CHANGES revision 101618
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6 7 *) [In 0.9.6g-engine release:] 8 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 9 [Lynn Gazis <lgazis@rainbow.com>] 10 11 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 12 13 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 14 and get fix the header length calculation. 15 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 16 Alon Kantor <alonk@checkpoint.com> (and others), 17 Steve Henson] 18 19 *) Use proper error handling instead of 'assertions' in buffer 20 overflow checks added in 0.9.6e. This prevents DoS (the 21 assertions could call abort()). 22 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 23 24 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 25 26 *) Fix cipher selection routines: ciphers without encryption had no flags 27 for the cipher strength set and where therefore not handled correctly 28 by the selection routines (PR #130). 29 [Lutz Jaenicke] 30 31 *) Fix EVP_dsa_sha macro. 32 [Nils Larsch] 33 34 *) New option 35 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 36 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 37 that was added in OpenSSL 0.9.6d. 38 39 As the countermeasure turned out to be incompatible with some 40 broken SSL implementations, the new option is part of SSL_OP_ALL. 41 SSL_OP_ALL is usually employed when compatibility with weird SSL 42 implementations is desired (e.g. '-bugs' option to 's_client' and 43 's_server'), so the new option is automatically set in many 44 applications. 45 [Bodo Moeller] 46 47 *) Changes in security patch: 48 49 Changes marked "(CHATS)" were sponsored by the Defense Advanced 50 Research Projects Agency (DARPA) and Air Force Research Laboratory, 51 Air Force Materiel Command, USAF, under agreement number 52 F30602-01-2-0537. 53 54 *) Add various sanity checks to asn1_get_length() to reject 55 the ASN1 length bytes if they exceed sizeof(long), will appear 56 negative or the content length exceeds the length of the 57 supplied buffer. 58 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 59 60 *) Assertions for various potential buffer overflows, not known to 61 happen in practice. 62 [Ben Laurie (CHATS)] 63 64 *) Various temporary buffers to hold ASCII versions of integers were 65 too small for 64 bit platforms. (CAN-2002-0655) 66 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 67 68 *) Remote buffer overflow in SSL3 protocol - an attacker could 69 supply an oversized session ID to a client. (CAN-2002-0656) 70 [Ben Laurie (CHATS)] 71 72 *) Remote buffer overflow in SSL2 protocol - an attacker could 73 supply an oversized client master key. (CAN-2002-0656) 74 [Ben Laurie (CHATS)] 75 76 Changes between 0.9.6c and 0.9.6d [9 May 2002] 77 78 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 79 encoded as NULL) with id-dsa-with-sha1. 80 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 81 82 *) Check various X509_...() return values in apps/req.c. 83 [Nils Larsch <nla@trustcenter.de>] 84 85 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 86 an end-of-file condition would erronously be flagged, when the CRLF 87 was just at the end of a processed block. The bug was discovered when 88 processing data through a buffering memory BIO handing the data to a 89 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 90 <ptsekov@syntrex.com> and Nedelcho Stanev. 91 [Lutz Jaenicke] 92 93 *) Implement a countermeasure against a vulnerability recently found 94 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 95 before application data chunks to avoid the use of known IVs 96 with data potentially chosen by the attacker. 97 [Bodo Moeller] 98 99 *) Fix length checks in ssl3_get_client_hello(). 100 [Bodo Moeller] 101 102 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 103 to prevent ssl3_read_internal() from incorrectly assuming that 104 ssl3_read_bytes() found application data while handshake 105 processing was enabled when in fact s->s3->in_read_app_data was 106 merely automatically cleared during the initial handshake. 107 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 108 109 *) Fix object definitions for Private and Enterprise: they were not 110 recognized in their shortname (=lowercase) representation. Extend 111 obj_dat.pl to issue an error when using undefined keywords instead 112 of silently ignoring the problem (Svenning Sorensen 113 <sss@sss.dnsalias.net>). 114 [Lutz Jaenicke] 115 116 *) Fix DH_generate_parameters() so that it works for 'non-standard' 117 generators, i.e. generators other than 2 and 5. (Previously, the 118 code did not properly initialise the 'add' and 'rem' values to 119 BN_generate_prime().) 120 121 In the new general case, we do not insist that 'generator' is 122 actually a primitive root: This requirement is rather pointless; 123 a generator of the order-q subgroup is just as good, if not 124 better. 125 [Bodo Moeller] 126 127 *) Map new X509 verification errors to alerts. Discovered and submitted by 128 Tom Wu <tom@arcot.com>. 129 [Lutz Jaenicke] 130 131 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 132 returning non-zero before the data has been completely received 133 when using non-blocking I/O. 134 [Bodo Moeller; problem pointed out by John Hughes] 135 136 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 137 [Ben Laurie, Lutz Jaenicke] 138 139 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 140 Yoram Zahavi <YoramZ@gilian.com>). 141 [Lutz Jaenicke] 142 143 *) Add information about CygWin 1.3 and on, and preserve proper 144 configuration for the versions before that. 145 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 146 147 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 148 check whether we deal with a copy of a session and do not delete from 149 the cache in this case. Problem reported by "Izhar Shoshani Levi" 150 <izhar@checkpoint.com>. 151 [Lutz Jaenicke] 152 153 *) Do not store session data into the internal session cache, if it 154 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 155 flag is set). Proposed by Aslam <aslam@funk.com>. 156 [Lutz Jaenicke] 157 158 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 159 value is 0. 160 [Richard Levitte] 161 162 *) [In 0.9.6c-engine release:] 163 Fix a crashbug and a logic bug in hwcrhk_load_pubkey() 164 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 165 166 *) Add the configuration target linux-s390x. 167 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 168 169 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 170 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 171 variable as an indication that a ClientHello message has been 172 received. As the flag value will be lost between multiple 173 invocations of ssl3_accept when using non-blocking I/O, the 174 function may not be aware that a handshake has actually taken 175 place, thus preventing a new session from being added to the 176 session cache. 177 178 To avoid this problem, we now set s->new_session to 2 instead of 179 using a local variable. 180 [Lutz Jaenicke, Bodo Moeller] 181 182 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 183 if the SSL_R_LENGTH_MISMATCH error is detected. 184 [Geoff Thorpe, Bodo Moeller] 185 186 *) New 'shared_ldflag' column in Configure platform table. 187 [Richard Levitte] 188 189 *) Fix EVP_CIPHER_mode macro. 190 ["Dan S. Camper" <dan@bti.net>] 191 192 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 193 type, we must throw them away by setting rr->length to 0. 194 [D P Chang <dpc@qualys.com>] 195 196 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 197 198 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 199 <Dominikus.Scherkl@biodata.com>. (The previous implementation 200 worked incorrectly for those cases where range = 10..._2 and 201 3*range is two bits longer than range.) 202 [Bodo Moeller] 203 204 *) Only add signing time to PKCS7 structures if it is not already 205 present. 206 [Steve Henson] 207 208 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 209 OBJ_ld_ce should be OBJ_id_ce. 210 Also some ip-pda OIDs in crypto/objects/objects.txt were 211 incorrect (cf. RFC 3039). 212 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 213 214 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 215 returns early because it has nothing to do. 216 [Andy Schneider <andy.schneider@bjss.co.uk>] 217 218 *) [In 0.9.6c-engine release:] 219 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 220 [Andy Schneider <andy.schneider@bjss.co.uk>] 221 222 *) [In 0.9.6c-engine release:] 223 Add support for Cryptographic Appliance's keyserver technology. 224 (Use engine 'keyclient') 225 [Cryptographic Appliances and Geoff Thorpe] 226 227 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 228 is called via tools/c89.sh because arguments have to be 229 rearranged (all '-L' options must appear before the first object 230 modules). 231 [Richard Shapiro <rshapiro@abinitio.com>] 232 233 *) [In 0.9.6c-engine release:] 234 Add support for Broadcom crypto accelerator cards, backported 235 from 0.9.7. 236 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 237 238 *) [In 0.9.6c-engine release:] 239 Add support for SureWare crypto accelerator cards from 240 Baltimore Technologies. (Use engine 'sureware') 241 [Baltimore Technologies and Mark Cox] 242 243 *) [In 0.9.6c-engine release:] 244 Add support for crypto accelerator cards from Accelerated 245 Encryption Processing, www.aep.ie. (Use engine 'aep') 246 [AEP Inc. and Mark Cox] 247 248 *) Add a configuration entry for gcc on UnixWare. 249 [Gary Benson <gbenson@redhat.com>] 250 251 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 252 messages are stored in a single piece (fixed-length part and 253 variable-length part combined) and fix various bugs found on the way. 254 [Bodo Moeller] 255 256 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 257 instead. BIO_gethostbyname() does not know what timeouts are 258 appropriate, so entries would stay in cache even when they have 259 become invalid. 260 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 261 262 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 263 faced with a pathologically small ClientHello fragment that does 264 not contain client_version: Instead of aborting with an error, 265 simply choose the highest available protocol version (i.e., 266 TLS 1.0 unless it is disabled). In practice, ClientHello 267 messages are never sent like this, but this change gives us 268 strictly correct behaviour at least for TLS. 269 [Bodo Moeller] 270 271 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 272 never resets s->method to s->ctx->method when called from within 273 one of the SSL handshake functions. 274 [Bodo Moeller; problem pointed out by Niko Baric] 275 276 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 277 (sent using the client's version number) if client_version is 278 smaller than the protocol version in use. Also change 279 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 280 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 281 the client will at least see that alert. 282 [Bodo Moeller] 283 284 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 285 correctly. 286 [Bodo Moeller] 287 288 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 289 client receives HelloRequest while in a handshake. 290 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 291 292 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 293 should end in 'break', not 'goto end' which circuments various 294 cleanups done in state SSL_ST_OK. But session related stuff 295 must be disabled for SSL_ST_OK in the case that we just sent a 296 HelloRequest. 297 298 Also avoid some overhead by not calling ssl_init_wbio_buffer() 299 before just sending a HelloRequest. 300 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 301 302 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 303 reveal whether illegal block cipher padding was found or a MAC 304 verification error occured. (Neither SSLerr() codes nor alerts 305 are directly visible to potential attackers, but the information 306 may leak via logfiles.) 307 308 Similar changes are not required for the SSL 2.0 implementation 309 because the number of padding bytes is sent in clear for SSL 2.0, 310 and the extra bytes are just ignored. However ssl/s2_pkt.c 311 failed to verify that the purported number of padding bytes is in 312 the legal range. 313 [Bodo Moeller] 314 315 *) Add OpenUNIX-8 support including shared libraries 316 (Boyd Lynn Gerber <gerberb@zenez.com>). 317 [Lutz Jaenicke] 318 319 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 320 'wristwatch attack' using huge encoding parameters (cf. 321 James H. Manger's CRYPTO 2001 paper). Note that the 322 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 323 encoding parameters and hence was not vulnerable. 324 [Bodo Moeller] 325 326 *) BN_sqr() bug fix. 327 [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 328 329 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 330 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 331 followed by modular reduction. 332 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 333 334 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 335 equivalent based on BN_pseudo_rand() instead of BN_rand(). 336 [Bodo Moeller] 337 338 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 339 This function was broken, as the check for a new client hello message 340 to handle SGC did not allow these large messages. 341 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 342 [Lutz Jaenicke] 343 344 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 345 [Lutz Jaenicke] 346 347 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 348 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 349 [Lutz Jaenicke] 350 351 *) Rework the configuration and shared library support for Tru64 Unix. 352 The configuration part makes use of modern compiler features and 353 still retains old compiler behavior for those that run older versions 354 of the OS. The shared library support part includes a variant that 355 uses the RPATH feature, and is available through the special 356 configuration target "alpha-cc-rpath", which will never be selected 357 automatically. 358 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 359 360 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 361 with the same message size as in ssl3_get_certificate_request(). 362 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 363 messages might inadvertently be reject as too long. 364 [Petr Lampa <lampa@fee.vutbr.cz>] 365 366 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 367 [Andy Polyakov] 368 369 *) Modified SSL library such that the verify_callback that has been set 370 specificly for an SSL object with SSL_set_verify() is actually being 371 used. Before the change, a verify_callback set with this function was 372 ignored and the verify_callback() set in the SSL_CTX at the time of 373 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 374 to allow the necessary settings. 375 [Lutz Jaenicke] 376 377 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 378 explicitly to NULL, as at least on Solaris 8 this seems not always to be 379 done automatically (in contradiction to the requirements of the C 380 standard). This made problems when used from OpenSSH. 381 [Lutz Jaenicke] 382 383 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 384 dh->length and always used 385 386 BN_rand_range(priv_key, dh->p). 387 388 BN_rand_range() is not necessary for Diffie-Hellman, and this 389 specific range makes Diffie-Hellman unnecessarily inefficient if 390 dh->length (recommended exponent length) is much smaller than the 391 length of dh->p. We could use BN_rand_range() if the order of 392 the subgroup was stored in the DH structure, but we only have 393 dh->length. 394 395 So switch back to 396 397 BN_rand(priv_key, l, ...) 398 399 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 400 otherwise. 401 [Bodo Moeller] 402 403 *) In 404 405 RSA_eay_public_encrypt 406 RSA_eay_private_decrypt 407 RSA_eay_private_encrypt (signing) 408 RSA_eay_public_decrypt (signature verification) 409 410 (default implementations for RSA_public_encrypt, 411 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 412 always reject numbers >= n. 413 [Bodo Moeller] 414 415 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 416 to synchronize access to 'locking_thread'. This is necessary on 417 systems where access to 'locking_thread' (an 'unsigned long' 418 variable) is not atomic. 419 [Bodo Moeller] 420 421 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 422 *before* setting the 'crypto_lock_rand' flag. The previous code had 423 a race condition if 0 is a valid thread ID. 424 [Travis Vitek <vitek@roguewave.com>] 425 426 *) Add support for shared libraries under Irix. 427 [Albert Chin-A-Young <china@thewrittenword.com>] 428 429 *) Add configuration option to build on Linux on both big-endian and 430 little-endian MIPS. 431 [Ralf Baechle <ralf@uni-koblenz.de>] 432 433 *) Add the possibility to create shared libraries on HP-UX. 434 [Richard Levitte] 435 436 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 437 438 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 439 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 440 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 441 PRNG state recovery was possible based on the output of 442 one PRNG request appropriately sized to gain knowledge on 443 'md' followed by enough consecutive 1-byte PRNG requests 444 to traverse all of 'state'. 445 446 1. When updating 'md_local' (the current thread's copy of 'md') 447 during PRNG output generation, hash all of the previous 448 'md_local' value, not just the half used for PRNG output. 449 450 2. Make the number of bytes from 'state' included into the hash 451 independent from the number of PRNG bytes requested. 452 453 The first measure alone would be sufficient to avoid 454 Markku-Juhani's attack. (Actually it had never occurred 455 to me that the half of 'md_local' used for chaining was the 456 half from which PRNG output bytes were taken -- I had always 457 assumed that the secret half would be used.) The second 458 measure makes sure that additional data from 'state' is never 459 mixed into 'md_local' in small portions; this heuristically 460 further strengthens the PRNG. 461 [Bodo Moeller] 462 463 *) Fix crypto/bn/asm/mips3.s. 464 [Andy Polyakov] 465 466 *) When only the key is given to "enc", the IV is undefined. Print out 467 an error message in this case. 468 [Lutz Jaenicke] 469 470 *) Handle special case when X509_NAME is empty in X509 printing routines. 471 [Steve Henson] 472 473 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 474 positive and less than q. 475 [Bodo Moeller] 476 477 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 478 used: it isn't thread safe and the add_lock_callback should handle 479 that itself. 480 [Paul Rose <Paul.Rose@bridge.com>] 481 482 *) Verify that incoming data obeys the block size in 483 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 484 [Bodo Moeller] 485 486 *) Fix OAEP check. 487 [Ulf M�ller, Bodo M�ller] 488 489 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 490 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 491 when fixing the server behaviour for backwards-compatible 'client 492 hello' messages. (Note that the attack is impractical against 493 SSL 3.0 and TLS 1.0 anyway because length and version checking 494 means that the probability of guessing a valid ciphertext is 495 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 496 paper.) 497 498 Before 0.9.5, the countermeasure (hide the error by generating a 499 random 'decryption result') did not work properly because 500 ERR_clear_error() was missing, meaning that SSL_get_error() would 501 detect the supposedly ignored error. 502 503 Both problems are now fixed. 504 [Bodo Moeller] 505 506 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 507 (previously it was 1024). 508 [Bodo Moeller] 509 510 *) Fix for compatibility mode trust settings: ignore trust settings 511 unless some valid trust or reject settings are present. 512 [Steve Henson] 513 514 *) Fix for blowfish EVP: its a variable length cipher. 515 [Steve Henson] 516 517 *) Fix various bugs related to DSA S/MIME verification. Handle missing 518 parameters in DSA public key structures and return an error in the 519 DSA routines if parameters are absent. 520 [Steve Henson] 521 522 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 523 in the current directory if neither $RANDFILE nor $HOME was set. 524 RAND_file_name() in 0.9.6a returned NULL in this case. This has 525 caused some confusion to Windows users who haven't defined $HOME. 526 Thus RAND_file_name() is changed again: e_os.h can define a 527 DEFAULT_HOME, which will be used if $HOME is not set. 528 For Windows, we use "C:"; on other platforms, we still require 529 environment variables. 530 531 *) Move 'if (!initialized) RAND_poll()' into regions protected by 532 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 533 having multiple threads call RAND_poll() concurrently. 534 [Bodo Moeller] 535 536 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 537 combination of a flag and a thread ID variable. 538 Otherwise while one thread is in ssleay_rand_bytes (which sets the 539 flag), *other* threads can enter ssleay_add_bytes without obeying 540 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 541 that they do not hold after the first thread unsets add_do_not_lock). 542 [Bodo Moeller] 543 544 *) Change bctest again: '-x' expressions are not available in all 545 versions of 'test'. 546 [Bodo Moeller] 547 548 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 549 550 *) Fix a couple of memory leaks in PKCS7_dataDecode() 551 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 552 553 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 554 the default extension for executables, if any. Also, make the perl 555 scripts that use symlink() to test if it really exists and use "cp" 556 if it doesn't. All this made OpenSSL compilable and installable in 557 CygWin. 558 [Richard Levitte] 559 560 *) Fix for asn1_GetSequence() for indefinite length constructed data. 561 If SEQUENCE is length is indefinite just set c->slen to the total 562 amount of data available. 563 [Steve Henson, reported by shige@FreeBSD.org] 564 [This change does not apply to 0.9.7.] 565 566 *) Change bctest to avoid here-documents inside command substitution 567 (workaround for FreeBSD /bin/sh bug). 568 For compatibility with Ultrix, avoid shell functions (introduced 569 in the bctest version that searches along $PATH). 570 [Bodo Moeller] 571 572 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 573 with des_encrypt() defined on some operating systems, like Solaris 574 and UnixWare. 