CHANGES revision 277597
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 6 7 *) Build fixes for the Windows and OpenVMS platforms 8 [Matt Caswell and Richard Levitte] 9 10 Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 11 12 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 13 message can cause a segmentation fault in OpenSSL due to a NULL pointer 14 dereference. This could lead to a Denial Of Service attack. Thanks to 15 Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 16 (CVE-2014-3571) 17 [Steve Henson] 18 19 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 20 dtls1_buffer_record function under certain conditions. In particular this 21 could occur if an attacker sent repeated DTLS records with the same 22 sequence number but for the next epoch. The memory leak could be exploited 23 by an attacker in a Denial of Service attack through memory exhaustion. 24 Thanks to Chris Mueller for reporting this issue. 25 (CVE-2015-0206) 26 [Matt Caswell] 27 28 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 29 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 30 method would be set to NULL which could later result in a NULL pointer 31 dereference. Thanks to Frank Schmirler for reporting this issue. 32 (CVE-2014-3569) 33 [Kurt Roeckx] 34 35 *) Abort handshake if server key exchange message is omitted for ephemeral 36 ECDH ciphersuites. 37 38 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 39 reporting this issue. 40 (CVE-2014-3572) 41 [Steve Henson] 42 43 *) Remove non-export ephemeral RSA code on client and server. This code 44 violated the TLS standard by allowing the use of temporary RSA keys in 45 non-export ciphersuites and could be used by a server to effectively 46 downgrade the RSA key length used to a value smaller than the server 47 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 48 INRIA or reporting this issue. 49 (CVE-2015-0204) 50 [Steve Henson] 51 52 *) Fixed issue where DH client certificates are accepted without verification. 53 An OpenSSL server will accept a DH certificate for client authentication 54 without the certificate verify message. This effectively allows a client to 55 authenticate without the use of a private key. This only affects servers 56 which trust a client certificate authority which issues certificates 57 containing DH keys: these are extremely rare and hardly ever encountered. 58 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 59 this issue. 60 (CVE-2015-0205) 61 [Steve Henson] 62 63 *) Ensure that the session ID context of an SSL is updated when its 64 SSL_CTX is updated via SSL_set_SSL_CTX. 65 66 The session ID context is typically set from the parent SSL_CTX, 67 and can vary with the CTX. 68 [Adam Langley] 69 70 *) Fix various certificate fingerprint issues. 71 72 By using non-DER or invalid encodings outside the signed portion of a 73 certificate the fingerprint can be changed without breaking the signature. 74 Although no details of the signed portion of the certificate can be changed 75 this can cause problems with some applications: e.g. those using the 76 certificate fingerprint for blacklists. 77 78 1. Reject signatures with non zero unused bits. 79 80 If the BIT STRING containing the signature has non zero unused bits reject 81 the signature. All current signature algorithms require zero unused bits. 82 83 2. Check certificate algorithm consistency. 84 85 Check the AlgorithmIdentifier inside TBS matches the one in the 86 certificate signature. NB: this will result in signature failure 87 errors for some broken certificates. 88 89 Thanks to Konrad Kraszewski from Google for reporting this issue. 90 91 3. Check DSA/ECDSA signatures use DER. 92 93 Reencode DSA/ECDSA signatures and compare with the original received 94 signature. Return an error if there is a mismatch. 95 96 This will reject various cases including garbage after signature 97 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 98 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 99 (negative or with leading zeroes). 100 101 Further analysis was conducted and fixes were developed by Stephen Henson 102 of the OpenSSL core team. 103 104 (CVE-2014-8275) 105 [Steve Henson] 106 107 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 108 results on some platforms, including x86_64. This bug occurs at random 109 with a very low probability, and is not known to be exploitable in any 110 way, though its exact impact is difficult to determine. Thanks to Pieter 111 Wuille (Blockstream) who reported this issue and also suggested an initial 112 fix. Further analysis was conducted by the OpenSSL development team and 113 Adam Langley of Google. The final fix was developed by Andy Polyakov of 114 the OpenSSL core team. 115 (CVE-2014-3570) 116 [Andy Polyakov] 117 118 *) Do not resume sessions on the server if the negotiated protocol 119 version does not match the session's version. Resuming with a different 120 version, while not strictly forbidden by the RFC, is of questionable 121 sanity and breaks all known clients. 122 [David Benjamin, Emilia K�sper] 123 124 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 125 early CCS messages during renegotiation. (Note that because 126 renegotiation is encrypted, this early CCS was not exploitable.) 127 [Emilia K�sper] 128 129 *) Tighten client-side session ticket handling during renegotiation: 130 ensure that the client only accepts a session ticket if the server sends 131 the extension anew in the ServerHello. Previously, a TLS client would 132 reuse the old extension state and thus accept a session ticket if one was 133 announced in the initial ServerHello. 134 135 Similarly, ensure that the client requires a session ticket if one 136 was advertised in the ServerHello. Previously, a TLS client would 137 ignore a missing NewSessionTicket message. 138 [Emilia K�sper] 139 140 Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 141 142 *) SRTP Memory Leak. 143 144 A flaw in the DTLS SRTP extension parsing code allows an attacker, who 145 sends a carefully crafted handshake message, to cause OpenSSL to fail 146 to free up to 64k of memory causing a memory leak. This could be 147 exploited in a Denial Of Service attack. This issue affects OpenSSL 148 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 149 whether SRTP is used or configured. Implementations of OpenSSL that 150 have been compiled with OPENSSL_NO_SRTP defined are not affected. 151 152 The fix was developed by the OpenSSL team. 153 (CVE-2014-3513) 154 [OpenSSL team] 155 156 *) Session Ticket Memory Leak. 157 158 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 159 integrity of that ticket is first verified. In the event of a session 160 ticket integrity check failing, OpenSSL will fail to free memory 161 causing a memory leak. By sending a large number of invalid session 162 tickets an attacker could exploit this issue in a Denial Of Service 163 attack. 164 (CVE-2014-3567) 165 [Steve Henson] 166 167 *) Build option no-ssl3 is incomplete. 168 169 When OpenSSL is configured with "no-ssl3" as a build option, servers 170 could accept and complete a SSL 3.0 handshake, and clients could be 171 configured to send them. 172 (CVE-2014-3568) 173 [Akamai and the OpenSSL team] 174 175 *) Add support for TLS_FALLBACK_SCSV. 176 Client applications doing fallback retries should call 177 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 178 (CVE-2014-3566) 179 [Adam Langley, Bodo Moeller] 180 181 *) Add additional DigestInfo checks. 182 183 Reencode DigestInto in DER and check against the original when 184 verifying RSA signature: this will reject any improperly encoded 185 DigestInfo structures. 186 187 Note: this is a precautionary measure and no attacks are currently known. 188 189 [Steve Henson] 190 191 Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 192 193 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 194 SRP code can be overrun an internal buffer. Add sanity check that 195 g, A, B < N to SRP code. 196 197 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 198 Group for discovering this issue. 199 (CVE-2014-3512) 200 [Steve Henson] 201 202 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 203 TLS 1.0 instead of higher protocol versions when the ClientHello message 204 is badly fragmented. This allows a man-in-the-middle attacker to force a 205 downgrade to TLS 1.0 even if both the server and the client support a 206 higher protocol version, by modifying the client's TLS records. 207 208 Thanks to David Benjamin and Adam Langley (Google) for discovering and 209 researching this issue. 210 (CVE-2014-3511) 211 [David Benjamin] 212 213 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 214 to a denial of service attack. A malicious server can crash the client 215 with a null pointer dereference (read) by specifying an anonymous (EC)DH 216 ciphersuite and sending carefully crafted handshake messages. 217 218 Thanks to Felix Gr�bert (Google) for discovering and researching this 219 issue. 220 (CVE-2014-3510) 221 [Emilia K�sper] 222 223 *) By sending carefully crafted DTLS packets an attacker could cause openssl 224 to leak memory. This can be exploited through a Denial of Service attack. 225 Thanks to Adam Langley for discovering and researching this issue. 226 (CVE-2014-3507) 227 [Adam Langley] 228 229 *) An attacker can force openssl to consume large amounts of memory whilst 230 processing DTLS handshake messages. This can be exploited through a 231 Denial of Service attack. 232 Thanks to Adam Langley for discovering and researching this issue. 233 (CVE-2014-3506) 234 [Adam Langley] 235 236 *) An attacker can force an error condition which causes openssl to crash 237 whilst processing DTLS packets due to memory being freed twice. This 238 can be exploited through a Denial of Service attack. 239 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 240 this issue. 241 (CVE-2014-3505) 242 [Adam Langley] 243 244 *) If a multithreaded client connects to a malicious server using a resumed 245 session and the server sends an ec point format extension it could write 246 up to 255 bytes to freed memory. 247 248 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 249 issue. 250 (CVE-2014-3509) 251 [Gabor Tyukasz] 252 253 *) A malicious server can crash an OpenSSL client with a null pointer 254 dereference (read) by specifying an SRP ciphersuite even though it was not 255 properly negotiated with the client. This can be exploited through a 256 Denial of Service attack. 257 258 Thanks to Joonas Kuorilehto and Riku Hietam�ki (Codenomicon) for 259 discovering and researching this issue. 260 (CVE-2014-5139) 261 [Steve Henson] 262 263 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 264 X509_name_oneline, X509_name_print_ex et al. to leak some information 265 from the stack. Applications may be affected if they echo pretty printing 266 output to the attacker. 267 268 Thanks to Ivan Fratric (Google) for discovering this issue. 269 (CVE-2014-3508) 270 [Emilia K�sper, and Steve Henson] 271 272 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 273 for corner cases. (Certain input points at infinity could lead to 274 bogus results, with non-infinity inputs mapped to infinity too.) 275 [Bodo Moeller] 276 277 Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 278 279 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 280 handshake can force the use of weak keying material in OpenSSL 281 SSL/TLS clients and servers. 282 283 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 284 researching this issue. (CVE-2014-0224) 285 [KIKUCHI Masashi, Steve Henson] 286 287 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 288 OpenSSL DTLS client the code can be made to recurse eventually crashing 289 in a DoS attack. 290 291 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 292 (CVE-2014-0221) 293 [Imre Rad, Steve Henson] 294 295 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 296 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 297 client or server. This is potentially exploitable to run arbitrary 298 code on a vulnerable client or server. 299 300 Thanks to J�ri Aedla for reporting this issue. (CVE-2014-0195) 301 [J�ri Aedla, Steve Henson] 302 303 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 304 are subject to a denial of service attack. 305 306 Thanks to Felix Gr�bert and Ivan Fratric at Google for discovering 307 this issue. (CVE-2014-3470) 308 [Felix Gr�bert, Ivan Fratric, Steve Henson] 309 310 *) Harmonize version and its documentation. -f flag is used to display 311 compilation flags. 312 [mancha <mancha1@zoho.com>] 313 314 *) Fix eckey_priv_encode so it immediately returns an error upon a failure 315 in i2d_ECPrivateKey. 316 [mancha <mancha1@zoho.com>] 317 318 *) Fix some double frees. These are not thought to be exploitable. 319 [mancha <mancha1@zoho.com>] 320 321 Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 322 323 *) A missing bounds check in the handling of the TLS heartbeat extension 324 can be used to reveal up to 64k of memory to a connected client or 325 server. 326 327 Thanks for Neel Mehta of Google Security for discovering this bug and to 328 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 329 preparing the fix (CVE-2014-0160) 330 [Adam Langley, Bodo Moeller] 331 332 *) Fix for the attack described in the paper "Recovering OpenSSL 333 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 334 by Yuval Yarom and Naomi Benger. Details can be obtained from: 335 http://eprint.iacr.org/2014/140 336 337 Thanks to Yuval Yarom and Naomi Benger for discovering this 338 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 339 [Yuval Yarom and Naomi Benger] 340 341 *) TLS pad extension: draft-agl-tls-padding-03 342 343 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 344 TLS client Hello record length value would otherwise be > 255 and 345 less that 512 pad with a dummy extension containing zeroes so it 346 is at least 512 bytes long. 347 348 [Adam Langley, Steve Henson] 349 350 Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 351 352 *) Fix for TLS record tampering bug. A carefully crafted invalid 353 handshake could crash OpenSSL with a NULL pointer exception. 354 Thanks to Anton Johansson for reporting this issues. 355 (CVE-2013-4353) 356 357 *) Keep original DTLS digest and encryption contexts in retransmission 358 structures so we can use the previous session parameters if they need 359 to be resent. (CVE-2013-6450) 360 [Steve Henson] 361 362 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 363 avoids preferring ECDHE-ECDSA ciphers when the client appears to be 364 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 365 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 366 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 367 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 368 [Rob Stradling, Adam Langley] 369 370 Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 371 372 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 373 supporting platforms or when small records were transferred. 374 [Andy Polyakov, Steve Henson] 375 376 Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 377 378 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 379 380 This addresses the flaw in CBC record processing discovered by 381 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 382 at: http://www.isg.rhul.ac.uk/tls/ 383 384 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 385 Security Group at Royal Holloway, University of London 386 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 387 Emilia K�sper for the initial patch. 388 (CVE-2013-0169) 389 [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 390 391 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 392 ciphersuites which can be exploited in a denial of service attack. 393 Thanks go to and to Adam Langley <agl@chromium.org> for discovering 394 and detecting this bug and to Wolfgang Ettlinger 395 <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 396 (CVE-2012-2686) 397 [Adam Langley] 398 399 *) Return an error when checking OCSP signatures when key is NULL. 400 This fixes a DoS attack. (CVE-2013-0166) 401 [Steve Henson] 402 403 *) Make openssl verify return errors. 404 [Chris Palmer <palmer@google.com> and Ben Laurie] 405 406 *) Call OCSP Stapling callback after ciphersuite has been chosen, so 407 the right response is stapled. Also change SSL_get_certificate() 408 so it returns the certificate actually sent. 409 See http://rt.openssl.org/Ticket/Display.html?id=2836. 410 [Rob Stradling <rob.stradling@comodo.com>] 411 412 *) Fix possible deadlock when decoding public keys. 413 [Steve Henson] 414 415 *) Don't use TLS 1.0 record version number in initial client hello 416 if renegotiating. 417 [Steve Henson] 418 419 Changes between 1.0.1b and 1.0.1c [10 May 2012] 420 421 *) Sanity check record length before skipping explicit IV in TLS 422 1.2, 1.1 and DTLS to fix DoS attack. 423 424 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 425 fuzzing as a service testing platform. 426 (CVE-2012-2333) 427 [Steve Henson] 428 429 *) Initialise tkeylen properly when encrypting CMS messages. 430 Thanks to Solar Designer of Openwall for reporting this issue. 431 [Steve Henson] 432 433 *) In FIPS mode don't try to use composite ciphers as they are not 434 approved. 435 [Steve Henson] 436 437 Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 438 439 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 440 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 441 mean any application compiled against OpenSSL 1.0.0 headers setting 442 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 443 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 444 0x10000000L Any application which was previously compiled against 445 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 446 will need to be recompiled as a result. Letting be results in 447 inability to disable specifically TLS 1.1 and in client context, 448 in unlike event, limit maximum offered version to TLS 1.0 [see below]. 449 [Steve Henson] 450 451 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 452 disable just protocol X, but all protocols above X *if* there are 453 protocols *below* X still enabled. In more practical terms it means 454 that if application wants to disable TLS1.0 in favor of TLS1.1 and 455 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 456 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 457 client side. 458 [Andy Polyakov] 459 460 Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 461 462 *) Check for potentially exploitable overflows in asn1_d2i_read_bio 463 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 464 in CRYPTO_realloc_clean. 465 466 Thanks to Tavis Ormandy, Google Security Team, for discovering this 467 issue and to Adam Langley <agl@chromium.org> for fixing it. 468 (CVE-2012-2110) 469 [Adam Langley (Google), Tavis Ormandy, Google Security Team] 470 471 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 472 [Adam Langley] 473 474 *) Workarounds for some broken servers that "hang" if a client hello 475 record length exceeds 255 bytes. 476 477 1. Do not use record version number > TLS 1.0 in initial client 478 hello: some (but not all) hanging servers will now work. 479 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 480 the number of ciphers sent in the client hello. This should be 481 set to an even number, such as 50, for example by passing: 482 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 483 Most broken servers should now work. 484 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 485 TLS 1.2 client support entirely. 486 [Steve Henson] 487 488 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 489 [Andy Polyakov] 490 491 Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 492 493 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 494 STRING form instead of a DigestInfo. 495 [Steve Henson] 496 497 *) The format used for MDC2 RSA signatures is inconsistent between EVP 498 and the RSA_sign/RSA_verify functions. This was made more apparent when 499 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 500 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 501 the correct format in RSA_verify so both forms transparently work. 502 [Steve Henson] 503 504 *) Some servers which support TLS 1.0 can choke if we initially indicate 505 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 506 encrypted premaster secret. As a workaround use the maximum pemitted 507 client version in client hello, this should keep such servers happy 508 and still work with previous versions of OpenSSL. 509 [Steve Henson] 510 511 *) Add support for TLS/DTLS heartbeats. 512 [Robin Seggelmann <seggelmann@fh-muenster.de>] 513 514 *) Add support for SCTP. 515 [Robin Seggelmann <seggelmann@fh-muenster.de>] 516 517 *) Improved PRNG seeding for VOS. 518 [Paul Green <Paul.Green@stratus.com>] 519 520 *) Extensive assembler packs updates, most notably: 521 522 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 523 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 524 - x86_64: bit-sliced AES implementation; 525 - ARM: NEON support, contemporary platforms optimizations; 526 - s390x: z196 support; 527 - *: GHASH and GF(2^m) multiplication implementations; 528 529 [Andy Polyakov] 530 531 *) Make TLS-SRP code conformant with RFC 5054 API cleanup 532 (removal of unnecessary code) 533 [Peter Sylvester <peter.sylvester@edelweb.fr>] 534 535 *) Add TLS key material exporter from RFC 5705. 536 [Eric Rescorla] 537 538 *) Add DTLS-SRTP negotiation from RFC 5764. 539 [Eric Rescorla] 540 541 *) Add Next Protocol Negotiation, 542 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 543 disabled with a no-npn flag to config or Configure. Code donated 544 by Google. 545 [Adam Langley <agl@google.com> and Ben Laurie] 546 547 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 548 NIST-P256, NIST-P521, with constant-time single point multiplication on 549 typical inputs. Compiler support for the nonstandard type __uint128_t is 550 required to use this (present in gcc 4.4 and later, for 64-bit builds). 551 Code made available under Apache License version 2.0. 552 553 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 554 line to include this in your build of OpenSSL, and run "make depend" (or 555 "make update"). This enables the following EC_METHODs: 556 557 EC_GFp_nistp224_method() 558 EC_GFp_nistp256_method() 559 EC_GFp_nistp521_method() 560 561 EC_GROUP_new_by_curve_name() will automatically use these (while 562 EC_GROUP_new_curve_GFp() currently prefers the more flexible 563 implementations). 564 [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 565 566 *) Use type ossl_ssize_t instad of ssize_t which isn't available on 567 all platforms. Move ssize_t definition from e_os.h to the public 568 header file e_os2.h as it now appears in public header file cms.h 569 [Steve Henson] 570 571 *) New -sigopt option to the ca, req and x509 utilities. Additional 572 signature parameters can be passed using this option and in 573 particular PSS. 574 [Steve Henson] 575 576 *) Add RSA PSS signing function. This will generate and set the 577 appropriate AlgorithmIdentifiers for PSS based on those in the 578 corresponding EVP_MD_CTX structure. No application support yet. 579 [Steve Henson] 580 581 *) Support for companion algorithm specific ASN1 signing routines. 582 New function ASN1_item_sign_ctx() signs a pre-initialised 583 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 584 the appropriate parameters. 585 [Steve Henson] 586 587 *) Add new algorithm specific ASN1 verification initialisation function 588 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 589 handling will be the same no matter what EVP_PKEY_METHOD is used. 590 Add a PSS handler to support verification of PSS signatures: checked 591 against a number of sample certificates. 592 [Steve Henson] 593 594 *) Add signature printing for PSS. Add PSS OIDs. 595 [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 596 597 *) Add algorithm specific signature printing. An individual ASN1 method 598 can now print out signatures instead of the standard hex dump. 599 600 More complex signatures (e.g. PSS) can print out more meaningful 601 information. Include DSA version that prints out the signature 602 parameters r, s. 603 [Steve Henson] 604 605 *) Password based recipient info support for CMS library: implementing 606 RFC3211. 607 [Steve Henson] 608 609 *) Split password based encryption into PBES2 and PBKDF2 functions. This 610 neatly separates the code into cipher and PBE sections and is required 611 for some algorithms that split PBES2 into separate pieces (such as 612 password based CMS). 613 [Steve Henson] 614 615 *) Session-handling fixes: 616 - Fix handling of connections that are resuming with a session ID, 617 but also support Session Tickets. 618 - Fix a bug that suppressed issuing of a new ticket if the client 619 presented a ticket with an expired session. 620 - Try to set the ticket lifetime hint to something reasonable. 621 - Make tickets shorter by excluding irrelevant information. 622 - On the client side, don't ignore renewed tickets. 623 [Adam Langley, Bodo Moeller (Google)] 624 625 *) Fix PSK session representation. 626 [Bodo Moeller] 627 628 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 629 630 This work was sponsored by Intel. 631 [Andy Polyakov] 632 633 *) Add GCM support to TLS library. Some custom code is needed to split 634 the IV between the fixed (from PRF) and explicit (from TLS record) 635 portions. This adds all GCM ciphersuites supported by RFC5288 and 636 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 637 add a special AESGCM string for GCM only. 638 [Steve Henson] 639 640 *) Expand range of ctrls for AES GCM. Permit setting invocation 641 field on decrypt and retrieval of invocation field only on encrypt. 642 [Steve Henson] 643 644 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 645 As required by RFC5289 these ciphersuites cannot be used if for 646 versions of TLS earlier than 1.2. 647 [Steve Henson] 648 649 *) For FIPS capable OpenSSL interpret a NULL default public key method 650 as unset and return the appopriate default but do *not* set the default. 651 This means we can return the appopriate method in applications that 652 swicth between FIPS and non-FIPS modes. 653 [Steve Henson] 654 655 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 656 ENGINE is used then we cannot handle that in the FIPS module so we 657 keep original code iff non-FIPS operations are allowed. 658 [Steve Henson] 659 660 *) Add -attime option to openssl utilities. 661 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 662 663 *) Redirect DSA and DH operations to FIPS module in FIPS mode. 664 [Steve Henson] 665 666 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 667 FIPS EC methods unconditionally for now. 668 [Steve Henson] 669 670 *) New build option no-ec2m to disable characteristic 2 code. 671 [Steve Henson] 672 673 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 674 all cases can be covered as some introduce binary incompatibilities. 675 [Steve Henson] 676 677 *) Redirect RSA operations to FIPS module including keygen, 678 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 679 [Steve Henson] 680 681 *) Add similar low level API blocking to ciphers. 682 [Steve Henson] 683 684 *) Low level digest APIs are not approved in FIPS mode: any attempt 685 to use these will cause a fatal error. Applications that *really* want 686 to use them can use the private_* version instead. 687 [Steve Henson] 688 689 *) Redirect cipher operations to FIPS module for FIPS builds. 690 [Steve Henson] 691 692 *) Redirect digest operations to FIPS module for FIPS builds. 693 [Steve Henson] 694 695 *) Update build system to add "fips" flag which will link in fipscanister.o 696 for static and shared library builds embedding a signature if needed. 697 [Steve Henson] 698 699 *) Output TLS supported curves in preference order instead of numerical 700 order. This is currently hardcoded for the highest order curves first. 701 This should be configurable so applications can judge speed vs strength. 702 [Steve Henson] 703 704 *) Add TLS v1.2 server support for client authentication. 705 [Steve Henson] 706 707 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 708 and enable MD5. 709 [Steve Henson] 710 711 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 712 FIPS modules versions. 713 [Steve Henson] 714 715 *) Add TLS v1.2 client side support for client authentication. Keep cache 716 of handshake records longer as we don't know the hash algorithm to use 717 until after the certificate request message is received. 718 [Steve Henson] 719 720 *) Initial TLS v1.2 client support. Add a default signature algorithms 721 extension including all the algorithms we support. Parse new signature 722 format in client key exchange. Relax some ECC signing restrictions for 723 TLS v1.2 as indicated in RFC5246. 724 [Steve Henson] 725 726 *) Add server support for TLS v1.2 signature algorithms extension. Switch 727 to new signature format when needed using client digest preference. 728 All server ciphersuites should now work correctly in TLS v1.2. No client 729 support yet and no support for client certificates. 730 [Steve Henson] 731 732 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 733 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 734 ciphersuites. At present only RSA key exchange ciphersuites work with 735 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 736 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 737 and version checking. 738 [Steve Henson] 739 740 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 741 with this defined it will not be affected by any changes to ssl internal 742 structures. Add several utility functions to allow openssl application 743 to work with OPENSSL_NO_SSL_INTERN defined. 744 [Steve Henson] 745 746 *) Add SRP support. 747 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 748 749 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 750 [Steve Henson] 751 752 *) Permit abbreviated handshakes when renegotiating using the function 753 SSL_renegotiate_abbreviated(). 754 [Robin Seggelmann <seggelmann@fh-muenster.de>] 755 756 *) Add call to ENGINE_register_all_complete() to 757 ENGINE_load_builtin_engines(), so some implementations get used 758 automatically instead of needing explicit application support. 759 [Steve Henson] 760 761 *) Add support for TLS key exporter as described in RFC5705. 762 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 763 764 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 765 a few changes are required: 766 767 Add SSL_OP_NO_TLSv1_1 flag. 768 Add TLSv1_1 methods. 769 Update version checking logic to handle version 1.1. 770 Add explicit IV handling (ported from DTLS code). 771 Add command line options to s_client/s_server. 772 [Steve Henson] 773 774 Changes between 1.0.0j and 1.0.0k [5 Feb 2013] 775 776 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 777 778 This addresses the flaw in CBC record processing discovered by 779 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 780 at: http://www.isg.rhul.ac.uk/tls/ 781 782 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 783 Security Group at Royal Holloway, University of London 784 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 785 Emilia K�sper for the initial patch. 786 (CVE-2013-0169) 787 [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 788 789 *) Return an error when checking OCSP signatures when key is NULL. 790 This fixes a DoS attack. (CVE-2013-0166) 791 [Steve Henson] 792 793 *) Call OCSP Stapling callback after ciphersuite has been chosen, so 794 the right response is stapled. Also change SSL_get_certificate() 795 so it returns the certificate actually sent. 796 See http://rt.openssl.org/Ticket/Display.html?id=2836. 797 (This is a backport) 798 [Rob Stradling <rob.stradling@comodo.com>] 799 800 *) Fix possible deadlock when decoding public keys. 801 [Steve Henson] 802 803 Changes between 1.0.0i and 1.0.0j [10 May 2012] 804 805 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after 806 OpenSSL 1.0.1.] 807 808 *) Sanity check record length before skipping explicit IV in DTLS 809 to fix DoS attack. 810 811 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 812 fuzzing as a service testing platform. 813 (CVE-2012-2333) 814 [Steve Henson] 815 816 *) Initialise tkeylen properly when encrypting CMS messages. 817 Thanks to Solar Designer of Openwall for reporting this issue. 818 [Steve Henson] 819 820 Changes between 1.0.0h and 1.0.0i [19 Apr 2012] 821 822 *) Check for potentially exploitable overflows in asn1_d2i_read_bio 823 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 824 in CRYPTO_realloc_clean. 825 826 Thanks to Tavis Ormandy, Google Security Team, for discovering this 827 issue and to Adam Langley <agl@chromium.org> for fixing it. 828 (CVE-2012-2110) 829 [Adam Langley (Google), Tavis Ormandy, Google Security Team] 830 831 Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 832 833 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 834 in CMS and PKCS7 code. When RSA decryption fails use a random key for 835 content decryption and always return the same error. Note: this attack 836 needs on average 2^20 messages so it only affects automated senders. The 837 old behaviour can be reenabled in the CMS code by setting the 838 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 839 an MMA defence is not necessary. 840 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 841 this issue. (CVE-2012-0884) 842 [Steve Henson] 843 844 *) Fix CVE-2011-4619: make sure we really are receiving a 845 client hello before rejecting multiple SGC restarts. Thanks to 846 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 847 [Steve Henson] 848 849 Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 850 851 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 852 Thanks to Antonio Martin, Enterprise Secure Access Research and 853 Development, Cisco Systems, Inc. for discovering this bug and 854 preparing a fix. (CVE-2012-0050) 855 [Antonio Martin] 856 857 Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 858 859 *) Nadhem Alfardan and Kenny Paterson have discovered an extension 860 of the Vaudenay padding oracle attack on CBC mode encryption 861 which enables an efficient plaintext recovery attack against 862 the OpenSSL implementation of DTLS. Their attack exploits timing 863 differences arising during decryption processing. A research 864 paper describing this attack can be found at: 865 http://www.isg.rhul.ac.uk/~kp/dtls.pdf 866 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 867 Security Group at Royal Holloway, University of London 868 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 869 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 870 for preparing the fix. (CVE-2011-4108) 871 [Robin Seggelmann, Michael Tuexen] 872 873 *) Clear bytes used for block padding of SSL 3.0 records. 874 (CVE-2011-4576) 875 [Adam Langley (Google)] 876 877 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 878 Kadianakis <desnacked@gmail.com> for discovering this issue and 879 Adam Langley for preparing the fix. (CVE-2011-4619) 880 [Adam Langley (Google)] 881 882 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 883 [Andrey Kulikov <amdeich@gmail.com>] 884 885 *) Prevent malformed RFC3779 data triggering an assertion failure. 886 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 887 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 888 [Rob Austein <sra@hactrn.net>] 889 890 *) Improved PRNG seeding for VOS. 891 [Paul Green <Paul.Green@stratus.com>] 892 893 *) Fix ssl_ciph.c set-up race. 894 [Adam Langley (Google)] 895 896 *) Fix spurious failures in ecdsatest.c. 897 [Emilia K�sper (Google)] 898 899 *) Fix the BIO_f_buffer() implementation (which was mixing different 900 interpretations of the '..._len' fields). 901 [Adam Langley (Google)] 902 903 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 904 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 905 threads won't reuse the same blinding coefficients. 906 907 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 908 lock to call BN_BLINDING_invert_ex, and avoids one use of 909 BN_BLINDING_update for each BN_BLINDING structure (previously, 910 the last update always remained unused). 911 [Emilia K�sper (Google)] 912 913 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 914 [Bob Buckholz (Google)] 915 916 Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 917 918 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 919 by initialising X509_STORE_CTX properly. (CVE-2011-3207) 920 [Kaspar Brand <ossl@velox.ch>] 921 922 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 923 for multi-threaded use of ECDH. (CVE-2011-3210) 924 [Adam Langley (Google)] 925 926 *) Fix x509_name_ex_d2i memory leak on bad inputs. 927 [Bodo Moeller] 928 929 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 930 signature public key algorithm by using OID xref utilities instead. 931 Before this you could only use some ECC ciphersuites with SHA1 only. 932 [Steve Henson] 933 934 *) Add protection against ECDSA timing attacks as mentioned in the paper 935 by Billy Bob Brumley and Nicola Tuveri, see: 936 937 http://eprint.iacr.org/2011/232.pdf 938 939 [Billy Bob Brumley and Nicola Tuveri] 940 941 Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 942 943 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 944 [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 945 946 *) Fix bug in string printing code: if *any* escaping is enabled we must 947 escape the escape character (backslash) or the resulting string is 948 ambiguous. 949 [Steve Henson] 950 951 Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 952 953 *) Disable code workaround for ancient and obsolete Netscape browsers 954 and servers: an attacker can use it in a ciphersuite downgrade attack. 955 Thanks to Martin Rex for discovering this bug. CVE-2010-4180 956 [Steve Henson] 957 958 *) Fixed J-PAKE implementation error, originally discovered by 959 Sebastien Martini, further info and confirmation from Stefan 960 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 961 [Ben Laurie] 962 963 Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 964 965 *) Fix extension code to avoid race conditions which can result in a buffer 966 overrun vulnerability: resumed sessions must not be modified as they can 967 be shared by multiple threads. CVE-2010-3864 968 [Steve Henson] 969 970 *) Fix WIN32 build system to correctly link an ENGINE directory into 971 a DLL. 972 [Steve Henson] 973 974 Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 975 976 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 977 (CVE-2010-1633) 978 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 979 980 Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 981 982 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 983 context. The operation can be customised via the ctrl mechanism in 984 case ENGINEs want to include additional functionality. 985 [Steve Henson] 986 987 *) Tolerate yet another broken PKCS#8 key format: private key value negative. 988 [Steve Henson] 989 990 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 991 output hashes compatible with older versions of OpenSSL. 992 [Willy Weisz <weisz@vcpc.univie.ac.at>] 993 994 *) Fix compression algorithm handling: if resuming a session use the 995 compression algorithm of the resumed session instead of determining 996 it from client hello again. Don't allow server to change algorithm. 997 [Steve Henson] 998 999 *) Add load_crls() function to apps tidying load_certs() too. Add option 1000 to verify utility to allow additional CRLs to be included. 1001 [Steve Henson] 1002 1003 *) Update OCSP request code to permit adding custom headers to the request: 1004 some responders need this. 1005 [Steve Henson] 1006 1007 *) The function EVP_PKEY_sign() returns <=0 on error: check return code 1008 correctly. 1009 [Julia Lawall <julia@diku.dk>] 1010 1011 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 1012 needlessly dereferenced structures, used obsolete functions and 1013 didn't handle all updated verify codes correctly. 1014 [Steve Henson] 1015 1016 *) Disable MD2 in the default configuration. 1017 [Steve Henson] 1018 1019 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 1020 indicate the initial BIO being pushed or popped. This makes it possible 1021 to determine whether the BIO is the one explicitly called or as a result 1022 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 1023 it handles reference counts correctly and doesn't zero out the I/O bio 1024 when it is not being explicitly popped. WARNING: applications which 1025 included workarounds for the old buggy behaviour will need to be modified 1026 or they could free up already freed BIOs. 1027 [Steve Henson] 1028 1029 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 1030 renaming to all platforms (within the 0.9.8 branch, this was 1031 done conditionally on Netware platforms to avoid a name clash). 1032 [Guenter <lists@gknw.net>] 1033 1034 *) Add ECDHE and PSK support to DTLS. 1035 [Michael Tuexen <tuexen@fh-muenster.de>] 1036 1037 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 1038 be used on C++. 1039 [Steve Henson] 1040 1041 *) Add "missing" function EVP_MD_flags() (without this the only way to 1042 retrieve a digest flags is by accessing the structure directly. Update 1043 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 1044 or cipher is registered as in the "from" argument. Print out all 1045 registered digests in the dgst usage message instead of manually 1046 attempting to work them out. 1047 [Steve Henson] 1048 1049 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 1050 this allows the use of compression and extensions. Change default cipher 1051 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 1052 by default unless an application cipher string requests it. 1053 [Steve Henson] 1054 1055 *) Alter match criteria in PKCS12_parse(). It used to try to use local 1056 key ids to find matching certificates and keys but some PKCS#12 files 1057 don't follow the (somewhat unwritten) rules and this strategy fails. 1058 Now just gather all certificates together and the first private key 1059 then look for the first certificate that matches the key. 1060 [Steve Henson] 1061 1062 *) Support use of registered digest and cipher names for dgst and cipher 1063 commands instead of having to add each one as a special case. So now 1064 you can do: 1065 1066 openssl sha256 foo 1067 1068 as well as: 1069 1070 openssl dgst -sha256 foo 1071 1072 and this works for ENGINE based algorithms too. 1073 1074 [Steve Henson] 1075 1076 *) Update Gost ENGINE to support parameter files. 1077 [Victor B. Wagner <vitus@cryptocom.ru>] 1078 1079 *) Support GeneralizedTime in ca utility. 1080 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 1081 1082 *) Enhance the hash format used for certificate directory links. The new 1083 form uses the canonical encoding (meaning equivalent names will work 1084 even if they aren't identical) and uses SHA1 instead of MD5. This form 1085 is incompatible with the older format and as a result c_rehash should 1086 be used to rebuild symbolic links. 1087 [Steve Henson] 1088 1089 *) Make PKCS#8 the default write format for private keys, replacing the 1090 traditional format. This form is standardised, more secure and doesn't 1091 include an implicit MD5 dependency. 1092 [Steve Henson] 1093 1094 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 1095 committed to OpenSSL should pass this lot as a minimum. 1096 [Steve Henson] 1097 1098 *) Add session ticket override functionality for use by EAP-FAST. 1099 [Jouni Malinen <j@w1.fi>] 1100 1101 *) Modify HMAC functions to return a value. Since these can be implemented 1102 in an ENGINE errors can occur. 1103 [Steve Henson] 1104 1105 *) Type-checked OBJ_bsearch_ex. 1106 [Ben Laurie] 1107 1108 *) Type-checked OBJ_bsearch. Also some constification necessitated 1109 by type-checking. Still to come: TXT_DB, bsearch(?), 1110 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 1111 CONF_VALUE. 1112 [Ben Laurie] 1113 1114 *) New function OPENSSL_gmtime_adj() to add a specific number of days and 1115 seconds to a tm structure directly, instead of going through OS 1116 specific date routines. This avoids any issues with OS routines such 1117 as the year 2038 bug. New *_adj() functions for ASN1 time structures 1118 and X509_time_adj_ex() to cover the extended range. The existing 1119 X509_time_adj() is still usable and will no longer have any date issues. 1120 [Steve Henson] 1121 1122 *) Delta CRL support. New use deltas option which will attempt to locate 1123 and search any appropriate delta CRLs available. 1124 1125 This work was sponsored by Google. 1126 [Steve Henson] 1127 1128 *) Support for CRLs partitioned by reason code. Reorganise CRL processing 1129 code and add additional score elements. Validate alternate CRL paths 1130 as part of the CRL checking and indicate a new error "CRL path validation 1131 error" in this case. Applications wanting additional details can use 1132 the verify callback and check the new "parent" field. If this is not 1133 NULL CRL path validation is taking place. Existing applications wont 1134 see this because it requires extended CRL support which is off by 1135 default. 1136 1137 This work was sponsored by Google. 1138 [Steve Henson] 1139 1140 *) Support for freshest CRL extension. 1141 1142 This work was sponsored by Google. 1143 [Steve Henson] 1144 1145 *) Initial indirect CRL support. Currently only supported in the CRLs 1146 passed directly and not via lookup. Process certificate issuer 1147 CRL entry extension and lookup CRL entries by bother issuer name 1148 and serial number. Check and process CRL issuer entry in IDP extension. 1149 1150 This work was sponsored by Google. 1151 [Steve Henson] 1152 1153 *) Add support for distinct certificate and CRL paths. The CRL issuer 1154 certificate is validated separately in this case. Only enabled if 1155 an extended CRL support flag is set: this flag will enable additional 1156 CRL functionality in future. 1157 1158 This work was sponsored by Google. 1159 [Steve Henson] 1160 1161 *) Add support for policy mappings extension. 1162 1163 This work was sponsored by Google. 1164 [Steve Henson] 1165 1166 *) Fixes to pathlength constraint, self issued certificate handling, 1167 policy processing to align with RFC3280 and PKITS tests. 1168 1169 This work was sponsored by Google. 1170 [Steve Henson] 1171 1172 *) Support for name constraints certificate extension. DN, email, DNS 1173 and URI types are currently supported. 1174 1175 This work was sponsored by Google. 1176 [Steve Henson] 1177 1178 *) To cater for systems that provide a pointer-based thread ID rather 1179 than numeric, deprecate the current numeric thread ID mechanism and 1180 replace it with a structure and associated callback type. This 1181 mechanism allows a numeric "hash" to be extracted from a thread ID in 1182 either case, and on platforms where pointers are larger than 'long', 1183 mixing is done to help ensure the numeric 'hash' is usable even if it 1184 can't be guaranteed unique. The default mechanism is to use "&errno" 1185 as a pointer-based thread ID to distinguish between threads. 1186 1187 Applications that want to provide their own thread IDs should now use 1188 CRYPTO_THREADID_set_callback() to register a callback that will call 1189 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1190 1191 Note that ERR_remove_state() is now deprecated, because it is tied 1192 to the assumption that thread IDs are numeric. ERR_remove_state(0) 1193 to free the current thread's error state should be replaced by 1194 ERR_remove_thread_state(NULL). 1195 1196 (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1197 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1198 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1199 application was previously providing a numeric thread callback that 1200 was inappropriate for distinguishing threads, then uniqueness might 1201 have been obtained with &errno that happened immediately in the 1202 intermediate development versions of OpenSSL; this is no longer the 1203 case, the numeric thread callback will now override the automatic use 1204 of &errno.) 1205 [Geoff Thorpe, with help from Bodo Moeller] 1206 1207 *) Initial support for different CRL issuing certificates. This covers a 1208 simple case where the self issued certificates in the chain exist and 1209 the real CRL issuer is higher in the existing chain. 1210 1211 This work was sponsored by Google. 1212 [Steve Henson] 1213 1214 *) Removed effectively defunct crypto/store from the build. 1215 [Ben Laurie] 1216 1217 *) Revamp of STACK to provide stronger type-checking. Still to come: 1218 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1219 ASN1_STRING, CONF_VALUE. 1220 [Ben Laurie] 1221 1222 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1223 RAM on SSL connections. This option can save about 34k per idle SSL. 1224 [Nick Mathewson] 1225 1226 *) Revamp of LHASH to provide stronger type-checking. Still to come: 1227 STACK, TXT_DB, bsearch, qsort. 1228 [Ben Laurie] 1229 1230 *) Initial support for Cryptographic Message Syntax (aka CMS) based 1231 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1232 support for data, signedData, compressedData, digestedData and 1233 encryptedData, envelopedData types included. Scripts to check against 1234 RFC4134 examples draft and interop and consistency checks of many 1235 content types and variants. 1236 [Steve Henson] 1237 1238 *) Add options to enc utility to support use of zlib compression BIO. 1239 [Steve Henson] 1240 1241 *) Extend mk1mf to support importing of options and assembly language 1242 files from Configure script, currently only included in VC-WIN32. 1243 The assembly language rules can now optionally generate the source 1244 files from the associated perl scripts. 1245 [Steve Henson] 1246 1247 *) Implement remaining functionality needed to support GOST ciphersuites. 1248 Interop testing has been performed using CryptoPro implementations. 1249 [Victor B. Wagner <vitus@cryptocom.ru>] 1250 1251 *) s390x assembler pack. 1252 [Andy Polyakov] 1253 1254 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1255 "family." 1256 [Andy Polyakov] 1257 1258 *) Implement Opaque PRF Input TLS extension as specified in 1259 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1260 official specification yet and no extension type assignment by 1261 IANA exists, this extension (for now) will have to be explicitly 1262 enabled when building OpenSSL by providing the extension number 1263 to use. For example, specify an option 1264 1265 -DTLSEXT_TYPE_opaque_prf_input=0x9527 1266 1267 to the "config" or "Configure" script to enable the extension, 1268 assuming extension number 0x9527 (which is a completely arbitrary 1269 and unofficial assignment based on the MD5 hash of the Internet 1270 Draft). Note that by doing so, you potentially lose 1271 interoperability with other TLS implementations since these might 1272 be using the same extension number for other purposes. 1273 1274 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1275 opaque PRF input value to use in the handshake. This will create 1276 an interal copy of the length-'len' string at 'src', and will 1277 return non-zero for success. 1278 1279 To get more control and flexibility, provide a callback function 1280 by using 1281 1282 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1283 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1284 1285 where 1286 1287 int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1288 void *arg; 1289 1290 Callback function 'cb' will be called in handshakes, and is 1291 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1292 Argument 'arg' is for application purposes (the value as given to 1293 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1294 be provided to the callback function). The callback function 1295 has to return non-zero to report success: usually 1 to use opaque 1296 PRF input just if possible, or 2 to enforce use of the opaque PRF 1297 input. In the latter case, the library will abort the handshake 1298 if opaque PRF input is not successfully negotiated. 1299 1300 Arguments 'peerinput' and 'len' given to the callback function 1301 will always be NULL and 0 in the case of a client. A server will 1302 see the client's opaque PRF input through these variables if 1303 available (NULL and 0 otherwise). Note that if the server 1304 provides an opaque PRF input, the length must be the same as the 1305 length of the client's opaque PRF input. 1306 1307 Note that the callback function will only be called when creating 1308 a new session (session resumption can resume whatever was 1309 previously negotiated), and will not be called in SSL 2.0 1310 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1311 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1312 for applications that need to enforce opaque PRF input. 1313 1314 [Bodo Moeller] 1315 1316 *) Update ssl code to support digests other than SHA1+MD5 for handshake 1317 MAC. 1318 1319 [Victor B. Wagner <vitus@cryptocom.ru>] 1320 1321 *) Add RFC4507 support to OpenSSL. This includes the corrections in 1322 RFC4507bis. The encrypted ticket format is an encrypted encoded 1323 SSL_SESSION structure, that way new session features are automatically 1324 supported. 1325 1326 If a client application caches session in an SSL_SESSION structure 1327 support is transparent because tickets are now stored in the encoded 1328 SSL_SESSION. 1329 1330 The SSL_CTX structure automatically generates keys for ticket 1331 protection in servers so again support should be possible 1332 with no application modification. 1333 1334 If a client or server wishes to disable RFC4507 support then the option 1335 SSL_OP_NO_TICKET can be set. 1336 1337 Add a TLS extension debugging callback to allow the contents of any client 1338 or server extensions to be examined. 1339 1340 This work was sponsored by Google. 1341 [Steve Henson] 1342 1343 *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1344 OpenSSL should now compile cleanly on gcc 4.2 1345 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1346 1347 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1348 support including streaming MAC support: this is required for GOST 1349 ciphersuite support. 1350 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1351 1352 *) Add option -stream to use PKCS#7 streaming in smime utility. New 1353 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1354 to output in BER and PEM format. 1355 [Steve Henson] 1356 1357 *) Experimental support for use of HMAC via EVP_PKEY interface. This 1358 allows HMAC to be handled via the EVP_DigestSign*() interface. The 1359 EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1360 ENGINE support for HMAC keys which are unextractable. New -mac and 1361 -macopt options to dgst utility. 1362 [Steve Henson] 1363 1364 *) New option -sigopt to dgst utility. Update dgst to use 1365 EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1366 alternative signing paramaters such as X9.31 or PSS in the dgst 1367 utility. 1368 [Steve Henson] 1369 1370 *) Change ssl_cipher_apply_rule(), the internal function that does 1371 the work each time a ciphersuite string requests enabling 1372 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1373 removing ("!foo+bar") a class of ciphersuites: Now it maintains 1374 the order of disabled ciphersuites such that those ciphersuites 1375 that most recently went from enabled to disabled not only stay 1376 in order with respect to each other, but also have higher priority 1377 than other disabled ciphersuites the next time ciphersuites are 1378 enabled again. 