CHANGES revision 284330
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 1.0.1n and 1.0.1o [12 Jun 2015] 6 7 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI 8 incompatibility in the handling of HMAC. The previous ABI has now been 9 restored. 10 11 Changes between 1.0.1m and 1.0.1n [11 Jun 2015] 12 13 *) Malformed ECParameters causes infinite loop 14 15 When processing an ECParameters structure OpenSSL enters an infinite loop 16 if the curve specified is over a specially malformed binary polynomial 17 field. 18 19 This can be used to perform denial of service against any 20 system which processes public keys, certificate requests or 21 certificates. This includes TLS clients and TLS servers with 22 client authentication enabled. 23 24 This issue was reported to OpenSSL by Joseph Barr-Pixton. 25 (CVE-2015-1788) 26 [Andy Polyakov] 27 28 *) Exploitable out-of-bounds read in X509_cmp_time 29 30 X509_cmp_time does not properly check the length of the ASN1_TIME 31 string and can read a few bytes out of bounds. In addition, 32 X509_cmp_time accepts an arbitrary number of fractional seconds in the 33 time string. 34 35 An attacker can use this to craft malformed certificates and CRLs of 36 various sizes and potentially cause a segmentation fault, resulting in 37 a DoS on applications that verify certificates or CRLs. TLS clients 38 that verify CRLs are affected. TLS clients and servers with client 39 authentication enabled may be affected if they use custom verification 40 callbacks. 41 42 This issue was reported to OpenSSL by Robert Swiecki (Google), and 43 independently by Hanno B�ck. 44 (CVE-2015-1789) 45 [Emilia K�sper] 46 47 *) PKCS7 crash with missing EnvelopedContent 48 49 The PKCS#7 parsing code does not handle missing inner EncryptedContent 50 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 51 with missing content and trigger a NULL pointer dereference on parsing. 52 53 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 54 structures from untrusted sources are affected. OpenSSL clients and 55 servers are not affected. 56 57 This issue was reported to OpenSSL by Michal Zalewski (Google). 58 (CVE-2015-1790) 59 [Emilia K�sper] 60 61 *) CMS verify infinite loop with unknown hash function 62 63 When verifying a signedData message the CMS code can enter an infinite loop 64 if presented with an unknown hash function OID. This can be used to perform 65 denial of service against any system which verifies signedData messages using 66 the CMS code. 67 This issue was reported to OpenSSL by Johannes Bauer. 68 (CVE-2015-1792) 69 [Stephen Henson] 70 71 *) Race condition handling NewSessionTicket 72 73 If a NewSessionTicket is received by a multi-threaded client when attempting to 74 reuse a previous ticket then a race condition can occur potentially leading to 75 a double free of the ticket data. 76 (CVE-2015-1791) 77 [Matt Caswell] 78 79 *) Reject DH handshakes with parameters shorter than 768 bits. 80 [Kurt Roeckx and Emilia Kasper] 81 82 Changes between 1.0.1l and 1.0.1m [19 Mar 2015] 83 84 *) Segmentation fault in ASN1_TYPE_cmp fix 85 86 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 87 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 88 certificate signature algorithm consistency this can be used to crash any 89 certificate verification operation and exploited in a DoS attack. Any 90 application which performs certificate verification is vulnerable including 91 OpenSSL clients and servers which enable client authentication. 92 (CVE-2015-0286) 93 [Stephen Henson] 94 95 *) ASN.1 structure reuse memory corruption fix 96 97 Reusing a structure in ASN.1 parsing may allow an attacker to cause 98 memory corruption via an invalid write. Such reuse is and has been 99 strongly discouraged and is believed to be rare. 100 101 Applications that parse structures containing CHOICE or ANY DEFINED BY 102 components may be affected. Certificate parsing (d2i_X509 and related 103 functions) are however not affected. OpenSSL clients and servers are 104 not affected. 105 (CVE-2015-0287) 106 [Stephen Henson] 107 108 *) PKCS7 NULL pointer dereferences fix 109 110 The PKCS#7 parsing code does not handle missing outer ContentInfo 111 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 112 missing content and trigger a NULL pointer dereference on parsing. 113 114 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 115 otherwise parse PKCS#7 structures from untrusted sources are 116 affected. OpenSSL clients and servers are not affected. 117 118 This issue was reported to OpenSSL by Michal Zalewski (Google). 119 (CVE-2015-0289) 120 [Emilia K�sper] 121 122 *) DoS via reachable assert in SSLv2 servers fix 123 124 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 125 servers that both support SSLv2 and enable export cipher suites by sending 126 a specially crafted SSLv2 CLIENT-MASTER-KEY message. 127 128 This issue was discovered by Sean Burford (Google) and Emilia K�sper 129 (OpenSSL development team). 130 (CVE-2015-0293) 131 [Emilia K�sper] 132 133 *) Use After Free following d2i_ECPrivatekey error fix 134 135 A malformed EC private key file consumed via the d2i_ECPrivateKey function 136 could cause a use after free condition. This, in turn, could cause a double 137 free in several private key parsing functions (such as d2i_PrivateKey 138 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 139 for applications that receive EC private keys from untrusted 140 sources. This scenario is considered rare. 141 142 This issue was discovered by the BoringSSL project and fixed in their 143 commit 517073cd4b. 144 (CVE-2015-0209) 145 [Matt Caswell] 146 147 *) X509_to_X509_REQ NULL pointer deref fix 148 149 The function X509_to_X509_REQ will crash with a NULL pointer dereference if 150 the certificate key is invalid. This function is rarely used in practice. 151 152 This issue was discovered by Brian Carpenter. 153 (CVE-2015-0288) 154 [Stephen Henson] 155 156 *) Removed the export ciphers from the DEFAULT ciphers 157 [Kurt Roeckx] 158 159 Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 160 161 *) Build fixes for the Windows and OpenVMS platforms 162 [Matt Caswell and Richard Levitte] 163 164 Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 165 166 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 167 message can cause a segmentation fault in OpenSSL due to a NULL pointer 168 dereference. This could lead to a Denial Of Service attack. Thanks to 169 Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 170 (CVE-2014-3571) 171 [Steve Henson] 172 173 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 174 dtls1_buffer_record function under certain conditions. In particular this 175 could occur if an attacker sent repeated DTLS records with the same 176 sequence number but for the next epoch. The memory leak could be exploited 177 by an attacker in a Denial of Service attack through memory exhaustion. 178 Thanks to Chris Mueller for reporting this issue. 179 (CVE-2015-0206) 180 [Matt Caswell] 181 182 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 183 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 184 method would be set to NULL which could later result in a NULL pointer 185 dereference. Thanks to Frank Schmirler for reporting this issue. 186 (CVE-2014-3569) 187 [Kurt Roeckx] 188 189 *) Abort handshake if server key exchange message is omitted for ephemeral 190 ECDH ciphersuites. 191 192 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 193 reporting this issue. 194 (CVE-2014-3572) 195 [Steve Henson] 196 197 *) Remove non-export ephemeral RSA code on client and server. This code 198 violated the TLS standard by allowing the use of temporary RSA keys in 199 non-export ciphersuites and could be used by a server to effectively 200 downgrade the RSA key length used to a value smaller than the server 201 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 202 INRIA or reporting this issue. 203 (CVE-2015-0204) 204 [Steve Henson] 205 206 *) Fixed issue where DH client certificates are accepted without verification. 207 An OpenSSL server will accept a DH certificate for client authentication 208 without the certificate verify message. This effectively allows a client to 209 authenticate without the use of a private key. This only affects servers 210 which trust a client certificate authority which issues certificates 211 containing DH keys: these are extremely rare and hardly ever encountered. 212 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 213 this issue. 214 (CVE-2015-0205) 215 [Steve Henson] 216 217 *) Ensure that the session ID context of an SSL is updated when its 218 SSL_CTX is updated via SSL_set_SSL_CTX. 219 220 The session ID context is typically set from the parent SSL_CTX, 221 and can vary with the CTX. 222 [Adam Langley] 223 224 *) Fix various certificate fingerprint issues. 225 226 By using non-DER or invalid encodings outside the signed portion of a 227 certificate the fingerprint can be changed without breaking the signature. 228 Although no details of the signed portion of the certificate can be changed 229 this can cause problems with some applications: e.g. those using the 230 certificate fingerprint for blacklists. 231 232 1. Reject signatures with non zero unused bits. 233 234 If the BIT STRING containing the signature has non zero unused bits reject 235 the signature. All current signature algorithms require zero unused bits. 236 237 2. Check certificate algorithm consistency. 238 239 Check the AlgorithmIdentifier inside TBS matches the one in the 240 certificate signature. NB: this will result in signature failure 241 errors for some broken certificates. 242 243 Thanks to Konrad Kraszewski from Google for reporting this issue. 244 245 3. Check DSA/ECDSA signatures use DER. 246 247 Reencode DSA/ECDSA signatures and compare with the original received 248 signature. Return an error if there is a mismatch. 249 250 This will reject various cases including garbage after signature 251 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 252 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 253 (negative or with leading zeroes). 254 255 Further analysis was conducted and fixes were developed by Stephen Henson 256 of the OpenSSL core team. 257 258 (CVE-2014-8275) 259 [Steve Henson] 260 261 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 262 results on some platforms, including x86_64. This bug occurs at random 263 with a very low probability, and is not known to be exploitable in any 264 way, though its exact impact is difficult to determine. Thanks to Pieter 265 Wuille (Blockstream) who reported this issue and also suggested an initial 266 fix. Further analysis was conducted by the OpenSSL development team and 267 Adam Langley of Google. The final fix was developed by Andy Polyakov of 268 the OpenSSL core team. 269 (CVE-2014-3570) 270 [Andy Polyakov] 271 272 *) Do not resume sessions on the server if the negotiated protocol 273 version does not match the session's version. Resuming with a different 274 version, while not strictly forbidden by the RFC, is of questionable 275 sanity and breaks all known clients. 276 [David Benjamin, Emilia K�sper] 277 278 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 279 early CCS messages during renegotiation. (Note that because 280 renegotiation is encrypted, this early CCS was not exploitable.) 281 [Emilia K�sper] 282 283 *) Tighten client-side session ticket handling during renegotiation: 284 ensure that the client only accepts a session ticket if the server sends 285 the extension anew in the ServerHello. Previously, a TLS client would 286 reuse the old extension state and thus accept a session ticket if one was 287 announced in the initial ServerHello. 288 289 Similarly, ensure that the client requires a session ticket if one 290 was advertised in the ServerHello. Previously, a TLS client would 291 ignore a missing NewSessionTicket message. 292 [Emilia K�sper] 293 294 Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 295 296 *) SRTP Memory Leak. 297 298 A flaw in the DTLS SRTP extension parsing code allows an attacker, who 299 sends a carefully crafted handshake message, to cause OpenSSL to fail 300 to free up to 64k of memory causing a memory leak. This could be 301 exploited in a Denial Of Service attack. This issue affects OpenSSL 302 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 303 whether SRTP is used or configured. Implementations of OpenSSL that 304 have been compiled with OPENSSL_NO_SRTP defined are not affected. 305 306 The fix was developed by the OpenSSL team. 307 (CVE-2014-3513) 308 [OpenSSL team] 309 310 *) Session Ticket Memory Leak. 311 312 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 313 integrity of that ticket is first verified. In the event of a session 314 ticket integrity check failing, OpenSSL will fail to free memory 315 causing a memory leak. By sending a large number of invalid session 316 tickets an attacker could exploit this issue in a Denial Of Service 317 attack. 318 (CVE-2014-3567) 319 [Steve Henson] 320 321 *) Build option no-ssl3 is incomplete. 322 323 When OpenSSL is configured with "no-ssl3" as a build option, servers 324 could accept and complete a SSL 3.0 handshake, and clients could be 325 configured to send them. 326 (CVE-2014-3568) 327 [Akamai and the OpenSSL team] 328 329 *) Add support for TLS_FALLBACK_SCSV. 330 Client applications doing fallback retries should call 331 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 332 (CVE-2014-3566) 333 [Adam Langley, Bodo Moeller] 334 335 *) Add additional DigestInfo checks. 336 337 Reencode DigestInto in DER and check against the original when 338 verifying RSA signature: this will reject any improperly encoded 339 DigestInfo structures. 340 341 Note: this is a precautionary measure and no attacks are currently known. 342 343 [Steve Henson] 344 345 Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 346 347 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 348 SRP code can be overrun an internal buffer. Add sanity check that 349 g, A, B < N to SRP code. 350 351 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 352 Group for discovering this issue. 353 (CVE-2014-3512) 354 [Steve Henson] 355 356 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 357 TLS 1.0 instead of higher protocol versions when the ClientHello message 358 is badly fragmented. This allows a man-in-the-middle attacker to force a 359 downgrade to TLS 1.0 even if both the server and the client support a 360 higher protocol version, by modifying the client's TLS records. 361 362 Thanks to David Benjamin and Adam Langley (Google) for discovering and 363 researching this issue. 364 (CVE-2014-3511) 365 [David Benjamin] 366 367 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 368 to a denial of service attack. A malicious server can crash the client 369 with a null pointer dereference (read) by specifying an anonymous (EC)DH 370 ciphersuite and sending carefully crafted handshake messages. 371 372 Thanks to Felix Gr�bert (Google) for discovering and researching this 373 issue. 374 (CVE-2014-3510) 375 [Emilia K�sper] 376 377 *) By sending carefully crafted DTLS packets an attacker could cause openssl 378 to leak memory. This can be exploited through a Denial of Service attack. 379 Thanks to Adam Langley for discovering and researching this issue. 380 (CVE-2014-3507) 381 [Adam Langley] 382 383 *) An attacker can force openssl to consume large amounts of memory whilst 384 processing DTLS handshake messages. This can be exploited through a 385 Denial of Service attack. 386 Thanks to Adam Langley for discovering and researching this issue. 387 (CVE-2014-3506) 388 [Adam Langley] 389 390 *) An attacker can force an error condition which causes openssl to crash 391 whilst processing DTLS packets due to memory being freed twice. This 392 can be exploited through a Denial of Service attack. 393 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 394 this issue. 395 (CVE-2014-3505) 396 [Adam Langley] 397 398 *) If a multithreaded client connects to a malicious server using a resumed 399 session and the server sends an ec point format extension it could write 400 up to 255 bytes to freed memory. 401 402 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 403 issue. 404 (CVE-2014-3509) 405 [Gabor Tyukasz] 406 407 *) A malicious server can crash an OpenSSL client with a null pointer 408 dereference (read) by specifying an SRP ciphersuite even though it was not 409 properly negotiated with the client. This can be exploited through a 410 Denial of Service attack. 411 412 Thanks to Joonas Kuorilehto and Riku Hietam�ki (Codenomicon) for 413 discovering and researching this issue. 414 (CVE-2014-5139) 415 [Steve Henson] 416 417 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 418 X509_name_oneline, X509_name_print_ex et al. to leak some information 419 from the stack. Applications may be affected if they echo pretty printing 420 output to the attacker. 421 422 Thanks to Ivan Fratric (Google) for discovering this issue. 423 (CVE-2014-3508) 424 [Emilia K�sper, and Steve Henson] 425 426 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 427 for corner cases. (Certain input points at infinity could lead to 428 bogus results, with non-infinity inputs mapped to infinity too.) 429 [Bodo Moeller] 430 431 Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 432 433 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 434 handshake can force the use of weak keying material in OpenSSL 435 SSL/TLS clients and servers. 436 437 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 438 researching this issue. (CVE-2014-0224) 439 [KIKUCHI Masashi, Steve Henson] 440 441 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 442 OpenSSL DTLS client the code can be made to recurse eventually crashing 443 in a DoS attack. 444 445 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 446 (CVE-2014-0221) 447 [Imre Rad, Steve Henson] 448 449 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 450 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 451 client or server. This is potentially exploitable to run arbitrary 452 code on a vulnerable client or server. 453 454 Thanks to J�ri Aedla for reporting this issue. (CVE-2014-0195) 455 [J�ri Aedla, Steve Henson] 456 457 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 458 are subject to a denial of service attack. 459 460 Thanks to Felix Gr�bert and Ivan Fratric at Google for discovering 461 this issue. (CVE-2014-3470) 462 [Felix Gr�bert, Ivan Fratric, Steve Henson] 463 464 *) Harmonize version and its documentation. -f flag is used to display 465 compilation flags. 466 [mancha <mancha1@zoho.com>] 467 468 *) Fix eckey_priv_encode so it immediately returns an error upon a failure 469 in i2d_ECPrivateKey. 470 [mancha <mancha1@zoho.com>] 471 472 *) Fix some double frees. These are not thought to be exploitable. 473 [mancha <mancha1@zoho.com>] 474 475 Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 476 477 *) A missing bounds check in the handling of the TLS heartbeat extension 478 can be used to reveal up to 64k of memory to a connected client or 479 server. 480 481 Thanks for Neel Mehta of Google Security for discovering this bug and to 482 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 483 preparing the fix (CVE-2014-0160) 484 [Adam Langley, Bodo Moeller] 485 486 *) Fix for the attack described in the paper "Recovering OpenSSL 487 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 488 by Yuval Yarom and Naomi Benger. Details can be obtained from: 489 http://eprint.iacr.org/2014/140 490 491 Thanks to Yuval Yarom and Naomi Benger for discovering this 492 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 493 [Yuval Yarom and Naomi Benger] 494 495 *) TLS pad extension: draft-agl-tls-padding-03 496 497 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 498 TLS client Hello record length value would otherwise be > 255 and 499 less that 512 pad with a dummy extension containing zeroes so it 500 is at least 512 bytes long. 501 502 [Adam Langley, Steve Henson] 503 504 Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 505 506 *) Fix for TLS record tampering bug. A carefully crafted invalid 507 handshake could crash OpenSSL with a NULL pointer exception. 508 Thanks to Anton Johansson for reporting this issues. 509 (CVE-2013-4353) 510 511 *) Keep original DTLS digest and encryption contexts in retransmission 512 structures so we can use the previous session parameters if they need 513 to be resent. (CVE-2013-6450) 514 [Steve Henson] 515 516 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 517 avoids preferring ECDHE-ECDSA ciphers when the client appears to be 518 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 519 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 520 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 521 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 522 [Rob Stradling, Adam Langley] 523 524 Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 525 526 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 527 supporting platforms or when small records were transferred. 528 [Andy Polyakov, Steve Henson] 529 530 Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 531 532 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 533 534 This addresses the flaw in CBC record processing discovered by 535 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 536 at: http://www.isg.rhul.ac.uk/tls/ 537 538 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 539 Security Group at Royal Holloway, University of London 540 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 541 Emilia K�sper for the initial patch. 542 (CVE-2013-0169) 543 [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 544 545 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 546 ciphersuites which can be exploited in a denial of service attack. 547 Thanks go to and to Adam Langley <agl@chromium.org> for discovering 548 and detecting this bug and to Wolfgang Ettlinger 549 <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 550 (CVE-2012-2686) 551 [Adam Langley] 552 553 *) Return an error when checking OCSP signatures when key is NULL. 554 This fixes a DoS attack. (CVE-2013-0166) 555 [Steve Henson] 556 557 *) Make openssl verify return errors. 558 [Chris Palmer <palmer@google.com> and Ben Laurie] 559 560 *) Call OCSP Stapling callback after ciphersuite has been chosen, so 561 the right response is stapled. Also change SSL_get_certificate() 562 so it returns the certificate actually sent. 563 See http://rt.openssl.org/Ticket/Display.html?id=2836. 564 [Rob Stradling <rob.stradling@comodo.com>] 565 566 *) Fix possible deadlock when decoding public keys. 567 [Steve Henson] 568 569 *) Don't use TLS 1.0 record version number in initial client hello 570 if renegotiating. 571 [Steve Henson] 572 573 Changes between 1.0.1b and 1.0.1c [10 May 2012] 574 575 *) Sanity check record length before skipping explicit IV in TLS 576 1.2, 1.1 and DTLS to fix DoS attack. 577 578 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 579 fuzzing as a service testing platform. 580 (CVE-2012-2333) 581 [Steve Henson] 582 583 *) Initialise tkeylen properly when encrypting CMS messages. 584 Thanks to Solar Designer of Openwall for reporting this issue. 585 [Steve Henson] 586 587 *) In FIPS mode don't try to use composite ciphers as they are not 588 approved. 589 [Steve Henson] 590 591 Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 592 593 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 594 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 595 mean any application compiled against OpenSSL 1.0.0 headers setting 596 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 597 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 598 0x10000000L Any application which was previously compiled against 599 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 600 will need to be recompiled as a result. Letting be results in 601 inability to disable specifically TLS 1.1 and in client context, 602 in unlike event, limit maximum offered version to TLS 1.0 [see below]. 603 [Steve Henson] 604 605 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 606 disable just protocol X, but all protocols above X *if* there are 607 protocols *below* X still enabled. In more practical terms it means 608 that if application wants to disable TLS1.0 in favor of TLS1.1 and 609 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 610 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 611 client side. 612 [Andy Polyakov] 613 614 Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 615 616 *) Check for potentially exploitable overflows in asn1_d2i_read_bio 617 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 618 in CRYPTO_realloc_clean. 619 620 Thanks to Tavis Ormandy, Google Security Team, for discovering this 621 issue and to Adam Langley <agl@chromium.org> for fixing it. 622 (CVE-2012-2110) 623 [Adam Langley (Google), Tavis Ormandy, Google Security Team] 624 625 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 626 [Adam Langley] 627 628 *) Workarounds for some broken servers that "hang" if a client hello 629 record length exceeds 255 bytes. 630 631 1. Do not use record version number > TLS 1.0 in initial client 632 hello: some (but not all) hanging servers will now work. 633 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 634 the number of ciphers sent in the client hello. This should be 635 set to an even number, such as 50, for example by passing: 636 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 637 Most broken servers should now work. 638 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 639 TLS 1.2 client support entirely. 640 [Steve Henson] 641 642 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 643 [Andy Polyakov] 644 645 Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 646 647 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 648 STRING form instead of a DigestInfo. 649 [Steve Henson] 650 651 *) The format used for MDC2 RSA signatures is inconsistent between EVP 652 and the RSA_sign/RSA_verify functions. This was made more apparent when 653 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 654 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 655 the correct format in RSA_verify so both forms transparently work. 656 [Steve Henson] 657 658 *) Some servers which support TLS 1.0 can choke if we initially indicate 659 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 660 encrypted premaster secret. As a workaround use the maximum pemitted 661 client version in client hello, this should keep such servers happy 662 and still work with previous versions of OpenSSL. 663 [Steve Henson] 664 665 *) Add support for TLS/DTLS heartbeats. 666 [Robin Seggelmann <seggelmann@fh-muenster.de>] 667 668 *) Add support for SCTP. 669 [Robin Seggelmann <seggelmann@fh-muenster.de>] 670 671 *) Improved PRNG seeding for VOS. 672 [Paul Green <Paul.Green@stratus.com>] 673 674 *) Extensive assembler packs updates, most notably: 675 676 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 677 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 678 - x86_64: bit-sliced AES implementation; 679 - ARM: NEON support, contemporary platforms optimizations; 680 - s390x: z196 support; 681 - *: GHASH and GF(2^m) multiplication implementations; 682 683 [Andy Polyakov] 684 685 *) Make TLS-SRP code conformant with RFC 5054 API cleanup 686 (removal of unnecessary code) 687 [Peter Sylvester <peter.sylvester@edelweb.fr>] 688 689 *) Add TLS key material exporter from RFC 5705. 690 [Eric Rescorla] 691 692 *) Add DTLS-SRTP negotiation from RFC 5764. 693 [Eric Rescorla] 694 695 *) Add Next Protocol Negotiation, 696 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 697 disabled with a no-npn flag to config or Configure. Code donated 698 by Google. 699 [Adam Langley <agl@google.com> and Ben Laurie] 700 701 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 702 NIST-P256, NIST-P521, with constant-time single point multiplication on 703 typical inputs. Compiler support for the nonstandard type __uint128_t is 704 required to use this (present in gcc 4.4 and later, for 64-bit builds). 705 Code made available under Apache License version 2.0. 706 707 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 708 line to include this in your build of OpenSSL, and run "make depend" (or 709 "make update"). This enables the following EC_METHODs: 710 711 EC_GFp_nistp224_method() 712 EC_GFp_nistp256_method() 713 EC_GFp_nistp521_method() 714 715 EC_GROUP_new_by_curve_name() will automatically use these (while 716 EC_GROUP_new_curve_GFp() currently prefers the more flexible 717 implementations). 718 [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 719 720 *) Use type ossl_ssize_t instad of ssize_t which isn't available on 721 all platforms. Move ssize_t definition from e_os.h to the public 722 header file e_os2.h as it now appears in public header file cms.h 723 [Steve Henson] 724 725 *) New -sigopt option to the ca, req and x509 utilities. Additional 726 signature parameters can be passed using this option and in 727 particular PSS. 728 [Steve Henson] 729 730 *) Add RSA PSS signing function. This will generate and set the 731 appropriate AlgorithmIdentifiers for PSS based on those in the 732 corresponding EVP_MD_CTX structure. No application support yet. 733 [Steve Henson] 734 735 *) Support for companion algorithm specific ASN1 signing routines. 736 New function ASN1_item_sign_ctx() signs a pre-initialised 737 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 738 the appropriate parameters. 739 [Steve Henson] 740 741 *) Add new algorithm specific ASN1 verification initialisation function 742 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 743 handling will be the same no matter what EVP_PKEY_METHOD is used. 744 Add a PSS handler to support verification of PSS signatures: checked 745 against a number of sample certificates. 746 [Steve Henson] 747 748 *) Add signature printing for PSS. Add PSS OIDs. 749 [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 750 751 *) Add algorithm specific signature printing. An individual ASN1 method 752 can now print out signatures instead of the standard hex dump. 753 754 More complex signatures (e.g. PSS) can print out more meaningful 755 information. Include DSA version that prints out the signature 756 parameters r, s. 757 [Steve Henson] 758 759 *) Password based recipient info support for CMS library: implementing 760 RFC3211. 761 [Steve Henson] 762 763 *) Split password based encryption into PBES2 and PBKDF2 functions. This 764 neatly separates the code into cipher and PBE sections and is required 765 for some algorithms that split PBES2 into separate pieces (such as 766 password based CMS). 767 [Steve Henson] 768 769 *) Session-handling fixes: 770 - Fix handling of connections that are resuming with a session ID, 771 but also support Session Tickets. 772 - Fix a bug that suppressed issuing of a new ticket if the client 773 presented a ticket with an expired session. 774 - Try to set the ticket lifetime hint to something reasonable. 775 - Make tickets shorter by excluding irrelevant information. 776 - On the client side, don't ignore renewed tickets. 777 [Adam Langley, Bodo Moeller (Google)] 778 779 *) Fix PSK session representation. 780 [Bodo Moeller] 781 782 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 783 784 This work was sponsored by Intel. 785 [Andy Polyakov] 786 787 *) Add GCM support to TLS library. Some custom code is needed to split 788 the IV between the fixed (from PRF) and explicit (from TLS record) 789 portions. This adds all GCM ciphersuites supported by RFC5288 and 790 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 791 add a special AESGCM string for GCM only. 792 [Steve Henson] 793 794 *) Expand range of ctrls for AES GCM. Permit setting invocation 795 field on decrypt and retrieval of invocation field only on encrypt. 796 [Steve Henson] 797 798 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 799 As required by RFC5289 these ciphersuites cannot be used if for 800 versions of TLS earlier than 1.2. 801 [Steve Henson] 802 803 *) For FIPS capable OpenSSL interpret a NULL default public key method 804 as unset and return the appopriate default but do *not* set the default. 805 This means we can return the appopriate method in applications that 806 swicth between FIPS and non-FIPS modes. 807 [Steve Henson] 808 809 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 810 ENGINE is used then we cannot handle that in the FIPS module so we 811 keep original code iff non-FIPS operations are allowed. 812 [Steve Henson] 813 814 *) Add -attime option to openssl utilities. 815 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 816 817 *) Redirect DSA and DH operations to FIPS module in FIPS mode. 818 [Steve Henson] 819 820 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 821 FIPS EC methods unconditionally for now. 822 [Steve Henson] 823 824 *) New build option no-ec2m to disable characteristic 2 code. 825 [Steve Henson] 826 827 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 828 all cases can be covered as some introduce binary incompatibilities. 829 [Steve Henson] 830 831 *) Redirect RSA operations to FIPS module including keygen, 832 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 833 [Steve Henson] 834 835 *) Add similar low level API blocking to ciphers. 