575 [Richard Levitte] 576 577 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 578 On the Importance of Eliminating Errors in Cryptographic 579 Computations, J. Cryptology 14 (2001) 2, 101-119, 580 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 581 [Ulf Moeller] 582 583 *) MIPS assembler BIGNUM division bug fix. 584 [Andy Polyakov] 585 586 *) Disabled incorrect Alpha assembler code. 587 [Richard Levitte] 588 589 *) Fix PKCS#7 decode routines so they correctly update the length 590 after reading an EOC for the EXPLICIT tag. 591 [Steve Henson] 592 [This change does not apply to 0.9.7.] 593 594 *) Fix bug in PKCS#12 key generation routines. This was triggered 595 if a 3DES key was generated with a 0 initial byte. Include 596 PKCS12_BROKEN_KEYGEN compilation option to retain the old 597 (but broken) behaviour. 598 [Steve Henson] 599 600 *) Enhance bctest to search for a working bc along $PATH and print 601 it when found. 602 [Tim Rice <tim@multitalents.net> via Richard Levitte] 603 604 *) Fix memory leaks in err.c: free err_data string if necessary; 605 don't write to the wrong index in ERR_set_error_data. 606 [Bodo Moeller] 607 608 *) Implement ssl23_peek (analogous to ssl23_read), which previously 609 did not exist. 610 [Bodo Moeller] 611 612 *) Replace rdtsc with _emit statements for VC++ version 5. 613 [Jeremy Cooper <jeremy@baymoo.org>] 614 615 *) Make it possible to reuse SSLv2 sessions. 616 [Richard Levitte] 617 618 *) In copy_email() check for >= 0 as a return value for 619 X509_NAME_get_index_by_NID() since 0 is a valid index. 620 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 621 622 *) Avoid coredump with unsupported or invalid public keys by checking if 623 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 624 PKCS7_verify() fails with non detached data. 625 [Steve Henson] 626 627 *) Don't use getenv in library functions when run as setuid/setgid. 628 New function OPENSSL_issetugid(). 629 [Ulf Moeller] 630 631 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 632 due to incorrect handling of multi-threading: 633 634 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 635 636 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 637 638 3. Count how many times MemCheck_off() has been called so that 639 nested use can be treated correctly. This also avoids 640 inband-signalling in the previous code (which relied on the 641 assumption that thread ID 0 is impossible). 642 [Bodo Moeller] 643 644 *) Add "-rand" option also to s_client and s_server. 645 [Lutz Jaenicke] 646 647 *) Fix CPU detection on Irix 6.x. 648 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 649 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 650 651 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 652 was empty. 653 [Steve Henson] 654 [This change does not apply to 0.9.7.] 655 656 *) Use the cached encoding of an X509_NAME structure rather than 657 copying it. This is apparently the reason for the libsafe "errors" 658 but the code is actually correct. 659 [Steve Henson] 660 661 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 662 Bleichenbacher's DSA attack. 663 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 664 to be set and top=0 forces the highest bit to be set; top=-1 is new 665 and leaves the highest bit random. 666 [Ulf Moeller, Bodo Moeller] 667 668 *) In the NCONF_...-based implementations for CONF_... queries 669 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 670 a temporary CONF structure with the data component set to NULL 671 (which gives segmentation faults in lh_retrieve). 672 Instead, use NULL for the CONF pointer in CONF_get_string and 673 CONF_get_number (which may use environment variables) and directly 674 return NULL from CONF_get_section. 675 [Bodo Moeller] 676 677 *) Fix potential buffer overrun for EBCDIC. 678 [Ulf Moeller] 679 680 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 681 keyUsage if basicConstraints absent for a CA. 682 [Steve Henson] 683 684 *) Make SMIME_write_PKCS7() write mail header values with a format that 685 is more generally accepted (no spaces before the semicolon), since 686 some programs can't parse those values properly otherwise. Also make 687 sure BIO's that break lines after each write do not create invalid 688 headers. 689 [Richard Levitte] 690 691 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 692 macros previously used would not encode an empty SEQUENCE OF 693 and break the signature. 694 [Steve Henson] 695 [This change does not apply to 0.9.7.] 696 697 *) Zero the premaster secret after deriving the master secret in 698 DH ciphersuites. 699 [Steve Henson] 700 701 *) Add some EVP_add_digest_alias registrations (as found in 702 OpenSSL_add_all_digests()) to SSL_library_init() 703 aka OpenSSL_add_ssl_algorithms(). This provides improved 704 compatibility with peers using X.509 certificates 705 with unconventional AlgorithmIdentifier OIDs. 706 [Bodo Moeller] 707 708 *) Fix for Irix with NO_ASM. 709 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 710 711 *) ./config script fixes. 712 [Ulf Moeller, Richard Levitte] 713 714 *) Fix 'openssl passwd -1'. 715 [Bodo Moeller] 716 717 *) Change PKCS12_key_gen_asc() so it can cope with non null 718 terminated strings whose length is passed in the passlen 719 parameter, for example from PEM callbacks. This was done 720 by adding an extra length parameter to asc2uni(). 721 [Steve Henson, reported by <oddissey@samsung.co.kr>] 722 723 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 724 call failed, free the DSA structure. 725 [Bodo Moeller] 726 727 *) Fix to uni2asc() to cope with zero length Unicode strings. 728 These are present in some PKCS#12 files. 729 [Steve Henson] 730 731 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 732 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 733 when writing a 32767 byte record. 734 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 735 736 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 737 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 738 739 (RSA objects have a reference count access to which is protected 740 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 741 so they are meant to be shared between threads.) 742 [Bodo Moeller, Geoff Thorpe; original patch submitted by 743 "Reddie, Steven" <Steven.Reddie@ca.com>] 744 745 *) Fix a deadlock in CRYPTO_mem_leaks(). 746 [Bodo Moeller] 747 748 *) Use better test patterns in bntest. 749 [Ulf M�ller] 750 751 *) rand_win.c fix for Borland C. 752 [Ulf M�ller] 753 754 *) BN_rshift bugfix for n == 0. 755 [Bodo Moeller] 756 757 *) Add a 'bctest' script that checks for some known 'bc' bugs 758 so that 'make test' does not abort just because 'bc' is broken. 759 [Bodo Moeller] 760 761 *) Store verify_result within SSL_SESSION also for client side to 762 avoid potential security hole. (Re-used sessions on the client side 763 always resulted in verify_result==X509_V_OK, not using the original 764 result of the server certificate verification.) 765 [Lutz Jaenicke] 766 767 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 768 SSL3_RT_APPLICATION_DATA, return 0. 769 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 770 [Bodo Moeller] 771 772 *) Fix SSL_peek: 773 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 774 releases, have been re-implemented by renaming the previous 775 implementations of ssl2_read and ssl3_read to ssl2_read_internal 776 and ssl3_read_internal, respectively, and adding 'peek' parameters 777 to them. The new ssl[23]_{read,peek} functions are calls to 778 ssl[23]_read_internal with the 'peek' flag set appropriately. 779 A 'peek' parameter has also been added to ssl3_read_bytes, which 780 does the actual work for ssl3_read_internal. 781 [Bodo Moeller] 782 783 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 784 the method-specific "init()" handler. Also clean up ex_data after 785 calling the method-specific "finish()" handler. Previously, this was 786 happening the other way round. 787 [Geoff Thorpe] 788 789 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 790 The previous value, 12, was not always sufficient for BN_mod_exp(). 791 [Bodo Moeller] 792 793 *) Make sure that shared libraries get the internal name engine with 794 the full version number and not just 0. This should mark the 795 shared libraries as not backward compatible. Of course, this should 796 be changed again when we can guarantee backward binary compatibility. 797 [Richard Levitte] 798 799 *) Fix typo in get_cert_by_subject() in by_dir.c 800 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 801 802 *) Rework the system to generate shared libraries: 803 804 - Make note of the expected extension for the shared libraries and 805 if there is a need for symbolic links from for example libcrypto.so.0 806 to libcrypto.so.0.9.7. There is extended info in Configure for 807 that. 808 809 - Make as few rebuilds of the shared libraries as possible. 810 811 - Still avoid linking the OpenSSL programs with the shared libraries. 812 813 - When installing, install the shared libraries separately from the 814 static ones. 815 [Richard Levitte] 816 817 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 818 819 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 820 and not in SSL_clear because the latter is also used by the 821 accept/connect functions; previously, the settings made by 822 SSL_set_read_ahead would be lost during the handshake. 823 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 824 825 *) Correct util/mkdef.pl to be selective about disabled algorithms. 826 Previously, it would create entries for disableed algorithms no 827 matter what. 828 [Richard Levitte] 829 830 *) Added several new manual pages for SSL_* function. 831 [Lutz Jaenicke] 832 833 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 834 835 *) In ssl23_get_client_hello, generate an error message when faced 836 with an initial SSL 3.0/TLS record that is too small to contain the 837 first two bytes of the ClientHello message, i.e. client_version. 838 (Note that this is a pathologic case that probably has never happened 839 in real life.) The previous approach was to use the version number 840 from the record header as a substitute; but our protocol choice 841 should not depend on that one because it is not authenticated 842 by the Finished messages. 843 [Bodo Moeller] 844 845 *) More robust randomness gathering functions for Windows. 846 [Jeffrey Altman <jaltman@columbia.edu>] 847 848 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 849 not set then we don't setup the error code for issuer check errors 850 to avoid possibly overwriting other errors which the callback does 851 handle. If an application does set the flag then we assume it knows 852 what it is doing and can handle the new informational codes 853 appropriately. 854 [Steve Henson] 855 856 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 857 a general "ANY" type, as such it should be able to decode anything 858 including tagged types. However it didn't check the class so it would 859 wrongly interpret tagged types in the same way as their universal 860 counterpart and unknown types were just rejected. Changed so that the 861 tagged and unknown types are handled in the same way as a SEQUENCE: 862 that is the encoding is stored intact. There is also a new type 863 "V_ASN1_OTHER" which is used when the class is not universal, in this 864 case we have no idea what the actual type is so we just lump them all 865 together. 866 [Steve Henson] 867 868 *) On VMS, stdout may very well lead to a file that is written to 869 in a record-oriented fashion. That means that every write() will 870 write a separate record, which will be read separately by the 871 programs trying to read from it. This can be very confusing. 872 873 The solution is to put a BIO filter in the way that will buffer 874 text until a linefeed is reached, and then write everything a 875 line at a time, so every record written will be an actual line, 876 not chunks of lines and not (usually doesn't happen, but I've 877 seen it once) several lines in one record. BIO_f_linebuffer() is 878 the answer. 879 880 Currently, it's a VMS-only method, because that's where it has 881 been tested well enough. 882 [Richard Levitte] 883 884 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 885 it can return incorrect results. 886 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 887 but it was in 0.9.6-beta[12].) 888 [Bodo Moeller] 889 890 *) Disable the check for content being present when verifying detached 891 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 892 include zero length content when signing messages. 893 [Steve Henson] 894 895 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 896 BIO_ctrl (for BIO pairs). 897 [Bodo M�ller] 898 899 *) Add DSO method for VMS. 900 [Richard Levitte] 901 902 *) Bug fix: Montgomery multiplication could produce results with the 903 wrong sign. 904 [Ulf M�ller] 905 906 *) Add RPM specification openssl.spec and modify it to build three 907 packages. The default package contains applications, application 908 documentation and run-time libraries. The devel package contains 909 include files, static libraries and function documentation. The 910 doc package contains the contents of the doc directory. The original 911 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 912 [Richard Levitte] 913 914 *) Add a large number of documentation files for many SSL routines. 915 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 916 917 *) Add a configuration entry for Sony News 4. 918 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 919 920 *) Don't set the two most significant bits to one when generating a 921 random number < q in the DSA library. 922 [Ulf M�ller] 923 924 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 925 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 926 the underlying transport is blocking) if a handshake took place. 927 (The default behaviour is needed by applications such as s_client 928 and s_server that use select() to determine when to use SSL_read; 929 but for applications that know in advance when to expect data, it 930 just makes things more complicated.) 931 [Bodo Moeller] 932 933 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 934 from EGD. 935 [Ben Laurie] 936 937 *) Add a few more EBCDIC conditionals that make `req' and `x509' 938 work better on such systems. 939 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 940 941 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 942 Update PKCS12_parse() so it copies the friendlyName and the 943 keyid to the certificates aux info. 944 [Steve Henson] 945 946 *) Fix bug in PKCS7_verify() which caused an infinite loop 947 if there was more than one signature. 948 [Sven Uszpelkat <su@celocom.de>] 949 950 *) Major change in util/mkdef.pl to include extra information 951 about each symbol, as well as presentig variables as well 952 as functions. This change means that there's n more need 953 to rebuild the .num files when some algorithms are excluded. 954 [Richard Levitte] 955 956 *) Allow the verify time to be set by an application, 957 rather than always using the current time. 958 [Steve Henson] 959 960 *) Phase 2 verify code reorganisation. The certificate 961 verify code now looks up an issuer certificate by a 962 number of criteria: subject name, authority key id 963 and key usage. It also verifies self signed certificates 964 by the same criteria. The main comparison function is 965 X509_check_issued() which performs these checks. 966 967 Lot of changes were necessary in order to support this 968 without completely rewriting the lookup code. 969 970 Authority and subject key identifier are now cached. 971 972 The LHASH 'certs' is X509_STORE has now been replaced 973 by a STACK_OF(X509_OBJECT). This is mainly because an 974 LHASH can't store or retrieve multiple objects with 975 the same hash value. 976 977 As a result various functions (which were all internal 978 use only) have changed to handle the new X509_STORE 979 structure. This will break anything that messed round 980 with X509_STORE internally. 981 982 The functions X509_STORE_add_cert() now checks for an 983 exact match, rather than just subject name. 984 985 The X509_STORE API doesn't directly support the retrieval 986 of multiple certificates matching a given criteria, however 987 this can be worked round by performing a lookup first 988 (which will fill the cache with candidate certificates) 989 and then examining the cache for matches. This is probably 990 the best we can do without throwing out X509_LOOKUP 991 entirely (maybe later...). 992 993 The X509_VERIFY_CTX structure has been enhanced considerably. 994 995 All certificate lookup operations now go via a get_issuer() 996 callback. Although this currently uses an X509_STORE it 997 can be replaced by custom lookups. This is a simple way 998 to bypass the X509_STORE hackery necessary to make this 999 work and makes it possible to use more efficient techniques 1000 in future. A very simple version which uses a simple 1001 STACK for its trusted certificate store is also provided 1002 using X509_STORE_CTX_trusted_stack(). 1003 1004 The verify_cb() and verify() callbacks now have equivalents 1005 in the X509_STORE_CTX structure. 1006 1007 X509_STORE_CTX also has a 'flags' field which can be used 1008 to customise the verify behaviour. 1009 [Steve Henson] 1010 1011 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 1012 excludes S/MIME capabilities. 1013 [Steve Henson] 1014 1015 *) When a certificate request is read in keep a copy of the 1016 original encoding of the signed data and use it when outputing 1017 again. Signatures then use the original encoding rather than 1018 a decoded, encoded version which may cause problems if the 1019 request is improperly encoded. 1020 [Steve Henson] 1021 1022 *) For consistency with other BIO_puts implementations, call 1023 buffer_write(b, ...) directly in buffer_puts instead of calling 1024 BIO_write(b, ...). 1025 1026 In BIO_puts, increment b->num_write as in BIO_write. 1027 [Peter.Sylvester@EdelWeb.fr] 1028 1029 *) Fix BN_mul_word for the case where the word is 0. (We have to use 1030 BN_zero, we may not return a BIGNUM with an array consisting of 1031 words set to zero.) 1032 [Bodo Moeller] 1033 1034 *) Avoid calling abort() from within the library when problems are 1035 detected, except if preprocessor symbols have been defined 1036 (such as REF_CHECK, BN_DEBUG etc.). 1037 [Bodo Moeller] 1038 1039 *) New openssl application 'rsautl'. This utility can be 1040 used for low level RSA operations. DER public key 1041 BIO/fp routines also added. 1042 [Steve Henson] 1043 1044 *) New Configure entry and patches for compiling on QNX 4. 1045 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 1046 1047 *) A demo state-machine implementation was sponsored by 1048 Nuron (http://www.nuron.com/) and is now available in 1049 demos/state_machine. 1050 [Ben Laurie] 1051 1052 *) New options added to the 'dgst' utility for signature 1053 generation and verification. 1054 [Steve Henson] 1055 1056 *) Unrecognized PKCS#7 content types are now handled via a 1057 catch all ASN1_TYPE structure. This allows unsupported 1058 types to be stored as a "blob" and an application can 1059 encode and decode it manually. 1060 [Steve Henson] 1061 1062 *) Fix various signed/unsigned issues to make a_strex.c 1063 compile under VC++. 1064 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 1065 1066 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 1067 length if passed a buffer. ASN1_INTEGER_to_BN failed 1068 if passed a NULL BN and its argument was negative. 1069 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 1070 1071 *) Modification to PKCS#7 encoding routines to output definite 1072 length encoding. Since currently the whole structures are in 1073 memory there's not real point in using indefinite length 1074 constructed encoding. However if OpenSSL is compiled with 1075 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 1076 [Steve Henson] 1077 1078 *) Added BIO_vprintf() and BIO_vsnprintf(). 1079 [Richard Levitte] 1080 1081 *) Added more prefixes to parse for in the the strings written 1082 through a logging bio, to cover all the levels that are available 1083 through syslog. The prefixes are now: 1084 1085 PANIC, EMERG, EMR => LOG_EMERG 1086 ALERT, ALR => LOG_ALERT 1087 CRIT, CRI => LOG_CRIT 1088 ERROR, ERR => LOG_ERR 1089 WARNING, WARN, WAR => LOG_WARNING 1090 NOTICE, NOTE, NOT => LOG_NOTICE 1091 INFO, INF => LOG_INFO 1092 DEBUG, DBG => LOG_DEBUG 1093 1094 and as before, if none of those prefixes are present at the 1095 beginning of the string, LOG_ERR is chosen. 1096 1097 On Win32, the LOG_* levels are mapped according to this: 1098 1099 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 1100 LOG_WARNING => EVENTLOG_WARNING_TYPE 1101 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 1102 1103 [Richard Levitte] 1104 1105 *) Made it possible to reconfigure with just the configuration 1106 argument "reconf" or "reconfigure". The command line arguments 1107 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 1108 and are retrieved from there when reconfiguring. 1109 [Richard Levitte] 1110 1111 *) MD4 implemented. 1112 [Assar Westerlund <assar@sics.se>, Richard Levitte] 1113 1114 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 1115 [Richard Levitte] 1116 1117 *) The obj_dat.pl script was messing up the sorting of object 1118 names. The reason was that it compared the quoted version 1119 of strings as a result "OCSP" > "OCSP Signing" because 1120 " > SPACE. Changed script to store unquoted versions of 1121 names and add quotes on output. It was also omitting some 1122 names from the lookup table if they were given a default 1123 value (that is if SN is missing it is given the same 1124 value as LN and vice versa), these are now added on the 1125 grounds that if an object has a name we should be able to 1126 look it up. Finally added warning output when duplicate 1127 short or long names are found. 1128 [Steve Henson] 1129 1130 *) Changes needed for Tandem NSK. 1131 [Scott Uroff <scott@xypro.com>] 1132 1133 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 1134 RSA_padding_check_SSLv23(), special padding was never detected 1135 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 1136 version rollback attacks was not effective. 1137 1138 In s23_clnt.c, don't use special rollback-attack detection padding 1139 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 1140 client; similarly, in s23_srvr.c, don't do the rollback check if 1141 SSL 2.0 is the only protocol enabled in the server. 1142 [Bodo Moeller] 1143 1144 *) Make it possible to get hexdumps of unprintable data with 'openssl 1145 asn1parse'. By implication, the functions ASN1_parse_dump() and 1146 BIO_dump_indent() are added. 