1379 1380 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1381 the same ciphersuites as with "HIGH" alone, but in a specific 1382 order where the PSK ciphersuites come first (since they are the 1383 most recently disabled ciphersuites when "HIGH" is parsed). 1384 1385 Also, change ssl_create_cipher_list() (using this new 1386 funcionality) such that between otherwise identical 1387 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1388 the default order. 1389 [Bodo Moeller] 1390 1391 *) Change ssl_create_cipher_list() so that it automatically 1392 arranges the ciphersuites in reasonable order before starting 1393 to process the rule string. Thus, the definition for "DEFAULT" 1394 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1395 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1396 This makes it much easier to arrive at a reasonable default order 1397 in applications for which anonymous ciphers are OK (meaning 1398 that you can't actually use DEFAULT). 1399 [Bodo Moeller; suggested by Victor Duchovni] 1400 1401 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1402 processing) into multiple integers instead of setting 1403 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1404 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1405 (These masks as well as the individual bit definitions are hidden 1406 away into the non-exported interface ssl/ssl_locl.h, so this 1407 change to the definition of the SSL_CIPHER structure shouldn't 1408 affect applications.) This give us more bits for each of these 1409 categories, so there is no longer a need to coagulate AES128 and 1410 AES256 into a single algorithm bit, and to coagulate Camellia128 1411 and Camellia256 into a single algorithm bit, which has led to all 1412 kinds of kludges. 1413 1414 Thus, among other things, the kludge introduced in 0.9.7m and 1415 0.9.8e for masking out AES256 independently of AES128 or masking 1416 out Camellia256 independently of AES256 is not needed here in 0.9.9. 1417 1418 With the change, we also introduce new ciphersuite aliases that 1419 so far were missing: "AES128", "AES256", "CAMELLIA128", and 1420 "CAMELLIA256". 1421 [Bodo Moeller] 1422 1423 *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1424 Use the leftmost N bytes of the signature input if the input is 1425 larger than the prime q (with N being the size in bytes of q). 1426 [Nils Larsch] 1427 1428 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1429 it yet and it is largely untested. 1430 [Steve Henson] 1431 1432 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1433 [Nils Larsch] 1434 1435 *) Initial incomplete changes to avoid need for function casts in OpenSSL 1436 some compilers (gcc 4.2 and later) reject their use. Safestack is 1437 reimplemented. Update ASN1 to avoid use of legacy functions. 1438 [Steve Henson] 1439 1440 *) Win32/64 targets are linked with Winsock2. 1441 [Andy Polyakov] 1442 1443 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1444 to external functions. This can be used to increase CRL handling 1445 efficiency especially when CRLs are very large by (for example) storing 1446 the CRL revoked certificates in a database. 1447 [Steve Henson] 1448 1449 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1450 new CRLs added to a directory can be used. New command line option 1451 -verify_return_error to s_client and s_server. This causes real errors 1452 to be returned by the verify callback instead of carrying on no matter 1453 what. This reflects the way a "real world" verify callback would behave. 1454 [Steve Henson] 1455 1456 *) GOST engine, supporting several GOST algorithms and public key formats. 1457 Kindly donated by Cryptocom. 1458 [Cryptocom] 1459 1460 *) Partial support for Issuing Distribution Point CRL extension. CRLs 1461 partitioned by DP are handled but no indirect CRL or reason partitioning 1462 (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1463 selected via a scoring technique which handles IDP and AKID in CRLs. 1464 [Steve Henson] 1465 1466 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1467 will ultimately be used for all verify operations: this will remove the 1468 X509_STORE dependency on certificate verification and allow alternative 1469 lookup methods. X509_STORE based implementations of these two callbacks. 1470 [Steve Henson] 1471 1472 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1473 Modify get_crl() to find a valid (unexpired) CRL if possible. 1474 [Steve Henson] 1475 1476 *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1477 this would be called X509_CRL_cmp() but that name is already used by 1478 a function that just compares CRL issuer names. Cache several CRL 1479 extensions in X509_CRL structure and cache CRLDP in X509. 1480 [Steve Henson] 1481 1482 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1483 this maps equivalent X509_NAME structures into a consistent structure. 1484 Name comparison can then be performed rapidly using memcmp(). 1485 [Steve Henson] 1486 1487 *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1488 utility. 1489 [Steve Henson] 1490 1491 *) Allow digests to supply their own micalg string for S/MIME type using 1492 the ctrl EVP_MD_CTRL_MICALG. 1493 [Steve Henson] 1494 1495 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1496 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1497 ctrl. It can then customise the structure before and/or after signing 1498 if necessary. 1499 [Steve Henson] 1500 1501 *) New function OBJ_add_sigid() to allow application defined signature OIDs 1502 to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1503 to free up any added signature OIDs. 1504 [Steve Henson] 1505 1506 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1507 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1508 digest and cipher tables. New options added to openssl utility: 1509 list-message-digest-algorithms and list-cipher-algorithms. 1510 [Steve Henson] 1511 1512 *) Change the array representation of binary polynomials: the list 1513 of degrees of non-zero coefficients is now terminated with -1. 1514 Previously it was terminated with 0, which was also part of the 1515 value; thus, the array representation was not applicable to 1516 polynomials where t^0 has coefficient zero. This change makes 1517 the array representation useful in a more general context. 1518 [Douglas Stebila] 1519 1520 *) Various modifications and fixes to SSL/TLS cipher string 1521 handling. For ECC, the code now distinguishes between fixed ECDH 1522 with RSA certificates on the one hand and with ECDSA certificates 1523 on the other hand, since these are separate ciphersuites. The 1524 unused code for Fortezza ciphersuites has been removed. 1525 1526 For consistency with EDH, ephemeral ECDH is now called "EECDH" 1527 (not "ECDHE"). For consistency with the code for DH 1528 certificates, use of ECDH certificates is now considered ECDH 1529 authentication, not RSA or ECDSA authentication (the latter is 1530 merely the CA's signing algorithm and not actively used in the 1531 protocol). 1532 1533 The temporary ciphersuite alias "ECCdraft" is no longer 1534 available, and ECC ciphersuites are no longer excluded from "ALL" 1535 and "DEFAULT". The following aliases now exist for RFC 4492 1536 ciphersuites, most of these by analogy with the DH case: 1537 1538 kECDHr - ECDH cert, signed with RSA 1539 kECDHe - ECDH cert, signed with ECDSA 1540 kECDH - ECDH cert (signed with either RSA or ECDSA) 1541 kEECDH - ephemeral ECDH 1542 ECDH - ECDH cert or ephemeral ECDH 1543 1544 aECDH - ECDH cert 1545 aECDSA - ECDSA cert 1546 ECDSA - ECDSA cert 1547 1548 AECDH - anonymous ECDH 1549 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1550 1551 [Bodo Moeller] 1552 1553 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1554 Use correct micalg parameters depending on digest(s) in signed message. 1555 [Steve Henson] 1556 1557 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1558 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1559 [Steve Henson] 1560 1561 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1562 an engine to register a method. Add ENGINE lookups for methods and 1563 functional reference processing. 1564 [Steve Henson] 1565 1566 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1567 EVP_{Sign,Verify}* which allow an application to customise the signature 1568 process. 1569 [Steve Henson] 1570 1571 *) New -resign option to smime utility. This adds one or more signers 1572 to an existing PKCS#7 signedData structure. Also -md option to use an 1573 alternative message digest algorithm for signing. 1574 [Steve Henson] 1575 1576 *) Tidy up PKCS#7 routines and add new functions to make it easier to 1577 create PKCS7 structures containing multiple signers. Update smime 1578 application to support multiple signers. 1579 [Steve Henson] 1580 1581 *) New -macalg option to pkcs12 utility to allow setting of an alternative 1582 digest MAC. 1583 [Steve Henson] 1584 1585 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1586 Reorganize PBE internals to lookup from a static table using NIDs, 1587 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1588 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1589 PRF which will be automatically used with PBES2. 1590 [Steve Henson] 1591 1592 *) Replace the algorithm specific calls to generate keys in "req" with the 1593 new API. 1594 [Steve Henson] 1595 1596 *) Update PKCS#7 enveloped data routines to use new API. This is now 1597 supported by any public key method supporting the encrypt operation. A 1598 ctrl is added to allow the public key algorithm to examine or modify 1599 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1600 a no op. 1601 [Steve Henson] 1602 1603 *) Add a ctrl to asn1 method to allow a public key algorithm to express 1604 a default digest type to use. In most cases this will be SHA1 but some 1605 algorithms (such as GOST) need to specify an alternative digest. The 1606 return value indicates how strong the prefernce is 1 means optional and 1607 2 is mandatory (that is it is the only supported type). Modify 1608 ASN1_item_sign() to accept a NULL digest argument to indicate it should 1609 use the default md. Update openssl utilities to use the default digest 1610 type for signing if it is not explicitly indicated. 1611 [Steve Henson] 1612 1613 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1614 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1615 signing method from the key type. This effectively removes the link 1616 between digests and public key types. 1617 [Steve Henson] 1618 1619 *) Add an OID cross reference table and utility functions. Its purpose is to 1620 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1621 rsaEncryption. This will allow some of the algorithm specific hackery 1622 needed to use the correct OID to be removed. 1623 [Steve Henson] 1624 1625 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1626 structures for PKCS7_sign(). They are now set up by the relevant public 1627 key ASN1 method. 1628 [Steve Henson] 1629 1630 *) Add provisional EC pkey method with support for ECDSA and ECDH. 1631 [Steve Henson] 1632 1633 *) Add support for key derivation (agreement) in the API, DH method and 1634 pkeyutl. 1635 [Steve Henson] 1636 1637 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1638 public and private key formats. As a side effect these add additional 1639 command line functionality not previously available: DSA signatures can be 1640 generated and verified using pkeyutl and DH key support and generation in 1641 pkey, genpkey. 1642 [Steve Henson] 1643 1644 *) BeOS support. 1645 [Oliver Tappe <zooey@hirschkaefer.de>] 1646 1647 *) New make target "install_html_docs" installs HTML renditions of the 1648 manual pages. 1649 [Oliver Tappe <zooey@hirschkaefer.de>] 1650 1651 *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1652 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1653 support key and parameter generation and add initial key generation 1654 functionality for RSA. 1655 [Steve Henson] 1656 1657 *) Add functions for main EVP_PKEY_method operations. The undocumented 1658 functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1659 EVP_PKEY_{encrypt,decrypt}_old. 1660 [Steve Henson] 1661 1662 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1663 key API, doesn't do much yet. 1664 [Steve Henson] 1665 1666 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1667 public key algorithms. New option to openssl utility: 1668 "list-public-key-algorithms" to print out info. 1669 [Steve Henson] 1670 1671 *) Implement the Supported Elliptic Curves Extension for 1672 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1673 [Douglas Stebila] 1674 1675 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1676 EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1677 [Steve Henson] 1678 1679 *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1680 utilities such as rsa, dsa, dsaparam etc except they process any key 1681 type. 1682 [Steve Henson] 1683 1684 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1685 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1686 EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1687 structure. 1688 [Steve Henson] 1689 1690 *) Initial support for pluggable public key ASN1. 1691 De-spaghettify the public key ASN1 handling. Move public and private 1692 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1693 algorithm specific handling to a single module within the relevant 1694 algorithm directory. Add functions to allow (near) opaque processing 1695 of public and private key structures. 1696 [Steve Henson] 1697 1698 *) Implement the Supported Point Formats Extension for 1699 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1700 [Douglas Stebila] 1701 1702 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1703 for the psk identity [hint] and the psk callback functions to the 1704 SSL_SESSION, SSL and SSL_CTX structure. 1705 1706 New ciphersuites: 1707 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1708 PSK-AES256-CBC-SHA 1709 1710 New functions: 1711 SSL_CTX_use_psk_identity_hint 1712 SSL_get_psk_identity_hint 1713 SSL_get_psk_identity 1714 SSL_use_psk_identity_hint 1715 1716 [Mika Kousa and Pasi Eronen of Nokia Corporation] 1717 1718 *) Add RFC 3161 compliant time stamp request creation, response generation 1719 and response verification functionality. 1720 [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1721 1722 *) Add initial support for TLS extensions, specifically for the server_name 1723 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1724 have new members for a host name. The SSL data structure has an 1725 additional member SSL_CTX *initial_ctx so that new sessions can be 1726 stored in that context to allow for session resumption, even after the 1727 SSL has been switched to a new SSL_CTX in reaction to a client's 1728 server_name extension. 1729 1730 New functions (subject to change): 1731 1732 SSL_get_servername() 1733 SSL_get_servername_type() 1734 SSL_set_SSL_CTX() 1735 1736 New CTRL codes and macros (subject to change): 1737 1738 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1739 - SSL_CTX_set_tlsext_servername_callback() 1740 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1741 - SSL_CTX_set_tlsext_servername_arg() 1742 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1743 1744 openssl s_client has a new '-servername ...' option. 1745 1746 openssl s_server has new options '-servername_host ...', '-cert2 ...', 1747 '-key2 ...', '-servername_fatal' (subject to change). This allows 1748 testing the HostName extension for a specific single host name ('-cert' 1749 and '-key' remain fallbacks for handshakes without HostName 1750 negotiation). If the unrecogninzed_name alert has to be sent, this by 1751 default is a warning; it becomes fatal with the '-servername_fatal' 1752 option. 1753 1754 [Peter Sylvester, Remy Allais, Christophe Renou] 1755 1756 *) Whirlpool hash implementation is added. 1757 [Andy Polyakov] 1758 1759 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1760 bn(64,32). Because of instruction set limitations it doesn't have 1761 any negative impact on performance. This was done mostly in order 1762 to make it possible to share assembler modules, such as bn_mul_mont 1763 implementations, between 32- and 64-bit builds without hassle. 1764 [Andy Polyakov] 1765 1766 *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1767 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1768 macro. 1769 [Bodo Moeller] 1770 1771 *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1772 dedicated Montgomery multiplication procedure, is introduced. 1773 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1774 "64-bit" performance on certain 32-bit targets. 1775 [Andy Polyakov] 1776 1777 *) New option SSL_OP_NO_COMP to disable use of compression selectively 1778 in SSL structures. New SSL ctrl to set maximum send fragment size. 1779 Save memory by seeting the I/O buffer sizes dynamically instead of 1780 using the maximum available value. 1781 [Steve Henson] 1782 1783 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1784 in addition to the text details. 1785 [Bodo Moeller] 1786 1787 *) Very, very preliminary EXPERIMENTAL support for printing of general 1788 ASN1 structures. This currently produces rather ugly output and doesn't 1789 handle several customised structures at all. 1790 [Steve Henson] 1791 1792 *) Integrated support for PVK file format and some related formats such 1793 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1794 these in the 'rsa' and 'dsa' utilities. 1795 [Steve Henson] 1796 1797 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1798 [Steve Henson] 1799 1800 *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1801 place for the (very old) "NETSCAPE" format certificates which are now 1802 handled using new ASN1 code equivalents. 1803 [Steve Henson] 1804 1805 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1806 pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1807 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1808 [Nils Larsch] 1809 1810 *) Modify CRL distribution points extension code to print out previously 1811 unsupported fields. Enhance extension setting code to allow setting of 1812 all fields. 1813 [Steve Henson] 1814 1815 *) Add print and set support for Issuing Distribution Point CRL extension. 1816 [Steve Henson] 1817 1818 *) Change 'Configure' script to enable Camellia by default. 1819 [NTT] 1820 1821 Changes between 0.9.8x and 0.9.8y [5 Feb 2013] 1822 1823 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1824 1825 This addresses the flaw in CBC record processing discovered by 1826 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1827 at: http://www.isg.rhul.ac.uk/tls/ 1828 1829 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1830 Security Group at Royal Holloway, University of London 1831 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1832 Emilia K�sper for the initial patch. 1833 (CVE-2013-0169) 1834 [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1835 1836 *) Return an error when checking OCSP signatures when key is NULL. 1837 This fixes a DoS attack. (CVE-2013-0166) 1838 [Steve Henson] 1839 1840 *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1841 the right response is stapled. Also change SSL_get_certificate() 1842 so it returns the certificate actually sent. 1843 See http://rt.openssl.org/Ticket/Display.html?id=2836. 1844 (This is a backport) 1845 [Rob Stradling <rob.stradling@comodo.com>] 1846 1847 *) Fix possible deadlock when decoding public keys. 1848 [Steve Henson] 1849 1850 Changes between 0.9.8w and 0.9.8x [10 May 2012] 1851 1852 *) Sanity check record length before skipping explicit IV in DTLS 1853 to fix DoS attack. 1854 1855 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1856 fuzzing as a service testing platform. 1857 (CVE-2012-2333) 1858 [Steve Henson] 1859 1860 *) Initialise tkeylen properly when encrypting CMS messages. 1861 Thanks to Solar Designer of Openwall for reporting this issue. 1862 [Steve Henson] 1863 1864 Changes between 0.9.8v and 0.9.8w [23 Apr 2012] 1865 1866 *) The fix for CVE-2012-2110 did not take into account that the 1867 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an 1868 int in OpenSSL 0.9.8, making it still vulnerable. Fix by 1869 rejecting negative len parameter. (CVE-2012-2131) 1870 [Tomas Hoger <thoger@redhat.com>] 1871 1872 Changes between 0.9.8u and 0.9.8v [19 Apr 2012] 1873 1874 *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1875 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1876 in CRYPTO_realloc_clean. 1877 1878 Thanks to Tavis Ormandy, Google Security Team, for discovering this 1879 issue and to Adam Langley <agl@chromium.org> for fixing it. 1880 (CVE-2012-2110) 1881 [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1882 1883 Changes between 0.9.8t and 0.9.8u [12 Mar 2012] 1884 1885 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1886 in CMS and PKCS7 code. When RSA decryption fails use a random key for 1887 content decryption and always return the same error. Note: this attack 1888 needs on average 2^20 messages so it only affects automated senders. The 1889 old behaviour can be reenabled in the CMS code by setting the 1890 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1891 an MMA defence is not necessary. 1892 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1893 this issue. (CVE-2012-0884) 1894 [Steve Henson] 1895 1896 *) Fix CVE-2011-4619: make sure we really are receiving a 1897 client hello before rejecting multiple SGC restarts. Thanks to 1898 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1899 [Steve Henson] 1900 1901 Changes between 0.9.8s and 0.9.8t [18 Jan 2012] 1902 1903 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1904 Thanks to Antonio Martin, Enterprise Secure Access Research and 1905 Development, Cisco Systems, Inc. for discovering this bug and 1906 preparing a fix. (CVE-2012-0050) 1907 [Antonio Martin] 1908 1909 Changes between 0.9.8r and 0.9.8s [4 Jan 2012] 1910 1911 *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1912 of the Vaudenay padding oracle attack on CBC mode encryption 1913 which enables an efficient plaintext recovery attack against 1914 the OpenSSL implementation of DTLS. Their attack exploits timing 1915 differences arising during decryption processing. A research 1916 paper describing this attack can be found at: 1917 http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1918 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1919 Security Group at Royal Holloway, University of London 1920 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1921 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1922 for preparing the fix. (CVE-2011-4108) 1923 [Robin Seggelmann, Michael Tuexen] 1924 1925 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109) 1926 [Ben Laurie, Kasper <ekasper@google.com>] 1927 1928 *) Clear bytes used for block padding of SSL 3.0 records. 1929 (CVE-2011-4576) 1930 [Adam Langley (Google)] 1931 1932 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1933 Kadianakis <desnacked@gmail.com> for discovering this issue and 1934 Adam Langley for preparing the fix. (CVE-2011-4619) 1935 [Adam Langley (Google)] 1936 1937 *) Prevent malformed RFC3779 data triggering an assertion failure. 1938 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1939 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1940 [Rob Austein <sra@hactrn.net>] 1941 1942 *) Fix ssl_ciph.c set-up race. 1943 [Adam Langley (Google)] 1944 1945 *) Fix spurious failures in ecdsatest.c. 1946 [Emilia K�sper (Google)] 1947 1948 *) Fix the BIO_f_buffer() implementation (which was mixing different 1949 interpretations of the '..._len' fields). 1950 [Adam Langley (Google)] 1951 1952 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1953 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1954 threads won't reuse the same blinding coefficients. 1955 1956 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1957 lock to call BN_BLINDING_invert_ex, and avoids one use of 1958 BN_BLINDING_update for each BN_BLINDING structure (previously, 1959 the last update always remained unused). 1960 [Emilia K�sper (Google)] 1961 1962 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1963 for multi-threaded use of ECDH. 1964 [Adam Langley (Google)] 1965 1966 *) Fix x509_name_ex_d2i memory leak on bad inputs. 1967 [Bodo Moeller] 1968 1969 *) Add protection against ECDSA timing attacks as mentioned in the paper 1970 by Billy Bob Brumley and Nicola Tuveri, see: 1971 1972 http://eprint.iacr.org/2011/232.pdf 1973 1974 [Billy Bob Brumley and Nicola Tuveri] 1975 1976 Changes between 0.9.8q and 0.9.8r [8 Feb 2011] 1977 1978 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1979 [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1980 1981 *) Fix bug in string printing code: if *any* escaping is enabled we must 1982 escape the escape character (backslash) or the resulting string is 1983 ambiguous. 1984 [Steve Henson] 1985 1986 Changes between 0.9.8p and 0.9.8q [2 Dec 2010] 1987 1988 *) Disable code workaround for ancient and obsolete Netscape browsers 1989 and servers: an attacker can use it in a ciphersuite downgrade attack. 1990 Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1991 [Steve Henson] 1992 1993 *) Fixed J-PAKE implementation error, originally discovered by 1994 Sebastien Martini, further info and confirmation from Stefan 1995 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1996 [Ben Laurie] 1997 1998 Changes between 0.9.8o and 0.9.8p [16 Nov 2010] 1999 2000 *) Fix extension code to avoid race conditions which can result in a buffer 2001 overrun vulnerability: resumed sessions must not be modified as they can 2002 be shared by multiple threads. CVE-2010-3864 2003 [Steve Henson] 2004 2005 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939 2006 [Steve Henson] 2007 2008 *) Don't reencode certificate when calculating signature: cache and use 2009 the original encoding instead. This makes signature verification of 2010 some broken encodings work correctly. 2011 [Steve Henson] 2012 2013 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT 2014 is also one of the inputs. 2015 [Emilia K�sper <emilia.kasper@esat.kuleuven.be> (Google)] 2016 2017 *) Don't repeatedly append PBE algorithms to table if they already exist. 2018 Sort table on each new add. This effectively makes the table read only 2019 after all algorithms are added and subsequent calls to PKCS12_pbe_add 2020 etc are non-op. 2021 [Steve Henson] 2022 2023 Changes between 0.9.8n and 0.9.8o [01 Jun 2010] 2024 2025 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after 2026 OpenSSL 1.0.0.] 2027 2028 *) Correct a typo in the CMS ASN1 module which can result in invalid memory 2029 access or freeing data twice (CVE-2010-0742) 2030 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>] 2031 2032 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more 2033 common in certificates and some applications which only call 2034 SSL_library_init and not OpenSSL_add_all_algorithms() will fail. 2035 [Steve Henson] 2036 2037 *) VMS fixes: 2038 Reduce copying into .apps and .test in makevms.com 2039 Don't try to use blank CA certificate in CA.com 2040 Allow use of C files from original directories in maketests.com 2041 [Steven M. Schweda" <sms@antinode.info>] 2042 2043 Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 2044 2045 *) When rejecting SSL/TLS records due to an incorrect version number, never 2046 update s->server with a new major version number. As of 2047 - OpenSSL 0.9.8m if 'short' is a 16-bit type, 2048 - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 2049 the previous behavior could result in a read attempt at NULL when 2050 receiving specific incorrect SSL/TLS records once record payload 2051 protection is active. (CVE-2010-0740) 2052 [Bodo Moeller, Adam Langley <agl@chromium.org>] 2053 2054 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 2055 could be crashed if the relevant tables were not present (e.g. chrooted). 2056 [Tomas Hoger <thoger@redhat.com>] 2057 2058 Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 2059 2060 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 2061 [Martin Olsson, Neel Mehta] 2062 2063 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 2064 accommodate for stack sorting, always a write lock!). 2065 [Bodo Moeller] 2066 2067 *) On some versions of WIN32 Heap32Next is very slow. This can cause 2068 excessive delays in the RAND_poll(): over a minute. As a workaround 2069 include a time check in the inner Heap32Next loop too. 2070 [Steve Henson] 2071 2072 *) The code that handled flushing of data in SSL/TLS originally used the 2073 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 2074 the problem outlined in PR#1949. The fix suggested there however can 2075 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 2076 of Apache). So instead simplify the code to flush unconditionally. 2077 This should be fine since flushing with no data to flush is a no op. 2078 [Steve Henson] 2079 2080 *) Handle TLS versions 2.0 and later properly and correctly use the 2081 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 2082 off ancient servers have a habit of sticking around for a while... 2083 [Steve Henson] 2084 2085 *) Modify compression code so it frees up structures without using the 2086 ex_data callbacks. This works around a problem where some applications 2087 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 2088 restarting) then use compression (e.g. SSL with compression) later. 2089 This results in significant per-connection memory leaks and 2090 has caused some security issues including CVE-2008-1678 and 2091 CVE-2009-4355. 2092 [Steve Henson] 2093 2094 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 2095 change when encrypting or decrypting. 2096 [Bodo Moeller] 2097 2098 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 2099 connect and renegotiate with servers which do not support RI. 2100 Until RI is more widely deployed this option is enabled by default. 2101 [Steve Henson] 2102 2103 *) Add "missing" ssl ctrls to clear options and mode. 2104 [Steve Henson] 2105 2106 *) If client attempts to renegotiate and doesn't support RI respond with 2107 a no_renegotiation alert as required by RFC5746. Some renegotiating 2108 TLS clients will continue a connection gracefully when they receive 2109 the alert. Unfortunately OpenSSL mishandled this alert and would hang 2110 waiting for a server hello which it will never receive. Now we treat a 2111 received no_renegotiation alert as a fatal error. This is because 2112 applications requesting a renegotiation might well expect it to succeed 2113 and would have no code in place to handle the server denying it so the 2114 only safe thing to do is to terminate the connection. 2115 [Steve Henson] 2116 2117 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 2118 peer supports secure renegotiation and 0 otherwise. Print out peer 2119 renegotiation support in s_client/s_server. 2120 [Steve Henson] 2121 2122 *) Replace the highly broken and deprecated SPKAC certification method with 2123 the updated NID creation version. This should correctly handle UTF8. 2124 [Steve Henson] 2125 2126 *) Implement RFC5746. Re-enable renegotiation but require the extension 2127 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 2128 turns out to be a bad idea. It has been replaced by 2129 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 2130 SSL_CTX_set_options(). This is really not recommended unless you 2131 know what you are doing. 2132 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 2133 2134 *) Fixes to stateless session resumption handling. Use initial_ctx when 2135 issuing and attempting to decrypt tickets in case it has changed during 2136 servername handling. Use a non-zero length session ID when attempting 2137 stateless session resumption: this makes it possible to determine if 2138 a resumption has occurred immediately after receiving server hello 2139 (several places in OpenSSL subtly assume this) instead of later in 2140 the handshake. 2141 [Steve Henson] 2142 2143 *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2144 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2145 fixes for a few places where the return code is not checked 2146 correctly. 2147 [Julia Lawall <julia@diku.dk>] 2148 2149 *) Add --strict-warnings option to Configure script to include devteam 2150 warnings in other configurations. 2151 [Steve Henson] 2152 2153 *) Add support for --libdir option and LIBDIR variable in makefiles. This 2154 makes it possible to install openssl libraries in locations which 2155 have names other than "lib", for example "/usr/lib64" which some 2156 systems need. 2157 [Steve Henson, based on patch from Jeremy Utley] 2158 2159 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 2160 X690 8.9.12 and can produce some misleading textual output of OIDs. 2161 [Steve Henson, reported by Dan Kaminsky] 2162 2163 *) Delete MD2 from algorithm tables. This follows the recommendation in 2164 several standards that it is not used in new applications due to 2165 several cryptographic weaknesses. For binary compatibility reasons 2166 the MD2 API is still compiled in by default. 2167 [Steve Henson] 2168 2169 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 2170 and restored. 2171 [Steve Henson] 2172 2173 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 2174 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 2175 clash. 2176 [Guenter <lists@gknw.net>] 2177 2178 *) Fix the server certificate chain building code to use X509_verify_cert(), 2179 it used to have an ad-hoc builder which was unable to cope with anything 2180 other than a simple chain. 2181 [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 2182 2183 *) Don't check self signed certificate signatures in X509_verify_cert() 2184 by default (a flag can override this): it just wastes time without 2185 adding any security. As a useful side effect self signed root CAs 2186 with non-FIPS digests are now usable in FIPS mode. 2187 [Steve Henson] 2188 2189 *) In dtls1_process_out_of_seq_message() the check if the current message 2190 is already buffered was missing. For every new message was memory 2191 allocated, allowing an attacker to perform an denial of service attack 2192 with sending out of seq handshake messages until there is no memory 2193 left. Additionally every future messege was buffered, even if the 2194 sequence number made no sense and would be part of another handshake. 2195 So only messages with sequence numbers less than 10 in advance will be 2196 buffered. (CVE-2009-1378) 2197 [Robin Seggelmann, discovered by Daniel Mentz] 2198 2199 *) Records are buffered if they arrive with a future epoch to be 2200 processed after finishing the corresponding handshake. There is 2201 currently no limitation to this buffer allowing an attacker to perform 2202 a DOS attack with sending records with future epochs until there is no 2203 memory left. This patch adds the pqueue_size() function to detemine 2204 the size of a buffer and limits the record buffer to 100 entries. 2205 (CVE-2009-1377) 2206 [Robin Seggelmann, discovered by Daniel Mentz] 2207 2208 *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2209 parent structure is freed. (CVE-2009-1379) 2210 [Daniel Mentz] 2211 2212 *) Handle non-blocking I/O properly in SSL_shutdown() call. 2213 [Darryl Miles <darryl-mailinglists@netbauds.net>] 2214 2215 *) Add 2.5.4.* OIDs 2216 [Ilya O. <vrghost@gmail.com>] 2217 2218 Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2219 2220 *) Disable renegotiation completely - this fixes a severe security 2221 problem (CVE-2009-3555) at the cost of breaking all 2222 renegotiation. Renegotiation can be re-enabled by setting 2223 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2224 run-time. This is really not recommended unless you know what 2225 you're doing. 2226 [Ben Laurie] 2227 2228 Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2229 2230 *) Don't set val to NULL when freeing up structures, it is freed up by 2231 underlying code. If sizeof(void *) > sizeof(long) this can result in 2232 zeroing past the valid field. (CVE-2009-0789) 2233 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2234 2235 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2236 checked correctly. This would allow some invalid signed attributes to 2237 appear to verify correctly. (CVE-2009-0591) 2238 [Ivan Nestlerode <inestlerode@us.ibm.com>] 2239 2240 *) Reject UniversalString and BMPString types with invalid lengths. This 2241 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2242 a legal length. (CVE-2009-0590) 2243 [Steve Henson] 2244 2245 *) Set S/MIME signing as the default purpose rather than setting it 2246 unconditionally. This allows applications to override it at the store 2247 level. 2248 [Steve Henson] 2249 2250 *) Permit restricted recursion of ASN1 strings. This is needed in practice 2251 to handle some structures. 2252 [Steve Henson] 2253 2254 *) Improve efficiency of mem_gets: don't search whole buffer each time 2255 for a '\n' 2256 [Jeremy Shapiro <jnshapir@us.ibm.com>] 2257 2258 *) New -hex option for openssl rand. 2259 [Matthieu Herrb] 2260 2261 *) Print out UTF8String and NumericString when parsing ASN1. 2262 [Steve Henson] 2263 2264 *) Support NumericString type for name components. 2265 [Steve Henson] 2266 2267 *) Allow CC in the environment to override the automatically chosen 2268 compiler. Note that nothing is done to ensure flags work with the 2269 chosen compiler. 2270 [Ben Laurie] 2271 2272 Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2273 2274 *) Properly check EVP_VerifyFinal() and similar return values 2275 (CVE-2008-5077). 2276 [Ben Laurie, Bodo Moeller, Google Security Team] 2277 2278 *) Enable TLS extensions by default. 2279 [Ben Laurie] 2280 2281 *) Allow the CHIL engine to be loaded, whether the application is 2282 multithreaded or not. (This does not release the developer from the 2283 obligation to set up the dynamic locking callbacks.) 2284 [Sander Temme <sander@temme.net>] 2285 2286 *) Use correct exit code if there is an error in dgst command. 2287 [Steve Henson; problem pointed out by Roland Dirlewanger] 2288 2289 *) Tweak Configure so that you need to say "experimental-jpake" to enable 2290 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2291 [Bodo Moeller] 2292 2293 *) Add experimental JPAKE support, including demo authentication in 2294 s_client and s_server. 2295 [Ben Laurie] 2296 2297 *) Set the comparison function in v3_addr_canonize(). 2298 [Rob Austein <sra@hactrn.net>] 2299 2300 *) Add support for XMPP STARTTLS in s_client. 2301 [Philip Paeps <philip@freebsd.org>] 2302 2303 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2304 to ensure that even with this option, only ciphersuites in the 2305 server's preference list will be accepted. (Note that the option 2306 applies only when resuming a session, so the earlier behavior was 2307 just about the algorithm choice for symmetric cryptography.) 2308 [Bodo Moeller] 2309 2310 Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2311 2312 *) Fix NULL pointer dereference if a DTLS server received 2313 ChangeCipherSpec as first record (CVE-2009-1386). 2314 [PR #1679] 2315 2316 *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2317 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2318 [Nagendra Modadugu] 2319 2320 *) The fix in 0.9.8c that supposedly got rid of unsafe 2321 double-checked locking was incomplete for RSA blinding, 2322 addressing just one layer of what turns out to have been 2323 doubly unsafe triple-checked locking. 2324 2325 So now fix this for real by retiring the MONT_HELPER macro 2326 in crypto/rsa/rsa_eay.c. 2327 2328 [Bodo Moeller; problem pointed out by Marius Schilder] 2329 2330 *) Various precautionary measures: 2331 2332 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2333 2334 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2335 (NB: This would require knowledge of the secret session ticket key 2336 to exploit, in which case you'd be SOL either way.) 2337 2338 - Change bn_nist.c so that it will properly handle input BIGNUMs 2339 outside the expected range. 2340 2341 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2342 builds. 2343 2344 [Neel Mehta, Bodo Moeller] 2345 2346 *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2347 the load fails. Useful for distros. 2348 [Ben Laurie and the FreeBSD team] 2349 2350 *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2351 [Steve Henson] 2352 2353 *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2354 [Huang Ying] 2355 2356 *) Expand ENGINE to support engine supplied SSL client certificate functions. 2357 2358 This work was sponsored by Logica. 2359 [Steve Henson] 2360 2361 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2362 keystores. Support for SSL/TLS client authentication too. 2363 Not compiled unless enable-capieng specified to Configure. 2364 2365 This work was sponsored by Logica. 2366 [Steve Henson] 2367 2368 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2369 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2370 attribute creation routines such as certifcate requests and PKCS#12 2371 files. 2372 [Steve Henson] 2373 2374 Changes between 0.9.8g and 0.9.8h [28 May 2008] 2375 2376 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2377 handshake which could lead to a cilent crash as found using the 2378 Codenomicon TLS test suite (CVE-2008-1672) 2379 [Steve Henson, Mark Cox] 2380 2381 *) Fix double free in TLS server name extensions which could lead to 2382 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2383 [Joe Orton] 2384 2385 *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2386 2387 Clear the error queue to ensure that error entries left from 2388 older function calls do not interfere with the correct operation. 2389 [Lutz Jaenicke, Erik de Castro Lopo] 2390 2391 *) Remove root CA certificates of commercial CAs: 2392 2393 The OpenSSL project does not recommend any specific CA and does not 2394 have any policy with respect to including or excluding any CA. 2395 Therefore it does not make any sense to ship an arbitrary selection 2396 of root CA certificates with the OpenSSL software. 2397 [Lutz Jaenicke] 2398 2399 *) RSA OAEP patches to fix two separate invalid memory reads. 2400 The first one involves inputs when 'lzero' is greater than 2401 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2402 before the beginning of from). The second one involves inputs where 2403 the 'db' section contains nothing but zeroes (there is a one-byte 2404 invalid read after the end of 'db'). 2405 [Ivan Nestlerode <inestlerode@us.ibm.com>] 2406 2407 *) Partial backport from 0.9.9-dev: 2408 2409 Introduce bn_mul_mont (dedicated Montgomery multiplication 2410 procedure) as a candidate for BIGNUM assembler implementation. 2411 While 0.9.9-dev uses assembler for various architectures, only 2412 x86_64 is available by default here in the 0.9.8 branch, and 2413 32-bit x86 is available through a compile-time setting. 2414 2415 To try the 32-bit x86 assembler implementation, use Configure 2416 option "enable-montasm" (which exists only for this backport). 2417 2418 As "enable-montasm" for 32-bit x86 disclaims code stability 2419 anyway, in this constellation we activate additional code 2420 backported from 0.9.9-dev for further performance improvements, 2421 namely BN_from_montgomery_word. (To enable this otherwise, 2422 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2423 2424 [Andy Polyakov (backport partially by Bodo Moeller)] 2425 2426 *) Add TLS session ticket callback. This allows an application to set 2427 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2428 values. This is useful for key rollover for example where several key 2429 sets may exist with different names. 2430 [Steve Henson] 2431 2432 *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2433 This was broken until now in 0.9.8 releases, such that the only way 2434 a registered ENGINE could be used (assuming it initialises 2435 successfully on the host) was to explicitly set it as the default 2436 for the relevant algorithms. This is in contradiction with 0.9.7 2437 behaviour and the documentation. With this fix, when an ENGINE is 2438 registered into a given algorithm's table of implementations, the 2439 'uptodate' flag is reset so that auto-discovery will be used next 2440 time a new context for that algorithm attempts to select an 2441 implementation. 2442 [Ian Lister (tweaked by Geoff Thorpe)] 2443 2444 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2445 implemention in the following ways: 2446 2447 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2448 hard coded. 2449 2450 Lack of BER streaming support means one pass streaming processing is 2451 only supported if data is detached: setting the streaming flag is 2452 ignored for embedded content. 2453 2454 CMS support is disabled by default and must be explicitly enabled 2455 with the enable-cms configuration option. 2456 [Steve Henson] 2457 2458 *) Update the GMP engine glue to do direct copies between BIGNUM and 2459 mpz_t when openssl and GMP use the same limb size. Otherwise the 2460 existing "conversion via a text string export" trick is still used. 2461 [Paul Sheer <paulsheer@gmail.com>] 2462 2463 *) Zlib compression BIO. This is a filter BIO which compressed and 2464 uncompresses any data passed through it. 2465 [Steve Henson] 2466 2467 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2468 RFC3394 compatible AES key wrapping. 2469 [Steve Henson] 2470 2471 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2472 sets string data without copying. X509_ALGOR_set0() and 2473 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2474 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2475 from an X509_ATTRIBUTE structure optionally checking it occurs only 2476 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2477 data. 2478 [Steve Henson] 2479 2480 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2481 to get the expected BN_FLG_CONSTTIME behavior. 2482 [Bodo Moeller (Google)] 2483 2484 *) Netware support: 2485 2486 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2487 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2488 - added some more tests to do_tests.pl 2489 - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2490 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2491 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2492 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2493 - various changes to netware.pl to enable gcc-cross builds on Win32 2494 platform 2495 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2496 - various changes to fix missing prototype warnings 2497 - fixed x86nasm.pl to create correct asm files for NASM COFF output 2498 - added AES, WHIRLPOOL and CPUID assembler code to build files 2499 - added missing AES assembler make rules to mk1mf.pl 2500 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2501 [Guenter Knauf <eflash@gmx.net>] 2502 2503 *) Implement certificate status request TLS extension defined in RFC3546. 2504 A client can set the appropriate parameters and receive the encoded 2505 OCSP response via a callback. A server can query the supplied parameters 2506 and set the encoded OCSP response in the callback. Add simplified examples 2507 to s_client and s_server. 2508 [Steve Henson] 2509 2510 Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2511 2512 *) Fix various bugs: 2513 + Binary incompatibility of ssl_ctx_st structure 2514 + DTLS interoperation with non-compliant servers 2515 + Don't call get_session_cb() without proposed session 2516 + Fix ia64 assembler code 2517 [Andy Polyakov, Steve Henson] 2518 2519 Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2520 2521 *) DTLS Handshake overhaul. There were longstanding issues with 2522 OpenSSL DTLS implementation, which were making it impossible for 2523 RFC 4347 compliant client to communicate with OpenSSL server. 2524 Unfortunately just fixing these incompatibilities would "cut off" 2525 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2526 server keeps tolerating non RFC compliant syntax. The opposite is 2527 not true, 0.9.8f client can not communicate with earlier server. 2528 This update even addresses CVE-2007-4995. 2529 [Andy Polyakov] 2530 2531 *) Changes to avoid need for function casts in OpenSSL: some compilers 2532 (gcc 4.2 and later) reject their use. 2533 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2534 Steve Henson] 2535 2536 *) Add RFC4507 support to OpenSSL. This includes the corrections in 2537 RFC4507bis. The encrypted ticket format is an encrypted encoded 2538 SSL_SESSION structure, that way new session features are automatically 2539 supported. 2540 2541 If a client application caches session in an SSL_SESSION structure 2542 support is transparent because tickets are now stored in the encoded 2543 SSL_SESSION. 2544 2545 The SSL_CTX structure automatically generates keys for ticket 2546 protection in servers so again support should be possible 2547 with no application modification. 2548 2549 If a client or server wishes to disable RFC4507 support then the option 2550 SSL_OP_NO_TICKET can be set. 2551 2552 Add a TLS extension debugging callback to allow the contents of any client 2553 or server extensions to be examined. 2554 2555 This work was sponsored by Google. 2556 [Steve Henson] 2557 2558 *) Add initial support for TLS extensions, specifically for the server_name 2559 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2560 have new members for a host name. The SSL data structure has an 2561 additional member SSL_CTX *initial_ctx so that new sessions can be 2562 stored in that context to allow for session resumption, even after the 2563 SSL has been switched to a new SSL_CTX in reaction to a client's 2564 server_name extension. 2565 2566 New functions (subject to change): 2567 2568 SSL_get_servername() 2569 SSL_get_servername_type() 2570 SSL_set_SSL_CTX() 2571 2572 New CTRL codes and macros (subject to change): 2573 2574 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2575 - SSL_CTX_set_tlsext_servername_callback() 2576 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2577 - SSL_CTX_set_tlsext_servername_arg() 2578 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2579 2580 openssl s_client has a new '-servername ...' option. 2581 2582 openssl s_server has new options '-servername_host ...', '-cert2 ...', 2583 '-key2 ...', '-servername_fatal' (subject to change). This allows 2584 testing the HostName extension for a specific single host name ('-cert' 2585 and '-key' remain fallbacks for handshakes without HostName 2586 negotiation). If the unrecogninzed_name alert has to be sent, this by 2587 default is a warning; it becomes fatal with the '-servername_fatal' 2588 option. 2589 2590 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2591 2592 *) Add AES and SSE2 assembly language support to VC++ build. 2593 [Steve Henson] 2594 2595 *) Mitigate attack on final subtraction in Montgomery reduction. 2596 [Andy Polyakov] 2597 2598 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2599 (which previously caused an internal error). 2600 [Bodo Moeller] 2601 2602 *) Squeeze another 10% out of IGE mode when in != out. 2603 [Ben Laurie] 2604 2605 *) AES IGE mode speedup. 2606 [Dean Gaudet (Google)] 2607 2608 *) Add the Korean symmetric 128-bit cipher SEED (see 2609 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2610 add SEED ciphersuites from RFC 4162: 2611 2612 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2613 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2614 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2615 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2616 2617 To minimize changes between patchlevels in the OpenSSL 0.9.8 2618 series, SEED remains excluded from compilation unless OpenSSL 2619 is configured with 'enable-seed'. 2620 [KISA, Bodo Moeller] 2621 2622 *) Mitigate branch prediction attacks, which can be practical if a 2623 single processor is shared, allowing a spy process to extract 2624 information. For detailed background information, see 2625 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2626 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2627 and Necessary Software Countermeasures"). The core of the change 2628 are new versions BN_div_no_branch() and 2629 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2630 respectively, which are slower, but avoid the security-relevant 2631 conditional branches. These are automatically called by BN_div() 2632 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2633 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2634 remove a conditional branch. 2635 2636 BN_FLG_CONSTTIME is the new name for the previous 2637 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2638 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2639 in the exponent causes BN_mod_exp_mont() to use the alternative 2640 implementation in BN_mod_exp_mont_consttime().) The old name 2641 remains as a deprecated alias. 2642 2643 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2644 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2645 constant-time implementations for more than just exponentiation. 2646 Here too the old name is kept as a deprecated alias. 2647 2648 BN_BLINDING_new() will now use BN_dup() for the modulus so that 2649 the BN_BLINDING structure gets an independent copy of the 2650 modulus. This means that the previous "BIGNUM *m" argument to 2651 BN_BLINDING_new() and to BN_BLINDING_create_param() now 2652 essentially becomes "const BIGNUM *m", although we can't actually 2653 change this in the header file before 0.9.9. It allows 2654 RSA_setup_blinding() to use BN_with_flags() on the modulus to 2655 enable BN_FLG_CONSTTIME. 2656 2657 [Matthew D Wood (Intel Corp)] 2658 2659 *) In the SSL/TLS server implementation, be strict about session ID 2660 context matching (which matters if an application uses a single 2661 external cache for different purposes). Previously, 2662 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2663 set. This did ensure strict client verification, but meant that, 2664 with applications using a single external cache for quite 2665 different requirements, clients could circumvent ciphersuite 2666 restrictions for a given session ID context by starting a session 2667 in a different context. 2668 [Bodo Moeller] 2669 2670 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2671 a ciphersuite string such as "DEFAULT:RSA" cannot enable 2672 authentication-only ciphersuites. 2673 [Bodo Moeller] 2674 2675 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2676 not complete and could lead to a possible single byte overflow 2677 (CVE-2007-5135) [Ben Laurie] 2678 2679 Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2680 2681 *) Since AES128 and AES256 (and similarly Camellia128 and 2682 Camellia256) share a single mask bit in the logic of 2683 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2684 kludge to work properly if AES128 is available and AES256 isn't 2685 (or if Camellia128 is available and Camellia256 isn't). 2686 [Victor Duchovni] 2687 2688 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2689 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2690 When a point or a seed is encoded in a BIT STRING, we need to 2691 prevent the removal of trailing zero bits to get the proper DER 2692 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2693 of a NamedBitList, for which trailing 0 bits need to be removed.) 2694 [Bodo Moeller] 2695 2696 *) Have SSL/TLS server implementation tolerate "mismatched" record 2697 protocol version while receiving ClientHello even if the 2698 ClientHello is fragmented. (The server can't insist on the 2699 particular protocol version it has chosen before the ServerHello 2700 message has informed the client about his choice.) 2701 [Bodo Moeller] 2702 2703 *) Add RFC 3779 support. 