836 [Steve Henson] 837 838 *) Low level digest APIs are not approved in FIPS mode: any attempt 839 to use these will cause a fatal error. Applications that *really* want 840 to use them can use the private_* version instead. 841 [Steve Henson] 842 843 *) Redirect cipher operations to FIPS module for FIPS builds. 844 [Steve Henson] 845 846 *) Redirect digest operations to FIPS module for FIPS builds. 847 [Steve Henson] 848 849 *) Update build system to add "fips" flag which will link in fipscanister.o 850 for static and shared library builds embedding a signature if needed. 851 [Steve Henson] 852 853 *) Output TLS supported curves in preference order instead of numerical 854 order. This is currently hardcoded for the highest order curves first. 855 This should be configurable so applications can judge speed vs strength. 856 [Steve Henson] 857 858 *) Add TLS v1.2 server support for client authentication. 859 [Steve Henson] 860 861 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 862 and enable MD5. 863 [Steve Henson] 864 865 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 866 FIPS modules versions. 867 [Steve Henson] 868 869 *) Add TLS v1.2 client side support for client authentication. Keep cache 870 of handshake records longer as we don't know the hash algorithm to use 871 until after the certificate request message is received. 872 [Steve Henson] 873 874 *) Initial TLS v1.2 client support. Add a default signature algorithms 875 extension including all the algorithms we support. Parse new signature 876 format in client key exchange. Relax some ECC signing restrictions for 877 TLS v1.2 as indicated in RFC5246. 878 [Steve Henson] 879 880 *) Add server support for TLS v1.2 signature algorithms extension. Switch 881 to new signature format when needed using client digest preference. 882 All server ciphersuites should now work correctly in TLS v1.2. No client 883 support yet and no support for client certificates. 884 [Steve Henson] 885 886 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 887 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 888 ciphersuites. At present only RSA key exchange ciphersuites work with 889 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 890 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 891 and version checking. 892 [Steve Henson] 893 894 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 895 with this defined it will not be affected by any changes to ssl internal 896 structures. Add several utility functions to allow openssl application 897 to work with OPENSSL_NO_SSL_INTERN defined. 898 [Steve Henson] 899 900 *) Add SRP support. 901 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 902 903 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 904 [Steve Henson] 905 906 *) Permit abbreviated handshakes when renegotiating using the function 907 SSL_renegotiate_abbreviated(). 908 [Robin Seggelmann <seggelmann@fh-muenster.de>] 909 910 *) Add call to ENGINE_register_all_complete() to 911 ENGINE_load_builtin_engines(), so some implementations get used 912 automatically instead of needing explicit application support. 913 [Steve Henson] 914 915 *) Add support for TLS key exporter as described in RFC5705. 916 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 917 918 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 919 a few changes are required: 920 921 Add SSL_OP_NO_TLSv1_1 flag. 922 Add TLSv1_1 methods. 923 Update version checking logic to handle version 1.1. 924 Add explicit IV handling (ported from DTLS code). 925 Add command line options to s_client/s_server. 926 [Steve Henson] 927 928 Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 929 930 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 931 in CMS and PKCS7 code. When RSA decryption fails use a random key for 932 content decryption and always return the same error. Note: this attack 933 needs on average 2^20 messages so it only affects automated senders. The 934 old behaviour can be reenabled in the CMS code by setting the 935 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 936 an MMA defence is not necessary. 937 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 938 this issue. (CVE-2012-0884) 939 [Steve Henson] 940 941 *) Fix CVE-2011-4619: make sure we really are receiving a 942 client hello before rejecting multiple SGC restarts. Thanks to 943 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 944 [Steve Henson] 945 946 Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 947 948 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 949 Thanks to Antonio Martin, Enterprise Secure Access Research and 950 Development, Cisco Systems, Inc. for discovering this bug and 951 preparing a fix. (CVE-2012-0050) 952 [Antonio Martin] 953 954 Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 955 956 *) Nadhem Alfardan and Kenny Paterson have discovered an extension 957 of the Vaudenay padding oracle attack on CBC mode encryption 958 which enables an efficient plaintext recovery attack against 959 the OpenSSL implementation of DTLS. Their attack exploits timing 960 differences arising during decryption processing. A research 961 paper describing this attack can be found at: 962 http://www.isg.rhul.ac.uk/~kp/dtls.pdf 963 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 964 Security Group at Royal Holloway, University of London 965 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 966 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 967 for preparing the fix. (CVE-2011-4108) 968 [Robin Seggelmann, Michael Tuexen] 969 970 *) Clear bytes used for block padding of SSL 3.0 records. 971 (CVE-2011-4576) 972 [Adam Langley (Google)] 973 974 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 975 Kadianakis <desnacked@gmail.com> for discovering this issue and 976 Adam Langley for preparing the fix. (CVE-2011-4619) 977 [Adam Langley (Google)] 978 979 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 980 [Andrey Kulikov <amdeich@gmail.com>] 981 982 *) Prevent malformed RFC3779 data triggering an assertion failure. 983 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 984 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 985 [Rob Austein <sra@hactrn.net>] 986 987 *) Improved PRNG seeding for VOS. 988 [Paul Green <Paul.Green@stratus.com>] 989 990 *) Fix ssl_ciph.c set-up race. 991 [Adam Langley (Google)] 992 993 *) Fix spurious failures in ecdsatest.c. 994 [Emilia K�sper (Google)] 995 996 *) Fix the BIO_f_buffer() implementation (which was mixing different 997 interpretations of the '..._len' fields). 998 [Adam Langley (Google)] 999 1000 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1001 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1002 threads won't reuse the same blinding coefficients. 1003 1004 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1005 lock to call BN_BLINDING_invert_ex, and avoids one use of 1006 BN_BLINDING_update for each BN_BLINDING structure (previously, 1007 the last update always remained unused). 1008 [Emilia K�sper (Google)] 1009 1010 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 1011 [Bob Buckholz (Google)] 1012 1013 Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 1014 1015 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 1016 by initialising X509_STORE_CTX properly. (CVE-2011-3207) 1017 [Kaspar Brand <ossl@velox.ch>] 1018 1019 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1020 for multi-threaded use of ECDH. (CVE-2011-3210) 1021 [Adam Langley (Google)] 1022 1023 *) Fix x509_name_ex_d2i memory leak on bad inputs. 1024 [Bodo Moeller] 1025 1026 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 1027 signature public key algorithm by using OID xref utilities instead. 1028 Before this you could only use some ECC ciphersuites with SHA1 only. 1029 [Steve Henson] 1030 1031 *) Add protection against ECDSA timing attacks as mentioned in the paper 1032 by Billy Bob Brumley and Nicola Tuveri, see: 1033 1034 http://eprint.iacr.org/2011/232.pdf 1035 1036 [Billy Bob Brumley and Nicola Tuveri] 1037 1038 Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 1039 1040 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1041 [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1042 1043 *) Fix bug in string printing code: if *any* escaping is enabled we must 1044 escape the escape character (backslash) or the resulting string is 1045 ambiguous. 1046 [Steve Henson] 1047 1048 Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 1049 1050 *) Disable code workaround for ancient and obsolete Netscape browsers 1051 and servers: an attacker can use it in a ciphersuite downgrade attack. 1052 Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1053 [Steve Henson] 1054 1055 *) Fixed J-PAKE implementation error, originally discovered by 1056 Sebastien Martini, further info and confirmation from Stefan 1057 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1058 [Ben Laurie] 1059 1060 Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 1061 1062 *) Fix extension code to avoid race conditions which can result in a buffer 1063 overrun vulnerability: resumed sessions must not be modified as they can 1064 be shared by multiple threads. CVE-2010-3864 1065 [Steve Henson] 1066 1067 *) Fix WIN32 build system to correctly link an ENGINE directory into 1068 a DLL. 1069 [Steve Henson] 1070 1071 Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 1072 1073 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 1074 (CVE-2010-1633) 1075 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 1076 1077 Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 1078 1079 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 1080 context. The operation can be customised via the ctrl mechanism in 1081 case ENGINEs want to include additional functionality. 1082 [Steve Henson] 1083 1084 *) Tolerate yet another broken PKCS#8 key format: private key value negative. 1085 [Steve Henson] 1086 1087 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 1088 output hashes compatible with older versions of OpenSSL. 1089 [Willy Weisz <weisz@vcpc.univie.ac.at>] 1090 1091 *) Fix compression algorithm handling: if resuming a session use the 1092 compression algorithm of the resumed session instead of determining 1093 it from client hello again. Don't allow server to change algorithm. 1094 [Steve Henson] 1095 1096 *) Add load_crls() function to apps tidying load_certs() too. Add option 1097 to verify utility to allow additional CRLs to be included. 1098 [Steve Henson] 1099 1100 *) Update OCSP request code to permit adding custom headers to the request: 1101 some responders need this. 1102 [Steve Henson] 1103 1104 *) The function EVP_PKEY_sign() returns <=0 on error: check return code 1105 correctly. 1106 [Julia Lawall <julia@diku.dk>] 1107 1108 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 1109 needlessly dereferenced structures, used obsolete functions and 1110 didn't handle all updated verify codes correctly. 1111 [Steve Henson] 1112 1113 *) Disable MD2 in the default configuration. 1114 [Steve Henson] 1115 1116 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 1117 indicate the initial BIO being pushed or popped. This makes it possible 1118 to determine whether the BIO is the one explicitly called or as a result 1119 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 1120 it handles reference counts correctly and doesn't zero out the I/O bio 1121 when it is not being explicitly popped. WARNING: applications which 1122 included workarounds for the old buggy behaviour will need to be modified 1123 or they could free up already freed BIOs. 1124 [Steve Henson] 1125 1126 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 1127 renaming to all platforms (within the 0.9.8 branch, this was 1128 done conditionally on Netware platforms to avoid a name clash). 1129 [Guenter <lists@gknw.net>] 1130 1131 *) Add ECDHE and PSK support to DTLS. 1132 [Michael Tuexen <tuexen@fh-muenster.de>] 1133 1134 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 1135 be used on C++. 1136 [Steve Henson] 1137 1138 *) Add "missing" function EVP_MD_flags() (without this the only way to 1139 retrieve a digest flags is by accessing the structure directly. Update 1140 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 1141 or cipher is registered as in the "from" argument. Print out all 1142 registered digests in the dgst usage message instead of manually 1143 attempting to work them out. 1144 [Steve Henson] 1145 1146 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 1147 this allows the use of compression and extensions. Change default cipher 1148 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 1149 by default unless an application cipher string requests it. 1150 [Steve Henson] 1151 1152 *) Alter match criteria in PKCS12_parse(). It used to try to use local 1153 key ids to find matching certificates and keys but some PKCS#12 files 1154 don't follow the (somewhat unwritten) rules and this strategy fails. 1155 Now just gather all certificates together and the first private key 1156 then look for the first certificate that matches the key. 1157 [Steve Henson] 1158 1159 *) Support use of registered digest and cipher names for dgst and cipher 1160 commands instead of having to add each one as a special case. So now 1161 you can do: 1162 1163 openssl sha256 foo 1164 1165 as well as: 1166 1167 openssl dgst -sha256 foo 1168 1169 and this works for ENGINE based algorithms too. 1170 1171 [Steve Henson] 1172 1173 *) Update Gost ENGINE to support parameter files. 1174 [Victor B. Wagner <vitus@cryptocom.ru>] 1175 1176 *) Support GeneralizedTime in ca utility. 1177 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 1178 1179 *) Enhance the hash format used for certificate directory links. The new 1180 form uses the canonical encoding (meaning equivalent names will work 1181 even if they aren't identical) and uses SHA1 instead of MD5. This form 1182 is incompatible with the older format and as a result c_rehash should 1183 be used to rebuild symbolic links. 1184 [Steve Henson] 1185 1186 *) Make PKCS#8 the default write format for private keys, replacing the 1187 traditional format. This form is standardised, more secure and doesn't 1188 include an implicit MD5 dependency. 1189 [Steve Henson] 1190 1191 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 1192 committed to OpenSSL should pass this lot as a minimum. 1193 [Steve Henson] 1194 1195 *) Add session ticket override functionality for use by EAP-FAST. 1196 [Jouni Malinen <j@w1.fi>] 1197 1198 *) Modify HMAC functions to return a value. Since these can be implemented 1199 in an ENGINE errors can occur. 1200 [Steve Henson] 1201 1202 *) Type-checked OBJ_bsearch_ex. 1203 [Ben Laurie] 1204 1205 *) Type-checked OBJ_bsearch. Also some constification necessitated 1206 by type-checking. Still to come: TXT_DB, bsearch(?), 1207 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 1208 CONF_VALUE. 1209 [Ben Laurie] 1210 1211 *) New function OPENSSL_gmtime_adj() to add a specific number of days and 1212 seconds to a tm structure directly, instead of going through OS 1213 specific date routines. This avoids any issues with OS routines such 1214 as the year 2038 bug. New *_adj() functions for ASN1 time structures 1215 and X509_time_adj_ex() to cover the extended range. The existing 1216 X509_time_adj() is still usable and will no longer have any date issues. 1217 [Steve Henson] 1218 1219 *) Delta CRL support. New use deltas option which will attempt to locate 1220 and search any appropriate delta CRLs available. 1221 1222 This work was sponsored by Google. 1223 [Steve Henson] 1224 1225 *) Support for CRLs partitioned by reason code. Reorganise CRL processing 1226 code and add additional score elements. Validate alternate CRL paths 1227 as part of the CRL checking and indicate a new error "CRL path validation 1228 error" in this case. Applications wanting additional details can use 1229 the verify callback and check the new "parent" field. If this is not 1230 NULL CRL path validation is taking place. Existing applications wont 1231 see this because it requires extended CRL support which is off by 1232 default. 1233 1234 This work was sponsored by Google. 1235 [Steve Henson] 1236 1237 *) Support for freshest CRL extension. 1238 1239 This work was sponsored by Google. 1240 [Steve Henson] 1241 1242 *) Initial indirect CRL support. Currently only supported in the CRLs 1243 passed directly and not via lookup. Process certificate issuer 1244 CRL entry extension and lookup CRL entries by bother issuer name 1245 and serial number. Check and process CRL issuer entry in IDP extension. 1246 1247 This work was sponsored by Google. 1248 [Steve Henson] 1249 1250 *) Add support for distinct certificate and CRL paths. The CRL issuer 1251 certificate is validated separately in this case. Only enabled if 1252 an extended CRL support flag is set: this flag will enable additional 1253 CRL functionality in future. 1254 1255 This work was sponsored by Google. 1256 [Steve Henson] 1257 1258 *) Add support for policy mappings extension. 1259 1260 This work was sponsored by Google. 1261 [Steve Henson] 1262 1263 *) Fixes to pathlength constraint, self issued certificate handling, 1264 policy processing to align with RFC3280 and PKITS tests. 1265 1266 This work was sponsored by Google. 1267 [Steve Henson] 1268 1269 *) Support for name constraints certificate extension. DN, email, DNS 1270 and URI types are currently supported. 1271 1272 This work was sponsored by Google. 1273 [Steve Henson] 1274 1275 *) To cater for systems that provide a pointer-based thread ID rather 1276 than numeric, deprecate the current numeric thread ID mechanism and 1277 replace it with a structure and associated callback type. This 1278 mechanism allows a numeric "hash" to be extracted from a thread ID in 1279 either case, and on platforms where pointers are larger than 'long', 1280 mixing is done to help ensure the numeric 'hash' is usable even if it 1281 can't be guaranteed unique. The default mechanism is to use "&errno" 1282 as a pointer-based thread ID to distinguish between threads. 1283 1284 Applications that want to provide their own thread IDs should now use 1285 CRYPTO_THREADID_set_callback() to register a callback that will call 1286 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1287 1288 Note that ERR_remove_state() is now deprecated, because it is tied 1289 to the assumption that thread IDs are numeric. ERR_remove_state(0) 1290 to free the current thread's error state should be replaced by 1291 ERR_remove_thread_state(NULL). 1292 1293 (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1294 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1295 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1296 application was previously providing a numeric thread callback that 1297 was inappropriate for distinguishing threads, then uniqueness might 1298 have been obtained with &errno that happened immediately in the 1299 intermediate development versions of OpenSSL; this is no longer the 1300 case, the numeric thread callback will now override the automatic use 1301 of &errno.) 1302 [Geoff Thorpe, with help from Bodo Moeller] 1303 1304 *) Initial support for different CRL issuing certificates. This covers a 1305 simple case where the self issued certificates in the chain exist and 1306 the real CRL issuer is higher in the existing chain. 1307 1308 This work was sponsored by Google. 1309 [Steve Henson] 1310 1311 *) Removed effectively defunct crypto/store from the build. 1312 [Ben Laurie] 1313 1314 *) Revamp of STACK to provide stronger type-checking. Still to come: 1315 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1316 ASN1_STRING, CONF_VALUE. 1317 [Ben Laurie] 1318 1319 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1320 RAM on SSL connections. This option can save about 34k per idle SSL. 1321 [Nick Mathewson] 1322 1323 *) Revamp of LHASH to provide stronger type-checking. Still to come: 1324 STACK, TXT_DB, bsearch, qsort. 1325 [Ben Laurie] 1326 1327 *) Initial support for Cryptographic Message Syntax (aka CMS) based 1328 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1329 support for data, signedData, compressedData, digestedData and 1330 encryptedData, envelopedData types included. Scripts to check against 1331 RFC4134 examples draft and interop and consistency checks of many 1332 content types and variants. 1333 [Steve Henson] 1334 1335 *) Add options to enc utility to support use of zlib compression BIO. 1336 [Steve Henson] 1337 1338 *) Extend mk1mf to support importing of options and assembly language 1339 files from Configure script, currently only included in VC-WIN32. 1340 The assembly language rules can now optionally generate the source 1341 files from the associated perl scripts. 1342 [Steve Henson] 1343 1344 *) Implement remaining functionality needed to support GOST ciphersuites. 1345 Interop testing has been performed using CryptoPro implementations. 1346 [Victor B. Wagner <vitus@cryptocom.ru>] 1347 1348 *) s390x assembler pack. 1349 [Andy Polyakov] 1350 1351 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1352 "family." 1353 [Andy Polyakov] 1354 1355 *) Implement Opaque PRF Input TLS extension as specified in 1356 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1357 official specification yet and no extension type assignment by 1358 IANA exists, this extension (for now) will have to be explicitly 1359 enabled when building OpenSSL by providing the extension number 1360 to use. For example, specify an option 1361 1362 -DTLSEXT_TYPE_opaque_prf_input=0x9527 1363 1364 to the "config" or "Configure" script to enable the extension, 1365 assuming extension number 0x9527 (which is a completely arbitrary 1366 and unofficial assignment based on the MD5 hash of the Internet 1367 Draft). Note that by doing so, you potentially lose 1368 interoperability with other TLS implementations since these might 1369 be using the same extension number for other purposes. 1370 1371 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1372 opaque PRF input value to use in the handshake. This will create 1373 an interal copy of the length-'len' string at 'src', and will 1374 return non-zero for success. 1375 1376 To get more control and flexibility, provide a callback function 1377 by using 1378 1379 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1380 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1381 1382 where 1383 1384 int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1385 void *arg; 1386 1387 Callback function 'cb' will be called in handshakes, and is 1388 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1389 Argument 'arg' is for application purposes (the value as given to 1390 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1391 be provided to the callback function). The callback function 1392 has to return non-zero to report success: usually 1 to use opaque 1393 PRF input just if possible, or 2 to enforce use of the opaque PRF 1394 input. In the latter case, the library will abort the handshake 1395 if opaque PRF input is not successfully negotiated. 1396 1397 Arguments 'peerinput' and 'len' given to the callback function 1398 will always be NULL and 0 in the case of a client. A server will 1399 see the client's opaque PRF input through these variables if 1400 available (NULL and 0 otherwise). Note that if the server 1401 provides an opaque PRF input, the length must be the same as the 1402 length of the client's opaque PRF input. 1403 1404 Note that the callback function will only be called when creating 1405 a new session (session resumption can resume whatever was 1406 previously negotiated), and will not be called in SSL 2.0 1407 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1408 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1409 for applications that need to enforce opaque PRF input. 1410 1411 [Bodo Moeller] 1412 1413 *) Update ssl code to support digests other than SHA1+MD5 for handshake 1414 MAC. 1415 1416 [Victor B. Wagner <vitus@cryptocom.ru>] 1417 1418 *) Add RFC4507 support to OpenSSL. This includes the corrections in 1419 RFC4507bis. The encrypted ticket format is an encrypted encoded 1420 SSL_SESSION structure, that way new session features are automatically 1421 supported. 1422 1423 If a client application caches session in an SSL_SESSION structure 1424 support is transparent because tickets are now stored in the encoded 1425 SSL_SESSION. 1426 1427 The SSL_CTX structure automatically generates keys for ticket 1428 protection in servers so again support should be possible 1429 with no application modification. 1430 1431 If a client or server wishes to disable RFC4507 support then the option 1432 SSL_OP_NO_TICKET can be set. 1433 1434 Add a TLS extension debugging callback to allow the contents of any client 1435 or server extensions to be examined. 1436 1437 This work was sponsored by Google. 1438 [Steve Henson] 1439 1440 *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1441 OpenSSL should now compile cleanly on gcc 4.2 1442 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1443 1444 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1445 support including streaming MAC support: this is required for GOST 1446 ciphersuite support. 1447 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1448 1449 *) Add option -stream to use PKCS#7 streaming in smime utility. New 1450 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1451 to output in BER and PEM format. 1452 [Steve Henson] 1453 1454 *) Experimental support for use of HMAC via EVP_PKEY interface. This 1455 allows HMAC to be handled via the EVP_DigestSign*() interface. The 1456 EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1457 ENGINE support for HMAC keys which are unextractable. New -mac and 1458 -macopt options to dgst utility. 1459 [Steve Henson] 1460 1461 *) New option -sigopt to dgst utility. Update dgst to use 1462 EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1463 alternative signing paramaters such as X9.31 or PSS in the dgst 1464 utility. 1465 [Steve Henson] 1466 1467 *) Change ssl_cipher_apply_rule(), the internal function that does 1468 the work each time a ciphersuite string requests enabling 1469 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1470 removing ("!foo+bar") a class of ciphersuites: Now it maintains 1471 the order of disabled ciphersuites such that those ciphersuites 1472 that most recently went from enabled to disabled not only stay 1473 in order with respect to each other, but also have higher priority 1474 than other disabled ciphersuites the next time ciphersuites are 1475 enabled again. 1476 1477 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1478 the same ciphersuites as with "HIGH" alone, but in a specific 1479 order where the PSK ciphersuites come first (since they are the 1480 most recently disabled ciphersuites when "HIGH" is parsed). 1481 1482 Also, change ssl_create_cipher_list() (using this new 1483 funcionality) such that between otherwise identical 1484 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1485 the default order. 1486 [Bodo Moeller] 1487 1488 *) Change ssl_create_cipher_list() so that it automatically 1489 arranges the ciphersuites in reasonable order before starting 1490 to process the rule string. Thus, the definition for "DEFAULT" 1491 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1492 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1493 This makes it much easier to arrive at a reasonable default order 1494 in applications for which anonymous ciphers are OK (meaning 1495 that you can't actually use DEFAULT). 1496 [Bodo Moeller; suggested by Victor Duchovni] 1497 1498 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1499 processing) into multiple integers instead of setting 1500 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1501 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1502 (These masks as well as the individual bit definitions are hidden 1503 away into the non-exported interface ssl/ssl_locl.h, so this 1504 change to the definition of the SSL_CIPHER structure shouldn't 1505 affect applications.) This give us more bits for each of these 1506 categories, so there is no longer a need to coagulate AES128 and 1507 AES256 into a single algorithm bit, and to coagulate Camellia128 1508 and Camellia256 into a single algorithm bit, which has led to all 1509 kinds of kludges. 1510 1511 Thus, among other things, the kludge introduced in 0.9.7m and 1512 0.9.8e for masking out AES256 independently of AES128 or masking 1513 out Camellia256 independently of AES256 is not needed here in 0.9.9. 1514 1515 With the change, we also introduce new ciphersuite aliases that 1516 so far were missing: "AES128", "AES256", "CAMELLIA128", and 1517 "CAMELLIA256". 1518 [Bodo Moeller] 1519 1520 *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1521 Use the leftmost N bytes of the signature input if the input is 1522 larger than the prime q (with N being the size in bytes of q). 1523 [Nils Larsch] 1524 1525 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1526 it yet and it is largely untested. 1527 [Steve Henson] 1528 1529 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1530 [Nils Larsch] 1531 1532 *) Initial incomplete changes to avoid need for function casts in OpenSSL 1533 some compilers (gcc 4.2 and later) reject their use. Safestack is 1534 reimplemented. Update ASN1 to avoid use of legacy functions. 1535 [Steve Henson] 1536 1537 *) Win32/64 targets are linked with Winsock2. 1538 [Andy Polyakov] 1539 1540 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1541 to external functions. This can be used to increase CRL handling 1542 efficiency especially when CRLs are very large by (for example) storing 1543 the CRL revoked certificates in a database. 1544 [Steve Henson] 1545 1546 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1547 new CRLs added to a directory can be used. New command line option 1548 -verify_return_error to s_client and s_server. This causes real errors 1549 to be returned by the verify callback instead of carrying on no matter 1550 what. This reflects the way a "real world" verify callback would behave. 1551 [Steve Henson] 1552 1553 *) GOST engine, supporting several GOST algorithms and public key formats. 1554 Kindly donated by Cryptocom. 1555 [Cryptocom] 1556 1557 *) Partial support for Issuing Distribution Point CRL extension. CRLs 1558 partitioned by DP are handled but no indirect CRL or reason partitioning 1559 (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1560 selected via a scoring technique which handles IDP and AKID in CRLs. 1561 [Steve Henson] 1562 1563 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1564 will ultimately be used for all verify operations: this will remove the 1565 X509_STORE dependency on certificate verification and allow alternative 1566 lookup methods. X509_STORE based implementations of these two callbacks. 1567 [Steve Henson] 1568 1569 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1570 Modify get_crl() to find a valid (unexpired) CRL if possible. 1571 [Steve Henson] 1572 1573 *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1574 this would be called X509_CRL_cmp() but that name is already used by 1575 a function that just compares CRL issuer names. Cache several CRL 1576 extensions in X509_CRL structure and cache CRLDP in X509. 1577 [Steve Henson] 1578 1579 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1580 this maps equivalent X509_NAME structures into a consistent structure. 1581 Name comparison can then be performed rapidly using memcmp(). 1582 [Steve Henson] 1583 1584 *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1585 utility. 1586 [Steve Henson] 1587 1588 *) Allow digests to supply their own micalg string for S/MIME type using 1589 the ctrl EVP_MD_CTRL_MICALG. 1590 [Steve Henson] 1591 1592 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1593 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1594 ctrl. It can then customise the structure before and/or after signing 1595 if necessary. 1596 [Steve Henson] 1597 1598 *) New function OBJ_add_sigid() to allow application defined signature OIDs 1599 to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1600 to free up any added signature OIDs. 1601 [Steve Henson] 1602 1603 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1604 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1605 digest and cipher tables. New options added to openssl utility: 1606 list-message-digest-algorithms and list-cipher-algorithms. 1607 [Steve Henson] 1608 1609 *) Change the array representation of binary polynomials: the list 1610 of degrees of non-zero coefficients is now terminated with -1. 1611 Previously it was terminated with 0, which was also part of the 1612 value; thus, the array representation was not applicable to 1613 polynomials where t^0 has coefficient zero. This change makes 1614 the array representation useful in a more general context. 1615 [Douglas Stebila] 1616 1617 *) Various modifications and fixes to SSL/TLS cipher string 1618 handling. For ECC, the code now distinguishes between fixed ECDH 1619 with RSA certificates on the one hand and with ECDSA certificates 1620 on the other hand, since these are separate ciphersuites. The 1621 unused code for Fortezza ciphersuites has been removed. 