1147 [Richard Levitte] 1148 1149 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 1150 these print out strings and name structures based on various 1151 flags including RFC2253 support and proper handling of 1152 multibyte characters. Added options to the 'x509' utility 1153 to allow the various flags to be set. 1154 [Steve Henson] 1155 1156 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 1157 Also change the functions X509_cmp_current_time() and 1158 X509_gmtime_adj() work with an ASN1_TIME structure, 1159 this will enable certificates using GeneralizedTime in validity 1160 dates to be checked. 1161 [Steve Henson] 1162 1163 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 1164 negative public key encodings) on by default, 1165 NO_NEG_PUBKEY_BUG can be set to disable it. 1166 [Steve Henson] 1167 1168 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 1169 content octets. An i2c_ASN1_OBJECT is unnecessary because 1170 the encoding can be trivially obtained from the structure. 1171 [Steve Henson] 1172 1173 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 1174 not read locks (CRYPTO_r_[un]lock). 1175 [Bodo Moeller] 1176 1177 *) A first attempt at creating official support for shared 1178 libraries through configuration. I've kept it so the 1179 default is static libraries only, and the OpenSSL programs 1180 are always statically linked for now, but there are 1181 preparations for dynamic linking in place. 1182 This has been tested on Linux and Tru64. 1183 [Richard Levitte] 1184 1185 *) Randomness polling function for Win9x, as described in: 1186 Peter Gutmann, Software Generation of Practically Strong 1187 Random Numbers. 1188 [Ulf M�ller] 1189 1190 *) Fix so PRNG is seeded in req if using an already existing 1191 DSA key. 1192 [Steve Henson] 1193 1194 *) New options to smime application. -inform and -outform 1195 allow alternative formats for the S/MIME message including 1196 PEM and DER. The -content option allows the content to be 1197 specified separately. This should allow things like Netscape 1198 form signing output easier to verify. 1199 [Steve Henson] 1200 1201 *) Fix the ASN1 encoding of tags using the 'long form'. 1202 [Steve Henson] 1203 1204 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 1205 STRING types. These convert content octets to and from the 1206 underlying type. The actual tag and length octets are 1207 already assumed to have been read in and checked. These 1208 are needed because all other string types have virtually 1209 identical handling apart from the tag. By having versions 1210 of the ASN1 functions that just operate on content octets 1211 IMPLICIT tagging can be handled properly. It also allows 1212 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 1213 and ASN1_INTEGER are identical apart from the tag. 1214 [Steve Henson] 1215 1216 *) Change the handling of OID objects as follows: 1217 1218 - New object identifiers are inserted in objects.txt, following 1219 the syntax given in objects.README. 1220 - objects.pl is used to process obj_mac.num and create a new 1221 obj_mac.h. 1222 - obj_dat.pl is used to create a new obj_dat.h, using the data in 1223 obj_mac.h. 1224 1225 This is currently kind of a hack, and the perl code in objects.pl 1226 isn't very elegant, but it works as I intended. The simplest way 1227 to check that it worked correctly is to look in obj_dat.h and 1228 check the array nid_objs and make sure the objects haven't moved 1229 around (this is important!). Additions are OK, as well as 1230 consistent name changes. 1231 [Richard Levitte] 1232 1233 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 1234 [Bodo Moeller] 1235 1236 *) Addition of the command line parameter '-rand file' to 'openssl req'. 1237 The given file adds to whatever has already been seeded into the 1238 random pool through the RANDFILE configuration file option or 1239 environment variable, or the default random state file. 1240 [Richard Levitte] 1241 1242 *) mkstack.pl now sorts each macro group into lexical order. 1243 Previously the output order depended on the order the files 1244 appeared in the directory, resulting in needless rewriting 1245 of safestack.h . 1246 [Steve Henson] 1247 1248 *) Patches to make OpenSSL compile under Win32 again. Mostly 1249 work arounds for the VC++ problem that it treats func() as 1250 func(void). Also stripped out the parts of mkdef.pl that 1251 added extra typesafe functions: these no longer exist. 1252 [Steve Henson] 1253 1254 *) Reorganisation of the stack code. The macros are now all 1255 collected in safestack.h . Each macro is defined in terms of 1256 a "stack macro" of the form SKM_<name>(type, a, b). The 1257 DEBUG_SAFESTACK is now handled in terms of function casts, 1258 this has the advantage of retaining type safety without the 1259 use of additional functions. If DEBUG_SAFESTACK is not defined 1260 then the non typesafe macros are used instead. Also modified the 1261 mkstack.pl script to handle the new form. Needs testing to see 1262 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 1263 the default if no major problems. Similar behaviour for ASN1_SET_OF 1264 and PKCS12_STACK_OF. 1265 [Steve Henson] 1266 1267 *) When some versions of IIS use the 'NET' form of private key the 1268 key derivation algorithm is different. Normally MD5(password) is 1269 used as a 128 bit RC4 key. In the modified case 1270 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 1271 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 1272 as the old Netscape_RSA functions except they have an additional 1273 'sgckey' parameter which uses the modified algorithm. Also added 1274 an -sgckey command line option to the rsa utility. Thanks to 1275 Adrian Peck <bertie@ncipher.com> for posting details of the modified 1276 algorithm to openssl-dev. 1277 [Steve Henson] 1278 1279 *) The evp_local.h macros were using 'c.##kname' which resulted in 1280 invalid expansion on some systems (SCO 5.0.5 for example). 1281 Corrected to 'c.kname'. 1282 [Phillip Porch <root@theporch.com>] 1283 1284 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 1285 a STACK of email addresses from a certificate or request, these look 1286 in the subject name and the subject alternative name extensions and 1287 omit any duplicate addresses. 1288 [Steve Henson] 1289 1290 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 1291 This makes DSA verification about 2 % faster. 1292 [Bodo Moeller] 1293 1294 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 1295 (meaning that now 2^5 values will be precomputed, which is only 4 KB 1296 plus overhead for 1024 bit moduli). 1297 This makes exponentiations about 0.5 % faster for 1024 bit 1298 exponents (as measured by "openssl speed rsa2048"). 1299 [Bodo Moeller] 1300 1301 *) Rename memory handling macros to avoid conflicts with other 1302 software: 1303 Malloc => OPENSSL_malloc 1304 Malloc_locked => OPENSSL_malloc_locked 1305 Realloc => OPENSSL_realloc 1306 Free => OPENSSL_free 1307 [Richard Levitte] 1308 1309 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 1310 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 1311 [Bodo Moeller] 1312 1313 *) CygWin32 support. 1314 [John Jarvie <jjarvie@newsguy.com>] 1315 1316 *) The type-safe stack code has been rejigged. It is now only compiled 1317 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 1318 by default all type-specific stack functions are "#define"d back to 1319 standard stack functions. This results in more streamlined output 1320 but retains the type-safety checking possibilities of the original 1321 approach. 1322 [Geoff Thorpe] 1323 1324 *) The STACK code has been cleaned up, and certain type declarations 1325 that didn't make a lot of sense have been brought in line. This has 1326 also involved a cleanup of sorts in safestack.h to more correctly 1327 map type-safe stack functions onto their plain stack counterparts. 1328 This work has also resulted in a variety of "const"ifications of 1329 lots of the code, especially "_cmp" operations which should normally 1330 be prototyped with "const" parameters anyway. 1331 [Geoff Thorpe] 1332 1333 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 1334 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 1335 (The PRNG state consists of two parts, the large pool 'state' and 'md', 1336 where all of 'md' is used each time the PRNG is used, but 'state' 1337 is used only indexed by a cyclic counter. As entropy may not be 1338 well distributed from the beginning, 'md' is important as a 1339 chaining variable. However, the output function chains only half 1340 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 1341 all of 'md', and seeding with STATE_SIZE dummy bytes will result 1342 in all of 'state' being rewritten, with the new values depending 1343 on virtually all of 'md'. This overcomes the 80 bit limitation.) 1344 [Bodo Moeller] 1345 1346 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 1347 the handshake is continued after ssl_verify_cert_chain(); 1348 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 1349 can lead to 'unexplainable' connection aborts later. 1350 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 1351 1352 *) Major EVP API cipher revision. 1353 Add hooks for extra EVP features. This allows various cipher 1354 parameters to be set in the EVP interface. Support added for variable 1355 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 1356 setting of RC2 and RC5 parameters. 1357 1358 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 1359 ciphers. 1360 1361 Remove lots of duplicated code from the EVP library. For example *every* 1362 cipher init() function handles the 'iv' in the same way according to the 1363 cipher mode. They also all do nothing if the 'key' parameter is NULL and 1364 for CFB and OFB modes they zero ctx->num. 1365 1366 New functionality allows removal of S/MIME code RC2 hack. 1367 1368 Most of the routines have the same form and so can be declared in terms 1369 of macros. 1370 1371 By shifting this to the top level EVP_CipherInit() it can be removed from 1372 all individual ciphers. If the cipher wants to handle IVs or keys 1373 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 1374 flags. 1375 1376 Change lots of functions like EVP_EncryptUpdate() to now return a 1377 value: although software versions of the algorithms cannot fail 1378 any installed hardware versions can. 1379 [Steve Henson] 1380 1381 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 1382 this option is set, tolerate broken clients that send the negotiated 1383 protocol version number instead of the requested protocol version 1384 number. 1385 [Bodo Moeller] 1386 1387 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 1388 i.e. non-zero for export ciphersuites, zero otherwise. 1389 Previous versions had this flag inverted, inconsistent with 1390 rsa_tmp_cb (..._TMP_RSA_CB). 1391 [Bodo Moeller; problem reported by Amit Chopra] 1392 1393 *) Add missing DSA library text string. Work around for some IIS 1394 key files with invalid SEQUENCE encoding. 1395 [Steve Henson] 1396 1397 *) Add a document (doc/standards.txt) that list all kinds of standards 1398 and so on that are implemented in OpenSSL. 1399 [Richard Levitte] 1400 1401 *) Enhance c_rehash script. Old version would mishandle certificates 1402 with the same subject name hash and wouldn't handle CRLs at all. 1403 Added -fingerprint option to crl utility, to support new c_rehash 1404 features. 1405 [Steve Henson] 1406 1407 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 1408 [Ulf M�ller] 1409 1410 *) Fix for SSL server purpose checking. Server checking was 1411 rejecting certificates which had extended key usage present 1412 but no ssl client purpose. 1413 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 1414 1415 *) Make PKCS#12 code work with no password. The PKCS#12 spec 1416 is a little unclear about how a blank password is handled. 1417 Since the password in encoded as a BMPString with terminating 1418 double NULL a zero length password would end up as just the 1419 double NULL. However no password at all is different and is 1420 handled differently in the PKCS#12 key generation code. NS 1421 treats a blank password as zero length. MSIE treats it as no 1422 password on export: but it will try both on import. We now do 1423 the same: PKCS12_parse() tries zero length and no password if 1424 the password is set to "" or NULL (NULL is now a valid password: 1425 it wasn't before) as does the pkcs12 application. 1426 [Steve Henson] 1427 1428 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 1429 perror when PEM_read_bio_X509_REQ fails, the error message must 1430 be obtained from the error queue. 1431 [Bodo Moeller] 1432 1433 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 1434 it in ERR_remove_state if appropriate, and change ERR_get_state 1435 accordingly to avoid race conditions (this is necessary because 1436 thread_hash is no longer constant once set). 1437 [Bodo Moeller] 1438 1439 *) Bugfix for linux-elf makefile.one. 1440 [Ulf M�ller] 1441 1442 *) RSA_get_default_method() will now cause a default 1443 RSA_METHOD to be chosen if one doesn't exist already. 1444 Previously this was only set during a call to RSA_new() 1445 or RSA_new_method(NULL) meaning it was possible for 1446 RSA_get_default_method() to return NULL. 1447 [Geoff Thorpe] 1448 1449 *) Added native name translation to the existing DSO code 1450 that will convert (if the flag to do so is set) filenames 1451 that are sufficiently small and have no path information 1452 into a canonical native form. Eg. "blah" converted to 1453 "libblah.so" or "blah.dll" etc. 1454 [Geoff Thorpe] 1455 1456 *) New function ERR_error_string_n(e, buf, len) which is like 1457 ERR_error_string(e, buf), but writes at most 'len' bytes 1458 including the 0 terminator. For ERR_error_string_n, 'buf' 1459 may not be NULL. 1460 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 1461 1462 *) CONF library reworked to become more general. A new CONF 1463 configuration file reader "class" is implemented as well as a 1464 new functions (NCONF_*, for "New CONF") to handle it. The now 1465 old CONF_* functions are still there, but are reimplemented to 1466 work in terms of the new functions. Also, a set of functions 1467 to handle the internal storage of the configuration data is 1468 provided to make it easier to write new configuration file 1469 reader "classes" (I can definitely see something reading a 1470 configuration file in XML format, for example), called _CONF_*, 1471 or "the configuration storage API"... 1472 1473 The new configuration file reading functions are: 1474 1475 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 1476 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 1477 1478 NCONF_default, NCONF_WIN32 1479 1480 NCONF_dump_fp, NCONF_dump_bio 1481 1482 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 1483 NCONF_new creates a new CONF object. This works in the same way 1484 as other interfaces in OpenSSL, like the BIO interface. 1485 NCONF_dump_* dump the internal storage of the configuration file, 1486 which is useful for debugging. All other functions take the same 1487 arguments as the old CONF_* functions wth the exception of the 1488 first that must be a `CONF *' instead of a `LHASH *'. 1489 1490 To make it easer to use the new classes with the old CONF_* functions, 1491 the function CONF_set_default_method is provided. 1492 [Richard Levitte] 1493 1494 *) Add '-tls1' option to 'openssl ciphers', which was already 1495 mentioned in the documentation but had not been implemented. 1496 (This option is not yet really useful because even the additional 1497 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 1498 [Bodo Moeller] 1499 1500 *) Initial DSO code added into libcrypto for letting OpenSSL (and 1501 OpenSSL-based applications) load shared libraries and bind to 1502 them in a portable way. 1503 [Geoff Thorpe, with contributions from Richard Levitte] 1504 1505 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 1506 1507 *) Make sure _lrotl and _lrotr are only used with MSVC. 1508 1509 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 1510 (the default implementation of RAND_status). 1511 1512 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 1513 to '-clrext' (= clear extensions), as intended and documented. 1514 [Bodo Moeller; inconsistency pointed out by Michael Attili 1515 <attili@amaxo.com>] 1516 1517 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 1518 was larger than the MD block size. 1519 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 1520 1521 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 1522 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 1523 using the passed key: if the passed key was a private key the result 1524 of X509_print(), for example, would be to print out all the private key 1525 components. 1526 [Steve Henson] 1527 1528 *) des_quad_cksum() byte order bug fix. 1529 [Ulf M�ller, using the problem description in krb4-0.9.7, where 1530 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 1531 1532 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 1533 discouraged. 1534 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 1535 1536 *) For easily testing in shell scripts whether some command 1537 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 1538 returns with exit code 0 iff no command of the given name is available. 1539 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 1540 the output goes to stdout and nothing is printed to stderr. 1541 Additional arguments are always ignored. 1542 1543 Since for each cipher there is a command of the same name, 1544 the 'no-cipher' compilation switches can be tested this way. 1545 1546 ('openssl no-XXX' is not able to detect pseudo-commands such 1547 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 1548 [Bodo Moeller] 1549 1550 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 1551 [Bodo Moeller] 1552 1553 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 1554 is set; it will be thrown away anyway because each handshake creates 1555 its own key. 1556 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 1557 to parameters -- in previous versions (since OpenSSL 0.9.3) the 1558 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 1559 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 1560 [Bodo Moeller] 1561 1562 *) New s_client option -ign_eof: EOF at stdin is ignored, and 1563 'Q' and 'R' lose their special meanings (quit/renegotiate). 1564 This is part of what -quiet does; unlike -quiet, -ign_eof 1565 does not suppress any output. 1566 [Richard Levitte] 1567 1568 *) Add compatibility options to the purpose and trust code. The 1569 purpose X509_PURPOSE_ANY is "any purpose" which automatically 1570 accepts a certificate or CA, this was the previous behaviour, 1571 with all the associated security issues. 1572 1573 X509_TRUST_COMPAT is the old trust behaviour: only and 1574 automatically trust self signed roots in certificate store. A 1575 new trust setting X509_TRUST_DEFAULT is used to specify that 1576 a purpose has no associated trust setting and it should instead 1577 use the value in the default purpose. 1578 [Steve Henson] 1579 1580 *) Fix the PKCS#8 DSA private key code so it decodes keys again 1581 and fix a memory leak. 1582 [Steve Henson] 1583 1584 *) In util/mkerr.pl (which implements 'make errors'), preserve 1585 reason strings from the previous version of the .c file, as 1586 the default to have only downcase letters (and digits) in 1587 automatically generated reasons codes is not always appropriate. 1588 [Bodo Moeller] 1589 1590 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 1591 using strerror. Previously, ERR_reason_error_string() returned 1592 library names as reason strings for SYSerr; but SYSerr is a special 1593 case where small numbers are errno values, not library numbers. 1594 [Bodo Moeller] 1595 1596 *) Add '-dsaparam' option to 'openssl dhparam' application. This 1597 converts DSA parameters into DH parameters. (When creating parameters, 1598 DSA_generate_parameters is used.) 1599 [Bodo Moeller] 1600 1601 *) Include 'length' (recommended exponent length) in C code generated 1602 by 'openssl dhparam -C'. 1603 [Bodo Moeller] 1604 1605 *) The second argument to set_label in perlasm was already being used 1606 so couldn't be used as a "file scope" flag. Moved to third argument 1607 which was free. 1608 [Steve Henson] 1609 1610 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 1611 instead of RAND_bytes for encryption IVs and salts. 1612 [Bodo Moeller] 1613 1614 *) Include RAND_status() into RAND_METHOD instead of implementing 1615 it only for md_rand.c Otherwise replacing the PRNG by calling 1616 RAND_set_rand_method would be impossible. 1617 [Bodo Moeller] 1618 1619 *) Don't let DSA_generate_key() enter an infinite loop if the random 1620 number generation fails. 1621 [Bodo Moeller] 1622 1623 *) New 'rand' application for creating pseudo-random output. 1624 [Bodo Moeller] 1625 1626 *) Added configuration support for Linux/IA64 1627 [Rolf Haberrecker <rolf@suse.de>] 1628 1629 *) Assembler module support for Mingw32. 1630 [Ulf M�ller] 1631 1632 *) Shared library support for HPUX (in shlib/). 1633 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 1634 1635 *) Shared library support for Solaris gcc. 1636 [Lutz Behnke <behnke@trustcenter.de>] 1637 1638 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 1639 1640 *) PKCS7_encrypt() was adding text MIME headers twice because they 1641 were added manually and by SMIME_crlf_copy(). 1642 [Steve Henson] 1643 1644 *) In bntest.c don't call BN_rand with zero bits argument. 1645 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 1646 1647 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 1648 case was implemented. This caused BN_div_recp() to fail occasionally. 1649 [Ulf M�ller] 1650 1651 *) Add an optional second argument to the set_label() in the perl 1652 assembly language builder. If this argument exists and is set 1653 to 1 it signals that the assembler should use a symbol whose 1654 scope is the entire file, not just the current function. This 1655 is needed with MASM which uses the format label:: for this scope. 1656 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 1657 1658 *) Change the ASN1 types so they are typedefs by default. Before 1659 almost all types were #define'd to ASN1_STRING which was causing 1660 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 1661 for example. 1662 [Steve Henson] 1663 1664 *) Change names of new functions to the new get1/get0 naming 1665 convention: After 'get1', the caller owns a reference count 1666 and has to call ..._free; 'get0' returns a pointer to some 1667 data structure without incrementing reference counters. 1668 (Some of the existing 'get' functions increment a reference 1669 counter, some don't.) 1670 Similarly, 'set1' and 'add1' functions increase reference 1671 counters or duplicate objects. 