2704 [Rob Austein for ARIN, Ben Laurie] 2705 2706 *) Load error codes if they are not already present instead of using a 2707 static variable. This allows them to be cleanly unloaded and reloaded. 2708 Improve header file function name parsing. 2709 [Steve Henson] 2710 2711 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2712 or CAPABILITY handshake as required by RFCs. 2713 [Goetz Babin-Ebell] 2714 2715 Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2716 2717 *) Introduce limits to prevent malicious keys being able to 2718 cause a denial of service. (CVE-2006-2940) 2719 [Steve Henson, Bodo Moeller] 2720 2721 *) Fix ASN.1 parsing of certain invalid structures that can result 2722 in a denial of service. (CVE-2006-2937) [Steve Henson] 2723 2724 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2725 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2726 2727 *) Fix SSL client code which could crash if connecting to a 2728 malicious SSLv2 server. (CVE-2006-4343) 2729 [Tavis Ormandy and Will Drewry, Google Security Team] 2730 2731 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2732 match only those. Before that, "AES256-SHA" would be interpreted 2733 as a pattern and match "AES128-SHA" too (since AES128-SHA got 2734 the same strength classification in 0.9.7h) as we currently only 2735 have a single AES bit in the ciphersuite description bitmap. 2736 That change, however, also applied to ciphersuite strings such as 2737 "RC4-MD5" that intentionally matched multiple ciphersuites -- 2738 namely, SSL 2.0 ciphersuites in addition to the more common ones 2739 from SSL 3.0/TLS 1.0. 2740 2741 So we change the selection algorithm again: Naming an explicit 2742 ciphersuite selects this one ciphersuite, and any other similar 2743 ciphersuite (same bitmap) from *other* protocol versions. 2744 Thus, "RC4-MD5" again will properly select both the SSL 2.0 2745 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2746 2747 Since SSL 2.0 does not have any ciphersuites for which the 2748 128/256 bit distinction would be relevant, this works for now. 2749 The proper fix will be to use different bits for AES128 and 2750 AES256, which would have avoided the problems from the beginning; 2751 however, bits are scarce, so we can only do this in a new release 2752 (not just a patchlevel) when we can change the SSL_CIPHER 2753 definition to split the single 'unsigned long mask' bitmap into 2754 multiple values to extend the available space. 2755 2756 [Bodo Moeller] 2757 2758 Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2759 2760 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2761 (CVE-2006-4339) [Ben Laurie and Google Security Team] 2762 2763 *) Add AES IGE and biIGE modes. 2764 [Ben Laurie] 2765 2766 *) Change the Unix randomness entropy gathering to use poll() when 2767 possible instead of select(), since the latter has some 2768 undesirable limitations. 2769 [Darryl Miles via Richard Levitte and Bodo Moeller] 2770 2771 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2772 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2773 cannot be implicitly activated as part of, e.g., the "AES" alias. 2774 However, please upgrade to OpenSSL 0.9.9[-dev] for 2775 non-experimental use of the ECC ciphersuites to get TLS extension 2776 support, which is required for curve and point format negotiation 2777 to avoid potential handshake problems. 2778 [Bodo Moeller] 2779 2780 *) Disable rogue ciphersuites: 2781 2782 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2783 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2784 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2785 2786 The latter two were purportedly from 2787 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2788 appear there. 2789 2790 Also deactivate the remaining ciphersuites from 2791 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2792 unofficial, and the ID has long expired. 2793 [Bodo Moeller] 2794 2795 *) Fix RSA blinding Heisenbug (problems sometimes occured on 2796 dual-core machines) and other potential thread-safety issues. 2797 [Bodo Moeller] 2798 2799 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2800 versions), which is now available for royalty-free use 2801 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2802 Also, add Camellia TLS ciphersuites from RFC 4132. 2803 2804 To minimize changes between patchlevels in the OpenSSL 0.9.8 2805 series, Camellia remains excluded from compilation unless OpenSSL 2806 is configured with 'enable-camellia'. 2807 [NTT] 2808 2809 *) Disable the padding bug check when compression is in use. The padding 2810 bug check assumes the first packet is of even length, this is not 2811 necessarily true if compresssion is enabled and can result in false 2812 positives causing handshake failure. The actual bug test is ancient 2813 code so it is hoped that implementations will either have fixed it by 2814 now or any which still have the bug do not support compression. 2815 [Steve Henson] 2816 2817 Changes between 0.9.8a and 0.9.8b [04 May 2006] 2818 2819 *) When applying a cipher rule check to see if string match is an explicit 2820 cipher suite and only match that one cipher suite if it is. 2821 [Steve Henson] 2822 2823 *) Link in manifests for VC++ if needed. 2824 [Austin Ziegler <halostatue@gmail.com>] 2825 2826 *) Update support for ECC-based TLS ciphersuites according to 2827 draft-ietf-tls-ecc-12.txt with proposed changes (but without 2828 TLS extensions, which are supported starting with the 0.9.9 2829 branch, not in the OpenSSL 0.9.8 branch). 2830 [Douglas Stebila] 2831 2832 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2833 opaque EVP_CIPHER_CTX handling. 2834 [Steve Henson] 2835 2836 *) Fixes and enhancements to zlib compression code. We now only use 2837 "zlib1.dll" and use the default __cdecl calling convention on Win32 2838 to conform with the standards mentioned here: 2839 http://www.zlib.net/DLL_FAQ.txt 2840 Static zlib linking now works on Windows and the new --with-zlib-include 2841 --with-zlib-lib options to Configure can be used to supply the location 2842 of the headers and library. Gracefully handle case where zlib library 2843 can't be loaded. 2844 [Steve Henson] 2845 2846 *) Several fixes and enhancements to the OID generation code. The old code 2847 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2848 handle numbers larger than ULONG_MAX, truncated printing and had a 2849 non standard OBJ_obj2txt() behaviour. 2850 [Steve Henson] 2851 2852 *) Add support for building of engines under engine/ as shared libraries 2853 under VC++ build system. 2854 [Steve Henson] 2855 2856 *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2857 Hopefully, we will not see any false combination of paths any more. 2858 [Richard Levitte] 2859 2860 Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2861 2862 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2863 (part of SSL_OP_ALL). This option used to disable the 2864 countermeasure against man-in-the-middle protocol-version 2865 rollback in the SSL 2.0 server implementation, which is a bad 2866 idea. (CVE-2005-2969) 2867 2868 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2869 for Information Security, National Institute of Advanced Industrial 2870 Science and Technology [AIST], Japan)] 2871 2872 *) Add two function to clear and return the verify parameter flags. 2873 [Steve Henson] 2874 2875 *) Keep cipherlists sorted in the source instead of sorting them at 2876 runtime, thus removing the need for a lock. 2877 [Nils Larsch] 2878 2879 *) Avoid some small subgroup attacks in Diffie-Hellman. 2880 [Nick Mathewson and Ben Laurie] 2881 2882 *) Add functions for well-known primes. 2883 [Nick Mathewson] 2884 2885 *) Extended Windows CE support. 2886 [Satoshi Nakamura and Andy Polyakov] 2887 2888 *) Initialize SSL_METHOD structures at compile time instead of during 2889 runtime, thus removing the need for a lock. 2890 [Steve Henson] 2891 2892 *) Make PKCS7_decrypt() work even if no certificate is supplied by 2893 attempting to decrypt each encrypted key in turn. Add support to 2894 smime utility. 2895 [Steve Henson] 2896 2897 Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2898 2899 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2900 OpenSSL 0.9.8.] 2901 2902 *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2903 [Richard Levitte] 2904 2905 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2906 key into the same file any more. 2907 [Richard Levitte] 2908 2909 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2910 [Andy Polyakov] 2911 2912 *) Add -utf8 command line and config file option to 'ca'. 2913 [Stefan <stf@udoma.org] 2914 2915 *) Removed the macro des_crypt(), as it seems to conflict with some 2916 libraries. Use DES_crypt(). 2917 [Richard Levitte] 2918 2919 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2920 involves renaming the source and generated shared-libs for 2921 both. The engines will accept the corrected or legacy ids 2922 ('ncipher' and '4758_cca' respectively) when binding. NB, 2923 this only applies when building 'shared'. 2924 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2925 2926 *) Add attribute functions to EVP_PKEY structure. Modify 2927 PKCS12_create() to recognize a CSP name attribute and 2928 use it. Make -CSP option work again in pkcs12 utility. 2929 [Steve Henson] 2930 2931 *) Add new functionality to the bn blinding code: 2932 - automatic re-creation of the BN_BLINDING parameters after 2933 a fixed number of uses (currently 32) 2934 - add new function for parameter creation 2935 - introduce flags to control the update behaviour of the 2936 BN_BLINDING parameters 2937 - hide BN_BLINDING structure 2938 Add a second BN_BLINDING slot to the RSA structure to improve 2939 performance when a single RSA object is shared among several 2940 threads. 2941 [Nils Larsch] 2942 2943 *) Add support for DTLS. 2944 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2945 2946 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2947 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2948 [Walter Goulet] 2949 2950 *) Remove buggy and incompletet DH cert support from 2951 ssl/ssl_rsa.c and ssl/s3_both.c 2952 [Nils Larsch] 2953 2954 *) Use SHA-1 instead of MD5 as the default digest algorithm for 2955 the apps/openssl applications. 2956 [Nils Larsch] 2957 2958 *) Compile clean with "-Wall -Wmissing-prototypes 2959 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2960 DEBUG_SAFESTACK must also be set. 2961 [Ben Laurie] 2962 2963 *) Change ./Configure so that certain algorithms can be disabled by default. 2964 The new counterpiece to "no-xxx" is "enable-xxx". 2965 2966 The patented RC5 and MDC2 algorithms will now be disabled unless 2967 "enable-rc5" and "enable-mdc2", respectively, are specified. 2968 2969 (IDEA remains enabled despite being patented. This is because IDEA 2970 is frequently required for interoperability, and there is no license 2971 fee for non-commercial use. As before, "no-idea" can be used to 2972 avoid this algorithm.) 2973 2974 [Bodo Moeller] 2975 2976 *) Add processing of proxy certificates (see RFC 3820). This work was 2977 sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2978 EGEE (Enabling Grids for E-science in Europe). 2979 [Richard Levitte] 2980 2981 *) RC4 performance overhaul on modern architectures/implementations, such 2982 as Intel P4, IA-64 and AMD64. 2983 [Andy Polyakov] 2984 2985 *) New utility extract-section.pl. This can be used specify an alternative 2986 section number in a pod file instead of having to treat each file as 2987 a separate case in Makefile. This can be done by adding two lines to the 2988 pod file: 2989 2990 =for comment openssl_section:XXX 2991 2992 The blank line is mandatory. 2993 2994 [Steve Henson] 2995 2996 *) New arguments -certform, -keyform and -pass for s_client and s_server 2997 to allow alternative format key and certificate files and passphrase 2998 sources. 2999 [Steve Henson] 3000 3001 *) New structure X509_VERIFY_PARAM which combines current verify parameters, 3002 update associated structures and add various utility functions. 3003 3004 Add new policy related verify parameters, include policy checking in 3005 standard verify code. Enhance 'smime' application with extra parameters 3006 to support policy checking and print out. 3007 [Steve Henson] 3008 3009 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 3010 Nehemiah processors. These extensions support AES encryption in hardware 3011 as well as RNG (though RNG support is currently disabled). 3012 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 3013 3014 *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 3015 [Geoff Thorpe] 3016 3017 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 3018 [Andy Polyakov and a number of other people] 3019 3020 *) Improved PowerPC platform support. Most notably BIGNUM assembler 3021 implementation contributed by IBM. 3022 [Suresh Chari, Peter Waltenberg, Andy Polyakov] 3023 3024 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 3025 exponent rather than 'unsigned long'. There is a corresponding change to 3026 the new 'rsa_keygen' element of the RSA_METHOD structure. 3027 [Jelte Jansen, Geoff Thorpe] 3028 3029 *) Functionality for creating the initial serial number file is now 3030 moved from CA.pl to the 'ca' utility with a new option -create_serial. 3031 3032 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 3033 number file to 1, which is bound to cause problems. To avoid 3034 the problems while respecting compatibility between different 0.9.7 3035 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 3036 CA.pl for serial number initialization. With the new release 0.9.8, 3037 we can fix the problem directly in the 'ca' utility.) 3038 [Steve Henson] 3039 3040 *) Reduced header interdepencies by declaring more opaque objects in 3041 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 3042 give fewer recursive includes, which could break lazy source code - so 3043 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 3044 developers should define this symbol when building and using openssl to 3045 ensure they track the recommended behaviour, interfaces, [etc], but 3046 backwards-compatible behaviour prevails when this isn't defined. 3047 [Geoff Thorpe] 3048 3049 *) New function X509_POLICY_NODE_print() which prints out policy nodes. 3050 [Steve Henson] 3051 3052 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 3053 This will generate a random key of the appropriate length based on the 3054 cipher context. The EVP_CIPHER can provide its own random key generation 3055 routine to support keys of a specific form. This is used in the des and 3056 3des routines to generate a key of the correct parity. Update S/MIME 3057 code to use new functions and hence generate correct parity DES keys. 3058 Add EVP_CHECK_DES_KEY #define to return an error if the key is not 3059 valid (weak or incorrect parity). 3060 [Steve Henson] 3061 3062 *) Add a local set of CRLs that can be used by X509_verify_cert() as well 3063 as looking them up. This is useful when the verified structure may contain 3064 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 3065 present unless the new PKCS7_NO_CRL flag is asserted. 3066 [Steve Henson] 3067 3068 *) Extend ASN1 oid configuration module. It now additionally accepts the 3069 syntax: 3070 3071 shortName = some long name, 1.2.3.4 3072 [Steve Henson] 3073 3074 *) Reimplemented the BN_CTX implementation. There is now no more static 3075 limitation on the number of variables it can handle nor the depth of the 3076 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 3077 information can now expand as required, and rather than having a single 3078 static array of bignums, BN_CTX now uses a linked-list of such arrays 3079 allowing it to expand on demand whilst maintaining the usefulness of 3080 BN_CTX's "bundling". 3081 [Geoff Thorpe] 3082 3083 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 3084 to allow all RSA operations to function using a single BN_CTX. 3085 [Geoff Thorpe] 3086 3087 *) Preliminary support for certificate policy evaluation and checking. This 3088 is initially intended to pass the tests outlined in "Conformance Testing 3089 of Relying Party Client Certificate Path Processing Logic" v1.07. 3090 [Steve Henson] 3091 3092 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 3093 remained unused and not that useful. A variety of other little bignum 3094 tweaks and fixes have also been made continuing on from the audit (see 3095 below). 3096 [Geoff Thorpe] 3097 3098 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 3099 associated ASN1, EVP and SSL functions and old ASN1 macros. 3100 [Richard Levitte] 3101 3102 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 3103 and this should never fail. So the return value from the use of 3104 BN_set_word() (which can fail due to needless expansion) is now deprecated; 3105 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 3106 [Geoff Thorpe] 3107 3108 *) BN_CTX_get() should return zero-valued bignums, providing the same 3109 initialised value as BN_new(). 3110 [Geoff Thorpe, suggested by Ulf M�ller] 3111 3112 *) Support for inhibitAnyPolicy certificate extension. 3113 [Steve Henson] 3114 3115 *) An audit of the BIGNUM code is underway, for which debugging code is 3116 enabled when BN_DEBUG is defined. This makes stricter enforcements on what 3117 is considered valid when processing BIGNUMs, and causes execution to 3118 assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 3119 further steps are taken to deliberately pollute unused data in BIGNUM 3120 structures to try and expose faulty code further on. For now, openssl will 3121 (in its default mode of operation) continue to tolerate the inconsistent 3122 forms that it has tolerated in the past, but authors and packagers should 3123 consider trying openssl and their own applications when compiled with 3124 these debugging symbols defined. It will help highlight potential bugs in 3125 their own code, and will improve the test coverage for OpenSSL itself. At 3126 some point, these tighter rules will become openssl's default to improve 3127 maintainability, though the assert()s and other overheads will remain only 3128 in debugging configurations. See bn.h for more details. 3129 [Geoff Thorpe, Nils Larsch, Ulf M�ller] 3130 3131 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 3132 that can only be obtained through BN_CTX_new() (which implicitly 3133 initialises it). The presence of this function only made it possible 3134 to overwrite an existing structure (and cause memory leaks). 3135 [Geoff Thorpe] 3136 3137 *) Because of the callback-based approach for implementing LHASH as a 3138 template type, lh_insert() adds opaque objects to hash-tables and 3139 lh_doall() or lh_doall_arg() are typically used with a destructor callback 3140 to clean up those corresponding objects before destroying the hash table 3141 (and losing the object pointers). So some over-zealous constifications in 3142 LHASH have been relaxed so that lh_insert() does not take (nor store) the 3143 objects as "const" and the lh_doall[_arg] callback wrappers are not 3144 prototyped to have "const" restrictions on the object pointers they are 3145 given (and so aren't required to cast them away any more). 3146 [Geoff Thorpe] 3147 3148 *) The tmdiff.h API was so ugly and minimal that our own timing utility 3149 (speed) prefers to use its own implementation. The two implementations 3150 haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3151 its object type properly exposed (MS_TM) instead of casting to/from "char 3152 *". This may still change yet if someone realises MS_TM and "ms_time_***" 3153 aren't necessarily the greatest nomenclatures - but this is what was used 3154 internally to the implementation so I've used that for now. 3155 [Geoff Thorpe] 3156 3157 *) Ensure that deprecated functions do not get compiled when 3158 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3159 the self-tests were still using deprecated key-generation functions so 3160 these have been updated also. 3161 [Geoff Thorpe] 3162 3163 *) Reorganise PKCS#7 code to separate the digest location functionality 3164 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 3165 New function PKCS7_set_digest() to set the digest type for PKCS#7 3166 digestedData type. Add additional code to correctly generate the 3167 digestedData type and add support for this type in PKCS7 initialization 3168 functions. 3169 [Steve Henson] 3170 3171 *) New function PKCS7_set0_type_other() this initializes a PKCS7 3172 structure of type "other". 3173 [Steve Henson] 3174 3175 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 3176 sure the loop does correctly stop and breaking ("division by zero") 3177 modulus operations are not performed. The (pre-generated) prime 3178 table crypto/bn/bn_prime.h was already correct, but it could not be 3179 re-generated on some platforms because of the "division by zero" 3180 situation in the script. 3181 [Ralf S. Engelschall] 3182 3183 *) Update support for ECC-based TLS ciphersuites according to 3184 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 3185 SHA-1 now is only used for "small" curves (where the 3186 representation of a field element takes up to 24 bytes); for 3187 larger curves, the field element resulting from ECDH is directly 3188 used as premaster secret. 3189 [Douglas Stebila (Sun Microsystems Laboratories)] 3190 3191 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3192 curve secp160r1 to the tests. 3193 [Douglas Stebila (Sun Microsystems Laboratories)] 3194 3195 *) Add the possibility to load symbols globally with DSO. 3196 [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3197 3198 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3199 control of the error stack. 3200 [Richard Levitte] 3201 3202 *) Add support for STORE in ENGINE. 3203 [Richard Levitte] 3204 3205 *) Add the STORE type. The intention is to provide a common interface 3206 to certificate and key stores, be they simple file-based stores, or 3207 HSM-type store, or LDAP stores, or... 3208 NOTE: The code is currently UNTESTED and isn't really used anywhere. 3209 [Richard Levitte] 3210 3211 *) Add a generic structure called OPENSSL_ITEM. This can be used to 3212 pass a list of arguments to any function as well as provide a way 3213 for a function to pass data back to the caller. 3214 [Richard Levitte] 3215 3216 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3217 works like BUF_strdup() but can be used to duplicate a portion of 3218 a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3219 a memory area. 3220 [Richard Levitte] 3221 3222 *) Add the function sk_find_ex() which works like sk_find(), but will 3223 return an index to an element even if an exact match couldn't be 3224 found. The index is guaranteed to point at the element where the 3225 searched-for key would be inserted to preserve sorting order. 3226 [Richard Levitte] 3227 3228 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3229 takes an extra flags argument for optional functionality. Currently, 3230 the following flags are defined: 3231 3232 OBJ_BSEARCH_VALUE_ON_NOMATCH 3233 This one gets OBJ_bsearch_ex() to return a pointer to the first 3234 element where the comparing function returns a negative or zero 3235 number. 3236 3237 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3238 This one gets OBJ_bsearch_ex() to return a pointer to the first 3239 element where the comparing function returns zero. This is useful 3240 if there are more than one element where the comparing function 3241 returns zero. 3242 [Richard Levitte] 3243 3244 *) Make it possible to create self-signed certificates with 'openssl ca' 3245 in such a way that the self-signed certificate becomes part of the 3246 CA database and uses the same mechanisms for serial number generation 3247 as all other certificate signing. The new flag '-selfsign' enables 3248 this functionality. Adapt CA.sh and CA.pl.in. 3249 [Richard Levitte] 3250 3251 *) Add functionality to check the public key of a certificate request 3252 against a given private. This is useful to check that a certificate 3253 request can be signed by that key (self-signing). 3254 [Richard Levitte] 3255 3256 *) Make it possible to have multiple active certificates with the same 3257 subject in the CA index file. This is done only if the keyword 3258 'unique_subject' is set to 'no' in the main CA section (default 3259 if 'CA_default') of the configuration file. The value is saved 3260 with the database itself in a separate index attribute file, 3261 named like the index file with '.attr' appended to the name. 3262 [Richard Levitte] 3263 3264 *) Generate muti valued AVAs using '+' notation in config files for 3265 req and dirName. 3266 [Steve Henson] 3267 3268 *) Support for nameConstraints certificate extension. 3269 [Steve Henson] 3270 3271 *) Support for policyConstraints certificate extension. 3272 [Steve Henson] 3273 3274 *) Support for policyMappings certificate extension. 3275 [Steve Henson] 3276 3277 *) Make sure the default DSA_METHOD implementation only uses its 3278 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3279 and change its own handlers to be NULL so as to remove unnecessary 3280 indirection. This lets alternative implementations fallback to the 3281 default implementation more easily. 3282 [Geoff Thorpe] 3283 3284 *) Support for directoryName in GeneralName related extensions 3285 in config files. 3286 [Steve Henson] 3287 3288 *) Make it possible to link applications using Makefile.shared. 3289 Make that possible even when linking against static libraries! 3290 [Richard Levitte] 3291 3292 *) Support for single pass processing for S/MIME signing. This now 3293 means that S/MIME signing can be done from a pipe, in addition 3294 cleartext signing (multipart/signed type) is effectively streaming 3295 and the signed data does not need to be all held in memory. 3296 3297 This is done with a new flag PKCS7_STREAM. When this flag is set 3298 PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3299 is done after the data is output (and digests calculated) in 3300 SMIME_write_PKCS7(). 3301 [Steve Henson] 3302 3303 *) Add full support for -rpath/-R, both in shared libraries and 3304 applications, at least on the platforms where it's known how 3305 to do it. 3306 [Richard Levitte] 3307 3308 *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3309 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3310 will now compute a table of multiples of the generator that 3311 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3312 faster (notably in the case of a single point multiplication, 3313 scalar * generator). 3314 [Nils Larsch, Bodo Moeller] 3315 3316 *) IPv6 support for certificate extensions. The various extensions 3317 which use the IP:a.b.c.d can now take IPv6 addresses using the 3318 formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3319 correctly. 3320 [Steve Henson] 3321 3322 *) Added an ENGINE that implements RSA by performing private key 3323 exponentiations with the GMP library. The conversions to and from 3324 GMP's mpz_t format aren't optimised nor are any montgomery forms 3325 cached, and on x86 it appears OpenSSL's own performance has caught up. 3326 However there are likely to be other architectures where GMP could 3327 provide a boost. This ENGINE is not built in by default, but it can be 3328 specified at Configure time and should be accompanied by the necessary 3329 linker additions, eg; 3330 ./config -DOPENSSL_USE_GMP -lgmp 3331 [Geoff Thorpe] 3332 3333 *) "openssl engine" will not display ENGINE/DSO load failure errors when 3334 testing availability of engines with "-t" - the old behaviour is 3335 produced by increasing the feature's verbosity with "-tt". 3336 [Geoff Thorpe] 3337 3338 *) ECDSA routines: under certain error conditions uninitialized BN objects 3339 could be freed. Solution: make sure initialization is performed early 3340 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3341 via PR#459) 3342 [Lutz Jaenicke] 3343 3344 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3345 and DH_METHOD (eg. by ENGINE implementations) to override the normal 3346 software implementations. For DSA and DH, parameter generation can 3347 also be overriden by providing the appropriate method callbacks. 3348 [Geoff Thorpe] 3349 3350 *) Change the "progress" mechanism used in key-generation and 3351 primality testing to functions that take a new BN_GENCB pointer in 3352 place of callback/argument pairs. The new API functions have "_ex" 3353 postfixes and the older functions are reimplemented as wrappers for 3354 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3355 declarations of the old functions to help (graceful) attempts to 3356 migrate to the new functions. Also, the new key-generation API 3357 functions operate on a caller-supplied key-structure and return 3358 success/failure rather than returning a key or NULL - this is to 3359 help make "keygen" another member function of RSA_METHOD etc. 3360 3361 Example for using the new callback interface: 3362 3363 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3364 void *my_arg = ...; 3365 BN_GENCB my_cb; 3366 3367 BN_GENCB_set(&my_cb, my_callback, my_arg); 3368 3369 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3370 /* For the meaning of a, b in calls to my_callback(), see the 3371 * documentation of the function that calls the callback. 3372 * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3373 * my_callback should return 1 if it wants BN_is_prime_ex() 3374 * to continue, or 0 to stop. 3375 */ 3376 3377 [Geoff Thorpe] 3378 3379 *) Change the ZLIB compression method to be stateful, and make it 3380 available to TLS with the number defined in 3381 draft-ietf-tls-compression-04.txt. 3382 [Richard Levitte] 3383 3384 *) Add the ASN.1 structures and functions for CertificatePair, which 3385 is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3386 3387 CertificatePair ::= SEQUENCE { 3388 forward [0] Certificate OPTIONAL, 3389 reverse [1] Certificate OPTIONAL, 3390 -- at least one of the pair shall be present -- } 3391 3392 Also implement the PEM functions to read and write certificate 3393 pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3394 3395 This needed to be defined, mostly for the sake of the LDAP 3396 attribute crossCertificatePair, but may prove useful elsewhere as 3397 well. 3398 [Richard Levitte] 3399 3400 *) Make it possible to inhibit symlinking of shared libraries in 3401 Makefile.shared, for Cygwin's sake. 3402 [Richard Levitte] 3403 3404 *) Extend the BIGNUM API by creating a function 3405 void BN_set_negative(BIGNUM *a, int neg); 3406 and a macro that behave like 3407 int BN_is_negative(const BIGNUM *a); 3408 3409 to avoid the need to access 'a->neg' directly in applications. 3410 [Nils Larsch] 3411 3412 *) Implement fast modular reduction for pseudo-Mersenne primes 3413 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3414 EC_GROUP_new_curve_GFp() will now automatically use this 3415 if applicable. 3416 [Nils Larsch <nla@trustcenter.de>] 3417 3418 *) Add new lock type (CRYPTO_LOCK_BN). 3419 [Bodo Moeller] 3420 3421 *) Change the ENGINE framework to automatically load engines 3422 dynamically from specific directories unless they could be 3423 found to already be built in or loaded. Move all the 3424 current engines except for the cryptodev one to a new 3425 directory engines/. 3426 The engines in engines/ are built as shared libraries if 3427 the "shared" options was given to ./Configure or ./config. 3428 Otherwise, they are inserted in libcrypto.a. 3429 /usr/local/ssl/engines is the default directory for dynamic 3430 engines, but that can be overriden at configure time through 3431 the usual use of --prefix and/or --openssldir, and at run 3432 time with the environment variable OPENSSL_ENGINES. 3433 [Geoff Thorpe and Richard Levitte] 3434 3435 *) Add Makefile.shared, a helper makefile to build shared 3436 libraries. Addapt Makefile.org. 3437 [Richard Levitte] 3438 3439 *) Add version info to Win32 DLLs. 3440 [Peter 'Luna' Runestig" <peter@runestig.com>] 3441 3442 *) Add new 'medium level' PKCS#12 API. Certificates and keys 3443 can be added using this API to created arbitrary PKCS#12 3444 files while avoiding the low level API. 3445 3446 New options to PKCS12_create(), key or cert can be NULL and 3447 will then be omitted from the output file. The encryption 3448 algorithm NIDs can be set to -1 for no encryption, the mac 3449 iteration count can be set to 0 to omit the mac. 3450 3451 Enhance pkcs12 utility by making the -nokeys and -nocerts 3452 options work when creating a PKCS#12 file. New option -nomac 3453 to omit the mac, NONE can be set for an encryption algorithm. 3454 New code is modified to use the enhanced PKCS12_create() 3455 instead of the low level API. 3456 [Steve Henson] 3457 3458 *) Extend ASN1 encoder to support indefinite length constructed 3459 encoding. This can output sequences tags and octet strings in 3460 this form. Modify pk7_asn1.c to support indefinite length 3461 encoding. This is experimental and needs additional code to 3462 be useful, such as an ASN1 bio and some enhanced streaming 3463 PKCS#7 code. 3464 3465 Extend template encode functionality so that tagging is passed 3466 down to the template encoder. 3467 [Steve Henson] 3468 3469 *) Let 'openssl req' fail if an argument to '-newkey' is not 3470 recognized instead of using RSA as a default. 3471 [Bodo Moeller] 3472 3473 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3474 As these are not official, they are not included in "ALL"; 3475 the "ECCdraft" ciphersuite group alias can be used to select them. 3476 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3477 3478 *) Add ECDH engine support. 3479 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3480 3481 *) Add ECDH in new directory crypto/ecdh/. 3482 [Douglas Stebila (Sun Microsystems Laboratories)] 3483 3484 *) Let BN_rand_range() abort with an error after 100 iterations 3485 without success (which indicates a broken PRNG). 3486 [Bodo Moeller] 3487 3488 *) Change BN_mod_sqrt() so that it verifies that the input value 3489 is really the square of the return value. (Previously, 3490 BN_mod_sqrt would show GIGO behaviour.) 3491 [Bodo Moeller] 3492 3493 *) Add named elliptic curves over binary fields from X9.62, SECG, 3494 and WAP/WTLS; add OIDs that were still missing. 3495 3496 [Sheueling Chang Shantz and Douglas Stebila 3497 (Sun Microsystems Laboratories)] 3498 3499 *) Extend the EC library for elliptic curves over binary fields 3500 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3501 New EC_METHOD: 3502 3503 EC_GF2m_simple_method 3504 3505 New API functions: 3506 3507 EC_GROUP_new_curve_GF2m 3508 EC_GROUP_set_curve_GF2m 3509 EC_GROUP_get_curve_GF2m 3510 EC_POINT_set_affine_coordinates_GF2m 3511 EC_POINT_get_affine_coordinates_GF2m 3512 EC_POINT_set_compressed_coordinates_GF2m 3513 3514 Point compression for binary fields is disabled by default for 3515 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3516 enable it). 3517 3518 As binary polynomials are represented as BIGNUMs, various members 3519 of the EC_GROUP and EC_POINT data structures can be shared 3520 between the implementations for prime fields and binary fields; 3521 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3522 are essentially identical to their ..._GFp counterparts. 3523 (For simplicity, the '..._GFp' prefix has been dropped from 3524 various internal method names.) 3525 3526 An internal 'field_div' method (similar to 'field_mul' and 3527 'field_sqr') has been added; this is used only for binary fields. 3528 3529 [Sheueling Chang Shantz and Douglas Stebila 3530 (Sun Microsystems Laboratories)] 3531 3532 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3533 through methods ('mul', 'precompute_mult'). 3534 3535 The generic implementations (now internally called 'ec_wNAF_mul' 3536 and 'ec_wNAF_precomputed_mult') remain the default if these 3537 methods are undefined. 3538 3539 [Sheueling Chang Shantz and Douglas Stebila 3540 (Sun Microsystems Laboratories)] 3541 3542 *) New function EC_GROUP_get_degree, which is defined through 3543 EC_METHOD. For curves over prime fields, this returns the bit 3544 length of the modulus. 3545 3546 [Sheueling Chang Shantz and Douglas Stebila 3547 (Sun Microsystems Laboratories)] 3548 3549 *) New functions EC_GROUP_dup, EC_POINT_dup. 3550 (These simply call ..._new and ..._copy). 3551 3552 [Sheueling Chang Shantz and Douglas Stebila 3553 (Sun Microsystems Laboratories)] 3554 3555 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3556 Polynomials are represented as BIGNUMs (where the sign bit is not 3557 used) in the following functions [macros]: 3558 3559 BN_GF2m_add 3560 BN_GF2m_sub [= BN_GF2m_add] 3561 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3562 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3563 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3564 BN_GF2m_mod_inv 3565 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3566 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3567 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3568 BN_GF2m_cmp [= BN_ucmp] 3569 3570 (Note that only the 'mod' functions are actually for fields GF(2^m). 3571 BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3572 3573 For some functions, an the irreducible polynomial defining a 3574 field can be given as an 'unsigned int[]' with strictly 3575 decreasing elements giving the indices of those bits that are set; 3576 i.e., p[] represents the polynomial 3577 f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3578 where 3579 p[0] > p[1] > ... > p[k] = 0. 3580 This applies to the following functions: 3581 3582 BN_GF2m_mod_arr 3583 BN_GF2m_mod_mul_arr 3584 BN_GF2m_mod_sqr_arr 3585 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3586 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3587 BN_GF2m_mod_exp_arr 3588 BN_GF2m_mod_sqrt_arr 3589 BN_GF2m_mod_solve_quad_arr 3590 BN_GF2m_poly2arr 3591 BN_GF2m_arr2poly 3592 3593 Conversion can be performed by the following functions: 3594 3595 BN_GF2m_poly2arr 3596 BN_GF2m_arr2poly 3597 3598 bntest.c has additional tests for binary polynomial arithmetic. 3599 3600 Two implementations for BN_GF2m_mod_div() are available. 3601 The default algorithm simply uses BN_GF2m_mod_inv() and 3602 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3603 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3604 copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3605 3606 [Sheueling Chang Shantz and Douglas Stebila 3607 (Sun Microsystems Laboratories)] 3608 3609 *) Add new error code 'ERR_R_DISABLED' that can be used when some 3610 functionality is disabled at compile-time. 3611 [Douglas Stebila <douglas.stebila@sun.com>] 3612 3613 *) Change default behaviour of 'openssl asn1parse' so that more 3614 information is visible when viewing, e.g., a certificate: 3615 3616 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3617 mode the content of non-printable OCTET STRINGs is output in a 3618 style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3619 avoid the appearance of a printable string. 3620 [Nils Larsch <nla@trustcenter.de>] 3621 3622 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3623 functions 3624 EC_GROUP_set_asn1_flag() 3625 EC_GROUP_get_asn1_flag() 3626 EC_GROUP_set_point_conversion_form() 3627 EC_GROUP_get_point_conversion_form() 3628 These control ASN1 encoding details: 3629 - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3630 has been set to OPENSSL_EC_NAMED_CURVE. 3631 - Points are encoded in uncompressed form by default; options for 3632 asn1_for are as for point2oct, namely 3633 POINT_CONVERSION_COMPRESSED 3634 POINT_CONVERSION_UNCOMPRESSED 3635 POINT_CONVERSION_HYBRID 3636 3637 Also add 'seed' and 'seed_len' members to EC_GROUP with access 3638 functions 3639 EC_GROUP_set_seed() 3640 EC_GROUP_get0_seed() 3641 EC_GROUP_get_seed_len() 3642 This is used only for ASN1 purposes (so far). 3643 [Nils Larsch <nla@trustcenter.de>] 3644 3645 *) Add 'field_type' member to EC_METHOD, which holds the NID 3646 of the appropriate field type OID. The new function 3647 EC_METHOD_get_field_type() returns this value. 3648 [Nils Larsch <nla@trustcenter.de>] 3649 3650 *) Add functions 3651 EC_POINT_point2bn() 3652 EC_POINT_bn2point() 3653 EC_POINT_point2hex() 3654 EC_POINT_hex2point() 3655 providing useful interfaces to EC_POINT_point2oct() and 3656 EC_POINT_oct2point(). 3657 [Nils Larsch <nla@trustcenter.de>] 3658 3659 *) Change internals of the EC library so that the functions 3660 EC_GROUP_set_generator() 3661 EC_GROUP_get_generator() 3662 EC_GROUP_get_order() 3663 EC_GROUP_get_cofactor() 3664 are implemented directly in crypto/ec/ec_lib.c and not dispatched 3665 to methods, which would lead to unnecessary code duplication when 3666 adding different types of curves. 3667 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3668 3669 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3670 arithmetic, and such that modified wNAFs are generated 3671 (which avoid length expansion in many cases). 3672 [Bodo Moeller] 3673 3674 *) Add a function EC_GROUP_check_discriminant() (defined via 3675 EC_METHOD) that verifies that the curve discriminant is non-zero. 3676 3677 Add a function EC_GROUP_check() that makes some sanity tests 3678 on a EC_GROUP, its generator and order. This includes 3679 EC_GROUP_check_discriminant(). 3680 [Nils Larsch <nla@trustcenter.de>] 3681 3682 *) Add ECDSA in new directory crypto/ecdsa/. 3683 3684 Add applications 'openssl ecparam' and 'openssl ecdsa' 3685 (these are based on 'openssl dsaparam' and 'openssl dsa'). 3686 3687 ECDSA support is also included in various other files across the 3688 library. Most notably, 3689 - 'openssl req' now has a '-newkey ecdsa:file' option; 3690 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3691 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3692 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3693 them suitable for ECDSA where domain parameters must be 3694 extracted before the specific public key; 3695 - ECDSA engine support has been added. 3696 [Nils Larsch <nla@trustcenter.de>] 3697 3698 *) Include some named elliptic curves, and add OIDs from X9.62, 3699 SECG, and WAP/WTLS. Each curve can be obtained from the new 3700 function 3701 EC_GROUP_new_by_curve_name(), 3702 and the list of available named curves can be obtained with 3703 EC_get_builtin_curves(). 3704 Also add a 'curve_name' member to EC_GROUP objects, which can be 3705 accessed via 3706 EC_GROUP_set_curve_name() 3707 EC_GROUP_get_curve_name() 3708 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3709 3710 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3711 was actually never needed) and in BN_mul(). The removal in BN_mul() 3712 required a small change in bn_mul_part_recursive() and the addition 3713 of the functions bn_cmp_part_words(), bn_sub_part_words() and 3714 bn_add_part_words(), which do the same thing as bn_cmp_words(), 3715 bn_sub_words() and bn_add_words() except they take arrays with 3716 differing sizes. 3717 [Richard Levitte] 3718 3719 Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3720 3721 *) Cleanse PEM buffers before freeing them since they may contain 3722 sensitive data. 3723 [Benjamin Bennett <ben@psc.edu>] 3724 3725 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3726 a ciphersuite string such as "DEFAULT:RSA" cannot enable 3727 authentication-only ciphersuites. 3728 [Bodo Moeller] 3729 3730 *) Since AES128 and AES256 share a single mask bit in the logic of 3731 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3732 kludge to work properly if AES128 is available and AES256 isn't. 3733 [Victor Duchovni] 3734 3735 *) Expand security boundary to match 1.1.1 module. 3736 [Steve Henson] 3737 3738 *) Remove redundant features: hash file source, editing of test vectors 3739 modify fipsld to use external fips_premain.c signature. 3740 [Steve Henson] 3741 3742 *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3743 run algorithm test programs. 3744 [Steve Henson] 3745 3746 *) Make algorithm test programs more tolerant of whitespace. 3747 [Steve Henson] 3748 3749 *) Have SSL/TLS server implementation tolerate "mismatched" record 3750 protocol version while receiving ClientHello even if the 3751 ClientHello is fragmented. (The server can't insist on the 3752 particular protocol version it has chosen before the ServerHello 3753 message has informed the client about his choice.) 3754 [Bodo Moeller] 3755 3756 *) Load error codes if they are not already present instead of using a 3757 static variable. This allows them to be cleanly unloaded and reloaded. 3758 [Steve Henson] 3759 3760 Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3761 3762 *) Introduce limits to prevent malicious keys being able to 3763 cause a denial of service. (CVE-2006-2940) 3764 [Steve Henson, Bodo Moeller] 3765 3766 *) Fix ASN.1 parsing of certain invalid structures that can result 3767 in a denial of service. (CVE-2006-2937) [Steve Henson] 3768 3769 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3770 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3771 3772 *) Fix SSL client code which could crash if connecting to a 3773 malicious SSLv2 server. (CVE-2006-4343) 3774 [Tavis Ormandy and Will Drewry, Google Security Team] 3775 3776 *) Change ciphersuite string processing so that an explicit 3777 ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3778 will no longer include "AES128-SHA"), and any other similar 3779 ciphersuite (same bitmap) from *other* protocol versions (so that 3780 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3781 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3782 changes from 0.9.8b and 0.9.8d. 3783 [Bodo Moeller] 3784 3785 Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3786 3787 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3788 (CVE-2006-4339) [Ben Laurie and Google Security Team] 3789 3790 *) Change the Unix randomness entropy gathering to use poll() when 3791 possible instead of select(), since the latter has some 3792 undesirable limitations. 3793 [Darryl Miles via Richard Levitte and Bodo Moeller] 3794 3795 *) Disable rogue ciphersuites: 3796 3797 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3798 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3799 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3800 3801 The latter two were purportedly from 3802 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3803 appear there. 3804 3805 Also deactive the remaining ciphersuites from 3806 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3807 unofficial, and the ID has long expired. 3808 [Bodo Moeller] 3809 3810 *) Fix RSA blinding Heisenbug (problems sometimes occured on 3811 dual-core machines) and other potential thread-safety issues. 3812 [Bodo Moeller] 3813 3814 Changes between 0.9.7i and 0.9.7j [04 May 2006] 3815 3816 *) Adapt fipsld and the build system to link against the validated FIPS 3817 module in FIPS mode. 3818 [Steve Henson] 3819 3820 *) Fixes for VC++ 2005 build under Windows. 3821 [Steve Henson] 3822 3823 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3824 from a Windows bash shell such as MSYS. It is autodetected from the 3825 "config" script when run from a VC++ environment. Modify standard VC++ 3826 build to use fipscanister.o from the GNU make build. 3827 [Steve Henson] 3828 3829 Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3830 3831 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3832 The value now differs depending on if you build for FIPS or not. 3833 BEWARE! A program linked with a shared FIPSed libcrypto can't be 3834 safely run with a non-FIPSed libcrypto, as it may crash because of 3835 the difference induced by this change. 3836 [Andy Polyakov] 3837 3838 Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3839 3840 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3841 (part of SSL_OP_ALL). This option used to disable the 3842 countermeasure against man-in-the-middle protocol-version 3843 rollback in the SSL 2.0 server implementation, which is a bad 3844 idea. (CVE-2005-2969) 3845 3846 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3847 for Information Security, National Institute of Advanced Industrial 3848 Science and Technology [AIST], Japan)] 3849 3850 *) Minimal support for X9.31 signatures and PSS padding modes. This is 3851 mainly for FIPS compliance and not fully integrated at this stage. 3852 [Steve Henson] 3853 3854 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3855 the exponentiation using a fixed-length exponent. (Otherwise, 3856 the information leaked through timing could expose the secret key 3857 after many signatures; cf. Bleichenbacher's attack on DSA with 3858 biased k.) 3859 [Bodo Moeller] 3860 3861 *) Make a new fixed-window mod_exp implementation the default for 3862 RSA, DSA, and DH private-key operations so that the sequence of 3863 squares and multiplies and the memory access pattern are 3864 independent of the particular secret key. This will mitigate 3865 cache-timing and potential related attacks. 3866 3867 BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3868 and this is automatically used by BN_mod_exp_mont() if the new flag 3869 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3870 will use this BN flag for private exponents unless the flag 3871 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3872 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3873 3874 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3875 3876 *) Change the client implementation for SSLv23_method() and 3877 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3878 Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3879 (Previously, the SSL 2.0 backwards compatible Client Hello 3880 message format would be used even with SSL_OP_NO_SSLv2.) 3881 [Bodo Moeller] 3882 3883 *) Add support for smime-type MIME parameter in S/MIME messages which some 3884 clients need. 3885 [Steve Henson] 3886 3887 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3888 a threadsafe manner. Modify rsa code to use new function and add calls 3889 to dsa and dh code (which had race conditions before). 3890 [Steve Henson] 3891 3892 *) Include the fixed error library code in the C error file definitions 3893 instead of fixing them up at runtime. This keeps the error code 3894 structures constant. 3895 [Steve Henson] 3896 3897 Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3898 3899 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3900 OpenSSL 0.9.8.] 3901 3902 *) Fixes for newer kerberos headers. NB: the casts are needed because 3903 the 'length' field is signed on one version and unsigned on another 3904 with no (?) obvious way to tell the difference, without these VC++ 3905 complains. Also the "definition" of FAR (blank) is no longer included 3906 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3907 some needed definitions. 3908 [Steve Henson] 3909 3910 *) Undo Cygwin change. 3911 [Ulf M�ller] 3912 3913 *) Added support for proxy certificates according to RFC 3820. 3914 Because they may be a security thread to unaware applications, 3915 they must be explicitely allowed in run-time. See 3916 docs/HOWTO/proxy_certificates.txt for further information. 3917 [Richard Levitte] 3918 3919 Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3920 3921 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3922 server and client random values. Previously 3923 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3924 less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3925 3926 This change has negligible security impact because: 3927 3928 1. Server and client random values still have 24 bytes of pseudo random 3929 data. 3930 3931 2. Server and client random values are sent in the clear in the initial 3932 handshake. 3933 3934 3. The master secret is derived using the premaster secret (48 bytes in 3935 size for static RSA ciphersuites) as well as client server and random 3936 values. 3937 3938 The OpenSSL team would like to thank the UK NISCC for bringing this issue 3939 to our attention. 3940 3941 [Stephen Henson, reported by UK NISCC] 3942 3943 *) Use Windows randomness collection on Cygwin. 3944 [Ulf M�ller] 3945 3946 *) Fix hang in EGD/PRNGD query when communication socket is closed 3947 prematurely by EGD/PRNGD. 3948 [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3949 3950 *) Prompt for pass phrases when appropriate for PKCS12 input format. 3951 [Steve Henson] 3952 3953 *) Back-port of selected performance improvements from development 3954 branch, as well as improved support for PowerPC platforms. 3955 [Andy Polyakov] 3956 3957 *) Add lots of checks for memory allocation failure, error codes to indicate 3958 failure and freeing up memory if a failure occurs. 3959 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3960 3961 *) Add new -passin argument to dgst. 3962 [Steve Henson] 3963 3964 *) Perform some character comparisons of different types in X509_NAME_cmp: 3965 this is needed for some certificates that reencode DNs into UTF8Strings 3966 (in violation of RFC3280) and can't or wont issue name rollover 3967 certificates. 3968 [Steve Henson] 3969 3970 *) Make an explicit check during certificate validation to see that 3971 the CA setting in each certificate on the chain is correct. As a 3972 side effect always do the following basic checks on extensions, 3973 not just when there's an associated purpose to the check: 3974 3975 - if there is an unhandled critical extension (unless the user 3976 has chosen to ignore this fault) 3977 - if the path length has been exceeded (if one is set at all) 3978 - that certain extensions fit the associated purpose (if one has 3979 been given) 3980 [Richard Levitte] 3981 3982 Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3983 3984 *) Avoid a race condition when CRLs are checked in a multi threaded 3985 environment. This would happen due to the reordering of the revoked 3986 entries during signature checking and serial number lookup. Now the 3987 encoding is cached and the serial number sort performed under a lock. 3988 Add new STACK function sk_is_sorted(). 3989 [Steve Henson] 3990 3991 *) Add Delta CRL to the extension code. 3992 [Steve Henson] 3993 3994 *) Various fixes to s3_pkt.c so alerts are sent properly. 3995 [David Holmes <d.holmes@f5.com>] 3996 3997 *) Reduce the chances of duplicate issuer name and serial numbers (in 3998 violation of RFC3280) using the OpenSSL certificate creation utilities. 3999 This is done by creating a random 64 bit value for the initial serial 4000 number when a serial number file is created or when a self signed 4001 certificate is created using 'openssl req -x509'. The initial serial 4002 number file is created using 'openssl x509 -next_serial' in CA.pl 4003 rather than being initialized to 1. 4004 [Steve Henson] 4005 4006 Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 4007 4008 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 4009 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 4010 [Joe Orton, Steve Henson] 4011 4012 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 4013 (CVE-2004-0112) 4014 [Joe Orton, Steve Henson] 4015 4016 *) Make it possible to have multiple active certificates with the same 4017 subject in the CA index file. This is done only if the keyword 4018 'unique_subject' is set to 'no' in the main CA section (default 4019 if 'CA_default') of the configuration file. The value is saved 4020 with the database itself in a separate index attribute file, 4021 named like the index file with '.attr' appended to the name. 4022 [Richard Levitte] 4023 4024 *) X509 verify fixes. Disable broken certificate workarounds when 4025 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 4026 keyUsage extension present. Don't accept CRLs with unhandled critical 4027 extensions: since verify currently doesn't process CRL extensions this 4028 rejects a CRL with *any* critical extensions. Add new verify error codes 4029 for these cases. 