or they are preserved original files where 3480 I've verified that the diff is also available in the CVS via "cvs diff 3481 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 3482 the crypto/md/ stuff). 3483 [Ralf S. Engelschall] 3484 3485 *) More extension code. Incomplete support for subject and issuer alt 3486 name, issuer and authority key id. Change the i2v function parameters 3487 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 3488 what that's for :-) Fix to ASN1 macro which messed up 3489 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 3490 [Steve Henson] 3491 3492 *) Preliminary support for ENUMERATED type. This is largely copied from the 3493 INTEGER code. 3494 [Steve Henson] 3495 3496 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 3497 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3498 3499 *) Make sure `make rehash' target really finds the `openssl' program. 3500 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 3501 3502 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 3503 like to hear about it if this slows down other processors. 3504 [Ben Laurie] 3505 3506 *) Add CygWin32 platform information to Configure script. 3507 [Alan Batie <batie@aahz.jf.intel.com>] 3508 3509 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 3510 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 3511 3512 *) New program nseq to manipulate netscape certificate sequences 3513 [Steve Henson] 3514 3515 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 3516 few typos. 3517 [Steve Henson] 3518 3519 *) Fixes to BN code. Previously the default was to define BN_RECURSION 3520 but the BN code had some problems that would cause failures when 3521 doing certificate verification and some other functions. 3522 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3523 3524 *) Add ASN1 and PEM code to support netscape certificate sequences. 3525 [Steve Henson] 3526 3527 *) Add ASN1 and PEM code to support netscape certificate sequences. 3528 [Steve Henson] 3529 3530 *) Add several PKIX and private extended key usage OIDs. 3531 [Steve Henson] 3532 3533 *) Modify the 'ca' program to handle the new extension code. Modify 3534 openssl.cnf for new extension format, add comments. 3535 [Steve Henson] 3536 3537 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 3538 and add a sample to openssl.cnf so req -x509 now adds appropriate 3539 CA extensions. 3540 [Steve Henson] 3541 3542 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 3543 error code, add initial support to X509_print() and x509 application. 3544 [Steve Henson] 3545 3546 *) Takes a deep breath and start addding X509 V3 extension support code. Add 3547 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 3548 stuff is currently isolated and isn't even compiled yet. 3549 [Steve Henson] 3550 3551 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 3552 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 3553 Removed the versions check from X509 routines when loading extensions: 3554 this allows certain broken certificates that don't set the version 3555 properly to be processed. 3556 [Steve Henson] 3557 3558 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 3559 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 3560 can still be regenerated with "make depend". 3561 [Ben Laurie] 3562 3563 *) Spelling mistake in C version of CAST-128. 3564 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 3565 3566 *) Changes to the error generation code. The perl script err-code.pl 3567 now reads in the old error codes and retains the old numbers, only 3568 adding new ones if necessary. It also only changes the .err files if new 3569 codes are added. The makefiles have been modified to only insert errors 3570 when needed (to avoid needlessly modifying header files). This is done 3571 by only inserting errors if the .err file is newer than the auto generated 3572 C file. To rebuild all the error codes from scratch (the old behaviour) 3573 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 3574 or delete all the .err files. 3575 [Steve Henson] 3576 3577 *) CAST-128 was incorrectly implemented for short keys. The C version has 3578 been fixed, but is untested. The assembler versions are also fixed, but 3579 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 3580 to regenerate it if needed. 3581 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 3582 Hagino <itojun@kame.net>] 3583 3584 *) File was opened incorrectly in randfile.c. 3585 [Ulf M�ller <ulf@fitug.de>] 3586 3587 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 3588 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 3589 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 3590 al: it's just almost always a UTCTime. Note this patch adds new error 3591 codes so do a "make errors" if there are problems. 3592 [Steve Henson] 3593 3594 *) Correct Linux 1 recognition in config. 3595 [Ulf M�ller <ulf@fitug.de>] 3596 3597 *) Remove pointless MD5 hash when using DSA keys in ca. 3598 [Anonymous <nobody@replay.com>] 3599 3600 *) Generate an error if given an empty string as a cert directory. Also 3601 generate an error if handed NULL (previously returned 0 to indicate an 3602 error, but didn't set one). 3603 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 3604 3605 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 3606 [Ben Laurie] 3607 3608 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 3609 parameters. This was causing a warning which killed off the Win32 compile. 