1622 1623 For consistency with EDH, ephemeral ECDH is now called "EECDH" 1624 (not "ECDHE"). For consistency with the code for DH 1625 certificates, use of ECDH certificates is now considered ECDH 1626 authentication, not RSA or ECDSA authentication (the latter is 1627 merely the CA's signing algorithm and not actively used in the 1628 protocol). 1629 1630 The temporary ciphersuite alias "ECCdraft" is no longer 1631 available, and ECC ciphersuites are no longer excluded from "ALL" 1632 and "DEFAULT". The following aliases now exist for RFC 4492 1633 ciphersuites, most of these by analogy with the DH case: 1634 1635 kECDHr - ECDH cert, signed with RSA 1636 kECDHe - ECDH cert, signed with ECDSA 1637 kECDH - ECDH cert (signed with either RSA or ECDSA) 1638 kEECDH - ephemeral ECDH 1639 ECDH - ECDH cert or ephemeral ECDH 1640 1641 aECDH - ECDH cert 1642 aECDSA - ECDSA cert 1643 ECDSA - ECDSA cert 1644 1645 AECDH - anonymous ECDH 1646 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1647 1648 [Bodo Moeller] 1649 1650 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1651 Use correct micalg parameters depending on digest(s) in signed message. 1652 [Steve Henson] 1653 1654 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1655 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1656 [Steve Henson] 1657 1658 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1659 an engine to register a method. Add ENGINE lookups for methods and 1660 functional reference processing. 1661 [Steve Henson] 1662 1663 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1664 EVP_{Sign,Verify}* which allow an application to customise the signature 1665 process. 1666 [Steve Henson] 1667 1668 *) New -resign option to smime utility. This adds one or more signers 1669 to an existing PKCS#7 signedData structure. Also -md option to use an 1670 alternative message digest algorithm for signing. 1671 [Steve Henson] 1672 1673 *) Tidy up PKCS#7 routines and add new functions to make it easier to 1674 create PKCS7 structures containing multiple signers. Update smime 1675 application to support multiple signers. 1676 [Steve Henson] 1677 1678 *) New -macalg option to pkcs12 utility to allow setting of an alternative 1679 digest MAC. 1680 [Steve Henson] 1681 1682 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1683 Reorganize PBE internals to lookup from a static table using NIDs, 1684 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1685 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1686 PRF which will be automatically used with PBES2. 1687 [Steve Henson] 1688 1689 *) Replace the algorithm specific calls to generate keys in "req" with the 1690 new API. 1691 [Steve Henson] 1692 1693 *) Update PKCS#7 enveloped data routines to use new API. This is now 1694 supported by any public key method supporting the encrypt operation. A 1695 ctrl is added to allow the public key algorithm to examine or modify 1696 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1697 a no op. 1698 [Steve Henson] 1699 1700 *) Add a ctrl to asn1 method to allow a public key algorithm to express 1701 a default digest type to use. In most cases this will be SHA1 but some 1702 algorithms (such as GOST) need to specify an alternative digest. The 1703 return value indicates how strong the prefernce is 1 means optional and 1704 2 is mandatory (that is it is the only supported type). Modify 1705 ASN1_item_sign() to accept a NULL digest argument to indicate it should 1706 use the default md. Update openssl utilities to use the default digest 1707 type for signing if it is not explicitly indicated. 1708 [Steve Henson] 1709 1710 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1711 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1712 signing method from the key type. This effectively removes the link 1713 between digests and public key types. 1714 [Steve Henson] 1715 1716 *) Add an OID cross reference table and utility functions. Its purpose is to 1717 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1718 rsaEncryption. This will allow some of the algorithm specific hackery 1719 needed to use the correct OID to be removed. 1720 [Steve Henson] 1721 1722 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1723 structures for PKCS7_sign(). They are now set up by the relevant public 1724 key ASN1 method. 1725 [Steve Henson] 1726 1727 *) Add provisional EC pkey method with support for ECDSA and ECDH. 1728 [Steve Henson] 1729 1730 *) Add support for key derivation (agreement) in the API, DH method and 1731 pkeyutl. 1732 [Steve Henson] 1733 1734 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1735 public and private key formats. As a side effect these add additional 1736 command line functionality not previously available: DSA signatures can be 1737 generated and verified using pkeyutl and DH key support and generation in 1738 pkey, genpkey. 1739 [Steve Henson] 1740 1741 *) BeOS support. 1742 [Oliver Tappe <zooey@hirschkaefer.de>] 1743 1744 *) New make target "install_html_docs" installs HTML renditions of the 1745 manual pages. 1746 [Oliver Tappe <zooey@hirschkaefer.de>] 1747 1748 *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1749 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1750 support key and parameter generation and add initial key generation 1751 functionality for RSA. 1752 [Steve Henson] 1753 1754 *) Add functions for main EVP_PKEY_method operations. The undocumented 1755 functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1756 EVP_PKEY_{encrypt,decrypt}_old. 1757 [Steve Henson] 1758 1759 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1760 key API, doesn't do much yet. 1761 [Steve Henson] 1762 1763 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1764 public key algorithms. New option to openssl utility: 1765 "list-public-key-algorithms" to print out info. 1766 [Steve Henson] 1767 1768 *) Implement the Supported Elliptic Curves Extension for 1769 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1770 [Douglas Stebila] 1771 1772 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1773 EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1774 [Steve Henson] 1775 1776 *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1777 utilities such as rsa, dsa, dsaparam etc except they process any key 1778 type. 1779 [Steve Henson] 1780 1781 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1782 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1783 EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1784 structure. 1785 [Steve Henson] 1786 1787 *) Initial support for pluggable public key ASN1. 1788 De-spaghettify the public key ASN1 handling. Move public and private 1789 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1790 algorithm specific handling to a single module within the relevant 1791 algorithm directory. Add functions to allow (near) opaque processing 1792 of public and private key structures. 1793 [Steve Henson] 1794 1795 *) Implement the Supported Point Formats Extension for 1796 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1797 [Douglas Stebila] 1798 1799 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1800 for the psk identity [hint] and the psk callback functions to the 1801 SSL_SESSION, SSL and SSL_CTX structure. 1802 1803 New ciphersuites: 1804 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1805 PSK-AES256-CBC-SHA 1806 1807 New functions: 1808 SSL_CTX_use_psk_identity_hint 1809 SSL_get_psk_identity_hint 1810 SSL_get_psk_identity 1811 SSL_use_psk_identity_hint 1812 1813 [Mika Kousa and Pasi Eronen of Nokia Corporation] 1814 1815 *) Add RFC 3161 compliant time stamp request creation, response generation 1816 and response verification functionality. 1817 [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1818 1819 *) Add initial support for TLS extensions, specifically for the server_name 1820 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1821 have new members for a host name. The SSL data structure has an 1822 additional member SSL_CTX *initial_ctx so that new sessions can be 1823 stored in that context to allow for session resumption, even after the 1824 SSL has been switched to a new SSL_CTX in reaction to a client's 1825 server_name extension. 1826 1827 New functions (subject to change): 1828 1829 SSL_get_servername() 1830 SSL_get_servername_type() 1831 SSL_set_SSL_CTX() 1832 1833 New CTRL codes and macros (subject to change): 1834 1835 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1836 - SSL_CTX_set_tlsext_servername_callback() 1837 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1838 - SSL_CTX_set_tlsext_servername_arg() 1839 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1840 1841 openssl s_client has a new '-servername ...' option. 1842 1843 openssl s_server has new options '-servername_host ...', '-cert2 ...', 1844 '-key2 ...', '-servername_fatal' (subject to change). This allows 1845 testing the HostName extension for a specific single host name ('-cert' 1846 and '-key' remain fallbacks for handshakes without HostName 1847 negotiation). If the unrecogninzed_name alert has to be sent, this by 1848 default is a warning; it becomes fatal with the '-servername_fatal' 1849 option. 1850 1851 [Peter Sylvester, Remy Allais, Christophe Renou] 1852 1853 *) Whirlpool hash implementation is added. 1854 [Andy Polyakov] 1855 1856 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1857 bn(64,32). Because of instruction set limitations it doesn't have 1858 any negative impact on performance. This was done mostly in order 1859 to make it possible to share assembler modules, such as bn_mul_mont 1860 implementations, between 32- and 64-bit builds without hassle. 1861 [Andy Polyakov] 1862 1863 *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1864 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1865 macro. 1866 [Bodo Moeller] 1867 1868 *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1869 dedicated Montgomery multiplication procedure, is introduced. 1870 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1871 "64-bit" performance on certain 32-bit targets. 1872 [Andy Polyakov] 1873 1874 *) New option SSL_OP_NO_COMP to disable use of compression selectively 1875 in SSL structures. New SSL ctrl to set maximum send fragment size. 1876 Save memory by seeting the I/O buffer sizes dynamically instead of 1877 using the maximum available value. 1878 [Steve Henson] 1879 1880 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1881 in addition to the text details. 1882 [Bodo Moeller] 1883 1884 *) Very, very preliminary EXPERIMENTAL support for printing of general 1885 ASN1 structures. This currently produces rather ugly output and doesn't 1886 handle several customised structures at all. 1887 [Steve Henson] 1888 1889 *) Integrated support for PVK file format and some related formats such 1890 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1891 these in the 'rsa' and 'dsa' utilities. 1892 [Steve Henson] 1893 1894 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1895 [Steve Henson] 1896 1897 *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1898 place for the (very old) "NETSCAPE" format certificates which are now 1899 handled using new ASN1 code equivalents. 1900 [Steve Henson] 1901 1902 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1903 pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1904 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1905 [Nils Larsch] 1906 1907 *) Modify CRL distribution points extension code to print out previously 1908 unsupported fields. Enhance extension setting code to allow setting of 1909 all fields. 1910 [Steve Henson] 1911 1912 *) Add print and set support for Issuing Distribution Point CRL extension. 1913 [Steve Henson] 1914 1915 *) Change 'Configure' script to enable Camellia by default. 1916 [NTT] 1917 1918 Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 1919 1920 *) When rejecting SSL/TLS records due to an incorrect version number, never 1921 update s->server with a new major version number. As of 1922 - OpenSSL 0.9.8m if 'short' is a 16-bit type, 1923 - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 1924 the previous behavior could result in a read attempt at NULL when 1925 receiving specific incorrect SSL/TLS records once record payload 1926 protection is active. (CVE-2010-0740) 1927 [Bodo Moeller, Adam Langley <agl@chromium.org>] 1928 1929 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 1930 could be crashed if the relevant tables were not present (e.g. chrooted). 1931 [Tomas Hoger <thoger@redhat.com>] 1932 1933 Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 1934 1935 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 1936 [Martin Olsson, Neel Mehta] 1937 1938 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 1939 accommodate for stack sorting, always a write lock!). 1940 [Bodo Moeller] 1941 1942 *) On some versions of WIN32 Heap32Next is very slow. This can cause 1943 excessive delays in the RAND_poll(): over a minute. As a workaround 1944 include a time check in the inner Heap32Next loop too. 1945 [Steve Henson] 1946 1947 *) The code that handled flushing of data in SSL/TLS originally used the 1948 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 1949 the problem outlined in PR#1949. The fix suggested there however can 1950 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 1951 of Apache). So instead simplify the code to flush unconditionally. 1952 This should be fine since flushing with no data to flush is a no op. 1953 [Steve Henson] 1954 1955 *) Handle TLS versions 2.0 and later properly and correctly use the 1956 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 1957 off ancient servers have a habit of sticking around for a while... 1958 [Steve Henson] 1959 1960 *) Modify compression code so it frees up structures without using the 1961 ex_data callbacks. This works around a problem where some applications 1962 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 1963 restarting) then use compression (e.g. SSL with compression) later. 1964 This results in significant per-connection memory leaks and 1965 has caused some security issues including CVE-2008-1678 and 1966 CVE-2009-4355. 1967 [Steve Henson] 1968 1969 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 1970 change when encrypting or decrypting. 1971 [Bodo Moeller] 1972 1973 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 1974 connect and renegotiate with servers which do not support RI. 1975 Until RI is more widely deployed this option is enabled by default. 1976 [Steve Henson] 1977 1978 *) Add "missing" ssl ctrls to clear options and mode. 1979 [Steve Henson] 1980 1981 *) If client attempts to renegotiate and doesn't support RI respond with 1982 a no_renegotiation alert as required by RFC5746. Some renegotiating 1983 TLS clients will continue a connection gracefully when they receive 1984 the alert. Unfortunately OpenSSL mishandled this alert and would hang 1985 waiting for a server hello which it will never receive. Now we treat a 1986 received no_renegotiation alert as a fatal error. This is because 1987 applications requesting a renegotiation might well expect it to succeed 1988 and would have no code in place to handle the server denying it so the 1989 only safe thing to do is to terminate the connection. 1990 [Steve Henson] 1991 1992 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 1993 peer supports secure renegotiation and 0 otherwise. Print out peer 1994 renegotiation support in s_client/s_server. 1995 [Steve Henson] 1996 1997 *) Replace the highly broken and deprecated SPKAC certification method with 1998 the updated NID creation version. This should correctly handle UTF8. 1999 [Steve Henson] 2000 2001 *) Implement RFC5746. Re-enable renegotiation but require the extension 2002 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 2003 turns out to be a bad idea. It has been replaced by 2004 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 2005 SSL_CTX_set_options(). This is really not recommended unless you 2006 know what you are doing. 2007 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 2008 2009 *) Fixes to stateless session resumption handling. Use initial_ctx when 2010 issuing and attempting to decrypt tickets in case it has changed during 2011 servername handling. Use a non-zero length session ID when attempting 2012 stateless session resumption: this makes it possible to determine if 2013 a resumption has occurred immediately after receiving server hello 2014 (several places in OpenSSL subtly assume this) instead of later in 2015 the handshake. 2016 [Steve Henson] 2017 2018 *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2019 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2020 fixes for a few places where the return code is not checked 2021 correctly. 2022 [Julia Lawall <julia@diku.dk>] 2023 2024 *) Add --strict-warnings option to Configure script to include devteam 2025 warnings in other configurations. 2026 [Steve Henson] 2027 2028 *) Add support for --libdir option and LIBDIR variable in makefiles. This 2029 makes it possible to install openssl libraries in locations which 2030 have names other than "lib", for example "/usr/lib64" which some 2031 systems need. 2032 [Steve Henson, based on patch from Jeremy Utley] 2033 2034 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 2035 X690 8.9.12 and can produce some misleading textual output of OIDs. 2036 [Steve Henson, reported by Dan Kaminsky] 2037 2038 *) Delete MD2 from algorithm tables. This follows the recommendation in 2039 several standards that it is not used in new applications due to 2040 several cryptographic weaknesses. For binary compatibility reasons 2041 the MD2 API is still compiled in by default. 2042 [Steve Henson] 2043 2044 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 2045 and restored. 2046 [Steve Henson] 2047 2048 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 2049 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 2050 clash. 2051 [Guenter <lists@gknw.net>] 2052 2053 *) Fix the server certificate chain building code to use X509_verify_cert(), 2054 it used to have an ad-hoc builder which was unable to cope with anything 2055 other than a simple chain. 2056 [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 2057 2058 *) Don't check self signed certificate signatures in X509_verify_cert() 2059 by default (a flag can override this): it just wastes time without 2060 adding any security. As a useful side effect self signed root CAs 2061 with non-FIPS digests are now usable in FIPS mode. 2062 [Steve Henson] 2063 2064 *) In dtls1_process_out_of_seq_message() the check if the current message 2065 is already buffered was missing. For every new message was memory 2066 allocated, allowing an attacker to perform an denial of service attack 2067 with sending out of seq handshake messages until there is no memory 2068 left. Additionally every future messege was buffered, even if the 2069 sequence number made no sense and would be part of another handshake. 2070 So only messages with sequence numbers less than 10 in advance will be 2071 buffered. (CVE-2009-1378) 2072 [Robin Seggelmann, discovered by Daniel Mentz] 2073 2074 *) Records are buffered if they arrive with a future epoch to be 2075 processed after finishing the corresponding handshake. There is 2076 currently no limitation to this buffer allowing an attacker to perform 2077 a DOS attack with sending records with future epochs until there is no 2078 memory left. This patch adds the pqueue_size() function to detemine 2079 the size of a buffer and limits the record buffer to 100 entries. 2080 (CVE-2009-1377) 2081 [Robin Seggelmann, discovered by Daniel Mentz] 2082 2083 *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2084 parent structure is freed. (CVE-2009-1379) 2085 [Daniel Mentz] 2086 2087 *) Handle non-blocking I/O properly in SSL_shutdown() call. 2088 [Darryl Miles <darryl-mailinglists@netbauds.net>] 2089 2090 *) Add 2.5.4.* OIDs 2091 [Ilya O. <vrghost@gmail.com>] 2092 2093 Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2094 2095 *) Disable renegotiation completely - this fixes a severe security 2096 problem (CVE-2009-3555) at the cost of breaking all 2097 renegotiation. Renegotiation can be re-enabled by setting 2098 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2099 run-time. This is really not recommended unless you know what 2100 you're doing. 2101 [Ben Laurie] 2102 2103 Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2104 2105 *) Don't set val to NULL when freeing up structures, it is freed up by 2106 underlying code. If sizeof(void *) > sizeof(long) this can result in 2107 zeroing past the valid field. (CVE-2009-0789) 2108 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2109 2110 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2111 checked correctly. This would allow some invalid signed attributes to 2112 appear to verify correctly. (CVE-2009-0591) 2113 [Ivan Nestlerode <inestlerode@us.ibm.com>] 2114 2115 *) Reject UniversalString and BMPString types with invalid lengths. This 2116 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2117 a legal length. (CVE-2009-0590) 2118 [Steve Henson] 2119 2120 *) Set S/MIME signing as the default purpose rather than setting it 2121 unconditionally. This allows applications to override it at the store 2122 level. 2123 [Steve Henson] 2124 2125 *) Permit restricted recursion of ASN1 strings. This is needed in practice 2126 to handle some structures. 2127 [Steve Henson] 2128 2129 *) Improve efficiency of mem_gets: don't search whole buffer each time 2130 for a '\n' 2131 [Jeremy Shapiro <jnshapir@us.ibm.com>] 2132 2133 *) New -hex option for openssl rand. 2134 [Matthieu Herrb] 2135 2136 *) Print out UTF8String and NumericString when parsing ASN1. 2137 [Steve Henson] 2138 2139 *) Support NumericString type for name components. 2140 [Steve Henson] 2141 2142 *) Allow CC in the environment to override the automatically chosen 2143 compiler. Note that nothing is done to ensure flags work with the 2144 chosen compiler. 2145 [Ben Laurie] 2146 2147 Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2148 2149 *) Properly check EVP_VerifyFinal() and similar return values 2150 (CVE-2008-5077). 2151 [Ben Laurie, Bodo Moeller, Google Security Team] 2152 2153 *) Enable TLS extensions by default. 2154 [Ben Laurie] 2155 2156 *) Allow the CHIL engine to be loaded, whether the application is 2157 multithreaded or not. (This does not release the developer from the 2158 obligation to set up the dynamic locking callbacks.) 2159 [Sander Temme <sander@temme.net>] 2160 2161 *) Use correct exit code if there is an error in dgst command. 2162 [Steve Henson; problem pointed out by Roland Dirlewanger] 2163 2164 *) Tweak Configure so that you need to say "experimental-jpake" to enable 2165 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2166 [Bodo Moeller] 2167 2168 *) Add experimental JPAKE support, including demo authentication in 2169 s_client and s_server. 2170 [Ben Laurie] 2171 2172 *) Set the comparison function in v3_addr_canonize(). 2173 [Rob Austein <sra@hactrn.net>] 2174 2175 *) Add support for XMPP STARTTLS in s_client. 2176 [Philip Paeps <philip@freebsd.org>] 2177 2178 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2179 to ensure that even with this option, only ciphersuites in the 2180 server's preference list will be accepted. (Note that the option 2181 applies only when resuming a session, so the earlier behavior was 2182 just about the algorithm choice for symmetric cryptography.) 2183 [Bodo Moeller] 2184 2185 Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2186 2187 *) Fix NULL pointer dereference if a DTLS server received 2188 ChangeCipherSpec as first record (CVE-2009-1386). 2189 [PR #1679] 2190 2191 *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2192 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2193 [Nagendra Modadugu] 2194 2195 *) The fix in 0.9.8c that supposedly got rid of unsafe 2196 double-checked locking was incomplete for RSA blinding, 2197 addressing just one layer of what turns out to have been 2198 doubly unsafe triple-checked locking. 2199 2200 So now fix this for real by retiring the MONT_HELPER macro 2201 in crypto/rsa/rsa_eay.c. 2202 2203 [Bodo Moeller; problem pointed out by Marius Schilder] 2204 2205 *) Various precautionary measures: 2206 2207 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2208 2209 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2210 (NB: This would require knowledge of the secret session ticket key 2211 to exploit, in which case you'd be SOL either way.) 2212 2213 - Change bn_nist.c so that it will properly handle input BIGNUMs 2214 outside the expected range. 2215 2216 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2217 builds. 2218 2219 [Neel Mehta, Bodo Moeller] 2220 2221 *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2222 the load fails. Useful for distros. 2223 [Ben Laurie and the FreeBSD team] 2224 2225 *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2226 [Steve Henson] 2227 2228 *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2229 [Huang Ying] 2230 2231 *) Expand ENGINE to support engine supplied SSL client certificate functions. 2232 2233 This work was sponsored by Logica. 2234 [Steve Henson] 2235 2236 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2237 keystores. Support for SSL/TLS client authentication too. 2238 Not compiled unless enable-capieng specified to Configure. 2239 2240 This work was sponsored by Logica. 2241 [Steve Henson] 2242 2243 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2244 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2245 attribute creation routines such as certifcate requests and PKCS#12 2246 files. 2247 [Steve Henson] 2248 2249 Changes between 0.9.8g and 0.9.8h [28 May 2008] 2250 2251 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2252 handshake which could lead to a cilent crash as found using the 2253 Codenomicon TLS test suite (CVE-2008-1672) 2254 [Steve Henson, Mark Cox] 2255 2256 *) Fix double free in TLS server name extensions which could lead to 2257 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2258 [Joe Orton] 2259 2260 *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2261 2262 Clear the error queue to ensure that error entries left from 2263 older function calls do not interfere with the correct operation. 2264 [Lutz Jaenicke, Erik de Castro Lopo] 2265 2266 *) Remove root CA certificates of commercial CAs: 2267 2268 The OpenSSL project does not recommend any specific CA and does not 2269 have any policy with respect to including or excluding any CA. 2270 Therefore it does not make any sense to ship an arbitrary selection 2271 of root CA certificates with the OpenSSL software. 2272 [Lutz Jaenicke] 2273 2274 *) RSA OAEP patches to fix two separate invalid memory reads. 2275 The first one involves inputs when 'lzero' is greater than 2276 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2277 before the beginning of from). The second one involves inputs where 2278 the 'db' section contains nothing but zeroes (there is a one-byte 2279 invalid read after the end of 'db'). 2280 [Ivan Nestlerode <inestlerode@us.ibm.com>] 2281 2282 *) Partial backport from 0.9.9-dev: 2283 2284 Introduce bn_mul_mont (dedicated Montgomery multiplication 2285 procedure) as a candidate for BIGNUM assembler implementation. 2286 While 0.9.9-dev uses assembler for various architectures, only 2287 x86_64 is available by default here in the 0.9.8 branch, and 2288 32-bit x86 is available through a compile-time setting. 2289 2290 To try the 32-bit x86 assembler implementation, use Configure 2291 option "enable-montasm" (which exists only for this backport). 2292 2293 As "enable-montasm" for 32-bit x86 disclaims code stability 2294 anyway, in this constellation we activate additional code 2295 backported from 0.9.9-dev for further performance improvements, 2296 namely BN_from_montgomery_word. (To enable this otherwise, 2297 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2298 2299 [Andy Polyakov (backport partially by Bodo Moeller)] 2300 2301 *) Add TLS session ticket callback. This allows an application to set 2302 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2303 values. This is useful for key rollover for example where several key 2304 sets may exist with different names. 2305 [Steve Henson] 2306 2307 *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2308 This was broken until now in 0.9.8 releases, such that the only way 2309 a registered ENGINE could be used (assuming it initialises 2310 successfully on the host) was to explicitly set it as the default 2311 for the relevant algorithms. This is in contradiction with 0.9.7 2312 behaviour and the documentation. With this fix, when an ENGINE is 2313 registered into a given algorithm's table of implementations, the 2314 'uptodate' flag is reset so that auto-discovery will be used next 2315 time a new context for that algorithm attempts to select an 2316 implementation. 2317 [Ian Lister (tweaked by Geoff Thorpe)] 2318 2319 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2320 implemention in the following ways: 2321 2322 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2323 hard coded. 2324 2325 Lack of BER streaming support means one pass streaming processing is 2326 only supported if data is detached: setting the streaming flag is 2327 ignored for embedded content. 2328 2329 CMS support is disabled by default and must be explicitly enabled 2330 with the enable-cms configuration option. 2331 [Steve Henson] 2332 2333 *) Update the GMP engine glue to do direct copies between BIGNUM and 2334 mpz_t when openssl and GMP use the same limb size. Otherwise the 2335 existing "conversion via a text string export" trick is still used. 2336 [Paul Sheer <paulsheer@gmail.com>] 2337 2338 *) Zlib compression BIO. This is a filter BIO which compressed and 2339 uncompresses any data passed through it. 2340 [Steve Henson] 2341 2342 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2343 RFC3394 compatible AES key wrapping. 2344 [Steve Henson] 2345 2346 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2347 sets string data without copying. X509_ALGOR_set0() and 2348 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2349 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2350 from an X509_ATTRIBUTE structure optionally checking it occurs only 2351 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2352 data. 2353 [Steve Henson] 2354 2355 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2356 to get the expected BN_FLG_CONSTTIME behavior. 2357 [Bodo Moeller (Google)] 2358 2359 *) Netware support: 2360 2361 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2362 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2363 - added some more tests to do_tests.pl 2364 - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2365 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2366 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2367 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2368 - various changes to netware.pl to enable gcc-cross builds on Win32 2369 platform 2370 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2371 - various changes to fix missing prototype warnings 2372 - fixed x86nasm.pl to create correct asm files for NASM COFF output 2373 - added AES, WHIRLPOOL and CPUID assembler code to build files 2374 - added missing AES assembler make rules to mk1mf.pl 2375 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2376 [Guenter Knauf <eflash@gmx.net>] 2377 2378 *) Implement certificate status request TLS extension defined in RFC3546. 2379 A client can set the appropriate parameters and receive the encoded 2380 OCSP response via a callback. A server can query the supplied parameters 2381 and set the encoded OCSP response in the callback. Add simplified examples 2382 to s_client and s_server. 2383 [Steve Henson] 2384 2385 Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2386 2387 *) Fix various bugs: 2388 + Binary incompatibility of ssl_ctx_st structure 2389 + DTLS interoperation with non-compliant servers 2390 + Don't call get_session_cb() without proposed session 2391 + Fix ia64 assembler code 2392 [Andy Polyakov, Steve Henson] 2393 2394 Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2395 2396 *) DTLS Handshake overhaul. There were longstanding issues with 2397 OpenSSL DTLS implementation, which were making it impossible for 2398 RFC 4347 compliant client to communicate with OpenSSL server. 2399 Unfortunately just fixing these incompatibilities would "cut off" 2400 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2401 server keeps tolerating non RFC compliant syntax. The opposite is 2402 not true, 0.9.8f client can not communicate with earlier server. 2403 This update even addresses CVE-2007-4995. 2404 [Andy Polyakov] 2405 2406 *) Changes to avoid need for function casts in OpenSSL: some compilers 2407 (gcc 4.2 and later) reject their use. 2408 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2409 Steve Henson] 2410 2411 *) Add RFC4507 support to OpenSSL. This includes the corrections in 2412 RFC4507bis. The encrypted ticket format is an encrypted encoded 2413 SSL_SESSION structure, that way new session features are automatically 2414 supported. 2415 2416 If a client application caches session in an SSL_SESSION structure 2417 support is transparent because tickets are now stored in the encoded 2418 SSL_SESSION. 