1672 [Steve Henson] 1673 1674 *) Allow for the possibility of temp RSA key generation failure: 1675 the code used to assume it always worked and crashed on failure. 1676 [Steve Henson] 1677 1678 *) Fix potential buffer overrun problem in BIO_printf(). 1679 [Ulf M�ller, using public domain code by Patrick Powell; problem 1680 pointed out by David Sacerdote <das33@cornell.edu>] 1681 1682 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 1683 RAND_egd() and RAND_status(). In the command line application, 1684 the EGD socket can be specified like a seed file using RANDFILE 1685 or -rand. 1686 [Ulf M�ller] 1687 1688 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 1689 Some CAs (e.g. Verisign) distribute certificates in this form. 1690 [Steve Henson] 1691 1692 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 1693 list to exclude them. This means that no special compilation option 1694 is needed to use anonymous DH: it just needs to be included in the 1695 cipher list. 1696 [Steve Henson] 1697 1698 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 1699 EVP_MD_type. The old functionality is available in a new macro called 1700 EVP_MD_md(). Change code that uses it and update docs. 1701 [Steve Henson] 1702 1703 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 1704 where the 'void *' argument is replaced by a function pointer argument. 1705 Previously 'void *' was abused to point to functions, which works on 1706 many platforms, but is not correct. As these functions are usually 1707 called by macros defined in OpenSSL header files, most source code 1708 should work without changes. 1709 [Richard Levitte] 1710 1711 *) <openssl/opensslconf.h> (which is created by Configure) now contains 1712 sections with information on -D... compiler switches used for 1713 compiling the library so that applications can see them. To enable 1714 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 1715 must be defined. E.g., 1716 #define OPENSSL_ALGORITHM_DEFINES 1717 #include <openssl/opensslconf.h> 1718 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 1719 [Richard Levitte, Ulf and Bodo M�ller] 1720 1721 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 1722 record layer. 1723 [Bodo Moeller] 1724 1725 *) Change the 'other' type in certificate aux info to a STACK_OF 1726 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 1727 the required ASN1 format: arbitrary types determined by an OID. 1728 [Steve Henson] 1729 1730 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 1731 argument to 'req'. This is not because the function is newer or 1732 better than others it just uses the work 'NEW' in the certificate 1733 request header lines. Some software needs this. 1734 [Steve Henson] 1735 1736 *) Reorganise password command line arguments: now passwords can be 1737 obtained from various sources. Delete the PEM_cb function and make 1738 it the default behaviour: i.e. if the callback is NULL and the 1739 usrdata argument is not NULL interpret it as a null terminated pass 1740 phrase. If usrdata and the callback are NULL then the pass phrase 1741 is prompted for as usual. 1742 [Steve Henson] 1743 1744 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 1745 the support is automatically enabled. The resulting binaries will 1746 autodetect the card and use it if present. 1747 [Ben Laurie and Compaq Inc.] 1748 1749 *) Work around for Netscape hang bug. This sends certificate request 1750 and server done in one record. Since this is perfectly legal in the 1751 SSL/TLS protocol it isn't a "bug" option and is on by default. See 1752 the bugs/SSLv3 entry for more info. 1753 [Steve Henson] 1754 1755 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 1756 [Andy Polyakov] 1757 1758 *) Add -rand argument to smime and pkcs12 applications and read/write 1759 of seed file. 1760 [Steve Henson] 1761 1762 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 1763 [Bodo Moeller] 1764 1765 *) Add command line password options to the remaining applications. 1766 [Steve Henson] 1767 1768 *) Bug fix for BN_div_recp() for numerators with an even number of 1769 bits. 1770 [Ulf M�ller] 1771 1772 *) More tests in bntest.c, and changed test_bn output. 1773 [Ulf M�ller] 1774 1775 *) ./config recognizes MacOS X now. 1776 [Andy Polyakov] 1777 1778 *) Bug fix for BN_div() when the first words of num and divsor are 1779 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 1780 [Ulf M�ller] 1781 1782 *) Add support for various broken PKCS#8 formats, and command line 1783 options to produce them. 1784 [Steve Henson] 1785 1786 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 1787 get temporary BIGNUMs from a BN_CTX. 1788 [Ulf M�ller] 1789 1790 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 1791 for p == 0. 1792 [Ulf M�ller] 1793 1794 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 1795 include a #define from the old name to the new. The original intent 1796 was that statically linked binaries could for example just call 1797 SSLeay_add_all_ciphers() to just add ciphers to the table and not 1798 link with digests. This never worked becayse SSLeay_add_all_digests() 1799 and SSLeay_add_all_ciphers() were in the same source file so calling 1800 one would link with the other. They are now in separate source files. 1801 [Steve Henson] 1802 1803 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 1804 [Steve Henson] 1805 1806 *) Use a less unusual form of the Miller-Rabin primality test (it used 1807 a binary algorithm for exponentiation integrated into the Miller-Rabin 1808 loop, our standard modexp algorithms are faster). 1809 [Bodo Moeller] 1810 1811 *) Support for the EBCDIC character set completed. 1812 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 1813 1814 *) Source code cleanups: use const where appropriate, eliminate casts, 1815 use void * instead of char * in lhash. 1816 [Ulf M�ller] 1817 1818 *) Bugfix: ssl3_send_server_key_exchange was not restartable 1819 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 1820 this the server could overwrite ephemeral keys that the client 1821 has already seen). 1822 [Bodo Moeller] 1823 1824 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 1825 using 50 iterations of the Rabin-Miller test. 1826 1827 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 1828 iterations of the Rabin-Miller test as required by the appendix 1829 to FIPS PUB 186[-1]) instead of DSA_is_prime. 1830 As BN_is_prime_fasttest includes trial division, DSA parameter 1831 generation becomes much faster. 1832 1833 This implies a change for the callback functions in DSA_is_prime 1834 and DSA_generate_parameters: The callback function is called once 1835 for each positive witness in the Rabin-Miller test, not just 1836 occasionally in the inner loop; and the parameters to the 1837 callback function now provide an iteration count for the outer 1838 loop rather than for the current invocation of the inner loop. 1839 DSA_generate_parameters additionally can call the callback 1840 function with an 'iteration count' of -1, meaning that a 1841 candidate has passed the trial division test (when q is generated 1842 from an application-provided seed, trial division is skipped). 1843 [Bodo Moeller] 1844 1845 *) New function BN_is_prime_fasttest that optionally does trial 1846 division before starting the Rabin-Miller test and has 1847 an additional BN_CTX * argument (whereas BN_is_prime always 1848 has to allocate at least one BN_CTX). 1849 'callback(1, -1, cb_arg)' is called when a number has passed the 1850 trial division stage. 1851 [Bodo Moeller] 1852 1853 *) Fix for bug in CRL encoding. The validity dates weren't being handled 1854 as ASN1_TIME. 1855 [Steve Henson] 1856 1857 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 1858 [Steve Henson] 1859 1860 *) New function BN_pseudo_rand(). 1861 [Ulf M�ller] 1862 1863 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 1864 bignum version of BN_from_montgomery() with the working code from 1865 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 1866 the comments. 1867 [Ulf M�ller] 1868 1869 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 1870 made it impossible to use the same SSL_SESSION data structure in 1871 SSL2 clients in multiple threads. 1872 [Bodo Moeller] 1873 1874 *) The return value of RAND_load_file() no longer counts bytes obtained 1875 by stat(). RAND_load_file(..., -1) is new and uses the complete file 1876 to seed the PRNG (previously an explicit byte count was required). 1877 [Ulf M�ller, Bodo M�ller] 1878 1879 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 1880 used (char *) instead of (void *) and had casts all over the place. 1881 [Steve Henson] 1882 1883 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 1884 [Ulf M�ller] 1885 1886 *) Retain source code compatibility for BN_prime_checks macro: 1887 BN_is_prime(..., BN_prime_checks, ...) now uses 1888 BN_prime_checks_for_size to determine the appropriate number of 1889 Rabin-Miller iterations. 1890 [Ulf M�ller] 1891 1892 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 1893 DH_CHECK_P_NOT_SAFE_PRIME. 1894 (Check if this is true? OpenPGP calls them "strong".) 1895 [Ulf M�ller] 1896 1897 *) Merge the functionality of "dh" and "gendh" programs into a new program 1898 "dhparam". The old programs are retained for now but will handle DH keys 1899 (instead of parameters) in future. 1900 [Steve Henson] 1901 1902 *) Make the ciphers, s_server and s_client programs check the return values 1903 when a new cipher list is set. 1904 [Steve Henson] 1905 1906 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 1907 ciphers. Before when the 56bit ciphers were enabled the sorting was 1908 wrong. 1909 1910 The syntax for the cipher sorting has been extended to support sorting by 1911 cipher-strength (using the strength_bits hard coded in the tables). 1912 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 1913 1914 Fix a bug in the cipher-command parser: when supplying a cipher command 1915 string with an "undefined" symbol (neither command nor alphanumeric 1916 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 1917 an error is flagged. 1918 1919 Due to the strength-sorting extension, the code of the 1920 ssl_create_cipher_list() function was completely rearranged. I hope that 1921 the readability was also increased :-) 1922 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 1923 1924 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 1925 for the first serial number and places 2 in the serial number file. This 1926 avoids problems when the root CA is created with serial number zero and 1927 the first user certificate has the same issuer name and serial number 1928 as the root CA. 1929 [Steve Henson] 1930 1931 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 1932 the new code. Add documentation for this stuff. 1933 [Steve Henson] 1934 1935 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 1936 X509_*() to X509at_*() on the grounds that they don't handle X509 1937 structures and behave in an analagous way to the X509v3 functions: 1938 they shouldn't be called directly but wrapper functions should be used 1939 instead. 1940 1941 So we also now have some wrapper functions that call the X509at functions 1942 when passed certificate requests. (TO DO: similar things can be done with 1943 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 1944 things. Some of these need some d2i or i2d and print functionality 1945 because they handle more complex structures.) 1946 [Steve Henson] 1947 1948 *) Add missing #ifndefs that caused missing symbols when building libssl 1949 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 1950 NO_RSA in ssl/s2*.c. 1951 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 1952 1953 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 1954 has a return value which indicates the quality of the random data 1955 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 1956 error queue. New function RAND_pseudo_bytes() generates output that is 1957 guaranteed to be unique but not unpredictable. RAND_add is like 1958 RAND_seed, but takes an extra argument for an entropy estimate 1959 (RAND_seed always assumes full entropy). 1960 [Ulf M�ller] 1961 1962 *) Do more iterations of Rabin-Miller probable prime test (specifically, 1963 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 1964 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 1965 in crypto/bn/bn_prime.c for the complete table). This guarantees a 1966 false-positive rate of at most 2^-80 for random input. 1967 [Bodo Moeller] 1968 1969 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 1970 [Bodo Moeller] 1971 1972 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 1973 in the 0.9.5 release), this returns the chain 1974 from an X509_CTX structure with a dup of the stack and all 1975 the X509 reference counts upped: so the stack will exist 1976 after X509_CTX_cleanup() has been called. Modify pkcs12.c 1977 to use this. 1978 1979 Also make SSL_SESSION_print() print out the verify return 1980 code. 1981 [Steve Henson] 1982 1983 *) Add manpage for the pkcs12 command. Also change the default 1984 behaviour so MAC iteration counts are used unless the new 1985 -nomaciter option is used. This improves file security and 1986 only older versions of MSIE (4.0 for example) need it. 1987 [Steve Henson] 1988 1989 *) Honor the no-xxx Configure options when creating .DEF files. 1990 [Ulf M�ller] 1991 1992 *) Add PKCS#10 attributes to field table: challengePassword, 1993 unstructuredName and unstructuredAddress. These are taken from 1994 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 1995 international characters are used. 1996 1997 More changes to X509_ATTRIBUTE code: allow the setting of types 1998 based on strings. Remove the 'loc' parameter when adding 1999 attributes because these will be a SET OF encoding which is sorted 2000 in ASN1 order. 2001 [Steve Henson] 2002 2003 *) Initial changes to the 'req' utility to allow request generation 2004 automation. This will allow an application to just generate a template 2005 file containing all the field values and have req construct the 2006 request. 2007 2008 Initial support for X509_ATTRIBUTE handling. Stacks of these are 2009 used all over the place including certificate requests and PKCS#7 2010 structures. They are currently handled manually where necessary with 2011 some primitive wrappers for PKCS#7. The new functions behave in a 2012 manner analogous to the X509 extension functions: they allow 2013 attributes to be looked up by NID and added. 2014 2015 Later something similar to the X509V3 code would be desirable to 2016 automatically handle the encoding, decoding and printing of the 2017 more complex types. The string types like challengePassword can 2018 be handled by the string table functions. 2019 2020 Also modified the multi byte string table handling. Now there is 2021 a 'global mask' which masks out certain types. The table itself 2022 can use the flag STABLE_NO_MASK to ignore the mask setting: this 2023 is useful when for example there is only one permissible type 2024 (as in countryName) and using the mask might result in no valid 2025 types at all. 2026 [Steve Henson] 2027 2028 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 2029 SSL_get_peer_finished to allow applications to obtain the latest 2030 Finished messages sent to the peer or expected from the peer, 2031 respectively. (SSL_get_peer_finished is usually the Finished message 2032 actually received from the peer, otherwise the protocol will be aborted.) 2033 2034 As the Finished message are message digests of the complete handshake 2035 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 2036 be used for external authentication procedures when the authentication 2037 provided by SSL/TLS is not desired or is not enough. 2038 [Bodo Moeller] 2039 2040 *) Enhanced support for Alpha Linux is added. Now ./config checks if 2041 the host supports BWX extension and if Compaq C is present on the 2042 $PATH. Just exploiting of the BWX extension results in 20-30% 2043 performance kick for some algorithms, e.g. DES and RC4 to mention 2044 a couple. Compaq C in turn generates ~20% faster code for MD5 and 2045 SHA1. 2046 [Andy Polyakov] 2047 2048 *) Add support for MS "fast SGC". This is arguably a violation of the 2049 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 2050 weak crypto and after checking the certificate is SGC a second one 2051 with strong crypto. MS SGC stops the first handshake after receiving 2052 the server certificate message and sends a second client hello. Since 2053 a server will typically do all the time consuming operations before 2054 expecting any further messages from the client (server key exchange 2055 is the most expensive) there is little difference between the two. 2056 2057 To get OpenSSL to support MS SGC we have to permit a second client 2058 hello message after we have sent server done. In addition we have to 2059 reset the MAC if we do get this second client hello. 2060 [Steve Henson] 2061 2062 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 2063 if a DER encoded private key is RSA or DSA traditional format. Changed 2064 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 2065 format DER encoded private key. Newer code should use PKCS#8 format which 2066 has the key type encoded in the ASN1 structure. Added DER private key 2067 support to pkcs8 application. 2068 [Steve Henson] 2069 2070 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 2071 ciphersuites has been selected (as required by the SSL 3/TLS 1 2072 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 2073 is set, we interpret this as a request to violate the specification 2074 (the worst that can happen is a handshake failure, and 'correct' 2075 behaviour would result in a handshake failure anyway). 2076 [Bodo Moeller] 2077 2078 *) In SSL_CTX_add_session, take into account that there might be multiple 2079 SSL_SESSION structures with the same session ID (e.g. when two threads 2080 concurrently obtain them from an external cache). 2081 The internal cache can handle only one SSL_SESSION with a given ID, 2082 so if there's a conflict, we now throw out the old one to achieve 2083 consistency. 2084 [Bodo Moeller] 2085 2086 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 2087 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 2088 some routines that use cipher OIDs: some ciphers do not have OIDs 2089 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 2090 example. 2091 [Steve Henson] 2092 2093 *) Simplify the trust setting structure and code. Now we just have 2094 two sequences of OIDs for trusted and rejected settings. These will 2095 typically have values the same as the extended key usage extension 2096 and any application specific purposes. 2097 2098 The trust checking code now has a default behaviour: it will just 2099 check for an object with the same NID as the passed id. Functions can 2100 be provided to override either the default behaviour or the behaviour 2101 for a given id. SSL client, server and email already have functions 2102 in place for compatibility: they check the NID and also return "trusted" 2103 if the certificate is self signed. 2104 [Steve Henson] 2105 2106 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 2107 traditional format into an EVP_PKEY structure. 2108 [Steve Henson] 2109 2110 *) Add a password callback function PEM_cb() which either prompts for 2111 a password if usr_data is NULL or otherwise assumes it is a null 2112 terminated password. Allow passwords to be passed on command line 2113 environment or config files in a few more utilities. 2114 [Steve Henson] 2115 2116 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 2117 keys. Add some short names for PKCS#8 PBE algorithms and allow them 2118 to be specified on the command line for the pkcs8 and pkcs12 utilities. 2119 Update documentation. 2120 [Steve Henson] 2121 2122 *) Support for ASN1 "NULL" type. This could be handled before by using 2123 ASN1_TYPE but there wasn't any function that would try to read a NULL 2124 and produce an error if it couldn't. For compatibility we also have 2125 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 2126 don't allocate anything because they don't need to. 2127 [Steve Henson] 2128 2129 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 2130 for details. 2131 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 2132 2133 *) Rebuild of the memory allocation routines used by OpenSSL code and 2134 possibly others as well. The purpose is to make an interface that 2135 provide hooks so anyone can build a separate set of allocation and 2136 deallocation routines to be used by OpenSSL, for example memory 2137 pool implementations, or something else, which was previously hard 2138 since Malloc(), Realloc() and Free() were defined as macros having 2139 the values malloc, realloc and free, respectively (except for Win32 2140 compilations). The same is provided for memory debugging code. 2141 OpenSSL already comes with functionality to find memory leaks, but 2142 this gives people a chance to debug other memory problems. 2143 2144 With these changes, a new set of functions and macros have appeared: 2145 2146 CRYPTO_set_mem_debug_functions() [F] 2147 CRYPTO_get_mem_debug_functions() [F] 2148 CRYPTO_dbg_set_options() [F] 2149 CRYPTO_dbg_get_options() [F] 2150 CRYPTO_malloc_debug_init() [M] 2151 2152 The memory debug functions are NULL by default, unless the library 2153 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 2154 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 2155 gives the standard debugging functions that come with OpenSSL) or 2156 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 2157 provided by the library user) must be used. When the standard 2158 debugging functions are used, CRYPTO_dbg_set_options can be used to 2159 request additional information: 2160 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 2161 the CRYPTO_MDEBUG_xxx macro when compiling the library. 2162 2163 Also, things like CRYPTO_set_mem_functions will always give the 2164 expected result (the new set of functions is used for allocation 2165 and deallocation) at all times, regardless of platform and compiler 2166 options. 2167 2168 To finish it up, some functions that were never use in any other 2169 way than through macros have a new API and new semantic: 2170 2171 CRYPTO_dbg_malloc() 2172 CRYPTO_dbg_realloc() 2173 CRYPTO_dbg_free() 2174 2175 All macros of value have retained their old syntax. 2176 [Richard Levitte and Bodo Moeller] 2177 2178 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 2179 ordering of SMIMECapabilities wasn't in "strength order" and there 2180 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 2181 algorithm. 