4030 [Steve Henson] 4031 4032 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 4033 A clarification of RFC2560 will require the use of OCTET STRINGs and 4034 some implementations cannot handle the current raw format. Since OpenSSL 4035 copies and compares OCSP nonces as opaque blobs without any attempt at 4036 parsing them this should not create any compatibility issues. 4037 [Steve Henson] 4038 4039 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 4040 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 4041 this HMAC (and other) operations are several times slower than OpenSSL 4042 < 0.9.7. 4043 [Steve Henson] 4044 4045 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 4046 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4047 4048 *) Use the correct content when signing type "other". 4049 [Steve Henson] 4050 4051 Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 4052 4053 *) Fix various bugs revealed by running the NISCC test suite: 4054 4055 Stop out of bounds reads in the ASN1 code when presented with 4056 invalid tags (CVE-2003-0543 and CVE-2003-0544). 4057 4058 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 4059 4060 If verify callback ignores invalid public key errors don't try to check 4061 certificate signature with the NULL public key. 4062 4063 [Steve Henson] 4064 4065 *) New -ignore_err option in ocsp application to stop the server 4066 exiting on the first error in a request. 4067 [Steve Henson] 4068 4069 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4070 if the server requested one: as stated in TLS 1.0 and SSL 3.0 4071 specifications. 4072 [Steve Henson] 4073 4074 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4075 extra data after the compression methods not only for TLS 1.0 4076 but also for SSL 3.0 (as required by the specification). 4077 [Bodo Moeller; problem pointed out by Matthias Loepfe] 4078 4079 *) Change X509_certificate_type() to mark the key as exported/exportable 4080 when it's 512 *bits* long, not 512 bytes. 4081 [Richard Levitte] 4082 4083 *) Change AES_cbc_encrypt() so it outputs exact multiple of 4084 blocks during encryption. 4085 [Richard Levitte] 4086 4087 *) Various fixes to base64 BIO and non blocking I/O. On write 4088 flushes were not handled properly if the BIO retried. On read 4089 data was not being buffered properly and had various logic bugs. 4090 This also affects blocking I/O when the data being decoded is a 4091 certain size. 4092 [Steve Henson] 4093 4094 *) Various S/MIME bugfixes and compatibility changes: 4095 output correct application/pkcs7 MIME type if 4096 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 4097 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 4098 of files as .eml work). Correctly handle very long lines in MIME 4099 parser. 4100 [Steve Henson] 4101 4102 Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 4103 4104 *) Countermeasure against the Klima-Pokorny-Rosa extension of 4105 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 4106 a protocol version number mismatch like a decryption error 4107 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 4108 [Bodo Moeller] 4109 4110 *) Turn on RSA blinding by default in the default implementation 4111 to avoid a timing attack. Applications that don't want it can call 4112 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4113 They would be ill-advised to do so in most cases. 4114 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4115 4116 *) Change RSA blinding code so that it works when the PRNG is not 4117 seeded (in this case, the secret RSA exponent is abused as 4118 an unpredictable seed -- if it is not unpredictable, there 4119 is no point in blinding anyway). Make RSA blinding thread-safe 4120 by remembering the creator's thread ID in rsa->blinding and 4121 having all other threads use local one-time blinding factors 4122 (this requires more computation than sharing rsa->blinding, but 4123 avoids excessive locking; and if an RSA object is not shared 4124 between threads, blinding will still be very fast). 4125 [Bodo Moeller] 4126 4127 *) Fixed a typo bug that would cause ENGINE_set_default() to set an 4128 ENGINE as defaults for all supported algorithms irrespective of 4129 the 'flags' parameter. 'flags' is now honoured, so applications 4130 should make sure they are passing it correctly. 4131 [Geoff Thorpe] 4132 4133 *) Target "mingw" now allows native Windows code to be generated in 4134 the Cygwin environment as well as with the MinGW compiler. 4135 [Ulf Moeller] 4136 4137 Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4138 4139 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4140 via timing by performing a MAC computation even if incorrrect 4141 block cipher padding has been found. This is a countermeasure 4142 against active attacks where the attacker has to distinguish 4143 between bad padding and a MAC verification error. (CVE-2003-0078) 4144 4145 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4146 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4147 Martin Vuagnoux (EPFL, Ilion)] 4148 4149 *) Make the no-err option work as intended. The intention with no-err 4150 is not to have the whole error stack handling routines removed from 4151 libcrypto, it's only intended to remove all the function name and 4152 reason texts, thereby removing some of the footprint that may not 4153 be interesting if those errors aren't displayed anyway. 4154 4155 NOTE: it's still possible for any application or module to have it's 4156 own set of error texts inserted. The routines are there, just not 4157 used by default when no-err is given. 4158 [Richard Levitte] 4159 4160 *) Add support for FreeBSD on IA64. 4161 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 4162 4163 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 4164 Kerberos function mit_des_cbc_cksum(). Before this change, 4165 the value returned by DES_cbc_cksum() was like the one from 4166 mit_des_cbc_cksum(), except the bytes were swapped. 4167 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 4168 4169 *) Allow an application to disable the automatic SSL chain building. 4170 Before this a rather primitive chain build was always performed in 4171 ssl3_output_cert_chain(): an application had no way to send the 4172 correct chain if the automatic operation produced an incorrect result. 4173 4174 Now the chain builder is disabled if either: 4175 4176 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 4177 4178 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 4179 4180 The reasoning behind this is that an application would not want the 4181 auto chain building to take place if extra chain certificates are 4182 present and it might also want a means of sending no additional 4183 certificates (for example the chain has two certificates and the 4184 root is omitted). 4185 [Steve Henson] 4186 4187 *) Add the possibility to build without the ENGINE framework. 4188 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4189 4190 *) Under Win32 gmtime() can return NULL: check return value in 4191 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4192 [Steve Henson] 4193 4194 *) DSA routines: under certain error conditions uninitialized BN objects 4195 could be freed. Solution: make sure initialization is performed early 4196 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4197 Nils Larsch <nla@trustcenter.de> via PR#459) 4198 [Lutz Jaenicke] 4199 4200 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4201 checked on reconnect on the client side, therefore session resumption 4202 could still fail with a "ssl session id is different" error. This 4203 behaviour is masked when SSL_OP_ALL is used due to 4204 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4205 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4206 followup to PR #377. 4207 [Lutz Jaenicke] 4208 4209 *) IA-32 assembler support enhancements: unified ELF targets, support 4210 for SCO/Caldera platforms, fix for Cygwin shared build. 4211 [Andy Polyakov] 4212 4213 *) Add support for FreeBSD on sparc64. As a consequence, support for 4214 FreeBSD on non-x86 processors is separate from x86 processors on 4215 the config script, much like the NetBSD support. 4216 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4217 4218 Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4219 4220 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4221 OpenSSL 0.9.7.] 4222 4223 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4224 code (06) was taken as the first octet of the session ID and the last 4225 octet was ignored consequently. As a result SSLv2 client side session 4226 caching could not have worked due to the session ID mismatch between 4227 client and server. 4228 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4229 PR #377. 4230 [Lutz Jaenicke] 4231 4232 *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4233 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4234 removed entirely. 4235 [Richard Levitte] 4236 4237 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4238 seems that in spite of existing for more than a year, many application 4239 author have done nothing to provide the necessary callbacks, which 4240 means that this particular engine will not work properly anywhere. 4241 This is a very unfortunate situation which forces us, in the name 4242 of usability, to give the hw_ncipher.c a static lock, which is part 4243 of libcrypto. 4244 NOTE: This is for the 0.9.7 series ONLY. This hack will never 4245 appear in 0.9.8 or later. We EXPECT application authors to have 4246 dealt properly with this when 0.9.8 is released (unless we actually 4247 make such changes in the libcrypto locking code that changes will 4248 have to be made anyway). 4249 [Richard Levitte] 4250 4251 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4252 octets have been read, EOF or an error occurs. Without this change 4253 some truncated ASN1 structures will not produce an error. 4254 [Steve Henson] 4255 4256 *) Disable Heimdal support, since it hasn't been fully implemented. 4257 Still give the possibility to force the use of Heimdal, but with 4258 warnings and a request that patches get sent to openssl-dev. 4259 [Richard Levitte] 4260 4261 *) Add the VC-CE target, introduce the WINCE sysname, and add 4262 INSTALL.WCE and appropriate conditionals to make it build. 4263 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4264 4265 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4266 cygssl-x.y.z.dll, where x, y and z are the major, minor and 4267 edit numbers of the version. 4268 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4269 4270 *) Introduce safe string copy and catenation functions 4271 (BUF_strlcpy() and BUF_strlcat()). 4272 [Ben Laurie (CHATS) and Richard Levitte] 4273 4274 *) Avoid using fixed-size buffers for one-line DNs. 4275 [Ben Laurie (CHATS)] 4276 4277 *) Add BUF_MEM_grow_clean() to avoid information leakage when 4278 resizing buffers containing secrets, and use where appropriate. 4279 [Ben Laurie (CHATS)] 4280 4281 *) Avoid using fixed size buffers for configuration file location. 4282 [Ben Laurie (CHATS)] 4283 4284 *) Avoid filename truncation for various CA files. 4285 [Ben Laurie (CHATS)] 4286 4287 *) Use sizeof in preference to magic numbers. 4288 [Ben Laurie (CHATS)] 4289 4290 *) Avoid filename truncation in cert requests. 4291 [Ben Laurie (CHATS)] 4292 4293 *) Add assertions to check for (supposedly impossible) buffer 4294 overflows. 4295 [Ben Laurie (CHATS)] 4296 4297 *) Don't cache truncated DNS entries in the local cache (this could 4298 potentially lead to a spoofing attack). 4299 [Ben Laurie (CHATS)] 4300 4301 *) Fix various buffers to be large enough for hex/decimal 4302 representations in a platform independent manner. 4303 [Ben Laurie (CHATS)] 4304 4305 *) Add CRYPTO_realloc_clean() to avoid information leakage when 4306 resizing buffers containing secrets, and use where appropriate. 4307 [Ben Laurie (CHATS)] 4308 4309 *) Add BIO_indent() to avoid much slightly worrying code to do 4310 indents. 4311 [Ben Laurie (CHATS)] 4312 4313 *) Convert sprintf()/BIO_puts() to BIO_printf(). 4314 [Ben Laurie (CHATS)] 4315 4316 *) buffer_gets() could terminate with the buffer only half 4317 full. Fixed. 4318 [Ben Laurie (CHATS)] 4319 4320 *) Add assertions to prevent user-supplied crypto functions from 4321 overflowing internal buffers by having large block sizes, etc. 4322 [Ben Laurie (CHATS)] 4323 4324 *) New OPENSSL_assert() macro (similar to assert(), but enabled 4325 unconditionally). 4326 [Ben Laurie (CHATS)] 4327 4328 *) Eliminate unused copy of key in RC4. 4329 [Ben Laurie (CHATS)] 4330 4331 *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4332 [Ben Laurie (CHATS)] 4333 4334 *) Fix off-by-one error in EGD path. 4335 [Ben Laurie (CHATS)] 4336 4337 *) If RANDFILE path is too long, ignore instead of truncating. 4338 [Ben Laurie (CHATS)] 4339 4340 *) Eliminate unused and incorrectly sized X.509 structure 4341 CBCParameter. 4342 [Ben Laurie (CHATS)] 4343 4344 *) Eliminate unused and dangerous function knumber(). 4345 [Ben Laurie (CHATS)] 4346 4347 *) Eliminate unused and dangerous structure, KSSL_ERR. 4348 [Ben Laurie (CHATS)] 4349 4350 *) Protect against overlong session ID context length in an encoded 4351 session object. Since these are local, this does not appear to be 4352 exploitable. 4353 [Ben Laurie (CHATS)] 4354 4355 *) Change from security patch (see 0.9.6e below) that did not affect 4356 the 0.9.6 release series: 4357 4358 Remote buffer overflow in SSL3 protocol - an attacker could 4359 supply an oversized master key in Kerberos-enabled versions. 4360 (CVE-2002-0657) 4361 [Ben Laurie (CHATS)] 4362 4363 *) Change the SSL kerb5 codes to match RFC 2712. 4364 [Richard Levitte] 4365 4366 *) Make -nameopt work fully for req and add -reqopt switch. 4367 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4368 4369 *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4370 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4371 4372 *) Make sure tests can be performed even if the corresponding algorithms 4373 have been removed entirely. This was also the last step to make 4374 OpenSSL compilable with DJGPP under all reasonable conditions. 4375 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4376 4377 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4378 to allow version independent disabling of normally unselected ciphers, 4379 which may be activated as a side-effect of selecting a single cipher. 4380 4381 (E.g., cipher list string "RSA" enables ciphersuites that are left 4382 out of "ALL" because they do not provide symmetric encryption. 4383 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4384 [Lutz Jaenicke, Bodo Moeller] 4385 4386 *) Add appropriate support for separate platform-dependent build 4387 directories. The recommended way to make a platform-dependent 4388 build directory is the following (tested on Linux), maybe with 4389 some local tweaks: 4390 4391 # Place yourself outside of the OpenSSL source tree. In 4392 # this example, the environment variable OPENSSL_SOURCE 4393 # is assumed to contain the absolute OpenSSL source directory. 4394 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4395 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4396 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4397 mkdir -p `dirname $F` 4398 ln -s $OPENSSL_SOURCE/$F $F 4399 done 4400 4401 To be absolutely sure not to disturb the source tree, a "make clean" 4402 is a good thing. If it isn't successfull, don't worry about it, 4403 it probably means the source directory is very clean. 4404 [Richard Levitte] 4405 4406 *) Make sure any ENGINE control commands make local copies of string 4407 pointers passed to them whenever necessary. Otherwise it is possible 4408 the caller may have overwritten (or deallocated) the original string 4409 data when a later ENGINE operation tries to use the stored values. 4410 [G�tz Babin-Ebell <babinebell@trustcenter.de>] 4411 4412 *) Improve diagnostics in file reading and command-line digests. 4413 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4414 4415 *) Add AES modes CFB and OFB to the object database. Correct an 4416 error in AES-CFB decryption. 4417 [Richard Levitte] 4418 4419 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4420 allows existing EVP_CIPHER_CTX structures to be reused after 4421 calling EVP_*Final(). This behaviour is used by encryption 4422 BIOs and some applications. This has the side effect that 4423 applications must explicitly clean up cipher contexts with 4424 EVP_CIPHER_CTX_cleanup() or they will leak memory. 4425 [Steve Henson] 4426 4427 *) Check the values of dna and dnb in bn_mul_recursive before calling 4428 bn_mul_comba (a non zero value means the a or b arrays do not contain 4429 n2 elements) and fallback to bn_mul_normal if either is not zero. 4430 [Steve Henson] 4431 4432 *) Fix escaping of non-ASCII characters when using the -subj option 4433 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4434 [Lutz Jaenicke] 4435 4436 *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4437 form for "surname", serialNumber has no short form. 4438 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4439 therefore remove "mail" short name for "internet 7". 4440 The OID for unique identifiers in X509 certificates is 4441 x500UniqueIdentifier, not uniqueIdentifier. 4442 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4443 [Lutz Jaenicke] 4444 4445 *) Add an "init" command to the ENGINE config module and auto initialize 4446 ENGINEs. Without any "init" command the ENGINE will be initialized 4447 after all ctrl commands have been executed on it. If init=1 the 4448 ENGINE is initailized at that point (ctrls before that point are run 4449 on the uninitialized ENGINE and after on the initialized one). If 4450 init=0 then the ENGINE will not be iniatialized at all. 4451 [Steve Henson] 4452 4453 *) Fix the 'app_verify_callback' interface so that the user-defined 4454 argument is actually passed to the callback: In the 4455 SSL_CTX_set_cert_verify_callback() prototype, the callback 4456 declaration has been changed from 4457 int (*cb)() 4458 into 4459 int (*cb)(X509_STORE_CTX *,void *); 4460 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4461 i=s->ctx->app_verify_callback(&ctx) 4462 has been changed into 4463 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4464 4465 To update applications using SSL_CTX_set_cert_verify_callback(), 4466 a dummy argument can be added to their callback functions. 4467 [D. K. Smetters <smetters@parc.xerox.com>] 4468 4469 *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4470 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4471 4472 *) Add and OPENSSL_LOAD_CONF define which will cause 4473 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4474 This allows older applications to transparently support certain 4475 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4476 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4477 load the config file and OPENSSL_add_all_algorithms_conf() which will 4478 always load it have also been added. 4479 [Steve Henson] 4480 4481 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4482 Adjust NIDs and EVP layer. 4483 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4484 4485 *) Config modules support in openssl utility. 4486 4487 Most commands now load modules from the config file, 4488 though in a few (such as version) this isn't done 4489 because it couldn't be used for anything. 4490 4491 In the case of ca and req the config file used is 4492 the same as the utility itself: that is the -config 4493 command line option can be used to specify an 4494 alternative file. 4495 [Steve Henson] 4496 4497 *) Move default behaviour from OPENSSL_config(). If appname is NULL 4498 use "openssl_conf" if filename is NULL use default openssl config file. 4499 [Steve Henson] 4500 4501 *) Add an argument to OPENSSL_config() to allow the use of an alternative 4502 config section name. Add a new flag to tolerate a missing config file 4503 and move code to CONF_modules_load_file(). 4504 [Steve Henson] 4505 4506 *) Support for crypto accelerator cards from Accelerated Encryption 4507 Processing, www.aep.ie. (Use engine 'aep') 4508 The support was copied from 0.9.6c [engine] and adapted/corrected 4509 to work with the new engine framework. 4510 [AEP Inc. and Richard Levitte] 4511 4512 *) Support for SureWare crypto accelerator cards from Baltimore 4513 Technologies. (Use engine 'sureware') 4514 The support was copied from 0.9.6c [engine] and adapted 4515 to work with the new engine framework. 4516 [Richard Levitte] 4517 4518 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4519 make the newer ENGINE framework commands for the CHIL engine work. 4520 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4521 4522 *) Make it possible to produce shared libraries on ReliantUNIX. 4523 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4524 4525 *) Add the configuration target debug-linux-ppro. 4526 Make 'openssl rsa' use the general key loading routines 4527 implemented in apps.c, and make those routines able to 4528 handle the key format FORMAT_NETSCAPE and the variant 4529 FORMAT_IISSGC. 4530 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4531 4532 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4533 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4534 4535 *) Add -keyform to rsautl, and document -engine. 4536 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4537 4538 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4539 BIO_R_NO_SUCH_FILE error code rather than the generic 4540 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4541 [Ben Laurie] 4542 4543 *) Add new functions 4544 ERR_peek_last_error 4545 ERR_peek_last_error_line 4546 ERR_peek_last_error_line_data. 4547 These are similar to 4548 ERR_peek_error 4549 ERR_peek_error_line 4550 ERR_peek_error_line_data, 4551 but report on the latest error recorded rather than the first one 4552 still in the error queue. 4553 [Ben Laurie, Bodo Moeller] 4554 4555 *) default_algorithms option in ENGINE config module. This allows things 4556 like: 4557 default_algorithms = ALL 4558 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4559 [Steve Henson] 4560 4561 *) Prelminary ENGINE config module. 4562 [Steve Henson] 4563 4564 *) New experimental application configuration code. 4565 [Steve Henson] 4566 4567 *) Change the AES code to follow the same name structure as all other 4568 symmetric ciphers, and behave the same way. Move everything to 4569 the directory crypto/aes, thereby obsoleting crypto/rijndael. 4570 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4571 4572 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4573 [Ben Laurie and Theo de Raadt] 4574 4575 *) Add option to output public keys in req command. 4576 [Massimiliano Pala madwolf@openca.org] 4577 4578 *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4579 (up to about 10% better than before for P-192 and P-224). 4580 [Bodo Moeller] 4581 4582 *) New functions/macros 4583 4584 SSL_CTX_set_msg_callback(ctx, cb) 4585 SSL_CTX_set_msg_callback_arg(ctx, arg) 4586 SSL_set_msg_callback(ssl, cb) 4587 SSL_set_msg_callback_arg(ssl, arg) 4588 4589 to request calling a callback function 4590 4591 void cb(int write_p, int version, int content_type, 4592 const void *buf, size_t len, SSL *ssl, void *arg) 4593 4594 whenever a protocol message has been completely received 4595 (write_p == 0) or sent (write_p == 1). Here 'version' is the 4596 protocol version according to which the SSL library interprets 4597 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4598 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4599 the content type as defined in the SSL 3.0/TLS 1.0 protocol 4600 specification (change_cipher_spec(20), alert(21), handshake(22)). 4601 'buf' and 'len' point to the actual message, 'ssl' to the 4602 SSL object, and 'arg' is the application-defined value set by 4603 SSL[_CTX]_set_msg_callback_arg(). 4604 4605 'openssl s_client' and 'openssl s_server' have new '-msg' options 4606 to enable a callback that displays all protocol messages. 4607 [Bodo Moeller] 4608 4609 *) Change the shared library support so shared libraries are built as 4610 soon as the corresponding static library is finished, and thereby get 4611 openssl and the test programs linked against the shared library. 4612 This still only happens when the keyword "shard" has been given to 4613 the configuration scripts. 4614 4615 NOTE: shared library support is still an experimental thing, and 4616 backward binary compatibility is still not guaranteed. 4617 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4618 4619 *) Add support for Subject Information Access extension. 4620 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4621 4622 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4623 additional bytes when new memory had to be allocated, not just 4624 when reusing an existing buffer. 4625 [Bodo Moeller] 4626 4627 *) New command line and configuration option 'utf8' for the req command. 4628 This allows field values to be specified as UTF8 strings. 4629 [Steve Henson] 4630 4631 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4632 runs for the former and machine-readable output for the latter. 4633 [Ben Laurie] 4634 4635 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4636 of the e-mail address in the DN (i.e., it will go into a certificate 4637 extension only). The new configuration file option 'email_in_dn = no' 4638 has the same effect. 4639 [Massimiliano Pala madwolf@openca.org] 4640 4641 *) Change all functions with names starting with des_ to be starting 4642 with DES_ instead. Add wrappers that are compatible with libdes, 4643 but are named _ossl_old_des_*. Finally, add macros that map the 4644 des_* symbols to the corresponding _ossl_old_des_* if libdes 4645 compatibility is desired. If OpenSSL 0.9.6c compatibility is 4646 desired, the des_* symbols will be mapped to DES_*, with one 4647 exception. 4648 4649 Since we provide two compatibility mappings, the user needs to 4650 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4651 compatibility is desired. The default (i.e., when that macro 4652 isn't defined) is OpenSSL 0.9.6c compatibility. 4653 4654 There are also macros that enable and disable the support of old 4655 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4656 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4657 are defined, the default will apply: to support the old des routines. 4658 4659 In either case, one must include openssl/des.h to get the correct 4660 definitions. Do not try to just include openssl/des_old.h, that 4661 won't work. 4662 4663 NOTE: This is a major break of an old API into a new one. Software 4664 authors are encouraged to switch to the DES_ style functions. Some 4665 time in the future, des_old.h and the libdes compatibility functions 4666 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4667 default), and then completely removed. 4668 [Richard Levitte] 4669 4670 *) Test for certificates which contain unsupported critical extensions. 4671 If such a certificate is found during a verify operation it is 4672 rejected by default: this behaviour can be overridden by either 4673 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4674 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4675 X509_supported_extension() has also been added which returns 1 if a 4676 particular extension is supported. 4677 [Steve Henson] 4678 4679 *) Modify the behaviour of EVP cipher functions in similar way to digests 4680 to retain compatibility with existing code. 4681 [Steve Henson] 4682 4683 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4684 compatibility with existing code. In particular the 'ctx' parameter does 4685 not have to be to be initialized before the call to EVP_DigestInit() and 4686 it is tidied up after a call to EVP_DigestFinal(). New function 4687 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4688 EVP_MD_CTX_copy() changed to not require the destination to be 4689 initialized valid and new function EVP_MD_CTX_copy_ex() added which 4690 requires the destination to be valid. 4691 4692 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4693 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4694 [Steve Henson] 4695 4696 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4697 so that complete 'Handshake' protocol structures are kept in memory 4698 instead of overwriting 'msg_type' and 'length' with 'body' data. 4699 [Bodo Moeller] 4700 4701 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4702 [Massimo Santin via Richard Levitte] 4703 4704 *) Major restructuring to the underlying ENGINE code. This includes 4705 reduction of linker bloat, separation of pure "ENGINE" manipulation 4706 (initialisation, etc) from functionality dealing with implementations 4707 of specific crypto iterfaces. This change also introduces integrated 4708 support for symmetric ciphers and digest implementations - so ENGINEs 4709 can now accelerate these by providing EVP_CIPHER and EVP_MD 4710 implementations of their own. This is detailed in crypto/engine/README 4711 as it couldn't be adequately described here. However, there are a few 4712 API changes worth noting - some RSA, DSA, DH, and RAND functions that 4713 were changed in the original introduction of ENGINE code have now 4714 reverted back - the hooking from this code to ENGINE is now a good 4715 deal more passive and at run-time, operations deal directly with 4716 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4717 dereferencing through an ENGINE pointer any more. Also, the ENGINE 4718 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4719 they were not being used by the framework as there is no concept of a 4720 BIGNUM_METHOD and they could not be generalised to the new 4721 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4722 ENGINE_cpy() has been removed as it cannot be consistently defined in 4723 the new code. 4724 [Geoff Thorpe] 4725 4726 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4727 [Steve Henson] 4728 4729 *) Change mkdef.pl to sort symbols that get the same entry number, 4730 and make sure the automatically generated functions ERR_load_* 4731 become part of libeay.num as well. 4732 [Richard Levitte] 4733 4734 *) New function SSL_renegotiate_pending(). This returns true once 4735 renegotiation has been requested (either SSL_renegotiate() call 4736 or HelloRequest/ClientHello receveived from the peer) and becomes 4737 false once a handshake has been completed. 4738 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4739 sends a HelloRequest, but does not ensure that a handshake takes 4740 place. SSL_renegotiate_pending() is useful for checking if the 4741 client has followed the request.) 4742 [Bodo Moeller] 4743 4744 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4745 By default, clients may request session resumption even during 4746 renegotiation (if session ID contexts permit); with this option, 4747 session resumption is possible only in the first handshake. 4748 4749 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4750 more bits available for options that should not be part of 4751 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4752 [Bodo Moeller] 4753 4754 *) Add some demos for certificate and certificate request creation. 4755 [Steve Henson] 4756 4757 *) Make maximum certificate chain size accepted from the peer application 4758 settable (SSL*_get/set_max_cert_list()), as proposed by 4759 "Douglas E. Engert" <deengert@anl.gov>. 4760 [Lutz Jaenicke] 4761 4762 *) Add support for shared libraries for Unixware-7 4763 (Boyd Lynn Gerber <gerberb@zenez.com>). 4764 [Lutz Jaenicke] 4765 4766 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4767 be done prior to destruction. Use this to unload error strings from 4768 ENGINEs that load their own error strings. NB: This adds two new API 4769 functions to "get" and "set" this destroy handler in an ENGINE. 4770 [Geoff Thorpe] 4771 4772 *) Alter all existing ENGINE implementations (except "openssl" and 4773 "openbsd") to dynamically instantiate their own error strings. This 4774 makes them more flexible to be built both as statically-linked ENGINEs 4775 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4776 Also, add stub code to each that makes building them as self-contained 4777 shared-libraries easier (see README.ENGINE). 4778 [Geoff Thorpe] 4779 4780 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4781 implementations into applications that are completely implemented in 4782 self-contained shared-libraries. The "dynamic" ENGINE exposes control 4783 commands that can be used to configure what shared-library to load and 4784 to control aspects of the way it is handled. Also, made an update to 4785 the README.ENGINE file that brings its information up-to-date and 4786 provides some information and instructions on the "dynamic" ENGINE 4787 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4788 [Geoff Thorpe] 4789 4790 *) Make it possible to unload ranges of ERR strings with a new 4791 "ERR_unload_strings" function. 4792 [Geoff Thorpe] 4793 4794 *) Add a copy() function to EVP_MD. 4795 [Ben Laurie] 4796 4797 *) Make EVP_MD routines take a context pointer instead of just the 4798 md_data void pointer. 4799 [Ben Laurie] 4800 4801 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4802 that the digest can only process a single chunk of data 4803 (typically because it is provided by a piece of 4804 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4805 is only going to provide a single chunk of data, and hence the 4806 framework needn't accumulate the data for oneshot drivers. 4807 [Ben Laurie] 4808 4809 *) As with "ERR", make it possible to replace the underlying "ex_data" 4810 functions. This change also alters the storage and management of global 4811 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4812 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4813 index counters. The API functions that use this state have been changed 4814 to take a "class_index" rather than pointers to the class's local STACK 4815 and counter, and there is now an API function to dynamically create new 4816 classes. This centralisation allows us to (a) plug a lot of the 4817 thread-safety problems that existed, and (b) makes it possible to clean 4818 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4819 such data would previously have always leaked in application code and 4820 workarounds were in place to make the memory debugging turn a blind eye 4821 to it. Application code that doesn't use this new function will still 4822 leak as before, but their memory debugging output will announce it now 4823 rather than letting it slide. 4824 4825 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4826 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4827 has a return value to indicate success or failure. 4828 [Geoff Thorpe] 4829 4830 *) Make it possible to replace the underlying "ERR" functions such that the 4831 global state (2 LHASH tables and 2 locks) is only used by the "default" 4832 implementation. This change also adds two functions to "get" and "set" 4833 the implementation prior to it being automatically set the first time 4834 any other ERR function takes place. Ie. an application can call "get", 4835 pass the return value to a module it has just loaded, and that module 4836 can call its own "set" function using that value. This means the 4837 module's "ERR" operations will use (and modify) the error state in the 4838 application and not in its own statically linked copy of OpenSSL code. 4839 [Geoff Thorpe] 4840 4841 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4842 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4843 the operation, and provides a more encapsulated way for external code 4844 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4845 to use these functions rather than manually incrementing the counts. 4846 4847 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4848 [Geoff Thorpe] 4849 4850 *) Add EVP test program. 4851 [Ben Laurie] 4852 4853 *) Add symmetric cipher support to ENGINE. Expect the API to change! 4854 [Ben Laurie] 4855 4856 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4857 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4858 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4859 These allow a CRL to be built without having to access X509_CRL fields 4860 directly. Modify 'ca' application to use new functions. 4861 [Steve Henson] 4862 4863 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4864 bug workarounds. Rollback attack detection is a security feature. 4865 The problem will only arise on OpenSSL servers when TLSv1 is not 4866 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4867 Software authors not wanting to support TLSv1 will have special reasons 4868 for their choice and can explicitly enable this option. 4869 [Bodo Moeller, Lutz Jaenicke] 4870 4871 *) Rationalise EVP so it can be extended: don't include a union of 4872 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4873 (similar to those existing for EVP_CIPHER_CTX). 4874 Usage example: 4875 4876 EVP_MD_CTX md; 4877 4878 EVP_MD_CTX_init(&md); /* new function call */ 4879 EVP_DigestInit(&md, EVP_sha1()); 4880 EVP_DigestUpdate(&md, in, len); 4881 EVP_DigestFinal(&md, out, NULL); 4882 EVP_MD_CTX_cleanup(&md); /* new function call */ 4883 4884 [Ben Laurie] 4885 4886 *) Make DES key schedule conform to the usual scheme, as well as 4887 correcting its structure. This means that calls to DES functions 4888 now have to pass a pointer to a des_key_schedule instead of a 4889 plain des_key_schedule (which was actually always a pointer 4890 anyway): E.g., 4891 4892 des_key_schedule ks; 4893 4894 des_set_key_checked(..., &ks); 4895 des_ncbc_encrypt(..., &ks, ...); 4896 4897 (Note that a later change renames 'des_...' into 'DES_...'.) 4898 [Ben Laurie] 4899 4900 *) Initial reduction of linker bloat: the use of some functions, such as 4901 PEM causes large amounts of unused functions to be linked in due to 4902 poor organisation. For example pem_all.c contains every PEM function 4903 which has a knock on effect of linking in large amounts of (unused) 4904 ASN1 code. Grouping together similar functions and splitting unrelated 4905 functions prevents this. 4906 [Steve Henson] 4907 4908 *) Cleanup of EVP macros. 4909 [Ben Laurie] 4910 4911 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4912 correct _ecb suffix. 4913 [Ben Laurie] 4914 4915 *) Add initial OCSP responder support to ocsp application. The 4916 revocation information is handled using the text based index 4917 use by the ca application. The responder can either handle 4918 requests generated internally, supplied in files (for example 4919 via a CGI script) or using an internal minimal server. 4920 [Steve Henson] 4921 4922 *) Add configuration choices to get zlib compression for TLS. 4923 [Richard Levitte] 4924 4925 *) Changes to Kerberos SSL for RFC 2712 compliance: 4926 1. Implemented real KerberosWrapper, instead of just using 4927 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4928 2. Implemented optional authenticator field of KerberosWrapper. 4929 4930 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4931 and authenticator structs; see crypto/krb5/. 4932 4933 Generalized Kerberos calls to support multiple Kerberos libraries. 4934 [Vern Staats <staatsvr@asc.hpc.mil>, 4935 Jeffrey Altman <jaltman@columbia.edu> 4936 via Richard Levitte] 4937 4938 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4939 already does with RSA. testdsa.h now has 'priv_key/pub_key' 4940 values for each of the key sizes rather than having just 4941 parameters (and 'speed' generating keys each time). 4942 [Geoff Thorpe] 4943 4944 *) Speed up EVP routines. 4945 Before: 4946encrypt 4947type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4948des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4949des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4950des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4951decrypt 4952des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4953des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4954des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4955 After: 4956encrypt 4957des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4958decrypt 4959des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4960 [Ben Laurie] 4961 4962 *) Added the OS2-EMX target. 4963 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4964 4965 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4966 to support NCONF routines in extension code. New function CONF_set_nconf() 4967 to allow functions which take an NCONF to also handle the old LHASH 4968 structure: this means that the old CONF compatible routines can be 4969 retained (in particular wrt extensions) without having to duplicate the 4970 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4971 [Steve Henson] 4972 4973 *) Enhance the general user interface with mechanisms for inner control 4974 and with possibilities to have yes/no kind of prompts. 4975 [Richard Levitte] 4976 4977 *) Change all calls to low level digest routines in the library and 4978 applications to use EVP. Add missing calls to HMAC_cleanup() and 4979 don't assume HMAC_CTX can be copied using memcpy(). 4980 [Verdon Walker <VWalker@novell.com>, Steve Henson] 4981 4982 *) Add the possibility to control engines through control names but with 4983 arbitrary arguments instead of just a string. 4984 Change the key loaders to take a UI_METHOD instead of a callback 4985 function pointer. NOTE: this breaks binary compatibility with earlier 4986 versions of OpenSSL [engine]. 4987 Adapt the nCipher code for these new conditions and add a card insertion 4988 callback. 4989 [Richard Levitte] 4990 4991 *) Enhance the general user interface with mechanisms to better support 4992 dialog box interfaces, application-defined prompts, the possibility 4993 to use defaults (for example default passwords from somewhere else) 4994 and interrupts/cancellations. 4995 [Richard Levitte] 4996 4997 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4998 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4999 [Steve Henson] 5000 5001 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 5002 tidy up some unnecessarily weird code in 'sk_new()'). 5003 [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 5004 5005 *) Change the key loading routines for ENGINEs to use the same kind 5006 callback (pem_password_cb) as all other routines that need this 5007 kind of callback. 5008 [Richard Levitte] 5009 5010 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 5011 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 5012 than this minimum value is recommended. 5013 [Lutz Jaenicke] 5014 5015 *) New random seeder for OpenVMS, using the system process statistics 5016 that are easily reachable. 5017 [Richard Levitte] 5018 5019 *) Windows apparently can't transparently handle global 5020 variables defined in DLLs. Initialisations such as: 5021 5022 const ASN1_ITEM *it = &ASN1_INTEGER_it; 5023 5024 wont compile. This is used by the any applications that need to 5025 declare their own ASN1 modules. This was fixed by adding the option 5026 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 5027 needed for static libraries under Win32. 5028 [Steve Henson] 5029 5030 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 5031 setting of purpose and trust fields. New X509_STORE trust and 5032 purpose functions and tidy up setting in other SSL functions. 5033 [Steve Henson] 5034 5035 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 5036 structure. These are inherited by X509_STORE_CTX when it is 5037 initialised. This allows various defaults to be set in the 5038 X509_STORE structure (such as flags for CRL checking and custom 5039 purpose or trust settings) for functions which only use X509_STORE_CTX 5040 internally such as S/MIME. 5041 5042 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 5043 trust settings if they are not set in X509_STORE. This allows X509_STORE 5044 purposes and trust (in S/MIME for example) to override any set by default. 5045 5046 Add command line options for CRL checking to smime, s_client and s_server 5047 applications. 5048 [Steve Henson] 5049 5050 *) Initial CRL based revocation checking. If the CRL checking flag(s) 5051 are set then the CRL is looked up in the X509_STORE structure and 5052 its validity and signature checked, then if the certificate is found 5053 in the CRL the verify fails with a revoked error. 5054 5055 Various new CRL related callbacks added to X509_STORE_CTX structure. 5056 5057 Command line options added to 'verify' application to support this. 5058 5059 This needs some additional work, such as being able to handle multiple 5060 CRLs with different times, extension based lookup (rather than just 5061 by subject name) and ultimately more complete V2 CRL extension 5062 handling. 5063 [Steve Henson] 5064 5065 *) Add a general user interface API (crypto/ui/). This is designed 5066 to replace things like des_read_password and friends (backward 5067 compatibility functions using this new API are provided). 5068 The purpose is to remove prompting functions from the DES code 5069 section as well as provide for prompting through dialog boxes in 5070 a window system and the like. 5071 [Richard Levitte] 5072 5073 *) Add "ex_data" support to ENGINE so implementations can add state at a 5074 per-structure level rather than having to store it globally. 5075 [Geoff] 5076 5077 *) Make it possible for ENGINE structures to be copied when retrieved by 5078 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 5079 This causes the "original" ENGINE structure to act like a template, 5080 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 5081 operational state can be localised to each ENGINE structure, despite the 5082 fact they all share the same "methods". New ENGINE structures returned in 5083 this case have no functional references and the return value is the single 5084 structural reference. This matches the single structural reference returned 5085 by ENGINE_by_id() normally, when it is incremented on the pre-existing 5086 ENGINE structure. 5087 [Geoff] 5088 5089 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 5090 needs to match any other type at all we need to manually clear the 5091 tag cache. 5092 [Steve Henson] 5093 5094 *) Changes to the "openssl engine" utility to include; 5095 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 5096 about an ENGINE's available control commands. 5097 - executing control commands from command line arguments using the 5098 '-pre' and '-post' switches. '-post' is only used if '-t' is 5099 specified and the ENGINE is successfully initialised. The syntax for 5100 the individual commands are colon-separated, for example; 5101 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 5102 [Geoff] 5103 5104 *) New dynamic control command support for ENGINEs. ENGINEs can now 5105 declare their own commands (numbers), names (strings), descriptions, 5106 and input types for run-time discovery by calling applications. A 5107 subset of these commands are implicitly classed as "executable" 5108 depending on their input type, and only these can be invoked through 5109 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 5110 can be based on user input, config files, etc). The distinction is 5111 that "executable" commands cannot return anything other than a boolean 5112 result and can only support numeric or string input, whereas some 5113 discoverable commands may only be for direct use through 5114 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 5115 pointers, or other custom uses. The "executable" commands are to 5116 support parameterisations of ENGINE behaviour that can be 5117 unambiguously defined by ENGINEs and used consistently across any 5118 OpenSSL-based application. Commands have been added to all the 5119 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 5120 control over shared-library paths without source code alterations. 5121 [Geoff] 5122 5123 *) Changed all ENGINE implementations to dynamically allocate their 5124 ENGINEs rather than declaring them statically. Apart from this being 5125 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 5126 this also allows the implementations to compile without using the 5127 internal engine_int.h header. 5128 [Geoff] 5129 5130 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 5131 'const' value. Any code that should be able to modify a RAND_METHOD 5132 should already have non-const pointers to it (ie. they should only 5133 modify their own ones). 5134 [Geoff] 5135 5136 *) Made a variety of little tweaks to the ENGINE code. 5137 - "atalla" and "ubsec" string definitions were moved from header files 5138 to C code. "nuron" string definitions were placed in variables 5139 rather than hard-coded - allowing parameterisation of these values 5140 later on via ctrl() commands. 5141 - Removed unused "#if 0"'d code. 5142 - Fixed engine list iteration code so it uses ENGINE_free() to release 5143 structural references. 5144 - Constified the RAND_METHOD element of ENGINE structures. 5145 - Constified various get/set functions as appropriate and added 5146 missing functions (including a catch-all ENGINE_cpy that duplicates 5147 all ENGINE values onto a new ENGINE except reference counts/state). 5148 - Removed NULL parameter checks in get/set functions. Setting a method 5149 or function to NULL is a way of cancelling out a previously set 5150 value. Passing a NULL ENGINE parameter is just plain stupid anyway 5151 and doesn't justify the extra error symbols and code. 5152 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5153 flags from engine_int.h to engine.h. 5154 - Changed prototypes for ENGINE handler functions (init(), finish(), 5155 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5156 [Geoff] 5157 5158 *) Implement binary inversion algorithm for BN_mod_inverse in addition 5159 to the algorithm using long division. The binary algorithm can be 5160 used only if the modulus is odd. On 32-bit systems, it is faster 5161 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 5162 roughly 5-15% for 256-bit moduli), so we use it only for moduli 5163 up to 450 bits. In 64-bit environments, the binary algorithm 5164 appears to be advantageous for much longer moduli; here we use it 5165 for moduli up to 2048 bits. 5166 [Bodo Moeller] 5167 5168 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 5169 could not support the combine flag in choice fields. 5170 [Steve Henson] 5171 5172 *) Add a 'copy_extensions' option to the 'ca' utility. This copies 5173 extensions from a certificate request to the certificate. 5174 [Steve Henson] 5175 5176 *) Allow multiple 'certopt' and 'nameopt' options to be separated 5177 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 5178 file: this allows the display of the certificate about to be 5179 signed to be customised, to allow certain fields to be included 5180 or excluded and extension details. The old system didn't display 5181 multicharacter strings properly, omitted fields not in the policy 5182 and couldn't display additional details such as extensions. 5183 [Steve Henson] 5184 5185 *) Function EC_POINTs_mul for multiple scalar multiplication 5186 of an arbitrary number of elliptic curve points 5187 \sum scalars[i]*points[i], 5188 optionally including the generator defined for the EC_GROUP: 5189 scalar*generator + \sum scalars[i]*points[i]. 5190 5191 EC_POINT_mul is a simple wrapper function for the typical case 5192 that the point list has just one item (besides the optional 5193 generator). 5194 [Bodo Moeller] 5195 5196 *) First EC_METHODs for curves over GF(p): 5197 5198 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5199 operations and provides various method functions that can also 5200 operate with faster implementations of modular arithmetic. 5201 5202 EC_GFp_mont_method() reuses most functions that are part of 5203 EC_GFp_simple_method, but uses Montgomery arithmetic. 5204 5205 [Bodo Moeller; point addition and point doubling 5206 implementation directly derived from source code provided by 5207 Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5208 5209 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5210 crypto/ec/ec_lib.c): 5211 5212 Curves are EC_GROUP objects (with an optional group generator) 5213 based on EC_METHODs that are built into the library. 5214 5215 Points are EC_POINT objects based on EC_GROUP objects. 5216 5217 Most of the framework would be able to handle curves over arbitrary 5218 finite fields, but as there are no obvious types for fields other 5219 than GF(p), some functions are limited to that for now. 5220 [Bodo Moeller] 5221 5222 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5223 that the file contains a complete HTTP response. 5224 [Richard Levitte] 5225 5226 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5227 change the def and num file printf format specifier from "%-40sXXX" 5228 to "%-39s XXX". The latter will always guarantee a space after the 5229 field while the former will cause them to run together if the field 5230 is 40 of more characters long. 5231 [Steve Henson] 5232 5233 *) Constify the cipher and digest 'method' functions and structures 5234 and modify related functions to take constant EVP_MD and EVP_CIPHER 5235 pointers. 5236 [Steve Henson] 5237 5238 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5239 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5240 [Bodo Moeller] 5241 5242 *) Modify EVP_Digest*() routines so they now return values. Although the 5243 internal software routines can never fail additional hardware versions 5244 might. 5245 [Steve Henson] 5246 5247 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5248 5249 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5250 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5251 5252 ASN1 error codes 5253 ERR_R_NESTED_ASN1_ERROR 5254 ... 5255 ERR_R_MISSING_ASN1_EOS 5256 were 4 .. 9, conflicting with 5257 ERR_LIB_RSA (= ERR_R_RSA_LIB) 5258 ... 5259 ERR_LIB_PEM (= ERR_R_PEM_LIB). 5260 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5261 5262 Add new error code 'ERR_R_INTERNAL_ERROR'. 5263 [Bodo Moeller] 5264 5265 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5266 suffices. 5267 [Bodo Moeller] 5268 5269 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5270 sets the subject name for a new request or supersedes the 5271 subject name in a given request. Formats that can be parsed are 5272 'CN=Some Name, OU=myOU, C=IT' 5273 and 5274 'CN=Some Name/OU=myOU/C=IT'. 5275 5276 Add options '-batch' and '-verbose' to 'openssl req'. 5277 [Massimiliano Pala <madwolf@hackmasters.net>] 5278 5279 *) Introduce the possibility to access global variables through 5280 functions on platform were that's the best way to handle exporting 5281 global variables in shared libraries. To enable this functionality, 5282 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5283 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5284 is normally done by Configure or something similar). 5285 5286 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5287 in the source file (foo.c) like this: 5288 5289 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5290 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5291 5292 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5293 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5294 5295 OPENSSL_DECLARE_GLOBAL(int,foo); 5296 #define foo OPENSSL_GLOBAL_REF(foo) 5297 OPENSSL_DECLARE_GLOBAL(double,bar); 5298 #define bar OPENSSL_GLOBAL_REF(bar) 5299 5300 The #defines are very important, and therefore so is including the 5301 header file everywhere where the defined globals are used. 5302 5303 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5304 of ASN.1 items, but that structure is a bit different. 