CHANGES revision 100936
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6 7 *) Fix cipher selection routines: ciphers without encryption had no flags 8 for the cipher strength set and where therefore not handled correctly 9 by the selection routines (PR #130). 10 [Lutz Jaenicke] 11 12 *) Fix EVP_dsa_sha macro. 13 [Nils Larsch] 14 15 *) New option 16 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 17 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 18 that was added in OpenSSL 0.9.6d. 19 20 As the countermeasure turned out to be incompatible with some 21 broken SSL implementations, the new option is part of SSL_OP_ALL. 22 SSL_OP_ALL is usually employed when compatibility with weird SSL 23 implementations is desired (e.g. '-bugs' option to 's_client' and 24 's_server'), so the new option is automatically set in many 25 applications. 26 [Bodo Moeller] 27 28 *) Changes in security patch: 29 30 Changes marked "(CHATS)" were sponsored by the Defense Advanced 31 Research Projects Agency (DARPA) and Air Force Research Laboratory, 32 Air Force Materiel Command, USAF, under agreement number 33 F30602-01-2-0537. 34 35 *) Add various sanity checks to asn1_get_length() to reject 36 the ASN1 length bytes if they exceed sizeof(long), will appear 37 negative or the content length exceeds the length of the 38 supplied buffer. 39 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 40 41 *) Assertions for various potential buffer overflows, not known to 42 happen in practice. 43 [Ben Laurie (CHATS)] 44 45 *) Various temporary buffers to hold ASCII versions of integers were 46 too small for 64 bit platforms. (CAN-2002-0655) 47 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 48 49 *) Remote buffer overflow in SSL3 protocol - an attacker could 50 supply an oversized session ID to a client. (CAN-2002-0656) 51 [Ben Laurie (CHATS)] 52 53 *) Remote buffer overflow in SSL2 protocol - an attacker could 54 supply an oversized client master key. (CAN-2002-0656) 55 [Ben Laurie (CHATS)] 56 57 Changes between 0.9.6c and 0.9.6d [9 May 2002] 58 59 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 60 encoded as NULL) with id-dsa-with-sha1. 61 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 62 63 *) Check various X509_...() return values in apps/req.c. 64 [Nils Larsch <nla@trustcenter.de>] 65 66 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 67 an end-of-file condition would erronously be flagged, when the CRLF 68 was just at the end of a processed block. The bug was discovered when 69 processing data through a buffering memory BIO handing the data to a 70 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 71 <ptsekov@syntrex.com> and Nedelcho Stanev. 72 [Lutz Jaenicke] 73 74 *) Implement a countermeasure against a vulnerability recently found 75 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 76 before application data chunks to avoid the use of known IVs 77 with data potentially chosen by the attacker. 78 [Bodo Moeller] 79 80 *) Fix length checks in ssl3_get_client_hello(). 81 [Bodo Moeller] 82 83 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 84 to prevent ssl3_read_internal() from incorrectly assuming that 85 ssl3_read_bytes() found application data while handshake 86 processing was enabled when in fact s->s3->in_read_app_data was 87 merely automatically cleared during the initial handshake. 88 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 89 90 *) Fix object definitions for Private and Enterprise: they were not 91 recognized in their shortname (=lowercase) representation. Extend 92 obj_dat.pl to issue an error when using undefined keywords instead 93 of silently ignoring the problem (Svenning Sorensen 94 <sss@sss.dnsalias.net>). 95 [Lutz Jaenicke] 96 97 *) Fix DH_generate_parameters() so that it works for 'non-standard' 98 generators, i.e. generators other than 2 and 5. (Previously, the 99 code did not properly initialise the 'add' and 'rem' values to 100 BN_generate_prime().) 101 102 In the new general case, we do not insist that 'generator' is 103 actually a primitive root: This requirement is rather pointless; 104 a generator of the order-q subgroup is just as good, if not 105 better. 106 [Bodo Moeller] 107 108 *) Map new X509 verification errors to alerts. Discovered and submitted by 109 Tom Wu <tom@arcot.com>. 110 [Lutz Jaenicke] 111 112 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 113 returning non-zero before the data has been completely received 114 when using non-blocking I/O. 115 [Bodo Moeller; problem pointed out by John Hughes] 116 117 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 118 [Ben Laurie, Lutz Jaenicke] 119 120 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 121 Yoram Zahavi <YoramZ@gilian.com>). 122 [Lutz Jaenicke] 123 124 *) Add information about CygWin 1.3 and on, and preserve proper 125 configuration for the versions before that. 126 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 127 128 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 129 check whether we deal with a copy of a session and do not delete from 130 the cache in this case. Problem reported by "Izhar Shoshani Levi" 131 <izhar@checkpoint.com>. 132 [Lutz Jaenicke] 133 134 *) Do not store session data into the internal session cache, if it 135 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 136 flag is set). Proposed by Aslam <aslam@funk.com>. 137 [Lutz Jaenicke] 138 139 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 140 value is 0. 141 [Richard Levitte] 142 143 *) [In 0.9.6c-engine release:] 144 Fix a crashbug and a logic bug in hwcrhk_load_pubkey() 145 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 146 147 *) Add the configuration target linux-s390x. 148 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 149 150 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 151 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 152 variable as an indication that a ClientHello message has been 153 received. As the flag value will be lost between multiple 154 invocations of ssl3_accept when using non-blocking I/O, the 155 function may not be aware that a handshake has actually taken 156 place, thus preventing a new session from being added to the 157 session cache. 158 159 To avoid this problem, we now set s->new_session to 2 instead of 160 using a local variable. 161 [Lutz Jaenicke, Bodo Moeller] 162 163 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 164 if the SSL_R_LENGTH_MISMATCH error is detected. 165 [Geoff Thorpe, Bodo Moeller] 166 167 *) New 'shared_ldflag' column in Configure platform table. 168 [Richard Levitte] 169 170 *) Fix EVP_CIPHER_mode macro. 171 ["Dan S. Camper" <dan@bti.net>] 172 173 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 174 type, we must throw them away by setting rr->length to 0. 175 [D P Chang <dpc@qualys.com>] 176 177 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 178 179 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 180 <Dominikus.Scherkl@biodata.com>. (The previous implementation 181 worked incorrectly for those cases where range = 10..._2 and 182 3*range is two bits longer than range.) 183 [Bodo Moeller] 184 185 *) Only add signing time to PKCS7 structures if it is not already 186 present. 187 [Steve Henson] 188 189 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 190 OBJ_ld_ce should be OBJ_id_ce. 191 Also some ip-pda OIDs in crypto/objects/objects.txt were 192 incorrect (cf. RFC 3039). 193 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 194 195 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 196 returns early because it has nothing to do. 197 [Andy Schneider <andy.schneider@bjss.co.uk>] 198 199 *) [In 0.9.6c-engine release:] 200 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 201 [Andy Schneider <andy.schneider@bjss.co.uk>] 202 203 *) [In 0.9.6c-engine release:] 204 Add support for Cryptographic Appliance's keyserver technology. 205 (Use engine 'keyclient') 206 [Cryptographic Appliances and Geoff Thorpe] 207 208 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 209 is called via tools/c89.sh because arguments have to be 210 rearranged (all '-L' options must appear before the first object 211 modules). 212 [Richard Shapiro <rshapiro@abinitio.com>] 213 214 *) [In 0.9.6c-engine release:] 215 Add support for Broadcom crypto accelerator cards, backported 216 from 0.9.7. 217 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 218 219 *) [In 0.9.6c-engine release:] 220 Add support for SureWare crypto accelerator cards from 221 Baltimore Technologies. (Use engine 'sureware') 222 [Baltimore Technologies and Mark Cox] 223 224 *) [In 0.9.6c-engine release:] 225 Add support for crypto accelerator cards from Accelerated 226 Encryption Processing, www.aep.ie. (Use engine 'aep') 227 [AEP Inc. and Mark Cox] 228 229 *) Add a configuration entry for gcc on UnixWare. 230 [Gary Benson <gbenson@redhat.com>] 231 232 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 233 messages are stored in a single piece (fixed-length part and 234 variable-length part combined) and fix various bugs found on the way. 235 [Bodo Moeller] 236 237 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 238 instead. BIO_gethostbyname() does not know what timeouts are 239 appropriate, so entries would stay in cache even when they have 240 become invalid. 241 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 242 243 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 244 faced with a pathologically small ClientHello fragment that does 245 not contain client_version: Instead of aborting with an error, 246 simply choose the highest available protocol version (i.e., 247 TLS 1.0 unless it is disabled). In practice, ClientHello 248 messages are never sent like this, but this change gives us 249 strictly correct behaviour at least for TLS. 250 [Bodo Moeller] 251 252 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 253 never resets s->method to s->ctx->method when called from within 254 one of the SSL handshake functions. 255 [Bodo Moeller; problem pointed out by Niko Baric] 256 257 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 258 (sent using the client's version number) if client_version is 259 smaller than the protocol version in use. Also change 260 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 261 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 262 the client will at least see that alert. 263 [Bodo Moeller] 264 265 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 266 correctly. 267 [Bodo Moeller] 268 269 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 270 client receives HelloRequest while in a handshake. 271 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 272 273 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 274 should end in 'break', not 'goto end' which circuments various 275 cleanups done in state SSL_ST_OK. But session related stuff 276 must be disabled for SSL_ST_OK in the case that we just sent a 277 HelloRequest. 278 279 Also avoid some overhead by not calling ssl_init_wbio_buffer() 280 before just sending a HelloRequest. 281 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 282 283 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 284 reveal whether illegal block cipher padding was found or a MAC 285 verification error occured. (Neither SSLerr() codes nor alerts 286 are directly visible to potential attackers, but the information 287 may leak via logfiles.) 288 289 Similar changes are not required for the SSL 2.0 implementation 290 because the number of padding bytes is sent in clear for SSL 2.0, 291 and the extra bytes are just ignored. However ssl/s2_pkt.c 292 failed to verify that the purported number of padding bytes is in 293 the legal range. 294 [Bodo Moeller] 295 296 *) Add OpenUNIX-8 support including shared libraries 297 (Boyd Lynn Gerber <gerberb@zenez.com>). 298 [Lutz Jaenicke] 299 300 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 301 'wristwatch attack' using huge encoding parameters (cf. 302 James H. Manger's CRYPTO 2001 paper). Note that the 303 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 304 encoding parameters and hence was not vulnerable. 305 [Bodo Moeller] 306 307 *) BN_sqr() bug fix. 308 [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 309 310 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 311 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 312 followed by modular reduction. 313 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 314 315 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 316 equivalent based on BN_pseudo_rand() instead of BN_rand(). 317 [Bodo Moeller] 318 319 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 320 This function was broken, as the check for a new client hello message 321 to handle SGC did not allow these large messages. 322 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 323 [Lutz Jaenicke] 324 325 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 326 [Lutz Jaenicke] 327 328 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 329 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 330 [Lutz Jaenicke] 331 332 *) Rework the configuration and shared library support for Tru64 Unix. 333 The configuration part makes use of modern compiler features and 334 still retains old compiler behavior for those that run older versions 335 of the OS. The shared library support part includes a variant that 336 uses the RPATH feature, and is available through the special 337 configuration target "alpha-cc-rpath", which will never be selected 338 automatically. 339 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 340 341 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 342 with the same message size as in ssl3_get_certificate_request(). 343 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 344 messages might inadvertently be reject as too long. 345 [Petr Lampa <lampa@fee.vutbr.cz>] 346 347 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 348 [Andy Polyakov] 349 350 *) Modified SSL library such that the verify_callback that has been set 351 specificly for an SSL object with SSL_set_verify() is actually being 352 used. Before the change, a verify_callback set with this function was 353 ignored and the verify_callback() set in the SSL_CTX at the time of 354 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 355 to allow the necessary settings. 356 [Lutz Jaenicke] 357 358 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 359 explicitly to NULL, as at least on Solaris 8 this seems not always to be 360 done automatically (in contradiction to the requirements of the C 361 standard). This made problems when used from OpenSSH. 362 [Lutz Jaenicke] 363 364 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 365 dh->length and always used 366 367 BN_rand_range(priv_key, dh->p). 368 369 BN_rand_range() is not necessary for Diffie-Hellman, and this 370 specific range makes Diffie-Hellman unnecessarily inefficient if 371 dh->length (recommended exponent length) is much smaller than the 372 length of dh->p. We could use BN_rand_range() if the order of 373 the subgroup was stored in the DH structure, but we only have 374 dh->length. 375 376 So switch back to 377 378 BN_rand(priv_key, l, ...) 379 380 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 381 otherwise. 382 [Bodo Moeller] 383 384 *) In 385 386 RSA_eay_public_encrypt 387 RSA_eay_private_decrypt 388 RSA_eay_private_encrypt (signing) 389 RSA_eay_public_decrypt (signature verification) 390 391 (default implementations for RSA_public_encrypt, 392 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 393 always reject numbers >= n. 394 [Bodo Moeller] 395 396 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 397 to synchronize access to 'locking_thread'. This is necessary on 398 systems where access to 'locking_thread' (an 'unsigned long' 399 variable) is not atomic. 400 [Bodo Moeller] 401 402 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 403 *before* setting the 'crypto_lock_rand' flag. The previous code had 404 a race condition if 0 is a valid thread ID. 405 [Travis Vitek <vitek@roguewave.com>] 406 407 *) Add support for shared libraries under Irix. 408 [Albert Chin-A-Young <china@thewrittenword.com>] 409 410 *) Add configuration option to build on Linux on both big-endian and 411 little-endian MIPS. 412 [Ralf Baechle <ralf@uni-koblenz.de>] 413 414 *) Add the possibility to create shared libraries on HP-UX. 415 [Richard Levitte] 416 417 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 418 419 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 420 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 421 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 422 PRNG state recovery was possible based on the output of 423 one PRNG request appropriately sized to gain knowledge on 424 'md' followed by enough consecutive 1-byte PRNG requests 425 to traverse all of 'state'. 426 427 1. When updating 'md_local' (the current thread's copy of 'md') 428 during PRNG output generation, hash all of the previous 429 'md_local' value, not just the half used for PRNG output. 430 431 2. Make the number of bytes from 'state' included into the hash 432 independent from the number of PRNG bytes requested. 433 434 The first measure alone would be sufficient to avoid 435 Markku-Juhani's attack. (Actually it had never occurred 436 to me that the half of 'md_local' used for chaining was the 437 half from which PRNG output bytes were taken -- I had always 438 assumed that the secret half would be used.) The second 439 measure makes sure that additional data from 'state' is never 440 mixed into 'md_local' in small portions; this heuristically 441 further strengthens the PRNG. 442 [Bodo Moeller] 443 444 *) Fix crypto/bn/asm/mips3.s. 445 [Andy Polyakov] 446 447 *) When only the key is given to "enc", the IV is undefined. Print out 448 an error message in this case. 449 [Lutz Jaenicke] 450 451 *) Handle special case when X509_NAME is empty in X509 printing routines. 452 [Steve Henson] 453 454 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 455 positive and less than q. 456 [Bodo Moeller] 457 458 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 459 used: it isn't thread safe and the add_lock_callback should handle 460 that itself. 461 [Paul Rose <Paul.Rose@bridge.com>] 462 463 *) Verify that incoming data obeys the block size in 464 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 465 [Bodo Moeller] 466 467 *) Fix OAEP check. 468 [Ulf M�ller, Bodo M�ller] 469 470 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 471 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 472 when fixing the server behaviour for backwards-compatible 'client 473 hello' messages. (Note that the attack is impractical against 474 SSL 3.0 and TLS 1.0 anyway because length and version checking 475 means that the probability of guessing a valid ciphertext is 476 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 477 paper.) 478 479 Before 0.9.5, the countermeasure (hide the error by generating a 480 random 'decryption result') did not work properly because 481 ERR_clear_error() was missing, meaning that SSL_get_error() would 482 detect the supposedly ignored error. 483 484 Both problems are now fixed. 485 [Bodo Moeller] 486 487 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 488 (previously it was 1024). 489 [Bodo Moeller] 490 491 *) Fix for compatibility mode trust settings: ignore trust settings 492 unless some valid trust or reject settings are present. 493 [Steve Henson] 494 495 *) Fix for blowfish EVP: its a variable length cipher. 496 [Steve Henson] 497 498 *) Fix various bugs related to DSA S/MIME verification. Handle missing 499 parameters in DSA public key structures and return an error in the 500 DSA routines if parameters are absent. 501 [Steve Henson] 502 503 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 504 in the current directory if neither $RANDFILE nor $HOME was set. 505 RAND_file_name() in 0.9.6a returned NULL in this case. This has 506 caused some confusion to Windows users who haven't defined $HOME. 507 Thus RAND_file_name() is changed again: e_os.h can define a 508 DEFAULT_HOME, which will be used if $HOME is not set. 509 For Windows, we use "C:"; on other platforms, we still require 510 environment variables. 511 512 *) Move 'if (!initialized) RAND_poll()' into regions protected by 513 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 514 having multiple threads call RAND_poll() concurrently. 515 [Bodo Moeller] 516 517 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 518 combination of a flag and a thread ID variable. 519 Otherwise while one thread is in ssleay_rand_bytes (which sets the 520 flag), *other* threads can enter ssleay_add_bytes without obeying 521 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 522 that they do not hold after the first thread unsets add_do_not_lock). 523 [Bodo Moeller] 524 525 *) Change bctest again: '-x' expressions are not available in all 526 versions of 'test'. 527 [Bodo Moeller] 528 529 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 530 531 *) Fix a couple of memory leaks in PKCS7_dataDecode() 532 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 533 534 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 535 the default extension for executables, if any. Also, make the perl 536 scripts that use symlink() to test if it really exists and use "cp" 537 if it doesn't. All this made OpenSSL compilable and installable in 538 CygWin. 539 [Richard Levitte] 540 541 *) Fix for asn1_GetSequence() for indefinite length constructed data. 542 If SEQUENCE is length is indefinite just set c->slen to the total 543 amount of data available. 544 [Steve Henson, reported by shige@FreeBSD.org] 545 [This change does not apply to 0.9.7.] 546 547 *) Change bctest to avoid here-documents inside command substitution 548 (workaround for FreeBSD /bin/sh bug). 549 For compatibility with Ultrix, avoid shell functions (introduced 550 in the bctest version that searches along $PATH). 551 [Bodo Moeller] 552 553 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 554 with des_encrypt() defined on some operating systems, like Solaris 555 and UnixWare. 556 [Richard Levitte] 557 558 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 559 On the Importance of Eliminating Errors in Cryptographic 560 Computations, J. Cryptology 14 (2001) 2, 101-119, 561 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 562 [Ulf Moeller] 563 564 *) MIPS assembler BIGNUM division bug fix. 565 [Andy Polyakov] 566 567 *) Disabled incorrect Alpha assembler code. 568 [Richard Levitte] 569 570 *) Fix PKCS#7 decode routines so they correctly update the length 571 after reading an EOC for the EXPLICIT tag. 572 [Steve Henson] 573 [This change does not apply to 0.9.7.] 