3610 [Steve Henson] 3611 3612 *) Remove C++ style comments from crypto/bn/bn_local.h. 3613 [Neil Costigan <neil.costigan@celocom.com>] 3614 3615 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 3616 based on a text string, looking up short and long names and finally 3617 "dot" format. The "dot" format stuff didn't work. Added new function 3618 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 3619 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 3620 OID is not part of the table. 3621 [Steve Henson] 3622 3623 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 3624 X509_LOOKUP_by_alias(). 3625 [Ben Laurie] 3626 3627 *) Sort openssl functions by name. 3628 [Ben Laurie] 3629 3630 *) Get the gendsa program working (hopefully) and add it to app list. Remove 3631 encryption from sample DSA keys (in case anyone is interested the password 3632 was "1234"). 3633 [Steve Henson] 3634 3635 *) Make _all_ *_free functions accept a NULL pointer. 3636 [Frans Heymans <fheymans@isaserver.be>] 3637 3638 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 3639 NULL pointers. 3640 [Anonymous <nobody@replay.com>] 3641 3642 *) s_server should send the CAfile as acceptable CAs, not its own cert. 3643 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 3644 3645 *) Don't blow it for numeric -newkey arguments to apps/req. 3646 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 3647 3648 *) Temp key "for export" tests were wrong in s3_srvr.c. 3649 [Anonymous <nobody@replay.com>] 3650 3651 *) Add prototype for temp key callback functions 3652 SSL_CTX_set_tmp_{rsa,dh}_callback(). 3653 [Ben Laurie] 3654 3655 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 3656 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 3657 [Steve Henson] 3658 3659 *) X509_name_add_entry() freed the wrong thing after an error. 3660 [Arne Ansper <arne@ats.cyber.ee>] 3661 3662 *) rsa_eay.c would attempt to free a NULL context. 3663 [Arne Ansper <arne@ats.cyber.ee>] 3664 3665 *) BIO_s_socket() had a broken should_retry() on Windoze. 3666 [Arne Ansper <arne@ats.cyber.ee>] 3667 3668 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 3669 [Arne Ansper <arne@ats.cyber.ee>] 3670 3671 *) Make sure the already existing X509_STORE->depth variable is initialized 3672 in X509_STORE_new(), but document the fact that this variable is still 3673 unused in the certificate verification process. 3674 [Ralf S. Engelschall] 3675 3676 *) Fix the various library and apps files to free up pkeys obtained from 3677 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 3678 [Steve Henson] 3679 3680 *) Fix reference counting in X509_PUBKEY_get(). This makes 3681 demos/maurice/example2.c work, amongst others, probably. 3682 [Steve Henson and Ben Laurie] 3683 3684 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 3685 `openssl' and second, the shortcut symlinks for the `openssl <command>' 3686 are no longer created. This way we have a single and consistent command 3687 line interface `openssl <command>', similar to `cvs <command>'. 3688 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 3689 3690 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 3691 BIT STRING wrapper always have zero unused bits. 3692 [Steve Henson] 3693 3694 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 3695 [Steve Henson] 3696 3697 *) Make the top-level INSTALL documentation easier to understand. 3698 [Paul Sutton] 3699 3700 *) Makefiles updated to exit if an error occurs in a sub-directory 3701 make (including if user presses ^C) [Paul Sutton] 3702 3703 *) Make Montgomery context stuff explicit in RSA data structure. 3704 [Ben Laurie] 3705 3706 *) Fix build order of pem and err to allow for generated pem.h. 3707 [Ben Laurie] 3708 3709 *) Fix renumbering bug in X509_NAME_delete_entry(). 3710 [Ben Laurie] 3711 3712 *) Enhanced the err-ins.pl script so it makes the error library number 3713 global and can add a library name. This is needed for external ASN1 and 3714 other error libraries. 3715 [Steve Henson] 3716 3717 *) Fixed sk_insert which never worked properly. 3718 [Steve Henson] 3719 3720 *) Fix ASN1 macros so they can handle indefinite length construted 3721 EXPLICIT tags. Some non standard certificates use these: they can now 3722 be read in. 3723 [Steve Henson] 3724 3725 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 3726 into a single doc/ssleay.txt bundle. This way the information is still 3727 preserved but no longer messes up this directory. Now it's new room for 3728 the new set of documenation files. 3729 [Ralf S. Engelschall] 3730 3731 *) SETs were incorrectly DER encoded. This was a major pain, because they 3732 shared code with SEQUENCEs, which aren't coded the same. This means that 3733 almost everything to do with SETs or SEQUENCEs has either changed name or 3734 number of arguments. 3735 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 3736 3737 *) Fix test data to work with the above. 3738 [Ben Laurie] 3739 3740 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 3741 was already fixed by Eric for 0.9.1 it seems. 3742 [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 3743 3744 *) Autodetect FreeBSD3. 