2419 2420 The SSL_CTX structure automatically generates keys for ticket 2421 protection in servers so again support should be possible 2422 with no application modification. 2423 2424 If a client or server wishes to disable RFC4507 support then the option 2425 SSL_OP_NO_TICKET can be set. 2426 2427 Add a TLS extension debugging callback to allow the contents of any client 2428 or server extensions to be examined. 2429 2430 This work was sponsored by Google. 2431 [Steve Henson] 2432 2433 *) Add initial support for TLS extensions, specifically for the server_name 2434 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2435 have new members for a host name. The SSL data structure has an 2436 additional member SSL_CTX *initial_ctx so that new sessions can be 2437 stored in that context to allow for session resumption, even after the 2438 SSL has been switched to a new SSL_CTX in reaction to a client's 2439 server_name extension. 2440 2441 New functions (subject to change): 2442 2443 SSL_get_servername() 2444 SSL_get_servername_type() 2445 SSL_set_SSL_CTX() 2446 2447 New CTRL codes and macros (subject to change): 2448 2449 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2450 - SSL_CTX_set_tlsext_servername_callback() 2451 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2452 - SSL_CTX_set_tlsext_servername_arg() 2453 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2454 2455 openssl s_client has a new '-servername ...' option. 2456 2457 openssl s_server has new options '-servername_host ...', '-cert2 ...', 2458 '-key2 ...', '-servername_fatal' (subject to change). This allows 2459 testing the HostName extension for a specific single host name ('-cert' 2460 and '-key' remain fallbacks for handshakes without HostName 2461 negotiation). If the unrecogninzed_name alert has to be sent, this by 2462 default is a warning; it becomes fatal with the '-servername_fatal' 2463 option. 2464 2465 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2466 2467 *) Add AES and SSE2 assembly language support to VC++ build. 2468 [Steve Henson] 2469 2470 *) Mitigate attack on final subtraction in Montgomery reduction. 2471 [Andy Polyakov] 2472 2473 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2474 (which previously caused an internal error). 2475 [Bodo Moeller] 2476 2477 *) Squeeze another 10% out of IGE mode when in != out. 2478 [Ben Laurie] 2479 2480 *) AES IGE mode speedup. 2481 [Dean Gaudet (Google)] 2482 2483 *) Add the Korean symmetric 128-bit cipher SEED (see 2484 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2485 add SEED ciphersuites from RFC 4162: 2486 2487 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2488 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2489 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2490 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2491 2492 To minimize changes between patchlevels in the OpenSSL 0.9.8 2493 series, SEED remains excluded from compilation unless OpenSSL 2494 is configured with 'enable-seed'. 2495 [KISA, Bodo Moeller] 2496 2497 *) Mitigate branch prediction attacks, which can be practical if a 2498 single processor is shared, allowing a spy process to extract 2499 information. For detailed background information, see 2500 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2501 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2502 and Necessary Software Countermeasures"). The core of the change 2503 are new versions BN_div_no_branch() and 2504 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2505 respectively, which are slower, but avoid the security-relevant 2506 conditional branches. These are automatically called by BN_div() 2507 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2508 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2509 remove a conditional branch. 2510 2511 BN_FLG_CONSTTIME is the new name for the previous 2512 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2513 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2514 in the exponent causes BN_mod_exp_mont() to use the alternative 2515 implementation in BN_mod_exp_mont_consttime().) The old name 2516 remains as a deprecated alias. 2517 2518 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2519 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2520 constant-time implementations for more than just exponentiation. 2521 Here too the old name is kept as a deprecated alias. 2522 2523 BN_BLINDING_new() will now use BN_dup() for the modulus so that 2524 the BN_BLINDING structure gets an independent copy of the 2525 modulus. This means that the previous "BIGNUM *m" argument to 2526 BN_BLINDING_new() and to BN_BLINDING_create_param() now 2527 essentially becomes "const BIGNUM *m", although we can't actually 2528 change this in the header file before 0.9.9. It allows 2529 RSA_setup_blinding() to use BN_with_flags() on the modulus to 2530 enable BN_FLG_CONSTTIME. 2531 2532 [Matthew D Wood (Intel Corp)] 2533 2534 *) In the SSL/TLS server implementation, be strict about session ID 2535 context matching (which matters if an application uses a single 2536 external cache for different purposes). Previously, 2537 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2538 set. This did ensure strict client verification, but meant that, 2539 with applications using a single external cache for quite 2540 different requirements, clients could circumvent ciphersuite 2541 restrictions for a given session ID context by starting a session 2542 in a different context. 2543 [Bodo Moeller] 2544 2545 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2546 a ciphersuite string such as "DEFAULT:RSA" cannot enable 2547 authentication-only ciphersuites. 2548 [Bodo Moeller] 2549 2550 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2551 not complete and could lead to a possible single byte overflow 2552 (CVE-2007-5135) [Ben Laurie] 2553 2554 Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2555 2556 *) Since AES128 and AES256 (and similarly Camellia128 and 2557 Camellia256) share a single mask bit in the logic of 2558 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2559 kludge to work properly if AES128 is available and AES256 isn't 2560 (or if Camellia128 is available and Camellia256 isn't). 2561 [Victor Duchovni] 2562 2563 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2564 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2565 When a point or a seed is encoded in a BIT STRING, we need to 2566 prevent the removal of trailing zero bits to get the proper DER 2567 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2568 of a NamedBitList, for which trailing 0 bits need to be removed.) 2569 [Bodo Moeller] 2570 2571 *) Have SSL/TLS server implementation tolerate "mismatched" record 2572 protocol version while receiving ClientHello even if the 2573 ClientHello is fragmented. (The server can't insist on the 2574 particular protocol version it has chosen before the ServerHello 2575 message has informed the client about his choice.) 2576 [Bodo Moeller] 2577 2578 *) Add RFC 3779 support. 2579 [Rob Austein for ARIN, Ben Laurie] 2580 2581 *) Load error codes if they are not already present instead of using a 2582 static variable. This allows them to be cleanly unloaded and reloaded. 2583 Improve header file function name parsing. 2584 [Steve Henson] 2585 2586 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2587 or CAPABILITY handshake as required by RFCs. 2588 [Goetz Babin-Ebell] 2589 2590 Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2591 2592 *) Introduce limits to prevent malicious keys being able to 2593 cause a denial of service. (CVE-2006-2940) 2594 [Steve Henson, Bodo Moeller] 2595 2596 *) Fix ASN.1 parsing of certain invalid structures that can result 2597 in a denial of service. (CVE-2006-2937) [Steve Henson] 2598 2599 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2600 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2601 2602 *) Fix SSL client code which could crash if connecting to a 2603 malicious SSLv2 server. (CVE-2006-4343) 2604 [Tavis Ormandy and Will Drewry, Google Security Team] 2605 2606 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2607 match only those. Before that, "AES256-SHA" would be interpreted 2608 as a pattern and match "AES128-SHA" too (since AES128-SHA got 2609 the same strength classification in 0.9.7h) as we currently only 2610 have a single AES bit in the ciphersuite description bitmap. 2611 That change, however, also applied to ciphersuite strings such as 2612 "RC4-MD5" that intentionally matched multiple ciphersuites -- 2613 namely, SSL 2.0 ciphersuites in addition to the more common ones 2614 from SSL 3.0/TLS 1.0. 2615 2616 So we change the selection algorithm again: Naming an explicit 2617 ciphersuite selects this one ciphersuite, and any other similar 2618 ciphersuite (same bitmap) from *other* protocol versions. 2619 Thus, "RC4-MD5" again will properly select both the SSL 2.0 2620 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2621 2622 Since SSL 2.0 does not have any ciphersuites for which the 2623 128/256 bit distinction would be relevant, this works for now. 2624 The proper fix will be to use different bits for AES128 and 2625 AES256, which would have avoided the problems from the beginning; 2626 however, bits are scarce, so we can only do this in a new release 2627 (not just a patchlevel) when we can change the SSL_CIPHER 2628 definition to split the single 'unsigned long mask' bitmap into 2629 multiple values to extend the available space. 2630 2631 [Bodo Moeller] 2632 2633 Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2634 2635 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2636 (CVE-2006-4339) [Ben Laurie and Google Security Team] 2637 2638 *) Add AES IGE and biIGE modes. 2639 [Ben Laurie] 2640 2641 *) Change the Unix randomness entropy gathering to use poll() when 2642 possible instead of select(), since the latter has some 2643 undesirable limitations. 2644 [Darryl Miles via Richard Levitte and Bodo Moeller] 2645 2646 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2647 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2648 cannot be implicitly activated as part of, e.g., the "AES" alias. 2649 However, please upgrade to OpenSSL 0.9.9[-dev] for 2650 non-experimental use of the ECC ciphersuites to get TLS extension 2651 support, which is required for curve and point format negotiation 2652 to avoid potential handshake problems. 2653 [Bodo Moeller] 2654 2655 *) Disable rogue ciphersuites: 2656 2657 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2658 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2659 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2660 2661 The latter two were purportedly from 2662 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2663 appear there. 2664 2665 Also deactivate the remaining ciphersuites from 2666 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2667 unofficial, and the ID has long expired. 2668 [Bodo Moeller] 2669 2670 *) Fix RSA blinding Heisenbug (problems sometimes occured on 2671 dual-core machines) and other potential thread-safety issues. 2672 [Bodo Moeller] 2673 2674 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2675 versions), which is now available for royalty-free use 2676 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2677 Also, add Camellia TLS ciphersuites from RFC 4132. 2678 2679 To minimize changes between patchlevels in the OpenSSL 0.9.8 2680 series, Camellia remains excluded from compilation unless OpenSSL 2681 is configured with 'enable-camellia'. 2682 [NTT] 2683 2684 *) Disable the padding bug check when compression is in use. The padding 2685 bug check assumes the first packet is of even length, this is not 2686 necessarily true if compresssion is enabled and can result in false 2687 positives causing handshake failure. The actual bug test is ancient 2688 code so it is hoped that implementations will either have fixed it by 2689 now or any which still have the bug do not support compression. 2690 [Steve Henson] 2691 2692 Changes between 0.9.8a and 0.9.8b [04 May 2006] 2693 2694 *) When applying a cipher rule check to see if string match is an explicit 2695 cipher suite and only match that one cipher suite if it is. 2696 [Steve Henson] 2697 2698 *) Link in manifests for VC++ if needed. 2699 [Austin Ziegler <halostatue@gmail.com>] 2700 2701 *) Update support for ECC-based TLS ciphersuites according to 2702 draft-ietf-tls-ecc-12.txt with proposed changes (but without 2703 TLS extensions, which are supported starting with the 0.9.9 2704 branch, not in the OpenSSL 0.9.8 branch). 2705 [Douglas Stebila] 2706 2707 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2708 opaque EVP_CIPHER_CTX handling. 2709 [Steve Henson] 2710 2711 *) Fixes and enhancements to zlib compression code. We now only use 2712 "zlib1.dll" and use the default __cdecl calling convention on Win32 2713 to conform with the standards mentioned here: 2714 http://www.zlib.net/DLL_FAQ.txt 2715 Static zlib linking now works on Windows and the new --with-zlib-include 2716 --with-zlib-lib options to Configure can be used to supply the location 2717 of the headers and library. Gracefully handle case where zlib library 2718 can't be loaded. 2719 [Steve Henson] 2720 2721 *) Several fixes and enhancements to the OID generation code. The old code 2722 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2723 handle numbers larger than ULONG_MAX, truncated printing and had a 2724 non standard OBJ_obj2txt() behaviour. 2725 [Steve Henson] 2726 2727 *) Add support for building of engines under engine/ as shared libraries 2728 under VC++ build system. 2729 [Steve Henson] 2730 2731 *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2732 Hopefully, we will not see any false combination of paths any more. 2733 [Richard Levitte] 2734 2735 Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2736 2737 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2738 (part of SSL_OP_ALL). This option used to disable the 2739 countermeasure against man-in-the-middle protocol-version 2740 rollback in the SSL 2.0 server implementation, which is a bad 2741 idea. (CVE-2005-2969) 2742 2743 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2744 for Information Security, National Institute of Advanced Industrial 2745 Science and Technology [AIST], Japan)] 2746 2747 *) Add two function to clear and return the verify parameter flags. 2748 [Steve Henson] 2749 2750 *) Keep cipherlists sorted in the source instead of sorting them at 2751 runtime, thus removing the need for a lock. 2752 [Nils Larsch] 2753 2754 *) Avoid some small subgroup attacks in Diffie-Hellman. 2755 [Nick Mathewson and Ben Laurie] 2756 2757 *) Add functions for well-known primes. 2758 [Nick Mathewson] 2759 2760 *) Extended Windows CE support. 2761 [Satoshi Nakamura and Andy Polyakov] 2762 2763 *) Initialize SSL_METHOD structures at compile time instead of during 2764 runtime, thus removing the need for a lock. 2765 [Steve Henson] 2766 2767 *) Make PKCS7_decrypt() work even if no certificate is supplied by 2768 attempting to decrypt each encrypted key in turn. Add support to 2769 smime utility. 2770 [Steve Henson] 2771 2772 Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2773 2774 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2775 OpenSSL 0.9.8.] 2776 2777 *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2778 [Richard Levitte] 2779 2780 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2781 key into the same file any more. 2782 [Richard Levitte] 2783 2784 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2785 [Andy Polyakov] 2786 2787 *) Add -utf8 command line and config file option to 'ca'. 2788 [Stefan <stf@udoma.org] 2789 2790 *) Removed the macro des_crypt(), as it seems to conflict with some 2791 libraries. Use DES_crypt(). 2792 [Richard Levitte] 2793 2794 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2795 involves renaming the source and generated shared-libs for 2796 both. The engines will accept the corrected or legacy ids 2797 ('ncipher' and '4758_cca' respectively) when binding. NB, 2798 this only applies when building 'shared'. 2799 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2800 2801 *) Add attribute functions to EVP_PKEY structure. Modify 2802 PKCS12_create() to recognize a CSP name attribute and 2803 use it. Make -CSP option work again in pkcs12 utility. 2804 [Steve Henson] 2805 2806 *) Add new functionality to the bn blinding code: 2807 - automatic re-creation of the BN_BLINDING parameters after 2808 a fixed number of uses (currently 32) 2809 - add new function for parameter creation 2810 - introduce flags to control the update behaviour of the 2811 BN_BLINDING parameters 2812 - hide BN_BLINDING structure 2813 Add a second BN_BLINDING slot to the RSA structure to improve 2814 performance when a single RSA object is shared among several 2815 threads. 2816 [Nils Larsch] 2817 2818 *) Add support for DTLS. 2819 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2820 2821 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2822 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2823 [Walter Goulet] 2824 2825 *) Remove buggy and incompletet DH cert support from 2826 ssl/ssl_rsa.c and ssl/s3_both.c 2827 [Nils Larsch] 2828 2829 *) Use SHA-1 instead of MD5 as the default digest algorithm for 2830 the apps/openssl applications. 2831 [Nils Larsch] 2832 2833 *) Compile clean with "-Wall -Wmissing-prototypes 2834 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2835 DEBUG_SAFESTACK must also be set. 2836 [Ben Laurie] 2837 2838 *) Change ./Configure so that certain algorithms can be disabled by default. 2839 The new counterpiece to "no-xxx" is "enable-xxx". 2840 2841 The patented RC5 and MDC2 algorithms will now be disabled unless 2842 "enable-rc5" and "enable-mdc2", respectively, are specified. 2843 2844 (IDEA remains enabled despite being patented. This is because IDEA 2845 is frequently required for interoperability, and there is no license 2846 fee for non-commercial use. As before, "no-idea" can be used to 2847 avoid this algorithm.) 2848 2849 [Bodo Moeller] 2850 2851 *) Add processing of proxy certificates (see RFC 3820). This work was 2852 sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2853 EGEE (Enabling Grids for E-science in Europe). 2854 [Richard Levitte] 2855 2856 *) RC4 performance overhaul on modern architectures/implementations, such 2857 as Intel P4, IA-64 and AMD64. 2858 [Andy Polyakov] 2859 2860 *) New utility extract-section.pl. This can be used specify an alternative 2861 section number in a pod file instead of having to treat each file as 2862 a separate case in Makefile. This can be done by adding two lines to the 2863 pod file: 2864 2865 =for comment openssl_section:XXX 2866 2867 The blank line is mandatory. 2868 2869 [Steve Henson] 2870 2871 *) New arguments -certform, -keyform and -pass for s_client and s_server 2872 to allow alternative format key and certificate files and passphrase 2873 sources. 2874 [Steve Henson] 2875 2876 *) New structure X509_VERIFY_PARAM which combines current verify parameters, 2877 update associated structures and add various utility functions. 2878 2879 Add new policy related verify parameters, include policy checking in 2880 standard verify code. Enhance 'smime' application with extra parameters 2881 to support policy checking and print out. 2882 [Steve Henson] 2883 2884 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 2885 Nehemiah processors. These extensions support AES encryption in hardware 2886 as well as RNG (though RNG support is currently disabled). 2887 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 2888 2889 *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 2890 [Geoff Thorpe] 2891 2892 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 2893 [Andy Polyakov and a number of other people] 2894 2895 *) Improved PowerPC platform support. Most notably BIGNUM assembler 2896 implementation contributed by IBM. 2897 [Suresh Chari, Peter Waltenberg, Andy Polyakov] 2898 2899 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 2900 exponent rather than 'unsigned long'. There is a corresponding change to 2901 the new 'rsa_keygen' element of the RSA_METHOD structure. 2902 [Jelte Jansen, Geoff Thorpe] 2903 2904 *) Functionality for creating the initial serial number file is now 2905 moved from CA.pl to the 'ca' utility with a new option -create_serial. 2906 2907 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 2908 number file to 1, which is bound to cause problems. To avoid 2909 the problems while respecting compatibility between different 0.9.7 2910 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 2911 CA.pl for serial number initialization. With the new release 0.9.8, 2912 we can fix the problem directly in the 'ca' utility.) 2913 [Steve Henson] 2914 2915 *) Reduced header interdepencies by declaring more opaque objects in 2916 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 2917 give fewer recursive includes, which could break lazy source code - so 2918 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 2919 developers should define this symbol when building and using openssl to 2920 ensure they track the recommended behaviour, interfaces, [etc], but 2921 backwards-compatible behaviour prevails when this isn't defined. 2922 [Geoff Thorpe] 2923 2924 *) New function X509_POLICY_NODE_print() which prints out policy nodes. 2925 [Steve Henson] 2926 2927 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 2928 This will generate a random key of the appropriate length based on the 2929 cipher context. The EVP_CIPHER can provide its own random key generation 2930 routine to support keys of a specific form. This is used in the des and 2931 3des routines to generate a key of the correct parity. Update S/MIME 2932 code to use new functions and hence generate correct parity DES keys. 2933 Add EVP_CHECK_DES_KEY #define to return an error if the key is not 2934 valid (weak or incorrect parity). 2935 [Steve Henson] 2936 2937 *) Add a local set of CRLs that can be used by X509_verify_cert() as well 2938 as looking them up. This is useful when the verified structure may contain 2939 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 2940 present unless the new PKCS7_NO_CRL flag is asserted. 2941 [Steve Henson] 2942 2943 *) Extend ASN1 oid configuration module. It now additionally accepts the 2944 syntax: 2945 2946 shortName = some long name, 1.2.3.4 2947 [Steve Henson] 2948 2949 *) Reimplemented the BN_CTX implementation. There is now no more static 2950 limitation on the number of variables it can handle nor the depth of the 2951 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 2952 information can now expand as required, and rather than having a single 2953 static array of bignums, BN_CTX now uses a linked-list of such arrays 2954 allowing it to expand on demand whilst maintaining the usefulness of 2955 BN_CTX's "bundling". 2956 [Geoff Thorpe] 2957 2958 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 2959 to allow all RSA operations to function using a single BN_CTX. 2960 [Geoff Thorpe] 2961 2962 *) Preliminary support for certificate policy evaluation and checking. This 2963 is initially intended to pass the tests outlined in "Conformance Testing 2964 of Relying Party Client Certificate Path Processing Logic" v1.07. 2965 [Steve Henson] 2966 2967 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 2968 remained unused and not that useful. A variety of other little bignum 2969 tweaks and fixes have also been made continuing on from the audit (see 2970 below). 2971 [Geoff Thorpe] 2972 2973 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 2974 associated ASN1, EVP and SSL functions and old ASN1 macros. 2975 [Richard Levitte] 2976 2977 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 2978 and this should never fail. So the return value from the use of 2979 BN_set_word() (which can fail due to needless expansion) is now deprecated; 2980 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 2981 [Geoff Thorpe] 2982 2983 *) BN_CTX_get() should return zero-valued bignums, providing the same 2984 initialised value as BN_new(). 2985 [Geoff Thorpe, suggested by Ulf M�ller] 2986 2987 *) Support for inhibitAnyPolicy certificate extension. 2988 [Steve Henson] 2989 2990 *) An audit of the BIGNUM code is underway, for which debugging code is 2991 enabled when BN_DEBUG is defined. This makes stricter enforcements on what 2992 is considered valid when processing BIGNUMs, and causes execution to 2993 assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 2994 further steps are taken to deliberately pollute unused data in BIGNUM 2995 structures to try and expose faulty code further on. For now, openssl will 2996 (in its default mode of operation) continue to tolerate the inconsistent 2997 forms that it has tolerated in the past, but authors and packagers should 2998 consider trying openssl and their own applications when compiled with 2999 these debugging symbols defined. It will help highlight potential bugs in 3000 their own code, and will improve the test coverage for OpenSSL itself. At 3001 some point, these tighter rules will become openssl's default to improve 3002 maintainability, though the assert()s and other overheads will remain only 3003 in debugging configurations. See bn.h for more details. 3004 [Geoff Thorpe, Nils Larsch, Ulf M�ller] 3005 3006 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 3007 that can only be obtained through BN_CTX_new() (which implicitly 3008 initialises it). The presence of this function only made it possible 3009 to overwrite an existing structure (and cause memory leaks). 3010 [Geoff Thorpe] 3011 3012 *) Because of the callback-based approach for implementing LHASH as a 3013 template type, lh_insert() adds opaque objects to hash-tables and 3014 lh_doall() or lh_doall_arg() are typically used with a destructor callback 3015 to clean up those corresponding objects before destroying the hash table 3016 (and losing the object pointers). So some over-zealous constifications in 3017 LHASH have been relaxed so that lh_insert() does not take (nor store) the 3018 objects as "const" and the lh_doall[_arg] callback wrappers are not 3019 prototyped to have "const" restrictions on the object pointers they are 3020 given (and so aren't required to cast them away any more). 3021 [Geoff Thorpe] 3022 3023 *) The tmdiff.h API was so ugly and minimal that our own timing utility 3024 (speed) prefers to use its own implementation. The two implementations 3025 haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3026 its object type properly exposed (MS_TM) instead of casting to/from "char 3027 *". This may still change yet if someone realises MS_TM and "ms_time_***" 3028 aren't necessarily the greatest nomenclatures - but this is what was used 3029 internally to the implementation so I've used that for now. 3030 [Geoff Thorpe] 3031 3032 *) Ensure that deprecated functions do not get compiled when 3033 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3034 the self-tests were still using deprecated key-generation functions so 3035 these have been updated also. 3036 [Geoff Thorpe] 3037 3038 *) Reorganise PKCS#7 code to separate the digest location functionality 3039 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 3040 New function PKCS7_set_digest() to set the digest type for PKCS#7 3041 digestedData type. Add additional code to correctly generate the 3042 digestedData type and add support for this type in PKCS7 initialization 3043 functions. 3044 [Steve Henson] 3045 3046 *) New function PKCS7_set0_type_other() this initializes a PKCS7 3047 structure of type "other". 3048 [Steve Henson] 3049 3050 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 3051 sure the loop does correctly stop and breaking ("division by zero") 3052 modulus operations are not performed. The (pre-generated) prime 3053 table crypto/bn/bn_prime.h was already correct, but it could not be 3054 re-generated on some platforms because of the "division by zero" 3055 situation in the script. 3056 [Ralf S. Engelschall] 3057 3058 *) Update support for ECC-based TLS ciphersuites according to 3059 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 3060 SHA-1 now is only used for "small" curves (where the 3061 representation of a field element takes up to 24 bytes); for 3062 larger curves, the field element resulting from ECDH is directly 3063 used as premaster secret. 3064 [Douglas Stebila (Sun Microsystems Laboratories)] 3065 3066 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3067 curve secp160r1 to the tests. 3068 [Douglas Stebila (Sun Microsystems Laboratories)] 3069 3070 *) Add the possibility to load symbols globally with DSO. 3071 [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3072 3073 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3074 control of the error stack. 3075 [Richard Levitte] 3076 3077 *) Add support for STORE in ENGINE. 3078 [Richard Levitte] 3079 3080 *) Add the STORE type. The intention is to provide a common interface 3081 to certificate and key stores, be they simple file-based stores, or 3082 HSM-type store, or LDAP stores, or... 3083 NOTE: The code is currently UNTESTED and isn't really used anywhere. 3084 [Richard Levitte] 3085 3086 *) Add a generic structure called OPENSSL_ITEM. This can be used to 3087 pass a list of arguments to any function as well as provide a way 3088 for a function to pass data back to the caller. 3089 [Richard Levitte] 3090 3091 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3092 works like BUF_strdup() but can be used to duplicate a portion of 3093 a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3094 a memory area. 3095 [Richard Levitte] 3096 3097 *) Add the function sk_find_ex() which works like sk_find(), but will 3098 return an index to an element even if an exact match couldn't be 3099 found. The index is guaranteed to point at the element where the 3100 searched-for key would be inserted to preserve sorting order. 3101 [Richard Levitte] 3102 3103 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3104 takes an extra flags argument for optional functionality. Currently, 3105 the following flags are defined: 3106 3107 OBJ_BSEARCH_VALUE_ON_NOMATCH 3108 This one gets OBJ_bsearch_ex() to return a pointer to the first 3109 element where the comparing function returns a negative or zero 3110 number. 3111 3112 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3113 This one gets OBJ_bsearch_ex() to return a pointer to the first 3114 element where the comparing function returns zero. This is useful 3115 if there are more than one element where the comparing function 3116 returns zero. 3117 [Richard Levitte] 3118 3119 *) Make it possible to create self-signed certificates with 'openssl ca' 3120 in such a way that the self-signed certificate becomes part of the 3121 CA database and uses the same mechanisms for serial number generation 3122 as all other certificate signing. The new flag '-selfsign' enables 3123 this functionality. Adapt CA.sh and CA.pl.in. 3124 [Richard Levitte] 3125 3126 *) Add functionality to check the public key of a certificate request 3127 against a given private. This is useful to check that a certificate 3128 request can be signed by that key (self-signing). 3129 [Richard Levitte] 3130 3131 *) Make it possible to have multiple active certificates with the same 3132 subject in the CA index file. This is done only if the keyword 3133 'unique_subject' is set to 'no' in the main CA section (default 3134 if 'CA_default') of the configuration file. The value is saved 3135 with the database itself in a separate index attribute file, 3136 named like the index file with '.attr' appended to the name. 3137 [Richard Levitte] 3138 3139 *) Generate muti valued AVAs using '+' notation in config files for 3140 req and dirName. 3141 [Steve Henson] 3142 3143 *) Support for nameConstraints certificate extension. 3144 [Steve Henson] 3145 3146 *) Support for policyConstraints certificate extension. 3147 [Steve Henson] 3148 3149 *) Support for policyMappings certificate extension. 3150 [Steve Henson] 3151 3152 *) Make sure the default DSA_METHOD implementation only uses its 3153 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3154 and change its own handlers to be NULL so as to remove unnecessary 3155 indirection. This lets alternative implementations fallback to the 3156 default implementation more easily. 3157 [Geoff Thorpe] 3158 3159 *) Support for directoryName in GeneralName related extensions 3160 in config files. 3161 [Steve Henson] 3162 3163 *) Make it possible to link applications using Makefile.shared. 3164 Make that possible even when linking against static libraries! 3165 [Richard Levitte] 3166 3167 *) Support for single pass processing for S/MIME signing. This now 3168 means that S/MIME signing can be done from a pipe, in addition 3169 cleartext signing (multipart/signed type) is effectively streaming 3170 and the signed data does not need to be all held in memory. 3171 3172 This is done with a new flag PKCS7_STREAM. When this flag is set 3173 PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3174 is done after the data is output (and digests calculated) in 3175 SMIME_write_PKCS7(). 