574 575 *) Fix bug in PKCS#12 key generation routines. This was triggered 576 if a 3DES key was generated with a 0 initial byte. Include 577 PKCS12_BROKEN_KEYGEN compilation option to retain the old 578 (but broken) behaviour. 579 [Steve Henson] 580 581 *) Enhance bctest to search for a working bc along $PATH and print 582 it when found. 583 [Tim Rice <tim@multitalents.net> via Richard Levitte] 584 585 *) Fix memory leaks in err.c: free err_data string if necessary; 586 don't write to the wrong index in ERR_set_error_data. 587 [Bodo Moeller] 588 589 *) Implement ssl23_peek (analogous to ssl23_read), which previously 590 did not exist. 591 [Bodo Moeller] 592 593 *) Replace rdtsc with _emit statements for VC++ version 5. 594 [Jeremy Cooper <jeremy@baymoo.org>] 595 596 *) Make it possible to reuse SSLv2 sessions. 597 [Richard Levitte] 598 599 *) In copy_email() check for >= 0 as a return value for 600 X509_NAME_get_index_by_NID() since 0 is a valid index. 601 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 602 603 *) Avoid coredump with unsupported or invalid public keys by checking if 604 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 605 PKCS7_verify() fails with non detached data. 606 [Steve Henson] 607 608 *) Don't use getenv in library functions when run as setuid/setgid. 609 New function OPENSSL_issetugid(). 610 [Ulf Moeller] 611 612 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 613 due to incorrect handling of multi-threading: 614 615 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 616 617 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 618 619 3. Count how many times MemCheck_off() has been called so that 620 nested use can be treated correctly. This also avoids 621 inband-signalling in the previous code (which relied on the 622 assumption that thread ID 0 is impossible). 623 [Bodo Moeller] 624 625 *) Add "-rand" option also to s_client and s_server. 626 [Lutz Jaenicke] 627 628 *) Fix CPU detection on Irix 6.x. 629 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 630 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 631 632 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 633 was empty. 634 [Steve Henson] 635 [This change does not apply to 0.9.7.] 636 637 *) Use the cached encoding of an X509_NAME structure rather than 638 copying it. This is apparently the reason for the libsafe "errors" 639 but the code is actually correct. 640 [Steve Henson] 641 642 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 643 Bleichenbacher's DSA attack. 644 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 645 to be set and top=0 forces the highest bit to be set; top=-1 is new 646 and leaves the highest bit random. 647 [Ulf Moeller, Bodo Moeller] 648 649 *) In the NCONF_...-based implementations for CONF_... queries 650 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 651 a temporary CONF structure with the data component set to NULL 652 (which gives segmentation faults in lh_retrieve). 653 Instead, use NULL for the CONF pointer in CONF_get_string and 654 CONF_get_number (which may use environment variables) and directly 655 return NULL from CONF_get_section. 656 [Bodo Moeller] 657 658 *) Fix potential buffer overrun for EBCDIC. 659 [Ulf Moeller] 660 661 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 662 keyUsage if basicConstraints absent for a CA. 663 [Steve Henson] 664 665 *) Make SMIME_write_PKCS7() write mail header values with a format that 666 is more generally accepted (no spaces before the semicolon), since 667 some programs can't parse those values properly otherwise. Also make 668 sure BIO's that break lines after each write do not create invalid 669 headers. 670 [Richard Levitte] 671 672 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 673 macros previously used would not encode an empty SEQUENCE OF 674 and break the signature. 675 [Steve Henson] 676 [This change does not apply to 0.9.7.] 677 678 *) Zero the premaster secret after deriving the master secret in 679 DH ciphersuites. 680 [Steve Henson] 681 682 *) Add some EVP_add_digest_alias registrations (as found in 683 OpenSSL_add_all_digests()) to SSL_library_init() 684 aka OpenSSL_add_ssl_algorithms(). This provides improved 685 compatibility with peers using X.509 certificates 686 with unconventional AlgorithmIdentifier OIDs. 687 [Bodo Moeller] 688 689 *) Fix for Irix with NO_ASM. 690 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 691 692 *) ./config script fixes. 693 [Ulf Moeller, Richard Levitte] 694 695 *) Fix 'openssl passwd -1'. 696 [Bodo Moeller] 697 698 *) Change PKCS12_key_gen_asc() so it can cope with non null 699 terminated strings whose length is passed in the passlen 700 parameter, for example from PEM callbacks. This was done 701 by adding an extra length parameter to asc2uni(). 702 [Steve Henson, reported by <oddissey@samsung.co.kr>] 703 704 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 705 call failed, free the DSA structure. 706 [Bodo Moeller] 707 708 *) Fix to uni2asc() to cope with zero length Unicode strings. 709 These are present in some PKCS#12 files. 710 [Steve Henson] 711 712 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 713 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 714 when writing a 32767 byte record. 715 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 716 717 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 718 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 719 720 (RSA objects have a reference count access to which is protected 721 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 722 so they are meant to be shared between threads.) 723 [Bodo Moeller, Geoff Thorpe; original patch submitted by 724 "Reddie, Steven" <Steven.Reddie@ca.com>] 725 726 *) Fix a deadlock in CRYPTO_mem_leaks(). 727 [Bodo Moeller] 728 729 *) Use better test patterns in bntest. 730 [Ulf M�ller] 731 732 *) rand_win.c fix for Borland C. 733 [Ulf M�ller] 734 735 *) BN_rshift bugfix for n == 0. 736 [Bodo Moeller] 737 738 *) Add a 'bctest' script that checks for some known 'bc' bugs 739 so that 'make test' does not abort just because 'bc' is broken. 740 [Bodo Moeller] 741 742 *) Store verify_result within SSL_SESSION also for client side to 743 avoid potential security hole. (Re-used sessions on the client side 744 always resulted in verify_result==X509_V_OK, not using the original 745 result of the server certificate verification.) 746 [Lutz Jaenicke] 747 748 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 749 SSL3_RT_APPLICATION_DATA, return 0. 750 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 751 [Bodo Moeller] 752 753 *) Fix SSL_peek: 754 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 755 releases, have been re-implemented by renaming the previous 756 implementations of ssl2_read and ssl3_read to ssl2_read_internal 757 and ssl3_read_internal, respectively, and adding 'peek' parameters 758 to them. The new ssl[23]_{read,peek} functions are calls to 759 ssl[23]_read_internal with the 'peek' flag set appropriately. 760 A 'peek' parameter has also been added to ssl3_read_bytes, which 761 does the actual work for ssl3_read_internal. 762 [Bodo Moeller] 763 764 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 765 the method-specific "init()" handler. Also clean up ex_data after 766 calling the method-specific "finish()" handler. Previously, this was 767 happening the other way round. 768 [Geoff Thorpe] 769 770 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 771 The previous value, 12, was not always sufficient for BN_mod_exp(). 772 [Bodo Moeller] 773 774 *) Make sure that shared libraries get the internal name engine with 775 the full version number and not just 0. This should mark the 776 shared libraries as not backward compatible. Of course, this should 777 be changed again when we can guarantee backward binary compatibility. 778 [Richard Levitte] 779 780 *) Fix typo in get_cert_by_subject() in by_dir.c 781 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 782 783 *) Rework the system to generate shared libraries: 784 785 - Make note of the expected extension for the shared libraries and 786 if there is a need for symbolic links from for example libcrypto.so.0 787 to libcrypto.so.0.9.7. There is extended info in Configure for 788 that. 789 790 - Make as few rebuilds of the shared libraries as possible. 791 792 - Still avoid linking the OpenSSL programs with the shared libraries. 793 794 - When installing, install the shared libraries separately from the 795 static ones. 796 [Richard Levitte] 797 798 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 799 800 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 801 and not in SSL_clear because the latter is also used by the 802 accept/connect functions; previously, the settings made by 803 SSL_set_read_ahead would be lost during the handshake. 804 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 805 806 *) Correct util/mkdef.pl to be selective about disabled algorithms. 807 Previously, it would create entries for disableed algorithms no 808 matter what. 809 [Richard Levitte] 810 811 *) Added several new manual pages for SSL_* function. 812 [Lutz Jaenicke] 813 814 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 815 816 *) In ssl23_get_client_hello, generate an error message when faced 817 with an initial SSL 3.0/TLS record that is too small to contain the 818 first two bytes of the ClientHello message, i.e. client_version. 819 (Note that this is a pathologic case that probably has never happened 820 in real life.) The previous approach was to use the version number 821 from the record header as a substitute; but our protocol choice 822 should not depend on that one because it is not authenticated 823 by the Finished messages. 824 [Bodo Moeller] 825 826 *) More robust randomness gathering functions for Windows. 827 [Jeffrey Altman <jaltman@columbia.edu>] 828 829 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 830 not set then we don't setup the error code for issuer check errors 831 to avoid possibly overwriting other errors which the callback does 832 handle. If an application does set the flag then we assume it knows 833 what it is doing and can handle the new informational codes 834 appropriately. 835 [Steve Henson] 836 837 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 838 a general "ANY" type, as such it should be able to decode anything 839 including tagged types. However it didn't check the class so it would 840 wrongly interpret tagged types in the same way as their universal 841 counterpart and unknown types were just rejected. Changed so that the 842 tagged and unknown types are handled in the same way as a SEQUENCE: 843 that is the encoding is stored intact. There is also a new type 844 "V_ASN1_OTHER" which is used when the class is not universal, in this 845 case we have no idea what the actual type is so we just lump them all 846 together. 847 [Steve Henson] 848 849 *) On VMS, stdout may very well lead to a file that is written to 850 in a record-oriented fashion. That means that every write() will 851 write a separate record, which will be read separately by the 852 programs trying to read from it. This can be very confusing. 853 854 The solution is to put a BIO filter in the way that will buffer 855 text until a linefeed is reached, and then write everything a 856 line at a time, so every record written will be an actual line, 857 not chunks of lines and not (usually doesn't happen, but I've 858 seen it once) several lines in one record. BIO_f_linebuffer() is 859 the answer. 860 861 Currently, it's a VMS-only method, because that's where it has 862 been tested well enough. 863 [Richard Levitte] 864 865 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 866 it can return incorrect results. 867 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 868 but it was in 0.9.6-beta[12].) 869 [Bodo Moeller] 870 871 *) Disable the check for content being present when verifying detached 872 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 873 include zero length content when signing messages. 874 [Steve Henson] 875 876 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 877 BIO_ctrl (for BIO pairs). 878 [Bodo M�ller] 879 880 *) Add DSO method for VMS. 881 [Richard Levitte] 882 883 *) Bug fix: Montgomery multiplication could produce results with the 884 wrong sign. 885 [Ulf M�ller] 886 887 *) Add RPM specification openssl.spec and modify it to build three 888 packages. The default package contains applications, application 889 documentation and run-time libraries. The devel package contains 890 include files, static libraries and function documentation. The 891 doc package contains the contents of the doc directory. The original 892 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 893 [Richard Levitte] 894 895 *) Add a large number of documentation files for many SSL routines. 896 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 897 898 *) Add a configuration entry for Sony News 4. 899 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 900 901 *) Don't set the two most significant bits to one when generating a 902 random number < q in the DSA library. 903 [Ulf M�ller] 904 905 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 906 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 907 the underlying transport is blocking) if a handshake took place. 908 (The default behaviour is needed by applications such as s_client 909 and s_server that use select() to determine when to use SSL_read; 910 but for applications that know in advance when to expect data, it 911 just makes things more complicated.) 912 [Bodo Moeller] 913 914 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 915 from EGD. 916 [Ben Laurie] 917 918 *) Add a few more EBCDIC conditionals that make `req' and `x509' 919 work better on such systems. 920 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 921 922 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 923 Update PKCS12_parse() so it copies the friendlyName and the 924 keyid to the certificates aux info. 925 [Steve Henson] 926 927 *) Fix bug in PKCS7_verify() which caused an infinite loop 928 if there was more than one signature. 929 [Sven Uszpelkat <su@celocom.de>] 930 931 *) Major change in util/mkdef.pl to include extra information 932 about each symbol, as well as presentig variables as well 933 as functions. This change means that there's n more need 934 to rebuild the .num files when some algorithms are excluded. 935 [Richard Levitte] 936 937 *) Allow the verify time to be set by an application, 938 rather than always using the current time. 939 [Steve Henson] 940 941 *) Phase 2 verify code reorganisation. The certificate 942 verify code now looks up an issuer certificate by a 943 number of criteria: subject name, authority key id 944 and key usage. It also verifies self signed certificates 945 by the same criteria. The main comparison function is 946 X509_check_issued() which performs these checks. 947 948 Lot of changes were necessary in order to support this 949 without completely rewriting the lookup code. 950 951 Authority and subject key identifier are now cached. 952 953 The LHASH 'certs' is X509_STORE has now been replaced 954 by a STACK_OF(X509_OBJECT). This is mainly because an 955 LHASH can't store or retrieve multiple objects with 956 the same hash value. 957 958 As a result various functions (which were all internal 959 use only) have changed to handle the new X509_STORE 960 structure. This will break anything that messed round 961 with X509_STORE internally. 962 963 The functions X509_STORE_add_cert() now checks for an 964 exact match, rather than just subject name. 965 966 The X509_STORE API doesn't directly support the retrieval 967 of multiple certificates matching a given criteria, however 968 this can be worked round by performing a lookup first 969 (which will fill the cache with candidate certificates) 970 and then examining the cache for matches. This is probably 971 the best we can do without throwing out X509_LOOKUP 972 entirely (maybe later...). 973 974 The X509_VERIFY_CTX structure has been enhanced considerably. 975 976 All certificate lookup operations now go via a get_issuer() 977 callback. Although this currently uses an X509_STORE it 978 can be replaced by custom lookups. This is a simple way 979 to bypass the X509_STORE hackery necessary to make this 980 work and makes it possible to use more efficient techniques 981 in future. A very simple version which uses a simple 982 STACK for its trusted certificate store is also provided 983 using X509_STORE_CTX_trusted_stack(). 984 985 The verify_cb() and verify() callbacks now have equivalents 986 in the X509_STORE_CTX structure. 987 988 X509_STORE_CTX also has a 'flags' field which can be used 989 to customise the verify behaviour. 990 [Steve Henson] 991 992 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 993 excludes S/MIME capabilities. 994 [Steve Henson] 995 996 *) When a certificate request is read in keep a copy of the 997 original encoding of the signed data and use it when outputing 998 again. Signatures then use the original encoding rather than 999 a decoded, encoded version which may cause problems if the 1000 request is improperly encoded. 1001 [Steve Henson] 1002 1003 *) For consistency with other BIO_puts implementations, call 1004 buffer_write(b, ...) directly in buffer_puts instead of calling 1005 BIO_write(b, ...). 1006 1007 In BIO_puts, increment b->num_write as in BIO_write. 1008 [Peter.Sylvester@EdelWeb.fr] 1009 1010 *) Fix BN_mul_word for the case where the word is 0. (We have to use 1011 BN_zero, we may not return a BIGNUM with an array consisting of 1012 words set to zero.) 1013 [Bodo Moeller] 1014 1015 *) Avoid calling abort() from within the library when problems are 1016 detected, except if preprocessor symbols have been defined 1017 (such as REF_CHECK, BN_DEBUG etc.). 1018 [Bodo Moeller] 1019 1020 *) New openssl application 'rsautl'. This utility can be 1021 used for low level RSA operations. DER public key 1022 BIO/fp routines also added. 1023 [Steve Henson] 1024 1025 *) New Configure entry and patches for compiling on QNX 4. 1026 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 1027 1028 *) A demo state-machine implementation was sponsored by 1029 Nuron (http://www.nuron.com/) and is now available in 1030 demos/state_machine. 1031 [Ben Laurie] 1032 1033 *) New options added to the 'dgst' utility for signature 1034 generation and verification. 1035 [Steve Henson] 1036 1037 *) Unrecognized PKCS#7 content types are now handled via a 1038 catch all ASN1_TYPE structure. This allows unsupported 1039 types to be stored as a "blob" and an application can 1040 encode and decode it manually. 1041 [Steve Henson] 1042 1043 *) Fix various signed/unsigned issues to make a_strex.c 1044 compile under VC++. 1045 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 1046 1047 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 1048 length if passed a buffer. ASN1_INTEGER_to_BN failed 1049 if passed a NULL BN and its argument was negative. 1050 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 1051 1052 *) Modification to PKCS#7 encoding routines to output definite 1053 length encoding. Since currently the whole structures are in 1054 memory there's not real point in using indefinite length 1055 constructed encoding. However if OpenSSL is compiled with 1056 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 1057 [Steve Henson] 1058 1059 *) Added BIO_vprintf() and BIO_vsnprintf(). 1060 [Richard Levitte] 1061 1062 *) Added more prefixes to parse for in the the strings written 1063 through a logging bio, to cover all the levels that are available 1064 through syslog. The prefixes are now: 1065 1066 PANIC, EMERG, EMR => LOG_EMERG 1067 ALERT, ALR => LOG_ALERT 1068 CRIT, CRI => LOG_CRIT 1069 ERROR, ERR => LOG_ERR 1070 WARNING, WARN, WAR => LOG_WARNING 1071 NOTICE, NOTE, NOT => LOG_NOTICE 1072 INFO, INF => LOG_INFO 1073 DEBUG, DBG => LOG_DEBUG 1074 1075 and as before, if none of those prefixes are present at the 1076 beginning of the string, LOG_ERR is chosen. 1077 1078 On Win32, the LOG_* levels are mapped according to this: 1079 1080 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 1081 LOG_WARNING => EVENTLOG_WARNING_TYPE 1082 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 1083 1084 [Richard Levitte] 1085 1086 *) Made it possible to reconfigure with just the configuration 1087 argument "reconf" or "reconfigure". The command line arguments 1088 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 1089 and are retrieved from there when reconfiguring. 1090 [Richard Levitte] 1091 1092 *) MD4 implemented. 1093 [Assar Westerlund <assar@sics.se>, Richard Levitte] 1094 1095 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 1096 [Richard Levitte] 1097 1098 *) The obj_dat.pl script was messing up the sorting of object 1099 names. The reason was that it compared the quoted version 1100 of strings as a result "OCSP" > "OCSP Signing" because 1101 " > SPACE. Changed script to store unquoted versions of 1102 names and add quotes on output. It was also omitting some 1103 names from the lookup table if they were given a default 1104 value (that is if SN is missing it is given the same 1105 value as LN and vice versa), these are now added on the 1106 grounds that if an object has a name we should be able to 1107 look it up. Finally added warning output when duplicate 1108 short or long names are found. 1109 [Steve Henson] 1110 1111 *) Changes needed for Tandem NSK. 1112 [Scott Uroff <scott@xypro.com>] 1113 1114 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 1115 RSA_padding_check_SSLv23(), special padding was never detected 1116 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 1117 version rollback attacks was not effective. 1118 1119 In s23_clnt.c, don't use special rollback-attack detection padding 1120 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 1121 client; similarly, in s23_srvr.c, don't do the rollback check if 1122 SSL 2.0 is the only protocol enabled in the server. 1123 [Bodo Moeller] 1124 1125 *) Make it possible to get hexdumps of unprintable data with 'openssl 1126 asn1parse'. By implication, the functions ASN1_parse_dump() and 1127 BIO_dump_indent() are added. 1128 [Richard Levitte] 1129 1130 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 1131 these print out strings and name structures based on various 1132 flags including RFC2253 support and proper handling of 1133 multibyte characters. Added options to the 'x509' utility 1134 to allow the various flags to be set. 1135 [Steve Henson] 1136 1137 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 1138 Also change the functions X509_cmp_current_time() and 1139 X509_gmtime_adj() work with an ASN1_TIME structure, 1140 this will enable certificates using GeneralizedTime in validity 1141 dates to be checked. 