5305 5306 The largest change is in util/mkdef.pl which has been enhanced with 5307 better and easier to understand logic to choose which symbols should 5308 go into the Windows .def files as well as a number of fixes and code 5309 cleanup (among others, algorithm keywords are now sorted 5310 lexicographically to avoid constant rewrites). 5311 [Richard Levitte] 5312 5313 *) In BN_div() keep a copy of the sign of 'num' before writing the 5314 result to 'rm' because if rm==num the value will be overwritten 5315 and produce the wrong result if 'num' is negative: this caused 5316 problems with BN_mod() and BN_nnmod(). 5317 [Steve Henson] 5318 5319 *) Function OCSP_request_verify(). This checks the signature on an 5320 OCSP request and verifies the signer certificate. The signer 5321 certificate is just checked for a generic purpose and OCSP request 5322 trust settings. 5323 [Steve Henson] 5324 5325 *) Add OCSP_check_validity() function to check the validity of OCSP 5326 responses. OCSP responses are prepared in real time and may only 5327 be a few seconds old. Simply checking that the current time lies 5328 between thisUpdate and nextUpdate max reject otherwise valid responses 5329 caused by either OCSP responder or client clock inaccuracy. Instead 5330 we allow thisUpdate and nextUpdate to fall within a certain period of 5331 the current time. The age of the response can also optionally be 5332 checked. Two new options -validity_period and -status_age added to 5333 ocsp utility. 5334 [Steve Henson] 5335 5336 *) If signature or public key algorithm is unrecognized print out its 5337 OID rather that just UNKNOWN. 5338 [Steve Henson] 5339 5340 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5341 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5342 ID to be generated from the issuer certificate alone which can then be 5343 passed to OCSP_id_issuer_cmp(). 5344 [Steve Henson] 5345 5346 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5347 ASN1 modules to export functions returning ASN1_ITEM pointers 5348 instead of the ASN1_ITEM structures themselves. This adds several 5349 new macros which allow the underlying ASN1 function/structure to 5350 be accessed transparently. As a result code should not use ASN1_ITEM 5351 references directly (such as &X509_it) but instead use the relevant 5352 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5353 use of the new ASN1 code on platforms where exporting structures 5354 is problematical (for example in shared libraries) but exporting 5355 functions returning pointers to structures is not. 5356 [Steve Henson] 5357 5358 *) Add support for overriding the generation of SSL/TLS session IDs. 5359 These callbacks can be registered either in an SSL_CTX or per SSL. 5360 The purpose of this is to allow applications to control, if they wish, 5361 the arbitrary values chosen for use as session IDs, particularly as it 5362 can be useful for session caching in multiple-server environments. A 5363 command-line switch for testing this (and any client code that wishes 5364 to use such a feature) has been added to "s_server". 5365 [Geoff Thorpe, Lutz Jaenicke] 5366 5367 *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5368 of the form '#if defined(...) || defined(...) || ...' and 5369 '#if !defined(...) && !defined(...) && ...'. This also avoids 5370 the growing number of special cases it was previously handling. 5371 [Richard Levitte] 5372 5373 *) Make all configuration macros available for application by making 5374 sure they are available in opensslconf.h, by giving them names starting 5375 with "OPENSSL_" to avoid conflicts with other packages and by making 5376 sure e_os2.h will cover all platform-specific cases together with 5377 opensslconf.h. 5378 Additionally, it is now possible to define configuration/platform- 5379 specific names (called "system identities"). In the C code, these 5380 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5381 macro with the name beginning with "OPENSSL_SYS_", which is determined 5382 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5383 what is available. 5384 [Richard Levitte] 5385 5386 *) New option -set_serial to 'req' and 'x509' this allows the serial 5387 number to use to be specified on the command line. Previously self 5388 signed certificates were hard coded with serial number 0 and the 5389 CA options of 'x509' had to use a serial number in a file which was 5390 auto incremented. 5391 [Steve Henson] 5392 5393 *) New options to 'ca' utility to support V2 CRL entry extensions. 5394 Currently CRL reason, invalidity date and hold instruction are 5395 supported. Add new CRL extensions to V3 code and some new objects. 5396 [Steve Henson] 5397 5398 *) New function EVP_CIPHER_CTX_set_padding() this is used to 5399 disable standard block padding (aka PKCS#5 padding) in the EVP 5400 API, which was previously mandatory. This means that the data is 5401 not padded in any way and so the total length much be a multiple 5402 of the block size, otherwise an error occurs. 5403 [Steve Henson] 5404 5405 *) Initial (incomplete) OCSP SSL support. 5406 [Steve Henson] 5407 5408 *) New function OCSP_parse_url(). This splits up a URL into its host, 5409 port and path components: primarily to parse OCSP URLs. New -url 5410 option to ocsp utility. 5411 [Steve Henson] 5412 5413 *) New nonce behavior. The return value of OCSP_check_nonce() now 5414 reflects the various checks performed. Applications can decide 5415 whether to tolerate certain situations such as an absent nonce 5416 in a response when one was present in a request: the ocsp application 5417 just prints out a warning. New function OCSP_add1_basic_nonce() 5418 this is to allow responders to include a nonce in a response even if 5419 the request is nonce-less. 5420 [Steve Henson] 5421 5422 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5423 skipped when using openssl x509 multiple times on a single input file, 5424 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5425 [Bodo Moeller] 5426 5427 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5428 set string type: to handle setting ASN1_TIME structures. Fix ca 5429 utility to correctly initialize revocation date of CRLs. 5430 [Steve Henson] 5431 5432 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5433 the clients preferred ciphersuites and rather use its own preferences. 5434 Should help to work around M$ SGC (Server Gated Cryptography) bug in 5435 Internet Explorer by ensuring unchanged hash method during stepup. 5436 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5437 [Lutz Jaenicke] 5438 5439 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5440 to aes and add a new 'exist' option to print out symbols that don't 5441 appear to exist. 5442 [Steve Henson] 5443 5444 *) Additional options to ocsp utility to allow flags to be set and 5445 additional certificates supplied. 5446 [Steve Henson] 5447 5448 *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5449 OCSP client a number of certificate to only verify the response 5450 signature against. 5451 [Richard Levitte] 5452 5453 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5454 handle the new API. Currently only ECB, CBC modes supported. Add new 5455 AES OIDs. 5456 5457 Add TLS AES ciphersuites as described in RFC3268, "Advanced 5458 Encryption Standard (AES) Ciphersuites for Transport Layer 5459 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5460 not enabled by default and were not part of the "ALL" ciphersuite 5461 alias because they were not yet official; they could be 5462 explicitly requested by specifying the "AESdraft" ciphersuite 5463 group alias. In the final release of OpenSSL 0.9.7, the group 5464 alias is called "AES" and is part of "ALL".) 5465 [Ben Laurie, Steve Henson, Bodo Moeller] 5466 5467 *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5468 request to response. 5469 [Steve Henson] 5470 5471 *) Functions for OCSP responders. OCSP_request_onereq_count(), 5472 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5473 extract information from a certificate request. OCSP_response_create() 5474 creates a response and optionally adds a basic response structure. 5475 OCSP_basic_add1_status() adds a complete single response to a basic 5476 response and returns the OCSP_SINGLERESP structure just added (to allow 5477 extensions to be included for example). OCSP_basic_add1_cert() adds a 5478 certificate to a basic response and OCSP_basic_sign() signs a basic 5479 response with various flags. New helper functions ASN1_TIME_check() 5480 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5481 (converts ASN1_TIME to GeneralizedTime). 5482 [Steve Henson] 5483 5484 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5485 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5486 structure from a certificate. X509_pubkey_digest() digests the public_key 5487 contents: this is used in various key identifiers. 5488 [Steve Henson] 5489 5490 *) Make sk_sort() tolerate a NULL argument. 5491 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5492 5493 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5494 passed by the function are trusted implicitly. If any of them signed the 5495 response then it is assumed to be valid and is not verified. 5496 [Steve Henson] 5497 5498 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5499 to data. This was previously part of the PKCS7 ASN1 code. This 5500 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5501 [Steve Henson, reported by Kenneth R. Robinette 5502 <support@securenetterm.com>] 5503 5504 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5505 routines: without these tracing memory leaks is very painful. 5506 Fix leaks in PKCS12 and PKCS7 routines. 5507 [Steve Henson] 5508 5509 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5510 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5511 effectively meant GeneralizedTime would never be used. Now it 5512 is initialised to -1 but X509_time_adj() now has to check the value 5513 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5514 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5515 [Steve Henson, reported by Kenneth R. Robinette 5516 <support@securenetterm.com>] 5517 5518 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5519 result in a zero length in the ASN1_INTEGER structure which was 5520 not consistent with the structure when d2i_ASN1_INTEGER() was used 5521 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5522 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5523 where it did not print out a minus for negative ASN1_INTEGER. 5524 [Steve Henson] 5525 5526 *) Add summary printout to ocsp utility. The various functions which 5527 convert status values to strings have been renamed to: 5528 OCSP_response_status_str(), OCSP_cert_status_str() and 5529 OCSP_crl_reason_str() and are no longer static. New options 5530 to verify nonce values and to disable verification. OCSP response 5531 printout format cleaned up. 5532 [Steve Henson] 5533 5534 *) Add additional OCSP certificate checks. These are those specified 5535 in RFC2560. This consists of two separate checks: the CA of the 5536 certificate being checked must either be the OCSP signer certificate 5537 or the issuer of the OCSP signer certificate. In the latter case the 5538 OCSP signer certificate must contain the OCSP signing extended key 5539 usage. This check is performed by attempting to match the OCSP 5540 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5541 in the OCSP_CERTID structures of the response. 5542 [Steve Henson] 5543 5544 *) Initial OCSP certificate verification added to OCSP_basic_verify() 5545 and related routines. This uses the standard OpenSSL certificate 5546 verify routines to perform initial checks (just CA validity) and 5547 to obtain the certificate chain. Then additional checks will be 5548 performed on the chain. Currently the root CA is checked to see 5549 if it is explicitly trusted for OCSP signing. This is used to set 5550 a root CA as a global signing root: that is any certificate that 5551 chains to that CA is an acceptable OCSP signing certificate. 5552 [Steve Henson] 5553 5554 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5555 extensions from a separate configuration file. 5556 As when reading extensions from the main configuration file, 5557 the '-extensions ...' option may be used for specifying the 5558 section to use. 5559 [Massimiliano Pala <madwolf@comune.modena.it>] 5560 5561 *) New OCSP utility. Allows OCSP requests to be generated or 5562 read. The request can be sent to a responder and the output 5563 parsed, outputed or printed in text form. Not complete yet: 5564 still needs to check the OCSP response validity. 5565 [Steve Henson] 5566 5567 *) New subcommands for 'openssl ca': 5568 'openssl ca -status <serial>' prints the status of the cert with 5569 the given serial number (according to the index file). 5570 'openssl ca -updatedb' updates the expiry status of certificates 5571 in the index file. 5572 [Massimiliano Pala <madwolf@comune.modena.it>] 5573 5574 *) New '-newreq-nodes' command option to CA.pl. This is like 5575 '-newreq', but calls 'openssl req' with the '-nodes' option 5576 so that the resulting key is not encrypted. 5577 [Damien Miller <djm@mindrot.org>] 5578 5579 *) New configuration for the GNU Hurd. 5580 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5581 5582 *) Initial code to implement OCSP basic response verify. This 5583 is currently incomplete. Currently just finds the signer's 5584 certificate and verifies the signature on the response. 5585 [Steve Henson] 5586 5587 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5588 value of OPENSSLDIR. This is available via the new '-d' option 5589 to 'openssl version', and is also included in 'openssl version -a'. 5590 [Bodo Moeller] 5591 5592 *) Allowing defining memory allocation callbacks that will be given 5593 file name and line number information in additional arguments 5594 (a const char* and an int). The basic functionality remains, as 5595 well as the original possibility to just replace malloc(), 5596 realloc() and free() by functions that do not know about these 5597 additional arguments. To register and find out the current 5598 settings for extended allocation functions, the following 5599 functions are provided: 5600 5601 CRYPTO_set_mem_ex_functions 5602 CRYPTO_set_locked_mem_ex_functions 5603 CRYPTO_get_mem_ex_functions 5604 CRYPTO_get_locked_mem_ex_functions 5605 5606 These work the same way as CRYPTO_set_mem_functions and friends. 5607 CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5608 extended allocation function is enabled. 5609 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5610 a conventional allocation function is enabled. 5611 [Richard Levitte, Bodo Moeller] 5612 5613 *) Finish off removing the remaining LHASH function pointer casts. 5614 There should no longer be any prototype-casting required when using 5615 the LHASH abstraction, and any casts that remain are "bugs". See 5616 the callback types and macros at the head of lhash.h for details 5617 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5618 [Geoff Thorpe] 5619 5620 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5621 If /dev/[u]random devices are not available or do not return enough 5622 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5623 be queried. 5624 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5625 /etc/entropy will be queried once each in this sequence, quering stops 5626 when enough entropy was collected without querying more sockets. 5627 [Lutz Jaenicke] 5628 5629 *) Change the Unix RAND_poll() variant to be able to poll several 5630 random devices, as specified by DEVRANDOM, until a sufficient amount 5631 of data has been collected. We spend at most 10 ms on each file 5632 (select timeout) and read in non-blocking mode. DEVRANDOM now 5633 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5634 (previously it was just the string "/dev/urandom"), so on typical 5635 platforms the 10 ms delay will never occur. 5636 Also separate out the Unix variant to its own file, rand_unix.c. 5637 For VMS, there's a currently-empty rand_vms.c. 5638 [Richard Levitte] 5639 5640 *) Move OCSP client related routines to ocsp_cl.c. These 5641 provide utility functions which an application needing 5642 to issue a request to an OCSP responder and analyse the 5643 response will typically need: as opposed to those which an 5644 OCSP responder itself would need which will be added later. 5645 5646 OCSP_request_sign() signs an OCSP request with an API similar 5647 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5648 response. OCSP_response_get1_basic() extracts basic response 5649 from response. OCSP_resp_find_status(): finds and extracts status 5650 information from an OCSP_CERTID structure (which will be created 5651 when the request structure is built). These are built from lower 5652 level functions which work on OCSP_SINGLERESP structures but 5653 wont normally be used unless the application wishes to examine 5654 extensions in the OCSP response for example. 5655 5656 Replace nonce routines with a pair of functions. 5657 OCSP_request_add1_nonce() adds a nonce value and optionally 5658 generates a random value. OCSP_check_nonce() checks the 5659 validity of the nonce in an OCSP response. 5660 [Steve Henson] 5661 5662 *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5663 This doesn't copy the supplied OCSP_CERTID and avoids the 5664 need to free up the newly created id. Change return type 5665 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5666 This can then be used to add extensions to the request. 5667 Deleted OCSP_request_new(), since most of its functionality 5668 is now in OCSP_REQUEST_new() (and the case insensitive name 5669 clash) apart from the ability to set the request name which 5670 will be added elsewhere. 5671 [Steve Henson] 5672 5673 *) Update OCSP API. Remove obsolete extensions argument from 5674 various functions. Extensions are now handled using the new 5675 OCSP extension code. New simple OCSP HTTP function which 5676 can be used to send requests and parse the response. 5677 [Steve Henson] 5678 5679 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5680 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5681 uses the special reorder version of SET OF to sort the attributes 5682 and reorder them to match the encoded order. This resolves a long 5683 standing problem: a verify on a PKCS7 structure just after signing 5684 it used to fail because the attribute order did not match the 5685 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5686 it uses the received order. This is necessary to tolerate some broken 5687 software that does not order SET OF. This is handled by encoding 5688 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5689 to produce the required SET OF. 5690 [Steve Henson] 5691 5692 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5693 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5694 files to get correct declarations of the ASN.1 item variables. 5695 [Richard Levitte] 5696 5697 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5698 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5699 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5700 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5701 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5702 ASN1_ITEM and no wrapper functions. 5703 [Steve Henson] 5704 5705 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5706 replace the old function pointer based I/O routines. Change most of 5707 the *_d2i_bio() and *_d2i_fp() functions to use these. 5708 [Steve Henson] 5709 5710 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5711 lines, recognice more "algorithms" that can be deselected, and make 5712 it complain about algorithm deselection that isn't recognised. 5713 [Richard Levitte] 5714 5715 *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5716 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5717 to use new functions. Add NO_ASN1_OLD which can be set to remove 5718 some old style ASN1 functions: this can be used to determine if old 5719 code will still work when these eventually go away. 5720 [Steve Henson] 5721 5722 *) New extension functions for OCSP structures, these follow the 5723 same conventions as certificates and CRLs. 5724 [Steve Henson] 5725 5726 *) New function X509V3_add1_i2d(). This automatically encodes and 5727 adds an extension. Its behaviour can be customised with various 5728 flags to append, replace or delete. Various wrappers added for 5729 certifcates and CRLs. 5730 [Steve Henson] 5731 5732 *) Fix to avoid calling the underlying ASN1 print routine when 5733 an extension cannot be parsed. Correct a typo in the 5734 OCSP_SERVICELOC extension. Tidy up print OCSP format. 5735 [Steve Henson] 5736 5737 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5738 entries for variables. 5739 [Steve Henson] 5740 5741 *) Add functionality to apps/openssl.c for detecting locking 5742 problems: As the program is single-threaded, all we have 5743 to do is register a locking callback using an array for 5744 storing which locks are currently held by the program. 5745 [Bodo Moeller] 5746 5747 *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5748 SSL_get_ex_data_X509_STORE_idx(), which is used in 5749 ssl_verify_cert_chain() and thus can be called at any time 5750 during TLS/SSL handshakes so that thread-safety is essential. 5751 Unfortunately, the ex_data design is not at all suited 5752 for multi-threaded use, so it probably should be abolished. 5753 [Bodo Moeller] 5754 5755 *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5756 [Broadcom, tweaked and integrated by Geoff Thorpe] 5757 5758 *) Move common extension printing code to new function 5759 X509V3_print_extensions(). Reorganise OCSP print routines and 5760 implement some needed OCSP ASN1 functions. Add OCSP extensions. 5761 [Steve Henson] 5762 5763 *) New function X509_signature_print() to remove duplication in some 5764 print routines. 5765 [Steve Henson] 5766 5767 *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5768 set (this was treated exactly the same as SET OF previously). This 5769 is used to reorder the STACK representing the structure to match the 5770 encoding. This will be used to get round a problem where a PKCS7 5771 structure which was signed could not be verified because the STACK 5772 order did not reflect the encoded order. 5773 [Steve Henson] 5774 5775 *) Reimplement the OCSP ASN1 module using the new code. 5776 [Steve Henson] 5777 5778 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5779 for its ASN1 operations. The old style function pointers still exist 5780 for now but they will eventually go away. 5781 [Steve Henson] 5782 5783 *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5784 completely replaces the old ASN1 functionality with a table driven 5785 encoder and decoder which interprets an ASN1_ITEM structure describing 5786 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5787 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5788 has also been converted to the new form. 5789 [Steve Henson] 5790 5791 *) Change BN_mod_exp_recp so that negative moduli are tolerated 5792 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5793 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5794 for negative moduli. 5795 [Bodo Moeller] 5796 5797 *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5798 of not touching the result's sign bit. 5799 [Bodo Moeller] 5800 5801 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5802 set. 5803 [Bodo Moeller] 5804 5805 *) Changed the LHASH code to use prototypes for callbacks, and created 5806 macros to declare and implement thin (optionally static) functions 5807 that provide type-safety and avoid function pointer casting for the 5808 type-specific callbacks. 5809 [Geoff Thorpe] 5810 5811 *) Added Kerberos Cipher Suites to be used with TLS, as written in 5812 RFC 2712. 5813 [Veers Staats <staatsvr@asc.hpc.mil>, 5814 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5815 5816 *) Reformat the FAQ so the different questions and answers can be divided 5817 in sections depending on the subject. 5818 [Richard Levitte] 5819 5820 *) Have the zlib compression code load ZLIB.DLL dynamically under 5821 Windows. 5822 [Richard Levitte] 5823 5824 *) New function BN_mod_sqrt for computing square roots modulo a prime 5825 (using the probabilistic Tonelli-Shanks algorithm unless 5826 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5827 be handled deterministically). 5828 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5829 5830 *) Make BN_mod_inverse faster by explicitly handling small quotients 5831 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5832 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5833 [Bodo Moeller] 5834 5835 *) New function BN_kronecker. 5836 [Bodo Moeller] 5837 5838 *) Fix BN_gcd so that it works on negative inputs; the result is 5839 positive unless both parameters are zero. 5840 Previously something reasonably close to an infinite loop was 5841 possible because numbers could be growing instead of shrinking 5842 in the implementation of Euclid's algorithm. 5843 [Bodo Moeller] 5844 5845 *) Fix BN_is_word() and BN_is_one() macros to take into account the 5846 sign of the number in question. 5847 5848 Fix BN_is_word(a,w) to work correctly for w == 0. 5849 5850 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5851 because its test if the absolute value of 'a' equals 'w'. 5852 Note that BN_abs_is_word does *not* handle w == 0 reliably; 5853 it exists mostly for use in the implementations of BN_is_zero(), 5854 BN_is_one(), and BN_is_word(). 5855 [Bodo Moeller] 5856 5857 *) New function BN_swap. 5858 [Bodo Moeller] 5859 5860 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5861 the exponentiation functions are more likely to produce reasonable 5862 results on negative inputs. 5863 [Bodo Moeller] 5864 5865 *) Change BN_mod_mul so that the result is always non-negative. 5866 Previously, it could be negative if one of the factors was negative; 5867 I don't think anyone really wanted that behaviour. 5868 [Bodo Moeller] 5869 5870 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5871 (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5872 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5873 and add new functions: 5874 5875 BN_nnmod 5876 BN_mod_sqr 5877 BN_mod_add 5878 BN_mod_add_quick 5879 BN_mod_sub 5880 BN_mod_sub_quick 5881 BN_mod_lshift1 5882 BN_mod_lshift1_quick 5883 BN_mod_lshift 5884 BN_mod_lshift_quick 5885 5886 These functions always generate non-negative results. 5887 5888 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5889 such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5890 5891 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5892 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5893 be reduced modulo m. 5894 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5895 5896#if 0 5897 The following entry accidentily appeared in the CHANGES file 5898 distributed with OpenSSL 0.9.7. The modifications described in 5899 it do *not* apply to OpenSSL 0.9.7. 5900 5901 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5902 was actually never needed) and in BN_mul(). The removal in BN_mul() 5903 required a small change in bn_mul_part_recursive() and the addition 5904 of the functions bn_cmp_part_words(), bn_sub_part_words() and 5905 bn_add_part_words(), which do the same thing as bn_cmp_words(), 5906 bn_sub_words() and bn_add_words() except they take arrays with 5907 differing sizes. 5908 [Richard Levitte] 5909#endif 5910 5911 *) In 'openssl passwd', verify passwords read from the terminal 5912 unless the '-salt' option is used (which usually means that 5913 verification would just waste user's time since the resulting 5914 hash is going to be compared with some given password hash) 5915 or the new '-noverify' option is used. 5916 5917 This is an incompatible change, but it does not affect 5918 non-interactive use of 'openssl passwd' (passwords on the command 5919 line, '-stdin' option, '-in ...' option) and thus should not 5920 cause any problems. 5921 [Bodo Moeller] 5922 5923 *) Remove all references to RSAref, since there's no more need for it. 5924 [Richard Levitte] 5925 5926 *) Make DSO load along a path given through an environment variable 5927 (SHLIB_PATH) with shl_load(). 5928 [Richard Levitte] 5929 5930 *) Constify the ENGINE code as a result of BIGNUM constification. 5931 Also constify the RSA code and most things related to it. In a 5932 few places, most notable in the depth of the ASN.1 code, ugly 5933 casts back to non-const were required (to be solved at a later 5934 time) 5935 [Richard Levitte] 5936 5937 *) Make it so the openssl application has all engines loaded by default. 5938 [Richard Levitte] 5939 5940 *) Constify the BIGNUM routines a little more. 5941 [Richard Levitte] 5942 5943 *) Add the following functions: 5944 5945 ENGINE_load_cswift() 5946 ENGINE_load_chil() 5947 ENGINE_load_atalla() 5948 ENGINE_load_nuron() 5949 ENGINE_load_builtin_engines() 5950 5951 That way, an application can itself choose if external engines that 5952 are built-in in OpenSSL shall ever be used or not. The benefit is 5953 that applications won't have to be linked with libdl or other dso 5954 libraries unless it's really needed. 5955 5956 Changed 'openssl engine' to load all engines on demand. 5957 Changed the engine header files to avoid the duplication of some 5958 declarations (they differed!). 5959 [Richard Levitte] 5960 5961 *) 'openssl engine' can now list capabilities. 5962 [Richard Levitte] 5963 5964 *) Better error reporting in 'openssl engine'. 5965 [Richard Levitte] 5966 5967 *) Never call load_dh_param(NULL) in s_server. 5968 [Bodo Moeller] 5969 5970 *) Add engine application. It can currently list engines by name and 5971 identity, and test if they are actually available. 5972 [Richard Levitte] 5973 5974 *) Improve RPM specification file by forcing symbolic linking and making 5975 sure the installed documentation is also owned by root.root. 5976 [Damien Miller <djm@mindrot.org>] 5977 5978 *) Give the OpenSSL applications more possibilities to make use of 5979 keys (public as well as private) handled by engines. 5980 [Richard Levitte] 5981 5982 *) Add OCSP code that comes from CertCo. 5983 [Richard Levitte] 5984 5985 *) Add VMS support for the Rijndael code. 5986 [Richard Levitte] 5987 5988 *) Added untested support for Nuron crypto accelerator. 5989 [Ben Laurie] 5990 5991 *) Add support for external cryptographic devices. This code was 5992 previously distributed separately as the "engine" branch. 5993 [Geoff Thorpe, Richard Levitte] 5994 5995 *) Rework the filename-translation in the DSO code. It is now possible to 5996 have far greater control over how a "name" is turned into a filename 5997 depending on the operating environment and any oddities about the 5998 different shared library filenames on each system. 5999 [Geoff Thorpe] 6000 6001 *) Support threads on FreeBSD-elf in Configure. 6002 [Richard Levitte] 6003 6004 *) Fix for SHA1 assembly problem with MASM: it produces 6005 warnings about corrupt line number information when assembling 6006 with debugging information. This is caused by the overlapping 6007 of two sections. 6008 [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 6009 6010 *) NCONF changes. 6011 NCONF_get_number() has no error checking at all. As a replacement, 6012 NCONF_get_number_e() is defined (_e for "error checking") and is 6013 promoted strongly. The old NCONF_get_number is kept around for 6014 binary backward compatibility. 6015 Make it possible for methods to load from something other than a BIO, 6016 by providing a function pointer that is given a name instead of a BIO. 6017 For example, this could be used to load configuration data from an 6018 LDAP server. 6019 [Richard Levitte] 6020 6021 *) Fix for non blocking accept BIOs. Added new I/O special reason 6022 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 6023 with non blocking I/O was not possible because no retry code was 6024 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 6025 this case. 6026 [Steve Henson] 6027 6028 *) Added the beginnings of Rijndael support. 6029 [Ben Laurie] 6030 6031 *) Fix for bug in DirectoryString mask setting. Add support for 6032 X509_NAME_print_ex() in 'req' and X509_print_ex() function 6033 to allow certificate printing to more controllable, additional 6034 'certopt' option to 'x509' to allow new printing options to be 6035 set. 6036 [Steve Henson] 6037 6038 *) Clean old EAY MD5 hack from e_os.h. 6039 [Richard Levitte] 6040 6041 Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 6042 6043 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 6044 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 6045 [Joe Orton, Steve Henson] 6046 6047 Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 6048 6049 *) Fix additional bug revealed by the NISCC test suite: 6050 6051 Stop bug triggering large recursion when presented with 6052 certain ASN.1 tags (CVE-2003-0851) 6053 [Steve Henson] 6054 6055 Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 6056 6057 *) Fix various bugs revealed by running the NISCC test suite: 6058 6059 Stop out of bounds reads in the ASN1 code when presented with 6060 invalid tags (CVE-2003-0543 and CVE-2003-0544). 6061 6062 If verify callback ignores invalid public key errors don't try to check 6063 certificate signature with the NULL public key. 6064 6065 [Steve Henson] 6066 6067 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 6068 if the server requested one: as stated in TLS 1.0 and SSL 3.0 6069 specifications. 6070 [Steve Henson] 6071 6072 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 6073 extra data after the compression methods not only for TLS 1.0 6074 but also for SSL 3.0 (as required by the specification). 6075 [Bodo Moeller; problem pointed out by Matthias Loepfe] 6076 6077 *) Change X509_certificate_type() to mark the key as exported/exportable 6078 when it's 512 *bits* long, not 512 bytes. 6079 [Richard Levitte] 6080 6081 Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 6082 6083 *) Countermeasure against the Klima-Pokorny-Rosa extension of 6084 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 6085 a protocol version number mismatch like a decryption error 6086 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 6087 [Bodo Moeller] 6088 6089 *) Turn on RSA blinding by default in the default implementation 6090 to avoid a timing attack. Applications that don't want it can call 6091 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 6092 They would be ill-advised to do so in most cases. 6093 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 6094 6095 *) Change RSA blinding code so that it works when the PRNG is not 6096 seeded (in this case, the secret RSA exponent is abused as 6097 an unpredictable seed -- if it is not unpredictable, there 6098 is no point in blinding anyway). Make RSA blinding thread-safe 6099 by remembering the creator's thread ID in rsa->blinding and 6100 having all other threads use local one-time blinding factors 6101 (this requires more computation than sharing rsa->blinding, but 6102 avoids excessive locking; and if an RSA object is not shared 6103 between threads, blinding will still be very fast). 6104 [Bodo Moeller] 6105 6106 Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 6107 6108 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 6109 via timing by performing a MAC computation even if incorrrect 6110 block cipher padding has been found. This is a countermeasure 6111 against active attacks where the attacker has to distinguish 6112 between bad padding and a MAC verification error. (CVE-2003-0078) 6113 6114 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 6115 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 6116 Martin Vuagnoux (EPFL, Ilion)] 6117 6118 Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 6119 6120 *) New function OPENSSL_cleanse(), which is used to cleanse a section of 6121 memory from it's contents. This is done with a counter that will 6122 place alternating values in each byte. This can be used to solve 6123 two issues: 1) the removal of calls to memset() by highly optimizing 6124 compilers, and 2) cleansing with other values than 0, since those can 6125 be read through on certain media, for example a swap space on disk. 6126 [Geoff Thorpe] 6127 6128 *) Bugfix: client side session caching did not work with external caching, 6129 because the session->cipher setting was not restored when reloading 6130 from the external cache. This problem was masked, when 6131 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 6132 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 6133 [Lutz Jaenicke] 6134 6135 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6136 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6137 [Zeev Lieber <zeev-l@yahoo.com>] 6138 6139 *) Undo an undocumented change introduced in 0.9.6e which caused 6140 repeated calls to OpenSSL_add_all_ciphers() and 6141 OpenSSL_add_all_digests() to be ignored, even after calling 6142 EVP_cleanup(). 6143 [Richard Levitte] 6144 6145 *) Change the default configuration reader to deal with last line not 6146 being properly terminated. 6147 [Richard Levitte] 6148 6149 *) Change X509_NAME_cmp() so it applies the special rules on handling 6150 DN values that are of type PrintableString, as well as RDNs of type 6151 emailAddress where the value has the type ia5String. 6152 [stefank@valicert.com via Richard Levitte] 6153 6154 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6155 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6156 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6157 the bitwise-OR of the two for use by the majority of applications 6158 wanting this behaviour, and update the docs. The documented 6159 behaviour and actual behaviour were inconsistent and had been 6160 changing anyway, so this is more a bug-fix than a behavioural 6161 change. 6162 [Geoff Thorpe, diagnosed by Nadav Har'El] 6163 6164 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 6165 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 6166 [Bodo Moeller] 6167 6168 *) Fix initialization code race conditions in 6169 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 6170 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 6171 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 6172 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 6173 ssl2_get_cipher_by_char(), 6174 ssl3_get_cipher_by_char(). 6175 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 6176 6177 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 6178 the cached sessions are flushed, as the remove_cb() might use ex_data 6179 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 6180 (see [openssl.org #212]). 6181 [Geoff Thorpe, Lutz Jaenicke] 6182 6183 *) Fix typo in OBJ_txt2obj which incorrectly passed the content 6184 length, instead of the encoding length to d2i_ASN1_OBJECT. 6185 [Steve Henson] 6186 6187 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6188 6189 *) [In 0.9.6g-engine release:] 6190 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6191 [Lynn Gazis <lgazis@rainbow.com>] 6192 6193 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6194 6195 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6196 and get fix the header length calculation. 6197 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6198 Alon Kantor <alonk@checkpoint.com> (and others), 6199 Steve Henson] 6200 6201 *) Use proper error handling instead of 'assertions' in buffer 6202 overflow checks added in 0.9.6e. This prevents DoS (the 6203 assertions could call abort()). 6204 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6205 6206 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6207 6208 *) Add various sanity checks to asn1_get_length() to reject 6209 the ASN1 length bytes if they exceed sizeof(long), will appear 6210 negative or the content length exceeds the length of the 6211 supplied buffer. 6212 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6213 6214 *) Fix cipher selection routines: ciphers without encryption had no flags 6215 for the cipher strength set and where therefore not handled correctly 6216 by the selection routines (PR #130). 6217 [Lutz Jaenicke] 6218 6219 *) Fix EVP_dsa_sha macro. 6220 [Nils Larsch] 6221 6222 *) New option 6223 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6224 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6225 that was added in OpenSSL 0.9.6d. 6226 6227 As the countermeasure turned out to be incompatible with some 6228 broken SSL implementations, the new option is part of SSL_OP_ALL. 6229 SSL_OP_ALL is usually employed when compatibility with weird SSL 6230 implementations is desired (e.g. '-bugs' option to 's_client' and 6231 's_server'), so the new option is automatically set in many 6232 applications. 6233 [Bodo Moeller] 6234 6235 *) Changes in security patch: 6236 6237 Changes marked "(CHATS)" were sponsored by the Defense Advanced 6238 Research Projects Agency (DARPA) and Air Force Research Laboratory, 6239 Air Force Materiel Command, USAF, under agreement number 6240 F30602-01-2-0537. 6241 6242 *) Add various sanity checks to asn1_get_length() to reject 6243 the ASN1 length bytes if they exceed sizeof(long), will appear 6244 negative or the content length exceeds the length of the 6245 supplied buffer. (CVE-2002-0659) 6246 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6247 6248 *) Assertions for various potential buffer overflows, not known to 6249 happen in practice. 6250 [Ben Laurie (CHATS)] 6251 6252 *) Various temporary buffers to hold ASCII versions of integers were 6253 too small for 64 bit platforms. (CVE-2002-0655) 6254 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6255 6256 *) Remote buffer overflow in SSL3 protocol - an attacker could 6257 supply an oversized session ID to a client. (CVE-2002-0656) 6258 [Ben Laurie (CHATS)] 6259 6260 *) Remote buffer overflow in SSL2 protocol - an attacker could 6261 supply an oversized client master key. (CVE-2002-0656) 6262 [Ben Laurie (CHATS)] 6263 6264 Changes between 0.9.6c and 0.9.6d [9 May 2002] 6265 6266 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6267 encoded as NULL) with id-dsa-with-sha1. 6268 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6269 6270 *) Check various X509_...() return values in apps/req.c. 6271 [Nils Larsch <nla@trustcenter.de>] 6272 6273 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6274 an end-of-file condition would erronously be flagged, when the CRLF 6275 was just at the end of a processed block. The bug was discovered when 6276 processing data through a buffering memory BIO handing the data to a 6277 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6278 <ptsekov@syntrex.com> and Nedelcho Stanev. 6279 [Lutz Jaenicke] 6280 6281 *) Implement a countermeasure against a vulnerability recently found 6282 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6283 before application data chunks to avoid the use of known IVs 6284 with data potentially chosen by the attacker. 6285 [Bodo Moeller] 6286 6287 *) Fix length checks in ssl3_get_client_hello(). 6288 [Bodo Moeller] 6289 6290 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6291 to prevent ssl3_read_internal() from incorrectly assuming that 6292 ssl3_read_bytes() found application data while handshake 6293 processing was enabled when in fact s->s3->in_read_app_data was 6294 merely automatically cleared during the initial handshake. 6295 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6296 6297 *) Fix object definitions for Private and Enterprise: they were not 6298 recognized in their shortname (=lowercase) representation. Extend 6299 obj_dat.pl to issue an error when using undefined keywords instead 6300 of silently ignoring the problem (Svenning Sorensen 6301 <sss@sss.dnsalias.net>). 6302 [Lutz Jaenicke] 6303 6304 *) Fix DH_generate_parameters() so that it works for 'non-standard' 6305 generators, i.e. generators other than 2 and 5. (Previously, the 6306 code did not properly initialise the 'add' and 'rem' values to 6307 BN_generate_prime().) 6308 6309 In the new general case, we do not insist that 'generator' is 6310 actually a primitive root: This requirement is rather pointless; 6311 a generator of the order-q subgroup is just as good, if not 6312 better. 6313 [Bodo Moeller] 6314 6315 *) Map new X509 verification errors to alerts. Discovered and submitted by 6316 Tom Wu <tom@arcot.com>. 6317 [Lutz Jaenicke] 6318 6319 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6320 returning non-zero before the data has been completely received 6321 when using non-blocking I/O. 6322 [Bodo Moeller; problem pointed out by John Hughes] 6323 6324 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6325 [Ben Laurie, Lutz Jaenicke] 6326 6327 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6328 Yoram Zahavi <YoramZ@gilian.com>). 6329 [Lutz Jaenicke] 6330 6331 *) Add information about CygWin 1.3 and on, and preserve proper 6332 configuration for the versions before that. 6333 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6334 6335 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6336 check whether we deal with a copy of a session and do not delete from 6337 the cache in this case. Problem reported by "Izhar Shoshani Levi" 6338 <izhar@checkpoint.com>. 6339 [Lutz Jaenicke] 6340 6341 *) Do not store session data into the internal session cache, if it 6342 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6343 flag is set). Proposed by Aslam <aslam@funk.com>. 6344 [Lutz Jaenicke] 6345 6346 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6347 value is 0. 6348 [Richard Levitte] 6349 6350 *) [In 0.9.6d-engine release:] 6351 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6352 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6353 6354 *) Add the configuration target linux-s390x. 6355 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6356 6357 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6358 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6359 variable as an indication that a ClientHello message has been 6360 received. As the flag value will be lost between multiple 6361 invocations of ssl3_accept when using non-blocking I/O, the 6362 function may not be aware that a handshake has actually taken 6363 place, thus preventing a new session from being added to the 6364 session cache. 6365 6366 To avoid this problem, we now set s->new_session to 2 instead of 6367 using a local variable. 6368 [Lutz Jaenicke, Bodo Moeller] 6369 6370 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6371 if the SSL_R_LENGTH_MISMATCH error is detected. 6372 [Geoff Thorpe, Bodo Moeller] 6373 6374 *) New 'shared_ldflag' column in Configure platform table. 6375 [Richard Levitte] 6376 6377 *) Fix EVP_CIPHER_mode macro. 6378 ["Dan S. Camper" <dan@bti.net>] 6379 6380 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6381 type, we must throw them away by setting rr->length to 0. 6382 [D P Chang <dpc@qualys.com>] 6383 6384 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 6385 6386 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 6387 <Dominikus.Scherkl@biodata.com>. (The previous implementation 6388 worked incorrectly for those cases where range = 10..._2 and 6389 3*range is two bits longer than range.) 6390 [Bodo Moeller] 6391 6392 *) Only add signing time to PKCS7 structures if it is not already 6393 present. 6394 [Steve Henson] 6395 6396 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 6397 OBJ_ld_ce should be OBJ_id_ce. 6398 Also some ip-pda OIDs in crypto/objects/objects.txt were 6399 incorrect (cf. RFC 3039). 6400 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 6401 6402 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 6403 returns early because it has nothing to do. 6404 [Andy Schneider <andy.schneider@bjss.co.uk>] 6405 6406 *) [In 0.9.6c-engine release:] 6407 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 6408 [Andy Schneider <andy.schneider@bjss.co.uk>] 6409 6410 *) [In 0.9.6c-engine release:] 6411 Add support for Cryptographic Appliance's keyserver technology. 6412 (Use engine 'keyclient') 6413 [Cryptographic Appliances and Geoff Thorpe] 6414 6415 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 6416 is called via tools/c89.sh because arguments have to be 6417 rearranged (all '-L' options must appear before the first object 6418 modules). 6419 [Richard Shapiro <rshapiro@abinitio.com>] 6420 6421 *) [In 0.9.6c-engine release:] 6422 Add support for Broadcom crypto accelerator cards, backported 6423 from 0.9.7. 6424 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 6425 6426 *) [In 0.9.6c-engine release:] 6427 Add support for SureWare crypto accelerator cards from 6428 Baltimore Technologies. (Use engine 'sureware') 6429 [Baltimore Technologies and Mark Cox] 6430 6431 *) [In 0.9.6c-engine release:] 6432 Add support for crypto accelerator cards from Accelerated 6433 Encryption Processing, www.aep.ie. (Use engine 'aep') 6434 [AEP Inc. and Mark Cox] 6435 6436 *) Add a configuration entry for gcc on UnixWare. 6437 [Gary Benson <gbenson@redhat.com>] 6438 6439 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 6440 messages are stored in a single piece (fixed-length part and 6441 variable-length part combined) and fix various bugs found on the way. 6442 [Bodo Moeller] 6443 6444 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 6445 instead. BIO_gethostbyname() does not know what timeouts are 6446 appropriate, so entries would stay in cache even when they have 6447 become invalid. 6448 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 6449 6450 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 6451 faced with a pathologically small ClientHello fragment that does 6452 not contain client_version: Instead of aborting with an error, 6453 simply choose the highest available protocol version (i.e., 6454 TLS 1.0 unless it is disabled). In practice, ClientHello 6455 messages are never sent like this, but this change gives us 6456 strictly correct behaviour at least for TLS. 6457 [Bodo Moeller] 6458 6459 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 6460 never resets s->method to s->ctx->method when called from within 6461 one of the SSL handshake functions. 6462 [Bodo Moeller; problem pointed out by Niko Baric] 6463 6464 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 6465 (sent using the client's version number) if client_version is 6466 smaller than the protocol version in use. Also change 6467 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 6468 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 6469 the client will at least see that alert. 6470 [Bodo Moeller] 6471 6472 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 6473 correctly. 6474 [Bodo Moeller] 6475 6476 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 6477 client receives HelloRequest while in a handshake. 6478 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 6479 6480 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 6481 should end in 'break', not 'goto end' which circuments various 6482 cleanups done in state SSL_ST_OK. But session related stuff 6483 must be disabled for SSL_ST_OK in the case that we just sent a 6484 HelloRequest. 6485 6486 Also avoid some overhead by not calling ssl_init_wbio_buffer() 6487 before just sending a HelloRequest. 6488 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 6489 6490 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 6491 reveal whether illegal block cipher padding was found or a MAC 6492 verification error occured. (Neither SSLerr() codes nor alerts 6493 are directly visible to potential attackers, but the information 6494 may leak via logfiles.) 6495 6496 Similar changes are not required for the SSL 2.0 implementation 6497 because the number of padding bytes is sent in clear for SSL 2.0, 6498 and the extra bytes are just ignored. However ssl/s2_pkt.c 6499 failed to verify that the purported number of padding bytes is in 6500 the legal range. 6501 [Bodo Moeller] 6502 6503 *) Add OpenUNIX-8 support including shared libraries 6504 (Boyd Lynn Gerber <gerberb@zenez.com>). 6505 [Lutz Jaenicke] 6506 6507 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 6508 'wristwatch attack' using huge encoding parameters (cf. 6509 James H. Manger's CRYPTO 2001 paper). Note that the 6510 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 6511 encoding parameters and hence was not vulnerable. 6512 [Bodo Moeller] 6513 6514 *) BN_sqr() bug fix. 6515 [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 6516 6517 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 6518 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 6519 followed by modular reduction. 6520 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 6521 6522 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 6523 equivalent based on BN_pseudo_rand() instead of BN_rand(). 6524 [Bodo Moeller] 6525 6526 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 6527 This function was broken, as the check for a new client hello message 6528 to handle SGC did not allow these large messages. 6529 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 6530 [Lutz Jaenicke] 6531 6532 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 6533 [Lutz Jaenicke] 6534 6535 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 6536 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 6537 [Lutz Jaenicke] 6538 6539 *) Rework the configuration and shared library support for Tru64 Unix. 6540 The configuration part makes use of modern compiler features and 6541 still retains old compiler behavior for those that run older versions 6542 of the OS. The shared library support part includes a variant that 6543 uses the RPATH feature, and is available through the special 6544 configuration target "alpha-cc-rpath", which will never be selected 6545 automatically. 6546 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 6547 6548 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 6549 with the same message size as in ssl3_get_certificate_request(). 6550 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 6551 messages might inadvertently be reject as too long. 6552 [Petr Lampa <lampa@fee.vutbr.cz>] 6553 6554 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 6555 [Andy Polyakov] 6556 6557 *) Modified SSL library such that the verify_callback that has been set 6558 specificly for an SSL object with SSL_set_verify() is actually being 6559 used. Before the change, a verify_callback set with this function was 6560 ignored and the verify_callback() set in the SSL_CTX at the time of 6561 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 6562 to allow the necessary settings. 6563 [Lutz Jaenicke] 6564 6565 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 6566 explicitly to NULL, as at least on Solaris 8 this seems not always to be 6567 done automatically (in contradiction to the requirements of the C 6568 standard). This made problems when used from OpenSSH. 6569 [Lutz Jaenicke] 6570 6571 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 6572 dh->length and always used 6573 6574 BN_rand_range(priv_key, dh->p). 6575 6576 BN_rand_range() is not necessary for Diffie-Hellman, and this 6577 specific range makes Diffie-Hellman unnecessarily inefficient if 6578 dh->length (recommended exponent length) is much smaller than the 6579 length of dh->p. We could use BN_rand_range() if the order of 6580 the subgroup was stored in the DH structure, but we only have 6581 dh->length. 6582 6583 So switch back to 6584 6585 BN_rand(priv_key, l, ...) 6586 6587 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 6588 otherwise. 