3745 [Ben Laurie] 3746 3747 *) Fix various bugs in Configure. This affects the following platforms: 3748 nextstep 3749 ncr-scde 3750 unixware-2.0 3751 unixware-2.0-pentium 3752 sco5-cc. 3753 [Ben Laurie] 3754 3755 *) Eliminate generated files from CVS. Reorder tests to regenerate files 3756 before they are needed. 3757 [Ben Laurie] 3758 3759 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 3760 [Ben Laurie] 3761 3762 3763 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 3764 3765 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 3766 changed SSLeay to OpenSSL in version strings. 3767 [Ralf S. Engelschall] 3768 3769 *) Some fixups to the top-level documents. 3770 [Paul Sutton] 3771 3772 *) Fixed the nasty bug where rsaref.h was not found under compile-time 3773 because the symlink to include/ was missing. 3774 [Ralf S. Engelschall] 3775 3776 *) Incorporated the popular no-RSA/DSA-only patches 3777 which allow to compile a RSA-free SSLeay. 3778 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 3779 3780 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 3781 when "ssleay" is still not found. 3782 [Ralf S. Engelschall] 3783 3784 *) Added more platforms to Configure: Cray T3E, HPUX 11, 3785 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 3786 3787 *) Updated the README file. 3788 [Ralf S. Engelschall] 3789 3790 *) Added various .cvsignore files in the CVS repository subdirs 3791 to make a "cvs update" really silent. 3792 [Ralf S. Engelschall] 3793 3794 *) Recompiled the error-definition header files and added 3795 missing symbols to the Win32 linker tables. 3796 [Ralf S. Engelschall] 3797 3798 *) Cleaned up the top-level documents; 3799 o new files: CHANGES and LICENSE 3800 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 3801 o merged COPYRIGHT into LICENSE 3802 o removed obsolete TODO file 3803 o renamed MICROSOFT to INSTALL.W32 3804 [Ralf S. Engelschall] 3805 3806 *) Removed dummy files from the 0.9.1b source tree: 3807 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 3808 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 3809 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 3810 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 3811 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 3812 [Ralf S. Engelschall] 3813 3814 *) Added various platform portability fixes. 3815 [Mark J. Cox] 3816 3817 *) The Genesis of the OpenSSL rpject: 3818 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 3819 Young and Tim J. Hudson created while they were working for C2Net until 3820 summer 1998. 3821 [The OpenSSL Project] 3822 3823 3824 Changes between 0.9.0b and 0.9.1b [not released] 3825 3826 *) Updated a few CA certificates under certs/ 3827 [Eric A. Young] 3828 3829 *) Changed some BIGNUM api stuff. 3830 [Eric A. Young] 3831 3832 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 3833 DGUX x86, Linux Alpha, etc. 3834 [Eric A. Young] 3835 3836 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 3837 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 3838 available). 3839 [Eric A. Young] 3840 3841 *) Add -strparse option to asn1pars program which parses nested 3842 binary structures 3843 [Dr Stephen Henson <shenson@bigfoot.com>] 3844 3845 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 3846 [Eric A. Young] 3847 3848 *) DSA fix for "ca" program. 3849 [Eric A. Young] 3850 3851 *) Added "-genkey" option to "dsaparam" program. 3852 [Eric A. Young] 3853 3854 *) Added RIPE MD160 (rmd160) message digest. 3855 [Eric A. Young] 3856 3857 *) Added -a (all) option to "ssleay version" command. 3858 [Eric A. Young] 3859 3860 *) Added PLATFORM define which is the id given to Configure. 3861 [Eric A. Young] 3862 3863 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 3864 [Eric A. Young] 3865 3866 *) Extended the ASN.1 parser routines. 3867 [Eric A. Young] 3868 3869 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 3870 [Eric A. Young] 3871 3872 *) Added a BN_CTX to the BN library. 3873 [Eric A. Young] 3874 3875 *) Fixed the weak key values in DES library 3876 [Eric A. Young] 3877 3878 *) Changed API in EVP library for cipher aliases. 3879 [Eric A. Young] 3880 3881 *) Added support for RC2/64bit cipher. 3882 [Eric A. Young] 3883 3884 *) Converted the lhash library to the crypto/mem.c functions. 3885 [Eric A. Young] 3886 3887 *) Added more recognized ASN.1 object ids. 3888 [Eric A. Young] 3889 3890 *) Added more RSA padding checks for SSL/TLS. 3891 [Eric A. Young] 3892 3893 *) Added BIO proxy/filter functionality. 3894 [Eric A. Young] 3895 3896 *) Added extra_certs to SSL_CTX which can be used 3897 send extra CA certificates to the client in the CA cert chain sending 3898 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 3899 [Eric A. Young] 3900 3901 *) Now Fortezza is denied in the authentication phase because 3902 this is key exchange mechanism is not supported by SSLeay at all. 3903 [Eric A. Young] 3904 3905 *) Additional PKCS1 checks. 3906 [Eric A. Young] 3907 3908 *) Support the string "TLSv1" for all TLS v1 ciphers. 3909 [Eric A. Young] 3910 3911 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 3912 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 3913 [Eric A. Young] 3914 3915 *) Fixed a few memory leaks. 3916 [Eric A. Young] 3917 3918 *) Fixed various code and comment typos. 3919 [Eric A. Young] 3920 3921 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 3922 bytes sent in the client random. 3923 [Edward Bishop <ebishop@spyglass.com>] 3924 3925