1142 [Steve Henson] 1143 1144 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 1145 negative public key encodings) on by default, 1146 NO_NEG_PUBKEY_BUG can be set to disable it. 1147 [Steve Henson] 1148 1149 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 1150 content octets. An i2c_ASN1_OBJECT is unnecessary because 1151 the encoding can be trivially obtained from the structure. 1152 [Steve Henson] 1153 1154 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 1155 not read locks (CRYPTO_r_[un]lock). 1156 [Bodo Moeller] 1157 1158 *) A first attempt at creating official support for shared 1159 libraries through configuration. I've kept it so the 1160 default is static libraries only, and the OpenSSL programs 1161 are always statically linked for now, but there are 1162 preparations for dynamic linking in place. 1163 This has been tested on Linux and Tru64. 1164 [Richard Levitte] 1165 1166 *) Randomness polling function for Win9x, as described in: 1167 Peter Gutmann, Software Generation of Practically Strong 1168 Random Numbers. 1169 [Ulf M�ller] 1170 1171 *) Fix so PRNG is seeded in req if using an already existing 1172 DSA key. 1173 [Steve Henson] 1174 1175 *) New options to smime application. -inform and -outform 1176 allow alternative formats for the S/MIME message including 1177 PEM and DER. The -content option allows the content to be 1178 specified separately. This should allow things like Netscape 1179 form signing output easier to verify. 1180 [Steve Henson] 1181 1182 *) Fix the ASN1 encoding of tags using the 'long form'. 1183 [Steve Henson] 1184 1185 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 1186 STRING types. These convert content octets to and from the 1187 underlying type. The actual tag and length octets are 1188 already assumed to have been read in and checked. These 1189 are needed because all other string types have virtually 1190 identical handling apart from the tag. By having versions 1191 of the ASN1 functions that just operate on content octets 1192 IMPLICIT tagging can be handled properly. It also allows 1193 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 1194 and ASN1_INTEGER are identical apart from the tag. 1195 [Steve Henson] 1196 1197 *) Change the handling of OID objects as follows: 1198 1199 - New object identifiers are inserted in objects.txt, following 1200 the syntax given in objects.README. 1201 - objects.pl is used to process obj_mac.num and create a new 1202 obj_mac.h. 1203 - obj_dat.pl is used to create a new obj_dat.h, using the data in 1204 obj_mac.h. 1205 1206 This is currently kind of a hack, and the perl code in objects.pl 1207 isn't very elegant, but it works as I intended. The simplest way 1208 to check that it worked correctly is to look in obj_dat.h and 1209 check the array nid_objs and make sure the objects haven't moved 1210 around (this is important!). Additions are OK, as well as 1211 consistent name changes. 1212 [Richard Levitte] 1213 1214 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 1215 [Bodo Moeller] 1216 1217 *) Addition of the command line parameter '-rand file' to 'openssl req'. 1218 The given file adds to whatever has already been seeded into the 1219 random pool through the RANDFILE configuration file option or 1220 environment variable, or the default random state file. 1221 [Richard Levitte] 1222 1223 *) mkstack.pl now sorts each macro group into lexical order. 1224 Previously the output order depended on the order the files 1225 appeared in the directory, resulting in needless rewriting 1226 of safestack.h . 1227 [Steve Henson] 1228 1229 *) Patches to make OpenSSL compile under Win32 again. Mostly 1230 work arounds for the VC++ problem that it treats func() as 1231 func(void). Also stripped out the parts of mkdef.pl that 1232 added extra typesafe functions: these no longer exist. 1233 [Steve Henson] 1234 1235 *) Reorganisation of the stack code. The macros are now all 1236 collected in safestack.h . Each macro is defined in terms of 1237 a "stack macro" of the form SKM_<name>(type, a, b). The 1238 DEBUG_SAFESTACK is now handled in terms of function casts, 1239 this has the advantage of retaining type safety without the 1240 use of additional functions. If DEBUG_SAFESTACK is not defined 1241 then the non typesafe macros are used instead. Also modified the 1242 mkstack.pl script to handle the new form. Needs testing to see 1243 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 1244 the default if no major problems. Similar behaviour for ASN1_SET_OF 1245 and PKCS12_STACK_OF. 1246 [Steve Henson] 1247 1248 *) When some versions of IIS use the 'NET' form of private key the 1249 key derivation algorithm is different. Normally MD5(password) is 1250 used as a 128 bit RC4 key. In the modified case 1251 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 1252 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 1253 as the old Netscape_RSA functions except they have an additional 1254 'sgckey' parameter which uses the modified algorithm. Also added 1255 an -sgckey command line option to the rsa utility. Thanks to 1256 Adrian Peck <bertie@ncipher.com> for posting details of the modified 1257 algorithm to openssl-dev. 1258 [Steve Henson] 1259 1260 *) The evp_local.h macros were using 'c.##kname' which resulted in 1261 invalid expansion on some systems (SCO 5.0.5 for example). 1262 Corrected to 'c.kname'. 1263 [Phillip Porch <root@theporch.com>] 1264 1265 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 1266 a STACK of email addresses from a certificate or request, these look 1267 in the subject name and the subject alternative name extensions and 1268 omit any duplicate addresses. 1269 [Steve Henson] 1270 1271 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 1272 This makes DSA verification about 2 % faster. 1273 [Bodo Moeller] 1274 1275 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 1276 (meaning that now 2^5 values will be precomputed, which is only 4 KB 1277 plus overhead for 1024 bit moduli). 1278 This makes exponentiations about 0.5 % faster for 1024 bit 1279 exponents (as measured by "openssl speed rsa2048"). 1280 [Bodo Moeller] 1281 1282 *) Rename memory handling macros to avoid conflicts with other 1283 software: 1284 Malloc => OPENSSL_malloc 1285 Malloc_locked => OPENSSL_malloc_locked 1286 Realloc => OPENSSL_realloc 1287 Free => OPENSSL_free 1288 [Richard Levitte] 1289 1290 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 1291 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 1292 [Bodo Moeller] 1293 1294 *) CygWin32 support. 1295 [John Jarvie <jjarvie@newsguy.com>] 1296 1297 *) The type-safe stack code has been rejigged. It is now only compiled 1298 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 1299 by default all type-specific stack functions are "#define"d back to 1300 standard stack functions. This results in more streamlined output 1301 but retains the type-safety checking possibilities of the original 1302 approach. 1303 [Geoff Thorpe] 1304 1305 *) The STACK code has been cleaned up, and certain type declarations 1306 that didn't make a lot of sense have been brought in line. This has 1307 also involved a cleanup of sorts in safestack.h to more correctly 1308 map type-safe stack functions onto their plain stack counterparts. 1309 This work has also resulted in a variety of "const"ifications of 1310 lots of the code, especially "_cmp" operations which should normally 1311 be prototyped with "const" parameters anyway. 1312 [Geoff Thorpe] 1313 1314 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 1315 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 1316 (The PRNG state consists of two parts, the large pool 'state' and 'md', 1317 where all of 'md' is used each time the PRNG is used, but 'state' 1318 is used only indexed by a cyclic counter. As entropy may not be 1319 well distributed from the beginning, 'md' is important as a 1320 chaining variable. However, the output function chains only half 1321 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 1322 all of 'md', and seeding with STATE_SIZE dummy bytes will result 1323 in all of 'state' being rewritten, with the new values depending 1324 on virtually all of 'md'. This overcomes the 80 bit limitation.) 1325 [Bodo Moeller] 1326 1327 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 1328 the handshake is continued after ssl_verify_cert_chain(); 1329 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 1330 can lead to 'unexplainable' connection aborts later. 1331 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 1332 1333 *) Major EVP API cipher revision. 1334 Add hooks for extra EVP features. This allows various cipher 1335 parameters to be set in the EVP interface. Support added for variable 1336 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 1337 setting of RC2 and RC5 parameters. 1338 1339 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 1340 ciphers. 1341 1342 Remove lots of duplicated code from the EVP library. For example *every* 1343 cipher init() function handles the 'iv' in the same way according to the 1344 cipher mode. They also all do nothing if the 'key' parameter is NULL and 1345 for CFB and OFB modes they zero ctx->num. 1346 1347 New functionality allows removal of S/MIME code RC2 hack. 1348 1349 Most of the routines have the same form and so can be declared in terms 1350 of macros. 1351 1352 By shifting this to the top level EVP_CipherInit() it can be removed from 1353 all individual ciphers. If the cipher wants to handle IVs or keys 1354 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 1355 flags. 1356 1357 Change lots of functions like EVP_EncryptUpdate() to now return a 1358 value: although software versions of the algorithms cannot fail 1359 any installed hardware versions can. 1360 [Steve Henson] 1361 1362 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 1363 this option is set, tolerate broken clients that send the negotiated 1364 protocol version number instead of the requested protocol version 1365 number. 1366 [Bodo Moeller] 1367 1368 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 1369 i.e. non-zero for export ciphersuites, zero otherwise. 1370 Previous versions had this flag inverted, inconsistent with 1371 rsa_tmp_cb (..._TMP_RSA_CB). 1372 [Bodo Moeller; problem reported by Amit Chopra] 1373 1374 *) Add missing DSA library text string. Work around for some IIS 1375 key files with invalid SEQUENCE encoding. 1376 [Steve Henson] 1377 1378 *) Add a document (doc/standards.txt) that list all kinds of standards 1379 and so on that are implemented in OpenSSL. 1380 [Richard Levitte] 1381 1382 *) Enhance c_rehash script. Old version would mishandle certificates 1383 with the same subject name hash and wouldn't handle CRLs at all. 1384 Added -fingerprint option to crl utility, to support new c_rehash 1385 features. 1386 [Steve Henson] 1387 1388 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 1389 [Ulf M�ller] 1390 1391 *) Fix for SSL server purpose checking. Server checking was 1392 rejecting certificates which had extended key usage present 1393 but no ssl client purpose. 1394 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 1395 1396 *) Make PKCS#12 code work with no password. The PKCS#12 spec 1397 is a little unclear about how a blank password is handled. 1398 Since the password in encoded as a BMPString with terminating 1399 double NULL a zero length password would end up as just the 1400 double NULL. However no password at all is different and is 1401 handled differently in the PKCS#12 key generation code. NS 1402 treats a blank password as zero length. MSIE treats it as no 1403 password on export: but it will try both on import. We now do 1404 the same: PKCS12_parse() tries zero length and no password if 1405 the password is set to "" or NULL (NULL is now a valid password: 1406 it wasn't before) as does the pkcs12 application. 1407 [Steve Henson] 1408 1409 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 1410 perror when PEM_read_bio_X509_REQ fails, the error message must 1411 be obtained from the error queue. 1412 [Bodo Moeller] 1413 1414 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 1415 it in ERR_remove_state if appropriate, and change ERR_get_state 1416 accordingly to avoid race conditions (this is necessary because 1417 thread_hash is no longer constant once set). 1418 [Bodo Moeller] 1419 1420 *) Bugfix for linux-elf makefile.one. 1421 [Ulf M�ller] 1422 1423 *) RSA_get_default_method() will now cause a default 1424 RSA_METHOD to be chosen if one doesn't exist already. 1425 Previously this was only set during a call to RSA_new() 1426 or RSA_new_method(NULL) meaning it was possible for 1427 RSA_get_default_method() to return NULL. 1428 [Geoff Thorpe] 1429 1430 *) Added native name translation to the existing DSO code 1431 that will convert (if the flag to do so is set) filenames 1432 that are sufficiently small and have no path information 1433 into a canonical native form. Eg. "blah" converted to 1434 "libblah.so" or "blah.dll" etc. 1435 [Geoff Thorpe] 1436 1437 *) New function ERR_error_string_n(e, buf, len) which is like 1438 ERR_error_string(e, buf), but writes at most 'len' bytes 1439 including the 0 terminator. For ERR_error_string_n, 'buf' 1440 may not be NULL. 1441 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 1442 1443 *) CONF library reworked to become more general. A new CONF 1444 configuration file reader "class" is implemented as well as a 1445 new functions (NCONF_*, for "New CONF") to handle it. The now 1446 old CONF_* functions are still there, but are reimplemented to 1447 work in terms of the new functions. Also, a set of functions 1448 to handle the internal storage of the configuration data is 1449 provided to make it easier to write new configuration file 1450 reader "classes" (I can definitely see something reading a 1451 configuration file in XML format, for example), called _CONF_*, 1452 or "the configuration storage API"... 1453 1454 The new configuration file reading functions are: 1455 1456 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 1457 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 1458 1459 NCONF_default, NCONF_WIN32 1460 1461 NCONF_dump_fp, NCONF_dump_bio 1462 1463 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 1464 NCONF_new creates a new CONF object. This works in the same way 1465 as other interfaces in OpenSSL, like the BIO interface. 1466 NCONF_dump_* dump the internal storage of the configuration file, 1467 which is useful for debugging. All other functions take the same 1468 arguments as the old CONF_* functions wth the exception of the 1469 first that must be a `CONF *' instead of a `LHASH *'. 1470 1471 To make it easer to use the new classes with the old CONF_* functions, 1472 the function CONF_set_default_method is provided. 1473 [Richard Levitte] 1474 1475 *) Add '-tls1' option to 'openssl ciphers', which was already 1476 mentioned in the documentation but had not been implemented. 1477 (This option is not yet really useful because even the additional 1478 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 1479 [Bodo Moeller] 1480 1481 *) Initial DSO code added into libcrypto for letting OpenSSL (and 1482 OpenSSL-based applications) load shared libraries and bind to 1483 them in a portable way. 1484 [Geoff Thorpe, with contributions from Richard Levitte] 1485 1486 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 1487 1488 *) Make sure _lrotl and _lrotr are only used with MSVC. 1489 1490 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 1491 (the default implementation of RAND_status). 1492 1493 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 1494 to '-clrext' (= clear extensions), as intended and documented. 1495 [Bodo Moeller; inconsistency pointed out by Michael Attili 1496 <attili@amaxo.com>] 1497 1498 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 1499 was larger than the MD block size. 1500 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 1501 1502 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 1503 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 1504 using the passed key: if the passed key was a private key the result 1505 of X509_print(), for example, would be to print out all the private key 1506 components. 1507 [Steve Henson] 1508 1509 *) des_quad_cksum() byte order bug fix. 1510 [Ulf M�ller, using the problem description in krb4-0.9.7, where 1511 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 1512 1513 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 1514 discouraged. 1515 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 1516 1517 *) For easily testing in shell scripts whether some command 1518 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 1519 returns with exit code 0 iff no command of the given name is available. 1520 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 1521 the output goes to stdout and nothing is printed to stderr. 1522 Additional arguments are always ignored. 1523 1524 Since for each cipher there is a command of the same name, 1525 the 'no-cipher' compilation switches can be tested this way. 1526 1527 ('openssl no-XXX' is not able to detect pseudo-commands such 1528 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 1529 [Bodo Moeller] 1530 1531 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 1532 [Bodo Moeller] 1533 1534 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 1535 is set; it will be thrown away anyway because each handshake creates 1536 its own key. 1537 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 1538 to parameters -- in previous versions (since OpenSSL 0.9.3) the 1539 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 1540 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 1541 [Bodo Moeller] 1542 1543 *) New s_client option -ign_eof: EOF at stdin is ignored, and 1544 'Q' and 'R' lose their special meanings (quit/renegotiate). 1545 This is part of what -quiet does; unlike -quiet, -ign_eof 1546 does not suppress any output. 1547 [Richard Levitte] 1548 1549 *) Add compatibility options to the purpose and trust code. The 1550 purpose X509_PURPOSE_ANY is "any purpose" which automatically 1551 accepts a certificate or CA, this was the previous behaviour, 1552 with all the associated security issues. 1553 1554 X509_TRUST_COMPAT is the old trust behaviour: only and 1555 automatically trust self signed roots in certificate store. A 1556 new trust setting X509_TRUST_DEFAULT is used to specify that 1557 a purpose has no associated trust setting and it should instead 1558 use the value in the default purpose. 1559 [Steve Henson] 1560 1561 *) Fix the PKCS#8 DSA private key code so it decodes keys again 1562 and fix a memory leak. 1563 [Steve Henson] 1564 1565 *) In util/mkerr.pl (which implements 'make errors'), preserve 1566 reason strings from the previous version of the .c file, as 1567 the default to have only downcase letters (and digits) in 1568 automatically generated reasons codes is not always appropriate. 1569 [Bodo Moeller] 1570 1571 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 1572 using strerror. Previously, ERR_reason_error_string() returned 1573 library names as reason strings for SYSerr; but SYSerr is a special 1574 case where small numbers are errno values, not library numbers. 1575 [Bodo Moeller] 1576 1577 *) Add '-dsaparam' option to 'openssl dhparam' application. This 1578 converts DSA parameters into DH parameters. (When creating parameters, 1579 DSA_generate_parameters is used.) 1580 [Bodo Moeller] 1581 1582 *) Include 'length' (recommended exponent length) in C code generated 1583 by 'openssl dhparam -C'. 1584 [Bodo Moeller] 1585 1586 *) The second argument to set_label in perlasm was already being used 1587 so couldn't be used as a "file scope" flag. Moved to third argument 1588 which was free. 1589 [Steve Henson] 1590 1591 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 1592 instead of RAND_bytes for encryption IVs and salts. 1593 [Bodo Moeller] 1594 1595 *) Include RAND_status() into RAND_METHOD instead of implementing 1596 it only for md_rand.c Otherwise replacing the PRNG by calling 1597 RAND_set_rand_method would be impossible. 1598 [Bodo Moeller] 1599 1600 *) Don't let DSA_generate_key() enter an infinite loop if the random 1601 number generation fails. 1602 [Bodo Moeller] 1603 1604 *) New 'rand' application for creating pseudo-random output. 1605 [Bodo Moeller] 1606 1607 *) Added configuration support for Linux/IA64 1608 [Rolf Haberrecker <rolf@suse.de>] 1609 1610 *) Assembler module support for Mingw32. 1611 [Ulf M�ller] 1612 1613 *) Shared library support for HPUX (in shlib/). 1614 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 1615 1616 *) Shared library support for Solaris gcc. 1617 [Lutz Behnke <behnke@trustcenter.de>] 1618 1619 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 1620 1621 *) PKCS7_encrypt() was adding text MIME headers twice because they 1622 were added manually and by SMIME_crlf_copy(). 1623 [Steve Henson] 1624 1625 *) In bntest.c don't call BN_rand with zero bits argument. 1626 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 1627 1628 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 1629 case was implemented. This caused BN_div_recp() to fail occasionally. 1630 [Ulf M�ller] 1631 1632 *) Add an optional second argument to the set_label() in the perl 1633 assembly language builder. If this argument exists and is set 1634 to 1 it signals that the assembler should use a symbol whose 1635 scope is the entire file, not just the current function. This 1636 is needed with MASM which uses the format label:: for this scope. 1637 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 1638 1639 *) Change the ASN1 types so they are typedefs by default. Before 1640 almost all types were #define'd to ASN1_STRING which was causing 1641 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 1642 for example. 1643 [Steve Henson] 1644 1645 *) Change names of new functions to the new get1/get0 naming 1646 convention: After 'get1', the caller owns a reference count 1647 and has to call ..._free; 'get0' returns a pointer to some 1648 data structure without incrementing reference counters. 1649 (Some of the existing 'get' functions increment a reference 1650 counter, some don't.) 1651 Similarly, 'set1' and 'add1' functions increase reference 1652 counters or duplicate objects. 1653 [Steve Henson] 1654 1655 *) Allow for the possibility of temp RSA key generation failure: 1656 the code used to assume it always worked and crashed on failure. 1657 [Steve Henson] 1658 1659 *) Fix potential buffer overrun problem in BIO_printf(). 