6589 [Bodo Moeller] 6590 6591 *) In 6592 6593 RSA_eay_public_encrypt 6594 RSA_eay_private_decrypt 6595 RSA_eay_private_encrypt (signing) 6596 RSA_eay_public_decrypt (signature verification) 6597 6598 (default implementations for RSA_public_encrypt, 6599 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 6600 always reject numbers >= n. 6601 [Bodo Moeller] 6602 6603 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 6604 to synchronize access to 'locking_thread'. This is necessary on 6605 systems where access to 'locking_thread' (an 'unsigned long' 6606 variable) is not atomic. 6607 [Bodo Moeller] 6608 6609 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 6610 *before* setting the 'crypto_lock_rand' flag. The previous code had 6611 a race condition if 0 is a valid thread ID. 6612 [Travis Vitek <vitek@roguewave.com>] 6613 6614 *) Add support for shared libraries under Irix. 6615 [Albert Chin-A-Young <china@thewrittenword.com>] 6616 6617 *) Add configuration option to build on Linux on both big-endian and 6618 little-endian MIPS. 6619 [Ralf Baechle <ralf@uni-koblenz.de>] 6620 6621 *) Add the possibility to create shared libraries on HP-UX. 6622 [Richard Levitte] 6623 6624 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 6625 6626 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 6627 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 6628 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 6629 PRNG state recovery was possible based on the output of 6630 one PRNG request appropriately sized to gain knowledge on 6631 'md' followed by enough consecutive 1-byte PRNG requests 6632 to traverse all of 'state'. 6633 6634 1. When updating 'md_local' (the current thread's copy of 'md') 6635 during PRNG output generation, hash all of the previous 6636 'md_local' value, not just the half used for PRNG output. 6637 6638 2. Make the number of bytes from 'state' included into the hash 6639 independent from the number of PRNG bytes requested. 6640 6641 The first measure alone would be sufficient to avoid 6642 Markku-Juhani's attack. (Actually it had never occurred 6643 to me that the half of 'md_local' used for chaining was the 6644 half from which PRNG output bytes were taken -- I had always 6645 assumed that the secret half would be used.) The second 6646 measure makes sure that additional data from 'state' is never 6647 mixed into 'md_local' in small portions; this heuristically 6648 further strengthens the PRNG. 6649 [Bodo Moeller] 6650 6651 *) Fix crypto/bn/asm/mips3.s. 6652 [Andy Polyakov] 6653 6654 *) When only the key is given to "enc", the IV is undefined. Print out 6655 an error message in this case. 6656 [Lutz Jaenicke] 6657 6658 *) Handle special case when X509_NAME is empty in X509 printing routines. 6659 [Steve Henson] 6660 6661 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 6662 positive and less than q. 6663 [Bodo Moeller] 6664 6665 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 6666 used: it isn't thread safe and the add_lock_callback should handle 6667 that itself. 6668 [Paul Rose <Paul.Rose@bridge.com>] 6669 6670 *) Verify that incoming data obeys the block size in 6671 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 6672 [Bodo Moeller] 6673 6674 *) Fix OAEP check. 6675 [Ulf M�ller, Bodo M�ller] 6676 6677 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 6678 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 6679 when fixing the server behaviour for backwards-compatible 'client 6680 hello' messages. (Note that the attack is impractical against 6681 SSL 3.0 and TLS 1.0 anyway because length and version checking 6682 means that the probability of guessing a valid ciphertext is 6683 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 6684 paper.) 6685 6686 Before 0.9.5, the countermeasure (hide the error by generating a 6687 random 'decryption result') did not work properly because 6688 ERR_clear_error() was missing, meaning that SSL_get_error() would 6689 detect the supposedly ignored error. 6690 6691 Both problems are now fixed. 6692 [Bodo Moeller] 6693 6694 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 6695 (previously it was 1024). 6696 [Bodo Moeller] 6697 6698 *) Fix for compatibility mode trust settings: ignore trust settings 6699 unless some valid trust or reject settings are present. 6700 [Steve Henson] 6701 6702 *) Fix for blowfish EVP: its a variable length cipher. 6703 [Steve Henson] 6704 6705 *) Fix various bugs related to DSA S/MIME verification. Handle missing 6706 parameters in DSA public key structures and return an error in the 6707 DSA routines if parameters are absent. 6708 [Steve Henson] 6709 6710 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 6711 in the current directory if neither $RANDFILE nor $HOME was set. 6712 RAND_file_name() in 0.9.6a returned NULL in this case. This has 6713 caused some confusion to Windows users who haven't defined $HOME. 6714 Thus RAND_file_name() is changed again: e_os.h can define a 6715 DEFAULT_HOME, which will be used if $HOME is not set. 6716 For Windows, we use "C:"; on other platforms, we still require 6717 environment variables. 6718 6719 *) Move 'if (!initialized) RAND_poll()' into regions protected by 6720 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 6721 having multiple threads call RAND_poll() concurrently. 6722 [Bodo Moeller] 6723 6724 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 6725 combination of a flag and a thread ID variable. 6726 Otherwise while one thread is in ssleay_rand_bytes (which sets the 6727 flag), *other* threads can enter ssleay_add_bytes without obeying 6728 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 6729 that they do not hold after the first thread unsets add_do_not_lock). 6730 [Bodo Moeller] 6731 6732 *) Change bctest again: '-x' expressions are not available in all 6733 versions of 'test'. 6734 [Bodo Moeller] 6735 6736 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 6737 6738 *) Fix a couple of memory leaks in PKCS7_dataDecode() 6739 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 6740 6741 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 6742 the default extension for executables, if any. Also, make the perl 6743 scripts that use symlink() to test if it really exists and use "cp" 6744 if it doesn't. All this made OpenSSL compilable and installable in 6745 CygWin. 6746 [Richard Levitte] 6747 6748 *) Fix for asn1_GetSequence() for indefinite length constructed data. 6749 If SEQUENCE is length is indefinite just set c->slen to the total 6750 amount of data available. 6751 [Steve Henson, reported by shige@FreeBSD.org] 6752 [This change does not apply to 0.9.7.] 6753 6754 *) Change bctest to avoid here-documents inside command substitution 6755 (workaround for FreeBSD /bin/sh bug). 6756 For compatibility with Ultrix, avoid shell functions (introduced 6757 in the bctest version that searches along $PATH). 6758 [Bodo Moeller] 6759 6760 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 6761 with des_encrypt() defined on some operating systems, like Solaris 6762 and UnixWare. 6763 [Richard Levitte] 6764 6765 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 6766 On the Importance of Eliminating Errors in Cryptographic 6767 Computations, J. Cryptology 14 (2001) 2, 101-119, 6768 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 6769 [Ulf Moeller] 6770 6771 *) MIPS assembler BIGNUM division bug fix. 6772 [Andy Polyakov] 6773 6774 *) Disabled incorrect Alpha assembler code. 6775 [Richard Levitte] 6776 6777 *) Fix PKCS#7 decode routines so they correctly update the length 6778 after reading an EOC for the EXPLICIT tag. 6779 [Steve Henson] 6780 [This change does not apply to 0.9.7.] 6781 6782 *) Fix bug in PKCS#12 key generation routines. This was triggered 6783 if a 3DES key was generated with a 0 initial byte. Include 6784 PKCS12_BROKEN_KEYGEN compilation option to retain the old 6785 (but broken) behaviour. 6786 [Steve Henson] 6787 6788 *) Enhance bctest to search for a working bc along $PATH and print 6789 it when found. 6790 [Tim Rice <tim@multitalents.net> via Richard Levitte] 6791 6792 *) Fix memory leaks in err.c: free err_data string if necessary; 6793 don't write to the wrong index in ERR_set_error_data. 6794 [Bodo Moeller] 6795 6796 *) Implement ssl23_peek (analogous to ssl23_read), which previously 6797 did not exist. 6798 [Bodo Moeller] 6799 6800 *) Replace rdtsc with _emit statements for VC++ version 5. 6801 [Jeremy Cooper <jeremy@baymoo.org>] 6802 6803 *) Make it possible to reuse SSLv2 sessions. 6804 [Richard Levitte] 6805 6806 *) In copy_email() check for >= 0 as a return value for 6807 X509_NAME_get_index_by_NID() since 0 is a valid index. 6808 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 6809 6810 *) Avoid coredump with unsupported or invalid public keys by checking if 6811 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 6812 PKCS7_verify() fails with non detached data. 6813 [Steve Henson] 6814 6815 *) Don't use getenv in library functions when run as setuid/setgid. 6816 New function OPENSSL_issetugid(). 6817 [Ulf Moeller] 6818 6819 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 6820 due to incorrect handling of multi-threading: 6821 6822 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 6823 6824 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 6825 6826 3. Count how many times MemCheck_off() has been called so that 6827 nested use can be treated correctly. This also avoids 6828 inband-signalling in the previous code (which relied on the 6829 assumption that thread ID 0 is impossible). 6830 [Bodo Moeller] 6831 6832 *) Add "-rand" option also to s_client and s_server. 6833 [Lutz Jaenicke] 6834 6835 *) Fix CPU detection on Irix 6.x. 6836 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 6837 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 6838 6839 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 6840 was empty. 6841 [Steve Henson] 6842 [This change does not apply to 0.9.7.] 6843 6844 *) Use the cached encoding of an X509_NAME structure rather than 6845 copying it. This is apparently the reason for the libsafe "errors" 6846 but the code is actually correct. 6847 [Steve Henson] 6848 6849 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 6850 Bleichenbacher's DSA attack. 6851 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 6852 to be set and top=0 forces the highest bit to be set; top=-1 is new 6853 and leaves the highest bit random. 6854 [Ulf Moeller, Bodo Moeller] 6855 6856 *) In the NCONF_...-based implementations for CONF_... queries 6857 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 6858 a temporary CONF structure with the data component set to NULL 6859 (which gives segmentation faults in lh_retrieve). 6860 Instead, use NULL for the CONF pointer in CONF_get_string and 6861 CONF_get_number (which may use environment variables) and directly 6862 return NULL from CONF_get_section. 6863 [Bodo Moeller] 6864 6865 *) Fix potential buffer overrun for EBCDIC. 6866 [Ulf Moeller] 6867 6868 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 6869 keyUsage if basicConstraints absent for a CA. 6870 [Steve Henson] 6871 6872 *) Make SMIME_write_PKCS7() write mail header values with a format that 6873 is more generally accepted (no spaces before the semicolon), since 6874 some programs can't parse those values properly otherwise. Also make 6875 sure BIO's that break lines after each write do not create invalid 6876 headers. 6877 [Richard Levitte] 6878 6879 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 6880 macros previously used would not encode an empty SEQUENCE OF 6881 and break the signature. 6882 [Steve Henson] 6883 [This change does not apply to 0.9.7.] 6884 6885 *) Zero the premaster secret after deriving the master secret in 6886 DH ciphersuites. 6887 [Steve Henson] 6888 6889 *) Add some EVP_add_digest_alias registrations (as found in 6890 OpenSSL_add_all_digests()) to SSL_library_init() 6891 aka OpenSSL_add_ssl_algorithms(). This provides improved 6892 compatibility with peers using X.509 certificates 6893 with unconventional AlgorithmIdentifier OIDs. 6894 [Bodo Moeller] 6895 6896 *) Fix for Irix with NO_ASM. 6897 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 6898 6899 *) ./config script fixes. 6900 [Ulf Moeller, Richard Levitte] 6901 6902 *) Fix 'openssl passwd -1'. 6903 [Bodo Moeller] 6904 6905 *) Change PKCS12_key_gen_asc() so it can cope with non null 6906 terminated strings whose length is passed in the passlen 6907 parameter, for example from PEM callbacks. This was done 6908 by adding an extra length parameter to asc2uni(). 6909 [Steve Henson, reported by <oddissey@samsung.co.kr>] 6910 6911 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 6912 call failed, free the DSA structure. 6913 [Bodo Moeller] 6914 6915 *) Fix to uni2asc() to cope with zero length Unicode strings. 6916 These are present in some PKCS#12 files. 6917 [Steve Henson] 6918 6919 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 6920 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 6921 when writing a 32767 byte record. 6922 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 6923 6924 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 6925 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 6926 6927 (RSA objects have a reference count access to which is protected 6928 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 6929 so they are meant to be shared between threads.) 6930 [Bodo Moeller, Geoff Thorpe; original patch submitted by 6931 "Reddie, Steven" <Steven.Reddie@ca.com>] 6932 6933 *) Fix a deadlock in CRYPTO_mem_leaks(). 6934 [Bodo Moeller] 6935 6936 *) Use better test patterns in bntest. 6937 [Ulf M�ller] 6938 6939 *) rand_win.c fix for Borland C. 6940 [Ulf M�ller] 6941 6942 *) BN_rshift bugfix for n == 0. 6943 [Bodo Moeller] 6944 6945 *) Add a 'bctest' script that checks for some known 'bc' bugs 6946 so that 'make test' does not abort just because 'bc' is broken. 6947 [Bodo Moeller] 6948 6949 *) Store verify_result within SSL_SESSION also for client side to 6950 avoid potential security hole. (Re-used sessions on the client side 6951 always resulted in verify_result==X509_V_OK, not using the original 6952 result of the server certificate verification.) 6953 [Lutz Jaenicke] 6954 6955 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 6956 SSL3_RT_APPLICATION_DATA, return 0. 6957 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 6958 [Bodo Moeller] 6959 6960 *) Fix SSL_peek: 6961 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 6962 releases, have been re-implemented by renaming the previous 6963 implementations of ssl2_read and ssl3_read to ssl2_read_internal 6964 and ssl3_read_internal, respectively, and adding 'peek' parameters 6965 to them. The new ssl[23]_{read,peek} functions are calls to 6966 ssl[23]_read_internal with the 'peek' flag set appropriately. 6967 A 'peek' parameter has also been added to ssl3_read_bytes, which 6968 does the actual work for ssl3_read_internal. 6969 [Bodo Moeller] 6970 6971 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 6972 the method-specific "init()" handler. Also clean up ex_data after 6973 calling the method-specific "finish()" handler. Previously, this was 6974 happening the other way round. 6975 [Geoff Thorpe] 6976 6977 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 6978 The previous value, 12, was not always sufficient for BN_mod_exp(). 6979 [Bodo Moeller] 6980 6981 *) Make sure that shared libraries get the internal name engine with 6982 the full version number and not just 0. This should mark the 6983 shared libraries as not backward compatible. Of course, this should 6984 be changed again when we can guarantee backward binary compatibility. 6985 [Richard Levitte] 6986 6987 *) Fix typo in get_cert_by_subject() in by_dir.c 6988 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 6989 6990 *) Rework the system to generate shared libraries: 6991 6992 - Make note of the expected extension for the shared libraries and 6993 if there is a need for symbolic links from for example libcrypto.so.0 6994 to libcrypto.so.0.9.7. There is extended info in Configure for 6995 that. 6996 6997 - Make as few rebuilds of the shared libraries as possible. 6998 6999 - Still avoid linking the OpenSSL programs with the shared libraries. 7000 7001 - When installing, install the shared libraries separately from the 7002 static ones. 7003 [Richard Levitte] 7004 7005 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 7006 7007 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 7008 and not in SSL_clear because the latter is also used by the 7009 accept/connect functions; previously, the settings made by 7010 SSL_set_read_ahead would be lost during the handshake. 7011 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 7012 7013 *) Correct util/mkdef.pl to be selective about disabled algorithms. 7014 Previously, it would create entries for disableed algorithms no 7015 matter what. 7016 [Richard Levitte] 7017 7018 *) Added several new manual pages for SSL_* function. 7019 [Lutz Jaenicke] 7020 7021 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 7022 7023 *) In ssl23_get_client_hello, generate an error message when faced 7024 with an initial SSL 3.0/TLS record that is too small to contain the 7025 first two bytes of the ClientHello message, i.e. client_version. 7026 (Note that this is a pathologic case that probably has never happened 7027 in real life.) The previous approach was to use the version number 7028 from the record header as a substitute; but our protocol choice 7029 should not depend on that one because it is not authenticated 7030 by the Finished messages. 7031 [Bodo Moeller] 7032 7033 *) More robust randomness gathering functions for Windows. 7034 [Jeffrey Altman <jaltman@columbia.edu>] 7035 7036 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 7037 not set then we don't setup the error code for issuer check errors 7038 to avoid possibly overwriting other errors which the callback does 7039 handle. If an application does set the flag then we assume it knows 7040 what it is doing and can handle the new informational codes 7041 appropriately. 7042 [Steve Henson] 7043 7044 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 7045 a general "ANY" type, as such it should be able to decode anything 7046 including tagged types. However it didn't check the class so it would 7047 wrongly interpret tagged types in the same way as their universal 7048 counterpart and unknown types were just rejected. Changed so that the 7049 tagged and unknown types are handled in the same way as a SEQUENCE: 7050 that is the encoding is stored intact. There is also a new type 7051 "V_ASN1_OTHER" which is used when the class is not universal, in this 7052 case we have no idea what the actual type is so we just lump them all 7053 together. 7054 [Steve Henson] 7055 7056 *) On VMS, stdout may very well lead to a file that is written to 7057 in a record-oriented fashion. That means that every write() will 7058 write a separate record, which will be read separately by the 7059 programs trying to read from it. This can be very confusing. 7060 7061 The solution is to put a BIO filter in the way that will buffer 7062 text until a linefeed is reached, and then write everything a 7063 line at a time, so every record written will be an actual line, 7064 not chunks of lines and not (usually doesn't happen, but I've 7065 seen it once) several lines in one record. BIO_f_linebuffer() is 7066 the answer. 7067 7068 Currently, it's a VMS-only method, because that's where it has 7069 been tested well enough. 7070 [Richard Levitte] 7071 7072 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 7073 it can return incorrect results. 7074 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 7075 but it was in 0.9.6-beta[12].) 7076 [Bodo Moeller] 7077 7078 *) Disable the check for content being present when verifying detached 7079 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 7080 include zero length content when signing messages. 7081 [Steve Henson] 7082 7083 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 7084 BIO_ctrl (for BIO pairs). 7085 [Bodo M�ller] 7086 7087 *) Add DSO method for VMS. 7088 [Richard Levitte] 7089 7090 *) Bug fix: Montgomery multiplication could produce results with the 7091 wrong sign. 7092 [Ulf M�ller] 7093 7094 *) Add RPM specification openssl.spec and modify it to build three 7095 packages. The default package contains applications, application 7096 documentation and run-time libraries. The devel package contains 7097 include files, static libraries and function documentation. The 7098 doc package contains the contents of the doc directory. The original 7099 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 7100 [Richard Levitte] 7101 7102 *) Add a large number of documentation files for many SSL routines. 7103 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 7104 7105 *) Add a configuration entry for Sony News 4. 7106 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 7107 7108 *) Don't set the two most significant bits to one when generating a 7109 random number < q in the DSA library. 7110 [Ulf M�ller] 7111 7112 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 7113 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 7114 the underlying transport is blocking) if a handshake took place. 7115 (The default behaviour is needed by applications such as s_client 7116 and s_server that use select() to determine when to use SSL_read; 7117 but for applications that know in advance when to expect data, it 7118 just makes things more complicated.) 7119 [Bodo Moeller] 7120 7121 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 7122 from EGD. 7123 [Ben Laurie] 7124 7125 *) Add a few more EBCDIC conditionals that make `req' and `x509' 7126 work better on such systems. 7127 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 7128 7129 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 7130 Update PKCS12_parse() so it copies the friendlyName and the 7131 keyid to the certificates aux info. 7132 [Steve Henson] 7133 7134 *) Fix bug in PKCS7_verify() which caused an infinite loop 7135 if there was more than one signature. 7136 [Sven Uszpelkat <su@celocom.de>] 7137 7138 *) Major change in util/mkdef.pl to include extra information 7139 about each symbol, as well as presentig variables as well 7140 as functions. This change means that there's n more need 7141 to rebuild the .num files when some algorithms are excluded. 7142 [Richard Levitte] 7143 7144 *) Allow the verify time to be set by an application, 7145 rather than always using the current time. 7146 [Steve Henson] 7147 7148 *) Phase 2 verify code reorganisation. The certificate 7149 verify code now looks up an issuer certificate by a 7150 number of criteria: subject name, authority key id 7151 and key usage. It also verifies self signed certificates 7152 by the same criteria. The main comparison function is 7153 X509_check_issued() which performs these checks. 7154 7155 Lot of changes were necessary in order to support this 7156 without completely rewriting the lookup code. 7157 7158 Authority and subject key identifier are now cached. 7159 7160 The LHASH 'certs' is X509_STORE has now been replaced 7161 by a STACK_OF(X509_OBJECT). This is mainly because an 7162 LHASH can't store or retrieve multiple objects with 7163 the same hash value. 7164 7165 As a result various functions (which were all internal 7166 use only) have changed to handle the new X509_STORE 7167 structure. This will break anything that messed round 7168 with X509_STORE internally. 7169 7170 The functions X509_STORE_add_cert() now checks for an 7171 exact match, rather than just subject name. 7172 7173 The X509_STORE API doesn't directly support the retrieval 7174 of multiple certificates matching a given criteria, however 7175 this can be worked round by performing a lookup first 7176 (which will fill the cache with candidate certificates) 7177 and then examining the cache for matches. This is probably 7178 the best we can do without throwing out X509_LOOKUP 7179 entirely (maybe later...). 7180 7181 The X509_VERIFY_CTX structure has been enhanced considerably. 7182 7183 All certificate lookup operations now go via a get_issuer() 7184 callback. Although this currently uses an X509_STORE it 7185 can be replaced by custom lookups. This is a simple way 7186 to bypass the X509_STORE hackery necessary to make this 7187 work and makes it possible to use more efficient techniques 7188 in future. A very simple version which uses a simple 7189 STACK for its trusted certificate store is also provided 7190 using X509_STORE_CTX_trusted_stack(). 7191 7192 The verify_cb() and verify() callbacks now have equivalents 7193 in the X509_STORE_CTX structure. 7194 7195 X509_STORE_CTX also has a 'flags' field which can be used 7196 to customise the verify behaviour. 7197 [Steve Henson] 7198 7199 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 7200 excludes S/MIME capabilities. 7201 [Steve Henson] 7202 7203 *) When a certificate request is read in keep a copy of the 7204 original encoding of the signed data and use it when outputing 7205 again. Signatures then use the original encoding rather than 7206 a decoded, encoded version which may cause problems if the 7207 request is improperly encoded. 7208 [Steve Henson] 7209 7210 *) For consistency with other BIO_puts implementations, call 7211 buffer_write(b, ...) directly in buffer_puts instead of calling 7212 BIO_write(b, ...). 7213 7214 In BIO_puts, increment b->num_write as in BIO_write. 7215 [Peter.Sylvester@EdelWeb.fr] 7216 7217 *) Fix BN_mul_word for the case where the word is 0. (We have to use 7218 BN_zero, we may not return a BIGNUM with an array consisting of 7219 words set to zero.) 7220 [Bodo Moeller] 7221 7222 *) Avoid calling abort() from within the library when problems are 7223 detected, except if preprocessor symbols have been defined 7224 (such as REF_CHECK, BN_DEBUG etc.). 7225 [Bodo Moeller] 7226 7227 *) New openssl application 'rsautl'. This utility can be 7228 used for low level RSA operations. DER public key 7229 BIO/fp routines also added. 7230 [Steve Henson] 7231 7232 *) New Configure entry and patches for compiling on QNX 4. 7233 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 7234 7235 *) A demo state-machine implementation was sponsored by 7236 Nuron (http://www.nuron.com/) and is now available in 7237 demos/state_machine. 7238 [Ben Laurie] 7239 7240 *) New options added to the 'dgst' utility for signature 7241 generation and verification. 7242 [Steve Henson] 7243 7244 *) Unrecognized PKCS#7 content types are now handled via a 7245 catch all ASN1_TYPE structure. This allows unsupported 7246 types to be stored as a "blob" and an application can 7247 encode and decode it manually. 7248 [Steve Henson] 7249 7250 *) Fix various signed/unsigned issues to make a_strex.c 7251 compile under VC++. 7252 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 7253 7254 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 7255 length if passed a buffer. ASN1_INTEGER_to_BN failed 7256 if passed a NULL BN and its argument was negative. 7257 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 7258 7259 *) Modification to PKCS#7 encoding routines to output definite 7260 length encoding. Since currently the whole structures are in 7261 memory there's not real point in using indefinite length 7262 constructed encoding. However if OpenSSL is compiled with 7263 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 7264 [Steve Henson] 7265 7266 *) Added BIO_vprintf() and BIO_vsnprintf(). 7267 [Richard Levitte] 7268 7269 *) Added more prefixes to parse for in the the strings written 7270 through a logging bio, to cover all the levels that are available 7271 through syslog. The prefixes are now: 7272 7273 PANIC, EMERG, EMR => LOG_EMERG 7274 ALERT, ALR => LOG_ALERT 7275 CRIT, CRI => LOG_CRIT 7276 ERROR, ERR => LOG_ERR 7277 WARNING, WARN, WAR => LOG_WARNING 7278 NOTICE, NOTE, NOT => LOG_NOTICE 7279 INFO, INF => LOG_INFO 7280 DEBUG, DBG => LOG_DEBUG 7281 7282 and as before, if none of those prefixes are present at the 7283 beginning of the string, LOG_ERR is chosen. 7284 7285 On Win32, the LOG_* levels are mapped according to this: 7286 7287 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 7288 LOG_WARNING => EVENTLOG_WARNING_TYPE 7289 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 7290 7291 [Richard Levitte] 7292 7293 *) Made it possible to reconfigure with just the configuration 7294 argument "reconf" or "reconfigure". The command line arguments 7295 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 7296 and are retrieved from there when reconfiguring. 7297 [Richard Levitte] 7298 7299 *) MD4 implemented. 7300 [Assar Westerlund <assar@sics.se>, Richard Levitte] 7301 7302 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 7303 [Richard Levitte] 7304 7305 *) The obj_dat.pl script was messing up the sorting of object 7306 names. The reason was that it compared the quoted version 7307 of strings as a result "OCSP" > "OCSP Signing" because 7308 " > SPACE. Changed script to store unquoted versions of 7309 names and add quotes on output. It was also omitting some 7310 names from the lookup table if they were given a default 7311 value (that is if SN is missing it is given the same 7312 value as LN and vice versa), these are now added on the 7313 grounds that if an object has a name we should be able to 7314 look it up. Finally added warning output when duplicate 7315 short or long names are found. 7316 [Steve Henson] 7317 7318 *) Changes needed for Tandem NSK. 7319 [Scott Uroff <scott@xypro.com>] 7320 7321 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 7322 RSA_padding_check_SSLv23(), special padding was never detected 7323 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 7324 version rollback attacks was not effective. 7325 7326 In s23_clnt.c, don't use special rollback-attack detection padding 7327 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 7328 client; similarly, in s23_srvr.c, don't do the rollback check if 7329 SSL 2.0 is the only protocol enabled in the server. 7330 [Bodo Moeller] 7331 7332 *) Make it possible to get hexdumps of unprintable data with 'openssl 7333 asn1parse'. By implication, the functions ASN1_parse_dump() and 7334 BIO_dump_indent() are added. 7335 [Richard Levitte] 7336 7337 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 7338 these print out strings and name structures based on various 7339 flags including RFC2253 support and proper handling of 7340 multibyte characters. Added options to the 'x509' utility 7341 to allow the various flags to be set. 7342 [Steve Henson] 7343 7344 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 7345 Also change the functions X509_cmp_current_time() and 7346 X509_gmtime_adj() work with an ASN1_TIME structure, 7347 this will enable certificates using GeneralizedTime in validity 7348 dates to be checked. 7349 [Steve Henson] 7350 7351 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 7352 negative public key encodings) on by default, 7353 NO_NEG_PUBKEY_BUG can be set to disable it. 7354 [Steve Henson] 7355 7356 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 7357 content octets. An i2c_ASN1_OBJECT is unnecessary because 7358 the encoding can be trivially obtained from the structure. 7359 [Steve Henson] 7360 7361 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 7362 not read locks (CRYPTO_r_[un]lock). 7363 [Bodo Moeller] 7364 7365 *) A first attempt at creating official support for shared 7366 libraries through configuration. I've kept it so the 7367 default is static libraries only, and the OpenSSL programs 7368 are always statically linked for now, but there are 7369 preparations for dynamic linking in place. 7370 This has been tested on Linux and Tru64. 7371 [Richard Levitte] 7372 7373 *) Randomness polling function for Win9x, as described in: 7374 Peter Gutmann, Software Generation of Practically Strong 7375 Random Numbers. 7376 [Ulf M�ller] 7377 7378 *) Fix so PRNG is seeded in req if using an already existing 7379 DSA key. 7380 [Steve Henson] 7381 7382 *) New options to smime application. -inform and -outform 7383 allow alternative formats for the S/MIME message including 7384 PEM and DER. The -content option allows the content to be 7385 specified separately. This should allow things like Netscape 7386 form signing output easier to verify. 7387 [Steve Henson] 7388 7389 *) Fix the ASN1 encoding of tags using the 'long form'. 7390 [Steve Henson] 7391 7392 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 7393 STRING types. These convert content octets to and from the 7394 underlying type. The actual tag and length octets are 7395 already assumed to have been read in and checked. These 7396 are needed because all other string types have virtually 7397 identical handling apart from the tag. By having versions 7398 of the ASN1 functions that just operate on content octets 7399 IMPLICIT tagging can be handled properly. It also allows 7400 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 7401 and ASN1_INTEGER are identical apart from the tag. 7402 [Steve Henson] 7403 7404 *) Change the handling of OID objects as follows: 7405 7406 - New object identifiers are inserted in objects.txt, following 7407 the syntax given in objects.README. 7408 - objects.pl is used to process obj_mac.num and create a new 7409 obj_mac.h. 7410 - obj_dat.pl is used to create a new obj_dat.h, using the data in 7411 obj_mac.h. 7412 7413 This is currently kind of a hack, and the perl code in objects.pl 7414 isn't very elegant, but it works as I intended. The simplest way 7415 to check that it worked correctly is to look in obj_dat.h and 7416 check the array nid_objs and make sure the objects haven't moved 7417 around (this is important!). Additions are OK, as well as 7418 consistent name changes. 7419 [Richard Levitte] 7420 7421 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 7422 [Bodo Moeller] 7423 7424 *) Addition of the command line parameter '-rand file' to 'openssl req'. 7425 The given file adds to whatever has already been seeded into the 7426 random pool through the RANDFILE configuration file option or 7427 environment variable, or the default random state file. 7428 [Richard Levitte] 7429 7430 *) mkstack.pl now sorts each macro group into lexical order. 7431 Previously the output order depended on the order the files 7432 appeared in the directory, resulting in needless rewriting 7433 of safestack.h . 7434 [Steve Henson] 7435 7436 *) Patches to make OpenSSL compile under Win32 again. Mostly 7437 work arounds for the VC++ problem that it treats func() as 7438 func(void). Also stripped out the parts of mkdef.pl that 7439 added extra typesafe functions: these no longer exist. 7440 [Steve Henson] 7441 7442 *) Reorganisation of the stack code. The macros are now all 7443 collected in safestack.h . Each macro is defined in terms of 7444 a "stack macro" of the form SKM_<name>(type, a, b). The 7445 DEBUG_SAFESTACK is now handled in terms of function casts, 7446 this has the advantage of retaining type safety without the 7447 use of additional functions. If DEBUG_SAFESTACK is not defined 7448 then the non typesafe macros are used instead. Also modified the 7449 mkstack.pl script to handle the new form. Needs testing to see 7450 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 7451 the default if no major problems. Similar behaviour for ASN1_SET_OF 7452 and PKCS12_STACK_OF. 7453 [Steve Henson] 7454 7455 *) When some versions of IIS use the 'NET' form of private key the 7456 key derivation algorithm is different. Normally MD5(password) is 7457 used as a 128 bit RC4 key. In the modified case 7458 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 7459 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 7460 as the old Netscape_RSA functions except they have an additional 7461 'sgckey' parameter which uses the modified algorithm. Also added 7462 an -sgckey command line option to the rsa utility. Thanks to 7463 Adrian Peck <bertie@ncipher.com> for posting details of the modified 7464 algorithm to openssl-dev. 7465 [Steve Henson] 7466 7467 *) The evp_local.h macros were using 'c.##kname' which resulted in 7468 invalid expansion on some systems (SCO 5.0.5 for example). 7469 Corrected to 'c.kname'. 7470 [Phillip Porch <root@theporch.com>] 7471 7472 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 7473 a STACK of email addresses from a certificate or request, these look 7474 in the subject name and the subject alternative name extensions and 7475 omit any duplicate addresses. 7476 [Steve Henson] 7477 7478 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 7479 This makes DSA verification about 2 % faster. 7480 [Bodo Moeller] 7481 7482 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 7483 (meaning that now 2^5 values will be precomputed, which is only 4 KB 7484 plus overhead for 1024 bit moduli). 7485 This makes exponentiations about 0.5 % faster for 1024 bit 7486 exponents (as measured by "openssl speed rsa2048"). 7487 [Bodo Moeller] 7488 7489 *) Rename memory handling macros to avoid conflicts with other 7490 software: 7491 Malloc => OPENSSL_malloc 7492 Malloc_locked => OPENSSL_malloc_locked 7493 Realloc => OPENSSL_realloc 7494 Free => OPENSSL_free 7495 [Richard Levitte] 7496 7497 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 7498 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 7499 [Bodo Moeller] 7500 7501 *) CygWin32 support. 7502 [John Jarvie <jjarvie@newsguy.com>] 7503 7504 *) The type-safe stack code has been rejigged. It is now only compiled 7505 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 7506 by default all type-specific stack functions are "#define"d back to 7507 standard stack functions. This results in more streamlined output 7508 but retains the type-safety checking possibilities of the original 7509 approach. 7510 [Geoff Thorpe] 7511 7512 *) The STACK code has been cleaned up, and certain type declarations 7513 that didn't make a lot of sense have been brought in line. This has 7514 also involved a cleanup of sorts in safestack.h to more correctly 7515 map type-safe stack functions onto their plain stack counterparts. 7516 This work has also resulted in a variety of "const"ifications of 7517 lots of the code, especially "_cmp" operations which should normally 7518 be prototyped with "const" parameters anyway. 7519 [Geoff Thorpe] 7520 7521 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 7522 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 7523 (The PRNG state consists of two parts, the large pool 'state' and 'md', 7524 where all of 'md' is used each time the PRNG is used, but 'state' 7525 is used only indexed by a cyclic counter. As entropy may not be 7526 well distributed from the beginning, 'md' is important as a 7527 chaining variable. However, the output function chains only half 7528 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 7529 all of 'md', and seeding with STATE_SIZE dummy bytes will result 7530 in all of 'state' being rewritten, with the new values depending 7531 on virtually all of 'md'. This overcomes the 80 bit limitation.) 7532 [Bodo Moeller] 7533 7534 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 7535 the handshake is continued after ssl_verify_cert_chain(); 7536 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 7537 can lead to 'unexplainable' connection aborts later. 7538 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 7539 7540 *) Major EVP API cipher revision. 7541 Add hooks for extra EVP features. This allows various cipher 7542 parameters to be set in the EVP interface. Support added for variable 7543 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 7544 setting of RC2 and RC5 parameters. 7545 7546 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 7547 ciphers. 7548 7549 Remove lots of duplicated code from the EVP library. For example *every* 7550 cipher init() function handles the 'iv' in the same way according to the 7551 cipher mode. They also all do nothing if the 'key' parameter is NULL and 7552 for CFB and OFB modes they zero ctx->num. 7553 7554 New functionality allows removal of S/MIME code RC2 hack. 7555 7556 Most of the routines have the same form and so can be declared in terms 7557 of macros. 7558 7559 By shifting this to the top level EVP_CipherInit() it can be removed from 7560 all individual ciphers. If the cipher wants to handle IVs or keys 7561 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 7562 flags. 7563 7564 Change lots of functions like EVP_EncryptUpdate() to now return a 7565 value: although software versions of the algorithms cannot fail 7566 any installed hardware versions can. 7567 [Steve Henson] 7568 7569 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 7570 this option is set, tolerate broken clients that send the negotiated 7571 protocol version number instead of the requested protocol version 7572 number. 7573 [Bodo Moeller] 7574 7575 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 7576 i.e. non-zero for export ciphersuites, zero otherwise. 7577 Previous versions had this flag inverted, inconsistent with 7578 rsa_tmp_cb (..._TMP_RSA_CB). 7579 [Bodo Moeller; problem reported by Amit Chopra] 7580 7581 *) Add missing DSA library text string. Work around for some IIS 7582 key files with invalid SEQUENCE encoding. 7583 [Steve Henson] 7584 7585 *) Add a document (doc/standards.txt) that list all kinds of standards 7586 and so on that are implemented in OpenSSL. 7587 [Richard Levitte] 7588 7589 *) Enhance c_rehash script. Old version would mishandle certificates 7590 with the same subject name hash and wouldn't handle CRLs at all. 7591 Added -fingerprint option to crl utility, to support new c_rehash 7592 features. 7593 [Steve Henson] 7594 7595 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 7596 [Ulf M�ller] 7597 7598 *) Fix for SSL server purpose checking. Server checking was 7599 rejecting certificates which had extended key usage present 7600 but no ssl client purpose. 7601 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 7602 7603 *) Make PKCS#12 code work with no password. The PKCS#12 spec 7604 is a little unclear about how a blank password is handled. 7605 Since the password in encoded as a BMPString with terminating 7606 double NULL a zero length password would end up as just the 7607 double NULL. However no password at all is different and is 7608 handled differently in the PKCS#12 key generation code. NS 7609 treats a blank password as zero length. MSIE treats it as no 7610 password on export: but it will try both on import. We now do 7611 the same: PKCS12_parse() tries zero length and no password if 7612 the password is set to "" or NULL (NULL is now a valid password: 7613 it wasn't before) as does the pkcs12 application. 7614 [Steve Henson] 7615 7616 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 7617 perror when PEM_read_bio_X509_REQ fails, the error message must 7618 be obtained from the error queue. 7619 [Bodo Moeller] 7620 7621 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 7622 it in ERR_remove_state if appropriate, and change ERR_get_state 7623 accordingly to avoid race conditions (this is necessary because 7624 thread_hash is no longer constant once set). 7625 [Bodo Moeller] 7626 7627 *) Bugfix for linux-elf makefile.one. 7628 [Ulf M�ller] 7629 7630 *) RSA_get_default_method() will now cause a default 7631 RSA_METHOD to be chosen if one doesn't exist already. 7632 Previously this was only set during a call to RSA_new() 7633 or RSA_new_method(NULL) meaning it was possible for 7634 RSA_get_default_method() to return NULL. 7635 [Geoff Thorpe] 7636 7637 *) Added native name translation to the existing DSO code 7638 that will convert (if the flag to do so is set) filenames 7639 that are sufficiently small and have no path information 7640 into a canonical native form. Eg. "blah" converted to 7641 "libblah.so" or "blah.dll" etc. 7642 [Geoff Thorpe] 7643 7644 *) New function ERR_error_string_n(e, buf, len) which is like 7645 ERR_error_string(e, buf), but writes at most 'len' bytes 7646 including the 0 terminator. For ERR_error_string_n, 'buf' 7647 may not be NULL. 7648 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 7649 7650 *) CONF library reworked to become more general. A new CONF 7651 configuration file reader "class" is implemented as well as a 7652 new functions (NCONF_*, for "New CONF") to handle it. The now 7653 old CONF_* functions are still there, but are reimplemented to 7654 work in terms of the new functions. Also, a set of functions 7655 to handle the internal storage of the configuration data is 7656 provided to make it easier to write new configuration file 7657 reader "classes" (I can definitely see something reading a 7658 configuration file in XML format, for example), called _CONF_*, 7659 or "the configuration storage API"... 7660 7661 The new configuration file reading functions are: 7662 7663 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 7664 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 7665 7666 NCONF_default, NCONF_WIN32 7667 7668 NCONF_dump_fp, NCONF_dump_bio 7669 7670 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 7671 NCONF_new creates a new CONF object. This works in the same way 7672 as other interfaces in OpenSSL, like the BIO interface. 7673 NCONF_dump_* dump the internal storage of the configuration file, 7674 which is useful for debugging. All other functions take the same 7675 arguments as the old CONF_* functions wth the exception of the 7676 first that must be a `CONF *' instead of a `LHASH *'. 7677 7678 To make it easer to use the new classes with the old CONF_* functions, 7679 the function CONF_set_default_method is provided. 7680 [Richard Levitte] 7681 7682 *) Add '-tls1' option to 'openssl ciphers', which was already 7683 mentioned in the documentation but had not been implemented. 7684 (This option is not yet really useful because even the additional 7685 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 7686 [Bodo Moeller] 7687 7688 *) Initial DSO code added into libcrypto for letting OpenSSL (and 7689 OpenSSL-based applications) load shared libraries and bind to 7690 them in a portable way. 7691 [Geoff Thorpe, with contributions from Richard Levitte] 7692 7693 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 7694 7695 *) Make sure _lrotl and _lrotr are only used with MSVC. 7696 7697 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 7698 (the default implementation of RAND_status). 7699 7700 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 7701 to '-clrext' (= clear extensions), as intended and documented. 7702 [Bodo Moeller; inconsistency pointed out by Michael Attili 7703 <attili@amaxo.com>] 7704 7705 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 7706 was larger than the MD block size. 7707 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 7708 7709 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 7710 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 7711 using the passed key: if the passed key was a private key the result 7712 of X509_print(), for example, would be to print out all the private key 7713 components. 7714 [Steve Henson] 7715 7716 *) des_quad_cksum() byte order bug fix. 7717 [Ulf M�ller, using the problem description in krb4-0.9.7, where 7718 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 7719 7720 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 7721 discouraged. 7722 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 7723 7724 *) For easily testing in shell scripts whether some command 7725 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 7726 returns with exit code 0 iff no command of the given name is available. 7727 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 7728 the output goes to stdout and nothing is printed to stderr. 7729 Additional arguments are always ignored. 7730 7731 Since for each cipher there is a command of the same name, 7732 the 'no-cipher' compilation switches can be tested this way. 7733 7734 ('openssl no-XXX' is not able to detect pseudo-commands such 7735 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 7736 [Bodo Moeller] 7737 7738 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 7739 [Bodo Moeller] 7740 7741 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 7742 is set; it will be thrown away anyway because each handshake creates 7743 its own key. 7744 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 7745 to parameters -- in previous versions (since OpenSSL 0.9.3) the 7746 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 7747 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 7748 [Bodo Moeller] 7749 7750 *) New s_client option -ign_eof: EOF at stdin is ignored, and 7751 'Q' and 'R' lose their special meanings (quit/renegotiate). 7752 This is part of what -quiet does; unlike -quiet, -ign_eof 7753 does not suppress any output. 7754 [Richard Levitte] 7755 7756 *) Add compatibility options to the purpose and trust code. The 7757 purpose X509_PURPOSE_ANY is "any purpose" which automatically 7758 accepts a certificate or CA, this was the previous behaviour, 7759 with all the associated security issues. 7760 7761 X509_TRUST_COMPAT is the old trust behaviour: only and 7762 automatically trust self signed roots in certificate store. A 7763 new trust setting X509_TRUST_DEFAULT is used to specify that 7764 a purpose has no associated trust setting and it should instead 7765 use the value in the default purpose. 7766 [Steve Henson] 7767 7768 *) Fix the PKCS#8 DSA private key code so it decodes keys again 7769 and fix a memory leak. 7770 [Steve Henson] 7771 7772 *) In util/mkerr.pl (which implements 'make errors'), preserve 7773 reason strings from the previous version of the .c file, as 7774 the default to have only downcase letters (and digits) in 7775 automatically generated reasons codes is not always appropriate. 7776 [Bodo Moeller] 7777 7778 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 7779 using strerror. Previously, ERR_reason_error_string() returned 7780 library names as reason strings for SYSerr; but SYSerr is a special 7781 case where small numbers are errno values, not library numbers. 7782 [Bodo Moeller] 7783 7784 *) Add '-dsaparam' option to 'openssl dhparam' application. This 7785 converts DSA parameters into DH parameters. (When creating parameters, 7786 DSA_generate_parameters is used.) 7787 [Bodo Moeller] 7788 7789 *) Include 'length' (recommended exponent length) in C code generated 7790 by 'openssl dhparam -C'. 7791 [Bodo Moeller] 7792 7793 *) The second argument to set_label in perlasm was already being used 7794 so couldn't be used as a "file scope" flag. Moved to third argument 7795 which was free. 7796 [Steve Henson] 7797 7798 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 7799 instead of RAND_bytes for encryption IVs and salts. 7800 [Bodo Moeller] 7801 7802 *) Include RAND_status() into RAND_METHOD instead of implementing 7803 it only for md_rand.c Otherwise replacing the PRNG by calling 7804 RAND_set_rand_method would be impossible. 7805 [Bodo Moeller] 7806 7807 *) Don't let DSA_generate_key() enter an infinite loop if the random 7808 number generation fails. 7809 [Bodo Moeller] 7810 7811 *) New 'rand' application for creating pseudo-random output. 7812 [Bodo Moeller] 7813 7814 *) Added configuration support for Linux/IA64 7815 [Rolf Haberrecker <rolf@suse.de>] 7816 7817 *) Assembler module support for Mingw32. 7818 [Ulf M�ller] 7819 7820 *) Shared library support for HPUX (in shlib/). 7821 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 7822 7823 *) Shared library support for Solaris gcc. 7824 [Lutz Behnke <behnke@trustcenter.de>] 7825 7826 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 7827 7828 *) PKCS7_encrypt() was adding text MIME headers twice because they 7829 were added manually and by SMIME_crlf_copy(). 7830 [Steve Henson] 7831 7832 *) In bntest.c don't call BN_rand with zero bits argument. 7833 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 7834 7835 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 7836 case was implemented. This caused BN_div_recp() to fail occasionally. 7837 [Ulf M�ller] 7838 7839 *) Add an optional second argument to the set_label() in the perl 7840 assembly language builder. If this argument exists and is set 7841 to 1 it signals that the assembler should use a symbol whose 7842 scope is the entire file, not just the current function. This 7843 is needed with MASM which uses the format label:: for this scope. 7844 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 7845 7846 *) Change the ASN1 types so they are typedefs by default. Before 7847 almost all types were #define'd to ASN1_STRING which was causing 7848 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 7849 for example. 7850 [Steve Henson] 7851 7852 *) Change names of new functions to the new get1/get0 naming 7853 convention: After 'get1', the caller owns a reference count 7854 and has to call ..._free; 'get0' returns a pointer to some 7855 data structure without incrementing reference counters. 7856 (Some of the existing 'get' functions increment a reference 7857 counter, some don't.) 7858 Similarly, 'set1' and 'add1' functions increase reference 7859 counters or duplicate objects. 7860 [Steve Henson] 7861 7862 *) Allow for the possibility of temp RSA key generation failure: 7863 the code used to assume it always worked and crashed on failure. 7864 [Steve Henson] 7865 7866 *) Fix potential buffer overrun problem in BIO_printf(). 7867 [Ulf M�ller, using public domain code by Patrick Powell; problem 7868 pointed out by David Sacerdote <das33@cornell.edu>] 7869 7870 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 7871 RAND_egd() and RAND_status(). In the command line application, 7872 the EGD socket can be specified like a seed file using RANDFILE 7873 or -rand. 7874 [Ulf M�ller] 7875 7876 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 7877 Some CAs (e.g. Verisign) distribute certificates in this form. 7878 [Steve Henson] 7879 7880 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 7881 list to exclude them. This means that no special compilation option 7882 is needed to use anonymous DH: it just needs to be included in the 7883 cipher list. 7884 [Steve Henson] 7885 7886 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 7887 EVP_MD_type. The old functionality is available in a new macro called 7888 EVP_MD_md(). Change code that uses it and update docs. 7889 [Steve Henson] 7890 7891 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 7892 where the 'void *' argument is replaced by a function pointer argument. 7893 Previously 'void *' was abused to point to functions, which works on 7894 many platforms, but is not correct. As these functions are usually 7895 called by macros defined in OpenSSL header files, most source code 7896 should work without changes. 7897 [Richard Levitte] 7898 7899 *) <openssl/opensslconf.h> (which is created by Configure) now contains 7900 sections with information on -D... compiler switches used for 7901 compiling the library so that applications can see them. To enable 7902 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 7903 must be defined. E.g., 7904 #define OPENSSL_ALGORITHM_DEFINES 7905 #include <openssl/opensslconf.h> 7906 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 7907 [Richard Levitte, Ulf and Bodo M�ller] 7908 7909 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 7910 record layer. 7911 [Bodo Moeller] 7912 7913 *) Change the 'other' type in certificate aux info to a STACK_OF 7914 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 7915 the required ASN1 format: arbitrary types determined by an OID. 7916 [Steve Henson] 7917 7918 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 7919 argument to 'req'. This is not because the function is newer or 7920 better than others it just uses the work 'NEW' in the certificate 7921 request header lines. Some software needs this. 7922 [Steve Henson] 7923 7924 *) Reorganise password command line arguments: now passwords can be 7925 obtained from various sources. Delete the PEM_cb function and make 7926 it the default behaviour: i.e. if the callback is NULL and the 7927 usrdata argument is not NULL interpret it as a null terminated pass 7928 phrase. If usrdata and the callback are NULL then the pass phrase 7929 is prompted for as usual. 