2182 [Steve Henson] 2183 2184 *) Some ASN1 types with illegal zero length encoding (INTEGER, 2185 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 2186 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 2187 2188 *) Merge in my S/MIME library for OpenSSL. This provides a simple 2189 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 2190 functionality to handle multipart/signed properly) and a utility 2191 called 'smime' to call all this stuff. This is based on code I 2192 originally wrote for Celo who have kindly allowed it to be 2193 included in OpenSSL. 2194 [Steve Henson] 2195 2196 *) Add variants des_set_key_checked and des_set_key_unchecked of 2197 des_set_key (aka des_key_sched). Global variable des_check_key 2198 decides which of these is called by des_set_key; this way 2199 des_check_key behaves as it always did, but applications and 2200 the library itself, which was buggy for des_check_key == 1, 2201 have a cleaner way to pick the version they need. 2202 [Bodo Moeller] 2203 2204 *) New function PKCS12_newpass() which changes the password of a 2205 PKCS12 structure. 2206 [Steve Henson] 2207 2208 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 2209 dynamic mix. In both cases the ids can be used as an index into the 2210 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 2211 functions so they accept a list of the field values and the 2212 application doesn't need to directly manipulate the X509_TRUST 2213 structure. 2214 [Steve Henson] 2215 2216 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 2217 need initialising. 2218 [Steve Henson] 2219 2220 *) Modify the way the V3 extension code looks up extensions. This now 2221 works in a similar way to the object code: we have some "standard" 2222 extensions in a static table which is searched with OBJ_bsearch() 2223 and the application can add dynamic ones if needed. The file 2224 crypto/x509v3/ext_dat.h now has the info: this file needs to be 2225 updated whenever a new extension is added to the core code and kept 2226 in ext_nid order. There is a simple program 'tabtest.c' which checks 2227 this. New extensions are not added too often so this file can readily 2228 be maintained manually. 2229 2230 There are two big advantages in doing things this way. The extensions 2231 can be looked up immediately and no longer need to be "added" using 2232 X509V3_add_standard_extensions(): this function now does nothing. 2233 [Side note: I get *lots* of email saying the extension code doesn't 2234 work because people forget to call this function] 2235 Also no dynamic allocation is done unless new extensions are added: 2236 so if we don't add custom extensions there is no need to call 2237 X509V3_EXT_cleanup(). 2238 [Steve Henson] 2239 2240 *) Modify enc utility's salting as follows: make salting the default. Add a 2241 magic header, so unsalted files fail gracefully instead of just decrypting 2242 to garbage. This is because not salting is a big security hole, so people 2243 should be discouraged from doing it. 2244 [Ben Laurie] 2245 2246 *) Fixes and enhancements to the 'x509' utility. It allowed a message 2247 digest to be passed on the command line but it only used this 2248 parameter when signing a certificate. Modified so all relevant 2249 operations are affected by the digest parameter including the 2250 -fingerprint and -x509toreq options. Also -x509toreq choked if a 2251 DSA key was used because it didn't fix the digest. 2252 [Steve Henson] 2253 2254 *) Initial certificate chain verify code. Currently tests the untrusted 2255 certificates for consistency with the verify purpose (which is set 2256 when the X509_STORE_CTX structure is set up) and checks the pathlength. 2257 2258 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 2259 this is because it will reject chains with invalid extensions whereas 2260 every previous version of OpenSSL and SSLeay made no checks at all. 2261 2262 Trust code: checks the root CA for the relevant trust settings. Trust 2263 settings have an initial value consistent with the verify purpose: e.g. 2264 if the verify purpose is for SSL client use it expects the CA to be 2265 trusted for SSL client use. However the default value can be changed to 2266 permit custom trust settings: one example of this would be to only trust 2267 certificates from a specific "secure" set of CAs. 2268 2269 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 2270 which should be used for version portability: especially since the 2271 verify structure is likely to change more often now. 2272 2273 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 2274 to set them. If not set then assume SSL clients will verify SSL servers 2275 and vice versa. 2276 2277 Two new options to the verify program: -untrusted allows a set of 2278 untrusted certificates to be passed in and -purpose which sets the 2279 intended purpose of the certificate. If a purpose is set then the 2280 new chain verify code is used to check extension consistency. 2281 [Steve Henson] 2282 2283 *) Support for the authority information access extension. 2284 [Steve Henson] 2285 2286 *) Modify RSA and DSA PEM read routines to transparently handle 2287 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 2288 public keys in a format compatible with certificate 2289 SubjectPublicKeyInfo structures. Unfortunately there were already 2290 functions called *_PublicKey_* which used various odd formats so 2291 these are retained for compatibility: however the DSA variants were 2292 never in a public release so they have been deleted. Changed dsa/rsa 2293 utilities to handle the new format: note no releases ever handled public 2294 keys so we should be OK. 2295 2296 The primary motivation for this change is to avoid the same fiasco 2297 that dogs private keys: there are several incompatible private key 2298 formats some of which are standard and some OpenSSL specific and 2299 require various evil hacks to allow partial transparent handling and 2300 even then it doesn't work with DER formats. Given the option anything 2301 other than PKCS#8 should be dumped: but the other formats have to 2302 stay in the name of compatibility. 2303 2304 With public keys and the benefit of hindsight one standard format 2305 is used which works with EVP_PKEY, RSA or DSA structures: though 2306 it clearly returns an error if you try to read the wrong kind of key. 2307 2308 Added a -pubkey option to the 'x509' utility to output the public key. 2309 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 2310 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 2311 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 2312 that do the same as the EVP_PKEY_assign_*() except they up the 2313 reference count of the added key (they don't "swallow" the 2314 supplied key). 2315 [Steve Henson] 2316 2317 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 2318 CRLs would fail if the file contained no certificates or no CRLs: 2319 added a new function to read in both types and return the number 2320 read: this means that if none are read it will be an error. The 2321 DER versions of the certificate and CRL reader would always fail 2322 because it isn't possible to mix certificates and CRLs in DER format 2323 without choking one or the other routine. Changed this to just read 2324 a certificate: this is the best we can do. Also modified the code 2325 in apps/verify.c to take notice of return codes: it was previously 2326 attempting to read in certificates from NULL pointers and ignoring 2327 any errors: this is one reason why the cert and CRL reader seemed 2328 to work. It doesn't check return codes from the default certificate 2329 routines: these may well fail if the certificates aren't installed. 2330 [Steve Henson] 2331 2332 *) Code to support otherName option in GeneralName. 2333 [Steve Henson] 2334 2335 *) First update to verify code. Change the verify utility 2336 so it warns if it is passed a self signed certificate: 2337 for consistency with the normal behaviour. X509_verify 2338 has been modified to it will now verify a self signed 2339 certificate if *exactly* the same certificate appears 2340 in the store: it was previously impossible to trust a 2341 single self signed certificate. This means that: 2342 openssl verify ss.pem 2343 now gives a warning about a self signed certificate but 2344 openssl verify -CAfile ss.pem ss.pem 2345 is OK. 2346 [Steve Henson] 2347 2348 *) For servers, store verify_result in SSL_SESSION data structure 2349 (and add it to external session representation). 2350 This is needed when client certificate verifications fails, 2351 but an application-provided verification callback (set by 2352 SSL_CTX_set_cert_verify_callback) allows accepting the session 2353 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 2354 but returns 1): When the session is reused, we have to set 2355 ssl->verify_result to the appropriate error code to avoid 2356 security holes. 2357 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 2358 2359 *) Fix a bug in the new PKCS#7 code: it didn't consider the 2360 case in PKCS7_dataInit() where the signed PKCS7 structure 2361 didn't contain any existing data because it was being created. 2362 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 2363 2364 *) Add a salt to the key derivation routines in enc.c. This 2365 forms the first 8 bytes of the encrypted file. Also add a 2366 -S option to allow a salt to be input on the command line. 2367 [Steve Henson] 2368 2369 *) New function X509_cmp(). Oddly enough there wasn't a function 2370 to compare two certificates. We do this by working out the SHA1 2371 hash and comparing that. X509_cmp() will be needed by the trust 2372 code. 2373 [Steve Henson] 2374 2375 *) SSL_get1_session() is like SSL_get_session(), but increments 2376 the reference count in the SSL_SESSION returned. 2377 [Geoff Thorpe <geoff@eu.c2.net>] 2378 2379 *) Fix for 'req': it was adding a null to request attributes. 2380 Also change the X509_LOOKUP and X509_INFO code to handle 2381 certificate auxiliary information. 2382 [Steve Henson] 2383 2384 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 2385 the 'enc' command. 2386 [Steve Henson] 2387 2388 *) Add the possibility to add extra information to the memory leak 2389 detecting output, to form tracebacks, showing from where each 2390 allocation was originated: CRYPTO_push_info("constant string") adds 2391 the string plus current file name and line number to a per-thread 2392 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 2393 is like calling CYRPTO_pop_info() until the stack is empty. 2394 Also updated memory leak detection code to be multi-thread-safe. 2395 [Richard Levitte] 2396 2397 *) Add options -text and -noout to pkcs7 utility and delete the 2398 encryption options which never did anything. Update docs. 2399 [Steve Henson] 2400 2401 *) Add options to some of the utilities to allow the pass phrase 2402 to be included on either the command line (not recommended on 2403 OSes like Unix) or read from the environment. Update the 2404 manpages and fix a few bugs. 2405 [Steve Henson] 2406 2407 *) Add a few manpages for some of the openssl commands. 2408 [Steve Henson] 2409 2410 *) Fix the -revoke option in ca. It was freeing up memory twice, 2411 leaking and not finding already revoked certificates. 2412 [Steve Henson] 2413 2414 *) Extensive changes to support certificate auxiliary information. 2415 This involves the use of X509_CERT_AUX structure and X509_AUX 2416 functions. An X509_AUX function such as PEM_read_X509_AUX() 2417 can still read in a certificate file in the usual way but it 2418 will also read in any additional "auxiliary information". By 2419 doing things this way a fair degree of compatibility can be 2420 retained: existing certificates can have this information added 2421 using the new 'x509' options. 2422 2423 Current auxiliary information includes an "alias" and some trust 2424 settings. The trust settings will ultimately be used in enhanced 2425 certificate chain verification routines: currently a certificate 2426 can only be trusted if it is self signed and then it is trusted 2427 for all purposes. 2428 [Steve Henson] 2429 2430 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 2431 The problem was that one of the replacement routines had not been working 2432 since SSLeay releases. For now the offending routine has been replaced 2433 with non-optimised assembler. Even so, this now gives around 95% 2434 performance improvement for 1024 bit RSA signs. 2435 [Mark Cox] 2436 2437 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 2438 handling. Most clients have the effective key size in bits equal to 2439 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 2440 A few however don't do this and instead use the size of the decrypted key 2441 to determine the RC2 key length and the AlgorithmIdentifier to determine 2442 the effective key length. In this case the effective key length can still 2443 be 40 bits but the key length can be 168 bits for example. This is fixed 2444 by manually forcing an RC2 key into the EVP_PKEY structure because the 2445 EVP code can't currently handle unusual RC2 key sizes: it always assumes 2446 the key length and effective key length are equal. 2447 [Steve Henson] 2448 2449 *) Add a bunch of functions that should simplify the creation of 2450 X509_NAME structures. Now you should be able to do: 2451 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 2452 and have it automatically work out the correct field type and fill in 2453 the structures. The more adventurous can try: 2454 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 2455 and it will (hopefully) work out the correct multibyte encoding. 2456 [Steve Henson] 2457 2458 *) Change the 'req' utility to use the new field handling and multibyte 2459 copy routines. Before the DN field creation was handled in an ad hoc 2460 way in req, ca, and x509 which was rather broken and didn't support 2461 BMPStrings or UTF8Strings. Since some software doesn't implement 2462 BMPStrings or UTF8Strings yet, they can be enabled using the config file 2463 using the dirstring_type option. See the new comment in the default 2464 openssl.cnf for more info. 2465 [Steve Henson] 2466 2467 *) Make crypto/rand/md_rand.c more robust: 2468 - Assure unique random numbers after fork(). 2469 - Make sure that concurrent threads access the global counter and 2470 md serializably so that we never lose entropy in them 2471 or use exactly the same state in multiple threads. 2472 Access to the large state is not always serializable because 2473 the additional locking could be a performance killer, and 2474 md should be large enough anyway. 2475 [Bodo Moeller] 2476 2477 *) New file apps/app_rand.c with commonly needed functionality 2478 for handling the random seed file. 2479 2480 Use the random seed file in some applications that previously did not: 2481 ca, 2482 dsaparam -genkey (which also ignored its '-rand' option), 2483 s_client, 2484 s_server, 2485 x509 (when signing). 2486 Except on systems with /dev/urandom, it is crucial to have a random 2487 seed file at least for key creation, DSA signing, and for DH exchanges; 2488 for RSA signatures we could do without one. 2489 2490 gendh and gendsa (unlike genrsa) used to read only the first byte 2491 of each file listed in the '-rand' option. The function as previously 2492 found in genrsa is now in app_rand.c and is used by all programs 2493 that support '-rand'. 2494 [Bodo Moeller] 2495 2496 *) In RAND_write_file, use mode 0600 for creating files; 2497 don't just chmod when it may be too late. 2498 [Bodo Moeller] 2499 2500 *) Report an error from X509_STORE_load_locations 2501 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 2502 [Bill Perry] 2503 2504 *) New function ASN1_mbstring_copy() this copies a string in either 2505 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 2506 into an ASN1_STRING type. A mask of permissible types is passed 2507 and it chooses the "minimal" type to use or an error if not type 2508 is suitable. 2509 [Steve Henson] 2510 2511 *) Add function equivalents to the various macros in asn1.h. The old 2512 macros are retained with an M_ prefix. Code inside the library can 2513 use the M_ macros. External code (including the openssl utility) 2514 should *NOT* in order to be "shared library friendly". 2515 [Steve Henson] 2516 2517 *) Add various functions that can check a certificate's extensions 2518 to see if it usable for various purposes such as SSL client, 2519 server or S/MIME and CAs of these types. This is currently 2520 VERY EXPERIMENTAL but will ultimately be used for certificate chain 2521 verification. Also added a -purpose flag to x509 utility to 2522 print out all the purposes. 2523 [Steve Henson] 2524 2525 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 2526 functions. 2527 [Steve Henson] 2528 2529 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 2530 for, obtain and decode and extension and obtain its critical flag. 2531 This allows all the necessary extension code to be handled in a 2532 single function call. 2533 [Steve Henson] 2534 2535 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 2536 platforms. See crypto/rc4/rc4_enc.c for further details. 2537 [Andy Polyakov] 2538 2539 *) New -noout option to asn1parse. This causes no output to be produced 2540 its main use is when combined with -strparse and -out to extract data 2541 from a file (which may not be in ASN.1 format). 2542 [Steve Henson] 2543 2544 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 2545 when producing the local key id. 2546 [Richard Levitte <levitte@stacken.kth.se>] 2547 2548 *) New option -dhparam in s_server. This allows a DH parameter file to be 2549 stated explicitly. If it is not stated then it tries the first server 2550 certificate file. The previous behaviour hard coded the filename 2551 "server.pem". 2552 [Steve Henson] 2553 2554 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 2555 a public key to be input or output. For example: 2556 openssl rsa -in key.pem -pubout -out pubkey.pem 2557 Also added necessary DSA public key functions to handle this. 2558 [Steve Henson] 2559 2560 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 2561 in the message. This was handled by allowing 2562 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 2563 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 2564 2565 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 2566 to the end of the strings whereas this didn't. This would cause problems 2567 if strings read with d2i_ASN1_bytes() were later modified. 2568 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 2569 2570 *) Fix for base64 decode bug. When a base64 bio reads only one line of 2571 data and it contains EOF it will end up returning an error. This is 2572 caused by input 46 bytes long. The cause is due to the way base64 2573 BIOs find the start of base64 encoded data. They do this by trying a 2574 trial decode on each line until they find one that works. When they 2575 do a flag is set and it starts again knowing it can pass all the 2576 data directly through the decoder. Unfortunately it doesn't reset 2577 the context it uses. This means that if EOF is reached an attempt 2578 is made to pass two EOFs through the context and this causes the 2579 resulting error. This can also cause other problems as well. As is 2580 usual with these problems it takes *ages* to find and the fix is 2581 trivial: move one line. 2582 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 2583 2584 *) Ugly workaround to get s_client and s_server working under Windows. The 2585 old code wouldn't work because it needed to select() on sockets and the 2586 tty (for keypresses and to see if data could be written). Win32 only 2587 supports select() on sockets so we select() with a 1s timeout on the 2588 sockets and then see if any characters are waiting to be read, if none 2589 are present then we retry, we also assume we can always write data to 2590 the tty. This isn't nice because the code then blocks until we've 2591 received a complete line of data and it is effectively polling the 2592 keyboard at 1s intervals: however it's quite a bit better than not 2593 working at all :-) A dedicated Windows application might handle this 2594 with an event loop for example. 2595 [Steve Henson] 2596 2597 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 2598 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 2599 will be called when RSA_sign() and RSA_verify() are used. This is useful 2600 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 2601 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 2602 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 2603 This necessitated the support of an extra signature type NID_md5_sha1 2604 for SSL signatures and modifications to the SSL library to use it instead 2605 of calling RSA_public_decrypt() and RSA_private_encrypt(). 2606 [Steve Henson] 2607 2608 *) Add new -verify -CAfile and -CApath options to the crl program, these 2609 will lookup a CRL issuers certificate and verify the signature in a 2610 similar way to the verify program. Tidy up the crl program so it 2611 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 2612 less strict. It will now permit CRL extensions even if it is not 2613 a V2 CRL: this will allow it to tolerate some broken CRLs. 2614 [Steve Henson] 2615 2616 *) Initialize all non-automatic variables each time one of the openssl 2617 sub-programs is started (this is necessary as they may be started 2618 multiple times from the "OpenSSL>" prompt). 2619 [Lennart Bang, Bodo Moeller] 2620 2621 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 2622 removing all other RSA functionality (this is what NO_RSA does). This 2623 is so (for example) those in the US can disable those operations covered 2624 by the RSA patent while allowing storage and parsing of RSA keys and RSA 2625 key generation. 2626 [Steve Henson] 2627 2628 *) Non-copying interface to BIO pairs. 2629 (still largely untested) 2630 [Bodo Moeller] 2631 2632 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 2633 ASCII string. This was handled independently in various places before. 