3176 [Steve Henson] 3177 3178 *) Add full support for -rpath/-R, both in shared libraries and 3179 applications, at least on the platforms where it's known how 3180 to do it. 3181 [Richard Levitte] 3182 3183 *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3184 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3185 will now compute a table of multiples of the generator that 3186 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3187 faster (notably in the case of a single point multiplication, 3188 scalar * generator). 3189 [Nils Larsch, Bodo Moeller] 3190 3191 *) IPv6 support for certificate extensions. The various extensions 3192 which use the IP:a.b.c.d can now take IPv6 addresses using the 3193 formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3194 correctly. 3195 [Steve Henson] 3196 3197 *) Added an ENGINE that implements RSA by performing private key 3198 exponentiations with the GMP library. The conversions to and from 3199 GMP's mpz_t format aren't optimised nor are any montgomery forms 3200 cached, and on x86 it appears OpenSSL's own performance has caught up. 3201 However there are likely to be other architectures where GMP could 3202 provide a boost. This ENGINE is not built in by default, but it can be 3203 specified at Configure time and should be accompanied by the necessary 3204 linker additions, eg; 3205 ./config -DOPENSSL_USE_GMP -lgmp 3206 [Geoff Thorpe] 3207 3208 *) "openssl engine" will not display ENGINE/DSO load failure errors when 3209 testing availability of engines with "-t" - the old behaviour is 3210 produced by increasing the feature's verbosity with "-tt". 3211 [Geoff Thorpe] 3212 3213 *) ECDSA routines: under certain error conditions uninitialized BN objects 3214 could be freed. Solution: make sure initialization is performed early 3215 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3216 via PR#459) 3217 [Lutz Jaenicke] 3218 3219 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3220 and DH_METHOD (eg. by ENGINE implementations) to override the normal 3221 software implementations. For DSA and DH, parameter generation can 3222 also be overriden by providing the appropriate method callbacks. 3223 [Geoff Thorpe] 3224 3225 *) Change the "progress" mechanism used in key-generation and 3226 primality testing to functions that take a new BN_GENCB pointer in 3227 place of callback/argument pairs. The new API functions have "_ex" 3228 postfixes and the older functions are reimplemented as wrappers for 3229 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3230 declarations of the old functions to help (graceful) attempts to 3231 migrate to the new functions. Also, the new key-generation API 3232 functions operate on a caller-supplied key-structure and return 3233 success/failure rather than returning a key or NULL - this is to 3234 help make "keygen" another member function of RSA_METHOD etc. 3235 3236 Example for using the new callback interface: 3237 3238 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3239 void *my_arg = ...; 3240 BN_GENCB my_cb; 3241 3242 BN_GENCB_set(&my_cb, my_callback, my_arg); 3243 3244 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3245 /* For the meaning of a, b in calls to my_callback(), see the 3246 * documentation of the function that calls the callback. 3247 * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3248 * my_callback should return 1 if it wants BN_is_prime_ex() 3249 * to continue, or 0 to stop. 3250 */ 3251 3252 [Geoff Thorpe] 3253 3254 *) Change the ZLIB compression method to be stateful, and make it 3255 available to TLS with the number defined in 3256 draft-ietf-tls-compression-04.txt. 3257 [Richard Levitte] 3258 3259 *) Add the ASN.1 structures and functions for CertificatePair, which 3260 is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3261 3262 CertificatePair ::= SEQUENCE { 3263 forward [0] Certificate OPTIONAL, 3264 reverse [1] Certificate OPTIONAL, 3265 -- at least one of the pair shall be present -- } 3266 3267 Also implement the PEM functions to read and write certificate 3268 pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3269 3270 This needed to be defined, mostly for the sake of the LDAP 3271 attribute crossCertificatePair, but may prove useful elsewhere as 3272 well. 3273 [Richard Levitte] 3274 3275 *) Make it possible to inhibit symlinking of shared libraries in 3276 Makefile.shared, for Cygwin's sake. 3277 [Richard Levitte] 3278 3279 *) Extend the BIGNUM API by creating a function 3280 void BN_set_negative(BIGNUM *a, int neg); 3281 and a macro that behave like 3282 int BN_is_negative(const BIGNUM *a); 3283 3284 to avoid the need to access 'a->neg' directly in applications. 3285 [Nils Larsch] 3286 3287 *) Implement fast modular reduction for pseudo-Mersenne primes 3288 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3289 EC_GROUP_new_curve_GFp() will now automatically use this 3290 if applicable. 3291 [Nils Larsch <nla@trustcenter.de>] 3292 3293 *) Add new lock type (CRYPTO_LOCK_BN). 3294 [Bodo Moeller] 3295 3296 *) Change the ENGINE framework to automatically load engines 3297 dynamically from specific directories unless they could be 3298 found to already be built in or loaded. Move all the 3299 current engines except for the cryptodev one to a new 3300 directory engines/. 3301 The engines in engines/ are built as shared libraries if 3302 the "shared" options was given to ./Configure or ./config. 3303 Otherwise, they are inserted in libcrypto.a. 3304 /usr/local/ssl/engines is the default directory for dynamic 3305 engines, but that can be overriden at configure time through 3306 the usual use of --prefix and/or --openssldir, and at run 3307 time with the environment variable OPENSSL_ENGINES. 3308 [Geoff Thorpe and Richard Levitte] 3309 3310 *) Add Makefile.shared, a helper makefile to build shared 3311 libraries. Addapt Makefile.org. 3312 [Richard Levitte] 3313 3314 *) Add version info to Win32 DLLs. 3315 [Peter 'Luna' Runestig" <peter@runestig.com>] 3316 3317 *) Add new 'medium level' PKCS#12 API. Certificates and keys 3318 can be added using this API to created arbitrary PKCS#12 3319 files while avoiding the low level API. 3320 3321 New options to PKCS12_create(), key or cert can be NULL and 3322 will then be omitted from the output file. The encryption 3323 algorithm NIDs can be set to -1 for no encryption, the mac 3324 iteration count can be set to 0 to omit the mac. 3325 3326 Enhance pkcs12 utility by making the -nokeys and -nocerts 3327 options work when creating a PKCS#12 file. New option -nomac 3328 to omit the mac, NONE can be set for an encryption algorithm. 3329 New code is modified to use the enhanced PKCS12_create() 3330 instead of the low level API. 3331 [Steve Henson] 3332 3333 *) Extend ASN1 encoder to support indefinite length constructed 3334 encoding. This can output sequences tags and octet strings in 3335 this form. Modify pk7_asn1.c to support indefinite length 3336 encoding. This is experimental and needs additional code to 3337 be useful, such as an ASN1 bio and some enhanced streaming 3338 PKCS#7 code. 3339 3340 Extend template encode functionality so that tagging is passed 3341 down to the template encoder. 3342 [Steve Henson] 3343 3344 *) Let 'openssl req' fail if an argument to '-newkey' is not 3345 recognized instead of using RSA as a default. 3346 [Bodo Moeller] 3347 3348 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3349 As these are not official, they are not included in "ALL"; 3350 the "ECCdraft" ciphersuite group alias can be used to select them. 3351 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3352 3353 *) Add ECDH engine support. 3354 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3355 3356 *) Add ECDH in new directory crypto/ecdh/. 3357 [Douglas Stebila (Sun Microsystems Laboratories)] 3358 3359 *) Let BN_rand_range() abort with an error after 100 iterations 3360 without success (which indicates a broken PRNG). 3361 [Bodo Moeller] 3362 3363 *) Change BN_mod_sqrt() so that it verifies that the input value 3364 is really the square of the return value. (Previously, 3365 BN_mod_sqrt would show GIGO behaviour.) 3366 [Bodo Moeller] 3367 3368 *) Add named elliptic curves over binary fields from X9.62, SECG, 3369 and WAP/WTLS; add OIDs that were still missing. 3370 3371 [Sheueling Chang Shantz and Douglas Stebila 3372 (Sun Microsystems Laboratories)] 3373 3374 *) Extend the EC library for elliptic curves over binary fields 3375 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3376 New EC_METHOD: 3377 3378 EC_GF2m_simple_method 3379 3380 New API functions: 3381 3382 EC_GROUP_new_curve_GF2m 3383 EC_GROUP_set_curve_GF2m 3384 EC_GROUP_get_curve_GF2m 3385 EC_POINT_set_affine_coordinates_GF2m 3386 EC_POINT_get_affine_coordinates_GF2m 3387 EC_POINT_set_compressed_coordinates_GF2m 3388 3389 Point compression for binary fields is disabled by default for 3390 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3391 enable it). 3392 3393 As binary polynomials are represented as BIGNUMs, various members 3394 of the EC_GROUP and EC_POINT data structures can be shared 3395 between the implementations for prime fields and binary fields; 3396 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3397 are essentially identical to their ..._GFp counterparts. 3398 (For simplicity, the '..._GFp' prefix has been dropped from 3399 various internal method names.) 3400 3401 An internal 'field_div' method (similar to 'field_mul' and 3402 'field_sqr') has been added; this is used only for binary fields. 3403 3404 [Sheueling Chang Shantz and Douglas Stebila 3405 (Sun Microsystems Laboratories)] 3406 3407 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3408 through methods ('mul', 'precompute_mult'). 3409 3410 The generic implementations (now internally called 'ec_wNAF_mul' 3411 and 'ec_wNAF_precomputed_mult') remain the default if these 3412 methods are undefined. 3413 3414 [Sheueling Chang Shantz and Douglas Stebila 3415 (Sun Microsystems Laboratories)] 3416 3417 *) New function EC_GROUP_get_degree, which is defined through 3418 EC_METHOD. For curves over prime fields, this returns the bit 3419 length of the modulus. 3420 3421 [Sheueling Chang Shantz and Douglas Stebila 3422 (Sun Microsystems Laboratories)] 3423 3424 *) New functions EC_GROUP_dup, EC_POINT_dup. 3425 (These simply call ..._new and ..._copy). 3426 3427 [Sheueling Chang Shantz and Douglas Stebila 3428 (Sun Microsystems Laboratories)] 3429 3430 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3431 Polynomials are represented as BIGNUMs (where the sign bit is not 3432 used) in the following functions [macros]: 3433 3434 BN_GF2m_add 3435 BN_GF2m_sub [= BN_GF2m_add] 3436 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3437 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3438 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3439 BN_GF2m_mod_inv 3440 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3441 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3442 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3443 BN_GF2m_cmp [= BN_ucmp] 3444 3445 (Note that only the 'mod' functions are actually for fields GF(2^m). 3446 BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3447 3448 For some functions, an the irreducible polynomial defining a 3449 field can be given as an 'unsigned int[]' with strictly 3450 decreasing elements giving the indices of those bits that are set; 3451 i.e., p[] represents the polynomial 3452 f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3453 where 3454 p[0] > p[1] > ... > p[k] = 0. 3455 This applies to the following functions: 3456 3457 BN_GF2m_mod_arr 3458 BN_GF2m_mod_mul_arr 3459 BN_GF2m_mod_sqr_arr 3460 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3461 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3462 BN_GF2m_mod_exp_arr 3463 BN_GF2m_mod_sqrt_arr 3464 BN_GF2m_mod_solve_quad_arr 3465 BN_GF2m_poly2arr 3466 BN_GF2m_arr2poly 3467 3468 Conversion can be performed by the following functions: 3469 3470 BN_GF2m_poly2arr 3471 BN_GF2m_arr2poly 3472 3473 bntest.c has additional tests for binary polynomial arithmetic. 3474 3475 Two implementations for BN_GF2m_mod_div() are available. 3476 The default algorithm simply uses BN_GF2m_mod_inv() and 3477 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3478 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3479 copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3480 3481 [Sheueling Chang Shantz and Douglas Stebila 3482 (Sun Microsystems Laboratories)] 3483 3484 *) Add new error code 'ERR_R_DISABLED' that can be used when some 3485 functionality is disabled at compile-time. 3486 [Douglas Stebila <douglas.stebila@sun.com>] 3487 3488 *) Change default behaviour of 'openssl asn1parse' so that more 3489 information is visible when viewing, e.g., a certificate: 3490 3491 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3492 mode the content of non-printable OCTET STRINGs is output in a 3493 style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3494 avoid the appearance of a printable string. 3495 [Nils Larsch <nla@trustcenter.de>] 3496 3497 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3498 functions 3499 EC_GROUP_set_asn1_flag() 3500 EC_GROUP_get_asn1_flag() 3501 EC_GROUP_set_point_conversion_form() 3502 EC_GROUP_get_point_conversion_form() 3503 These control ASN1 encoding details: 3504 - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3505 has been set to OPENSSL_EC_NAMED_CURVE. 3506 - Points are encoded in uncompressed form by default; options for 3507 asn1_for are as for point2oct, namely 3508 POINT_CONVERSION_COMPRESSED 3509 POINT_CONVERSION_UNCOMPRESSED 3510 POINT_CONVERSION_HYBRID 3511 3512 Also add 'seed' and 'seed_len' members to EC_GROUP with access 3513 functions 3514 EC_GROUP_set_seed() 3515 EC_GROUP_get0_seed() 3516 EC_GROUP_get_seed_len() 3517 This is used only for ASN1 purposes (so far). 3518 [Nils Larsch <nla@trustcenter.de>] 3519 3520 *) Add 'field_type' member to EC_METHOD, which holds the NID 3521 of the appropriate field type OID. The new function 3522 EC_METHOD_get_field_type() returns this value. 3523 [Nils Larsch <nla@trustcenter.de>] 3524 3525 *) Add functions 3526 EC_POINT_point2bn() 3527 EC_POINT_bn2point() 3528 EC_POINT_point2hex() 3529 EC_POINT_hex2point() 3530 providing useful interfaces to EC_POINT_point2oct() and 3531 EC_POINT_oct2point(). 3532 [Nils Larsch <nla@trustcenter.de>] 3533 3534 *) Change internals of the EC library so that the functions 3535 EC_GROUP_set_generator() 3536 EC_GROUP_get_generator() 3537 EC_GROUP_get_order() 3538 EC_GROUP_get_cofactor() 3539 are implemented directly in crypto/ec/ec_lib.c and not dispatched 3540 to methods, which would lead to unnecessary code duplication when 3541 adding different types of curves. 3542 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3543 3544 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3545 arithmetic, and such that modified wNAFs are generated 3546 (which avoid length expansion in many cases). 3547 [Bodo Moeller] 3548 3549 *) Add a function EC_GROUP_check_discriminant() (defined via 3550 EC_METHOD) that verifies that the curve discriminant is non-zero. 3551 3552 Add a function EC_GROUP_check() that makes some sanity tests 3553 on a EC_GROUP, its generator and order. This includes 3554 EC_GROUP_check_discriminant(). 3555 [Nils Larsch <nla@trustcenter.de>] 3556 3557 *) Add ECDSA in new directory crypto/ecdsa/. 3558 3559 Add applications 'openssl ecparam' and 'openssl ecdsa' 3560 (these are based on 'openssl dsaparam' and 'openssl dsa'). 3561 3562 ECDSA support is also included in various other files across the 3563 library. Most notably, 3564 - 'openssl req' now has a '-newkey ecdsa:file' option; 3565 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3566 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3567 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3568 them suitable for ECDSA where domain parameters must be 3569 extracted before the specific public key; 3570 - ECDSA engine support has been added. 3571 [Nils Larsch <nla@trustcenter.de>] 3572 3573 *) Include some named elliptic curves, and add OIDs from X9.62, 3574 SECG, and WAP/WTLS. Each curve can be obtained from the new 3575 function 3576 EC_GROUP_new_by_curve_name(), 3577 and the list of available named curves can be obtained with 3578 EC_get_builtin_curves(). 3579 Also add a 'curve_name' member to EC_GROUP objects, which can be 3580 accessed via 3581 EC_GROUP_set_curve_name() 3582 EC_GROUP_get_curve_name() 3583 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3584 3585 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3586 was actually never needed) and in BN_mul(). The removal in BN_mul() 3587 required a small change in bn_mul_part_recursive() and the addition 3588 of the functions bn_cmp_part_words(), bn_sub_part_words() and 3589 bn_add_part_words(), which do the same thing as bn_cmp_words(), 3590 bn_sub_words() and bn_add_words() except they take arrays with 3591 differing sizes. 3592 [Richard Levitte] 3593 3594 Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3595 3596 *) Cleanse PEM buffers before freeing them since they may contain 3597 sensitive data. 3598 [Benjamin Bennett <ben@psc.edu>] 3599 3600 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3601 a ciphersuite string such as "DEFAULT:RSA" cannot enable 3602 authentication-only ciphersuites. 3603 [Bodo Moeller] 3604 3605 *) Since AES128 and AES256 share a single mask bit in the logic of 3606 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3607 kludge to work properly if AES128 is available and AES256 isn't. 3608 [Victor Duchovni] 3609 3610 *) Expand security boundary to match 1.1.1 module. 3611 [Steve Henson] 3612 3613 *) Remove redundant features: hash file source, editing of test vectors 3614 modify fipsld to use external fips_premain.c signature. 3615 [Steve Henson] 3616 3617 *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3618 run algorithm test programs. 3619 [Steve Henson] 3620 3621 *) Make algorithm test programs more tolerant of whitespace. 3622 [Steve Henson] 3623 3624 *) Have SSL/TLS server implementation tolerate "mismatched" record 3625 protocol version while receiving ClientHello even if the 3626 ClientHello is fragmented. (The server can't insist on the 3627 particular protocol version it has chosen before the ServerHello 3628 message has informed the client about his choice.) 3629 [Bodo Moeller] 3630 3631 *) Load error codes if they are not already present instead of using a 3632 static variable. This allows them to be cleanly unloaded and reloaded. 3633 [Steve Henson] 3634 3635 Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3636 3637 *) Introduce limits to prevent malicious keys being able to 3638 cause a denial of service. (CVE-2006-2940) 3639 [Steve Henson, Bodo Moeller] 3640 3641 *) Fix ASN.1 parsing of certain invalid structures that can result 3642 in a denial of service. (CVE-2006-2937) [Steve Henson] 3643 3644 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3645 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3646 3647 *) Fix SSL client code which could crash if connecting to a 3648 malicious SSLv2 server. (CVE-2006-4343) 3649 [Tavis Ormandy and Will Drewry, Google Security Team] 3650 3651 *) Change ciphersuite string processing so that an explicit 3652 ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3653 will no longer include "AES128-SHA"), and any other similar 3654 ciphersuite (same bitmap) from *other* protocol versions (so that 3655 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3656 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3657 changes from 0.9.8b and 0.9.8d. 3658 [Bodo Moeller] 3659 3660 Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3661 3662 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3663 (CVE-2006-4339) [Ben Laurie and Google Security Team] 3664 3665 *) Change the Unix randomness entropy gathering to use poll() when 3666 possible instead of select(), since the latter has some 3667 undesirable limitations. 3668 [Darryl Miles via Richard Levitte and Bodo Moeller] 3669 3670 *) Disable rogue ciphersuites: 3671 3672 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3673 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3674 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3675 3676 The latter two were purportedly from 3677 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3678 appear there. 3679 3680 Also deactive the remaining ciphersuites from 3681 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3682 unofficial, and the ID has long expired. 3683 [Bodo Moeller] 3684 3685 *) Fix RSA blinding Heisenbug (problems sometimes occured on 3686 dual-core machines) and other potential thread-safety issues. 3687 [Bodo Moeller] 3688 3689 Changes between 0.9.7i and 0.9.7j [04 May 2006] 3690 3691 *) Adapt fipsld and the build system to link against the validated FIPS 3692 module in FIPS mode. 3693 [Steve Henson] 3694 3695 *) Fixes for VC++ 2005 build under Windows. 3696 [Steve Henson] 3697 3698 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3699 from a Windows bash shell such as MSYS. It is autodetected from the 3700 "config" script when run from a VC++ environment. Modify standard VC++ 3701 build to use fipscanister.o from the GNU make build. 3702 [Steve Henson] 3703 3704 Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3705 3706 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3707 The value now differs depending on if you build for FIPS or not. 3708 BEWARE! A program linked with a shared FIPSed libcrypto can't be 3709 safely run with a non-FIPSed libcrypto, as it may crash because of 3710 the difference induced by this change. 3711 [Andy Polyakov] 3712 3713 Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3714 3715 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3716 (part of SSL_OP_ALL). This option used to disable the 3717 countermeasure against man-in-the-middle protocol-version 3718 rollback in the SSL 2.0 server implementation, which is a bad 3719 idea. (CVE-2005-2969) 3720 3721 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3722 for Information Security, National Institute of Advanced Industrial 3723 Science and Technology [AIST], Japan)] 3724 3725 *) Minimal support for X9.31 signatures and PSS padding modes. This is 3726 mainly for FIPS compliance and not fully integrated at this stage. 3727 [Steve Henson] 3728 3729 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3730 the exponentiation using a fixed-length exponent. (Otherwise, 3731 the information leaked through timing could expose the secret key 3732 after many signatures; cf. Bleichenbacher's attack on DSA with 3733 biased k.) 3734 [Bodo Moeller] 3735 3736 *) Make a new fixed-window mod_exp implementation the default for 3737 RSA, DSA, and DH private-key operations so that the sequence of 3738 squares and multiplies and the memory access pattern are 3739 independent of the particular secret key. This will mitigate 3740 cache-timing and potential related attacks. 3741 3742 BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3743 and this is automatically used by BN_mod_exp_mont() if the new flag 3744 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3745 will use this BN flag for private exponents unless the flag 3746 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3747 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3748 3749 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3750 3751 *) Change the client implementation for SSLv23_method() and 3752 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3753 Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3754 (Previously, the SSL 2.0 backwards compatible Client Hello 3755 message format would be used even with SSL_OP_NO_SSLv2.) 3756 [Bodo Moeller] 3757 3758 *) Add support for smime-type MIME parameter in S/MIME messages which some 3759 clients need. 3760 [Steve Henson] 3761 3762 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3763 a threadsafe manner. Modify rsa code to use new function and add calls 3764 to dsa and dh code (which had race conditions before). 3765 [Steve Henson] 3766 3767 *) Include the fixed error library code in the C error file definitions 3768 instead of fixing them up at runtime. This keeps the error code 3769 structures constant. 3770 [Steve Henson] 3771 3772 Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3773 3774 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3775 OpenSSL 0.9.8.] 3776 3777 *) Fixes for newer kerberos headers. NB: the casts are needed because 3778 the 'length' field is signed on one version and unsigned on another 3779 with no (?) obvious way to tell the difference, without these VC++ 3780 complains. Also the "definition" of FAR (blank) is no longer included 3781 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3782 some needed definitions. 3783 [Steve Henson] 3784 3785 *) Undo Cygwin change. 3786 [Ulf M�ller] 3787 3788 *) Added support for proxy certificates according to RFC 3820. 3789 Because they may be a security thread to unaware applications, 3790 they must be explicitely allowed in run-time. See 3791 docs/HOWTO/proxy_certificates.txt for further information. 3792 [Richard Levitte] 3793 3794 Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3795 3796 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3797 server and client random values. Previously 3798 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3799 less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3800 3801 This change has negligible security impact because: 3802 3803 1. Server and client random values still have 24 bytes of pseudo random 3804 data. 3805 3806 2. Server and client random values are sent in the clear in the initial 3807 handshake. 3808 3809 3. The master secret is derived using the premaster secret (48 bytes in 3810 size for static RSA ciphersuites) as well as client server and random 3811 values. 3812 3813 The OpenSSL team would like to thank the UK NISCC for bringing this issue 3814 to our attention. 3815 3816 [Stephen Henson, reported by UK NISCC] 3817 3818 *) Use Windows randomness collection on Cygwin. 3819 [Ulf M�ller] 3820 3821 *) Fix hang in EGD/PRNGD query when communication socket is closed 3822 prematurely by EGD/PRNGD. 3823 [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3824 3825 *) Prompt for pass phrases when appropriate for PKCS12 input format. 3826 [Steve Henson] 3827 3828 *) Back-port of selected performance improvements from development 3829 branch, as well as improved support for PowerPC platforms. 3830 [Andy Polyakov] 3831 3832 *) Add lots of checks for memory allocation failure, error codes to indicate 3833 failure and freeing up memory if a failure occurs. 3834 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3835 3836 *) Add new -passin argument to dgst. 3837 [Steve Henson] 3838 3839 *) Perform some character comparisons of different types in X509_NAME_cmp: 3840 this is needed for some certificates that reencode DNs into UTF8Strings 3841 (in violation of RFC3280) and can't or wont issue name rollover 3842 certificates. 3843 [Steve Henson] 3844 3845 *) Make an explicit check during certificate validation to see that 3846 the CA setting in each certificate on the chain is correct. As a 3847 side effect always do the following basic checks on extensions, 3848 not just when there's an associated purpose to the check: 3849 3850 - if there is an unhandled critical extension (unless the user 3851 has chosen to ignore this fault) 3852 - if the path length has been exceeded (if one is set at all) 3853 - that certain extensions fit the associated purpose (if one has 3854 been given) 3855 [Richard Levitte] 3856 3857 Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3858 3859 *) Avoid a race condition when CRLs are checked in a multi threaded 3860 environment. This would happen due to the reordering of the revoked 3861 entries during signature checking and serial number lookup. Now the 3862 encoding is cached and the serial number sort performed under a lock. 3863 Add new STACK function sk_is_sorted(). 3864 [Steve Henson] 3865 3866 *) Add Delta CRL to the extension code. 3867 [Steve Henson] 3868 3869 *) Various fixes to s3_pkt.c so alerts are sent properly. 3870 [David Holmes <d.holmes@f5.com>] 3871 3872 *) Reduce the chances of duplicate issuer name and serial numbers (in 3873 violation of RFC3280) using the OpenSSL certificate creation utilities. 3874 This is done by creating a random 64 bit value for the initial serial 3875 number when a serial number file is created or when a self signed 3876 certificate is created using 'openssl req -x509'. The initial serial 3877 number file is created using 'openssl x509 -next_serial' in CA.pl 3878 rather than being initialized to 1. 3879 [Steve Henson] 3880 3881 Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 3882 3883 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3884 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3885 [Joe Orton, Steve Henson] 3886 3887 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 3888 (CVE-2004-0112) 3889 [Joe Orton, Steve Henson] 3890 3891 *) Make it possible to have multiple active certificates with the same 3892 subject in the CA index file. This is done only if the keyword 3893 'unique_subject' is set to 'no' in the main CA section (default 3894 if 'CA_default') of the configuration file. The value is saved 3895 with the database itself in a separate index attribute file, 3896 named like the index file with '.attr' appended to the name. 3897 [Richard Levitte] 3898 3899 *) X509 verify fixes. Disable broken certificate workarounds when 3900 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 3901 keyUsage extension present. Don't accept CRLs with unhandled critical 3902 extensions: since verify currently doesn't process CRL extensions this 3903 rejects a CRL with *any* critical extensions. Add new verify error codes 3904 for these cases. 3905 [Steve Henson] 3906 3907 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 3908 A clarification of RFC2560 will require the use of OCTET STRINGs and 3909 some implementations cannot handle the current raw format. Since OpenSSL 3910 copies and compares OCSP nonces as opaque blobs without any attempt at 3911 parsing them this should not create any compatibility issues. 3912 [Steve Henson] 3913 3914 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 3915 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 3916 this HMAC (and other) operations are several times slower than OpenSSL 3917 < 0.9.7. 3918 [Steve Henson] 3919 3920 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 3921 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 3922 3923 *) Use the correct content when signing type "other". 3924 [Steve Henson] 3925 3926 Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 3927 3928 *) Fix various bugs revealed by running the NISCC test suite: 3929 3930 Stop out of bounds reads in the ASN1 code when presented with 3931 invalid tags (CVE-2003-0543 and CVE-2003-0544). 3932 3933 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 3934 3935 If verify callback ignores invalid public key errors don't try to check 3936 certificate signature with the NULL public key. 3937 3938 [Steve Henson] 3939 3940 *) New -ignore_err option in ocsp application to stop the server 3941 exiting on the first error in a request. 3942 [Steve Henson] 3943 3944 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3945 if the server requested one: as stated in TLS 1.0 and SSL 3.0 3946 specifications. 3947 [Steve Henson] 3948 3949 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3950 extra data after the compression methods not only for TLS 1.0 3951 but also for SSL 3.0 (as required by the specification). 3952 [Bodo Moeller; problem pointed out by Matthias Loepfe] 3953 3954 *) Change X509_certificate_type() to mark the key as exported/exportable 3955 when it's 512 *bits* long, not 512 bytes. 3956 [Richard Levitte] 3957 3958 *) Change AES_cbc_encrypt() so it outputs exact multiple of 3959 blocks during encryption. 3960 [Richard Levitte] 3961 3962 *) Various fixes to base64 BIO and non blocking I/O. On write 3963 flushes were not handled properly if the BIO retried. On read 3964 data was not being buffered properly and had various logic bugs. 3965 This also affects blocking I/O when the data being decoded is a 3966 certain size. 3967 [Steve Henson] 3968 3969 *) Various S/MIME bugfixes and compatibility changes: 3970 output correct application/pkcs7 MIME type if 3971 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 3972 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 3973 of files as .eml work). Correctly handle very long lines in MIME 3974 parser. 3975 [Steve Henson] 3976 3977 Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 3978 3979 *) Countermeasure against the Klima-Pokorny-Rosa extension of 3980 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3981 a protocol version number mismatch like a decryption error 3982 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3983 [Bodo Moeller] 3984 3985 *) Turn on RSA blinding by default in the default implementation 3986 to avoid a timing attack. Applications that don't want it can call 3987 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3988 They would be ill-advised to do so in most cases. 