1660 [Ulf M�ller, using public domain code by Patrick Powell; problem 1661 pointed out by David Sacerdote <das33@cornell.edu>] 1662 1663 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 1664 RAND_egd() and RAND_status(). In the command line application, 1665 the EGD socket can be specified like a seed file using RANDFILE 1666 or -rand. 1667 [Ulf M�ller] 1668 1669 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 1670 Some CAs (e.g. Verisign) distribute certificates in this form. 1671 [Steve Henson] 1672 1673 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 1674 list to exclude them. This means that no special compilation option 1675 is needed to use anonymous DH: it just needs to be included in the 1676 cipher list. 1677 [Steve Henson] 1678 1679 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 1680 EVP_MD_type. The old functionality is available in a new macro called 1681 EVP_MD_md(). Change code that uses it and update docs. 1682 [Steve Henson] 1683 1684 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 1685 where the 'void *' argument is replaced by a function pointer argument. 1686 Previously 'void *' was abused to point to functions, which works on 1687 many platforms, but is not correct. As these functions are usually 1688 called by macros defined in OpenSSL header files, most source code 1689 should work without changes. 1690 [Richard Levitte] 1691 1692 *) <openssl/opensslconf.h> (which is created by Configure) now contains 1693 sections with information on -D... compiler switches used for 1694 compiling the library so that applications can see them. To enable 1695 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 1696 must be defined. E.g., 1697 #define OPENSSL_ALGORITHM_DEFINES 1698 #include <openssl/opensslconf.h> 1699 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 1700 [Richard Levitte, Ulf and Bodo M�ller] 1701 1702 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 1703 record layer. 1704 [Bodo Moeller] 1705 1706 *) Change the 'other' type in certificate aux info to a STACK_OF 1707 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 1708 the required ASN1 format: arbitrary types determined by an OID. 1709 [Steve Henson] 1710 1711 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 1712 argument to 'req'. This is not because the function is newer or 1713 better than others it just uses the work 'NEW' in the certificate 1714 request header lines. Some software needs this. 1715 [Steve Henson] 1716 1717 *) Reorganise password command line arguments: now passwords can be 1718 obtained from various sources. Delete the PEM_cb function and make 1719 it the default behaviour: i.e. if the callback is NULL and the 1720 usrdata argument is not NULL interpret it as a null terminated pass 1721 phrase. If usrdata and the callback are NULL then the pass phrase 1722 is prompted for as usual. 1723 [Steve Henson] 1724 1725 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 1726 the support is automatically enabled. The resulting binaries will 1727 autodetect the card and use it if present. 1728 [Ben Laurie and Compaq Inc.] 1729 1730 *) Work around for Netscape hang bug. This sends certificate request 1731 and server done in one record. Since this is perfectly legal in the 1732 SSL/TLS protocol it isn't a "bug" option and is on by default. See 1733 the bugs/SSLv3 entry for more info. 1734 [Steve Henson] 1735 1736 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 1737 [Andy Polyakov] 1738 1739 *) Add -rand argument to smime and pkcs12 applications and read/write 1740 of seed file. 1741 [Steve Henson] 1742 1743 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 1744 [Bodo Moeller] 1745 1746 *) Add command line password options to the remaining applications. 1747 [Steve Henson] 1748 1749 *) Bug fix for BN_div_recp() for numerators with an even number of 1750 bits. 1751 [Ulf M�ller] 1752 1753 *) More tests in bntest.c, and changed test_bn output. 1754 [Ulf M�ller] 1755 1756 *) ./config recognizes MacOS X now. 1757 [Andy Polyakov] 1758 1759 *) Bug fix for BN_div() when the first words of num and divsor are 1760 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 1761 [Ulf M�ller] 1762 1763 *) Add support for various broken PKCS#8 formats, and command line 1764 options to produce them. 1765 [Steve Henson] 1766 1767 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 1768 get temporary BIGNUMs from a BN_CTX. 1769 [Ulf M�ller] 1770 1771 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 1772 for p == 0. 1773 [Ulf M�ller] 1774 1775 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 1776 include a #define from the old name to the new. The original intent 1777 was that statically linked binaries could for example just call 1778 SSLeay_add_all_ciphers() to just add ciphers to the table and not 1779 link with digests. This never worked becayse SSLeay_add_all_digests() 1780 and SSLeay_add_all_ciphers() were in the same source file so calling 1781 one would link with the other. They are now in separate source files. 1782 [Steve Henson] 1783 1784 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 1785 [Steve Henson] 1786 1787 *) Use a less unusual form of the Miller-Rabin primality test (it used 1788 a binary algorithm for exponentiation integrated into the Miller-Rabin 1789 loop, our standard modexp algorithms are faster). 1790 [Bodo Moeller] 1791 1792 *) Support for the EBCDIC character set completed. 1793 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 1794 1795 *) Source code cleanups: use const where appropriate, eliminate casts, 1796 use void * instead of char * in lhash. 1797 [Ulf M�ller] 1798 1799 *) Bugfix: ssl3_send_server_key_exchange was not restartable 1800 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 1801 this the server could overwrite ephemeral keys that the client 1802 has already seen). 1803 [Bodo Moeller] 1804 1805 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 1806 using 50 iterations of the Rabin-Miller test. 1807 1808 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 1809 iterations of the Rabin-Miller test as required by the appendix 1810 to FIPS PUB 186[-1]) instead of DSA_is_prime. 1811 As BN_is_prime_fasttest includes trial division, DSA parameter 1812 generation becomes much faster. 1813 1814 This implies a change for the callback functions in DSA_is_prime 1815 and DSA_generate_parameters: The callback function is called once 1816 for each positive witness in the Rabin-Miller test, not just 1817 occasionally in the inner loop; and the parameters to the 1818 callback function now provide an iteration count for the outer 1819 loop rather than for the current invocation of the inner loop. 1820 DSA_generate_parameters additionally can call the callback 1821 function with an 'iteration count' of -1, meaning that a 1822 candidate has passed the trial division test (when q is generated 1823 from an application-provided seed, trial division is skipped). 1824 [Bodo Moeller] 1825 1826 *) New function BN_is_prime_fasttest that optionally does trial 1827 division before starting the Rabin-Miller test and has 1828 an additional BN_CTX * argument (whereas BN_is_prime always 1829 has to allocate at least one BN_CTX). 1830 'callback(1, -1, cb_arg)' is called when a number has passed the 1831 trial division stage. 1832 [Bodo Moeller] 1833 1834 *) Fix for bug in CRL encoding. The validity dates weren't being handled 1835 as ASN1_TIME. 1836 [Steve Henson] 1837 1838 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 1839 [Steve Henson] 1840 1841 *) New function BN_pseudo_rand(). 1842 [Ulf M�ller] 1843 1844 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 1845 bignum version of BN_from_montgomery() with the working code from 1846 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 1847 the comments. 1848 [Ulf M�ller] 1849 1850 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 1851 made it impossible to use the same SSL_SESSION data structure in 1852 SSL2 clients in multiple threads. 1853 [Bodo Moeller] 1854 1855 *) The return value of RAND_load_file() no longer counts bytes obtained 1856 by stat(). RAND_load_file(..., -1) is new and uses the complete file 1857 to seed the PRNG (previously an explicit byte count was required). 1858 [Ulf M�ller, Bodo M�ller] 1859 1860 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 1861 used (char *) instead of (void *) and had casts all over the place. 1862 [Steve Henson] 1863 1864 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 1865 [Ulf M�ller] 1866 1867 *) Retain source code compatibility for BN_prime_checks macro: 1868 BN_is_prime(..., BN_prime_checks, ...) now uses 1869 BN_prime_checks_for_size to determine the appropriate number of 1870 Rabin-Miller iterations. 1871 [Ulf M�ller] 1872 1873 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 1874 DH_CHECK_P_NOT_SAFE_PRIME. 1875 (Check if this is true? OpenPGP calls them "strong".) 1876 [Ulf M�ller] 1877 1878 *) Merge the functionality of "dh" and "gendh" programs into a new program 1879 "dhparam". The old programs are retained for now but will handle DH keys 1880 (instead of parameters) in future. 1881 [Steve Henson] 1882 1883 *) Make the ciphers, s_server and s_client programs check the return values 1884 when a new cipher list is set. 1885 [Steve Henson] 1886 1887 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 1888 ciphers. Before when the 56bit ciphers were enabled the sorting was 1889 wrong. 1890 1891 The syntax for the cipher sorting has been extended to support sorting by 1892 cipher-strength (using the strength_bits hard coded in the tables). 1893 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 1894 1895 Fix a bug in the cipher-command parser: when supplying a cipher command 1896 string with an "undefined" symbol (neither command nor alphanumeric 1897 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 1898 an error is flagged. 1899 1900 Due to the strength-sorting extension, the code of the 1901 ssl_create_cipher_list() function was completely rearranged. I hope that 1902 the readability was also increased :-) 1903 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 1904 1905 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 1906 for the first serial number and places 2 in the serial number file. This 1907 avoids problems when the root CA is created with serial number zero and 1908 the first user certificate has the same issuer name and serial number 1909 as the root CA. 1910 [Steve Henson] 1911 1912 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 1913 the new code. Add documentation for this stuff. 1914 [Steve Henson] 1915 1916 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 1917 X509_*() to X509at_*() on the grounds that they don't handle X509 1918 structures and behave in an analagous way to the X509v3 functions: 1919 they shouldn't be called directly but wrapper functions should be used 1920 instead. 1921 1922 So we also now have some wrapper functions that call the X509at functions 1923 when passed certificate requests. (TO DO: similar things can be done with 1924 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 1925 things. Some of these need some d2i or i2d and print functionality 1926 because they handle more complex structures.) 1927 [Steve Henson] 1928 1929 *) Add missing #ifndefs that caused missing symbols when building libssl 1930 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 1931 NO_RSA in ssl/s2*.c. 1932 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 1933 1934 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 1935 has a return value which indicates the quality of the random data 1936 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 1937 error queue. New function RAND_pseudo_bytes() generates output that is 1938 guaranteed to be unique but not unpredictable. RAND_add is like 1939 RAND_seed, but takes an extra argument for an entropy estimate 1940 (RAND_seed always assumes full entropy). 1941 [Ulf M�ller] 1942 1943 *) Do more iterations of Rabin-Miller probable prime test (specifically, 1944 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 1945 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 1946 in crypto/bn/bn_prime.c for the complete table). This guarantees a 1947 false-positive rate of at most 2^-80 for random input. 1948 [Bodo Moeller] 1949 1950 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 1951 [Bodo Moeller] 1952 1953 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 1954 in the 0.9.5 release), this returns the chain 1955 from an X509_CTX structure with a dup of the stack and all 1956 the X509 reference counts upped: so the stack will exist 1957 after X509_CTX_cleanup() has been called. Modify pkcs12.c 1958 to use this. 1959 1960 Also make SSL_SESSION_print() print out the verify return 1961 code. 1962 [Steve Henson] 1963 1964 *) Add manpage for the pkcs12 command. Also change the default 1965 behaviour so MAC iteration counts are used unless the new 1966 -nomaciter option is used. This improves file security and 1967 only older versions of MSIE (4.0 for example) need it. 1968 [Steve Henson] 1969 1970 *) Honor the no-xxx Configure options when creating .DEF files. 1971 [Ulf M�ller] 1972 1973 *) Add PKCS#10 attributes to field table: challengePassword, 1974 unstructuredName and unstructuredAddress. These are taken from 1975 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 1976 international characters are used. 1977 1978 More changes to X509_ATTRIBUTE code: allow the setting of types 1979 based on strings. Remove the 'loc' parameter when adding 1980 attributes because these will be a SET OF encoding which is sorted 1981 in ASN1 order. 1982 [Steve Henson] 1983 1984 *) Initial changes to the 'req' utility to allow request generation 1985 automation. This will allow an application to just generate a template 1986 file containing all the field values and have req construct the 1987 request. 1988 1989 Initial support for X509_ATTRIBUTE handling. Stacks of these are 1990 used all over the place including certificate requests and PKCS#7 1991 structures. They are currently handled manually where necessary with 1992 some primitive wrappers for PKCS#7. The new functions behave in a 1993 manner analogous to the X509 extension functions: they allow 1994 attributes to be looked up by NID and added. 1995 1996 Later something similar to the X509V3 code would be desirable to 1997 automatically handle the encoding, decoding and printing of the 1998 more complex types. The string types like challengePassword can 1999 be handled by the string table functions. 2000 2001 Also modified the multi byte string table handling. Now there is 2002 a 'global mask' which masks out certain types. The table itself 2003 can use the flag STABLE_NO_MASK to ignore the mask setting: this 2004 is useful when for example there is only one permissible type 2005 (as in countryName) and using the mask might result in no valid 2006 types at all. 2007 [Steve Henson] 2008 2009 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 2010 SSL_get_peer_finished to allow applications to obtain the latest 2011 Finished messages sent to the peer or expected from the peer, 2012 respectively. (SSL_get_peer_finished is usually the Finished message 2013 actually received from the peer, otherwise the protocol will be aborted.) 2014 2015 As the Finished message are message digests of the complete handshake 2016 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 2017 be used for external authentication procedures when the authentication 2018 provided by SSL/TLS is not desired or is not enough. 2019 [Bodo Moeller] 2020 2021 *) Enhanced support for Alpha Linux is added. Now ./config checks if 2022 the host supports BWX extension and if Compaq C is present on the 2023 $PATH. Just exploiting of the BWX extension results in 20-30% 2024 performance kick for some algorithms, e.g. DES and RC4 to mention 2025 a couple. Compaq C in turn generates ~20% faster code for MD5 and 2026 SHA1. 2027 [Andy Polyakov] 2028 2029 *) Add support for MS "fast SGC". This is arguably a violation of the 2030 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 2031 weak crypto and after checking the certificate is SGC a second one 2032 with strong crypto. MS SGC stops the first handshake after receiving 2033 the server certificate message and sends a second client hello. Since 2034 a server will typically do all the time consuming operations before 2035 expecting any further messages from the client (server key exchange 2036 is the most expensive) there is little difference between the two. 2037 2038 To get OpenSSL to support MS SGC we have to permit a second client 2039 hello message after we have sent server done. In addition we have to 2040 reset the MAC if we do get this second client hello. 2041 [Steve Henson] 2042 2043 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 2044 if a DER encoded private key is RSA or DSA traditional format. Changed 2045 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 2046 format DER encoded private key. Newer code should use PKCS#8 format which 2047 has the key type encoded in the ASN1 structure. Added DER private key 2048 support to pkcs8 application. 2049 [Steve Henson] 2050 2051 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 2052 ciphersuites has been selected (as required by the SSL 3/TLS 1 2053 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 2054 is set, we interpret this as a request to violate the specification 2055 (the worst that can happen is a handshake failure, and 'correct' 2056 behaviour would result in a handshake failure anyway). 2057 [Bodo Moeller] 2058 2059 *) In SSL_CTX_add_session, take into account that there might be multiple 2060 SSL_SESSION structures with the same session ID (e.g. when two threads 2061 concurrently obtain them from an external cache). 2062 The internal cache can handle only one SSL_SESSION with a given ID, 2063 so if there's a conflict, we now throw out the old one to achieve 2064 consistency. 2065 [Bodo Moeller] 2066 2067 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 2068 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 2069 some routines that use cipher OIDs: some ciphers do not have OIDs 2070 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 2071 example. 2072 [Steve Henson] 2073 2074 *) Simplify the trust setting structure and code. Now we just have 2075 two sequences of OIDs for trusted and rejected settings. These will 2076 typically have values the same as the extended key usage extension 2077 and any application specific purposes. 2078 2079 The trust checking code now has a default behaviour: it will just 2080 check for an object with the same NID as the passed id. Functions can 2081 be provided to override either the default behaviour or the behaviour 2082 for a given id. SSL client, server and email already have functions 2083 in place for compatibility: they check the NID and also return "trusted" 2084 if the certificate is self signed. 2085 [Steve Henson] 2086 2087 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 2088 traditional format into an EVP_PKEY structure. 2089 [Steve Henson] 2090 2091 *) Add a password callback function PEM_cb() which either prompts for 2092 a password if usr_data is NULL or otherwise assumes it is a null 2093 terminated password. Allow passwords to be passed on command line 2094 environment or config files in a few more utilities. 2095 [Steve Henson] 2096 2097 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 2098 keys. Add some short names for PKCS#8 PBE algorithms and allow them 2099 to be specified on the command line for the pkcs8 and pkcs12 utilities. 2100 Update documentation. 2101 [Steve Henson] 2102 2103 *) Support for ASN1 "NULL" type. This could be handled before by using 2104 ASN1_TYPE but there wasn't any function that would try to read a NULL 2105 and produce an error if it couldn't. For compatibility we also have 2106 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 2107 don't allocate anything because they don't need to. 2108 [Steve Henson] 2109 2110 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 2111 for details. 2112 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 2113 2114 *) Rebuild of the memory allocation routines used by OpenSSL code and 2115 possibly others as well. The purpose is to make an interface that 2116 provide hooks so anyone can build a separate set of allocation and 2117 deallocation routines to be used by OpenSSL, for example memory 2118 pool implementations, or something else, which was previously hard 2119 since Malloc(), Realloc() and Free() were defined as macros having 2120 the values malloc, realloc and free, respectively (except for Win32 2121 compilations). The same is provided for memory debugging code. 2122 OpenSSL already comes with functionality to find memory leaks, but 2123 this gives people a chance to debug other memory problems. 2124 2125 With these changes, a new set of functions and macros have appeared: 2126 2127 CRYPTO_set_mem_debug_functions() [F] 2128 CRYPTO_get_mem_debug_functions() [F] 2129 CRYPTO_dbg_set_options() [F] 2130 CRYPTO_dbg_get_options() [F] 2131 CRYPTO_malloc_debug_init() [M] 2132 2133 The memory debug functions are NULL by default, unless the library 2134 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 2135 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 2136 gives the standard debugging functions that come with OpenSSL) or 2137 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 2138 provided by the library user) must be used. When the standard 2139 debugging functions are used, CRYPTO_dbg_set_options can be used to 2140 request additional information: 2141 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 2142 the CRYPTO_MDEBUG_xxx macro when compiling the library. 2143 2144 Also, things like CRYPTO_set_mem_functions will always give the 2145 expected result (the new set of functions is used for allocation 2146 and deallocation) at all times, regardless of platform and compiler 2147 options. 2148 2149 To finish it up, some functions that were never use in any other 2150 way than through macros have a new API and new semantic: 2151 2152 CRYPTO_dbg_malloc() 2153 CRYPTO_dbg_realloc() 2154 CRYPTO_dbg_free() 2155 2156 All macros of value have retained their old syntax. 2157 [Richard Levitte and Bodo Moeller] 2158 2159 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 2160 ordering of SMIMECapabilities wasn't in "strength order" and there 2161 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 2162 algorithm. 2163 [Steve Henson] 2164 2165 *) Some ASN1 types with illegal zero length encoding (INTEGER, 2166 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 2167 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 2168 2169 *) Merge in my S/MIME library for OpenSSL. This provides a simple 2170 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 2171 functionality to handle multipart/signed properly) and a utility 2172 called 'smime' to call all this stuff. This is based on code I 2173 originally wrote for Celo who have kindly allowed it to be 2174 included in OpenSSL. 