2634 [Steve Henson] 2635 2636 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 2637 UTF8 strings a character at a time. 2638 [Steve Henson] 2639 2640 *) Use client_version from client hello to select the protocol 2641 (s23_srvr.c) and for RSA client key exchange verification 2642 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 2643 [Bodo Moeller] 2644 2645 *) Add various utility functions to handle SPKACs, these were previously 2646 handled by poking round in the structure internals. Added new function 2647 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 2648 print, verify and generate SPKACs. Based on an original idea from 2649 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 2650 [Steve Henson] 2651 2652 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 2653 [Andy Polyakov] 2654 2655 *) Allow the config file extension section to be overwritten on the 2656 command line. Based on an original idea from Massimiliano Pala 2657 <madwolf@comune.modena.it>. The new option is called -extensions 2658 and can be applied to ca, req and x509. Also -reqexts to override 2659 the request extensions in req and -crlexts to override the crl extensions 2660 in ca. 2661 [Steve Henson] 2662 2663 *) Add new feature to the SPKAC handling in ca. Now you can include 2664 the same field multiple times by preceding it by "XXXX." for example: 2665 1.OU="Unit name 1" 2666 2.OU="Unit name 2" 2667 this is the same syntax as used in the req config file. 2668 [Steve Henson] 2669 2670 *) Allow certificate extensions to be added to certificate requests. These 2671 are specified in a 'req_extensions' option of the req section of the 2672 config file. They can be printed out with the -text option to req but 2673 are otherwise ignored at present. 2674 [Steve Henson] 2675 2676 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 2677 data read consists of only the final block it would not decrypted because 2678 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 2679 A misplaced 'break' also meant the decrypted final block might not be 2680 copied until the next read. 2681 [Steve Henson] 2682 2683 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 2684 a few extra parameters to the DH structure: these will be useful if 2685 for example we want the value of 'q' or implement X9.42 DH. 2686 [Steve Henson] 2687 2688 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 2689 provides hooks that allow the default DSA functions or functions on a 2690 "per key" basis to be replaced. This allows hardware acceleration and 2691 hardware key storage to be handled without major modification to the 2692 library. Also added low level modexp hooks and CRYPTO_EX structure and 2693 associated functions. 2694 [Steve Henson] 2695 2696 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 2697 as "read only": it can't be written to and the buffer it points to will 2698 not be freed. Reading from a read only BIO is much more efficient than 2699 a normal memory BIO. This was added because there are several times when 2700 an area of memory needs to be read from a BIO. The previous method was 2701 to create a memory BIO and write the data to it, this results in two 2702 copies of the data and an O(n^2) reading algorithm. There is a new 2703 function BIO_new_mem_buf() which creates a read only memory BIO from 2704 an area of memory. Also modified the PKCS#7 routines to use read only 2705 memory BIOs. 2706 [Steve Henson] 2707 2708 *) Bugfix: ssl23_get_client_hello did not work properly when called in 2709 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 2710 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 2711 but a retry condition occured while trying to read the rest. 2712 [Bodo Moeller] 2713 2714 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 2715 NID_pkcs7_encrypted by default: this was wrong since this should almost 2716 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 2717 the encrypted data type: this is a more sensible place to put it and it 2718 allows the PKCS#12 code to be tidied up that duplicated this 2719 functionality. 2720 [Steve Henson] 2721 2722 *) Changed obj_dat.pl script so it takes its input and output files on 2723 the command line. This should avoid shell escape redirection problems 2724 under Win32. 2725 [Steve Henson] 2726 2727 *) Initial support for certificate extension requests, these are included 2728 in things like Xenroll certificate requests. Included functions to allow 2729 extensions to be obtained and added. 2730 [Steve Henson] 2731 2732 *) -crlf option to s_client and s_server for sending newlines as 2733 CRLF (as required by many protocols). 2734 [Bodo Moeller] 2735 2736 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 2737 2738 *) Install libRSAglue.a when OpenSSL is built with RSAref. 2739 [Ralf S. Engelschall] 2740 2741 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 2742 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 2743 2744 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 2745 program. 2746 [Steve Henson] 2747 2748 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 2749 DH parameters/keys (q is lost during that conversion, but the resulting 2750 DH parameters contain its length). 2751 2752 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 2753 much faster than DH_generate_parameters (which creates parameters 2754 where p = 2*q + 1), and also the smaller q makes DH computations 2755 much more efficient (160-bit exponentiation instead of 1024-bit 2756 exponentiation); so this provides a convenient way to support DHE 2757 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 2758 utter importance to use 2759 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 2760 or 2761 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 2762 when such DH parameters are used, because otherwise small subgroup 2763 attacks may become possible! 2764 [Bodo Moeller] 2765 2766 *) Avoid memory leak in i2d_DHparams. 2767 [Bodo Moeller] 2768 2769 *) Allow the -k option to be used more than once in the enc program: 2770 this allows the same encrypted message to be read by multiple recipients. 2771 [Steve Henson] 2772 2773 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 2774 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 2775 it will always use the numerical form of the OID, even if it has a short 2776 or long name. 2777 [Steve Henson] 2778 2779 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 2780 method only got called if p,q,dmp1,dmq1,iqmp components were present, 2781 otherwise bn_mod_exp was called. In the case of hardware keys for example 2782 no private key components need be present and it might store extra data 2783 in the RSA structure, which cannot be accessed from bn_mod_exp. 2784 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 2785 private key operations. 2786 [Steve Henson] 2787 2788 *) Added support for SPARC Linux. 2789 [Andy Polyakov] 2790 2791 *) pem_password_cb function type incompatibly changed from 2792 typedef int pem_password_cb(char *buf, int size, int rwflag); 2793 to 2794 ....(char *buf, int size, int rwflag, void *userdata); 2795 so that applications can pass data to their callbacks: 2796 The PEM[_ASN1]_{read,write}... functions and macros now take an 2797 additional void * argument, which is just handed through whenever 2798 the password callback is called. 2799 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 2800 2801 New function SSL_CTX_set_default_passwd_cb_userdata. 2802 2803 Compatibility note: As many C implementations push function arguments 2804 onto the stack in reverse order, the new library version is likely to 2805 interoperate with programs that have been compiled with the old 2806 pem_password_cb definition (PEM_whatever takes some data that 2807 happens to be on the stack as its last argument, and the callback 2808 just ignores this garbage); but there is no guarantee whatsoever that 2809 this will work. 2810 2811 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 2812 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 2813 problems not only on Windows, but also on some Unix platforms. 2814 To avoid problematic command lines, these definitions are now in an 2815 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 2816 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 2817 [Bodo Moeller] 2818 2819 *) MIPS III/IV assembler module is reimplemented. 2820 [Andy Polyakov] 2821 2822 *) More DES library cleanups: remove references to srand/rand and 2823 delete an unused file. 2824 [Ulf M�ller] 2825 2826 *) Add support for the the free Netwide assembler (NASM) under Win32, 2827 since not many people have MASM (ml) and it can be hard to obtain. 2828 This is currently experimental but it seems to work OK and pass all 2829 the tests. Check out INSTALL.W32 for info. 2830 [Steve Henson] 2831 2832 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 2833 without temporary keys kept an extra copy of the server key, 2834 and connections with temporary keys did not free everything in case 2835 of an error. 2836 [Bodo Moeller] 2837 2838 *) New function RSA_check_key and new openssl rsa option -check 2839 for verifying the consistency of RSA keys. 2840 [Ulf Moeller, Bodo Moeller] 2841 2842 *) Various changes to make Win32 compile work: 2843 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 2844 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 2845 comparison" warnings. 2846 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 2847 [Steve Henson] 2848 2849 *) Add a debugging option to PKCS#5 v2 key generation function: when 2850 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 2851 derived keys are printed to stderr. 2852 [Steve Henson] 2853 2854 *) Copy the flags in ASN1_STRING_dup(). 2855 [Roman E. Pavlov <pre@mo.msk.ru>] 2856 2857 *) The x509 application mishandled signing requests containing DSA 2858 keys when the signing key was also DSA and the parameters didn't match. 2859 2860 It was supposed to omit the parameters when they matched the signing key: 2861 the verifying software was then supposed to automatically use the CA's 2862 parameters if they were absent from the end user certificate. 2863 2864 Omitting parameters is no longer recommended. The test was also 2865 the wrong way round! This was probably due to unusual behaviour in 2866 EVP_cmp_parameters() which returns 1 if the parameters match. 2867 This meant that parameters were omitted when they *didn't* match and 2868 the certificate was useless. Certificates signed with 'ca' didn't have 2869 this bug. 2870 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 2871 2872 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 2873 The interface is as follows: 2874 Applications can use 2875 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 2876 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 2877 "off" is now the default. 2878 The library internally uses 2879 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 2880 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 2881 to disable memory-checking temporarily. 2882 2883 Some inconsistent states that previously were possible (and were 2884 even the default) are now avoided. 2885 2886 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 2887 with each memory chunk allocated; this is occasionally more helpful 2888 than just having a counter. 2889 2890 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 2891 2892 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 2893 extensions. 2894 [Bodo Moeller] 2895 2896 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 2897 which largely parallels "options", but is for changing API behaviour, 2898 whereas "options" are about protocol behaviour. 2899 Initial "mode" flags are: 2900 2901 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 2902 a single record has been written. 2903 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 2904 retries use the same buffer location. 2905 (But all of the contents must be 2906 copied!) 2907 [Bodo Moeller] 2908 2909 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 2910 worked. 2911 2912 *) Fix problems with no-hmac etc. 2913 [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 2914 2915 *) New functions RSA_get_default_method(), RSA_set_method() and 2916 RSA_get_method(). These allows replacement of RSA_METHODs without having 2917 to mess around with the internals of an RSA structure. 2918 [Steve Henson] 2919 2920 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 2921 Also really enable memory leak checks in openssl.c and in some 2922 test programs. 2923 [Chad C. Mulligan, Bodo Moeller] 2924 2925 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 2926 up the length of negative integers. This has now been simplified to just 2927 store the length when it is first determined and use it later, rather 2928 than trying to keep track of where data is copied and updating it to 2929 point to the end. 2930 [Steve Henson, reported by Brien Wheeler 2931 <bwheeler@authentica-security.com>] 2932 2933 *) Add a new function PKCS7_signatureVerify. This allows the verification 2934 of a PKCS#7 signature but with the signing certificate passed to the 2935 function itself. This contrasts with PKCS7_dataVerify which assumes the 2936 certificate is present in the PKCS#7 structure. This isn't always the 2937 case: certificates can be omitted from a PKCS#7 structure and be 2938 distributed by "out of band" means (such as a certificate database). 2939 [Steve Henson] 2940 2941 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 2942 function prototypes in pem.h, also change util/mkdef.pl to add the 2943 necessary function names. 2944 [Steve Henson] 2945 2946 *) mk1mf.pl (used by Windows builds) did not properly read the 2947 options set by Configure in the top level Makefile, and Configure 2948 was not even able to write more than one option correctly. 2949 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 2950 [Bodo Moeller] 2951 2952 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 2953 file to be loaded from a BIO or FILE pointer. The BIO version will 2954 for example allow memory BIOs to contain config info. 2955 [Steve Henson] 2956 2957 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 2958 Whoever hopes to achieve shared-library compatibility across versions 2959 must use this, not the compile-time macro. 2960 (Exercise 0.9.4: Which is the minimum library version required by 2961 such programs?) 2962 Note: All this applies only to multi-threaded programs, others don't 2963 need locks. 2964 [Bodo Moeller] 2965 2966 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 2967 through a BIO pair triggered the default case, i.e. 2968 SSLerr(...,SSL_R_UNKNOWN_STATE). 2969 [Bodo Moeller] 2970 2971 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 2972 can use the SSL library even if none of the specific BIOs is 2973 appropriate. 2974 [Bodo Moeller] 2975 2976 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 2977 for the encoded length. 2978 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 2979 2980 *) Add initial documentation of the X509V3 functions. 2981 [Steve Henson] 2982 2983 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 2984 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 2985 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 2986 secure PKCS#8 private key format with a high iteration count. 2987 [Steve Henson] 2988 2989 *) Fix determination of Perl interpreter: A perl or perl5 2990 _directory_ in $PATH was also accepted as the interpreter. 2991 [Ralf S. Engelschall] 2992 2993 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 2994 wrong with it but it was very old and did things like calling 2995 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 2996 unusual formatting. 2997 [Steve Henson] 2998 2999 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 3000 to use the new extension code. 3001 [Steve Henson] 3002 3003 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 3004 with macros. This should make it easier to change their form, add extra 3005 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 3006 constant. 3007 [Steve Henson] 3008 3009 *) Add to configuration table a new entry that can specify an alternative 3010 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 3011 according to Mark Crispin <MRC@Panda.COM>. 3012 [Bodo Moeller] 3013 3014#if 0 3015 *) DES CBC did not update the IV. Weird. 3016 [Ben Laurie] 3017#else 3018 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 3019 Changing the behaviour of the former might break existing programs -- 3020 where IV updating is needed, des_ncbc_encrypt can be used. 3021#endif 3022 3023 *) When bntest is run from "make test" it drives bc to check its 3024 calculations, as well as internally checking them. If an internal check 3025 fails, it needs to cause bc to give a non-zero result or make test carries 3026 on without noticing the failure. Fixed. 3027 [Ben Laurie] 3028 3029 *) DES library cleanups. 3030 [Ulf M�ller] 3031 3032 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 3033 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 3034 ciphers. NOTE: although the key derivation function has been verified 3035 against some published test vectors it has not been extensively tested 3036 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 3037 of v2.0. 3038 [Steve Henson] 3039 3040 *) Instead of "mkdir -p", which is not fully portable, use new 3041 Perl script "util/mkdir-p.pl". 3042 [Bodo Moeller] 3043 3044 *) Rewrite the way password based encryption (PBE) is handled. It used to 3045 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 3046 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 3047 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 3048 the 'parameter' field of the AlgorithmIdentifier is passed to the 3049 underlying key generation function so it must do its own ASN1 parsing. 3050 This has also changed the EVP_PBE_CipherInit() function which now has a 3051 'parameter' argument instead of literal salt and iteration count values 3052 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 3053 [Steve Henson] 3054 3055 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 3056 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 3057 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 3058 KEY" because this clashed with PKCS#8 unencrypted string. Since this 3059 value was just used as a "magic string" and not used directly its 3060 value doesn't matter. 3061 [Steve Henson] 3062 3063 *) Introduce some semblance of const correctness to BN. Shame C doesn't 3064 support mutable. 3065 [Ben Laurie] 3066 3067 *) "linux-sparc64" configuration (ultrapenguin). 3068 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 3069 "linux-sparc" configuration. 3070 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 3071 3072 *) config now generates no-xxx options for missing ciphers. 3073 [Ulf M�ller] 3074 3075 *) Support the EBCDIC character set (work in progress). 3076 File ebcdic.c not yet included because it has a different license. 3077 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 3078 3079 *) Support BS2000/OSD-POSIX. 3080 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 3081 3082 *) Make callbacks for key generation use void * instead of char *. 3083 [Ben Laurie] 3084 3085 *) Make S/MIME samples compile (not yet tested). 3086 [Ben Laurie] 3087 3088 *) Additional typesafe stacks. 3089 [Ben Laurie] 3090 3091 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 3092 [Bodo Moeller] 3093 3094 3095 Changes between 0.9.3 and 0.9.3a [29 May 1999] 3096 3097 *) New configuration variant "sco5-gcc". 3098 3099 *) Updated some demos. 3100 [Sean O Riordain, Wade Scholine] 3101 3102 *) Add missing BIO_free at exit of pkcs12 application. 3103 [Wu Zhigang] 3104 3105 *) Fix memory leak in conf.c. 3106 [Steve Henson] 3107 3108 *) Updates for Win32 to assembler version of MD5. 3109 [Steve Henson] 3110 3111 *) Set #! path to perl in apps/der_chop to where we found it 3112 instead of using a fixed path. 3113 [Bodo Moeller] 3114 3115 *) SHA library changes for irix64-mips4-cc. 3116 [Andy Polyakov] 3117 3118 *) Improvements for VMS support. 3119 [Richard Levitte] 3120 3121 3122 Changes between 0.9.2b and 0.9.3 [24 May 1999] 3123 3124 *) Bignum library bug fix. IRIX 6 passes "make test" now! 3125 This also avoids the problems with SC4.2 and unpatched SC5. 3126 [Andy Polyakov <appro@fy.chalmers.se>] 3127 3128 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 3129 These are required because of the typesafe stack would otherwise break 3130 existing code. If old code used a structure member which used to be STACK 3131 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 3132 sk_num or sk_value it would produce an error because the num, data members 3133 are not present in STACK_OF. Now it just produces a warning. sk_set 3134 replaces the old method of assigning a value to sk_value 3135 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 3136 that does this will no longer work (and should use sk_set instead) but 3137 this could be regarded as a "questionable" behaviour anyway. 3138 [Steve Henson] 3139 3140 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 3141 correctly handle encrypted S/MIME data. 3142 [Steve Henson] 3143 3144 *) Change type of various DES function arguments from des_cblock 3145 (which means, in function argument declarations, pointer to char) 3146 to des_cblock * (meaning pointer to array with 8 char elements), 3147 which allows the compiler to do more typechecking; it was like 3148 that back in SSLeay, but with lots of ugly casts. 3149 3150 Introduce new type const_des_cblock. 3151 [Bodo Moeller] 3152 3153 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 3154 problems: find RecipientInfo structure that matches recipient certificate 3155 and initialise the ASN1 structures properly based on passed cipher. 3156 [Steve Henson] 3157 3158 *) Belatedly make the BN tests actually check the results. 3159 [Ben Laurie] 3160 3161 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 3162 to and from BNs: it was completely broken. New compilation option 3163 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 3164 key elements as negative integers. 3165 [Steve Henson] 3166 3167 *) Reorganize and speed up MD5. 3168 [Andy Polyakov <appro@fy.chalmers.se>] 3169 3170 *) VMS support. 3171 [Richard Levitte <richard@levitte.org>] 3172 3173 *) New option -out to asn1parse to allow the parsed structure to be 3174 output to a file. This is most useful when combined with the -strparse 3175 option to examine the output of things like OCTET STRINGS. 3176 [Steve Henson] 3177 3178 *) Make SSL library a little more fool-proof by not requiring any longer 3179 that SSL_set_{accept,connect}_state be called before 3180 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 3181 in many applications because usually everything *appeared* to work as 3182 intended anyway -- now it really works as intended). 3183 [Bodo Moeller] 3184 3185 *) Move openssl.cnf out of lib/. 3186 [Ulf M�ller] 3187 3188 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 3189 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 3190 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 3191 [Ralf S. Engelschall] 3192 3193 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 3194 handle PKCS#7 enveloped data properly. 3195 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 3196 3197 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 3198 copying pointers. The cert_st handling is changed by this in 3199 various ways (and thus what used to be known as ctx->default_cert 3200 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 3201 any longer when s->cert does not give us what we need). 