3989 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3990 3991 *) Change RSA blinding code so that it works when the PRNG is not 3992 seeded (in this case, the secret RSA exponent is abused as 3993 an unpredictable seed -- if it is not unpredictable, there 3994 is no point in blinding anyway). Make RSA blinding thread-safe 3995 by remembering the creator's thread ID in rsa->blinding and 3996 having all other threads use local one-time blinding factors 3997 (this requires more computation than sharing rsa->blinding, but 3998 avoids excessive locking; and if an RSA object is not shared 3999 between threads, blinding will still be very fast). 4000 [Bodo Moeller] 4001 4002 *) Fixed a typo bug that would cause ENGINE_set_default() to set an 4003 ENGINE as defaults for all supported algorithms irrespective of 4004 the 'flags' parameter. 'flags' is now honoured, so applications 4005 should make sure they are passing it correctly. 4006 [Geoff Thorpe] 4007 4008 *) Target "mingw" now allows native Windows code to be generated in 4009 the Cygwin environment as well as with the MinGW compiler. 4010 [Ulf Moeller] 4011 4012 Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4013 4014 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4015 via timing by performing a MAC computation even if incorrrect 4016 block cipher padding has been found. This is a countermeasure 4017 against active attacks where the attacker has to distinguish 4018 between bad padding and a MAC verification error. (CVE-2003-0078) 4019 4020 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4021 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4022 Martin Vuagnoux (EPFL, Ilion)] 4023 4024 *) Make the no-err option work as intended. The intention with no-err 4025 is not to have the whole error stack handling routines removed from 4026 libcrypto, it's only intended to remove all the function name and 4027 reason texts, thereby removing some of the footprint that may not 4028 be interesting if those errors aren't displayed anyway. 4029 4030 NOTE: it's still possible for any application or module to have it's 4031 own set of error texts inserted. The routines are there, just not 4032 used by default when no-err is given. 4033 [Richard Levitte] 4034 4035 *) Add support for FreeBSD on IA64. 4036 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 4037 4038 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 4039 Kerberos function mit_des_cbc_cksum(). Before this change, 4040 the value returned by DES_cbc_cksum() was like the one from 4041 mit_des_cbc_cksum(), except the bytes were swapped. 4042 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 4043 4044 *) Allow an application to disable the automatic SSL chain building. 4045 Before this a rather primitive chain build was always performed in 4046 ssl3_output_cert_chain(): an application had no way to send the 4047 correct chain if the automatic operation produced an incorrect result. 4048 4049 Now the chain builder is disabled if either: 4050 4051 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 4052 4053 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 4054 4055 The reasoning behind this is that an application would not want the 4056 auto chain building to take place if extra chain certificates are 4057 present and it might also want a means of sending no additional 4058 certificates (for example the chain has two certificates and the 4059 root is omitted). 4060 [Steve Henson] 4061 4062 *) Add the possibility to build without the ENGINE framework. 4063 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4064 4065 *) Under Win32 gmtime() can return NULL: check return value in 4066 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4067 [Steve Henson] 4068 4069 *) DSA routines: under certain error conditions uninitialized BN objects 4070 could be freed. Solution: make sure initialization is performed early 4071 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4072 Nils Larsch <nla@trustcenter.de> via PR#459) 4073 [Lutz Jaenicke] 4074 4075 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4076 checked on reconnect on the client side, therefore session resumption 4077 could still fail with a "ssl session id is different" error. This 4078 behaviour is masked when SSL_OP_ALL is used due to 4079 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4080 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4081 followup to PR #377. 4082 [Lutz Jaenicke] 4083 4084 *) IA-32 assembler support enhancements: unified ELF targets, support 4085 for SCO/Caldera platforms, fix for Cygwin shared build. 4086 [Andy Polyakov] 4087 4088 *) Add support for FreeBSD on sparc64. As a consequence, support for 4089 FreeBSD on non-x86 processors is separate from x86 processors on 4090 the config script, much like the NetBSD support. 4091 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4092 4093 Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4094 4095 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4096 OpenSSL 0.9.7.] 4097 4098 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4099 code (06) was taken as the first octet of the session ID and the last 4100 octet was ignored consequently. As a result SSLv2 client side session 4101 caching could not have worked due to the session ID mismatch between 4102 client and server. 4103 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4104 PR #377. 4105 [Lutz Jaenicke] 4106 4107 *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4108 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4109 removed entirely. 4110 [Richard Levitte] 4111 4112 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4113 seems that in spite of existing for more than a year, many application 4114 author have done nothing to provide the necessary callbacks, which 4115 means that this particular engine will not work properly anywhere. 4116 This is a very unfortunate situation which forces us, in the name 4117 of usability, to give the hw_ncipher.c a static lock, which is part 4118 of libcrypto. 4119 NOTE: This is for the 0.9.7 series ONLY. This hack will never 4120 appear in 0.9.8 or later. We EXPECT application authors to have 4121 dealt properly with this when 0.9.8 is released (unless we actually 4122 make such changes in the libcrypto locking code that changes will 4123 have to be made anyway). 4124 [Richard Levitte] 4125 4126 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4127 octets have been read, EOF or an error occurs. Without this change 4128 some truncated ASN1 structures will not produce an error. 4129 [Steve Henson] 4130 4131 *) Disable Heimdal support, since it hasn't been fully implemented. 4132 Still give the possibility to force the use of Heimdal, but with 4133 warnings and a request that patches get sent to openssl-dev. 4134 [Richard Levitte] 4135 4136 *) Add the VC-CE target, introduce the WINCE sysname, and add 4137 INSTALL.WCE and appropriate conditionals to make it build. 4138 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4139 4140 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4141 cygssl-x.y.z.dll, where x, y and z are the major, minor and 4142 edit numbers of the version. 4143 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4144 4145 *) Introduce safe string copy and catenation functions 4146 (BUF_strlcpy() and BUF_strlcat()). 4147 [Ben Laurie (CHATS) and Richard Levitte] 4148 4149 *) Avoid using fixed-size buffers for one-line DNs. 4150 [Ben Laurie (CHATS)] 4151 4152 *) Add BUF_MEM_grow_clean() to avoid information leakage when 4153 resizing buffers containing secrets, and use where appropriate. 4154 [Ben Laurie (CHATS)] 4155 4156 *) Avoid using fixed size buffers for configuration file location. 4157 [Ben Laurie (CHATS)] 4158 4159 *) Avoid filename truncation for various CA files. 4160 [Ben Laurie (CHATS)] 4161 4162 *) Use sizeof in preference to magic numbers. 4163 [Ben Laurie (CHATS)] 4164 4165 *) Avoid filename truncation in cert requests. 4166 [Ben Laurie (CHATS)] 4167 4168 *) Add assertions to check for (supposedly impossible) buffer 4169 overflows. 4170 [Ben Laurie (CHATS)] 4171 4172 *) Don't cache truncated DNS entries in the local cache (this could 4173 potentially lead to a spoofing attack). 4174 [Ben Laurie (CHATS)] 4175 4176 *) Fix various buffers to be large enough for hex/decimal 4177 representations in a platform independent manner. 4178 [Ben Laurie (CHATS)] 4179 4180 *) Add CRYPTO_realloc_clean() to avoid information leakage when 4181 resizing buffers containing secrets, and use where appropriate. 4182 [Ben Laurie (CHATS)] 4183 4184 *) Add BIO_indent() to avoid much slightly worrying code to do 4185 indents. 4186 [Ben Laurie (CHATS)] 4187 4188 *) Convert sprintf()/BIO_puts() to BIO_printf(). 4189 [Ben Laurie (CHATS)] 4190 4191 *) buffer_gets() could terminate with the buffer only half 4192 full. Fixed. 4193 [Ben Laurie (CHATS)] 4194 4195 *) Add assertions to prevent user-supplied crypto functions from 4196 overflowing internal buffers by having large block sizes, etc. 4197 [Ben Laurie (CHATS)] 4198 4199 *) New OPENSSL_assert() macro (similar to assert(), but enabled 4200 unconditionally). 4201 [Ben Laurie (CHATS)] 4202 4203 *) Eliminate unused copy of key in RC4. 4204 [Ben Laurie (CHATS)] 4205 4206 *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4207 [Ben Laurie (CHATS)] 4208 4209 *) Fix off-by-one error in EGD path. 4210 [Ben Laurie (CHATS)] 4211 4212 *) If RANDFILE path is too long, ignore instead of truncating. 4213 [Ben Laurie (CHATS)] 4214 4215 *) Eliminate unused and incorrectly sized X.509 structure 4216 CBCParameter. 4217 [Ben Laurie (CHATS)] 4218 4219 *) Eliminate unused and dangerous function knumber(). 4220 [Ben Laurie (CHATS)] 4221 4222 *) Eliminate unused and dangerous structure, KSSL_ERR. 4223 [Ben Laurie (CHATS)] 4224 4225 *) Protect against overlong session ID context length in an encoded 4226 session object. Since these are local, this does not appear to be 4227 exploitable. 4228 [Ben Laurie (CHATS)] 4229 4230 *) Change from security patch (see 0.9.6e below) that did not affect 4231 the 0.9.6 release series: 4232 4233 Remote buffer overflow in SSL3 protocol - an attacker could 4234 supply an oversized master key in Kerberos-enabled versions. 4235 (CVE-2002-0657) 4236 [Ben Laurie (CHATS)] 4237 4238 *) Change the SSL kerb5 codes to match RFC 2712. 4239 [Richard Levitte] 4240 4241 *) Make -nameopt work fully for req and add -reqopt switch. 4242 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4243 4244 *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4245 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4246 4247 *) Make sure tests can be performed even if the corresponding algorithms 4248 have been removed entirely. This was also the last step to make 4249 OpenSSL compilable with DJGPP under all reasonable conditions. 4250 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4251 4252 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4253 to allow version independent disabling of normally unselected ciphers, 4254 which may be activated as a side-effect of selecting a single cipher. 4255 4256 (E.g., cipher list string "RSA" enables ciphersuites that are left 4257 out of "ALL" because they do not provide symmetric encryption. 4258 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4259 [Lutz Jaenicke, Bodo Moeller] 4260 4261 *) Add appropriate support for separate platform-dependent build 4262 directories. The recommended way to make a platform-dependent 4263 build directory is the following (tested on Linux), maybe with 4264 some local tweaks: 4265 4266 # Place yourself outside of the OpenSSL source tree. In 4267 # this example, the environment variable OPENSSL_SOURCE 4268 # is assumed to contain the absolute OpenSSL source directory. 4269 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4270 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4271 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4272 mkdir -p `dirname $F` 4273 ln -s $OPENSSL_SOURCE/$F $F 4274 done 4275 4276 To be absolutely sure not to disturb the source tree, a "make clean" 4277 is a good thing. If it isn't successfull, don't worry about it, 4278 it probably means the source directory is very clean. 4279 [Richard Levitte] 4280 4281 *) Make sure any ENGINE control commands make local copies of string 4282 pointers passed to them whenever necessary. Otherwise it is possible 4283 the caller may have overwritten (or deallocated) the original string 4284 data when a later ENGINE operation tries to use the stored values. 4285 [G�tz Babin-Ebell <babinebell@trustcenter.de>] 4286 4287 *) Improve diagnostics in file reading and command-line digests. 4288 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4289 4290 *) Add AES modes CFB and OFB to the object database. Correct an 4291 error in AES-CFB decryption. 4292 [Richard Levitte] 4293 4294 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4295 allows existing EVP_CIPHER_CTX structures to be reused after 4296 calling EVP_*Final(). This behaviour is used by encryption 4297 BIOs and some applications. This has the side effect that 4298 applications must explicitly clean up cipher contexts with 4299 EVP_CIPHER_CTX_cleanup() or they will leak memory. 4300 [Steve Henson] 4301 4302 *) Check the values of dna and dnb in bn_mul_recursive before calling 4303 bn_mul_comba (a non zero value means the a or b arrays do not contain 4304 n2 elements) and fallback to bn_mul_normal if either is not zero. 4305 [Steve Henson] 4306 4307 *) Fix escaping of non-ASCII characters when using the -subj option 4308 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4309 [Lutz Jaenicke] 4310 4311 *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4312 form for "surname", serialNumber has no short form. 4313 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4314 therefore remove "mail" short name for "internet 7". 4315 The OID for unique identifiers in X509 certificates is 4316 x500UniqueIdentifier, not uniqueIdentifier. 4317 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4318 [Lutz Jaenicke] 4319 4320 *) Add an "init" command to the ENGINE config module and auto initialize 4321 ENGINEs. Without any "init" command the ENGINE will be initialized 4322 after all ctrl commands have been executed on it. If init=1 the 4323 ENGINE is initailized at that point (ctrls before that point are run 4324 on the uninitialized ENGINE and after on the initialized one). If 4325 init=0 then the ENGINE will not be iniatialized at all. 4326 [Steve Henson] 4327 4328 *) Fix the 'app_verify_callback' interface so that the user-defined 4329 argument is actually passed to the callback: In the 4330 SSL_CTX_set_cert_verify_callback() prototype, the callback 4331 declaration has been changed from 4332 int (*cb)() 4333 into 4334 int (*cb)(X509_STORE_CTX *,void *); 4335 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4336 i=s->ctx->app_verify_callback(&ctx) 4337 has been changed into 4338 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4339 4340 To update applications using SSL_CTX_set_cert_verify_callback(), 4341 a dummy argument can be added to their callback functions. 4342 [D. K. Smetters <smetters@parc.xerox.com>] 4343 4344 *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4345 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4346 4347 *) Add and OPENSSL_LOAD_CONF define which will cause 4348 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4349 This allows older applications to transparently support certain 4350 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4351 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4352 load the config file and OPENSSL_add_all_algorithms_conf() which will 4353 always load it have also been added. 4354 [Steve Henson] 4355 4356 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4357 Adjust NIDs and EVP layer. 4358 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4359 4360 *) Config modules support in openssl utility. 4361 4362 Most commands now load modules from the config file, 4363 though in a few (such as version) this isn't done 4364 because it couldn't be used for anything. 4365 4366 In the case of ca and req the config file used is 4367 the same as the utility itself: that is the -config 4368 command line option can be used to specify an 4369 alternative file. 4370 [Steve Henson] 4371 4372 *) Move default behaviour from OPENSSL_config(). If appname is NULL 4373 use "openssl_conf" if filename is NULL use default openssl config file. 4374 [Steve Henson] 4375 4376 *) Add an argument to OPENSSL_config() to allow the use of an alternative 4377 config section name. Add a new flag to tolerate a missing config file 4378 and move code to CONF_modules_load_file(). 4379 [Steve Henson] 4380 4381 *) Support for crypto accelerator cards from Accelerated Encryption 4382 Processing, www.aep.ie. (Use engine 'aep') 4383 The support was copied from 0.9.6c [engine] and adapted/corrected 4384 to work with the new engine framework. 4385 [AEP Inc. and Richard Levitte] 4386 4387 *) Support for SureWare crypto accelerator cards from Baltimore 4388 Technologies. (Use engine 'sureware') 4389 The support was copied from 0.9.6c [engine] and adapted 4390 to work with the new engine framework. 4391 [Richard Levitte] 4392 4393 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4394 make the newer ENGINE framework commands for the CHIL engine work. 4395 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4396 4397 *) Make it possible to produce shared libraries on ReliantUNIX. 4398 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4399 4400 *) Add the configuration target debug-linux-ppro. 4401 Make 'openssl rsa' use the general key loading routines 4402 implemented in apps.c, and make those routines able to 4403 handle the key format FORMAT_NETSCAPE and the variant 4404 FORMAT_IISSGC. 4405 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4406 4407 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4408 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4409 4410 *) Add -keyform to rsautl, and document -engine. 4411 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4412 4413 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4414 BIO_R_NO_SUCH_FILE error code rather than the generic 4415 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4416 [Ben Laurie] 4417 4418 *) Add new functions 4419 ERR_peek_last_error 4420 ERR_peek_last_error_line 4421 ERR_peek_last_error_line_data. 4422 These are similar to 4423 ERR_peek_error 4424 ERR_peek_error_line 4425 ERR_peek_error_line_data, 4426 but report on the latest error recorded rather than the first one 4427 still in the error queue. 4428 [Ben Laurie, Bodo Moeller] 4429 4430 *) default_algorithms option in ENGINE config module. This allows things 4431 like: 4432 default_algorithms = ALL 4433 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4434 [Steve Henson] 4435 4436 *) Prelminary ENGINE config module. 4437 [Steve Henson] 4438 4439 *) New experimental application configuration code. 4440 [Steve Henson] 4441 4442 *) Change the AES code to follow the same name structure as all other 4443 symmetric ciphers, and behave the same way. Move everything to 4444 the directory crypto/aes, thereby obsoleting crypto/rijndael. 4445 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4446 4447 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4448 [Ben Laurie and Theo de Raadt] 4449 4450 *) Add option to output public keys in req command. 4451 [Massimiliano Pala madwolf@openca.org] 4452 4453 *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4454 (up to about 10% better than before for P-192 and P-224). 4455 [Bodo Moeller] 4456 4457 *) New functions/macros 4458 4459 SSL_CTX_set_msg_callback(ctx, cb) 4460 SSL_CTX_set_msg_callback_arg(ctx, arg) 4461 SSL_set_msg_callback(ssl, cb) 4462 SSL_set_msg_callback_arg(ssl, arg) 4463 4464 to request calling a callback function 4465 4466 void cb(int write_p, int version, int content_type, 4467 const void *buf, size_t len, SSL *ssl, void *arg) 4468 4469 whenever a protocol message has been completely received 4470 (write_p == 0) or sent (write_p == 1). Here 'version' is the 4471 protocol version according to which the SSL library interprets 4472 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4473 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4474 the content type as defined in the SSL 3.0/TLS 1.0 protocol 4475 specification (change_cipher_spec(20), alert(21), handshake(22)). 4476 'buf' and 'len' point to the actual message, 'ssl' to the 4477 SSL object, and 'arg' is the application-defined value set by 4478 SSL[_CTX]_set_msg_callback_arg(). 4479 4480 'openssl s_client' and 'openssl s_server' have new '-msg' options 4481 to enable a callback that displays all protocol messages. 4482 [Bodo Moeller] 4483 4484 *) Change the shared library support so shared libraries are built as 4485 soon as the corresponding static library is finished, and thereby get 4486 openssl and the test programs linked against the shared library. 4487 This still only happens when the keyword "shard" has been given to 4488 the configuration scripts. 4489 4490 NOTE: shared library support is still an experimental thing, and 4491 backward binary compatibility is still not guaranteed. 4492 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4493 4494 *) Add support for Subject Information Access extension. 4495 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4496 4497 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4498 additional bytes when new memory had to be allocated, not just 4499 when reusing an existing buffer. 4500 [Bodo Moeller] 4501 4502 *) New command line and configuration option 'utf8' for the req command. 4503 This allows field values to be specified as UTF8 strings. 4504 [Steve Henson] 4505 4506 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4507 runs for the former and machine-readable output for the latter. 4508 [Ben Laurie] 4509 4510 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4511 of the e-mail address in the DN (i.e., it will go into a certificate 4512 extension only). The new configuration file option 'email_in_dn = no' 4513 has the same effect. 4514 [Massimiliano Pala madwolf@openca.org] 4515 4516 *) Change all functions with names starting with des_ to be starting 4517 with DES_ instead. Add wrappers that are compatible with libdes, 4518 but are named _ossl_old_des_*. Finally, add macros that map the 4519 des_* symbols to the corresponding _ossl_old_des_* if libdes 4520 compatibility is desired. If OpenSSL 0.9.6c compatibility is 4521 desired, the des_* symbols will be mapped to DES_*, with one 4522 exception. 4523 4524 Since we provide two compatibility mappings, the user needs to 4525 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4526 compatibility is desired. The default (i.e., when that macro 4527 isn't defined) is OpenSSL 0.9.6c compatibility. 4528 4529 There are also macros that enable and disable the support of old 4530 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4531 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4532 are defined, the default will apply: to support the old des routines. 4533 4534 In either case, one must include openssl/des.h to get the correct 4535 definitions. Do not try to just include openssl/des_old.h, that 4536 won't work. 4537 4538 NOTE: This is a major break of an old API into a new one. Software 4539 authors are encouraged to switch to the DES_ style functions. Some 4540 time in the future, des_old.h and the libdes compatibility functions 4541 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4542 default), and then completely removed. 4543 [Richard Levitte] 4544 4545 *) Test for certificates which contain unsupported critical extensions. 4546 If such a certificate is found during a verify operation it is 4547 rejected by default: this behaviour can be overridden by either 4548 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4549 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4550 X509_supported_extension() has also been added which returns 1 if a 4551 particular extension is supported. 4552 [Steve Henson] 4553 4554 *) Modify the behaviour of EVP cipher functions in similar way to digests 4555 to retain compatibility with existing code. 4556 [Steve Henson] 4557 4558 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4559 compatibility with existing code. In particular the 'ctx' parameter does 4560 not have to be to be initialized before the call to EVP_DigestInit() and 4561 it is tidied up after a call to EVP_DigestFinal(). New function 4562 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4563 EVP_MD_CTX_copy() changed to not require the destination to be 4564 initialized valid and new function EVP_MD_CTX_copy_ex() added which 4565 requires the destination to be valid. 4566 4567 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4568 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4569 [Steve Henson] 4570 4571 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4572 so that complete 'Handshake' protocol structures are kept in memory 4573 instead of overwriting 'msg_type' and 'length' with 'body' data. 4574 [Bodo Moeller] 4575 4576 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4577 [Massimo Santin via Richard Levitte] 4578 4579 *) Major restructuring to the underlying ENGINE code. This includes 4580 reduction of linker bloat, separation of pure "ENGINE" manipulation 4581 (initialisation, etc) from functionality dealing with implementations 4582 of specific crypto iterfaces. This change also introduces integrated 4583 support for symmetric ciphers and digest implementations - so ENGINEs 4584 can now accelerate these by providing EVP_CIPHER and EVP_MD 4585 implementations of their own. This is detailed in crypto/engine/README 4586 as it couldn't be adequately described here. However, there are a few 4587 API changes worth noting - some RSA, DSA, DH, and RAND functions that 4588 were changed in the original introduction of ENGINE code have now 4589 reverted back - the hooking from this code to ENGINE is now a good 4590 deal more passive and at run-time, operations deal directly with 4591 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4592 dereferencing through an ENGINE pointer any more. Also, the ENGINE 4593 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4594 they were not being used by the framework as there is no concept of a 4595 BIGNUM_METHOD and they could not be generalised to the new 4596 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4597 ENGINE_cpy() has been removed as it cannot be consistently defined in 4598 the new code. 4599 [Geoff Thorpe] 4600 4601 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4602 [Steve Henson] 4603 4604 *) Change mkdef.pl to sort symbols that get the same entry number, 4605 and make sure the automatically generated functions ERR_load_* 4606 become part of libeay.num as well. 4607 [Richard Levitte] 4608 4609 *) New function SSL_renegotiate_pending(). This returns true once 4610 renegotiation has been requested (either SSL_renegotiate() call 4611 or HelloRequest/ClientHello receveived from the peer) and becomes 4612 false once a handshake has been completed. 4613 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4614 sends a HelloRequest, but does not ensure that a handshake takes 4615 place. SSL_renegotiate_pending() is useful for checking if the 4616 client has followed the request.) 4617 [Bodo Moeller] 4618 4619 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4620 By default, clients may request session resumption even during 4621 renegotiation (if session ID contexts permit); with this option, 4622 session resumption is possible only in the first handshake. 4623 4624 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4625 more bits available for options that should not be part of 4626 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4627 [Bodo Moeller] 4628 4629 *) Add some demos for certificate and certificate request creation. 4630 [Steve Henson] 4631 4632 *) Make maximum certificate chain size accepted from the peer application 4633 settable (SSL*_get/set_max_cert_list()), as proposed by 4634 "Douglas E. Engert" <deengert@anl.gov>. 4635 [Lutz Jaenicke] 4636 4637 *) Add support for shared libraries for Unixware-7 4638 (Boyd Lynn Gerber <gerberb@zenez.com>). 4639 [Lutz Jaenicke] 4640 4641 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4642 be done prior to destruction. Use this to unload error strings from 4643 ENGINEs that load their own error strings. NB: This adds two new API 4644 functions to "get" and "set" this destroy handler in an ENGINE. 4645 [Geoff Thorpe] 4646 4647 *) Alter all existing ENGINE implementations (except "openssl" and 4648 "openbsd") to dynamically instantiate their own error strings. This 4649 makes them more flexible to be built both as statically-linked ENGINEs 4650 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4651 Also, add stub code to each that makes building them as self-contained 4652 shared-libraries easier (see README.ENGINE). 4653 [Geoff Thorpe] 4654 4655 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4656 implementations into applications that are completely implemented in 4657 self-contained shared-libraries. The "dynamic" ENGINE exposes control 4658 commands that can be used to configure what shared-library to load and 4659 to control aspects of the way it is handled. Also, made an update to 4660 the README.ENGINE file that brings its information up-to-date and 4661 provides some information and instructions on the "dynamic" ENGINE 4662 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4663 [Geoff Thorpe] 4664 4665 *) Make it possible to unload ranges of ERR strings with a new 4666 "ERR_unload_strings" function. 4667 [Geoff Thorpe] 4668 4669 *) Add a copy() function to EVP_MD. 4670 [Ben Laurie] 4671 4672 *) Make EVP_MD routines take a context pointer instead of just the 4673 md_data void pointer. 4674 [Ben Laurie] 4675 4676 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4677 that the digest can only process a single chunk of data 4678 (typically because it is provided by a piece of 4679 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4680 is only going to provide a single chunk of data, and hence the 4681 framework needn't accumulate the data for oneshot drivers. 4682 [Ben Laurie] 4683 4684 *) As with "ERR", make it possible to replace the underlying "ex_data" 4685 functions. This change also alters the storage and management of global 4686 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4687 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4688 index counters. The API functions that use this state have been changed 4689 to take a "class_index" rather than pointers to the class's local STACK 4690 and counter, and there is now an API function to dynamically create new 4691 classes. This centralisation allows us to (a) plug a lot of the 4692 thread-safety problems that existed, and (b) makes it possible to clean 4693 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4694 such data would previously have always leaked in application code and 4695 workarounds were in place to make the memory debugging turn a blind eye 4696 to it. Application code that doesn't use this new function will still 4697 leak as before, but their memory debugging output will announce it now 4698 rather than letting it slide. 4699 4700 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4701 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4702 has a return value to indicate success or failure. 4703 [Geoff Thorpe] 4704 4705 *) Make it possible to replace the underlying "ERR" functions such that the 4706 global state (2 LHASH tables and 2 locks) is only used by the "default" 4707 implementation. This change also adds two functions to "get" and "set" 4708 the implementation prior to it being automatically set the first time 4709 any other ERR function takes place. Ie. an application can call "get", 4710 pass the return value to a module it has just loaded, and that module 4711 can call its own "set" function using that value. This means the 4712 module's "ERR" operations will use (and modify) the error state in the 4713 application and not in its own statically linked copy of OpenSSL code. 4714 [Geoff Thorpe] 4715 4716 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4717 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4718 the operation, and provides a more encapsulated way for external code 4719 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4720 to use these functions rather than manually incrementing the counts. 4721 4722 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4723 [Geoff Thorpe] 4724 4725 *) Add EVP test program. 4726 [Ben Laurie] 4727 4728 *) Add symmetric cipher support to ENGINE. Expect the API to change! 4729 [Ben Laurie] 4730 4731 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4732 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4733 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4734 These allow a CRL to be built without having to access X509_CRL fields 4735 directly. Modify 'ca' application to use new functions. 4736 [Steve Henson] 4737 4738 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4739 bug workarounds. Rollback attack detection is a security feature. 4740 The problem will only arise on OpenSSL servers when TLSv1 is not 4741 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4742 Software authors not wanting to support TLSv1 will have special reasons 4743 for their choice and can explicitly enable this option. 