2175 [Steve Henson] 2176 2177 *) Add variants des_set_key_checked and des_set_key_unchecked of 2178 des_set_key (aka des_key_sched). Global variable des_check_key 2179 decides which of these is called by des_set_key; this way 2180 des_check_key behaves as it always did, but applications and 2181 the library itself, which was buggy for des_check_key == 1, 2182 have a cleaner way to pick the version they need. 2183 [Bodo Moeller] 2184 2185 *) New function PKCS12_newpass() which changes the password of a 2186 PKCS12 structure. 2187 [Steve Henson] 2188 2189 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 2190 dynamic mix. In both cases the ids can be used as an index into the 2191 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 2192 functions so they accept a list of the field values and the 2193 application doesn't need to directly manipulate the X509_TRUST 2194 structure. 2195 [Steve Henson] 2196 2197 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 2198 need initialising. 2199 [Steve Henson] 2200 2201 *) Modify the way the V3 extension code looks up extensions. This now 2202 works in a similar way to the object code: we have some "standard" 2203 extensions in a static table which is searched with OBJ_bsearch() 2204 and the application can add dynamic ones if needed. The file 2205 crypto/x509v3/ext_dat.h now has the info: this file needs to be 2206 updated whenever a new extension is added to the core code and kept 2207 in ext_nid order. There is a simple program 'tabtest.c' which checks 2208 this. New extensions are not added too often so this file can readily 2209 be maintained manually. 2210 2211 There are two big advantages in doing things this way. The extensions 2212 can be looked up immediately and no longer need to be "added" using 2213 X509V3_add_standard_extensions(): this function now does nothing. 2214 [Side note: I get *lots* of email saying the extension code doesn't 2215 work because people forget to call this function] 2216 Also no dynamic allocation is done unless new extensions are added: 2217 so if we don't add custom extensions there is no need to call 2218 X509V3_EXT_cleanup(). 2219 [Steve Henson] 2220 2221 *) Modify enc utility's salting as follows: make salting the default. Add a 2222 magic header, so unsalted files fail gracefully instead of just decrypting 2223 to garbage. This is because not salting is a big security hole, so people 2224 should be discouraged from doing it. 2225 [Ben Laurie] 2226 2227 *) Fixes and enhancements to the 'x509' utility. It allowed a message 2228 digest to be passed on the command line but it only used this 2229 parameter when signing a certificate. Modified so all relevant 2230 operations are affected by the digest parameter including the 2231 -fingerprint and -x509toreq options. Also -x509toreq choked if a 2232 DSA key was used because it didn't fix the digest. 2233 [Steve Henson] 2234 2235 *) Initial certificate chain verify code. Currently tests the untrusted 2236 certificates for consistency with the verify purpose (which is set 2237 when the X509_STORE_CTX structure is set up) and checks the pathlength. 2238 2239 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 2240 this is because it will reject chains with invalid extensions whereas 2241 every previous version of OpenSSL and SSLeay made no checks at all. 2242 2243 Trust code: checks the root CA for the relevant trust settings. Trust 2244 settings have an initial value consistent with the verify purpose: e.g. 2245 if the verify purpose is for SSL client use it expects the CA to be 2246 trusted for SSL client use. However the default value can be changed to 2247 permit custom trust settings: one example of this would be to only trust 2248 certificates from a specific "secure" set of CAs. 2249 2250 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 2251 which should be used for version portability: especially since the 2252 verify structure is likely to change more often now. 2253 2254 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 2255 to set them. If not set then assume SSL clients will verify SSL servers 2256 and vice versa. 2257 2258 Two new options to the verify program: -untrusted allows a set of 2259 untrusted certificates to be passed in and -purpose which sets the 2260 intended purpose of the certificate. If a purpose is set then the 2261 new chain verify code is used to check extension consistency. 2262 [Steve Henson] 2263 2264 *) Support for the authority information access extension. 2265 [Steve Henson] 2266 2267 *) Modify RSA and DSA PEM read routines to transparently handle 2268 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 2269 public keys in a format compatible with certificate 2270 SubjectPublicKeyInfo structures. Unfortunately there were already 2271 functions called *_PublicKey_* which used various odd formats so 2272 these are retained for compatibility: however the DSA variants were 2273 never in a public release so they have been deleted. Changed dsa/rsa 2274 utilities to handle the new format: note no releases ever handled public 2275 keys so we should be OK. 2276 2277 The primary motivation for this change is to avoid the same fiasco 2278 that dogs private keys: there are several incompatible private key 2279 formats some of which are standard and some OpenSSL specific and 2280 require various evil hacks to allow partial transparent handling and 2281 even then it doesn't work with DER formats. Given the option anything 2282 other than PKCS#8 should be dumped: but the other formats have to 2283 stay in the name of compatibility. 2284 2285 With public keys and the benefit of hindsight one standard format 2286 is used which works with EVP_PKEY, RSA or DSA structures: though 2287 it clearly returns an error if you try to read the wrong kind of key. 2288 2289 Added a -pubkey option to the 'x509' utility to output the public key. 2290 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 2291 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 2292 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 2293 that do the same as the EVP_PKEY_assign_*() except they up the 2294 reference count of the added key (they don't "swallow" the 2295 supplied key). 2296 [Steve Henson] 2297 2298 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 2299 CRLs would fail if the file contained no certificates or no CRLs: 2300 added a new function to read in both types and return the number 2301 read: this means that if none are read it will be an error. The 2302 DER versions of the certificate and CRL reader would always fail 2303 because it isn't possible to mix certificates and CRLs in DER format 2304 without choking one or the other routine. Changed this to just read 2305 a certificate: this is the best we can do. Also modified the code 2306 in apps/verify.c to take notice of return codes: it was previously 2307 attempting to read in certificates from NULL pointers and ignoring 2308 any errors: this is one reason why the cert and CRL reader seemed 2309 to work. It doesn't check return codes from the default certificate 2310 routines: these may well fail if the certificates aren't installed. 2311 [Steve Henson] 2312 2313 *) Code to support otherName option in GeneralName. 2314 [Steve Henson] 2315 2316 *) First update to verify code. Change the verify utility 2317 so it warns if it is passed a self signed certificate: 2318 for consistency with the normal behaviour. X509_verify 2319 has been modified to it will now verify a self signed 2320 certificate if *exactly* the same certificate appears 2321 in the store: it was previously impossible to trust a 2322 single self signed certificate. This means that: 2323 openssl verify ss.pem 2324 now gives a warning about a self signed certificate but 2325 openssl verify -CAfile ss.pem ss.pem 2326 is OK. 2327 [Steve Henson] 2328 2329 *) For servers, store verify_result in SSL_SESSION data structure 2330 (and add it to external session representation). 2331 This is needed when client certificate verifications fails, 2332 but an application-provided verification callback (set by 2333 SSL_CTX_set_cert_verify_callback) allows accepting the session 2334 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 2335 but returns 1): When the session is reused, we have to set 2336 ssl->verify_result to the appropriate error code to avoid 2337 security holes. 2338 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 2339 2340 *) Fix a bug in the new PKCS#7 code: it didn't consider the 2341 case in PKCS7_dataInit() where the signed PKCS7 structure 2342 didn't contain any existing data because it was being created. 2343 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 2344 2345 *) Add a salt to the key derivation routines in enc.c. This 2346 forms the first 8 bytes of the encrypted file. Also add a 2347 -S option to allow a salt to be input on the command line. 2348 [Steve Henson] 2349 2350 *) New function X509_cmp(). Oddly enough there wasn't a function 2351 to compare two certificates. We do this by working out the SHA1 2352 hash and comparing that. X509_cmp() will be needed by the trust 2353 code. 2354 [Steve Henson] 2355 2356 *) SSL_get1_session() is like SSL_get_session(), but increments 2357 the reference count in the SSL_SESSION returned. 2358 [Geoff Thorpe <geoff@eu.c2.net>] 2359 2360 *) Fix for 'req': it was adding a null to request attributes. 2361 Also change the X509_LOOKUP and X509_INFO code to handle 2362 certificate auxiliary information. 2363 [Steve Henson] 2364 2365 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 2366 the 'enc' command. 2367 [Steve Henson] 2368 2369 *) Add the possibility to add extra information to the memory leak 2370 detecting output, to form tracebacks, showing from where each 2371 allocation was originated: CRYPTO_push_info("constant string") adds 2372 the string plus current file name and line number to a per-thread 2373 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 2374 is like calling CYRPTO_pop_info() until the stack is empty. 2375 Also updated memory leak detection code to be multi-thread-safe. 2376 [Richard Levitte] 2377 2378 *) Add options -text and -noout to pkcs7 utility and delete the 2379 encryption options which never did anything. Update docs. 2380 [Steve Henson] 2381 2382 *) Add options to some of the utilities to allow the pass phrase 2383 to be included on either the command line (not recommended on 2384 OSes like Unix) or read from the environment. Update the 2385 manpages and fix a few bugs. 2386 [Steve Henson] 2387 2388 *) Add a few manpages for some of the openssl commands. 2389 [Steve Henson] 2390 2391 *) Fix the -revoke option in ca. It was freeing up memory twice, 2392 leaking and not finding already revoked certificates. 2393 [Steve Henson] 2394 2395 *) Extensive changes to support certificate auxiliary information. 2396 This involves the use of X509_CERT_AUX structure and X509_AUX 2397 functions. An X509_AUX function such as PEM_read_X509_AUX() 2398 can still read in a certificate file in the usual way but it 2399 will also read in any additional "auxiliary information". By 2400 doing things this way a fair degree of compatibility can be 2401 retained: existing certificates can have this information added 2402 using the new 'x509' options. 2403 2404 Current auxiliary information includes an "alias" and some trust 2405 settings. The trust settings will ultimately be used in enhanced 2406 certificate chain verification routines: currently a certificate 2407 can only be trusted if it is self signed and then it is trusted 2408 for all purposes. 2409 [Steve Henson] 2410 2411 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 2412 The problem was that one of the replacement routines had not been working 2413 since SSLeay releases. For now the offending routine has been replaced 2414 with non-optimised assembler. Even so, this now gives around 95% 2415 performance improvement for 1024 bit RSA signs. 2416 [Mark Cox] 2417 2418 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 2419 handling. Most clients have the effective key size in bits equal to 2420 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 2421 A few however don't do this and instead use the size of the decrypted key 2422 to determine the RC2 key length and the AlgorithmIdentifier to determine 2423 the effective key length. In this case the effective key length can still 2424 be 40 bits but the key length can be 168 bits for example. This is fixed 2425 by manually forcing an RC2 key into the EVP_PKEY structure because the 2426 EVP code can't currently handle unusual RC2 key sizes: it always assumes 2427 the key length and effective key length are equal. 2428 [Steve Henson] 2429 2430 *) Add a bunch of functions that should simplify the creation of 2431 X509_NAME structures. Now you should be able to do: 2432 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 2433 and have it automatically work out the correct field type and fill in 2434 the structures. The more adventurous can try: 2435 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 2436 and it will (hopefully) work out the correct multibyte encoding. 2437 [Steve Henson] 2438 2439 *) Change the 'req' utility to use the new field handling and multibyte 2440 copy routines. Before the DN field creation was handled in an ad hoc 2441 way in req, ca, and x509 which was rather broken and didn't support 2442 BMPStrings or UTF8Strings. Since some software doesn't implement 2443 BMPStrings or UTF8Strings yet, they can be enabled using the config file 2444 using the dirstring_type option. See the new comment in the default 2445 openssl.cnf for more info. 2446 [Steve Henson] 2447 2448 *) Make crypto/rand/md_rand.c more robust: 2449 - Assure unique random numbers after fork(). 2450 - Make sure that concurrent threads access the global counter and 2451 md serializably so that we never lose entropy in them 2452 or use exactly the same state in multiple threads. 2453 Access to the large state is not always serializable because 2454 the additional locking could be a performance killer, and 2455 md should be large enough anyway. 2456 [Bodo Moeller] 2457 2458 *) New file apps/app_rand.c with commonly needed functionality 2459 for handling the random seed file. 2460 2461 Use the random seed file in some applications that previously did not: 2462 ca, 2463 dsaparam -genkey (which also ignored its '-rand' option), 2464 s_client, 2465 s_server, 2466 x509 (when signing). 2467 Except on systems with /dev/urandom, it is crucial to have a random 2468 seed file at least for key creation, DSA signing, and for DH exchanges; 2469 for RSA signatures we could do without one. 2470 2471 gendh and gendsa (unlike genrsa) used to read only the first byte 2472 of each file listed in the '-rand' option. The function as previously 2473 found in genrsa is now in app_rand.c and is used by all programs 2474 that support '-rand'. 2475 [Bodo Moeller] 2476 2477 *) In RAND_write_file, use mode 0600 for creating files; 2478 don't just chmod when it may be too late. 2479 [Bodo Moeller] 2480 2481 *) Report an error from X509_STORE_load_locations 2482 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 2483 [Bill Perry] 2484 2485 *) New function ASN1_mbstring_copy() this copies a string in either 2486 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 2487 into an ASN1_STRING type. A mask of permissible types is passed 2488 and it chooses the "minimal" type to use or an error if not type 2489 is suitable. 2490 [Steve Henson] 2491 2492 *) Add function equivalents to the various macros in asn1.h. The old 2493 macros are retained with an M_ prefix. Code inside the library can 2494 use the M_ macros. External code (including the openssl utility) 2495 should *NOT* in order to be "shared library friendly". 2496 [Steve Henson] 2497 2498 *) Add various functions that can check a certificate's extensions 2499 to see if it usable for various purposes such as SSL client, 2500 server or S/MIME and CAs of these types. This is currently 2501 VERY EXPERIMENTAL but will ultimately be used for certificate chain 2502 verification. Also added a -purpose flag to x509 utility to 2503 print out all the purposes. 2504 [Steve Henson] 2505 2506 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 2507 functions. 2508 [Steve Henson] 2509 2510 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 2511 for, obtain and decode and extension and obtain its critical flag. 2512 This allows all the necessary extension code to be handled in a 2513 single function call. 2514 [Steve Henson] 2515 2516 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 2517 platforms. See crypto/rc4/rc4_enc.c for further details. 2518 [Andy Polyakov] 2519 2520 *) New -noout option to asn1parse. This causes no output to be produced 2521 its main use is when combined with -strparse and -out to extract data 2522 from a file (which may not be in ASN.1 format). 2523 [Steve Henson] 2524 2525 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 2526 when producing the local key id. 2527 [Richard Levitte <levitte@stacken.kth.se>] 2528 2529 *) New option -dhparam in s_server. This allows a DH parameter file to be 2530 stated explicitly. If it is not stated then it tries the first server 2531 certificate file. The previous behaviour hard coded the filename 2532 "server.pem". 2533 [Steve Henson] 2534 2535 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 2536 a public key to be input or output. For example: 2537 openssl rsa -in key.pem -pubout -out pubkey.pem 2538 Also added necessary DSA public key functions to handle this. 2539 [Steve Henson] 2540 2541 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 2542 in the message. This was handled by allowing 2543 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 2544 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 2545 2546 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 2547 to the end of the strings whereas this didn't. This would cause problems 2548 if strings read with d2i_ASN1_bytes() were later modified. 2549 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 2550 2551 *) Fix for base64 decode bug. When a base64 bio reads only one line of 2552 data and it contains EOF it will end up returning an error. This is 2553 caused by input 46 bytes long. The cause is due to the way base64 2554 BIOs find the start of base64 encoded data. They do this by trying a 2555 trial decode on each line until they find one that works. When they 2556 do a flag is set and it starts again knowing it can pass all the 2557 data directly through the decoder. Unfortunately it doesn't reset 2558 the context it uses. This means that if EOF is reached an attempt 2559 is made to pass two EOFs through the context and this causes the 2560 resulting error. This can also cause other problems as well. As is 2561 usual with these problems it takes *ages* to find and the fix is 2562 trivial: move one line. 2563 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 2564 2565 *) Ugly workaround to get s_client and s_server working under Windows. The 2566 old code wouldn't work because it needed to select() on sockets and the 2567 tty (for keypresses and to see if data could be written). Win32 only 2568 supports select() on sockets so we select() with a 1s timeout on the 2569 sockets and then see if any characters are waiting to be read, if none 2570 are present then we retry, we also assume we can always write data to 2571 the tty. This isn't nice because the code then blocks until we've 2572 received a complete line of data and it is effectively polling the 2573 keyboard at 1s intervals: however it's quite a bit better than not 2574 working at all :-) A dedicated Windows application might handle this 2575 with an event loop for example. 2576 [Steve Henson] 2577 2578 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 2579 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 2580 will be called when RSA_sign() and RSA_verify() are used. This is useful 2581 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 2582 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 2583 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 2584 This necessitated the support of an extra signature type NID_md5_sha1 2585 for SSL signatures and modifications to the SSL library to use it instead 2586 of calling RSA_public_decrypt() and RSA_private_encrypt(). 2587 [Steve Henson] 2588 2589 *) Add new -verify -CAfile and -CApath options to the crl program, these 2590 will lookup a CRL issuers certificate and verify the signature in a 2591 similar way to the verify program. Tidy up the crl program so it 2592 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 2593 less strict. It will now permit CRL extensions even if it is not 2594 a V2 CRL: this will allow it to tolerate some broken CRLs. 2595 [Steve Henson] 2596 2597 *) Initialize all non-automatic variables each time one of the openssl 2598 sub-programs is started (this is necessary as they may be started 2599 multiple times from the "OpenSSL>" prompt). 2600 [Lennart Bang, Bodo Moeller] 2601 2602 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 2603 removing all other RSA functionality (this is what NO_RSA does). This 2604 is so (for example) those in the US can disable those operations covered 2605 by the RSA patent while allowing storage and parsing of RSA keys and RSA 2606 key generation. 2607 [Steve Henson] 2608 2609 *) Non-copying interface to BIO pairs. 2610 (still largely untested) 2611 [Bodo Moeller] 2612 2613 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 2614 ASCII string. This was handled independently in various places before. 2615 [Steve Henson] 2616 2617 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 2618 UTF8 strings a character at a time. 2619 [Steve Henson] 2620 2621 *) Use client_version from client hello to select the protocol 2622 (s23_srvr.c) and for RSA client key exchange verification 2623 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 2624 [Bodo Moeller] 2625 2626 *) Add various utility functions to handle SPKACs, these were previously 2627 handled by poking round in the structure internals. Added new function 2628 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 2629 print, verify and generate SPKACs. Based on an original idea from 2630 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 2631 [Steve Henson] 2632 2633 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 2634 [Andy Polyakov] 2635 2636 *) Allow the config file extension section to be overwritten on the 2637 command line. Based on an original idea from Massimiliano Pala 2638 <madwolf@comune.modena.it>. The new option is called -extensions 2639 and can be applied to ca, req and x509. Also -reqexts to override 2640 the request extensions in req and -crlexts to override the crl extensions 2641 in ca. 2642 [Steve Henson] 2643 2644 *) Add new feature to the SPKAC handling in ca. Now you can include 2645 the same field multiple times by preceding it by "XXXX." for example: 2646 1.OU="Unit name 1" 2647 2.OU="Unit name 2" 2648 this is the same syntax as used in the req config file. 2649 [Steve Henson] 2650 2651 *) Allow certificate extensions to be added to certificate requests. These 2652 are specified in a 'req_extensions' option of the req section of the 2653 config file. They can be printed out with the -text option to req but 2654 are otherwise ignored at present. 