7930 [Steve Henson] 7931 7932 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 7933 the support is automatically enabled. The resulting binaries will 7934 autodetect the card and use it if present. 7935 [Ben Laurie and Compaq Inc.] 7936 7937 *) Work around for Netscape hang bug. This sends certificate request 7938 and server done in one record. Since this is perfectly legal in the 7939 SSL/TLS protocol it isn't a "bug" option and is on by default. See 7940 the bugs/SSLv3 entry for more info. 7941 [Steve Henson] 7942 7943 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 7944 [Andy Polyakov] 7945 7946 *) Add -rand argument to smime and pkcs12 applications and read/write 7947 of seed file. 7948 [Steve Henson] 7949 7950 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 7951 [Bodo Moeller] 7952 7953 *) Add command line password options to the remaining applications. 7954 [Steve Henson] 7955 7956 *) Bug fix for BN_div_recp() for numerators with an even number of 7957 bits. 7958 [Ulf M�ller] 7959 7960 *) More tests in bntest.c, and changed test_bn output. 7961 [Ulf M�ller] 7962 7963 *) ./config recognizes MacOS X now. 7964 [Andy Polyakov] 7965 7966 *) Bug fix for BN_div() when the first words of num and divsor are 7967 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 7968 [Ulf M�ller] 7969 7970 *) Add support for various broken PKCS#8 formats, and command line 7971 options to produce them. 7972 [Steve Henson] 7973 7974 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 7975 get temporary BIGNUMs from a BN_CTX. 7976 [Ulf M�ller] 7977 7978 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 7979 for p == 0. 7980 [Ulf M�ller] 7981 7982 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 7983 include a #define from the old name to the new. The original intent 7984 was that statically linked binaries could for example just call 7985 SSLeay_add_all_ciphers() to just add ciphers to the table and not 7986 link with digests. This never worked becayse SSLeay_add_all_digests() 7987 and SSLeay_add_all_ciphers() were in the same source file so calling 7988 one would link with the other. They are now in separate source files. 7989 [Steve Henson] 7990 7991 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 7992 [Steve Henson] 7993 7994 *) Use a less unusual form of the Miller-Rabin primality test (it used 7995 a binary algorithm for exponentiation integrated into the Miller-Rabin 7996 loop, our standard modexp algorithms are faster). 7997 [Bodo Moeller] 7998 7999 *) Support for the EBCDIC character set completed. 8000 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 8001 8002 *) Source code cleanups: use const where appropriate, eliminate casts, 8003 use void * instead of char * in lhash. 8004 [Ulf M�ller] 8005 8006 *) Bugfix: ssl3_send_server_key_exchange was not restartable 8007 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 8008 this the server could overwrite ephemeral keys that the client 8009 has already seen). 8010 [Bodo Moeller] 8011 8012 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 8013 using 50 iterations of the Rabin-Miller test. 8014 8015 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 8016 iterations of the Rabin-Miller test as required by the appendix 8017 to FIPS PUB 186[-1]) instead of DSA_is_prime. 8018 As BN_is_prime_fasttest includes trial division, DSA parameter 8019 generation becomes much faster. 8020 8021 This implies a change for the callback functions in DSA_is_prime 8022 and DSA_generate_parameters: The callback function is called once 8023 for each positive witness in the Rabin-Miller test, not just 8024 occasionally in the inner loop; and the parameters to the 8025 callback function now provide an iteration count for the outer 8026 loop rather than for the current invocation of the inner loop. 8027 DSA_generate_parameters additionally can call the callback 8028 function with an 'iteration count' of -1, meaning that a 8029 candidate has passed the trial division test (when q is generated 8030 from an application-provided seed, trial division is skipped). 8031 [Bodo Moeller] 8032 8033 *) New function BN_is_prime_fasttest that optionally does trial 8034 division before starting the Rabin-Miller test and has 8035 an additional BN_CTX * argument (whereas BN_is_prime always 8036 has to allocate at least one BN_CTX). 8037 'callback(1, -1, cb_arg)' is called when a number has passed the 8038 trial division stage. 8039 [Bodo Moeller] 8040 8041 *) Fix for bug in CRL encoding. The validity dates weren't being handled 8042 as ASN1_TIME. 8043 [Steve Henson] 8044 8045 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 8046 [Steve Henson] 8047 8048 *) New function BN_pseudo_rand(). 8049 [Ulf M�ller] 8050 8051 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 8052 bignum version of BN_from_montgomery() with the working code from 8053 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 8054 the comments. 8055 [Ulf M�ller] 8056 8057 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 8058 made it impossible to use the same SSL_SESSION data structure in 8059 SSL2 clients in multiple threads. 8060 [Bodo Moeller] 8061 8062 *) The return value of RAND_load_file() no longer counts bytes obtained 8063 by stat(). RAND_load_file(..., -1) is new and uses the complete file 8064 to seed the PRNG (previously an explicit byte count was required). 8065 [Ulf M�ller, Bodo M�ller] 8066 8067 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 8068 used (char *) instead of (void *) and had casts all over the place. 8069 [Steve Henson] 8070 8071 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 8072 [Ulf M�ller] 8073 8074 *) Retain source code compatibility for BN_prime_checks macro: 8075 BN_is_prime(..., BN_prime_checks, ...) now uses 8076 BN_prime_checks_for_size to determine the appropriate number of 8077 Rabin-Miller iterations. 8078 [Ulf M�ller] 8079 8080 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 8081 DH_CHECK_P_NOT_SAFE_PRIME. 8082 (Check if this is true? OpenPGP calls them "strong".) 8083 [Ulf M�ller] 8084 8085 *) Merge the functionality of "dh" and "gendh" programs into a new program 8086 "dhparam". The old programs are retained for now but will handle DH keys 8087 (instead of parameters) in future. 8088 [Steve Henson] 8089 8090 *) Make the ciphers, s_server and s_client programs check the return values 8091 when a new cipher list is set. 8092 [Steve Henson] 8093 8094 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 8095 ciphers. Before when the 56bit ciphers were enabled the sorting was 8096 wrong. 8097 8098 The syntax for the cipher sorting has been extended to support sorting by 8099 cipher-strength (using the strength_bits hard coded in the tables). 8100 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 8101 8102 Fix a bug in the cipher-command parser: when supplying a cipher command 8103 string with an "undefined" symbol (neither command nor alphanumeric 8104 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 8105 an error is flagged. 8106 8107 Due to the strength-sorting extension, the code of the 8108 ssl_create_cipher_list() function was completely rearranged. I hope that 8109 the readability was also increased :-) 8110 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 8111 8112 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 8113 for the first serial number and places 2 in the serial number file. This 8114 avoids problems when the root CA is created with serial number zero and 8115 the first user certificate has the same issuer name and serial number 8116 as the root CA. 8117 [Steve Henson] 8118 8119 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 8120 the new code. Add documentation for this stuff. 8121 [Steve Henson] 8122 8123 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 8124 X509_*() to X509at_*() on the grounds that they don't handle X509 8125 structures and behave in an analagous way to the X509v3 functions: 8126 they shouldn't be called directly but wrapper functions should be used 8127 instead. 8128 8129 So we also now have some wrapper functions that call the X509at functions 8130 when passed certificate requests. (TO DO: similar things can be done with 8131 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 8132 things. Some of these need some d2i or i2d and print functionality 8133 because they handle more complex structures.) 8134 [Steve Henson] 8135 8136 *) Add missing #ifndefs that caused missing symbols when building libssl 8137 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 8138 NO_RSA in ssl/s2*.c. 8139 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 8140 8141 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 8142 has a return value which indicates the quality of the random data 8143 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 8144 error queue. New function RAND_pseudo_bytes() generates output that is 8145 guaranteed to be unique but not unpredictable. RAND_add is like 8146 RAND_seed, but takes an extra argument for an entropy estimate 8147 (RAND_seed always assumes full entropy). 8148 [Ulf M�ller] 8149 8150 *) Do more iterations of Rabin-Miller probable prime test (specifically, 8151 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 8152 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 8153 in crypto/bn/bn_prime.c for the complete table). This guarantees a 8154 false-positive rate of at most 2^-80 for random input. 8155 [Bodo Moeller] 8156 8157 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 8158 [Bodo Moeller] 8159 8160 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 8161 in the 0.9.5 release), this returns the chain 8162 from an X509_CTX structure with a dup of the stack and all 8163 the X509 reference counts upped: so the stack will exist 8164 after X509_CTX_cleanup() has been called. Modify pkcs12.c 8165 to use this. 8166 8167 Also make SSL_SESSION_print() print out the verify return 8168 code. 8169 [Steve Henson] 8170 8171 *) Add manpage for the pkcs12 command. Also change the default 8172 behaviour so MAC iteration counts are used unless the new 8173 -nomaciter option is used. This improves file security and 8174 only older versions of MSIE (4.0 for example) need it. 8175 [Steve Henson] 8176 8177 *) Honor the no-xxx Configure options when creating .DEF files. 8178 [Ulf M�ller] 8179 8180 *) Add PKCS#10 attributes to field table: challengePassword, 8181 unstructuredName and unstructuredAddress. These are taken from 8182 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 8183 international characters are used. 8184 8185 More changes to X509_ATTRIBUTE code: allow the setting of types 8186 based on strings. Remove the 'loc' parameter when adding 8187 attributes because these will be a SET OF encoding which is sorted 8188 in ASN1 order. 8189 [Steve Henson] 8190 8191 *) Initial changes to the 'req' utility to allow request generation 8192 automation. This will allow an application to just generate a template 8193 file containing all the field values and have req construct the 8194 request. 8195 8196 Initial support for X509_ATTRIBUTE handling. Stacks of these are 8197 used all over the place including certificate requests and PKCS#7 8198 structures. They are currently handled manually where necessary with 8199 some primitive wrappers for PKCS#7. The new functions behave in a 8200 manner analogous to the X509 extension functions: they allow 8201 attributes to be looked up by NID and added. 8202 8203 Later something similar to the X509V3 code would be desirable to 8204 automatically handle the encoding, decoding and printing of the 8205 more complex types. The string types like challengePassword can 8206 be handled by the string table functions. 8207 8208 Also modified the multi byte string table handling. Now there is 8209 a 'global mask' which masks out certain types. The table itself 8210 can use the flag STABLE_NO_MASK to ignore the mask setting: this 8211 is useful when for example there is only one permissible type 8212 (as in countryName) and using the mask might result in no valid 8213 types at all. 8214 [Steve Henson] 8215 8216 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 8217 SSL_get_peer_finished to allow applications to obtain the latest 8218 Finished messages sent to the peer or expected from the peer, 8219 respectively. (SSL_get_peer_finished is usually the Finished message 8220 actually received from the peer, otherwise the protocol will be aborted.) 8221 8222 As the Finished message are message digests of the complete handshake 8223 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 8224 be used for external authentication procedures when the authentication 8225 provided by SSL/TLS is not desired or is not enough. 8226 [Bodo Moeller] 8227 8228 *) Enhanced support for Alpha Linux is added. Now ./config checks if 8229 the host supports BWX extension and if Compaq C is present on the 8230 $PATH. Just exploiting of the BWX extension results in 20-30% 8231 performance kick for some algorithms, e.g. DES and RC4 to mention 8232 a couple. Compaq C in turn generates ~20% faster code for MD5 and 8233 SHA1. 8234 [Andy Polyakov] 8235 8236 *) Add support for MS "fast SGC". This is arguably a violation of the 8237 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 8238 weak crypto and after checking the certificate is SGC a second one 8239 with strong crypto. MS SGC stops the first handshake after receiving 8240 the server certificate message and sends a second client hello. Since 8241 a server will typically do all the time consuming operations before 8242 expecting any further messages from the client (server key exchange 8243 is the most expensive) there is little difference between the two. 8244 8245 To get OpenSSL to support MS SGC we have to permit a second client 8246 hello message after we have sent server done. In addition we have to 8247 reset the MAC if we do get this second client hello. 8248 [Steve Henson] 8249 8250 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 8251 if a DER encoded private key is RSA or DSA traditional format. Changed 8252 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 8253 format DER encoded private key. Newer code should use PKCS#8 format which 8254 has the key type encoded in the ASN1 structure. Added DER private key 8255 support to pkcs8 application. 8256 [Steve Henson] 8257 8258 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 8259 ciphersuites has been selected (as required by the SSL 3/TLS 1 8260 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 8261 is set, we interpret this as a request to violate the specification 8262 (the worst that can happen is a handshake failure, and 'correct' 8263 behaviour would result in a handshake failure anyway). 8264 [Bodo Moeller] 8265 8266 *) In SSL_CTX_add_session, take into account that there might be multiple 8267 SSL_SESSION structures with the same session ID (e.g. when two threads 8268 concurrently obtain them from an external cache). 8269 The internal cache can handle only one SSL_SESSION with a given ID, 8270 so if there's a conflict, we now throw out the old one to achieve 8271 consistency. 8272 [Bodo Moeller] 8273 8274 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 8275 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 8276 some routines that use cipher OIDs: some ciphers do not have OIDs 8277 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 8278 example. 8279 [Steve Henson] 8280 8281 *) Simplify the trust setting structure and code. Now we just have 8282 two sequences of OIDs for trusted and rejected settings. These will 8283 typically have values the same as the extended key usage extension 8284 and any application specific purposes. 8285 8286 The trust checking code now has a default behaviour: it will just 8287 check for an object with the same NID as the passed id. Functions can 8288 be provided to override either the default behaviour or the behaviour 8289 for a given id. SSL client, server and email already have functions 8290 in place for compatibility: they check the NID and also return "trusted" 8291 if the certificate is self signed. 8292 [Steve Henson] 8293 8294 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 8295 traditional format into an EVP_PKEY structure. 8296 [Steve Henson] 8297 8298 *) Add a password callback function PEM_cb() which either prompts for 8299 a password if usr_data is NULL or otherwise assumes it is a null 8300 terminated password. Allow passwords to be passed on command line 8301 environment or config files in a few more utilities. 8302 [Steve Henson] 8303 8304 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 8305 keys. Add some short names for PKCS#8 PBE algorithms and allow them 8306 to be specified on the command line for the pkcs8 and pkcs12 utilities. 8307 Update documentation. 8308 [Steve Henson] 8309 8310 *) Support for ASN1 "NULL" type. This could be handled before by using 8311 ASN1_TYPE but there wasn't any function that would try to read a NULL 8312 and produce an error if it couldn't. For compatibility we also have 8313 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 8314 don't allocate anything because they don't need to. 8315 [Steve Henson] 8316 8317 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 8318 for details. 8319 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 8320 8321 *) Rebuild of the memory allocation routines used by OpenSSL code and 8322 possibly others as well. The purpose is to make an interface that 8323 provide hooks so anyone can build a separate set of allocation and 8324 deallocation routines to be used by OpenSSL, for example memory 8325 pool implementations, or something else, which was previously hard 8326 since Malloc(), Realloc() and Free() were defined as macros having 8327 the values malloc, realloc and free, respectively (except for Win32 8328 compilations). The same is provided for memory debugging code. 8329 OpenSSL already comes with functionality to find memory leaks, but 8330 this gives people a chance to debug other memory problems. 8331 8332 With these changes, a new set of functions and macros have appeared: 8333 8334 CRYPTO_set_mem_debug_functions() [F] 8335 CRYPTO_get_mem_debug_functions() [F] 8336 CRYPTO_dbg_set_options() [F] 8337 CRYPTO_dbg_get_options() [F] 8338 CRYPTO_malloc_debug_init() [M] 8339 8340 The memory debug functions are NULL by default, unless the library 8341 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 8342 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 8343 gives the standard debugging functions that come with OpenSSL) or 8344 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 8345 provided by the library user) must be used. When the standard 8346 debugging functions are used, CRYPTO_dbg_set_options can be used to 8347 request additional information: 8348 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 8349 the CRYPTO_MDEBUG_xxx macro when compiling the library. 8350 8351 Also, things like CRYPTO_set_mem_functions will always give the 8352 expected result (the new set of functions is used for allocation 8353 and deallocation) at all times, regardless of platform and compiler 8354 options. 8355 8356 To finish it up, some functions that were never use in any other 8357 way than through macros have a new API and new semantic: 8358 8359 CRYPTO_dbg_malloc() 8360 CRYPTO_dbg_realloc() 8361 CRYPTO_dbg_free() 8362 8363 All macros of value have retained their old syntax. 8364 [Richard Levitte and Bodo Moeller] 8365 8366 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 8367 ordering of SMIMECapabilities wasn't in "strength order" and there 8368 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 8369 algorithm. 8370 [Steve Henson] 8371 8372 *) Some ASN1 types with illegal zero length encoding (INTEGER, 8373 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 8374 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 8375 8376 *) Merge in my S/MIME library for OpenSSL. This provides a simple 8377 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 8378 functionality to handle multipart/signed properly) and a utility 8379 called 'smime' to call all this stuff. This is based on code I 8380 originally wrote for Celo who have kindly allowed it to be 8381 included in OpenSSL. 8382 [Steve Henson] 8383 8384 *) Add variants des_set_key_checked and des_set_key_unchecked of 8385 des_set_key (aka des_key_sched). Global variable des_check_key 8386 decides which of these is called by des_set_key; this way 8387 des_check_key behaves as it always did, but applications and 8388 the library itself, which was buggy for des_check_key == 1, 8389 have a cleaner way to pick the version they need. 8390 [Bodo Moeller] 8391 8392 *) New function PKCS12_newpass() which changes the password of a 8393 PKCS12 structure. 8394 [Steve Henson] 8395 8396 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 8397 dynamic mix. In both cases the ids can be used as an index into the 8398 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 8399 functions so they accept a list of the field values and the 8400 application doesn't need to directly manipulate the X509_TRUST 8401 structure. 8402 [Steve Henson] 8403 8404 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 8405 need initialising. 8406 [Steve Henson] 8407 8408 *) Modify the way the V3 extension code looks up extensions. This now 8409 works in a similar way to the object code: we have some "standard" 8410 extensions in a static table which is searched with OBJ_bsearch() 8411 and the application can add dynamic ones if needed. The file 8412 crypto/x509v3/ext_dat.h now has the info: this file needs to be 8413 updated whenever a new extension is added to the core code and kept 8414 in ext_nid order. There is a simple program 'tabtest.c' which checks 8415 this. New extensions are not added too often so this file can readily 8416 be maintained manually. 8417 8418 There are two big advantages in doing things this way. The extensions 8419 can be looked up immediately and no longer need to be "added" using 8420 X509V3_add_standard_extensions(): this function now does nothing. 8421 [Side note: I get *lots* of email saying the extension code doesn't 8422 work because people forget to call this function] 8423 Also no dynamic allocation is done unless new extensions are added: 8424 so if we don't add custom extensions there is no need to call 8425 X509V3_EXT_cleanup(). 8426 [Steve Henson] 8427 8428 *) Modify enc utility's salting as follows: make salting the default. Add a 8429 magic header, so unsalted files fail gracefully instead of just decrypting 8430 to garbage. This is because not salting is a big security hole, so people 8431 should be discouraged from doing it. 8432 [Ben Laurie] 8433 8434 *) Fixes and enhancements to the 'x509' utility. It allowed a message 8435 digest to be passed on the command line but it only used this 8436 parameter when signing a certificate. Modified so all relevant 8437 operations are affected by the digest parameter including the 8438 -fingerprint and -x509toreq options. Also -x509toreq choked if a 8439 DSA key was used because it didn't fix the digest. 8440 [Steve Henson] 8441 8442 *) Initial certificate chain verify code. Currently tests the untrusted 8443 certificates for consistency with the verify purpose (which is set 8444 when the X509_STORE_CTX structure is set up) and checks the pathlength. 8445 8446 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 8447 this is because it will reject chains with invalid extensions whereas 8448 every previous version of OpenSSL and SSLeay made no checks at all. 8449 8450 Trust code: checks the root CA for the relevant trust settings. Trust 8451 settings have an initial value consistent with the verify purpose: e.g. 8452 if the verify purpose is for SSL client use it expects the CA to be 8453 trusted for SSL client use. However the default value can be changed to 8454 permit custom trust settings: one example of this would be to only trust 8455 certificates from a specific "secure" set of CAs. 8456 8457 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 8458 which should be used for version portability: especially since the 8459 verify structure is likely to change more often now. 8460 8461 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 8462 to set them. If not set then assume SSL clients will verify SSL servers 8463 and vice versa. 8464 8465 Two new options to the verify program: -untrusted allows a set of 8466 untrusted certificates to be passed in and -purpose which sets the 8467 intended purpose of the certificate. If a purpose is set then the 8468 new chain verify code is used to check extension consistency. 8469 [Steve Henson] 8470 8471 *) Support for the authority information access extension. 8472 [Steve Henson] 8473 8474 *) Modify RSA and DSA PEM read routines to transparently handle 8475 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 8476 public keys in a format compatible with certificate 8477 SubjectPublicKeyInfo structures. Unfortunately there were already 8478 functions called *_PublicKey_* which used various odd formats so 8479 these are retained for compatibility: however the DSA variants were 8480 never in a public release so they have been deleted. Changed dsa/rsa 8481 utilities to handle the new format: note no releases ever handled public 8482 keys so we should be OK. 8483 8484 The primary motivation for this change is to avoid the same fiasco 8485 that dogs private keys: there are several incompatible private key 8486 formats some of which are standard and some OpenSSL specific and 8487 require various evil hacks to allow partial transparent handling and 8488 even then it doesn't work with DER formats. Given the option anything 8489 other than PKCS#8 should be dumped: but the other formats have to 8490 stay in the name of compatibility. 8491 8492 With public keys and the benefit of hindsight one standard format 8493 is used which works with EVP_PKEY, RSA or DSA structures: though 8494 it clearly returns an error if you try to read the wrong kind of key. 8495 8496 Added a -pubkey option to the 'x509' utility to output the public key. 8497 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 8498 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 8499 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 8500 that do the same as the EVP_PKEY_assign_*() except they up the 8501 reference count of the added key (they don't "swallow" the 8502 supplied key). 8503 [Steve Henson] 8504 8505 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 8506 CRLs would fail if the file contained no certificates or no CRLs: 8507 added a new function to read in both types and return the number 8508 read: this means that if none are read it will be an error. The 8509 DER versions of the certificate and CRL reader would always fail 8510 because it isn't possible to mix certificates and CRLs in DER format 8511 without choking one or the other routine. Changed this to just read 8512 a certificate: this is the best we can do. Also modified the code 8513 in apps/verify.c to take notice of return codes: it was previously 8514 attempting to read in certificates from NULL pointers and ignoring 8515 any errors: this is one reason why the cert and CRL reader seemed 8516 to work. It doesn't check return codes from the default certificate 8517 routines: these may well fail if the certificates aren't installed. 8518 [Steve Henson] 8519 8520 *) Code to support otherName option in GeneralName. 8521 [Steve Henson] 8522 8523 *) First update to verify code. Change the verify utility 8524 so it warns if it is passed a self signed certificate: 8525 for consistency with the normal behaviour. X509_verify 8526 has been modified to it will now verify a self signed 8527 certificate if *exactly* the same certificate appears 8528 in the store: it was previously impossible to trust a 8529 single self signed certificate. This means that: 8530 openssl verify ss.pem 8531 now gives a warning about a self signed certificate but 8532 openssl verify -CAfile ss.pem ss.pem 8533 is OK. 8534 [Steve Henson] 8535 8536 *) For servers, store verify_result in SSL_SESSION data structure 8537 (and add it to external session representation). 8538 This is needed when client certificate verifications fails, 8539 but an application-provided verification callback (set by 8540 SSL_CTX_set_cert_verify_callback) allows accepting the session 8541 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 8542 but returns 1): When the session is reused, we have to set 8543 ssl->verify_result to the appropriate error code to avoid 8544 security holes. 8545 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 8546 8547 *) Fix a bug in the new PKCS#7 code: it didn't consider the 8548 case in PKCS7_dataInit() where the signed PKCS7 structure 8549 didn't contain any existing data because it was being created. 8550 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 8551 8552 *) Add a salt to the key derivation routines in enc.c. This 8553 forms the first 8 bytes of the encrypted file. Also add a 8554 -S option to allow a salt to be input on the command line. 8555 [Steve Henson] 8556 8557 *) New function X509_cmp(). Oddly enough there wasn't a function 8558 to compare two certificates. We do this by working out the SHA1 8559 hash and comparing that. X509_cmp() will be needed by the trust 8560 code. 8561 [Steve Henson] 8562 8563 *) SSL_get1_session() is like SSL_get_session(), but increments 8564 the reference count in the SSL_SESSION returned. 8565 [Geoff Thorpe <geoff@eu.c2.net>] 8566 8567 *) Fix for 'req': it was adding a null to request attributes. 8568 Also change the X509_LOOKUP and X509_INFO code to handle 8569 certificate auxiliary information. 8570 [Steve Henson] 8571 8572 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 8573 the 'enc' command. 8574 [Steve Henson] 8575 8576 *) Add the possibility to add extra information to the memory leak 8577 detecting output, to form tracebacks, showing from where each 8578 allocation was originated: CRYPTO_push_info("constant string") adds 8579 the string plus current file name and line number to a per-thread 8580 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 8581 is like calling CYRPTO_pop_info() until the stack is empty. 8582 Also updated memory leak detection code to be multi-thread-safe. 8583 [Richard Levitte] 8584 8585 *) Add options -text and -noout to pkcs7 utility and delete the 8586 encryption options which never did anything. Update docs. 8587 [Steve Henson] 8588 8589 *) Add options to some of the utilities to allow the pass phrase 8590 to be included on either the command line (not recommended on 8591 OSes like Unix) or read from the environment. Update the 8592 manpages and fix a few bugs. 8593 [Steve Henson] 8594 8595 *) Add a few manpages for some of the openssl commands. 8596 [Steve Henson] 8597 8598 *) Fix the -revoke option in ca. It was freeing up memory twice, 8599 leaking and not finding already revoked certificates. 8600 [Steve Henson] 8601 8602 *) Extensive changes to support certificate auxiliary information. 8603 This involves the use of X509_CERT_AUX structure and X509_AUX 8604 functions. An X509_AUX function such as PEM_read_X509_AUX() 8605 can still read in a certificate file in the usual way but it 8606 will also read in any additional "auxiliary information". By 8607 doing things this way a fair degree of compatibility can be 8608 retained: existing certificates can have this information added 8609 using the new 'x509' options. 8610 8611 Current auxiliary information includes an "alias" and some trust 8612 settings. The trust settings will ultimately be used in enhanced 8613 certificate chain verification routines: currently a certificate 8614 can only be trusted if it is self signed and then it is trusted 8615 for all purposes. 8616 [Steve Henson] 8617 8618 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 8619 The problem was that one of the replacement routines had not been working 8620 since SSLeay releases. For now the offending routine has been replaced 8621 with non-optimised assembler. Even so, this now gives around 95% 8622 performance improvement for 1024 bit RSA signs. 8623 [Mark Cox] 8624 8625 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 8626 handling. Most clients have the effective key size in bits equal to 8627 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 8628 A few however don't do this and instead use the size of the decrypted key 8629 to determine the RC2 key length and the AlgorithmIdentifier to determine 8630 the effective key length. In this case the effective key length can still 8631 be 40 bits but the key length can be 168 bits for example. This is fixed 8632 by manually forcing an RC2 key into the EVP_PKEY structure because the 8633 EVP code can't currently handle unusual RC2 key sizes: it always assumes 8634 the key length and effective key length are equal. 8635 [Steve Henson] 8636 8637 *) Add a bunch of functions that should simplify the creation of 8638 X509_NAME structures. Now you should be able to do: 8639 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 8640 and have it automatically work out the correct field type and fill in 8641 the structures. The more adventurous can try: 8642 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 8643 and it will (hopefully) work out the correct multibyte encoding. 8644 [Steve Henson] 8645 8646 *) Change the 'req' utility to use the new field handling and multibyte 8647 copy routines. Before the DN field creation was handled in an ad hoc 8648 way in req, ca, and x509 which was rather broken and didn't support 8649 BMPStrings or UTF8Strings. Since some software doesn't implement 8650 BMPStrings or UTF8Strings yet, they can be enabled using the config file 8651 using the dirstring_type option. See the new comment in the default 8652 openssl.cnf for more info. 8653 [Steve Henson] 8654 8655 *) Make crypto/rand/md_rand.c more robust: 8656 - Assure unique random numbers after fork(). 8657 - Make sure that concurrent threads access the global counter and 8658 md serializably so that we never lose entropy in them 8659 or use exactly the same state in multiple threads. 8660 Access to the large state is not always serializable because 8661 the additional locking could be a performance killer, and 8662 md should be large enough anyway. 8663 [Bodo Moeller] 8664 8665 *) New file apps/app_rand.c with commonly needed functionality 8666 for handling the random seed file. 8667 8668 Use the random seed file in some applications that previously did not: 8669 ca, 8670 dsaparam -genkey (which also ignored its '-rand' option), 8671 s_client, 8672 s_server, 8673 x509 (when signing). 8674 Except on systems with /dev/urandom, it is crucial to have a random 8675 seed file at least for key creation, DSA signing, and for DH exchanges; 8676 for RSA signatures we could do without one. 8677 8678 gendh and gendsa (unlike genrsa) used to read only the first byte 8679 of each file listed in the '-rand' option. The function as previously 8680 found in genrsa is now in app_rand.c and is used by all programs 8681 that support '-rand'. 8682 [Bodo Moeller] 8683 8684 *) In RAND_write_file, use mode 0600 for creating files; 8685 don't just chmod when it may be too late. 8686 [Bodo Moeller] 8687 8688 *) Report an error from X509_STORE_load_locations 8689 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 8690 [Bill Perry] 8691 8692 *) New function ASN1_mbstring_copy() this copies a string in either 8693 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 8694 into an ASN1_STRING type. A mask of permissible types is passed 8695 and it chooses the "minimal" type to use or an error if not type 8696 is suitable. 8697 [Steve Henson] 8698 8699 *) Add function equivalents to the various macros in asn1.h. The old 8700 macros are retained with an M_ prefix. Code inside the library can 8701 use the M_ macros. External code (including the openssl utility) 8702 should *NOT* in order to be "shared library friendly". 8703 [Steve Henson] 8704 8705 *) Add various functions that can check a certificate's extensions 8706 to see if it usable for various purposes such as SSL client, 8707 server or S/MIME and CAs of these types. This is currently 8708 VERY EXPERIMENTAL but will ultimately be used for certificate chain 8709 verification. Also added a -purpose flag to x509 utility to 8710 print out all the purposes. 8711 [Steve Henson] 8712 8713 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 8714 functions. 8715 [Steve Henson] 8716 8717 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 8718 for, obtain and decode and extension and obtain its critical flag. 8719 This allows all the necessary extension code to be handled in a 8720 single function call. 8721 [Steve Henson] 8722 8723 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 8724 platforms. See crypto/rc4/rc4_enc.c for further details. 8725 [Andy Polyakov] 8726 8727 *) New -noout option to asn1parse. This causes no output to be produced 8728 its main use is when combined with -strparse and -out to extract data 8729 from a file (which may not be in ASN.1 format). 8730 [Steve Henson] 8731 8732 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 8733 when producing the local key id. 8734 [Richard Levitte <levitte@stacken.kth.se>] 8735 8736 *) New option -dhparam in s_server. This allows a DH parameter file to be 8737 stated explicitly. If it is not stated then it tries the first server 8738 certificate file. The previous behaviour hard coded the filename 8739 "server.pem". 8740 [Steve Henson] 8741 8742 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 8743 a public key to be input or output. For example: 8744 openssl rsa -in key.pem -pubout -out pubkey.pem 8745 Also added necessary DSA public key functions to handle this. 8746 [Steve Henson] 8747 8748 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 8749 in the message. This was handled by allowing 8750 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 8751 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 8752 8753 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 8754 to the end of the strings whereas this didn't. This would cause problems 8755 if strings read with d2i_ASN1_bytes() were later modified. 8756 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 8757 8758 *) Fix for base64 decode bug. When a base64 bio reads only one line of 8759 data and it contains EOF it will end up returning an error. This is 8760 caused by input 46 bytes long. The cause is due to the way base64 8761 BIOs find the start of base64 encoded data. They do this by trying a 8762 trial decode on each line until they find one that works. When they 8763 do a flag is set and it starts again knowing it can pass all the 8764 data directly through the decoder. Unfortunately it doesn't reset 8765 the context it uses. This means that if EOF is reached an attempt 8766 is made to pass two EOFs through the context and this causes the 8767 resulting error. This can also cause other problems as well. As is 8768 usual with these problems it takes *ages* to find and the fix is 8769 trivial: move one line. 8770 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 8771 8772 *) Ugly workaround to get s_client and s_server working under Windows. The 8773 old code wouldn't work because it needed to select() on sockets and the 8774 tty (for keypresses and to see if data could be written). Win32 only 8775 supports select() on sockets so we select() with a 1s timeout on the 8776 sockets and then see if any characters are waiting to be read, if none 8777 are present then we retry, we also assume we can always write data to 8778 the tty. This isn't nice because the code then blocks until we've 8779 received a complete line of data and it is effectively polling the 8780 keyboard at 1s intervals: however it's quite a bit better than not 8781 working at all :-) A dedicated Windows application might handle this 8782 with an event loop for example. 8783 [Steve Henson] 8784 8785 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 8786 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 8787 will be called when RSA_sign() and RSA_verify() are used. This is useful 8788 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 8789 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 8790 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 8791 This necessitated the support of an extra signature type NID_md5_sha1 8792 for SSL signatures and modifications to the SSL library to use it instead 8793 of calling RSA_public_decrypt() and RSA_private_encrypt(). 8794 [Steve Henson] 8795 8796 *) Add new -verify -CAfile and -CApath options to the crl program, these 8797 will lookup a CRL issuers certificate and verify the signature in a 8798 similar way to the verify program. Tidy up the crl program so it 8799 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 8800 less strict. It will now permit CRL extensions even if it is not 8801 a V2 CRL: this will allow it to tolerate some broken CRLs. 8802 [Steve Henson] 8803 8804 *) Initialize all non-automatic variables each time one of the openssl 8805 sub-programs is started (this is necessary as they may be started 8806 multiple times from the "OpenSSL>" prompt). 8807 [Lennart Bang, Bodo Moeller] 8808 8809 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 8810 removing all other RSA functionality (this is what NO_RSA does). This 8811 is so (for example) those in the US can disable those operations covered 8812 by the RSA patent while allowing storage and parsing of RSA keys and RSA 8813 key generation. 8814 [Steve Henson] 8815 8816 *) Non-copying interface to BIO pairs. 8817 (still largely untested) 8818 [Bodo Moeller] 8819 8820 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 8821 ASCII string. This was handled independently in various places before. 8822 [Steve Henson] 8823 8824 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 8825 UTF8 strings a character at a time. 8826 [Steve Henson] 8827 8828 *) Use client_version from client hello to select the protocol 8829 (s23_srvr.c) and for RSA client key exchange verification 8830 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 8831 [Bodo Moeller] 8832 8833 *) Add various utility functions to handle SPKACs, these were previously 8834 handled by poking round in the structure internals. Added new function 8835 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 8836 print, verify and generate SPKACs. Based on an original idea from 8837 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 8838 [Steve Henson] 8839 8840 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 8841 [Andy Polyakov] 8842 8843 *) Allow the config file extension section to be overwritten on the 8844 command line. Based on an original idea from Massimiliano Pala 8845 <madwolf@comune.modena.it>. The new option is called -extensions 8846 and can be applied to ca, req and x509. Also -reqexts to override 8847 the request extensions in req and -crlexts to override the crl extensions 8848 in ca. 8849 [Steve Henson] 8850 8851 *) Add new feature to the SPKAC handling in ca. Now you can include 8852 the same field multiple times by preceding it by "XXXX." for example: 8853 1.OU="Unit name 1" 8854 2.OU="Unit name 2" 8855 this is the same syntax as used in the req config file. 8856 [Steve Henson] 8857 8858 *) Allow certificate extensions to be added to certificate requests. These 8859 are specified in a 'req_extensions' option of the req section of the 8860 config file. They can be printed out with the -text option to req but 8861 are otherwise ignored at present. 8862 [Steve Henson] 8863 8864 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 8865 data read consists of only the final block it would not decrypted because 8866 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 8867 A misplaced 'break' also meant the decrypted final block might not be 8868 copied until the next read. 8869 [Steve Henson] 8870 8871 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 8872 a few extra parameters to the DH structure: these will be useful if 8873 for example we want the value of 'q' or implement X9.42 DH. 8874 [Steve Henson] 8875 8876 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 8877 provides hooks that allow the default DSA functions or functions on a 8878 "per key" basis to be replaced. This allows hardware acceleration and 8879 hardware key storage to be handled without major modification to the 8880 library. Also added low level modexp hooks and CRYPTO_EX structure and 8881 associated functions. 8882 [Steve Henson] 8883 8884 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 8885 as "read only": it can't be written to and the buffer it points to will 8886 not be freed. Reading from a read only BIO is much more efficient than 8887 a normal memory BIO. This was added because there are several times when 8888 an area of memory needs to be read from a BIO. The previous method was 8889 to create a memory BIO and write the data to it, this results in two 8890 copies of the data and an O(n^2) reading algorithm. There is a new 8891 function BIO_new_mem_buf() which creates a read only memory BIO from 8892 an area of memory. Also modified the PKCS#7 routines to use read only 8893 memory BIOs. 8894 [Steve Henson] 8895 8896 *) Bugfix: ssl23_get_client_hello did not work properly when called in 8897 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 8898 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 8899 but a retry condition occured while trying to read the rest. 8900 [Bodo Moeller] 8901 8902 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 8903 NID_pkcs7_encrypted by default: this was wrong since this should almost 8904 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 8905 the encrypted data type: this is a more sensible place to put it and it 8906 allows the PKCS#12 code to be tidied up that duplicated this 8907 functionality. 8908 [Steve Henson] 8909 8910 *) Changed obj_dat.pl script so it takes its input and output files on 8911 the command line. This should avoid shell escape redirection problems 8912 under Win32. 8913 [Steve Henson] 8914 8915 *) Initial support for certificate extension requests, these are included 8916 in things like Xenroll certificate requests. Included functions to allow 8917 extensions to be obtained and added. 8918 [Steve Henson] 8919 8920 *) -crlf option to s_client and s_server for sending newlines as 8921 CRLF (as required by many protocols). 8922 [Bodo Moeller] 8923 8924 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 8925 8926 *) Install libRSAglue.a when OpenSSL is built with RSAref. 8927 [Ralf S. Engelschall] 8928 8929 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 8930 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 8931 8932 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 8933 program. 8934 [Steve Henson] 8935 8936 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 8937 DH parameters/keys (q is lost during that conversion, but the resulting 8938 DH parameters contain its length). 8939 8940 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 8941 much faster than DH_generate_parameters (which creates parameters 8942 where p = 2*q + 1), and also the smaller q makes DH computations 8943 much more efficient (160-bit exponentiation instead of 1024-bit 8944 exponentiation); so this provides a convenient way to support DHE 8945 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 8946 utter importance to use 8947 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 8948 or 8949 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 8950 when such DH parameters are used, because otherwise small subgroup 8951 attacks may become possible! 8952 [Bodo Moeller] 8953 8954 *) Avoid memory leak in i2d_DHparams. 8955 [Bodo Moeller] 8956 8957 *) Allow the -k option to be used more than once in the enc program: 8958 this allows the same encrypted message to be read by multiple recipients. 8959 [Steve Henson] 8960 8961 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 8962 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 8963 it will always use the numerical form of the OID, even if it has a short 8964 or long name. 8965 [Steve Henson] 8966 8967 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 8968 method only got called if p,q,dmp1,dmq1,iqmp components were present, 8969 otherwise bn_mod_exp was called. In the case of hardware keys for example 8970 no private key components need be present and it might store extra data 8971 in the RSA structure, which cannot be accessed from bn_mod_exp. 8972 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 8973 private key operations. 8974 [Steve Henson] 8975 8976 *) Added support for SPARC Linux. 8977 [Andy Polyakov] 8978 8979 *) pem_password_cb function type incompatibly changed from 8980 typedef int pem_password_cb(char *buf, int size, int rwflag); 8981 to 8982 ....(char *buf, int size, int rwflag, void *userdata); 8983 so that applications can pass data to their callbacks: 8984 The PEM[_ASN1]_{read,write}... functions and macros now take an 8985 additional void * argument, which is just handed through whenever 8986 the password callback is called. 8987 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 8988 8989 New function SSL_CTX_set_default_passwd_cb_userdata. 8990 8991 Compatibility note: As many C implementations push function arguments 8992 onto the stack in reverse order, the new library version is likely to 8993 interoperate with programs that have been compiled with the old 8994 pem_password_cb definition (PEM_whatever takes some data that 8995 happens to be on the stack as its last argument, and the callback 8996 just ignores this garbage); but there is no guarantee whatsoever that 8997 this will work. 8998 8999 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 9000 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 9001 problems not only on Windows, but also on some Unix platforms. 9002 To avoid problematic command lines, these definitions are now in an 9003 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 9004 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 9005 [Bodo Moeller] 9006 9007 *) MIPS III/IV assembler module is reimplemented. 9008 [Andy Polyakov] 9009 9010 *) More DES library cleanups: remove references to srand/rand and 9011 delete an unused file. 9012 [Ulf M�ller] 9013 9014 *) Add support for the the free Netwide assembler (NASM) under Win32, 9015 since not many people have MASM (ml) and it can be hard to obtain. 9016 This is currently experimental but it seems to work OK and pass all 9017 the tests. Check out INSTALL.W32 for info. 9018 [Steve Henson] 9019 9020 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 9021 without temporary keys kept an extra copy of the server key, 9022 and connections with temporary keys did not free everything in case 9023 of an error. 9024 [Bodo Moeller] 9025 9026 *) New function RSA_check_key and new openssl rsa option -check 9027 for verifying the consistency of RSA keys. 9028 [Ulf Moeller, Bodo Moeller] 9029 9030 *) Various changes to make Win32 compile work: 9031 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 9032 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 9033 comparison" warnings. 9034 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 9035 [Steve Henson] 9036 9037 *) Add a debugging option to PKCS#5 v2 key generation function: when 9038 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 9039 derived keys are printed to stderr. 9040 [Steve Henson] 9041 9042 *) Copy the flags in ASN1_STRING_dup(). 9043 [Roman E. Pavlov <pre@mo.msk.ru>] 9044 9045 *) The x509 application mishandled signing requests containing DSA 9046 keys when the signing key was also DSA and the parameters didn't match. 9047 9048 It was supposed to omit the parameters when they matched the signing key: 9049 the verifying software was then supposed to automatically use the CA's 9050 parameters if they were absent from the end user certificate. 9051 9052 Omitting parameters is no longer recommended. The test was also 9053 the wrong way round! This was probably due to unusual behaviour in 9054 EVP_cmp_parameters() which returns 1 if the parameters match. 9055 This meant that parameters were omitted when they *didn't* match and 9056 the certificate was useless. Certificates signed with 'ca' didn't have 9057 this bug. 9058 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 9059 9060 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 9061 The interface is as follows: 9062 Applications can use 9063 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 9064 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 9065 "off" is now the default. 9066 The library internally uses 9067 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 9068 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 9069 to disable memory-checking temporarily. 9070 9071 Some inconsistent states that previously were possible (and were 9072 even the default) are now avoided. 9073 9074 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 9075 with each memory chunk allocated; this is occasionally more helpful 9076 than just having a counter. 9077 9078 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 9079 9080 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 9081 extensions. 9082 [Bodo Moeller] 9083 9084 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 9085 which largely parallels "options", but is for changing API behaviour, 9086 whereas "options" are about protocol behaviour. 9087 Initial "mode" flags are: 9088 9089 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 9090 a single record has been written. 9091 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 9092 retries use the same buffer location. 9093 (But all of the contents must be 9094 copied!) 9095 [Bodo Moeller] 9096 9097 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 9098 worked. 9099 9100 *) Fix problems with no-hmac etc. 9101 [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 9102 9103 *) New functions RSA_get_default_method(), RSA_set_method() and 9104 RSA_get_method(). These allows replacement of RSA_METHODs without having 9105 to mess around with the internals of an RSA structure. 9106 [Steve Henson] 9107 9108 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 9109 Also really enable memory leak checks in openssl.c and in some 9110 test programs. 9111 [Chad C. Mulligan, Bodo Moeller] 9112 9113 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 9114 up the length of negative integers. This has now been simplified to just 9115 store the length when it is first determined and use it later, rather 9116 than trying to keep track of where data is copied and updating it to 9117 point to the end. 9118 [Steve Henson, reported by Brien Wheeler 9119 <bwheeler@authentica-security.com>] 9120 9121 *) Add a new function PKCS7_signatureVerify. This allows the verification 9122 of a PKCS#7 signature but with the signing certificate passed to the 9123 function itself. This contrasts with PKCS7_dataVerify which assumes the 9124 certificate is present in the PKCS#7 structure. This isn't always the 9125 case: certificates can be omitted from a PKCS#7 structure and be 9126 distributed by "out of band" means (such as a certificate database). 9127 [Steve Henson] 9128 9129 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 9130 function prototypes in pem.h, also change util/mkdef.pl to add the 9131 necessary function names. 9132 [Steve Henson] 9133 9134 *) mk1mf.pl (used by Windows builds) did not properly read the 9135 options set by Configure in the top level Makefile, and Configure 9136 was not even able to write more than one option correctly. 