4744 [Bodo Moeller, Lutz Jaenicke] 4745 4746 *) Rationalise EVP so it can be extended: don't include a union of 4747 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4748 (similar to those existing for EVP_CIPHER_CTX). 4749 Usage example: 4750 4751 EVP_MD_CTX md; 4752 4753 EVP_MD_CTX_init(&md); /* new function call */ 4754 EVP_DigestInit(&md, EVP_sha1()); 4755 EVP_DigestUpdate(&md, in, len); 4756 EVP_DigestFinal(&md, out, NULL); 4757 EVP_MD_CTX_cleanup(&md); /* new function call */ 4758 4759 [Ben Laurie] 4760 4761 *) Make DES key schedule conform to the usual scheme, as well as 4762 correcting its structure. This means that calls to DES functions 4763 now have to pass a pointer to a des_key_schedule instead of a 4764 plain des_key_schedule (which was actually always a pointer 4765 anyway): E.g., 4766 4767 des_key_schedule ks; 4768 4769 des_set_key_checked(..., &ks); 4770 des_ncbc_encrypt(..., &ks, ...); 4771 4772 (Note that a later change renames 'des_...' into 'DES_...'.) 4773 [Ben Laurie] 4774 4775 *) Initial reduction of linker bloat: the use of some functions, such as 4776 PEM causes large amounts of unused functions to be linked in due to 4777 poor organisation. For example pem_all.c contains every PEM function 4778 which has a knock on effect of linking in large amounts of (unused) 4779 ASN1 code. Grouping together similar functions and splitting unrelated 4780 functions prevents this. 4781 [Steve Henson] 4782 4783 *) Cleanup of EVP macros. 4784 [Ben Laurie] 4785 4786 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4787 correct _ecb suffix. 4788 [Ben Laurie] 4789 4790 *) Add initial OCSP responder support to ocsp application. The 4791 revocation information is handled using the text based index 4792 use by the ca application. The responder can either handle 4793 requests generated internally, supplied in files (for example 4794 via a CGI script) or using an internal minimal server. 4795 [Steve Henson] 4796 4797 *) Add configuration choices to get zlib compression for TLS. 4798 [Richard Levitte] 4799 4800 *) Changes to Kerberos SSL for RFC 2712 compliance: 4801 1. Implemented real KerberosWrapper, instead of just using 4802 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4803 2. Implemented optional authenticator field of KerberosWrapper. 4804 4805 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4806 and authenticator structs; see crypto/krb5/. 4807 4808 Generalized Kerberos calls to support multiple Kerberos libraries. 4809 [Vern Staats <staatsvr@asc.hpc.mil>, 4810 Jeffrey Altman <jaltman@columbia.edu> 4811 via Richard Levitte] 4812 4813 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4814 already does with RSA. testdsa.h now has 'priv_key/pub_key' 4815 values for each of the key sizes rather than having just 4816 parameters (and 'speed' generating keys each time). 4817 [Geoff Thorpe] 4818 4819 *) Speed up EVP routines. 4820 Before: 4821encrypt 4822type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4823des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4824des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4825des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4826decrypt 4827des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4828des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4829des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4830 After: 4831encrypt 4832des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4833decrypt 4834des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4835 [Ben Laurie] 4836 4837 *) Added the OS2-EMX target. 4838 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4839 4840 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4841 to support NCONF routines in extension code. New function CONF_set_nconf() 4842 to allow functions which take an NCONF to also handle the old LHASH 4843 structure: this means that the old CONF compatible routines can be 4844 retained (in particular wrt extensions) without having to duplicate the 4845 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4846 [Steve Henson] 4847 4848 *) Enhance the general user interface with mechanisms for inner control 4849 and with possibilities to have yes/no kind of prompts. 4850 [Richard Levitte] 4851 4852 *) Change all calls to low level digest routines in the library and 4853 applications to use EVP. Add missing calls to HMAC_cleanup() and 4854 don't assume HMAC_CTX can be copied using memcpy(). 4855 [Verdon Walker <VWalker@novell.com>, Steve Henson] 4856 4857 *) Add the possibility to control engines through control names but with 4858 arbitrary arguments instead of just a string. 4859 Change the key loaders to take a UI_METHOD instead of a callback 4860 function pointer. NOTE: this breaks binary compatibility with earlier 4861 versions of OpenSSL [engine]. 4862 Adapt the nCipher code for these new conditions and add a card insertion 4863 callback. 4864 [Richard Levitte] 4865 4866 *) Enhance the general user interface with mechanisms to better support 4867 dialog box interfaces, application-defined prompts, the possibility 4868 to use defaults (for example default passwords from somewhere else) 4869 and interrupts/cancellations. 4870 [Richard Levitte] 4871 4872 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4873 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4874 [Steve Henson] 4875 4876 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 4877 tidy up some unnecessarily weird code in 'sk_new()'). 4878 [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 4879 4880 *) Change the key loading routines for ENGINEs to use the same kind 4881 callback (pem_password_cb) as all other routines that need this 4882 kind of callback. 4883 [Richard Levitte] 4884 4885 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 4886 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 4887 than this minimum value is recommended. 4888 [Lutz Jaenicke] 4889 4890 *) New random seeder for OpenVMS, using the system process statistics 4891 that are easily reachable. 4892 [Richard Levitte] 4893 4894 *) Windows apparently can't transparently handle global 4895 variables defined in DLLs. Initialisations such as: 4896 4897 const ASN1_ITEM *it = &ASN1_INTEGER_it; 4898 4899 wont compile. This is used by the any applications that need to 4900 declare their own ASN1 modules. This was fixed by adding the option 4901 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 4902 needed for static libraries under Win32. 4903 [Steve Henson] 4904 4905 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 4906 setting of purpose and trust fields. New X509_STORE trust and 4907 purpose functions and tidy up setting in other SSL functions. 4908 [Steve Henson] 4909 4910 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 4911 structure. These are inherited by X509_STORE_CTX when it is 4912 initialised. This allows various defaults to be set in the 4913 X509_STORE structure (such as flags for CRL checking and custom 4914 purpose or trust settings) for functions which only use X509_STORE_CTX 4915 internally such as S/MIME. 4916 4917 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 4918 trust settings if they are not set in X509_STORE. This allows X509_STORE 4919 purposes and trust (in S/MIME for example) to override any set by default. 4920 4921 Add command line options for CRL checking to smime, s_client and s_server 4922 applications. 4923 [Steve Henson] 4924 4925 *) Initial CRL based revocation checking. If the CRL checking flag(s) 4926 are set then the CRL is looked up in the X509_STORE structure and 4927 its validity and signature checked, then if the certificate is found 4928 in the CRL the verify fails with a revoked error. 4929 4930 Various new CRL related callbacks added to X509_STORE_CTX structure. 4931 4932 Command line options added to 'verify' application to support this. 4933 4934 This needs some additional work, such as being able to handle multiple 4935 CRLs with different times, extension based lookup (rather than just 4936 by subject name) and ultimately more complete V2 CRL extension 4937 handling. 4938 [Steve Henson] 4939 4940 *) Add a general user interface API (crypto/ui/). This is designed 4941 to replace things like des_read_password and friends (backward 4942 compatibility functions using this new API are provided). 4943 The purpose is to remove prompting functions from the DES code 4944 section as well as provide for prompting through dialog boxes in 4945 a window system and the like. 4946 [Richard Levitte] 4947 4948 *) Add "ex_data" support to ENGINE so implementations can add state at a 4949 per-structure level rather than having to store it globally. 4950 [Geoff] 4951 4952 *) Make it possible for ENGINE structures to be copied when retrieved by 4953 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 4954 This causes the "original" ENGINE structure to act like a template, 4955 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 4956 operational state can be localised to each ENGINE structure, despite the 4957 fact they all share the same "methods". New ENGINE structures returned in 4958 this case have no functional references and the return value is the single 4959 structural reference. This matches the single structural reference returned 4960 by ENGINE_by_id() normally, when it is incremented on the pre-existing 4961 ENGINE structure. 4962 [Geoff] 4963 4964 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 4965 needs to match any other type at all we need to manually clear the 4966 tag cache. 4967 [Steve Henson] 4968 4969 *) Changes to the "openssl engine" utility to include; 4970 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 4971 about an ENGINE's available control commands. 4972 - executing control commands from command line arguments using the 4973 '-pre' and '-post' switches. '-post' is only used if '-t' is 4974 specified and the ENGINE is successfully initialised. The syntax for 4975 the individual commands are colon-separated, for example; 4976 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 4977 [Geoff] 4978 4979 *) New dynamic control command support for ENGINEs. ENGINEs can now 4980 declare their own commands (numbers), names (strings), descriptions, 4981 and input types for run-time discovery by calling applications. A 4982 subset of these commands are implicitly classed as "executable" 4983 depending on their input type, and only these can be invoked through 4984 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 4985 can be based on user input, config files, etc). The distinction is 4986 that "executable" commands cannot return anything other than a boolean 4987 result and can only support numeric or string input, whereas some 4988 discoverable commands may only be for direct use through 4989 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 4990 pointers, or other custom uses. The "executable" commands are to 4991 support parameterisations of ENGINE behaviour that can be 4992 unambiguously defined by ENGINEs and used consistently across any 4993 OpenSSL-based application. Commands have been added to all the 4994 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 4995 control over shared-library paths without source code alterations. 4996 [Geoff] 4997 4998 *) Changed all ENGINE implementations to dynamically allocate their 4999 ENGINEs rather than declaring them statically. Apart from this being 5000 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 5001 this also allows the implementations to compile without using the 5002 internal engine_int.h header. 5003 [Geoff] 5004 5005 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 5006 'const' value. Any code that should be able to modify a RAND_METHOD 5007 should already have non-const pointers to it (ie. they should only 5008 modify their own ones). 5009 [Geoff] 5010 5011 *) Made a variety of little tweaks to the ENGINE code. 5012 - "atalla" and "ubsec" string definitions were moved from header files 5013 to C code. "nuron" string definitions were placed in variables 5014 rather than hard-coded - allowing parameterisation of these values 5015 later on via ctrl() commands. 5016 - Removed unused "#if 0"'d code. 5017 - Fixed engine list iteration code so it uses ENGINE_free() to release 5018 structural references. 5019 - Constified the RAND_METHOD element of ENGINE structures. 5020 - Constified various get/set functions as appropriate and added 5021 missing functions (including a catch-all ENGINE_cpy that duplicates 5022 all ENGINE values onto a new ENGINE except reference counts/state). 5023 - Removed NULL parameter checks in get/set functions. Setting a method 5024 or function to NULL is a way of cancelling out a previously set 5025 value. Passing a NULL ENGINE parameter is just plain stupid anyway 5026 and doesn't justify the extra error symbols and code. 5027 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5028 flags from engine_int.h to engine.h. 5029 - Changed prototypes for ENGINE handler functions (init(), finish(), 5030 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5031 [Geoff] 5032 5033 *) Implement binary inversion algorithm for BN_mod_inverse in addition 5034 to the algorithm using long division. The binary algorithm can be 5035 used only if the modulus is odd. On 32-bit systems, it is faster 5036 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 5037 roughly 5-15% for 256-bit moduli), so we use it only for moduli 5038 up to 450 bits. In 64-bit environments, the binary algorithm 5039 appears to be advantageous for much longer moduli; here we use it 5040 for moduli up to 2048 bits. 5041 [Bodo Moeller] 5042 5043 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 5044 could not support the combine flag in choice fields. 5045 [Steve Henson] 5046 5047 *) Add a 'copy_extensions' option to the 'ca' utility. This copies 5048 extensions from a certificate request to the certificate. 5049 [Steve Henson] 5050 5051 *) Allow multiple 'certopt' and 'nameopt' options to be separated 5052 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 5053 file: this allows the display of the certificate about to be 5054 signed to be customised, to allow certain fields to be included 5055 or excluded and extension details. The old system didn't display 5056 multicharacter strings properly, omitted fields not in the policy 5057 and couldn't display additional details such as extensions. 5058 [Steve Henson] 5059 5060 *) Function EC_POINTs_mul for multiple scalar multiplication 5061 of an arbitrary number of elliptic curve points 5062 \sum scalars[i]*points[i], 5063 optionally including the generator defined for the EC_GROUP: 5064 scalar*generator + \sum scalars[i]*points[i]. 5065 5066 EC_POINT_mul is a simple wrapper function for the typical case 5067 that the point list has just one item (besides the optional 5068 generator). 5069 [Bodo Moeller] 5070 5071 *) First EC_METHODs for curves over GF(p): 5072 5073 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5074 operations and provides various method functions that can also 5075 operate with faster implementations of modular arithmetic. 5076 5077 EC_GFp_mont_method() reuses most functions that are part of 5078 EC_GFp_simple_method, but uses Montgomery arithmetic. 5079 5080 [Bodo Moeller; point addition and point doubling 5081 implementation directly derived from source code provided by 5082 Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5083 5084 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5085 crypto/ec/ec_lib.c): 5086 5087 Curves are EC_GROUP objects (with an optional group generator) 5088 based on EC_METHODs that are built into the library. 5089 5090 Points are EC_POINT objects based on EC_GROUP objects. 5091 5092 Most of the framework would be able to handle curves over arbitrary 5093 finite fields, but as there are no obvious types for fields other 5094 than GF(p), some functions are limited to that for now. 5095 [Bodo Moeller] 5096 5097 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5098 that the file contains a complete HTTP response. 5099 [Richard Levitte] 5100 5101 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5102 change the def and num file printf format specifier from "%-40sXXX" 5103 to "%-39s XXX". The latter will always guarantee a space after the 5104 field while the former will cause them to run together if the field 5105 is 40 of more characters long. 5106 [Steve Henson] 5107 5108 *) Constify the cipher and digest 'method' functions and structures 5109 and modify related functions to take constant EVP_MD and EVP_CIPHER 5110 pointers. 5111 [Steve Henson] 5112 5113 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5114 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5115 [Bodo Moeller] 5116 5117 *) Modify EVP_Digest*() routines so they now return values. Although the 5118 internal software routines can never fail additional hardware versions 5119 might. 5120 [Steve Henson] 5121 5122 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5123 5124 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5125 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5126 5127 ASN1 error codes 5128 ERR_R_NESTED_ASN1_ERROR 5129 ... 5130 ERR_R_MISSING_ASN1_EOS 5131 were 4 .. 9, conflicting with 5132 ERR_LIB_RSA (= ERR_R_RSA_LIB) 5133 ... 5134 ERR_LIB_PEM (= ERR_R_PEM_LIB). 5135 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5136 5137 Add new error code 'ERR_R_INTERNAL_ERROR'. 5138 [Bodo Moeller] 5139 5140 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5141 suffices. 5142 [Bodo Moeller] 5143 5144 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5145 sets the subject name for a new request or supersedes the 5146 subject name in a given request. Formats that can be parsed are 5147 'CN=Some Name, OU=myOU, C=IT' 5148 and 5149 'CN=Some Name/OU=myOU/C=IT'. 5150 5151 Add options '-batch' and '-verbose' to 'openssl req'. 5152 [Massimiliano Pala <madwolf@hackmasters.net>] 5153 5154 *) Introduce the possibility to access global variables through 5155 functions on platform were that's the best way to handle exporting 5156 global variables in shared libraries. To enable this functionality, 5157 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5158 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5159 is normally done by Configure or something similar). 5160 5161 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5162 in the source file (foo.c) like this: 5163 5164 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5165 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5166 5167 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5168 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5169 5170 OPENSSL_DECLARE_GLOBAL(int,foo); 5171 #define foo OPENSSL_GLOBAL_REF(foo) 5172 OPENSSL_DECLARE_GLOBAL(double,bar); 5173 #define bar OPENSSL_GLOBAL_REF(bar) 5174 5175 The #defines are very important, and therefore so is including the 5176 header file everywhere where the defined globals are used. 5177 5178 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5179 of ASN.1 items, but that structure is a bit different. 5180 5181 The largest change is in util/mkdef.pl which has been enhanced with 5182 better and easier to understand logic to choose which symbols should 5183 go into the Windows .def files as well as a number of fixes and code 5184 cleanup (among others, algorithm keywords are now sorted 5185 lexicographically to avoid constant rewrites). 5186 [Richard Levitte] 5187 5188 *) In BN_div() keep a copy of the sign of 'num' before writing the 5189 result to 'rm' because if rm==num the value will be overwritten 5190 and produce the wrong result if 'num' is negative: this caused 5191 problems with BN_mod() and BN_nnmod(). 5192 [Steve Henson] 5193 5194 *) Function OCSP_request_verify(). This checks the signature on an 5195 OCSP request and verifies the signer certificate. The signer 5196 certificate is just checked for a generic purpose and OCSP request 5197 trust settings. 5198 [Steve Henson] 5199 5200 *) Add OCSP_check_validity() function to check the validity of OCSP 5201 responses. OCSP responses are prepared in real time and may only 5202 be a few seconds old. Simply checking that the current time lies 5203 between thisUpdate and nextUpdate max reject otherwise valid responses 5204 caused by either OCSP responder or client clock inaccuracy. Instead 5205 we allow thisUpdate and nextUpdate to fall within a certain period of 5206 the current time. The age of the response can also optionally be 5207 checked. Two new options -validity_period and -status_age added to 5208 ocsp utility. 5209 [Steve Henson] 5210 5211 *) If signature or public key algorithm is unrecognized print out its 5212 OID rather that just UNKNOWN. 5213 [Steve Henson] 5214 5215 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5216 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5217 ID to be generated from the issuer certificate alone which can then be 5218 passed to OCSP_id_issuer_cmp(). 5219 [Steve Henson] 5220 5221 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5222 ASN1 modules to export functions returning ASN1_ITEM pointers 5223 instead of the ASN1_ITEM structures themselves. This adds several 5224 new macros which allow the underlying ASN1 function/structure to 5225 be accessed transparently. As a result code should not use ASN1_ITEM 5226 references directly (such as &X509_it) but instead use the relevant 5227 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5228 use of the new ASN1 code on platforms where exporting structures 5229 is problematical (for example in shared libraries) but exporting 5230 functions returning pointers to structures is not. 5231 [Steve Henson] 5232 5233 *) Add support for overriding the generation of SSL/TLS session IDs. 5234 These callbacks can be registered either in an SSL_CTX or per SSL. 5235 The purpose of this is to allow applications to control, if they wish, 5236 the arbitrary values chosen for use as session IDs, particularly as it 5237 can be useful for session caching in multiple-server environments. A 5238 command-line switch for testing this (and any client code that wishes 5239 to use such a feature) has been added to "s_server". 5240 [Geoff Thorpe, Lutz Jaenicke] 5241 5242 *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5243 of the form '#if defined(...) || defined(...) || ...' and 5244 '#if !defined(...) && !defined(...) && ...'. This also avoids 5245 the growing number of special cases it was previously handling. 5246 [Richard Levitte] 5247 5248 *) Make all configuration macros available for application by making 5249 sure they are available in opensslconf.h, by giving them names starting 5250 with "OPENSSL_" to avoid conflicts with other packages and by making 5251 sure e_os2.h will cover all platform-specific cases together with 5252 opensslconf.h. 5253 Additionally, it is now possible to define configuration/platform- 5254 specific names (called "system identities"). In the C code, these 5255 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5256 macro with the name beginning with "OPENSSL_SYS_", which is determined 5257 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5258 what is available. 5259 [Richard Levitte] 5260 5261 *) New option -set_serial to 'req' and 'x509' this allows the serial 5262 number to use to be specified on the command line. Previously self 5263 signed certificates were hard coded with serial number 0 and the 5264 CA options of 'x509' had to use a serial number in a file which was 5265 auto incremented. 5266 [Steve Henson] 5267 5268 *) New options to 'ca' utility to support V2 CRL entry extensions. 5269 Currently CRL reason, invalidity date and hold instruction are 5270 supported. Add new CRL extensions to V3 code and some new objects. 5271 [Steve Henson] 5272 5273 *) New function EVP_CIPHER_CTX_set_padding() this is used to 5274 disable standard block padding (aka PKCS#5 padding) in the EVP 5275 API, which was previously mandatory. This means that the data is 5276 not padded in any way and so the total length much be a multiple 5277 of the block size, otherwise an error occurs. 5278 [Steve Henson] 5279 5280 *) Initial (incomplete) OCSP SSL support. 5281 [Steve Henson] 5282 5283 *) New function OCSP_parse_url(). This splits up a URL into its host, 5284 port and path components: primarily to parse OCSP URLs. New -url 5285 option to ocsp utility. 5286 [Steve Henson] 5287 5288 *) New nonce behavior. The return value of OCSP_check_nonce() now 5289 reflects the various checks performed. Applications can decide 5290 whether to tolerate certain situations such as an absent nonce 5291 in a response when one was present in a request: the ocsp application 5292 just prints out a warning. New function OCSP_add1_basic_nonce() 5293 this is to allow responders to include a nonce in a response even if 5294 the request is nonce-less. 5295 [Steve Henson] 5296 5297 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5298 skipped when using openssl x509 multiple times on a single input file, 5299 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5300 [Bodo Moeller] 5301 5302 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5303 set string type: to handle setting ASN1_TIME structures. Fix ca 5304 utility to correctly initialize revocation date of CRLs. 5305 [Steve Henson] 5306 5307 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5308 the clients preferred ciphersuites and rather use its own preferences. 5309 Should help to work around M$ SGC (Server Gated Cryptography) bug in 5310 Internet Explorer by ensuring unchanged hash method during stepup. 5311 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5312 [Lutz Jaenicke] 5313 5314 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5315 to aes and add a new 'exist' option to print out symbols that don't 5316 appear to exist. 5317 [Steve Henson] 5318 5319 *) Additional options to ocsp utility to allow flags to be set and 5320 additional certificates supplied. 5321 [Steve Henson] 5322 5323 *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5324 OCSP client a number of certificate to only verify the response 5325 signature against. 5326 [Richard Levitte] 5327 5328 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5329 handle the new API. Currently only ECB, CBC modes supported. Add new 5330 AES OIDs. 5331 5332 Add TLS AES ciphersuites as described in RFC3268, "Advanced 5333 Encryption Standard (AES) Ciphersuites for Transport Layer 5334 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5335 not enabled by default and were not part of the "ALL" ciphersuite 5336 alias because they were not yet official; they could be 5337 explicitly requested by specifying the "AESdraft" ciphersuite 5338 group alias. In the final release of OpenSSL 0.9.7, the group 5339 alias is called "AES" and is part of "ALL".) 5340 [Ben Laurie, Steve Henson, Bodo Moeller] 5341 5342 *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5343 request to response. 5344 [Steve Henson] 5345 5346 *) Functions for OCSP responders. OCSP_request_onereq_count(), 5347 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5348 extract information from a certificate request. OCSP_response_create() 5349 creates a response and optionally adds a basic response structure. 5350 OCSP_basic_add1_status() adds a complete single response to a basic 5351 response and returns the OCSP_SINGLERESP structure just added (to allow 5352 extensions to be included for example). OCSP_basic_add1_cert() adds a 5353 certificate to a basic response and OCSP_basic_sign() signs a basic 5354 response with various flags. New helper functions ASN1_TIME_check() 5355 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5356 (converts ASN1_TIME to GeneralizedTime). 5357 [Steve Henson] 5358 5359 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5360 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5361 structure from a certificate. X509_pubkey_digest() digests the public_key 5362 contents: this is used in various key identifiers. 5363 [Steve Henson] 5364 5365 *) Make sk_sort() tolerate a NULL argument. 5366 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5367 5368 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5369 passed by the function are trusted implicitly. If any of them signed the 5370 response then it is assumed to be valid and is not verified. 5371 [Steve Henson] 5372 5373 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5374 to data. This was previously part of the PKCS7 ASN1 code. This 5375 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5376 [Steve Henson, reported by Kenneth R. Robinette 5377 <support@securenetterm.com>] 5378 5379 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5380 routines: without these tracing memory leaks is very painful. 5381 Fix leaks in PKCS12 and PKCS7 routines. 5382 [Steve Henson] 5383 5384 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5385 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5386 effectively meant GeneralizedTime would never be used. Now it 5387 is initialised to -1 but X509_time_adj() now has to check the value 5388 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5389 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5390 [Steve Henson, reported by Kenneth R. Robinette 5391 <support@securenetterm.com>] 5392 5393 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5394 result in a zero length in the ASN1_INTEGER structure which was 5395 not consistent with the structure when d2i_ASN1_INTEGER() was used 5396 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5397 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5398 where it did not print out a minus for negative ASN1_INTEGER. 5399 [Steve Henson] 5400 5401 *) Add summary printout to ocsp utility. The various functions which 5402 convert status values to strings have been renamed to: 5403 OCSP_response_status_str(), OCSP_cert_status_str() and 5404 OCSP_crl_reason_str() and are no longer static. New options 5405 to verify nonce values and to disable verification. OCSP response 5406 printout format cleaned up. 5407 [Steve Henson] 5408 5409 *) Add additional OCSP certificate checks. These are those specified 5410 in RFC2560. This consists of two separate checks: the CA of the 5411 certificate being checked must either be the OCSP signer certificate 5412 or the issuer of the OCSP signer certificate. In the latter case the 5413 OCSP signer certificate must contain the OCSP signing extended key 5414 usage. This check is performed by attempting to match the OCSP 5415 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5416 in the OCSP_CERTID structures of the response. 5417 [Steve Henson] 5418 5419 *) Initial OCSP certificate verification added to OCSP_basic_verify() 5420 and related routines. This uses the standard OpenSSL certificate 5421 verify routines to perform initial checks (just CA validity) and 5422 to obtain the certificate chain. Then additional checks will be 5423 performed on the chain. Currently the root CA is checked to see 5424 if it is explicitly trusted for OCSP signing. This is used to set 5425 a root CA as a global signing root: that is any certificate that 5426 chains to that CA is an acceptable OCSP signing certificate. 5427 [Steve Henson] 5428 5429 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5430 extensions from a separate configuration file. 5431 As when reading extensions from the main configuration file, 5432 the '-extensions ...' option may be used for specifying the 5433 section to use. 5434 [Massimiliano Pala <madwolf@comune.modena.it>] 5435 5436 *) New OCSP utility. Allows OCSP requests to be generated or 5437 read. The request can be sent to a responder and the output 5438 parsed, outputed or printed in text form. Not complete yet: 5439 still needs to check the OCSP response validity. 5440 [Steve Henson] 5441 5442 *) New subcommands for 'openssl ca': 5443 'openssl ca -status <serial>' prints the status of the cert with 5444 the given serial number (according to the index file). 5445 'openssl ca -updatedb' updates the expiry status of certificates 5446 in the index file. 5447 [Massimiliano Pala <madwolf@comune.modena.it>] 5448 5449 *) New '-newreq-nodes' command option to CA.pl. This is like 5450 '-newreq', but calls 'openssl req' with the '-nodes' option 5451 so that the resulting key is not encrypted. 5452 [Damien Miller <djm@mindrot.org>] 5453 5454 *) New configuration for the GNU Hurd. 5455 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5456 5457 *) Initial code to implement OCSP basic response verify. This 5458 is currently incomplete. Currently just finds the signer's 5459 certificate and verifies the signature on the response. 5460 [Steve Henson] 5461 5462 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5463 value of OPENSSLDIR. This is available via the new '-d' option 5464 to 'openssl version', and is also included in 'openssl version -a'. 5465 [Bodo Moeller] 5466 5467 *) Allowing defining memory allocation callbacks that will be given 5468 file name and line number information in additional arguments 5469 (a const char* and an int). The basic functionality remains, as 5470 well as the original possibility to just replace malloc(), 5471 realloc() and free() by functions that do not know about these 5472 additional arguments. To register and find out the current 5473 settings for extended allocation functions, the following 5474 functions are provided: 5475 5476 CRYPTO_set_mem_ex_functions 5477 CRYPTO_set_locked_mem_ex_functions 5478 CRYPTO_get_mem_ex_functions 5479 CRYPTO_get_locked_mem_ex_functions 5480 5481 These work the same way as CRYPTO_set_mem_functions and friends. 