2655 [Steve Henson] 2656 2657 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 2658 data read consists of only the final block it would not decrypted because 2659 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 2660 A misplaced 'break' also meant the decrypted final block might not be 2661 copied until the next read. 2662 [Steve Henson] 2663 2664 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 2665 a few extra parameters to the DH structure: these will be useful if 2666 for example we want the value of 'q' or implement X9.42 DH. 2667 [Steve Henson] 2668 2669 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 2670 provides hooks that allow the default DSA functions or functions on a 2671 "per key" basis to be replaced. This allows hardware acceleration and 2672 hardware key storage to be handled without major modification to the 2673 library. Also added low level modexp hooks and CRYPTO_EX structure and 2674 associated functions. 2675 [Steve Henson] 2676 2677 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 2678 as "read only": it can't be written to and the buffer it points to will 2679 not be freed. Reading from a read only BIO is much more efficient than 2680 a normal memory BIO. This was added because there are several times when 2681 an area of memory needs to be read from a BIO. The previous method was 2682 to create a memory BIO and write the data to it, this results in two 2683 copies of the data and an O(n^2) reading algorithm. There is a new 2684 function BIO_new_mem_buf() which creates a read only memory BIO from 2685 an area of memory. Also modified the PKCS#7 routines to use read only 2686 memory BIOs. 2687 [Steve Henson] 2688 2689 *) Bugfix: ssl23_get_client_hello did not work properly when called in 2690 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 2691 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 2692 but a retry condition occured while trying to read the rest. 2693 [Bodo Moeller] 2694 2695 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 2696 NID_pkcs7_encrypted by default: this was wrong since this should almost 2697 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 2698 the encrypted data type: this is a more sensible place to put it and it 2699 allows the PKCS#12 code to be tidied up that duplicated this 2700 functionality. 2701 [Steve Henson] 2702 2703 *) Changed obj_dat.pl script so it takes its input and output files on 2704 the command line. This should avoid shell escape redirection problems 2705 under Win32. 2706 [Steve Henson] 2707 2708 *) Initial support for certificate extension requests, these are included 2709 in things like Xenroll certificate requests. Included functions to allow 2710 extensions to be obtained and added. 2711 [Steve Henson] 2712 2713 *) -crlf option to s_client and s_server for sending newlines as 2714 CRLF (as required by many protocols). 2715 [Bodo Moeller] 2716 2717 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 2718 2719 *) Install libRSAglue.a when OpenSSL is built with RSAref. 2720 [Ralf S. Engelschall] 2721 2722 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 2723 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 2724 2725 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 2726 program. 2727 [Steve Henson] 2728 2729 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 2730 DH parameters/keys (q is lost during that conversion, but the resulting 2731 DH parameters contain its length). 2732 2733 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 2734 much faster than DH_generate_parameters (which creates parameters 2735 where p = 2*q + 1), and also the smaller q makes DH computations 2736 much more efficient (160-bit exponentiation instead of 1024-bit 2737 exponentiation); so this provides a convenient way to support DHE 2738 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 2739 utter importance to use 2740 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 2741 or 2742 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 2743 when such DH parameters are used, because otherwise small subgroup 2744 attacks may become possible! 2745 [Bodo Moeller] 2746 2747 *) Avoid memory leak in i2d_DHparams. 2748 [Bodo Moeller] 2749 2750 *) Allow the -k option to be used more than once in the enc program: 2751 this allows the same encrypted message to be read by multiple recipients. 2752 [Steve Henson] 2753 2754 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 2755 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 2756 it will always use the numerical form of the OID, even if it has a short 2757 or long name. 2758 [Steve Henson] 2759 2760 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 2761 method only got called if p,q,dmp1,dmq1,iqmp components were present, 2762 otherwise bn_mod_exp was called. In the case of hardware keys for example 2763 no private key components need be present and it might store extra data 2764 in the RSA structure, which cannot be accessed from bn_mod_exp. 2765 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 2766 private key operations. 2767 [Steve Henson] 2768 2769 *) Added support for SPARC Linux. 2770 [Andy Polyakov] 2771 2772 *) pem_password_cb function type incompatibly changed from 2773 typedef int pem_password_cb(char *buf, int size, int rwflag); 2774 to 2775 ....(char *buf, int size, int rwflag, void *userdata); 2776 so that applications can pass data to their callbacks: 2777 The PEM[_ASN1]_{read,write}... functions and macros now take an 2778 additional void * argument, which is just handed through whenever 2779 the password callback is called. 2780 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 2781 2782 New function SSL_CTX_set_default_passwd_cb_userdata. 2783 2784 Compatibility note: As many C implementations push function arguments 2785 onto the stack in reverse order, the new library version is likely to 2786 interoperate with programs that have been compiled with the old 2787 pem_password_cb definition (PEM_whatever takes some data that 2788 happens to be on the stack as its last argument, and the callback 2789 just ignores this garbage); but there is no guarantee whatsoever that 2790 this will work. 2791 2792 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 2793 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 2794 problems not only on Windows, but also on some Unix platforms. 2795 To avoid problematic command lines, these definitions are now in an 2796 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 2797 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 2798 [Bodo Moeller] 2799 2800 *) MIPS III/IV assembler module is reimplemented. 2801 [Andy Polyakov] 2802 2803 *) More DES library cleanups: remove references to srand/rand and 2804 delete an unused file. 2805 [Ulf M�ller] 2806 2807 *) Add support for the the free Netwide assembler (NASM) under Win32, 2808 since not many people have MASM (ml) and it can be hard to obtain. 2809 This is currently experimental but it seems to work OK and pass all 2810 the tests. Check out INSTALL.W32 for info. 2811 [Steve Henson] 2812 2813 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 2814 without temporary keys kept an extra copy of the server key, 2815 and connections with temporary keys did not free everything in case 2816 of an error. 2817 [Bodo Moeller] 2818 2819 *) New function RSA_check_key and new openssl rsa option -check 2820 for verifying the consistency of RSA keys. 2821 [Ulf Moeller, Bodo Moeller] 2822 2823 *) Various changes to make Win32 compile work: 2824 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 2825 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 2826 comparison" warnings. 2827 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 2828 [Steve Henson] 2829 2830 *) Add a debugging option to PKCS#5 v2 key generation function: when 2831 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 2832 derived keys are printed to stderr. 2833 [Steve Henson] 2834 2835 *) Copy the flags in ASN1_STRING_dup(). 2836 [Roman E. Pavlov <pre@mo.msk.ru>] 2837 2838 *) The x509 application mishandled signing requests containing DSA 2839 keys when the signing key was also DSA and the parameters didn't match. 2840 2841 It was supposed to omit the parameters when they matched the signing key: 2842 the verifying software was then supposed to automatically use the CA's 2843 parameters if they were absent from the end user certificate. 2844 2845 Omitting parameters is no longer recommended. The test was also 2846 the wrong way round! This was probably due to unusual behaviour in 2847 EVP_cmp_parameters() which returns 1 if the parameters match. 2848 This meant that parameters were omitted when they *didn't* match and 2849 the certificate was useless. Certificates signed with 'ca' didn't have 2850 this bug. 2851 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 2852 2853 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 2854 The interface is as follows: 2855 Applications can use 2856 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 2857 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 2858 "off" is now the default. 2859 The library internally uses 2860 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 2861 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 2862 to disable memory-checking temporarily. 2863 2864 Some inconsistent states that previously were possible (and were 2865 even the default) are now avoided. 2866 2867 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 2868 with each memory chunk allocated; this is occasionally more helpful 2869 than just having a counter. 2870 2871 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 2872 2873 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 2874 extensions. 2875 [Bodo Moeller] 2876 2877 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 2878 which largely parallels "options", but is for changing API behaviour, 2879 whereas "options" are about protocol behaviour. 2880 Initial "mode" flags are: 2881 2882 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 2883 a single record has been written. 2884 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 2885 retries use the same buffer location. 2886 (But all of the contents must be 2887 copied!) 2888 [Bodo Moeller] 2889 2890 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 2891 worked. 2892 2893 *) Fix problems with no-hmac etc. 2894 [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 2895 2896 *) New functions RSA_get_default_method(), RSA_set_method() and 2897 RSA_get_method(). These allows replacement of RSA_METHODs without having 2898 to mess around with the internals of an RSA structure. 2899 [Steve Henson] 2900 2901 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 2902 Also really enable memory leak checks in openssl.c and in some 2903 test programs. 2904 [Chad C. Mulligan, Bodo Moeller] 2905 2906 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 2907 up the length of negative integers. This has now been simplified to just 2908 store the length when it is first determined and use it later, rather 2909 than trying to keep track of where data is copied and updating it to 2910 point to the end. 2911 [Steve Henson, reported by Brien Wheeler 2912 <bwheeler@authentica-security.com>] 2913 2914 *) Add a new function PKCS7_signatureVerify. This allows the verification 2915 of a PKCS#7 signature but with the signing certificate passed to the 2916 function itself. This contrasts with PKCS7_dataVerify which assumes the 2917 certificate is present in the PKCS#7 structure. This isn't always the 2918 case: certificates can be omitted from a PKCS#7 structure and be 2919 distributed by "out of band" means (such as a certificate database). 2920 [Steve Henson] 2921 2922 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 2923 function prototypes in pem.h, also change util/mkdef.pl to add the 2924 necessary function names. 2925 [Steve Henson] 2926 2927 *) mk1mf.pl (used by Windows builds) did not properly read the 2928 options set by Configure in the top level Makefile, and Configure 2929 was not even able to write more than one option correctly. 2930 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 2931 [Bodo Moeller] 2932 2933 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 2934 file to be loaded from a BIO or FILE pointer. The BIO version will 2935 for example allow memory BIOs to contain config info. 2936 [Steve Henson] 2937 2938 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 2939 Whoever hopes to achieve shared-library compatibility across versions 2940 must use this, not the compile-time macro. 2941 (Exercise 0.9.4: Which is the minimum library version required by 2942 such programs?) 2943 Note: All this applies only to multi-threaded programs, others don't 2944 need locks. 2945 [Bodo Moeller] 2946 2947 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 2948 through a BIO pair triggered the default case, i.e. 2949 SSLerr(...,SSL_R_UNKNOWN_STATE). 2950 [Bodo Moeller] 2951 2952 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 2953 can use the SSL library even if none of the specific BIOs is 2954 appropriate. 2955 [Bodo Moeller] 2956 2957 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 2958 for the encoded length. 2959 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 2960 2961 *) Add initial documentation of the X509V3 functions. 2962 [Steve Henson] 2963 2964 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 2965 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 2966 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 2967 secure PKCS#8 private key format with a high iteration count. 2968 [Steve Henson] 2969 2970 *) Fix determination of Perl interpreter: A perl or perl5 2971 _directory_ in $PATH was also accepted as the interpreter. 2972 [Ralf S. Engelschall] 2973 2974 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 2975 wrong with it but it was very old and did things like calling 2976 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 2977 unusual formatting. 2978 [Steve Henson] 2979 2980 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 2981 to use the new extension code. 2982 [Steve Henson] 2983 2984 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 2985 with macros. This should make it easier to change their form, add extra 2986 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 2987 constant. 2988 [Steve Henson] 2989 2990 *) Add to configuration table a new entry that can specify an alternative 2991 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 2992 according to Mark Crispin <MRC@Panda.COM>. 2993 [Bodo Moeller] 2994 2995#if 0 2996 *) DES CBC did not update the IV. Weird. 2997 [Ben Laurie] 2998#else 2999 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 3000 Changing the behaviour of the former might break existing programs -- 3001 where IV updating is needed, des_ncbc_encrypt can be used. 3002#endif 3003 3004 *) When bntest is run from "make test" it drives bc to check its 3005 calculations, as well as internally checking them. If an internal check 3006 fails, it needs to cause bc to give a non-zero result or make test carries 3007 on without noticing the failure. Fixed. 3008 [Ben Laurie] 3009 3010 *) DES library cleanups. 3011 [Ulf M�ller] 3012 3013 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 3014 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 3015 ciphers. NOTE: although the key derivation function has been verified 3016 against some published test vectors it has not been extensively tested 3017 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 3018 of v2.0. 3019 [Steve Henson] 3020 3021 *) Instead of "mkdir -p", which is not fully portable, use new 3022 Perl script "util/mkdir-p.pl". 3023 [Bodo Moeller] 3024 3025 *) Rewrite the way password based encryption (PBE) is handled. It used to 3026 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 3027 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 3028 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 3029 the 'parameter' field of the AlgorithmIdentifier is passed to the 3030 underlying key generation function so it must do its own ASN1 parsing. 3031 This has also changed the EVP_PBE_CipherInit() function which now has a 3032 'parameter' argument instead of literal salt and iteration count values 3033 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 3034 [Steve Henson] 3035 3036 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 3037 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 3038 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 3039 KEY" because this clashed with PKCS#8 unencrypted string. Since this 3040 value was just used as a "magic string" and not used directly its 3041 value doesn't matter. 3042 [Steve Henson] 3043 3044 *) Introduce some semblance of const correctness to BN. Shame C doesn't 3045 support mutable. 3046 [Ben Laurie] 3047 3048 *) "linux-sparc64" configuration (ultrapenguin). 3049 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 3050 "linux-sparc" configuration. 3051 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 3052 3053 *) config now generates no-xxx options for missing ciphers. 3054 [Ulf M�ller] 3055 3056 *) Support the EBCDIC character set (work in progress). 3057 File ebcdic.c not yet included because it has a different license. 3058 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 3059 3060 *) Support BS2000/OSD-POSIX. 3061 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 3062 3063 *) Make callbacks for key generation use void * instead of char *. 3064 [Ben Laurie] 3065 3066 *) Make S/MIME samples compile (not yet tested). 3067 [Ben Laurie] 3068 3069 *) Additional typesafe stacks. 3070 [Ben Laurie] 3071 3072 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 3073 [Bodo Moeller] 3074 3075 3076 Changes between 0.9.3 and 0.9.3a [29 May 1999] 3077 3078 *) New configuration variant "sco5-gcc". 3079 3080 *) Updated some demos. 3081 [Sean O Riordain, Wade Scholine] 3082 3083 *) Add missing BIO_free at exit of pkcs12 application. 3084 [Wu Zhigang] 3085 3086 *) Fix memory leak in conf.c. 3087 [Steve Henson] 3088 3089 *) Updates for Win32 to assembler version of MD5. 3090 [Steve Henson] 3091 3092 *) Set #! path to perl in apps/der_chop to where we found it 3093 instead of using a fixed path. 3094 [Bodo Moeller] 3095 3096 *) SHA library changes for irix64-mips4-cc. 3097 [Andy Polyakov] 3098 3099 *) Improvements for VMS support. 3100 [Richard Levitte] 3101 3102 3103 Changes between 0.9.2b and 0.9.3 [24 May 1999] 3104 3105 *) Bignum library bug fix. IRIX 6 passes "make test" now! 3106 This also avoids the problems with SC4.2 and unpatched SC5. 3107 [Andy Polyakov <appro@fy.chalmers.se>] 3108 3109 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 3110 These are required because of the typesafe stack would otherwise break 3111 existing code. If old code used a structure member which used to be STACK 3112 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 3113 sk_num or sk_value it would produce an error because the num, data members 3114 are not present in STACK_OF. Now it just produces a warning. sk_set 3115 replaces the old method of assigning a value to sk_value 3116 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 3117 that does this will no longer work (and should use sk_set instead) but 3118 this could be regarded as a "questionable" behaviour anyway. 3119 [Steve Henson] 3120 3121 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 3122 correctly handle encrypted S/MIME data. 3123 [Steve Henson] 3124 3125 *) Change type of various DES function arguments from des_cblock 3126 (which means, in function argument declarations, pointer to char) 3127 to des_cblock * (meaning pointer to array with 8 char elements), 3128 which allows the compiler to do more typechecking; it was like 3129 that back in SSLeay, but with lots of ugly casts. 3130 3131 Introduce new type const_des_cblock. 3132 [Bodo Moeller] 3133 3134 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 3135 problems: find RecipientInfo structure that matches recipient certificate 3136 and initialise the ASN1 structures properly based on passed cipher. 3137 [Steve Henson] 3138 3139 *) Belatedly make the BN tests actually check the results. 3140 [Ben Laurie] 3141 3142 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 3143 to and from BNs: it was completely broken. New compilation option 3144 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 3145 key elements as negative integers. 3146 [Steve Henson] 3147 3148 *) Reorganize and speed up MD5. 3149 [Andy Polyakov <appro@fy.chalmers.se>] 3150 3151 *) VMS support. 3152 [Richard Levitte <richard@levitte.org>] 3153 3154 *) New option -out to asn1parse to allow the parsed structure to be 3155 output to a file. This is most useful when combined with the -strparse 3156 option to examine the output of things like OCTET STRINGS. 3157 [Steve Henson] 3158 3159 *) Make SSL library a little more fool-proof by not requiring any longer 3160 that SSL_set_{accept,connect}_state be called before 3161 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 3162 in many applications because usually everything *appeared* to work as 3163 intended anyway -- now it really works as intended). 3164 [Bodo Moeller] 3165 3166 *) Move openssl.cnf out of lib/. 3167 [Ulf M�ller] 3168 3169 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 3170 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 3171 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 3172 [Ralf S. Engelschall] 3173 3174 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 3175 handle PKCS#7 enveloped data properly. 3176 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 3177 3178 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 3179 copying pointers. The cert_st handling is changed by this in 3180 various ways (and thus what used to be known as ctx->default_cert 3181 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 3182 any longer when s->cert does not give us what we need). 3183 ssl_cert_instantiate becomes obsolete by this change. 3184 As soon as we've got the new code right (possibly it already is?), 3185 we have solved a couple of bugs of the earlier code where s->cert 3186 was used as if it could not have been shared with other SSL structures. 