3202 ssl_cert_instantiate becomes obsolete by this change. 3203 As soon as we've got the new code right (possibly it already is?), 3204 we have solved a couple of bugs of the earlier code where s->cert 3205 was used as if it could not have been shared with other SSL structures. 3206 3207 Note that using the SSL API in certain dirty ways now will result 3208 in different behaviour than observed with earlier library versions: 3209 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 3210 does not influence s as it used to. 3211 3212 In order to clean up things more thoroughly, inside SSL_SESSION 3213 we don't use CERT any longer, but a new structure SESS_CERT 3214 that holds per-session data (if available); currently, this is 3215 the peer's certificate chain and, for clients, the server's certificate 3216 and temporary key. CERT holds only those values that can have 3217 meaningful defaults in an SSL_CTX. 3218 [Bodo Moeller] 3219 3220 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 3221 from the internal representation. Various PKCS#7 fixes: remove some 3222 evil casts and set the enc_dig_alg field properly based on the signing 3223 key type. 3224 [Steve Henson] 3225 3226 *) Allow PKCS#12 password to be set from the command line or the 3227 environment. Let 'ca' get its config file name from the environment 3228 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 3229 and 'x509'). 3230 [Steve Henson] 3231 3232 *) Allow certificate policies extension to use an IA5STRING for the 3233 organization field. This is contrary to the PKIX definition but 3234 VeriSign uses it and IE5 only recognises this form. Document 'x509' 3235 extension option. 3236 [Steve Henson] 3237 3238 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 3239 without disallowing inline assembler and the like for non-pedantic builds. 3240 [Ben Laurie] 3241 3242 *) Support Borland C++ builder. 3243 [Janez Jere <jj@void.si>, modified by Ulf M�ller] 3244 3245 *) Support Mingw32. 3246 [Ulf M�ller] 3247 3248 *) SHA-1 cleanups and performance enhancements. 3249 [Andy Polyakov <appro@fy.chalmers.se>] 3250 3251 *) Sparc v8plus assembler for the bignum library. 3252 [Andy Polyakov <appro@fy.chalmers.se>] 3253 3254 *) Accept any -xxx and +xxx compiler options in Configure. 3255 [Ulf M�ller] 3256 3257 *) Update HPUX configuration. 3258 [Anonymous] 3259 3260 *) Add missing sk_<type>_unshift() function to safestack.h 3261 [Ralf S. Engelschall] 3262 3263 *) New function SSL_CTX_use_certificate_chain_file that sets the 3264 "extra_cert"s in addition to the certificate. (This makes sense 3265 only for "PEM" format files, as chains as a whole are not 3266 DER-encoded.) 3267 [Bodo Moeller] 3268 3269 *) Support verify_depth from the SSL API. 3270 x509_vfy.c had what can be considered an off-by-one-error: 3271 Its depth (which was not part of the external interface) 3272 was actually counting the number of certificates in a chain; 3273 now it really counts the depth. 3274 [Bodo Moeller] 3275 3276 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 3277 instead of X509err, which often resulted in confusing error 3278 messages since the error codes are not globally unique 3279 (e.g. an alleged error in ssl3_accept when a certificate 3280 didn't match the private key). 3281 3282 *) New function SSL_CTX_set_session_id_context that allows to set a default 3283 value (so that you don't need SSL_set_session_id_context for each 3284 connection using the SSL_CTX). 3285 [Bodo Moeller] 3286 3287 *) OAEP decoding bug fix. 3288 [Ulf M�ller] 3289 3290 *) Support INSTALL_PREFIX for package builders, as proposed by 3291 David Harris. 3292 [Bodo Moeller] 3293 3294 *) New Configure options "threads" and "no-threads". For systems 3295 where the proper compiler options are known (currently Solaris 3296 and Linux), "threads" is the default. 3297 [Bodo Moeller] 3298 3299 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 3300 [Bodo Moeller] 3301 3302 *) Install various scripts to $(OPENSSLDIR)/misc, not to 3303 $(INSTALLTOP)/bin -- they shouldn't clutter directories 3304 such as /usr/local/bin. 3305 [Bodo Moeller] 3306 3307 *) "make linux-shared" to build shared libraries. 3308 [Niels Poppe <niels@netbox.org>] 3309 3310 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 3311 [Ulf M�ller] 3312 3313 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 3314 extension adding in x509 utility. 3315 [Steve Henson] 3316 3317 *) Remove NOPROTO sections and error code comments. 3318 [Ulf M�ller] 3319 3320 *) Partial rewrite of the DEF file generator to now parse the ANSI 3321 prototypes. 3322 [Steve Henson] 3323 3324 *) New Configure options --prefix=DIR and --openssldir=DIR. 3325 [Ulf M�ller] 3326 3327 *) Complete rewrite of the error code script(s). It is all now handled 3328 by one script at the top level which handles error code gathering, 3329 header rewriting and C source file generation. It should be much better 3330 than the old method: it now uses a modified version of Ulf's parser to 3331 read the ANSI prototypes in all header files (thus the old K&R definitions 3332 aren't needed for error creation any more) and do a better job of 3333 translating function codes into names. The old 'ASN1 error code imbedded 3334 in a comment' is no longer necessary and it doesn't use .err files which 3335 have now been deleted. Also the error code call doesn't have to appear all 3336 on one line (which resulted in some large lines...). 3337 [Steve Henson] 3338 3339 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 3340 [Bodo Moeller] 3341 3342 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 3343 0 (which usually indicates a closed connection), but continue reading. 3344 [Bodo Moeller] 3345 3346 *) Fix some race conditions. 3347 [Bodo Moeller] 3348 3349 *) Add support for CRL distribution points extension. Add Certificate 3350 Policies and CRL distribution points documentation. 3351 [Steve Henson] 3352 3353 *) Move the autogenerated header file parts to crypto/opensslconf.h. 3354 [Ulf M�ller] 3355 3356 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 3357 8 of keying material. Merlin has also confirmed interop with this fix 3358 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 3359 [Merlin Hughes <merlin@baltimore.ie>] 3360 3361 *) Fix lots of warnings. 3362 [Richard Levitte <levitte@stacken.kth.se>] 3363 3364 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 3365 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 3366 [Richard Levitte <levitte@stacken.kth.se>] 3367 3368 *) Fix problems with sizeof(long) == 8. 3369 [Andy Polyakov <appro@fy.chalmers.se>] 3370 3371 *) Change functions to ANSI C. 3372 [Ulf M�ller] 3373 3374 *) Fix typos in error codes. 3375 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 3376 3377 *) Remove defunct assembler files from Configure. 3378 [Ulf M�ller] 3379 3380 *) SPARC v8 assembler BIGNUM implementation. 3381 [Andy Polyakov <appro@fy.chalmers.se>] 3382 3383 *) Support for Certificate Policies extension: both print and set. 3384 Various additions to support the r2i method this uses. 3385 [Steve Henson] 3386 3387 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 3388 return a const string when you are expecting an allocated buffer. 3389 [Ben Laurie] 3390 3391 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 3392 types DirectoryString and DisplayText. 3393 [Steve Henson] 3394 3395 *) Add code to allow r2i extensions to access the configuration database, 3396 add an LHASH database driver and add several ctx helper functions. 3397 [Steve Henson] 3398 3399 *) Fix an evil bug in bn_expand2() which caused various BN functions to 3400 fail when they extended the size of a BIGNUM. 3401 [Steve Henson] 3402 3403 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 3404 support typesafe stack. 3405 [Steve Henson] 3406 3407 *) Fix typo in SSL_[gs]et_options(). 3408 [Nils Frostberg <nils@medcom.se>] 3409 3410 *) Delete various functions and files that belonged to the (now obsolete) 3411 old X509V3 handling code. 3412 [Steve Henson] 3413 3414 *) New Configure option "rsaref". 3415 [Ulf M�ller] 3416 3417 *) Don't auto-generate pem.h. 3418 [Bodo Moeller] 3419 3420 *) Introduce type-safe ASN.1 SETs. 3421 [Ben Laurie] 3422 3423 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 3424 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 3425 3426 *) Introduce type-safe STACKs. This will almost certainly break lots of code 3427 that links with OpenSSL (well at least cause lots of warnings), but fear 3428 not: the conversion is trivial, and it eliminates loads of evil casts. A 3429 few STACKed things have been converted already. Feel free to convert more. 3430 In the fullness of time, I'll do away with the STACK type altogether. 3431 [Ben Laurie] 3432 3433 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 3434 specified in <certfile> by updating the entry in the index.txt file. 3435 This way one no longer has to edit the index.txt file manually for 3436 revoking a certificate. The -revoke option does the gory details now. 3437 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 3438 3439 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 3440 `-text' option at all and this way the `-noout -text' combination was 3441 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 3442 [Ralf S. Engelschall] 3443 3444 *) Make sure a corresponding plain text error message exists for the 3445 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 3446 verify callback function determined that a certificate was revoked. 3447 [Ralf S. Engelschall] 3448 3449 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 3450 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 3451 all available cipers including rc5, which was forgotten until now. 3452 In order to let the testing shell script know which algorithms 3453 are available, a new (up to now undocumented) command 3454 "openssl list-cipher-commands" is used. 3455 [Bodo Moeller] 3456 3457 *) Bugfix: s_client occasionally would sleep in select() when 3458 it should have checked SSL_pending() first. 3459 [Bodo Moeller] 3460 3461 *) New functions DSA_do_sign and DSA_do_verify to provide access to 3462 the raw DSA values prior to ASN.1 encoding. 3463 [Ulf M�ller] 3464 3465 *) Tweaks to Configure 3466 [Niels Poppe <niels@netbox.org>] 3467 3468 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 3469 yet... 3470 [Steve Henson] 3471 3472 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 3473 [Ulf M�ller] 3474 3475 *) New config option to avoid instructions that are illegal on the 80386. 3476 The default code is faster, but requires at least a 486. 3477 [Ulf M�ller] 3478 3479 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 3480 SSL2_SERVER_VERSION (not used at all) macros, which are now the 3481 same as SSL2_VERSION anyway. 3482 [Bodo Moeller] 3483 3484 *) New "-showcerts" option for s_client. 3485 [Bodo Moeller] 3486 3487 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 3488 application. Various cleanups and fixes. 3489 [Steve Henson] 3490 3491 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 3492 modify error routines to work internally. Add error codes and PBE init 3493 to library startup routines. 3494 [Steve Henson] 3495 3496 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 3497 packing functions to asn1 and evp. Changed function names and error 3498 codes along the way. 3499 [Steve Henson] 3500 3501 *) PKCS12 integration: and so it begins... First of several patches to 3502 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 3503 objects to objects.h 3504 [Steve Henson] 3505 3506 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 3507 and display support for Thawte strong extranet extension. 3508 [Steve Henson] 3509 3510 *) Add LinuxPPC support. 3511 [Jeff Dubrule <igor@pobox.org>] 3512 3513 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 3514 bn_div_words in alpha.s. 3515 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 3516 3517 *) Make sure the RSA OAEP test is skipped under -DRSAref because 3518 OAEP isn't supported when OpenSSL is built with RSAref. 3519 [Ulf Moeller <ulf@fitug.de>] 3520 3521 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 3522 so they no longer are missing under -DNOPROTO. 3523 [Soren S. Jorvang <soren@t.dk>] 3524 3525 3526 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 3527 3528 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 3529 doesn't work when the session is reused. Coming soon! 3530 [Ben Laurie] 3531 3532 *) Fix a security hole, that allows sessions to be reused in the wrong 3533 context thus bypassing client cert protection! All software that uses 3534 client certs and session caches in multiple contexts NEEDS PATCHING to 3535 allow session reuse! A fuller solution is in the works. 3536 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 3537 3538 *) Some more source tree cleanups (removed obsolete files 3539 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 3540 permission on "config" script to be executable) and a fix for the INSTALL 3541 document. 3542 [Ulf Moeller <ulf@fitug.de>] 3543 3544 *) Remove some legacy and erroneous uses of malloc, free instead of 3545 Malloc, Free. 3546 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 3547 3548 *) Make rsa_oaep_test return non-zero on error. 3549 [Ulf Moeller <ulf@fitug.de>] 3550 3551 *) Add support for native Solaris shared libraries. Configure 3552 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 3553 if someone would make that last step automatic. 3554 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 3555 3556 *) ctx_size was not built with the right compiler during "make links". Fixed. 3557 [Ben Laurie] 3558 3559 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 3560 except NULL ciphers". This means the default cipher list will no longer 3561 enable NULL ciphers. They need to be specifically enabled e.g. with 3562 the string "DEFAULT:eNULL". 3563 [Steve Henson] 3564 3565 *) Fix to RSA private encryption routines: if p < q then it would 3566 occasionally produce an invalid result. This will only happen with 3567 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 3568 [Steve Henson] 3569 3570 *) Be less restrictive and allow also `perl util/perlpath.pl 3571 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 3572 because this way one can also use an interpreter named `perl5' (which is 3573 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 3574 installed as `perl'). 3575 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 3576 3577 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 3578 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 3579 3580 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 3581 advapi32.lib to Win32 build and change the pem test comparision 3582 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 3583 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 3584 and crypto/des/ede_cbcm_enc.c. 3585 [Steve Henson] 3586 3587 *) DES quad checksum was broken on big-endian architectures. Fixed. 3588 [Ben Laurie] 3589 3590 *) Comment out two functions in bio.h that aren't implemented. Fix up the 3591 Win32 test batch file so it (might) work again. The Win32 test batch file 3592 is horrible: I feel ill.... 3593 [Steve Henson] 3594 3595 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 3596 in e_os.h. Audit of header files to check ANSI and non ANSI 3597 sections: 10 functions were absent from non ANSI section and not exported 3598 from Windows DLLs. Fixed up libeay.num for new functions. 3599 [Steve Henson] 3600 3601 *) Make `openssl version' output lines consistent. 3602 [Ralf S. Engelschall] 3603 3604 *) Fix Win32 symbol export lists for BIO functions: Added 3605 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 3606 to ms/libeay{16,32}.def. 3607 [Ralf S. Engelschall] 3608 3609 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 3610 fine under Unix and passes some trivial tests I've now added. But the 3611 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 3612 added to make sure no one expects that this stuff really works in the 3613 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 3614 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 3615 openssl_bio.xs. 3616 [Ralf S. Engelschall] 3617 3618 *) Fix the generation of two part addresses in perl. 3619 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 3620 3621 *) Add config entry for Linux on MIPS. 3622 [John Tobey <jtobey@channel1.com>] 3623 3624 *) Make links whenever Configure is run, unless we are on Windoze. 3625 [Ben Laurie] 3626 3627 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 3628 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 3629 in CRLs. 3630 [Steve Henson] 3631 3632 *) Add a useful kludge to allow package maintainers to specify compiler and 3633 other platforms details on the command line without having to patch the 3634 Configure script everytime: One now can use ``perl Configure 3635 <id>:<details>'', i.e. platform ids are allowed to have details appended 3636 to them (seperated by colons). This is treated as there would be a static 3637 pre-configured entry in Configure's %table under key <id> with value 3638 <details> and ``perl Configure <id>'' is called. So, when you want to 3639 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 3640 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 3641 now, which overrides the FreeBSD-elf entry on-the-fly. 3642 [Ralf S. Engelschall] 3643 3644 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 3645 [Ben Laurie] 3646 3647 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 3648 on the `perl Configure ...' command line. This way one can compile 3649 OpenSSL libraries with Position Independent Code (PIC) which is needed 3650 for linking it into DSOs. 3651 [Ralf S. Engelschall] 3652 3653 *) Remarkably, export ciphers were totally broken and no-one had noticed! 3654 Fixed. 3655 [Ben Laurie] 3656 3657 *) Cleaned up the LICENSE document: The official contact for any license 3658 questions now is the OpenSSL core team under openssl-core@openssl.org. 3659 And add a paragraph about the dual-license situation to make sure people 3660 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 3661 to the OpenSSL toolkit. 3662 [Ralf S. Engelschall] 3663 3664 *) General source tree makefile cleanups: Made `making xxx in yyy...' 3665 display consistent in the source tree and replaced `/bin/rm' by `rm'. 3666 Additonally cleaned up the `make links' target: Remove unnecessary 3667 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 3668 to speed processing and no longer clutter the display with confusing 3669 stuff. Instead only the actually done links are displayed. 3670 [Ralf S. Engelschall] 3671 3672 *) Permit null encryption ciphersuites, used for authentication only. It used 3673 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 3674 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 3675 encryption. 3676 [Ben Laurie] 3677 3678 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 3679 signed attributes when verifying signatures (this would break them), 3680 the detached data encoding was wrong and public keys obtained using 3681 X509_get_pubkey() weren't freed. 3682 [Steve Henson] 3683 3684 *) Add text documentation for the BUFFER functions. Also added a work around 3685 to a Win95 console bug. This was triggered by the password read stuff: the 3686 last character typed gets carried over to the next fread(). If you were 3687 generating a new cert request using 'req' for example then the last 3688 character of the passphrase would be CR which would then enter the first 3689 field as blank. 3690 [Steve Henson] 3691 3692 *) Added the new `Includes OpenSSL Cryptography Software' button as 3693 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 3694 button and can be used by applications based on OpenSSL to show the 3695 relationship to the OpenSSL project. 3696 [Ralf S. Engelschall] 3697 3698 *) Remove confusing variables in function signatures in files 3699 ssl/ssl_lib.c and ssl/ssl.h. 3700 [Lennart Bong <lob@kulthea.stacken.kth.se>] 3701 3702 *) Don't install bss_file.c under PREFIX/include/ 3703 [Lennart Bong <lob@kulthea.stacken.kth.se>] 3704 3705 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 3706 functions that return function pointers and has support for NT specific 3707 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 3708 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 3709 unsigned to signed types: this was killing the Win32 compile. 3710 [Steve Henson] 3711 3712 *) Add new certificate file to stack functions, 3713 SSL_add_dir_cert_subjects_to_stack() and 3714 SSL_add_file_cert_subjects_to_stack(). These largely supplant 3715 SSL_load_client_CA_file(), and can be used to add multiple certs easily 3716 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 3717 This means that Apache-SSL and similar packages don't have to mess around 3718 to add as many CAs as they want to the preferred list. 3719 [Ben Laurie] 3720 3721 *) Experiment with doxygen documentation. Currently only partially applied to 3722 ssl/ssl_lib.c. 3723 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 3724 openssl.doxy as the configuration file. 3725 [Ben Laurie] 3726 3727 *) Get rid of remaining C++-style comments which strict C compilers hate. 3728 [Ralf S. Engelschall, pointed out by Carlos Amengual] 3729 3730 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 3731 compiled in by default: it has problems with large keys. 3732 [Steve Henson] 3733 3734 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 3735 DH private keys and/or callback functions which directly correspond to 3736 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 3737 is needed for applications which have to configure certificates on a 3738 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 3739 (e.g. s_server). 3740 For the RSA certificate situation is makes no difference, but 3741 for the DSA certificate situation this fixes the "no shared cipher" 3742 problem where the OpenSSL cipher selection procedure failed because the 3743 temporary keys were not overtaken from the context and the API provided 3744 no way to reconfigure them. 3745 The new functions now let applications reconfigure the stuff and they 3746 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 3747 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 3748 non-public-API function ssl_cert_instantiate() is used as a helper 3749 function and also to reduce code redundancy inside ssl_rsa.c. 3750 [Ralf S. Engelschall] 3751 3752 *) Move s_server -dcert and -dkey options out of the undocumented feature 3753 area because they are useful for the DSA situation and should be 3754 recognized by the users. 3755 [Ralf S. Engelschall] 3756 3757 *) Fix the cipher decision scheme for export ciphers: the export bits are 3758 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 3759 SSL_EXP_MASK. So, the original variable has to be used instead of the 3760 already masked variable. 