9137 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 9138 [Bodo Moeller] 9139 9140 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 9141 file to be loaded from a BIO or FILE pointer. The BIO version will 9142 for example allow memory BIOs to contain config info. 9143 [Steve Henson] 9144 9145 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 9146 Whoever hopes to achieve shared-library compatibility across versions 9147 must use this, not the compile-time macro. 9148 (Exercise 0.9.4: Which is the minimum library version required by 9149 such programs?) 9150 Note: All this applies only to multi-threaded programs, others don't 9151 need locks. 9152 [Bodo Moeller] 9153 9154 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 9155 through a BIO pair triggered the default case, i.e. 9156 SSLerr(...,SSL_R_UNKNOWN_STATE). 9157 [Bodo Moeller] 9158 9159 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 9160 can use the SSL library even if none of the specific BIOs is 9161 appropriate. 9162 [Bodo Moeller] 9163 9164 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 9165 for the encoded length. 9166 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 9167 9168 *) Add initial documentation of the X509V3 functions. 9169 [Steve Henson] 9170 9171 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 9172 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 9173 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 9174 secure PKCS#8 private key format with a high iteration count. 9175 [Steve Henson] 9176 9177 *) Fix determination of Perl interpreter: A perl or perl5 9178 _directory_ in $PATH was also accepted as the interpreter. 9179 [Ralf S. Engelschall] 9180 9181 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 9182 wrong with it but it was very old and did things like calling 9183 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 9184 unusual formatting. 9185 [Steve Henson] 9186 9187 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 9188 to use the new extension code. 9189 [Steve Henson] 9190 9191 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 9192 with macros. This should make it easier to change their form, add extra 9193 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 9194 constant. 9195 [Steve Henson] 9196 9197 *) Add to configuration table a new entry that can specify an alternative 9198 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 9199 according to Mark Crispin <MRC@Panda.COM>. 9200 [Bodo Moeller] 9201 9202#if 0 9203 *) DES CBC did not update the IV. Weird. 9204 [Ben Laurie] 9205#else 9206 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 9207 Changing the behaviour of the former might break existing programs -- 9208 where IV updating is needed, des_ncbc_encrypt can be used. 9209#endif 9210 9211 *) When bntest is run from "make test" it drives bc to check its 9212 calculations, as well as internally checking them. If an internal check 9213 fails, it needs to cause bc to give a non-zero result or make test carries 9214 on without noticing the failure. Fixed. 9215 [Ben Laurie] 9216 9217 *) DES library cleanups. 9218 [Ulf M�ller] 9219 9220 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 9221 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 9222 ciphers. NOTE: although the key derivation function has been verified 9223 against some published test vectors it has not been extensively tested 9224 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 9225 of v2.0. 9226 [Steve Henson] 9227 9228 *) Instead of "mkdir -p", which is not fully portable, use new 9229 Perl script "util/mkdir-p.pl". 9230 [Bodo Moeller] 9231 9232 *) Rewrite the way password based encryption (PBE) is handled. It used to 9233 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 9234 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 9235 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 9236 the 'parameter' field of the AlgorithmIdentifier is passed to the 9237 underlying key generation function so it must do its own ASN1 parsing. 9238 This has also changed the EVP_PBE_CipherInit() function which now has a 9239 'parameter' argument instead of literal salt and iteration count values 9240 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 9241 [Steve Henson] 9242 9243 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 9244 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 9245 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 9246 KEY" because this clashed with PKCS#8 unencrypted string. Since this 9247 value was just used as a "magic string" and not used directly its 9248 value doesn't matter. 9249 [Steve Henson] 9250 9251 *) Introduce some semblance of const correctness to BN. Shame C doesn't 9252 support mutable. 9253 [Ben Laurie] 9254 9255 *) "linux-sparc64" configuration (ultrapenguin). 9256 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 9257 "linux-sparc" configuration. 9258 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 9259 9260 *) config now generates no-xxx options for missing ciphers. 9261 [Ulf M�ller] 9262 9263 *) Support the EBCDIC character set (work in progress). 9264 File ebcdic.c not yet included because it has a different license. 9265 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 9266 9267 *) Support BS2000/OSD-POSIX. 9268 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 9269 9270 *) Make callbacks for key generation use void * instead of char *. 9271 [Ben Laurie] 9272 9273 *) Make S/MIME samples compile (not yet tested). 9274 [Ben Laurie] 9275 9276 *) Additional typesafe stacks. 9277 [Ben Laurie] 9278 9279 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 9280 [Bodo Moeller] 9281 9282 9283 Changes between 0.9.3 and 0.9.3a [29 May 1999] 9284 9285 *) New configuration variant "sco5-gcc". 9286 9287 *) Updated some demos. 9288 [Sean O Riordain, Wade Scholine] 9289 9290 *) Add missing BIO_free at exit of pkcs12 application. 9291 [Wu Zhigang] 9292 9293 *) Fix memory leak in conf.c. 9294 [Steve Henson] 9295 9296 *) Updates for Win32 to assembler version of MD5. 9297 [Steve Henson] 9298 9299 *) Set #! path to perl in apps/der_chop to where we found it 9300 instead of using a fixed path. 9301 [Bodo Moeller] 9302 9303 *) SHA library changes for irix64-mips4-cc. 9304 [Andy Polyakov] 9305 9306 *) Improvements for VMS support. 9307 [Richard Levitte] 9308 9309 9310 Changes between 0.9.2b and 0.9.3 [24 May 1999] 9311 9312 *) Bignum library bug fix. IRIX 6 passes "make test" now! 9313 This also avoids the problems with SC4.2 and unpatched SC5. 9314 [Andy Polyakov <appro@fy.chalmers.se>] 9315 9316 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 9317 These are required because of the typesafe stack would otherwise break 9318 existing code. If old code used a structure member which used to be STACK 9319 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 9320 sk_num or sk_value it would produce an error because the num, data members 9321 are not present in STACK_OF. Now it just produces a warning. sk_set 9322 replaces the old method of assigning a value to sk_value 9323 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 9324 that does this will no longer work (and should use sk_set instead) but 9325 this could be regarded as a "questionable" behaviour anyway. 9326 [Steve Henson] 9327 9328 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 9329 correctly handle encrypted S/MIME data. 9330 [Steve Henson] 9331 9332 *) Change type of various DES function arguments from des_cblock 9333 (which means, in function argument declarations, pointer to char) 9334 to des_cblock * (meaning pointer to array with 8 char elements), 9335 which allows the compiler to do more typechecking; it was like 9336 that back in SSLeay, but with lots of ugly casts. 9337 9338 Introduce new type const_des_cblock. 9339 [Bodo Moeller] 9340 9341 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 9342 problems: find RecipientInfo structure that matches recipient certificate 9343 and initialise the ASN1 structures properly based on passed cipher. 9344 [Steve Henson] 9345 9346 *) Belatedly make the BN tests actually check the results. 9347 [Ben Laurie] 9348 9349 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 9350 to and from BNs: it was completely broken. New compilation option 9351 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 9352 key elements as negative integers. 9353 [Steve Henson] 9354 9355 *) Reorganize and speed up MD5. 9356 [Andy Polyakov <appro@fy.chalmers.se>] 9357 9358 *) VMS support. 9359 [Richard Levitte <richard@levitte.org>] 9360 9361 *) New option -out to asn1parse to allow the parsed structure to be 9362 output to a file. This is most useful when combined with the -strparse 9363 option to examine the output of things like OCTET STRINGS. 9364 [Steve Henson] 9365 9366 *) Make SSL library a little more fool-proof by not requiring any longer 9367 that SSL_set_{accept,connect}_state be called before 9368 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 9369 in many applications because usually everything *appeared* to work as 9370 intended anyway -- now it really works as intended). 9371 [Bodo Moeller] 9372 9373 *) Move openssl.cnf out of lib/. 9374 [Ulf M�ller] 9375 9376 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 9377 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 9378 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 9379 [Ralf S. Engelschall] 9380 9381 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 9382 handle PKCS#7 enveloped data properly. 9383 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 9384 9385 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 9386 copying pointers. The cert_st handling is changed by this in 9387 various ways (and thus what used to be known as ctx->default_cert 9388 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 9389 any longer when s->cert does not give us what we need). 9390 ssl_cert_instantiate becomes obsolete by this change. 9391 As soon as we've got the new code right (possibly it already is?), 9392 we have solved a couple of bugs of the earlier code where s->cert 9393 was used as if it could not have been shared with other SSL structures. 9394 9395 Note that using the SSL API in certain dirty ways now will result 9396 in different behaviour than observed with earlier library versions: 9397 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 9398 does not influence s as it used to. 9399 9400 In order to clean up things more thoroughly, inside SSL_SESSION 9401 we don't use CERT any longer, but a new structure SESS_CERT 9402 that holds per-session data (if available); currently, this is 9403 the peer's certificate chain and, for clients, the server's certificate 9404 and temporary key. CERT holds only those values that can have 9405 meaningful defaults in an SSL_CTX. 9406 [Bodo Moeller] 9407 9408 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 9409 from the internal representation. Various PKCS#7 fixes: remove some 9410 evil casts and set the enc_dig_alg field properly based on the signing 9411 key type. 9412 [Steve Henson] 9413 9414 *) Allow PKCS#12 password to be set from the command line or the 9415 environment. Let 'ca' get its config file name from the environment 9416 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 9417 and 'x509'). 9418 [Steve Henson] 9419 9420 *) Allow certificate policies extension to use an IA5STRING for the 9421 organization field. This is contrary to the PKIX definition but 9422 VeriSign uses it and IE5 only recognises this form. Document 'x509' 9423 extension option. 9424 [Steve Henson] 9425 9426 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 9427 without disallowing inline assembler and the like for non-pedantic builds. 9428 [Ben Laurie] 9429 9430 *) Support Borland C++ builder. 9431 [Janez Jere <jj@void.si>, modified by Ulf M�ller] 9432 9433 *) Support Mingw32. 9434 [Ulf M�ller] 9435 9436 *) SHA-1 cleanups and performance enhancements. 9437 [Andy Polyakov <appro@fy.chalmers.se>] 9438 9439 *) Sparc v8plus assembler for the bignum library. 9440 [Andy Polyakov <appro@fy.chalmers.se>] 9441 9442 *) Accept any -xxx and +xxx compiler options in Configure. 9443 [Ulf M�ller] 9444 9445 *) Update HPUX configuration. 9446 [Anonymous] 9447 9448 *) Add missing sk_<type>_unshift() function to safestack.h 9449 [Ralf S. Engelschall] 9450 9451 *) New function SSL_CTX_use_certificate_chain_file that sets the 9452 "extra_cert"s in addition to the certificate. (This makes sense 9453 only for "PEM" format files, as chains as a whole are not 9454 DER-encoded.) 9455 [Bodo Moeller] 9456 9457 *) Support verify_depth from the SSL API. 9458 x509_vfy.c had what can be considered an off-by-one-error: 9459 Its depth (which was not part of the external interface) 9460 was actually counting the number of certificates in a chain; 9461 now it really counts the depth. 9462 [Bodo Moeller] 9463 9464 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 9465 instead of X509err, which often resulted in confusing error 9466 messages since the error codes are not globally unique 9467 (e.g. an alleged error in ssl3_accept when a certificate 9468 didn't match the private key). 9469 9470 *) New function SSL_CTX_set_session_id_context that allows to set a default 9471 value (so that you don't need SSL_set_session_id_context for each 9472 connection using the SSL_CTX). 9473 [Bodo Moeller] 9474 9475 *) OAEP decoding bug fix. 9476 [Ulf M�ller] 9477 9478 *) Support INSTALL_PREFIX for package builders, as proposed by 9479 David Harris. 9480 [Bodo Moeller] 9481 9482 *) New Configure options "threads" and "no-threads". For systems 9483 where the proper compiler options are known (currently Solaris 9484 and Linux), "threads" is the default. 9485 [Bodo Moeller] 9486 9487 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 9488 [Bodo Moeller] 9489 9490 *) Install various scripts to $(OPENSSLDIR)/misc, not to 9491 $(INSTALLTOP)/bin -- they shouldn't clutter directories 9492 such as /usr/local/bin. 9493 [Bodo Moeller] 9494 9495 *) "make linux-shared" to build shared libraries. 9496 [Niels Poppe <niels@netbox.org>] 9497 9498 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 9499 [Ulf M�ller] 9500 9501 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 9502 extension adding in x509 utility. 9503 [Steve Henson] 9504 9505 *) Remove NOPROTO sections and error code comments. 9506 [Ulf M�ller] 9507 9508 *) Partial rewrite of the DEF file generator to now parse the ANSI 9509 prototypes. 9510 [Steve Henson] 9511 9512 *) New Configure options --prefix=DIR and --openssldir=DIR. 9513 [Ulf M�ller] 9514 9515 *) Complete rewrite of the error code script(s). It is all now handled 9516 by one script at the top level which handles error code gathering, 9517 header rewriting and C source file generation. It should be much better 9518 than the old method: it now uses a modified version of Ulf's parser to 9519 read the ANSI prototypes in all header files (thus the old K&R definitions 9520 aren't needed for error creation any more) and do a better job of 9521 translating function codes into names. The old 'ASN1 error code imbedded 9522 in a comment' is no longer necessary and it doesn't use .err files which 9523 have now been deleted. Also the error code call doesn't have to appear all 9524 on one line (which resulted in some large lines...). 9525 [Steve Henson] 9526 9527 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 9528 [Bodo Moeller] 9529 9530 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 9531 0 (which usually indicates a closed connection), but continue reading. 9532 [Bodo Moeller] 9533 9534 *) Fix some race conditions. 9535 [Bodo Moeller] 9536 9537 *) Add support for CRL distribution points extension. Add Certificate 9538 Policies and CRL distribution points documentation. 9539 [Steve Henson] 9540 9541 *) Move the autogenerated header file parts to crypto/opensslconf.h. 9542 [Ulf M�ller] 9543 9544 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 9545 8 of keying material. Merlin has also confirmed interop with this fix 9546 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 9547 [Merlin Hughes <merlin@baltimore.ie>] 9548 9549 *) Fix lots of warnings. 9550 [Richard Levitte <levitte@stacken.kth.se>] 9551 9552 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 9553 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 9554 [Richard Levitte <levitte@stacken.kth.se>] 9555 9556 *) Fix problems with sizeof(long) == 8. 9557 [Andy Polyakov <appro@fy.chalmers.se>] 9558 9559 *) Change functions to ANSI C. 9560 [Ulf M�ller] 9561 9562 *) Fix typos in error codes. 9563 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 9564 9565 *) Remove defunct assembler files from Configure. 9566 [Ulf M�ller] 9567 9568 *) SPARC v8 assembler BIGNUM implementation. 9569 [Andy Polyakov <appro@fy.chalmers.se>] 9570 9571 *) Support for Certificate Policies extension: both print and set. 9572 Various additions to support the r2i method this uses. 9573 [Steve Henson] 9574 9575 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 9576 return a const string when you are expecting an allocated buffer. 9577 [Ben Laurie] 9578 9579 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 9580 types DirectoryString and DisplayText. 9581 [Steve Henson] 9582 9583 *) Add code to allow r2i extensions to access the configuration database, 9584 add an LHASH database driver and add several ctx helper functions. 9585 [Steve Henson] 9586 9587 *) Fix an evil bug in bn_expand2() which caused various BN functions to 9588 fail when they extended the size of a BIGNUM. 9589 [Steve Henson] 9590 9591 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 9592 support typesafe stack. 9593 [Steve Henson] 9594 9595 *) Fix typo in SSL_[gs]et_options(). 9596 [Nils Frostberg <nils@medcom.se>] 9597 9598 *) Delete various functions and files that belonged to the (now obsolete) 9599 old X509V3 handling code. 9600 [Steve Henson] 9601 9602 *) New Configure option "rsaref". 9603 [Ulf M�ller] 9604 9605 *) Don't auto-generate pem.h. 9606 [Bodo Moeller] 9607 9608 *) Introduce type-safe ASN.1 SETs. 9609 [Ben Laurie] 9610 9611 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 9612 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 9613 9614 *) Introduce type-safe STACKs. This will almost certainly break lots of code 9615 that links with OpenSSL (well at least cause lots of warnings), but fear 9616 not: the conversion is trivial, and it eliminates loads of evil casts. A 9617 few STACKed things have been converted already. Feel free to convert more. 9618 In the fullness of time, I'll do away with the STACK type altogether. 9619 [Ben Laurie] 9620 9621 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 9622 specified in <certfile> by updating the entry in the index.txt file. 9623 This way one no longer has to edit the index.txt file manually for 9624 revoking a certificate. The -revoke option does the gory details now. 9625 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 9626 9627 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 9628 `-text' option at all and this way the `-noout -text' combination was 9629 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 9630 [Ralf S. Engelschall] 9631 9632 *) Make sure a corresponding plain text error message exists for the 9633 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 9634 verify callback function determined that a certificate was revoked. 9635 [Ralf S. Engelschall] 9636 9637 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 9638 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 9639 all available cipers including rc5, which was forgotten until now. 9640 In order to let the testing shell script know which algorithms 9641 are available, a new (up to now undocumented) command 9642 "openssl list-cipher-commands" is used. 9643 [Bodo Moeller] 9644 9645 *) Bugfix: s_client occasionally would sleep in select() when 9646 it should have checked SSL_pending() first. 9647 [Bodo Moeller] 9648 9649 *) New functions DSA_do_sign and DSA_do_verify to provide access to 9650 the raw DSA values prior to ASN.1 encoding. 9651 [Ulf M�ller] 9652 9653 *) Tweaks to Configure 9654 [Niels Poppe <niels@netbox.org>] 9655 9656 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 9657 yet... 9658 [Steve Henson] 9659 9660 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 9661 [Ulf M�ller] 9662 9663 *) New config option to avoid instructions that are illegal on the 80386. 9664 The default code is faster, but requires at least a 486. 9665 [Ulf M�ller] 9666 9667 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 9668 SSL2_SERVER_VERSION (not used at all) macros, which are now the 9669 same as SSL2_VERSION anyway. 9670 [Bodo Moeller] 9671 9672 *) New "-showcerts" option for s_client. 9673 [Bodo Moeller] 9674 9675 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 9676 application. Various cleanups and fixes. 9677 [Steve Henson] 9678 9679 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 9680 modify error routines to work internally. Add error codes and PBE init 9681 to library startup routines. 9682 [Steve Henson] 9683 9684 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 9685 packing functions to asn1 and evp. Changed function names and error 9686 codes along the way. 9687 [Steve Henson] 9688 9689 *) PKCS12 integration: and so it begins... First of several patches to 9690 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 9691 objects to objects.h 9692 [Steve Henson] 9693 9694 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 9695 and display support for Thawte strong extranet extension. 9696 [Steve Henson] 9697 9698 *) Add LinuxPPC support. 9699 [Jeff Dubrule <igor@pobox.org>] 9700 9701 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 9702 bn_div_words in alpha.s. 9703 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 9704 9705 *) Make sure the RSA OAEP test is skipped under -DRSAref because 9706 OAEP isn't supported when OpenSSL is built with RSAref. 9707 [Ulf Moeller <ulf@fitug.de>] 9708 9709 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 9710 so they no longer are missing under -DNOPROTO. 9711 [Soren S. Jorvang <soren@t.dk>] 9712 9713 9714 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 9715 9716 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 9717 doesn't work when the session is reused. Coming soon! 9718 [Ben Laurie] 9719 9720 *) Fix a security hole, that allows sessions to be reused in the wrong 9721 context thus bypassing client cert protection! All software that uses 9722 client certs and session caches in multiple contexts NEEDS PATCHING to 9723 allow session reuse! A fuller solution is in the works. 9724 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 9725 9726 *) Some more source tree cleanups (removed obsolete files 9727 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 9728 permission on "config" script to be executable) and a fix for the INSTALL 9729 document. 9730 [Ulf Moeller <ulf@fitug.de>] 9731 9732 *) Remove some legacy and erroneous uses of malloc, free instead of 9733 Malloc, Free. 9734 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 9735 9736 *) Make rsa_oaep_test return non-zero on error. 9737 [Ulf Moeller <ulf@fitug.de>] 9738 9739 *) Add support for native Solaris shared libraries. Configure 9740 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 9741 if someone would make that last step automatic. 9742 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 9743 9744 *) ctx_size was not built with the right compiler during "make links". Fixed. 9745 [Ben Laurie] 9746 9747 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 9748 except NULL ciphers". This means the default cipher list will no longer 9749 enable NULL ciphers. They need to be specifically enabled e.g. with 9750 the string "DEFAULT:eNULL". 9751 [Steve Henson] 9752 9753 *) Fix to RSA private encryption routines: if p < q then it would 9754 occasionally produce an invalid result. This will only happen with 9755 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 9756 [Steve Henson] 9757 9758 *) Be less restrictive and allow also `perl util/perlpath.pl 9759 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 9760 because this way one can also use an interpreter named `perl5' (which is 9761 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 9762 installed as `perl'). 9763 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 9764 9765 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 9766 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 9767 9768 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 9769 advapi32.lib to Win32 build and change the pem test comparision 9770 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 9771 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 9772 and crypto/des/ede_cbcm_enc.c. 9773 [Steve Henson] 9774 9775 *) DES quad checksum was broken on big-endian architectures. Fixed. 9776 [Ben Laurie] 9777 9778 *) Comment out two functions in bio.h that aren't implemented. Fix up the 9779 Win32 test batch file so it (might) work again. The Win32 test batch file 9780 is horrible: I feel ill.... 9781 [Steve Henson] 9782 9783 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 9784 in e_os.h. Audit of header files to check ANSI and non ANSI 9785 sections: 10 functions were absent from non ANSI section and not exported 9786 from Windows DLLs. Fixed up libeay.num for new functions. 9787 [Steve Henson] 9788 9789 *) Make `openssl version' output lines consistent. 9790 [Ralf S. Engelschall] 9791 9792 *) Fix Win32 symbol export lists for BIO functions: Added 9793 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 9794 to ms/libeay{16,32}.def. 9795 [Ralf S. Engelschall] 9796 9797 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 9798 fine under Unix and passes some trivial tests I've now added. But the 9799 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 9800 added to make sure no one expects that this stuff really works in the 9801 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 9802 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 9803 openssl_bio.xs. 9804 [Ralf S. Engelschall] 9805 9806 *) Fix the generation of two part addresses in perl. 9807 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 9808 9809 *) Add config entry for Linux on MIPS. 9810 [John Tobey <jtobey@channel1.com>] 9811 9812 *) Make links whenever Configure is run, unless we are on Windoze. 9813 [Ben Laurie] 9814 9815 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 9816 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 9817 in CRLs. 9818 [Steve Henson] 9819 9820 *) Add a useful kludge to allow package maintainers to specify compiler and 9821 other platforms details on the command line without having to patch the 9822 Configure script everytime: One now can use ``perl Configure 9823 <id>:<details>'', i.e. platform ids are allowed to have details appended 9824 to them (seperated by colons). This is treated as there would be a static 9825 pre-configured entry in Configure's %table under key <id> with value 9826 <details> and ``perl Configure <id>'' is called. So, when you want to 9827 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 9828 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 9829 now, which overrides the FreeBSD-elf entry on-the-fly. 9830 [Ralf S. Engelschall] 9831 9832 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 9833 [Ben Laurie] 9834 9835 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 9836 on the `perl Configure ...' command line. This way one can compile 9837 OpenSSL libraries with Position Independent Code (PIC) which is needed 9838 for linking it into DSOs. 9839 [Ralf S. Engelschall] 9840 9841 *) Remarkably, export ciphers were totally broken and no-one had noticed! 9842 Fixed. 9843 [Ben Laurie] 9844 9845 *) Cleaned up the LICENSE document: The official contact for any license 9846 questions now is the OpenSSL core team under openssl-core@openssl.org. 9847 And add a paragraph about the dual-license situation to make sure people 9848 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 9849 to the OpenSSL toolkit. 9850 [Ralf S. Engelschall] 9851 9852 *) General source tree makefile cleanups: Made `making xxx in yyy...' 9853 display consistent in the source tree and replaced `/bin/rm' by `rm'. 9854 Additonally cleaned up the `make links' target: Remove unnecessary 9855 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 9856 to speed processing and no longer clutter the display with confusing 9857 stuff. Instead only the actually done links are displayed. 9858 [Ralf S. Engelschall] 9859 9860 *) Permit null encryption ciphersuites, used for authentication only. It used 9861 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 9862 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 9863 encryption. 9864 [Ben Laurie] 9865 9866 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 9867 signed attributes when verifying signatures (this would break them), 9868 the detached data encoding was wrong and public keys obtained using 9869 X509_get_pubkey() weren't freed. 9870 [Steve Henson] 9871 9872 *) Add text documentation for the BUFFER functions. Also added a work around 9873 to a Win95 console bug. This was triggered by the password read stuff: the 9874 last character typed gets carried over to the next fread(). If you were 9875 generating a new cert request using 'req' for example then the last 9876 character of the passphrase would be CR which would then enter the first 9877 field as blank. 9878 [Steve Henson] 9879 9880 *) Added the new `Includes OpenSSL Cryptography Software' button as 9881 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 9882 button and can be used by applications based on OpenSSL to show the 9883 relationship to the OpenSSL project. 9884 [Ralf S. Engelschall] 9885 9886 *) Remove confusing variables in function signatures in files 9887 ssl/ssl_lib.c and ssl/ssl.h. 9888 [Lennart Bong <lob@kulthea.stacken.kth.se>] 9889 9890 *) Don't install bss_file.c under PREFIX/include/ 9891 [Lennart Bong <lob@kulthea.stacken.kth.se>] 9892 9893 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 9894 functions that return function pointers and has support for NT specific 9895 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 9896 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 9897 unsigned to signed types: this was killing the Win32 compile. 9898 [Steve Henson] 9899 9900 *) Add new certificate file to stack functions, 9901 SSL_add_dir_cert_subjects_to_stack() and 9902 SSL_add_file_cert_subjects_to_stack(). These largely supplant 9903 SSL_load_client_CA_file(), and can be used to add multiple certs easily 9904 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 9905 This means that Apache-SSL and similar packages don't have to mess around 9906 to add as many CAs as they want to the preferred list. 9907 [Ben Laurie] 9908 9909 *) Experiment with doxygen documentation. Currently only partially applied to 9910 ssl/ssl_lib.c. 9911 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 9912 openssl.doxy as the configuration file. 9913 [Ben Laurie] 9914 9915 *) Get rid of remaining C++-style comments which strict C compilers hate. 9916 [Ralf S. Engelschall, pointed out by Carlos Amengual] 9917 9918 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 9919 compiled in by default: it has problems with large keys. 9920 [Steve Henson] 9921 9922 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 9923 DH private keys and/or callback functions which directly correspond to 9924 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 9925 is needed for applications which have to configure certificates on a 9926 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 9927 (e.g. s_server). 9928 For the RSA certificate situation is makes no difference, but 9929 for the DSA certificate situation this fixes the "no shared cipher" 9930 problem where the OpenSSL cipher selection procedure failed because the 9931 temporary keys were not overtaken from the context and the API provided 9932 no way to reconfigure them. 9933 The new functions now let applications reconfigure the stuff and they 9934 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 9935 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 9936 non-public-API function ssl_cert_instantiate() is used as a helper 9937 function and also to reduce code redundancy inside ssl_rsa.c. 9938 [Ralf S. Engelschall] 9939 9940 *) Move s_server -dcert and -dkey options out of the undocumented feature 9941 area because they are useful for the DSA situation and should be 9942 recognized by the users. 9943 [Ralf S. Engelschall] 9944 9945 *) Fix the cipher decision scheme for export ciphers: the export bits are 9946 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 9947 SSL_EXP_MASK. So, the original variable has to be used instead of the 9948 already masked variable. 9949 [Richard Levitte <levitte@stacken.kth.se>] 9950 9951 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 9952 [Richard Levitte <levitte@stacken.kth.se>] 9953 9954 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 9955 from `int' to `unsigned int' because it's a length and initialized by 9956 EVP_DigestFinal() which expects an `unsigned int *'. 9957 [Richard Levitte <levitte@stacken.kth.se>] 9958 9959 *) Don't hard-code path to Perl interpreter on shebang line of Configure 9960 script. Instead use the usual Shell->Perl transition trick. 9961 [Ralf S. Engelschall] 9962 9963 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 9964 (in addition to RSA certificates) to match the behaviour of `openssl dsa 9965 -noout -modulus' as it's already the case for `openssl rsa -noout 9966 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 9967 currently the public key is printed (a decision which was already done by 9968 `openssl dsa -modulus' in the past) which serves a similar purpose. 9969 Additionally the NO_RSA no longer completely removes the whole -modulus 9970 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 9971 now, too. 9972 [Ralf S. Engelschall] 9973 9974 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 9975 BIO. See the source (crypto/evp/bio_ok.c) for more info. 9976 [Arne Ansper <arne@ats.cyber.ee>] 9977 9978 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 9979 to be added. Now both 'req' and 'ca' can use new objects defined in the 9980 config file. 9981 [Steve Henson] 9982 9983 *) Add cool BIO that does syslog (or event log on NT). 9984 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 9985 9986 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 9987 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 9988 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 9989 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 9990 [Ben Laurie] 9991 9992 *) Add preliminary config info for new extension code. 9993 [Steve Henson] 9994 9995 *) Make RSA_NO_PADDING really use no padding. 9996 [Ulf Moeller <ulf@fitug.de>] 9997 9998 *) Generate errors when private/public key check is done. 9999 [Ben Laurie] 10000 10001 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 10002 for some CRL extensions and new objects added. 10003 [Steve Henson] 10004 10005 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 10006 key usage extension and fuller support for authority key id. 10007 [Steve Henson] 10008 10009 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 10010 padding method for RSA, which is recommended for new applications in PKCS 10011 #1 v2.0 (RFC 2437, October 1998). 10012 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 10013 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 10014 against Bleichbacher's attack on RSA. 10015 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 10016 Ben Laurie] 10017 10018 *) Updates to the new SSL compression code 10019 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10020 10021 *) Fix so that the version number in the master secret, when passed 10022 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 10023 (because the server will not accept higher), that the version number 10024 is 0x03,0x01, not 0x03,0x00 10025 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10026 10027 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 10028 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 10029 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 10030 [Steve Henson] 10031 10032 *) Support for RAW extensions where an arbitrary extension can be 10033 created by including its DER encoding. See apps/openssl.cnf for 10034 an example. 10035 [Steve Henson] 10036 10037 *) Make sure latest Perl versions don't interpret some generated C array 10038 code as Perl array code in the crypto/err/err_genc.pl script. 10039 [Lars Weber <3weber@informatik.uni-hamburg.de>] 10040 10041 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 10042 not many people have the assembler. Various Win32 compilation fixes and 10043 update to the INSTALL.W32 file with (hopefully) more accurate Win32 10044 build instructions. 10045 [Steve Henson] 10046 10047 *) Modify configure script 'Configure' to automatically create crypto/date.h 10048 file under Win32 and also build pem.h from pem.org. New script 10049 util/mkfiles.pl to create the MINFO file on environments that can't do a 10050 'make files': perl util/mkfiles.pl >MINFO should work. 10051 [Steve Henson] 10052 10053 *) Major rework of DES function declarations, in the pursuit of correctness 10054 and purity. As a result, many evil casts evaporated, and some weirdness, 10055 too. You may find this causes warnings in your code. Zapping your evil 10056 casts will probably fix them. Mostly. 10057 [Ben Laurie] 10058 10059 *) Fix for a typo in asn1.h. Bug fix to object creation script 10060 obj_dat.pl. It considered a zero in an object definition to mean 10061 "end of object": none of the objects in objects.h have any zeros 10062 so it wasn't spotted. 10063 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 10064 10065 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 10066 Masking (CBCM). In the absence of test vectors, the best I have been able 10067 to do is check that the decrypt undoes the encrypt, so far. Send me test 10068 vectors if you have them. 10069 [Ben Laurie] 10070 10071 *) Correct calculation of key length for export ciphers (too much space was 10072 allocated for null ciphers). This has not been tested! 10073 [Ben Laurie] 10074 10075 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 10076 message is now correct (it understands "crypto" and "ssl" on its 10077 command line). There is also now an "update" option. This will update 10078 the util/ssleay.num and util/libeay.num files with any new functions. 10079 If you do a: 10080 perl util/mkdef.pl crypto ssl update 10081 it will update them. 10082 [Steve Henson] 10083 10084 *) Overhauled the Perl interface (perl/*): 10085 - ported BN stuff to OpenSSL's different BN library 10086 - made the perl/ source tree CVS-aware 10087 - renamed the package from SSLeay to OpenSSL (the files still contain 10088 their history because I've copied them in the repository) 10089 - removed obsolete files (the test scripts will be replaced 10090 by better Test::Harness variants in the future) 10091 [Ralf S. Engelschall] 10092 10093 *) First cut for a very conservative source tree cleanup: 10094 1. merge various obsolete readme texts into doc/ssleay.txt 10095 where we collect the old documents and readme texts. 10096 2. remove the first part of files where I'm already sure that we no 10097 longer need them because of three reasons: either they are just temporary 10098 files which were left by Eric or they are preserved original files where 10099 I've verified that the diff is also available in the CVS via "cvs diff 10100 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 10101 the crypto/md/ stuff). 10102 [Ralf S. Engelschall] 10103 10104 *) More extension code. Incomplete support for subject and issuer alt 10105 name, issuer and authority key id. Change the i2v function parameters 10106 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 10107 what that's for :-) Fix to ASN1 macro which messed up 10108 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 10109 [Steve Henson] 10110 10111 *) Preliminary support for ENUMERATED type. This is largely copied from the 10112 INTEGER code. 10113 [Steve Henson] 10114 10115 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 10116 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10117 10118 *) Make sure `make rehash' target really finds the `openssl' program. 10119 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10120 10121 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 10122 like to hear about it if this slows down other processors. 10123 [Ben Laurie] 10124 10125 *) Add CygWin32 platform information to Configure script. 10126 [Alan Batie <batie@aahz.jf.intel.com>] 10127 10128 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 10129 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 10130 10131 *) New program nseq to manipulate netscape certificate sequences 10132 [Steve Henson] 10133 10134 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 10135 few typos. 10136 [Steve Henson] 10137 10138 *) Fixes to BN code. Previously the default was to define BN_RECURSION 10139 but the BN code had some problems that would cause failures when 10140 doing certificate verification and some other functions. 10141 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10142 10143 *) Add ASN1 and PEM code to support netscape certificate sequences. 10144 [Steve Henson] 10145 10146 *) Add ASN1 and PEM code to support netscape certificate sequences. 10147 [Steve Henson] 10148 10149 *) Add several PKIX and private extended key usage OIDs. 10150 [Steve Henson] 10151 10152 *) Modify the 'ca' program to handle the new extension code. Modify 10153 openssl.cnf for new extension format, add comments. 10154 [Steve Henson] 10155 10156 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 10157 and add a sample to openssl.cnf so req -x509 now adds appropriate 10158 CA extensions. 10159 [Steve Henson] 10160 10161 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 10162 error code, add initial support to X509_print() and x509 application. 10163 [Steve Henson] 10164 10165 *) Takes a deep breath and start addding X509 V3 extension support code. Add 10166 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 10167 stuff is currently isolated and isn't even compiled yet. 10168 [Steve Henson] 10169 10170 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 10171 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 10172 Removed the versions check from X509 routines when loading extensions: 10173 this allows certain broken certificates that don't set the version 10174 properly to be processed. 10175 [Steve Henson] 10176 10177 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 10178 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 10179 can still be regenerated with "make depend". 10180 [Ben Laurie] 10181 10182 *) Spelling mistake in C version of CAST-128. 10183 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 10184 10185 *) Changes to the error generation code. The perl script err-code.pl 10186 now reads in the old error codes and retains the old numbers, only 10187 adding new ones if necessary. It also only changes the .err files if new 10188 codes are added. The makefiles have been modified to only insert errors 10189 when needed (to avoid needlessly modifying header files). This is done 10190 by only inserting errors if the .err file is newer than the auto generated 10191 C file. To rebuild all the error codes from scratch (the old behaviour) 10192 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 10193 or delete all the .err files. 10194 [Steve Henson] 10195 10196 *) CAST-128 was incorrectly implemented for short keys. The C version has 10197 been fixed, but is untested. The assembler versions are also fixed, but 10198 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 10199 to regenerate it if needed. 10200 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 10201 Hagino <itojun@kame.net>] 10202 10203 *) File was opened incorrectly in randfile.c. 10204 [Ulf M�ller <ulf@fitug.de>] 10205 10206 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 10207 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 10208 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 10209 al: it's just almost always a UTCTime. Note this patch adds new error 10210 codes so do a "make errors" if there are problems. 10211 [Steve Henson] 10212 10213 *) Correct Linux 1 recognition in config. 10214 [Ulf M�ller <ulf@fitug.de>] 10215 10216 *) Remove pointless MD5 hash when using DSA keys in ca. 10217 [Anonymous <nobody@replay.com>] 10218 10219 *) Generate an error if given an empty string as a cert directory. Also 10220 generate an error if handed NULL (previously returned 0 to indicate an 10221 error, but didn't set one). 10222 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 10223 10224 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 10225 [Ben Laurie] 10226 10227 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 10228 parameters. This was causing a warning which killed off the Win32 compile. 10229 [Steve Henson] 10230 10231 *) Remove C++ style comments from crypto/bn/bn_local.h. 10232 [Neil Costigan <neil.costigan@celocom.com>] 10233 10234 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 10235 based on a text string, looking up short and long names and finally 10236 "dot" format. The "dot" format stuff didn't work. Added new function 10237 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 10238 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 10239 OID is not part of the table. 10240 [Steve Henson] 10241 10242 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 10243 X509_LOOKUP_by_alias(). 10244 [Ben Laurie] 10245 10246 *) Sort openssl functions by name. 10247 [Ben Laurie] 10248 10249 *) Get the gendsa program working (hopefully) and add it to app list. Remove 10250 encryption from sample DSA keys (in case anyone is interested the password 10251 was "1234"). 10252 [Steve Henson] 10253 10254 *) Make _all_ *_free functions accept a NULL pointer. 10255 [Frans Heymans <fheymans@isaserver.be>] 10256 10257 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 10258 NULL pointers. 10259 [Anonymous <nobody@replay.com>] 10260 10261 *) s_server should send the CAfile as acceptable CAs, not its own cert. 10262 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 10263 10264 *) Don't blow it for numeric -newkey arguments to apps/req. 10265 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 10266 10267 *) Temp key "for export" tests were wrong in s3_srvr.c. 10268 [Anonymous <nobody@replay.com>] 10269 10270 *) Add prototype for temp key callback functions 10271 SSL_CTX_set_tmp_{rsa,dh}_callback(). 10272 [Ben Laurie] 10273 10274 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 10275 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 10276 [Steve Henson] 10277 10278 *) X509_name_add_entry() freed the wrong thing after an error. 10279 [Arne Ansper <arne@ats.cyber.ee>] 10280 10281 *) rsa_eay.c would attempt to free a NULL context. 10282 [Arne Ansper <arne@ats.cyber.ee>] 10283 10284 *) BIO_s_socket() had a broken should_retry() on Windoze. 10285 [Arne Ansper <arne@ats.cyber.ee>] 10286 10287 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 10288 [Arne Ansper <arne@ats.cyber.ee>] 10289 10290 *) Make sure the already existing X509_STORE->depth variable is initialized 10291 in X509_STORE_new(), but document the fact that this variable is still 10292 unused in the certificate verification process. 10293 [Ralf S. Engelschall] 10294 10295 *) Fix the various library and apps files to free up pkeys obtained from 10296 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 10297 [Steve Henson] 10298 10299 *) Fix reference counting in X509_PUBKEY_get(). This makes 10300 demos/maurice/example2.c work, amongst others, probably. 10301 [Steve Henson and Ben Laurie] 10302 10303 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 10304 `openssl' and second, the shortcut symlinks for the `openssl <command>' 10305 are no longer created. This way we have a single and consistent command 10306 line interface `openssl <command>', similar to `cvs <command>'. 10307 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 10308 10309 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 10310 BIT STRING wrapper always have zero unused bits. 10311 [Steve Henson] 10312 10313 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 10314 [Steve Henson] 10315 10316 *) Make the top-level INSTALL documentation easier to understand. 10317 [Paul Sutton] 10318 10319 *) Makefiles updated to exit if an error occurs in a sub-directory 10320 make (including if user presses ^C) [Paul Sutton] 10321 10322 *) Make Montgomery context stuff explicit in RSA data structure. 10323 [Ben Laurie] 10324 10325 *) Fix build order of pem and err to allow for generated pem.h. 10326 [Ben Laurie] 10327 10328 *) Fix renumbering bug in X509_NAME_delete_entry(). 10329 [Ben Laurie] 10330 10331 *) Enhanced the err-ins.pl script so it makes the error library number 10332 global and can add a library name. This is needed for external ASN1 and 10333 other error libraries. 10334 [Steve Henson] 10335 10336 *) Fixed sk_insert which never worked properly. 10337 [Steve Henson] 10338 10339 *) Fix ASN1 macros so they can handle indefinite length construted 10340 EXPLICIT tags. Some non standard certificates use these: they can now 10341 be read in. 10342 [Steve Henson] 10343 10344 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 10345 into a single doc/ssleay.txt bundle. This way the information is still 10346 preserved but no longer messes up this directory. Now it's new room for 10347 the new set of documenation files. 10348 [Ralf S. Engelschall] 10349 10350 *) SETs were incorrectly DER encoded. This was a major pain, because they 10351 shared code with SEQUENCEs, which aren't coded the same. This means that 10352 almost everything to do with SETs or SEQUENCEs has either changed name or 10353 number of arguments. 10354 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 10355 10356 *) Fix test data to work with the above. 10357 [Ben Laurie] 10358 10359 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 10360 was already fixed by Eric for 0.9.1 it seems. 10361 [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 10362 10363 *) Autodetect FreeBSD3. 10364 [Ben Laurie] 10365 10366 *) Fix various bugs in Configure. This affects the following platforms: 10367 nextstep 10368 ncr-scde 10369 unixware-2.0 10370 unixware-2.0-pentium 10371 sco5-cc. 10372 [Ben Laurie] 10373 10374 *) Eliminate generated files from CVS. Reorder tests to regenerate files 10375 before they are needed. 10376 [Ben Laurie] 10377 10378 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 10379 [Ben Laurie] 10380 10381 10382 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 10383 10384 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 10385 changed SSLeay to OpenSSL in version strings. 10386 [Ralf S. Engelschall] 10387 10388 *) Some fixups to the top-level documents. 10389 [Paul Sutton] 10390 10391 *) Fixed the nasty bug where rsaref.h was not found under compile-time 10392 because the symlink to include/ was missing. 10393 [Ralf S. Engelschall] 10394 10395 *) Incorporated the popular no-RSA/DSA-only patches 10396 which allow to compile a RSA-free SSLeay. 10397 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 10398 10399 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 10400 when "ssleay" is still not found. 10401 [Ralf S. Engelschall] 10402 10403 *) Added more platforms to Configure: Cray T3E, HPUX 11, 10404 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 10405 10406 *) Updated the README file. 10407 [Ralf S. Engelschall] 10408 10409 *) Added various .cvsignore files in the CVS repository subdirs 10410 to make a "cvs update" really silent. 10411 [Ralf S. Engelschall] 10412 10413 *) Recompiled the error-definition header files and added 10414 missing symbols to the Win32 linker tables. 10415 [Ralf S. Engelschall] 10416 10417 *) Cleaned up the top-level documents; 10418 o new files: CHANGES and LICENSE 10419 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 10420 o merged COPYRIGHT into LICENSE 10421 o removed obsolete TODO file 10422 o renamed MICROSOFT to INSTALL.W32 10423 [Ralf S. Engelschall] 10424 10425 *) Removed dummy files from the 0.9.1b source tree: 10426 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 10427 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 10428 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 10429 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 10430 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 10431 [Ralf S. Engelschall] 10432 10433 *) Added various platform portability fixes. 10434 [Mark J. Cox] 10435 10436 *) The Genesis of the OpenSSL rpject: 10437 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 10438 Young and Tim J. Hudson created while they were working for C2Net until 10439 summer 1998. 10440 [The OpenSSL Project] 10441 10442 10443 Changes between 0.9.0b and 0.9.1b [not released] 10444 10445 *) Updated a few CA certificates under certs/ 10446 [Eric A. Young] 10447 10448 *) Changed some BIGNUM api stuff. 10449 [Eric A. Young] 10450 10451 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 10452 DGUX x86, Linux Alpha, etc. 10453 [Eric A. Young] 10454 10455 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 10456 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 10457 available). 10458 [Eric A. Young] 10459 10460 *) Add -strparse option to asn1pars program which parses nested 10461 binary structures 10462 [Dr Stephen Henson <shenson@bigfoot.com>] 10463 10464 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 10465 [Eric A. Young] 10466 10467 *) DSA fix for "ca" program. 10468 [Eric A. Young] 10469 10470 *) Added "-genkey" option to "dsaparam" program. 10471 [Eric A. Young] 10472 10473 *) Added RIPE MD160 (rmd160) message digest. 10474 [Eric A. Young] 10475 10476 *) Added -a (all) option to "ssleay version" command. 10477 [Eric A. Young] 10478 10479 *) Added PLATFORM define which is the id given to Configure. 10480 [Eric A. Young] 10481 10482 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 10483 [Eric A. Young] 10484 10485 *) Extended the ASN.1 parser routines. 10486 [Eric A. Young] 10487 10488 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 10489 [Eric A. Young] 10490 10491 *) Added a BN_CTX to the BN library. 10492 [Eric A. Young] 10493 10494 *) Fixed the weak key values in DES library 10495 [Eric A. Young] 10496 10497 *) Changed API in EVP library for cipher aliases. 10498 [Eric A. Young] 10499 10500 *) Added support for RC2/64bit cipher. 10501 [Eric A. Young] 10502 10503 *) Converted the lhash library to the crypto/mem.c functions. 10504 [Eric A. Young] 10505 10506 *) Added more recognized ASN.1 object ids. 10507 [Eric A. Young] 10508 10509 *) Added more RSA padding checks for SSL/TLS. 10510 [Eric A. Young] 10511 10512 *) Added BIO proxy/filter functionality. 10513 [Eric A. Young] 10514 10515 *) Added extra_certs to SSL_CTX which can be used 10516 send extra CA certificates to the client in the CA cert chain sending 10517 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 10518 [Eric A. Young] 10519 10520 *) Now Fortezza is denied in the authentication phase because 10521 this is key exchange mechanism is not supported by SSLeay at all. 10522 [Eric A. Young] 10523 10524 *) Additional PKCS1 checks. 10525 [Eric A. Young] 10526 10527 *) Support the string "TLSv1" for all TLS v1 ciphers. 10528 [Eric A. Young] 10529 10530 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 10531 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 10532 [Eric A. Young] 10533 10534 *) Fixed a few memory leaks. 10535 [Eric A. Young] 10536 10537 *) Fixed various code and comment typos. 10538 [Eric A. Young] 10539 10540 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 10541 bytes sent in the client random. 10542 [Edward Bishop <ebishop@spyglass.com>] 10543 10544