3187 3188 Note that using the SSL API in certain dirty ways now will result 3189 in different behaviour than observed with earlier library versions: 3190 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 3191 does not influence s as it used to. 3192 3193 In order to clean up things more thoroughly, inside SSL_SESSION 3194 we don't use CERT any longer, but a new structure SESS_CERT 3195 that holds per-session data (if available); currently, this is 3196 the peer's certificate chain and, for clients, the server's certificate 3197 and temporary key. CERT holds only those values that can have 3198 meaningful defaults in an SSL_CTX. 3199 [Bodo Moeller] 3200 3201 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 3202 from the internal representation. Various PKCS#7 fixes: remove some 3203 evil casts and set the enc_dig_alg field properly based on the signing 3204 key type. 3205 [Steve Henson] 3206 3207 *) Allow PKCS#12 password to be set from the command line or the 3208 environment. Let 'ca' get its config file name from the environment 3209 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 3210 and 'x509'). 3211 [Steve Henson] 3212 3213 *) Allow certificate policies extension to use an IA5STRING for the 3214 organization field. This is contrary to the PKIX definition but 3215 VeriSign uses it and IE5 only recognises this form. Document 'x509' 3216 extension option. 3217 [Steve Henson] 3218 3219 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 3220 without disallowing inline assembler and the like for non-pedantic builds. 3221 [Ben Laurie] 3222 3223 *) Support Borland C++ builder. 3224 [Janez Jere <jj@void.si>, modified by Ulf M�ller] 3225 3226 *) Support Mingw32. 3227 [Ulf M�ller] 3228 3229 *) SHA-1 cleanups and performance enhancements. 3230 [Andy Polyakov <appro@fy.chalmers.se>] 3231 3232 *) Sparc v8plus assembler for the bignum library. 3233 [Andy Polyakov <appro@fy.chalmers.se>] 3234 3235 *) Accept any -xxx and +xxx compiler options in Configure. 3236 [Ulf M�ller] 3237 3238 *) Update HPUX configuration. 3239 [Anonymous] 3240 3241 *) Add missing sk_<type>_unshift() function to safestack.h 3242 [Ralf S. Engelschall] 3243 3244 *) New function SSL_CTX_use_certificate_chain_file that sets the 3245 "extra_cert"s in addition to the certificate. (This makes sense 3246 only for "PEM" format files, as chains as a whole are not 3247 DER-encoded.) 3248 [Bodo Moeller] 3249 3250 *) Support verify_depth from the SSL API. 3251 x509_vfy.c had what can be considered an off-by-one-error: 3252 Its depth (which was not part of the external interface) 3253 was actually counting the number of certificates in a chain; 3254 now it really counts the depth. 3255 [Bodo Moeller] 3256 3257 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 3258 instead of X509err, which often resulted in confusing error 3259 messages since the error codes are not globally unique 3260 (e.g. an alleged error in ssl3_accept when a certificate 3261 didn't match the private key). 3262 3263 *) New function SSL_CTX_set_session_id_context that allows to set a default 3264 value (so that you don't need SSL_set_session_id_context for each 3265 connection using the SSL_CTX). 3266 [Bodo Moeller] 3267 3268 *) OAEP decoding bug fix. 3269 [Ulf M�ller] 3270 3271 *) Support INSTALL_PREFIX for package builders, as proposed by 3272 David Harris. 3273 [Bodo Moeller] 3274 3275 *) New Configure options "threads" and "no-threads". For systems 3276 where the proper compiler options are known (currently Solaris 3277 and Linux), "threads" is the default. 3278 [Bodo Moeller] 3279 3280 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 3281 [Bodo Moeller] 3282 3283 *) Install various scripts to $(OPENSSLDIR)/misc, not to 3284 $(INSTALLTOP)/bin -- they shouldn't clutter directories 3285 such as /usr/local/bin. 3286 [Bodo Moeller] 3287 3288 *) "make linux-shared" to build shared libraries. 3289 [Niels Poppe <niels@netbox.org>] 3290 3291 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 3292 [Ulf M�ller] 3293 3294 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 3295 extension adding in x509 utility. 3296 [Steve Henson] 3297 3298 *) Remove NOPROTO sections and error code comments. 3299 [Ulf M�ller] 3300 3301 *) Partial rewrite of the DEF file generator to now parse the ANSI 3302 prototypes. 3303 [Steve Henson] 3304 3305 *) New Configure options --prefix=DIR and --openssldir=DIR. 3306 [Ulf M�ller] 3307 3308 *) Complete rewrite of the error code script(s). It is all now handled 3309 by one script at the top level which handles error code gathering, 3310 header rewriting and C source file generation. It should be much better 3311 than the old method: it now uses a modified version of Ulf's parser to 3312 read the ANSI prototypes in all header files (thus the old K&R definitions 3313 aren't needed for error creation any more) and do a better job of 3314 translating function codes into names. The old 'ASN1 error code imbedded 3315 in a comment' is no longer necessary and it doesn't use .err files which 3316 have now been deleted. Also the error code call doesn't have to appear all 3317 on one line (which resulted in some large lines...). 3318 [Steve Henson] 3319 3320 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 3321 [Bodo Moeller] 3322 3323 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 3324 0 (which usually indicates a closed connection), but continue reading. 3325 [Bodo Moeller] 3326 3327 *) Fix some race conditions. 3328 [Bodo Moeller] 3329 3330 *) Add support for CRL distribution points extension. Add Certificate 3331 Policies and CRL distribution points documentation. 3332 [Steve Henson] 3333 3334 *) Move the autogenerated header file parts to crypto/opensslconf.h. 3335 [Ulf M�ller] 3336 3337 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 3338 8 of keying material. Merlin has also confirmed interop with this fix 3339 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 3340 [Merlin Hughes <merlin@baltimore.ie>] 3341 3342 *) Fix lots of warnings. 3343 [Richard Levitte <levitte@stacken.kth.se>] 3344 3345 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 3346 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 3347 [Richard Levitte <levitte@stacken.kth.se>] 3348 3349 *) Fix problems with sizeof(long) == 8. 3350 [Andy Polyakov <appro@fy.chalmers.se>] 3351 3352 *) Change functions to ANSI C. 3353 [Ulf M�ller] 3354 3355 *) Fix typos in error codes. 3356 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 3357 3358 *) Remove defunct assembler files from Configure. 3359 [Ulf M�ller] 3360 3361 *) SPARC v8 assembler BIGNUM implementation. 3362 [Andy Polyakov <appro@fy.chalmers.se>] 3363 3364 *) Support for Certificate Policies extension: both print and set. 3365 Various additions to support the r2i method this uses. 3366 [Steve Henson] 3367 3368 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 3369 return a const string when you are expecting an allocated buffer. 3370 [Ben Laurie] 3371 3372 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 3373 types DirectoryString and DisplayText. 3374 [Steve Henson] 3375 3376 *) Add code to allow r2i extensions to access the configuration database, 3377 add an LHASH database driver and add several ctx helper functions. 3378 [Steve Henson] 3379 3380 *) Fix an evil bug in bn_expand2() which caused various BN functions to 3381 fail when they extended the size of a BIGNUM. 3382 [Steve Henson] 3383 3384 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 3385 support typesafe stack. 3386 [Steve Henson] 3387 3388 *) Fix typo in SSL_[gs]et_options(). 3389 [Nils Frostberg <nils@medcom.se>] 3390 3391 *) Delete various functions and files that belonged to the (now obsolete) 3392 old X509V3 handling code. 3393 [Steve Henson] 3394 3395 *) New Configure option "rsaref". 3396 [Ulf M�ller] 3397 3398 *) Don't auto-generate pem.h. 3399 [Bodo Moeller] 3400 3401 *) Introduce type-safe ASN.1 SETs. 3402 [Ben Laurie] 3403 3404 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 3405 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 3406 3407 *) Introduce type-safe STACKs. This will almost certainly break lots of code 3408 that links with OpenSSL (well at least cause lots of warnings), but fear 3409 not: the conversion is trivial, and it eliminates loads of evil casts. A 3410 few STACKed things have been converted already. Feel free to convert more. 3411 In the fullness of time, I'll do away with the STACK type altogether. 3412 [Ben Laurie] 3413 3414 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 3415 specified in <certfile> by updating the entry in the index.txt file. 3416 This way one no longer has to edit the index.txt file manually for 3417 revoking a certificate. The -revoke option does the gory details now. 3418 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 3419 3420 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 3421 `-text' option at all and this way the `-noout -text' combination was 3422 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 3423 [Ralf S. Engelschall] 3424 3425 *) Make sure a corresponding plain text error message exists for the 3426 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 3427 verify callback function determined that a certificate was revoked. 3428 [Ralf S. Engelschall] 3429 3430 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 3431 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 3432 all available cipers including rc5, which was forgotten until now. 3433 In order to let the testing shell script know which algorithms 3434 are available, a new (up to now undocumented) command 3435 "openssl list-cipher-commands" is used. 3436 [Bodo Moeller] 3437 3438 *) Bugfix: s_client occasionally would sleep in select() when 3439 it should have checked SSL_pending() first. 3440 [Bodo Moeller] 3441 3442 *) New functions DSA_do_sign and DSA_do_verify to provide access to 3443 the raw DSA values prior to ASN.1 encoding. 3444 [Ulf M�ller] 3445 3446 *) Tweaks to Configure 3447 [Niels Poppe <niels@netbox.org>] 3448 3449 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 3450 yet... 3451 [Steve Henson] 3452 3453 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 3454 [Ulf M�ller] 3455 3456 *) New config option to avoid instructions that are illegal on the 80386. 3457 The default code is faster, but requires at least a 486. 3458 [Ulf M�ller] 3459 3460 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 3461 SSL2_SERVER_VERSION (not used at all) macros, which are now the 3462 same as SSL2_VERSION anyway. 3463 [Bodo Moeller] 3464 3465 *) New "-showcerts" option for s_client. 3466 [Bodo Moeller] 3467 3468 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 3469 application. Various cleanups and fixes. 3470 [Steve Henson] 3471 3472 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 3473 modify error routines to work internally. Add error codes and PBE init 3474 to library startup routines. 3475 [Steve Henson] 3476 3477 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 3478 packing functions to asn1 and evp. Changed function names and error 3479 codes along the way. 3480 [Steve Henson] 3481 3482 *) PKCS12 integration: and so it begins... First of several patches to 3483 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 3484 objects to objects.h 3485 [Steve Henson] 3486 3487 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 3488 and display support for Thawte strong extranet extension. 3489 [Steve Henson] 3490 3491 *) Add LinuxPPC support. 3492 [Jeff Dubrule <igor@pobox.org>] 3493 3494 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 3495 bn_div_words in alpha.s. 3496 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 3497 3498 *) Make sure the RSA OAEP test is skipped under -DRSAref because 3499 OAEP isn't supported when OpenSSL is built with RSAref. 3500 [Ulf Moeller <ulf@fitug.de>] 3501 3502 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 3503 so they no longer are missing under -DNOPROTO. 3504 [Soren S. Jorvang <soren@t.dk>] 3505 3506 3507 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 3508 3509 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 3510 doesn't work when the session is reused. Coming soon! 3511 [Ben Laurie] 3512 3513 *) Fix a security hole, that allows sessions to be reused in the wrong 3514 context thus bypassing client cert protection! All software that uses 3515 client certs and session caches in multiple contexts NEEDS PATCHING to 3516 allow session reuse! A fuller solution is in the works. 3517 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 3518 3519 *) Some more source tree cleanups (removed obsolete files 3520 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 3521 permission on "config" script to be executable) and a fix for the INSTALL 3522 document. 3523 [Ulf Moeller <ulf@fitug.de>] 3524 3525 *) Remove some legacy and erroneous uses of malloc, free instead of 3526 Malloc, Free. 3527 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 3528 3529 *) Make rsa_oaep_test return non-zero on error. 3530 [Ulf Moeller <ulf@fitug.de>] 3531 3532 *) Add support for native Solaris shared libraries. Configure 3533 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 3534 if someone would make that last step automatic. 3535 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 3536 3537 *) ctx_size was not built with the right compiler during "make links". Fixed. 3538 [Ben Laurie] 3539 3540 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 3541 except NULL ciphers". This means the default cipher list will no longer 3542 enable NULL ciphers. They need to be specifically enabled e.g. with 3543 the string "DEFAULT:eNULL". 3544 [Steve Henson] 3545 3546 *) Fix to RSA private encryption routines: if p < q then it would 3547 occasionally produce an invalid result. This will only happen with 3548 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 3549 [Steve Henson] 3550 3551 *) Be less restrictive and allow also `perl util/perlpath.pl 3552 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 3553 because this way one can also use an interpreter named `perl5' (which is 3554 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 3555 installed as `perl'). 3556 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 3557 3558 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 3559 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 3560 3561 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 3562 advapi32.lib to Win32 build and change the pem test comparision 3563 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 3564 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 3565 and crypto/des/ede_cbcm_enc.c. 3566 [Steve Henson] 3567 3568 *) DES quad checksum was broken on big-endian architectures. Fixed. 3569 [Ben Laurie] 3570 3571 *) Comment out two functions in bio.h that aren't implemented. Fix up the 3572 Win32 test batch file so it (might) work again. The Win32 test batch file 3573 is horrible: I feel ill.... 3574 [Steve Henson] 3575 3576 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 3577 in e_os.h. Audit of header files to check ANSI and non ANSI 3578 sections: 10 functions were absent from non ANSI section and not exported 3579 from Windows DLLs. Fixed up libeay.num for new functions. 3580 [Steve Henson] 3581 3582 *) Make `openssl version' output lines consistent. 3583 [Ralf S. Engelschall] 3584 3585 *) Fix Win32 symbol export lists for BIO functions: Added 3586 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 3587 to ms/libeay{16,32}.def. 3588 [Ralf S. Engelschall] 3589 3590 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 3591 fine under Unix and passes some trivial tests I've now added. But the 3592 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 3593 added to make sure no one expects that this stuff really works in the 3594 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 3595 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 3596 openssl_bio.xs. 3597 [Ralf S. Engelschall] 3598 3599 *) Fix the generation of two part addresses in perl. 3600 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 3601 3602 *) Add config entry for Linux on MIPS. 3603 [John Tobey <jtobey@channel1.com>] 3604 3605 *) Make links whenever Configure is run, unless we are on Windoze. 3606 [Ben Laurie] 3607 3608 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 3609 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 3610 in CRLs. 3611 [Steve Henson] 3612 3613 *) Add a useful kludge to allow package maintainers to specify compiler and 3614 other platforms details on the command line without having to patch the 3615 Configure script everytime: One now can use ``perl Configure 3616 <id>:<details>'', i.e. platform ids are allowed to have details appended 3617 to them (seperated by colons). This is treated as there would be a static 3618 pre-configured entry in Configure's %table under key <id> with value 3619 <details> and ``perl Configure <id>'' is called. So, when you want to 3620 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 3621 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 3622 now, which overrides the FreeBSD-elf entry on-the-fly. 3623 [Ralf S. Engelschall] 3624 3625 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 3626 [Ben Laurie] 3627 3628 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 3629 on the `perl Configure ...' command line. This way one can compile 3630 OpenSSL libraries with Position Independent Code (PIC) which is needed 3631 for linking it into DSOs. 3632 [Ralf S. Engelschall] 3633 3634 *) Remarkably, export ciphers were totally broken and no-one had noticed! 3635 Fixed. 3636 [Ben Laurie] 3637 3638 *) Cleaned up the LICENSE document: The official contact for any license 3639 questions now is the OpenSSL core team under openssl-core@openssl.org. 3640 And add a paragraph about the dual-license situation to make sure people 3641 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 3642 to the OpenSSL toolkit. 3643 [Ralf S. Engelschall] 3644 3645 *) General source tree makefile cleanups: Made `making xxx in yyy...' 3646 display consistent in the source tree and replaced `/bin/rm' by `rm'. 3647 Additonally cleaned up the `make links' target: Remove unnecessary 3648 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 3649 to speed processing and no longer clutter the display with confusing 3650 stuff. Instead only the actually done links are displayed. 3651 [Ralf S. Engelschall] 3652 3653 *) Permit null encryption ciphersuites, used for authentication only. It used 3654 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 3655 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 3656 encryption. 3657 [Ben Laurie] 3658 3659 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 3660 signed attributes when verifying signatures (this would break them), 3661 the detached data encoding was wrong and public keys obtained using 3662 X509_get_pubkey() weren't freed. 3663 [Steve Henson] 3664 3665 *) Add text documentation for the BUFFER functions. Also added a work around 3666 to a Win95 console bug. This was triggered by the password read stuff: the 3667 last character typed gets carried over to the next fread(). If you were 3668 generating a new cert request using 'req' for example then the last 3669 character of the passphrase would be CR which would then enter the first 3670 field as blank. 3671 [Steve Henson] 3672 3673 *) Added the new `Includes OpenSSL Cryptography Software' button as 3674 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 3675 button and can be used by applications based on OpenSSL to show the 3676 relationship to the OpenSSL project. 3677 [Ralf S. Engelschall] 3678 3679 *) Remove confusing variables in function signatures in files 3680 ssl/ssl_lib.c and ssl/ssl.h. 3681 [Lennart Bong <lob@kulthea.stacken.kth.se>] 3682 3683 *) Don't install bss_file.c under PREFIX/include/ 3684 [Lennart Bong <lob@kulthea.stacken.kth.se>] 3685 3686 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 3687 functions that return function pointers and has support for NT specific 3688 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 3689 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 3690 unsigned to signed types: this was killing the Win32 compile. 3691 [Steve Henson] 3692 3693 *) Add new certificate file to stack functions, 3694 SSL_add_dir_cert_subjects_to_stack() and 3695 SSL_add_file_cert_subjects_to_stack(). These largely supplant 3696 SSL_load_client_CA_file(), and can be used to add multiple certs easily 3697 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 3698 This means that Apache-SSL and similar packages don't have to mess around 3699 to add as many CAs as they want to the preferred list. 3700 [Ben Laurie] 3701 3702 *) Experiment with doxygen documentation. Currently only partially applied to 3703 ssl/ssl_lib.c. 3704 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 3705 openssl.doxy as the configuration file. 3706 [Ben Laurie] 3707 3708 *) Get rid of remaining C++-style comments which strict C compilers hate. 3709 [Ralf S. Engelschall, pointed out by Carlos Amengual] 3710 3711 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 3712 compiled in by default: it has problems with large keys. 3713 [Steve Henson] 3714 3715 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 3716 DH private keys and/or callback functions which directly correspond to 3717 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 3718 is needed for applications which have to configure certificates on a 3719 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 3720 (e.g. s_server). 3721 For the RSA certificate situation is makes no difference, but 3722 for the DSA certificate situation this fixes the "no shared cipher" 3723 problem where the OpenSSL cipher selection procedure failed because the 3724 temporary keys were not overtaken from the context and the API provided 3725 no way to reconfigure them. 3726 The new functions now let applications reconfigure the stuff and they 3727 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 3728 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 3729 non-public-API function ssl_cert_instantiate() is used as a helper 3730 function and also to reduce code redundancy inside ssl_rsa.c. 3731 [Ralf S. Engelschall] 3732 3733 *) Move s_server -dcert and -dkey options out of the undocumented feature 3734 area because they are useful for the DSA situation and should be 3735 recognized by the users. 3736 [Ralf S. Engelschall] 3737 3738 *) Fix the cipher decision scheme for export ciphers: the export bits are 3739 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 3740 SSL_EXP_MASK. So, the original variable has to be used instead of the 3741 already masked variable. 