3761 [Richard Levitte <levitte@stacken.kth.se>] 3762 3763 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 3764 [Richard Levitte <levitte@stacken.kth.se>] 3765 3766 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 3767 from `int' to `unsigned int' because it's a length and initialized by 3768 EVP_DigestFinal() which expects an `unsigned int *'. 3769 [Richard Levitte <levitte@stacken.kth.se>] 3770 3771 *) Don't hard-code path to Perl interpreter on shebang line of Configure 3772 script. Instead use the usual Shell->Perl transition trick. 3773 [Ralf S. Engelschall] 3774 3775 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 3776 (in addition to RSA certificates) to match the behaviour of `openssl dsa 3777 -noout -modulus' as it's already the case for `openssl rsa -noout 3778 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 3779 currently the public key is printed (a decision which was already done by 3780 `openssl dsa -modulus' in the past) which serves a similar purpose. 3781 Additionally the NO_RSA no longer completely removes the whole -modulus 3782 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 3783 now, too. 3784 [Ralf S. Engelschall] 3785 3786 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 3787 BIO. See the source (crypto/evp/bio_ok.c) for more info. 3788 [Arne Ansper <arne@ats.cyber.ee>] 3789 3790 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 3791 to be added. Now both 'req' and 'ca' can use new objects defined in the 3792 config file. 3793 [Steve Henson] 3794 3795 *) Add cool BIO that does syslog (or event log on NT). 3796 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 3797 3798 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 3799 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 3800 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 3801 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 3802 [Ben Laurie] 3803 3804 *) Add preliminary config info for new extension code. 3805 [Steve Henson] 3806 3807 *) Make RSA_NO_PADDING really use no padding. 3808 [Ulf Moeller <ulf@fitug.de>] 3809 3810 *) Generate errors when private/public key check is done. 3811 [Ben Laurie] 3812 3813 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 3814 for some CRL extensions and new objects added. 3815 [Steve Henson] 3816 3817 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 3818 key usage extension and fuller support for authority key id. 3819 [Steve Henson] 3820 3821 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 3822 padding method for RSA, which is recommended for new applications in PKCS 3823 #1 v2.0 (RFC 2437, October 1998). 3824 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 3825 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 3826 against Bleichbacher's attack on RSA. 3827 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 3828 Ben Laurie] 3829 3830 *) Updates to the new SSL compression code 3831 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3832 3833 *) Fix so that the version number in the master secret, when passed 3834 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 3835 (because the server will not accept higher), that the version number 3836 is 0x03,0x01, not 0x03,0x00 3837 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3838 3839 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 3840 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 3841 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 3842 [Steve Henson] 3843 3844 *) Support for RAW extensions where an arbitrary extension can be 3845 created by including its DER encoding. See apps/openssl.cnf for 3846 an example. 3847 [Steve Henson] 3848 3849 *) Make sure latest Perl versions don't interpret some generated C array 3850 code as Perl array code in the crypto/err/err_genc.pl script. 3851 [Lars Weber <3weber@informatik.uni-hamburg.de>] 3852 3853 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 3854 not many people have the assembler. Various Win32 compilation fixes and 3855 update to the INSTALL.W32 file with (hopefully) more accurate Win32 3856 build instructions. 3857 [Steve Henson] 3858 3859 *) Modify configure script 'Configure' to automatically create crypto/date.h 3860 file under Win32 and also build pem.h from pem.org. New script 3861 util/mkfiles.pl to create the MINFO file on environments that can't do a 3862 'make files': perl util/mkfiles.pl >MINFO should work. 3863 [Steve Henson] 3864 3865 *) Major rework of DES function declarations, in the pursuit of correctness 3866 and purity. As a result, many evil casts evaporated, and some weirdness, 3867 too. You may find this causes warnings in your code. Zapping your evil 3868 casts will probably fix them. Mostly. 3869 [Ben Laurie] 3870 3871 *) Fix for a typo in asn1.h. Bug fix to object creation script 3872 obj_dat.pl. It considered a zero in an object definition to mean 3873 "end of object": none of the objects in objects.h have any zeros 3874 so it wasn't spotted. 3875 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 3876 3877 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 3878 Masking (CBCM). In the absence of test vectors, the best I have been able 3879 to do is check that the decrypt undoes the encrypt, so far. Send me test 3880 vectors if you have them. 3881 [Ben Laurie] 3882 3883 *) Correct calculation of key length for export ciphers (too much space was 3884 allocated for null ciphers). This has not been tested! 3885 [Ben Laurie] 3886 3887 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 3888 message is now correct (it understands "crypto" and "ssl" on its 3889 command line). There is also now an "update" option. This will update 3890 the util/ssleay.num and util/libeay.num files with any new functions. 3891 If you do a: 3892 perl util/mkdef.pl crypto ssl update 3893 it will update them. 3894 [Steve Henson] 3895 3896 *) Overhauled the Perl interface (perl/*): 3897 - ported BN stuff to OpenSSL's different BN library 3898 - made the perl/ source tree CVS-aware 3899 - renamed the package from SSLeay to OpenSSL (the files still contain 3900 their history because I've copied them in the repository) 3901 - removed obsolete files (the test scripts will be replaced 3902 by better Test::Harness variants in the future) 3903 [Ralf S. Engelschall] 3904 3905 *) First cut for a very conservative source tree cleanup: 3906 1. merge various obsolete readme texts into doc/ssleay.txt 3907 where we collect the old documents and readme texts. 3908 2. remove the first part of files where I'm already sure that we no 3909 longer need them because of three reasons: either they are just temporary 3910 files which were left by Eric or they are preserved original files where 3911 I've verified that the diff is also available in the CVS via "cvs diff 3912 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 3913 the crypto/md/ stuff). 3914 [Ralf S. Engelschall] 3915 3916 *) More extension code. Incomplete support for subject and issuer alt 3917 name, issuer and authority key id. Change the i2v function parameters 3918 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 3919 what that's for :-) Fix to ASN1 macro which messed up 3920 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 3921 [Steve Henson] 3922 3923 *) Preliminary support for ENUMERATED type. This is largely copied from the 3924 INTEGER code. 3925 [Steve Henson] 3926 3927 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 3928 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3929 3930 *) Make sure `make rehash' target really finds the `openssl' program. 3931 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 3932 3933 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 3934 like to hear about it if this slows down other processors. 3935 [Ben Laurie] 3936 3937 *) Add CygWin32 platform information to Configure script. 3938 [Alan Batie <batie@aahz.jf.intel.com>] 3939 3940 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 3941 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 3942 3943 *) New program nseq to manipulate netscape certificate sequences 3944 [Steve Henson] 3945 3946 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 3947 few typos. 3948 [Steve Henson] 3949 3950 *) Fixes to BN code. Previously the default was to define BN_RECURSION 3951 but the BN code had some problems that would cause failures when 3952 doing certificate verification and some other functions. 3953 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 3954 3955 *) Add ASN1 and PEM code to support netscape certificate sequences. 3956 [Steve Henson] 3957 3958 *) Add ASN1 and PEM code to support netscape certificate sequences. 3959 [Steve Henson] 3960 3961 *) Add several PKIX and private extended key usage OIDs. 3962 [Steve Henson] 3963 3964 *) Modify the 'ca' program to handle the new extension code. Modify 3965 openssl.cnf for new extension format, add comments. 3966 [Steve Henson] 3967 3968 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 3969 and add a sample to openssl.cnf so req -x509 now adds appropriate 3970 CA extensions. 3971 [Steve Henson] 3972 3973 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 3974 error code, add initial support to X509_print() and x509 application. 3975 [Steve Henson] 3976 3977 *) Takes a deep breath and start addding X509 V3 extension support code. Add 3978 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 3979 stuff is currently isolated and isn't even compiled yet. 3980 [Steve Henson] 3981 3982 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 3983 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 3984 Removed the versions check from X509 routines when loading extensions: 3985 this allows certain broken certificates that don't set the version 3986 properly to be processed. 3987 [Steve Henson] 3988 3989 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 3990 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 3991 can still be regenerated with "make depend". 3992 [Ben Laurie] 3993 3994 *) Spelling mistake in C version of CAST-128. 3995 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 3996 3997 *) Changes to the error generation code. The perl script err-code.pl 3998 now reads in the old error codes and retains the old numbers, only 3999 adding new ones if necessary. It also only changes the .err files if new 4000 codes are added. The makefiles have been modified to only insert errors 4001 when needed (to avoid needlessly modifying header files). This is done 4002 by only inserting errors if the .err file is newer than the auto generated 4003 C file. To rebuild all the error codes from scratch (the old behaviour) 4004 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 4005 or delete all the .err files. 4006 [Steve Henson] 4007 4008 *) CAST-128 was incorrectly implemented for short keys. The C version has 4009 been fixed, but is untested. The assembler versions are also fixed, but 4010 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 4011 to regenerate it if needed. 4012 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 4013 Hagino <itojun@kame.net>] 4014 4015 *) File was opened incorrectly in randfile.c. 4016 [Ulf M�ller <ulf@fitug.de>] 4017 4018 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 4019 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 4020 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 4021 al: it's just almost always a UTCTime. Note this patch adds new error 4022 codes so do a "make errors" if there are problems. 4023 [Steve Henson] 4024 4025 *) Correct Linux 1 recognition in config. 4026 [Ulf M�ller <ulf@fitug.de>] 4027 4028 *) Remove pointless MD5 hash when using DSA keys in ca. 4029 [Anonymous <nobody@replay.com>] 4030 4031 *) Generate an error if given an empty string as a cert directory. Also 4032 generate an error if handed NULL (previously returned 0 to indicate an 4033 error, but didn't set one). 4034 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 4035 4036 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 4037 [Ben Laurie] 4038 4039 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 4040 parameters. This was causing a warning which killed off the Win32 compile. 4041 [Steve Henson] 4042 4043 *) Remove C++ style comments from crypto/bn/bn_local.h. 4044 [Neil Costigan <neil.costigan@celocom.com>] 4045 4046 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 4047 based on a text string, looking up short and long names and finally 4048 "dot" format. The "dot" format stuff didn't work. Added new function 4049 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 4050 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 4051 OID is not part of the table. 4052 [Steve Henson] 4053 4054 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 4055 X509_LOOKUP_by_alias(). 4056 [Ben Laurie] 4057 4058 *) Sort openssl functions by name. 4059 [Ben Laurie] 4060 4061 *) Get the gendsa program working (hopefully) and add it to app list. Remove 4062 encryption from sample DSA keys (in case anyone is interested the password 4063 was "1234"). 4064 [Steve Henson] 4065 4066 *) Make _all_ *_free functions accept a NULL pointer. 4067 [Frans Heymans <fheymans@isaserver.be>] 4068 4069 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 4070 NULL pointers. 4071 [Anonymous <nobody@replay.com>] 4072 4073 *) s_server should send the CAfile as acceptable CAs, not its own cert. 4074 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 4075 4076 *) Don't blow it for numeric -newkey arguments to apps/req. 4077 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 4078 4079 *) Temp key "for export" tests were wrong in s3_srvr.c. 4080 [Anonymous <nobody@replay.com>] 4081 4082 *) Add prototype for temp key callback functions 4083 SSL_CTX_set_tmp_{rsa,dh}_callback(). 4084 [Ben Laurie] 4085 4086 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 4087 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 4088 [Steve Henson] 4089 4090 *) X509_name_add_entry() freed the wrong thing after an error. 4091 [Arne Ansper <arne@ats.cyber.ee>] 4092 4093 *) rsa_eay.c would attempt to free a NULL context. 4094 [Arne Ansper <arne@ats.cyber.ee>] 4095 4096 *) BIO_s_socket() had a broken should_retry() on Windoze. 4097 [Arne Ansper <arne@ats.cyber.ee>] 4098 4099 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 4100 [Arne Ansper <arne@ats.cyber.ee>] 4101 4102 *) Make sure the already existing X509_STORE->depth variable is initialized 4103 in X509_STORE_new(), but document the fact that this variable is still 4104 unused in the certificate verification process. 4105 [Ralf S. Engelschall] 4106 4107 *) Fix the various library and apps files to free up pkeys obtained from 4108 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 4109 [Steve Henson] 4110 4111 *) Fix reference counting in X509_PUBKEY_get(). This makes 4112 demos/maurice/example2.c work, amongst others, probably. 4113 [Steve Henson and Ben Laurie] 4114 4115 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 4116 `openssl' and second, the shortcut symlinks for the `openssl <command>' 4117 are no longer created. This way we have a single and consistent command 4118 line interface `openssl <command>', similar to `cvs <command>'. 4119 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 4120 4121 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 4122 BIT STRING wrapper always have zero unused bits. 4123 [Steve Henson] 4124 4125 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 4126 [Steve Henson] 4127 4128 *) Make the top-level INSTALL documentation easier to understand. 4129 [Paul Sutton] 4130 4131 *) Makefiles updated to exit if an error occurs in a sub-directory 4132 make (including if user presses ^C) [Paul Sutton] 4133 4134 *) Make Montgomery context stuff explicit in RSA data structure. 4135 [Ben Laurie] 4136 4137 *) Fix build order of pem and err to allow for generated pem.h. 4138 [Ben Laurie] 4139 4140 *) Fix renumbering bug in X509_NAME_delete_entry(). 4141 [Ben Laurie] 4142 4143 *) Enhanced the err-ins.pl script so it makes the error library number 4144 global and can add a library name. This is needed for external ASN1 and 4145 other error libraries. 4146 [Steve Henson] 4147 4148 *) Fixed sk_insert which never worked properly. 4149 [Steve Henson] 4150 4151 *) Fix ASN1 macros so they can handle indefinite length construted 4152 EXPLICIT tags. Some non standard certificates use these: they can now 4153 be read in. 4154 [Steve Henson] 4155 4156 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 4157 into a single doc/ssleay.txt bundle. This way the information is still 4158 preserved but no longer messes up this directory. Now it's new room for 4159 the new set of documenation files. 4160 [Ralf S. Engelschall] 4161 4162 *) SETs were incorrectly DER encoded. This was a major pain, because they 4163 shared code with SEQUENCEs, which aren't coded the same. This means that 4164 almost everything to do with SETs or SEQUENCEs has either changed name or 4165 number of arguments. 4166 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 4167 4168 *) Fix test data to work with the above. 4169 [Ben Laurie] 4170 4171 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 4172 was already fixed by Eric for 0.9.1 it seems. 4173 [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 4174 4175 *) Autodetect FreeBSD3. 4176 [Ben Laurie] 4177 4178 *) Fix various bugs in Configure. This affects the following platforms: 4179 nextstep 4180 ncr-scde 4181 unixware-2.0 4182 unixware-2.0-pentium 4183 sco5-cc. 4184 [Ben Laurie] 4185 4186 *) Eliminate generated files from CVS. Reorder tests to regenerate files 4187 before they are needed. 4188 [Ben Laurie] 4189 4190 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 4191 [Ben Laurie] 4192 4193 4194 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 4195 4196 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 4197 changed SSLeay to OpenSSL in version strings. 4198 [Ralf S. Engelschall] 4199 4200 *) Some fixups to the top-level documents. 4201 [Paul Sutton] 4202 4203 *) Fixed the nasty bug where rsaref.h was not found under compile-time 4204 because the symlink to include/ was missing. 4205 [Ralf S. Engelschall] 4206 4207 *) Incorporated the popular no-RSA/DSA-only patches 4208 which allow to compile a RSA-free SSLeay. 4209 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 4210 4211 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 4212 when "ssleay" is still not found. 4213 [Ralf S. Engelschall] 4214 4215 *) Added more platforms to Configure: Cray T3E, HPUX 11, 4216 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 4217 4218 *) Updated the README file. 4219 [Ralf S. Engelschall] 4220 4221 *) Added various .cvsignore files in the CVS repository subdirs 4222 to make a "cvs update" really silent. 4223 [Ralf S. Engelschall] 4224 4225 *) Recompiled the error-definition header files and added 4226 missing symbols to the Win32 linker tables. 4227 [Ralf S. Engelschall] 4228 4229 *) Cleaned up the top-level documents; 4230 o new files: CHANGES and LICENSE 4231 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 4232 o merged COPYRIGHT into LICENSE 4233 o removed obsolete TODO file 4234 o renamed MICROSOFT to INSTALL.W32 4235 [Ralf S. Engelschall] 4236 4237 *) Removed dummy files from the 0.9.1b source tree: 4238 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 4239 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 4240 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 4241 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 4242 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 4243 [Ralf S. Engelschall] 4244 4245 *) Added various platform portability fixes. 4246 [Mark J. Cox] 4247 4248 *) The Genesis of the OpenSSL rpject: 4249 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 4250 Young and Tim J. Hudson created while they were working for C2Net until 4251 summer 1998. 4252 [The OpenSSL Project] 4253 4254 4255 Changes between 0.9.0b and 0.9.1b [not released] 4256 4257 *) Updated a few CA certificates under certs/ 4258 [Eric A. Young] 4259 4260 *) Changed some BIGNUM api stuff. 4261 [Eric A. Young] 4262 4263 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 4264 DGUX x86, Linux Alpha, etc. 4265 [Eric A. Young] 4266 4267 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 4268 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 4269 available). 4270 [Eric A. Young] 4271 4272 *) Add -strparse option to asn1pars program which parses nested 4273 binary structures 4274 [Dr Stephen Henson <shenson@bigfoot.com>] 4275 4276 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 4277 [Eric A. Young] 4278 4279 *) DSA fix for "ca" program. 4280 [Eric A. Young] 4281 4282 *) Added "-genkey" option to "dsaparam" program. 4283 [Eric A. Young] 4284 4285 *) Added RIPE MD160 (rmd160) message digest. 4286 [Eric A. Young] 4287 4288 *) Added -a (all) option to "ssleay version" command. 4289 [Eric A. Young] 4290 4291 *) Added PLATFORM define which is the id given to Configure. 4292 [Eric A. Young] 4293 4294 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 4295 [Eric A. Young] 4296 4297 *) Extended the ASN.1 parser routines. 4298 [Eric A. Young] 4299 4300 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 4301 [Eric A. Young] 4302 4303 *) Added a BN_CTX to the BN library. 4304 [Eric A. Young] 4305 4306 *) Fixed the weak key values in DES library 4307 [Eric A. Young] 4308 4309 *) Changed API in EVP library for cipher aliases. 4310 [Eric A. Young] 4311 4312 *) Added support for RC2/64bit cipher. 4313 [Eric A. Young] 4314 4315 *) Converted the lhash library to the crypto/mem.c functions. 4316 [Eric A. Young] 4317 4318 *) Added more recognized ASN.1 object ids. 4319 [Eric A. Young] 4320 4321 *) Added more RSA padding checks for SSL/TLS. 4322 [Eric A. Young] 4323 4324 *) Added BIO proxy/filter functionality. 4325 [Eric A. Young] 4326 4327 *) Added extra_certs to SSL_CTX which can be used 4328 send extra CA certificates to the client in the CA cert chain sending 4329 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 4330 [Eric A. Young] 4331 4332 *) Now Fortezza is denied in the authentication phase because 4333 this is key exchange mechanism is not supported by SSLeay at all. 4334 [Eric A. Young] 4335 4336 *) Additional PKCS1 checks. 4337 [Eric A. Young] 4338 4339 *) Support the string "TLSv1" for all TLS v1 ciphers. 4340 [Eric A. Young] 4341 4342 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 4343 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 4344 [Eric A. Young] 4345 4346 *) Fixed a few memory leaks. 4347 [Eric A. Young] 4348 4349 *) Fixed various code and comment typos. 4350 [Eric A. Young] 4351 4352 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 4353 bytes sent in the client random. 4354 [Edward Bishop <ebishop@spyglass.com>] 4355 4356