3742 [Richard Levitte <levitte@stacken.kth.se>] 3743 3744 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 3745 [Richard Levitte <levitte@stacken.kth.se>] 3746 3747 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 3748 from `int' to `unsigned int' because it's a length and initialized by 3749 EVP_DigestFinal() which expects an `unsigned int *'. 3750 [Richard Levitte <levitte@stacken.kth.se>] 3751 3752 *) Don't hard-code path to Perl interpreter on shebang line of Configure 3753 script. Instead use the usual Shell->Perl transition trick. 3754 [Ralf S. Engelschall] 3755 3756 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 3757 (in addition to RSA certificates) to match the behaviour of `openssl dsa 3758 -noout -modulus' as it's already the case for `openssl rsa -noout 3759 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 3760 currently the public key is printed (a decision which was already done by 3761 `openssl dsa -modulus' in the past) which serves a similar purpose. 3762 Additionally the NO_RSA no longer completely removes the whole -modulus 3763 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 3764 now, too. 3765 [Ralf S. Engelschall] 3766 3767 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 3768 BIO. See the source (crypto/evp/bio_ok.c) for more info. 3769 [Arne Ansper <arne@ats.cyber.ee>] 3770 3771 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 3772 to be added. Now both 'req' and 'ca' can use new objects defined in the 3773 config file. 3774 [Steve Henson] 3775 3776 *) Add cool BIO that does syslog (or event log on NT). 3777 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 3778 3779 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 3780 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 3781 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 3782 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 3783 [Ben Laurie] 3784 3785 *) Add preliminary config info for new extension code. 3786 [Steve Henson] 3787 3788 *) Make RSA_NO_PADDING really use no padding. 3789 [Ulf Moeller <ulf@fitug.de>] 3790 3791 *) Generate errors when private/public key check is done. 3792 [Ben Laurie] 3793 3794 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 3795 for some CRL extensions and new objects added. 3796 [Steve Henson] 3797 3798 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 3799 key usage extension and fuller support for authority key id. 3800 [Steve Henson] 3801 3802 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 3803 padding method for RSA, which is recommended for new applications in PKCS 3804 #1 v2.0 (RFC 2437, October 1998). 3805 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 3806 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 3807 against Bleichbacher's attack on RSA. 3808 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 3809 Ben Laurie] 3810 3811 *) Updates to the new SSL compression code 3812 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3813 3814 *) Fix so that the version number in the master secret, when passed 3815 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 3816 (because the server will not accept higher), that the version number 3817 is 0x03,0x01, not 0x03,0x00 3818 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3819 3820 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 3821 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 3822 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 3823 [Steve Henson] 3824 3825 *) Support for RAW extensions where an arbitrary extension can be 3826 created by including its DER encoding. See apps/openssl.cnf for 3827 an example. 3828 [Steve Henson] 3829 3830 *) Make sure latest Perl versions don't interpret some generated C array 3831 code as Perl array code in the crypto/err/err_genc.pl script. 3832 [Lars Weber <3weber@informatik.uni-hamburg.de>] 3833 3834 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 3835 not many people have the assembler. Various Win32 compilation fixes and 3836 update to the INSTALL.W32 file with (hopefully) more accurate Win32 3837 build instructions. 3838 [Steve Henson] 3839 3840 *) Modify configure script 'Configure' to automatically create crypto/date.h 3841 file under Win32 and also build pem.h from pem.org. New script 3842 util/mkfiles.pl to create the MINFO file on environments that can't do a 3843 'make files': perl util/mkfiles.pl >MINFO should work. 3844 [Steve Henson] 3845 3846 *) Major rework of DES function declarations, in the pursuit of correctness 3847 and purity. As a result, many evil casts evaporated, and some weirdness, 3848 too. You may find this causes warnings in your code. Zapping your evil 3849 casts will probably fix them. Mostly. 3850 [Ben Laurie] 3851 3852 *) Fix for a typo in asn1.h. Bug fix to object creation script 3853 obj_dat.pl. It considered a zero in an object definition to mean 3854 "end of object": none of the objects in objects.h have any zeros 3855 so it wasn't spotted. 3856 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 3857 3858 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 3859 Masking (CBCM). In the absence of test vectors, the best I have been able 3860 to do is check that the decrypt undoes the encrypt, so far. Send me test 3861 vectors if you have them. 3862 [Ben Laurie] 3863 3864 *) Correct calculation of key length for export ciphers (too much space was 3865 allocated for null ciphers). This has not been tested! 3866 [Ben Laurie] 3867 3868 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 3869 message is now correct (it understands "crypto" and "ssl" on its 3870 command line). There is also now an "update" option. This will update 3871 the util/ssleay.num and util/libeay.num files with any new functions. 3872 If you do a: 3873 perl util/mkdef.pl crypto ssl update 3874 it will update them. 3875 [Steve Henson] 3876 3877 *) Overhauled the Perl interface (perl/*): 3878 - ported BN stuff to OpenSSL's different BN library 3879 - made the perl/ source tree CVS-aware 3880 - renamed the package from SSLeay to OpenSSL (the files still contain 3881 their history because I've copied them in the repository) 3882 - removed obsolete files (the test scripts will be replaced 3883 by better Test::Harness variants in the future) 3884 [Ralf S. Engelschall] 3885 3886 *) First cut for a very conservative source tree cleanup: 3887 1. merge various obsolete readme texts into doc/ssleay.txt 3888 where we collect the old documents and readme texts. 3889 2. remove the first part of files where I'm already sure that we no 3890 longer need them because of three reasons: either they are just temporary 3891 files which were left by Eric or they are preserved original files where 3892 I've verified that the diff is also available in the CVS via "cvs diff 3893 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 3894 the crypto/md/ stuff). 3895 [Ralf S. Engelschall] 3896 3897 *) More extension code. Incomplete support for subject and issuer alt 3898 name, issuer and authority key id. Change the i2v function parameters 3899 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 3900 what that's for :-) Fix to ASN1 macro which messed up 3901 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 3902 [Steve Henson] 3903 3904 *) Preliminary support for ENUMERATED type. This is largely copied from the 3905 INTEGER code. 3906 [Steve Henson] 3907 3908 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 3909 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3910 3911 *) Make sure `make rehash' target really finds the `openssl' program. 3912 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 3913 3914 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 3915 like to hear about it if this slows down other processors. 3916 [Ben Laurie] 3917 3918 *) Add CygWin32 platform information to Configure script. 3919 [Alan Batie <batie@aahz.jf.intel.com>] 3920 3921 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 3922 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 3923 3924 *) New program nseq to manipulate netscape certificate sequences 3925 [Steve Henson] 3926 3927 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 3928 few typos. 3929 [Steve Henson] 3930 3931 *) Fixes to BN code. Previously the default was to define BN_RECURSION 3932 but the BN code had some problems that would cause failures when 3933 doing certificate verification and some other functions. 3934 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3935 3936 *) Add ASN1 and PEM code to support netscape certificate sequences. 3937 [Steve Henson] 3938 3939 *) Add ASN1 and PEM code to support netscape certificate sequences. 3940 [Steve Henson] 3941 3942 *) Add several PKIX and private extended key usage OIDs. 3943 [Steve Henson] 3944 3945 *) Modify the 'ca' program to handle the new extension code. Modify 3946 openssl.cnf for new extension format, add comments. 3947 [Steve Henson] 3948 3949 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 3950 and add a sample to openssl.cnf so req -x509 now adds appropriate 3951 CA extensions. 3952 [Steve Henson] 3953 3954 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 3955 error code, add initial support to X509_print() and x509 application. 3956 [Steve Henson] 3957 3958 *) Takes a deep breath and start addding X509 V3 extension support code. Add 3959 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 3960 stuff is currently isolated and isn't even compiled yet. 3961 [Steve Henson] 3962 3963 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 3964 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 3965 Removed the versions check from X509 routines when loading extensions: 3966 this allows certain broken certificates that don't set the version 3967 properly to be processed. 3968 [Steve Henson] 3969 3970 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 3971 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 3972 can still be regenerated with "make depend". 3973 [Ben Laurie] 3974 3975 *) Spelling mistake in C version of CAST-128. 3976 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 3977 3978 *) Changes to the error generation code. The perl script err-code.pl 3979 now reads in the old error codes and retains the old numbers, only 3980 adding new ones if necessary. It also only changes the .err files if new 3981 codes are added. The makefiles have been modified to only insert errors 3982 when needed (to avoid needlessly modifying header files). This is done 3983 by only inserting errors if the .err file is newer than the auto generated 3984 C file. To rebuild all the error codes from scratch (the old behaviour) 3985 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 3986 or delete all the .err files. 3987 [Steve Henson] 3988 3989 *) CAST-128 was incorrectly implemented for short keys. The C version has 3990 been fixed, but is untested. The assembler versions are also fixed, but 3991 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 3992 to regenerate it if needed. 3993 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 3994 Hagino <itojun@kame.net>] 3995 3996 *) File was opened incorrectly in randfile.c. 3997 [Ulf M�ller <ulf@fitug.de>] 3998 3999 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 4000 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 4001 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 4002 al: it's just almost always a UTCTime. Note this patch adds new error 4003 codes so do a "make errors" if there are problems. 4004 [Steve Henson] 4005 4006 *) Correct Linux 1 recognition in config. 4007 [Ulf M�ller <ulf@fitug.de>] 4008 4009 *) Remove pointless MD5 hash when using DSA keys in ca. 4010 [Anonymous <nobody@replay.com>] 4011 4012 *) Generate an error if given an empty string as a cert directory. Also 4013 generate an error if handed NULL (previously returned 0 to indicate an 4014 error, but didn't set one). 4015 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 4016 4017 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 4018 [Ben Laurie] 4019 4020 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 4021 parameters. This was causing a warning which killed off the Win32 compile. 4022 [Steve Henson] 4023 4024 *) Remove C++ style comments from crypto/bn/bn_local.h. 4025 [Neil Costigan <neil.costigan@celocom.com>] 4026 4027 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 4028 based on a text string, looking up short and long names and finally 4029 "dot" format. The "dot" format stuff didn't work. Added new function 4030 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 4031 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 4032 OID is not part of the table. 4033 [Steve Henson] 4034 4035 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 4036 X509_LOOKUP_by_alias(). 4037 [Ben Laurie] 4038 4039 *) Sort openssl functions by name. 4040 [Ben Laurie] 4041 4042 *) Get the gendsa program working (hopefully) and add it to app list. Remove 4043 encryption from sample DSA keys (in case anyone is interested the password 4044 was "1234"). 4045 [Steve Henson] 4046 4047 *) Make _all_ *_free functions accept a NULL pointer. 4048 [Frans Heymans <fheymans@isaserver.be>] 4049 4050 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 4051 NULL pointers. 4052 [Anonymous <nobody@replay.com>] 4053 4054 *) s_server should send the CAfile as acceptable CAs, not its own cert. 4055 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 4056 4057 *) Don't blow it for numeric -newkey arguments to apps/req. 4058 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 4059 4060 *) Temp key "for export" tests were wrong in s3_srvr.c. 4061 [Anonymous <nobody@replay.com>] 4062 4063 *) Add prototype for temp key callback functions 4064 SSL_CTX_set_tmp_{rsa,dh}_callback(). 4065 [Ben Laurie] 4066 4067 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 4068 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 4069 [Steve Henson] 4070 4071 *) X509_name_add_entry() freed the wrong thing after an error. 4072 [Arne Ansper <arne@ats.cyber.ee>] 4073 4074 *) rsa_eay.c would attempt to free a NULL context. 4075 [Arne Ansper <arne@ats.cyber.ee>] 4076 4077 *) BIO_s_socket() had a broken should_retry() on Windoze. 4078 [Arne Ansper <arne@ats.cyber.ee>] 4079 4080 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 4081 [Arne Ansper <arne@ats.cyber.ee>] 4082 4083 *) Make sure the already existing X509_STORE->depth variable is initialized 4084 in X509_STORE_new(), but document the fact that this variable is still 4085 unused in the certificate verification process. 4086 [Ralf S. Engelschall] 4087 4088 *) Fix the various library and apps files to free up pkeys obtained from 4089 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 4090 [Steve Henson] 4091 4092 *) Fix reference counting in X509_PUBKEY_get(). This makes 4093 demos/maurice/example2.c work, amongst others, probably. 4094 [Steve Henson and Ben Laurie] 4095 4096 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 4097 `openssl' and second, the shortcut symlinks for the `openssl <command>' 4098 are no longer created. This way we have a single and consistent command 4099 line interface `openssl <command>', similar to `cvs <command>'. 4100 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 4101 4102 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 4103 BIT STRING wrapper always have zero unused bits. 4104 [Steve Henson] 4105 4106 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 4107 [Steve Henson] 4108 4109 *) Make the top-level INSTALL documentation easier to understand. 4110 [Paul Sutton] 4111 4112 *) Makefiles updated to exit if an error occurs in a sub-directory 4113 make (including if user presses ^C) [Paul Sutton] 4114 4115 *) Make Montgomery context stuff explicit in RSA data structure. 4116 [Ben Laurie] 4117 4118 *) Fix build order of pem and err to allow for generated pem.h. 4119 [Ben Laurie] 4120 4121 *) Fix renumbering bug in X509_NAME_delete_entry(). 4122 [Ben Laurie] 4123 4124 *) Enhanced the err-ins.pl script so it makes the error library number 4125 global and can add a library name. This is needed for external ASN1 and 4126 other error libraries. 4127 [Steve Henson] 4128 4129 *) Fixed sk_insert which never worked properly. 4130 [Steve Henson] 4131 4132 *) Fix ASN1 macros so they can handle indefinite length construted 4133 EXPLICIT tags. Some non standard certificates use these: they can now 4134 be read in. 4135 [Steve Henson] 4136 4137 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 4138 into a single doc/ssleay.txt bundle. This way the information is still 4139 preserved but no longer messes up this directory. Now it's new room for 4140 the new set of documenation files. 4141 [Ralf S. Engelschall] 4142 4143 *) SETs were incorrectly DER encoded. This was a major pain, because they 4144 shared code with SEQUENCEs, which aren't coded the same. This means that 4145 almost everything to do with SETs or SEQUENCEs has either changed name or 4146 number of arguments. 4147 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 4148 4149 *) Fix test data to work with the above. 4150 [Ben Laurie] 4151 4152 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 4153 was already fixed by Eric for 0.9.1 it seems. 4154 [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 4155 4156 *) Autodetect FreeBSD3. 4157 [Ben Laurie] 4158 4159 *) Fix various bugs in Configure. This affects the following platforms: 4160 nextstep 4161 ncr-scde 4162 unixware-2.0 4163 unixware-2.0-pentium 4164 sco5-cc. 4165 [Ben Laurie] 4166 4167 *) Eliminate generated files from CVS. Reorder tests to regenerate files 4168 before they are needed. 4169 [Ben Laurie] 4170 4171 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 4172 [Ben Laurie] 4173 4174 4175 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 4176 4177 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 4178 changed SSLeay to OpenSSL in version strings. 4179 [Ralf S. Engelschall] 4180 4181 *) Some fixups to the top-level documents. 4182 [Paul Sutton] 4183 4184 *) Fixed the nasty bug where rsaref.h was not found under compile-time 4185 because the symlink to include/ was missing. 4186 [Ralf S. Engelschall] 4187 4188 *) Incorporated the popular no-RSA/DSA-only patches 4189 which allow to compile a RSA-free SSLeay. 4190 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 4191 4192 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 4193 when "ssleay" is still not found. 4194 [Ralf S. Engelschall] 4195 4196 *) Added more platforms to Configure: Cray T3E, HPUX 11, 4197 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 4198 4199 *) Updated the README file. 4200 [Ralf S. Engelschall] 4201 4202 *) Added various .cvsignore files in the CVS repository subdirs 4203 to make a "cvs update" really silent. 4204 [Ralf S. Engelschall] 4205 4206 *) Recompiled the error-definition header files and added 4207 missing symbols to the Win32 linker tables. 4208 [Ralf S. Engelschall] 4209 4210 *) Cleaned up the top-level documents; 4211 o new files: CHANGES and LICENSE 4212 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 4213 o merged COPYRIGHT into LICENSE 4214 o removed obsolete TODO file 4215 o renamed MICROSOFT to INSTALL.W32 4216 [Ralf S. Engelschall] 4217 4218 *) Removed dummy files from the 0.9.1b source tree: 4219 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 4220 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 4221 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 4222 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 4223 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 4224 [Ralf S. Engelschall] 4225 4226 *) Added various platform portability fixes. 4227 [Mark J. Cox] 4228 4229 *) The Genesis of the OpenSSL rpject: 4230 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 4231 Young and Tim J. Hudson created while they were working for C2Net until 4232 summer 1998. 4233 [The OpenSSL Project] 4234 4235 4236 Changes between 0.9.0b and 0.9.1b [not released] 4237 4238 *) Updated a few CA certificates under certs/ 4239 [Eric A. Young] 4240 4241 *) Changed some BIGNUM api stuff. 4242 [Eric A. Young] 4243 4244 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 4245 DGUX x86, Linux Alpha, etc. 4246 [Eric A. Young] 4247 4248 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 4249 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 4250 available). 4251 [Eric A. Young] 4252 4253 *) Add -strparse option to asn1pars program which parses nested 4254 binary structures 4255 [Dr Stephen Henson <shenson@bigfoot.com>] 4256 4257 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 4258 [Eric A. Young] 4259 4260 *) DSA fix for "ca" program. 4261 [Eric A. Young] 4262 4263 *) Added "-genkey" option to "dsaparam" program. 4264 [Eric A. Young] 4265 4266 *) Added RIPE MD160 (rmd160) message digest. 4267 [Eric A. Young] 4268 4269 *) Added -a (all) option to "ssleay version" command. 4270 [Eric A. Young] 4271 4272 *) Added PLATFORM define which is the id given to Configure. 4273 [Eric A. Young] 4274 4275 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 4276 [Eric A. Young] 4277 4278 *) Extended the ASN.1 parser routines. 4279 [Eric A. Young] 4280 4281 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 4282 [Eric A. Young] 4283 4284 *) Added a BN_CTX to the BN library. 4285 [Eric A. Young] 4286 4287 *) Fixed the weak key values in DES library 4288 [Eric A. Young] 4289 4290 *) Changed API in EVP library for cipher aliases. 4291 [Eric A. Young] 4292 4293 *) Added support for RC2/64bit cipher. 4294 [Eric A. Young] 4295 4296 *) Converted the lhash library to the crypto/mem.c functions. 4297 [Eric A. Young] 4298 4299 *) Added more recognized ASN.1 object ids. 4300 [Eric A. Young] 4301 4302 *) Added more RSA padding checks for SSL/TLS. 4303 [Eric A. Young] 4304 4305 *) Added BIO proxy/filter functionality. 4306 [Eric A. Young] 4307 4308 *) Added extra_certs to SSL_CTX which can be used 4309 send extra CA certificates to the client in the CA cert chain sending 4310 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 4311 [Eric A. Young] 4312 4313 *) Now Fortezza is denied in the authentication phase because 4314 this is key exchange mechanism is not supported by SSLeay at all. 4315 [Eric A. Young] 4316 4317 *) Additional PKCS1 checks. 4318 [Eric A. Young] 4319 4320 *) Support the string "TLSv1" for all TLS v1 ciphers. 4321 [Eric A. Young] 4322 4323 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 4324 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 4325 [Eric A. Young] 4326 4327 *) Fixed a few memory leaks. 4328 [Eric A. Young] 4329 4330 *) Fixed various code and comment typos. 4331 [Eric A. Young] 4332 4333 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 4334 bytes sent in the client random. 4335 [Edward Bishop <ebishop@spyglass.com>] 4336 4337