5482 CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5483 extended allocation function is enabled. 5484 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5485 a conventional allocation function is enabled. 5486 [Richard Levitte, Bodo Moeller] 5487 5488 *) Finish off removing the remaining LHASH function pointer casts. 5489 There should no longer be any prototype-casting required when using 5490 the LHASH abstraction, and any casts that remain are "bugs". See 5491 the callback types and macros at the head of lhash.h for details 5492 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5493 [Geoff Thorpe] 5494 5495 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5496 If /dev/[u]random devices are not available or do not return enough 5497 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5498 be queried. 5499 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5500 /etc/entropy will be queried once each in this sequence, quering stops 5501 when enough entropy was collected without querying more sockets. 5502 [Lutz Jaenicke] 5503 5504 *) Change the Unix RAND_poll() variant to be able to poll several 5505 random devices, as specified by DEVRANDOM, until a sufficient amount 5506 of data has been collected. We spend at most 10 ms on each file 5507 (select timeout) and read in non-blocking mode. DEVRANDOM now 5508 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5509 (previously it was just the string "/dev/urandom"), so on typical 5510 platforms the 10 ms delay will never occur. 5511 Also separate out the Unix variant to its own file, rand_unix.c. 5512 For VMS, there's a currently-empty rand_vms.c. 5513 [Richard Levitte] 5514 5515 *) Move OCSP client related routines to ocsp_cl.c. These 5516 provide utility functions which an application needing 5517 to issue a request to an OCSP responder and analyse the 5518 response will typically need: as opposed to those which an 5519 OCSP responder itself would need which will be added later. 5520 5521 OCSP_request_sign() signs an OCSP request with an API similar 5522 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5523 response. OCSP_response_get1_basic() extracts basic response 5524 from response. OCSP_resp_find_status(): finds and extracts status 5525 information from an OCSP_CERTID structure (which will be created 5526 when the request structure is built). These are built from lower 5527 level functions which work on OCSP_SINGLERESP structures but 5528 wont normally be used unless the application wishes to examine 5529 extensions in the OCSP response for example. 5530 5531 Replace nonce routines with a pair of functions. 5532 OCSP_request_add1_nonce() adds a nonce value and optionally 5533 generates a random value. OCSP_check_nonce() checks the 5534 validity of the nonce in an OCSP response. 5535 [Steve Henson] 5536 5537 *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5538 This doesn't copy the supplied OCSP_CERTID and avoids the 5539 need to free up the newly created id. Change return type 5540 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5541 This can then be used to add extensions to the request. 5542 Deleted OCSP_request_new(), since most of its functionality 5543 is now in OCSP_REQUEST_new() (and the case insensitive name 5544 clash) apart from the ability to set the request name which 5545 will be added elsewhere. 5546 [Steve Henson] 5547 5548 *) Update OCSP API. Remove obsolete extensions argument from 5549 various functions. Extensions are now handled using the new 5550 OCSP extension code. New simple OCSP HTTP function which 5551 can be used to send requests and parse the response. 5552 [Steve Henson] 5553 5554 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5555 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5556 uses the special reorder version of SET OF to sort the attributes 5557 and reorder them to match the encoded order. This resolves a long 5558 standing problem: a verify on a PKCS7 structure just after signing 5559 it used to fail because the attribute order did not match the 5560 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5561 it uses the received order. This is necessary to tolerate some broken 5562 software that does not order SET OF. This is handled by encoding 5563 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5564 to produce the required SET OF. 5565 [Steve Henson] 5566 5567 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5568 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5569 files to get correct declarations of the ASN.1 item variables. 5570 [Richard Levitte] 5571 5572 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5573 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5574 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5575 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5576 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5577 ASN1_ITEM and no wrapper functions. 5578 [Steve Henson] 5579 5580 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5581 replace the old function pointer based I/O routines. Change most of 5582 the *_d2i_bio() and *_d2i_fp() functions to use these. 5583 [Steve Henson] 5584 5585 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5586 lines, recognice more "algorithms" that can be deselected, and make 5587 it complain about algorithm deselection that isn't recognised. 5588 [Richard Levitte] 5589 5590 *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5591 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5592 to use new functions. Add NO_ASN1_OLD which can be set to remove 5593 some old style ASN1 functions: this can be used to determine if old 5594 code will still work when these eventually go away. 5595 [Steve Henson] 5596 5597 *) New extension functions for OCSP structures, these follow the 5598 same conventions as certificates and CRLs. 5599 [Steve Henson] 5600 5601 *) New function X509V3_add1_i2d(). This automatically encodes and 5602 adds an extension. Its behaviour can be customised with various 5603 flags to append, replace or delete. Various wrappers added for 5604 certifcates and CRLs. 5605 [Steve Henson] 5606 5607 *) Fix to avoid calling the underlying ASN1 print routine when 5608 an extension cannot be parsed. Correct a typo in the 5609 OCSP_SERVICELOC extension. Tidy up print OCSP format. 5610 [Steve Henson] 5611 5612 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5613 entries for variables. 5614 [Steve Henson] 5615 5616 *) Add functionality to apps/openssl.c for detecting locking 5617 problems: As the program is single-threaded, all we have 5618 to do is register a locking callback using an array for 5619 storing which locks are currently held by the program. 5620 [Bodo Moeller] 5621 5622 *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5623 SSL_get_ex_data_X509_STORE_idx(), which is used in 5624 ssl_verify_cert_chain() and thus can be called at any time 5625 during TLS/SSL handshakes so that thread-safety is essential. 5626 Unfortunately, the ex_data design is not at all suited 5627 for multi-threaded use, so it probably should be abolished. 5628 [Bodo Moeller] 5629 5630 *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5631 [Broadcom, tweaked and integrated by Geoff Thorpe] 5632 5633 *) Move common extension printing code to new function 5634 X509V3_print_extensions(). Reorganise OCSP print routines and 5635 implement some needed OCSP ASN1 functions. Add OCSP extensions. 5636 [Steve Henson] 5637 5638 *) New function X509_signature_print() to remove duplication in some 5639 print routines. 5640 [Steve Henson] 5641 5642 *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5643 set (this was treated exactly the same as SET OF previously). This 5644 is used to reorder the STACK representing the structure to match the 5645 encoding. This will be used to get round a problem where a PKCS7 5646 structure which was signed could not be verified because the STACK 5647 order did not reflect the encoded order. 5648 [Steve Henson] 5649 5650 *) Reimplement the OCSP ASN1 module using the new code. 5651 [Steve Henson] 5652 5653 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5654 for its ASN1 operations. The old style function pointers still exist 5655 for now but they will eventually go away. 5656 [Steve Henson] 5657 5658 *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5659 completely replaces the old ASN1 functionality with a table driven 5660 encoder and decoder which interprets an ASN1_ITEM structure describing 5661 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5662 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5663 has also been converted to the new form. 5664 [Steve Henson] 5665 5666 *) Change BN_mod_exp_recp so that negative moduli are tolerated 5667 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5668 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5669 for negative moduli. 5670 [Bodo Moeller] 5671 5672 *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5673 of not touching the result's sign bit. 5674 [Bodo Moeller] 5675 5676 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5677 set. 5678 [Bodo Moeller] 5679 5680 *) Changed the LHASH code to use prototypes for callbacks, and created 5681 macros to declare and implement thin (optionally static) functions 5682 that provide type-safety and avoid function pointer casting for the 5683 type-specific callbacks. 5684 [Geoff Thorpe] 5685 5686 *) Added Kerberos Cipher Suites to be used with TLS, as written in 5687 RFC 2712. 5688 [Veers Staats <staatsvr@asc.hpc.mil>, 5689 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5690 5691 *) Reformat the FAQ so the different questions and answers can be divided 5692 in sections depending on the subject. 5693 [Richard Levitte] 5694 5695 *) Have the zlib compression code load ZLIB.DLL dynamically under 5696 Windows. 5697 [Richard Levitte] 5698 5699 *) New function BN_mod_sqrt for computing square roots modulo a prime 5700 (using the probabilistic Tonelli-Shanks algorithm unless 5701 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5702 be handled deterministically). 5703 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5704 5705 *) Make BN_mod_inverse faster by explicitly handling small quotients 5706 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5707 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5708 [Bodo Moeller] 5709 5710 *) New function BN_kronecker. 5711 [Bodo Moeller] 5712 5713 *) Fix BN_gcd so that it works on negative inputs; the result is 5714 positive unless both parameters are zero. 5715 Previously something reasonably close to an infinite loop was 5716 possible because numbers could be growing instead of shrinking 5717 in the implementation of Euclid's algorithm. 5718 [Bodo Moeller] 5719 5720 *) Fix BN_is_word() and BN_is_one() macros to take into account the 5721 sign of the number in question. 5722 5723 Fix BN_is_word(a,w) to work correctly for w == 0. 5724 5725 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5726 because its test if the absolute value of 'a' equals 'w'. 5727 Note that BN_abs_is_word does *not* handle w == 0 reliably; 5728 it exists mostly for use in the implementations of BN_is_zero(), 5729 BN_is_one(), and BN_is_word(). 5730 [Bodo Moeller] 5731 5732 *) New function BN_swap. 5733 [Bodo Moeller] 5734 5735 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5736 the exponentiation functions are more likely to produce reasonable 5737 results on negative inputs. 5738 [Bodo Moeller] 5739 5740 *) Change BN_mod_mul so that the result is always non-negative. 5741 Previously, it could be negative if one of the factors was negative; 5742 I don't think anyone really wanted that behaviour. 5743 [Bodo Moeller] 5744 5745 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5746 (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5747 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5748 and add new functions: 5749 5750 BN_nnmod 5751 BN_mod_sqr 5752 BN_mod_add 5753 BN_mod_add_quick 5754 BN_mod_sub 5755 BN_mod_sub_quick 5756 BN_mod_lshift1 5757 BN_mod_lshift1_quick 5758 BN_mod_lshift 5759 BN_mod_lshift_quick 5760 5761 These functions always generate non-negative results. 5762 5763 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5764 such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5765 5766 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5767 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5768 be reduced modulo m. 5769 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5770 5771#if 0 5772 The following entry accidentily appeared in the CHANGES file 5773 distributed with OpenSSL 0.9.7. The modifications described in 5774 it do *not* apply to OpenSSL 0.9.7. 5775 5776 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5777 was actually never needed) and in BN_mul(). The removal in BN_mul() 5778 required a small change in bn_mul_part_recursive() and the addition 5779 of the functions bn_cmp_part_words(), bn_sub_part_words() and 5780 bn_add_part_words(), which do the same thing as bn_cmp_words(), 5781 bn_sub_words() and bn_add_words() except they take arrays with 5782 differing sizes. 5783 [Richard Levitte] 5784#endif 5785 5786 *) In 'openssl passwd', verify passwords read from the terminal 5787 unless the '-salt' option is used (which usually means that 5788 verification would just waste user's time since the resulting 5789 hash is going to be compared with some given password hash) 5790 or the new '-noverify' option is used. 5791 5792 This is an incompatible change, but it does not affect 5793 non-interactive use of 'openssl passwd' (passwords on the command 5794 line, '-stdin' option, '-in ...' option) and thus should not 5795 cause any problems. 5796 [Bodo Moeller] 5797 5798 *) Remove all references to RSAref, since there's no more need for it. 5799 [Richard Levitte] 5800 5801 *) Make DSO load along a path given through an environment variable 5802 (SHLIB_PATH) with shl_load(). 5803 [Richard Levitte] 5804 5805 *) Constify the ENGINE code as a result of BIGNUM constification. 5806 Also constify the RSA code and most things related to it. In a 5807 few places, most notable in the depth of the ASN.1 code, ugly 5808 casts back to non-const were required (to be solved at a later 5809 time) 5810 [Richard Levitte] 5811 5812 *) Make it so the openssl application has all engines loaded by default. 5813 [Richard Levitte] 5814 5815 *) Constify the BIGNUM routines a little more. 5816 [Richard Levitte] 5817 5818 *) Add the following functions: 5819 5820 ENGINE_load_cswift() 5821 ENGINE_load_chil() 5822 ENGINE_load_atalla() 5823 ENGINE_load_nuron() 5824 ENGINE_load_builtin_engines() 5825 5826 That way, an application can itself choose if external engines that 5827 are built-in in OpenSSL shall ever be used or not. The benefit is 5828 that applications won't have to be linked with libdl or other dso 5829 libraries unless it's really needed. 5830 5831 Changed 'openssl engine' to load all engines on demand. 5832 Changed the engine header files to avoid the duplication of some 5833 declarations (they differed!). 5834 [Richard Levitte] 5835 5836 *) 'openssl engine' can now list capabilities. 5837 [Richard Levitte] 5838 5839 *) Better error reporting in 'openssl engine'. 5840 [Richard Levitte] 5841 5842 *) Never call load_dh_param(NULL) in s_server. 5843 [Bodo Moeller] 5844 5845 *) Add engine application. It can currently list engines by name and 5846 identity, and test if they are actually available. 5847 [Richard Levitte] 5848 5849 *) Improve RPM specification file by forcing symbolic linking and making 5850 sure the installed documentation is also owned by root.root. 5851 [Damien Miller <djm@mindrot.org>] 5852 5853 *) Give the OpenSSL applications more possibilities to make use of 5854 keys (public as well as private) handled by engines. 5855 [Richard Levitte] 5856 5857 *) Add OCSP code that comes from CertCo. 5858 [Richard Levitte] 5859 5860 *) Add VMS support for the Rijndael code. 5861 [Richard Levitte] 5862 5863 *) Added untested support for Nuron crypto accelerator. 5864 [Ben Laurie] 5865 5866 *) Add support for external cryptographic devices. This code was 5867 previously distributed separately as the "engine" branch. 5868 [Geoff Thorpe, Richard Levitte] 5869 5870 *) Rework the filename-translation in the DSO code. It is now possible to 5871 have far greater control over how a "name" is turned into a filename 5872 depending on the operating environment and any oddities about the 5873 different shared library filenames on each system. 5874 [Geoff Thorpe] 5875 5876 *) Support threads on FreeBSD-elf in Configure. 5877 [Richard Levitte] 5878 5879 *) Fix for SHA1 assembly problem with MASM: it produces 5880 warnings about corrupt line number information when assembling 5881 with debugging information. This is caused by the overlapping 5882 of two sections. 5883 [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 5884 5885 *) NCONF changes. 5886 NCONF_get_number() has no error checking at all. As a replacement, 5887 NCONF_get_number_e() is defined (_e for "error checking") and is 5888 promoted strongly. The old NCONF_get_number is kept around for 5889 binary backward compatibility. 5890 Make it possible for methods to load from something other than a BIO, 5891 by providing a function pointer that is given a name instead of a BIO. 5892 For example, this could be used to load configuration data from an 5893 LDAP server. 5894 [Richard Levitte] 5895 5896 *) Fix for non blocking accept BIOs. Added new I/O special reason 5897 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 5898 with non blocking I/O was not possible because no retry code was 5899 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 5900 this case. 5901 [Steve Henson] 5902 5903 *) Added the beginnings of Rijndael support. 5904 [Ben Laurie] 5905 5906 *) Fix for bug in DirectoryString mask setting. Add support for 5907 X509_NAME_print_ex() in 'req' and X509_print_ex() function 5908 to allow certificate printing to more controllable, additional 5909 'certopt' option to 'x509' to allow new printing options to be 5910 set. 5911 [Steve Henson] 5912 5913 *) Clean old EAY MD5 hack from e_os.h. 5914 [Richard Levitte] 5915 5916 Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 5917 5918 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 5919 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 5920 [Joe Orton, Steve Henson] 5921 5922 Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 5923 5924 *) Fix additional bug revealed by the NISCC test suite: 5925 5926 Stop bug triggering large recursion when presented with 5927 certain ASN.1 tags (CVE-2003-0851) 5928 [Steve Henson] 5929 5930 Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 5931 5932 *) Fix various bugs revealed by running the NISCC test suite: 5933 5934 Stop out of bounds reads in the ASN1 code when presented with 5935 invalid tags (CVE-2003-0543 and CVE-2003-0544). 5936 5937 If verify callback ignores invalid public key errors don't try to check 5938 certificate signature with the NULL public key. 5939 5940 [Steve Henson] 5941 5942 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 5943 if the server requested one: as stated in TLS 1.0 and SSL 3.0 5944 specifications. 5945 [Steve Henson] 5946 5947 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 5948 extra data after the compression methods not only for TLS 1.0 5949 but also for SSL 3.0 (as required by the specification). 5950 [Bodo Moeller; problem pointed out by Matthias Loepfe] 5951 5952 *) Change X509_certificate_type() to mark the key as exported/exportable 5953 when it's 512 *bits* long, not 512 bytes. 5954 [Richard Levitte] 5955 5956 Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 5957 5958 *) Countermeasure against the Klima-Pokorny-Rosa extension of 5959 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 5960 a protocol version number mismatch like a decryption error 5961 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 5962 [Bodo Moeller] 5963 5964 *) Turn on RSA blinding by default in the default implementation 5965 to avoid a timing attack. Applications that don't want it can call 5966 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 5967 They would be ill-advised to do so in most cases. 5968 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 5969 5970 *) Change RSA blinding code so that it works when the PRNG is not 5971 seeded (in this case, the secret RSA exponent is abused as 5972 an unpredictable seed -- if it is not unpredictable, there 5973 is no point in blinding anyway). Make RSA blinding thread-safe 5974 by remembering the creator's thread ID in rsa->blinding and 5975 having all other threads use local one-time blinding factors 5976 (this requires more computation than sharing rsa->blinding, but 5977 avoids excessive locking; and if an RSA object is not shared 5978 between threads, blinding will still be very fast). 5979 [Bodo Moeller] 5980 5981 Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 5982 5983 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 5984 via timing by performing a MAC computation even if incorrrect 5985 block cipher padding has been found. This is a countermeasure 5986 against active attacks where the attacker has to distinguish 5987 between bad padding and a MAC verification error. (CVE-2003-0078) 5988 5989 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 5990 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 5991 Martin Vuagnoux (EPFL, Ilion)] 5992 5993 Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 5994 5995 *) New function OPENSSL_cleanse(), which is used to cleanse a section of 5996 memory from it's contents. This is done with a counter that will 5997 place alternating values in each byte. This can be used to solve 5998 two issues: 1) the removal of calls to memset() by highly optimizing 5999 compilers, and 2) cleansing with other values than 0, since those can 6000 be read through on certain media, for example a swap space on disk. 6001 [Geoff Thorpe] 6002 6003 *) Bugfix: client side session caching did not work with external caching, 6004 because the session->cipher setting was not restored when reloading 6005 from the external cache. This problem was masked, when 6006 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 6007 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 6008 [Lutz Jaenicke] 6009 6010 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6011 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6012 [Zeev Lieber <zeev-l@yahoo.com>] 6013 6014 *) Undo an undocumented change introduced in 0.9.6e which caused 6015 repeated calls to OpenSSL_add_all_ciphers() and 6016 OpenSSL_add_all_digests() to be ignored, even after calling 6017 EVP_cleanup(). 6018 [Richard Levitte] 6019 6020 *) Change the default configuration reader to deal with last line not 6021 being properly terminated. 6022 [Richard Levitte] 6023 6024 *) Change X509_NAME_cmp() so it applies the special rules on handling 6025 DN values that are of type PrintableString, as well as RDNs of type 6026 emailAddress where the value has the type ia5String. 6027 [stefank@valicert.com via Richard Levitte] 6028 6029 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6030 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6031 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6032 the bitwise-OR of the two for use by the majority of applications 6033 wanting this behaviour, and update the docs. The documented 6034 behaviour and actual behaviour were inconsistent and had been 6035 changing anyway, so this is more a bug-fix than a behavioural 6036 change. 6037 [Geoff Thorpe, diagnosed by Nadav Har'El] 6038 6039 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 6040 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 6041 [Bodo Moeller] 6042 6043 *) Fix initialization code race conditions in 6044 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 6045 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 6046 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 6047 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 6048 ssl2_get_cipher_by_char(), 6049 ssl3_get_cipher_by_char(). 6050 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 6051 6052 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 6053 the cached sessions are flushed, as the remove_cb() might use ex_data 6054 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 6055 (see [openssl.org #212]). 6056 [Geoff Thorpe, Lutz Jaenicke] 6057 6058 *) Fix typo in OBJ_txt2obj which incorrectly passed the content 6059 length, instead of the encoding length to d2i_ASN1_OBJECT. 6060 [Steve Henson] 6061 6062 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6063 6064 *) [In 0.9.6g-engine release:] 6065 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6066 [Lynn Gazis <lgazis@rainbow.com>] 6067 6068 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6069 6070 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6071 and get fix the header length calculation. 6072 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6073 Alon Kantor <alonk@checkpoint.com> (and others), 6074 Steve Henson] 6075 6076 *) Use proper error handling instead of 'assertions' in buffer 6077 overflow checks added in 0.9.6e. This prevents DoS (the 6078 assertions could call abort()). 6079 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6080 6081 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6082 6083 *) Add various sanity checks to asn1_get_length() to reject 6084 the ASN1 length bytes if they exceed sizeof(long), will appear 6085 negative or the content length exceeds the length of the 6086 supplied buffer. 6087 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6088 6089 *) Fix cipher selection routines: ciphers without encryption had no flags 6090 for the cipher strength set and where therefore not handled correctly 6091 by the selection routines (PR #130). 6092 [Lutz Jaenicke] 6093 6094 *) Fix EVP_dsa_sha macro. 6095 [Nils Larsch] 6096 6097 *) New option 6098 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6099 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6100 that was added in OpenSSL 0.9.6d. 6101 6102 As the countermeasure turned out to be incompatible with some 6103 broken SSL implementations, the new option is part of SSL_OP_ALL. 6104 SSL_OP_ALL is usually employed when compatibility with weird SSL 6105 implementations is desired (e.g. '-bugs' option to 's_client' and 6106 's_server'), so the new option is automatically set in many 6107 applications. 6108 [Bodo Moeller] 6109 6110 *) Changes in security patch: 6111 6112 Changes marked "(CHATS)" were sponsored by the Defense Advanced 6113 Research Projects Agency (DARPA) and Air Force Research Laboratory, 6114 Air Force Materiel Command, USAF, under agreement number 6115 F30602-01-2-0537. 6116 6117 *) Add various sanity checks to asn1_get_length() to reject 6118 the ASN1 length bytes if they exceed sizeof(long), will appear 6119 negative or the content length exceeds the length of the 6120 supplied buffer. (CVE-2002-0659) 6121 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6122 6123 *) Assertions for various potential buffer overflows, not known to 6124 happen in practice. 6125 [Ben Laurie (CHATS)] 6126 6127 *) Various temporary buffers to hold ASCII versions of integers were 6128 too small for 64 bit platforms. (CVE-2002-0655) 6129 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6130 6131 *) Remote buffer overflow in SSL3 protocol - an attacker could 6132 supply an oversized session ID to a client. (CVE-2002-0656) 6133 [Ben Laurie (CHATS)] 6134 6135 *) Remote buffer overflow in SSL2 protocol - an attacker could 6136 supply an oversized client master key. (CVE-2002-0656) 6137 [Ben Laurie (CHATS)] 6138 6139 Changes between 0.9.6c and 0.9.6d [9 May 2002] 6140 6141 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6142 encoded as NULL) with id-dsa-with-sha1. 6143 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6144 6145 *) Check various X509_...() return values in apps/req.c. 6146 [Nils Larsch <nla@trustcenter.de>] 6147 6148 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6149 an end-of-file condition would erronously be flagged, when the CRLF 6150 was just at the end of a processed block. The bug was discovered when 6151 processing data through a buffering memory BIO handing the data to a 6152 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6153 <ptsekov@syntrex.com> and Nedelcho Stanev. 6154 [Lutz Jaenicke] 6155 6156 *) Implement a countermeasure against a vulnerability recently found 6157 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6158 before application data chunks to avoid the use of known IVs 6159 with data potentially chosen by the attacker. 6160 [Bodo Moeller] 6161 6162 *) Fix length checks in ssl3_get_client_hello(). 6163 [Bodo Moeller] 6164 6165 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6166 to prevent ssl3_read_internal() from incorrectly assuming that 6167 ssl3_read_bytes() found application data while handshake 6168 processing was enabled when in fact s->s3->in_read_app_data was 6169 merely automatically cleared during the initial handshake. 6170 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6171 6172 *) Fix object definitions for Private and Enterprise: they were not 6173 recognized in their shortname (=lowercase) representation. Extend 6174 obj_dat.pl to issue an error when using undefined keywords instead 6175 of silently ignoring the problem (Svenning Sorensen 6176 <sss@sss.dnsalias.net>). 6177 [Lutz Jaenicke] 6178 6179 *) Fix DH_generate_parameters() so that it works for 'non-standard' 6180 generators, i.e. generators other than 2 and 5. (Previously, the 6181 code did not properly initialise the 'add' and 'rem' values to 6182 BN_generate_prime().) 6183 6184 In the new general case, we do not insist that 'generator' is 6185 actually a primitive root: This requirement is rather pointless; 6186 a generator of the order-q subgroup is just as good, if not 6187 better. 6188 [Bodo Moeller] 6189 6190 *) Map new X509 verification errors to alerts. Discovered and submitted by 6191 Tom Wu <tom@arcot.com>. 6192 [Lutz Jaenicke] 6193 6194 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6195 returning non-zero before the data has been completely received 6196 when using non-blocking I/O. 6197 [Bodo Moeller; problem pointed out by John Hughes] 6198 6199 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6200 [Ben Laurie, Lutz Jaenicke] 6201 6202 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6203 Yoram Zahavi <YoramZ@gilian.com>). 6204 [Lutz Jaenicke] 6205 6206 *) Add information about CygWin 1.3 and on, and preserve proper 6207 configuration for the versions before that. 6208 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6209 6210 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6211 check whether we deal with a copy of a session and do not delete from 6212 the cache in this case. Problem reported by "Izhar Shoshani Levi" 6213 <izhar@checkpoint.com>. 6214 [Lutz Jaenicke] 6215 6216 *) Do not store session data into the internal session cache, if it 6217 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6218 flag is set). Proposed by Aslam <aslam@funk.com>. 6219 [Lutz Jaenicke] 6220 6221 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6222 value is 0. 6223 [Richard Levitte] 6224 6225 *) [In 0.9.6d-engine release:] 6226 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6227 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6228 6229 *) Add the configuration target linux-s390x. 6230 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6231 6232 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6233 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6234 variable as an indication that a ClientHello message has been 6235 received. As the flag value will be lost between multiple 6236 invocations of ssl3_accept when using non-blocking I/O, the 6237 function may not be aware that a handshake has actually taken 6238 place, thus preventing a new session from being added to the 6239 session cache. 6240 6241 To avoid this problem, we now set s->new_session to 2 instead of 6242 using a local variable. 6243 [Lutz Jaenicke, Bodo Moeller] 6244 6245 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6246 if the SSL_R_LENGTH_MISMATCH error is detected. 6247 [Geoff Thorpe, Bodo Moeller] 6248 6249 *) New 'shared_ldflag' column in Configure platform table. 6250 [Richard Levitte] 6251 6252 *) Fix EVP_CIPHER_mode macro. 6253 ["Dan S. Camper" <dan@bti.net>] 6254 6255 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6256 type, we must throw them away by setting rr->length to 0. 6257 [D P Chang <dpc@qualys.com>] 6258 6259 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 6260 6261 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 6262 <Dominikus.Scherkl@biodata.com>. (The previous implementation 6263 worked incorrectly for those cases where range = 10..._2 and 6264 3*range is two bits longer than range.) 6265 [Bodo Moeller] 6266 6267 *) Only add signing time to PKCS7 structures if it is not already 6268 present. 6269 [Steve Henson] 6270 6271 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 6272 OBJ_ld_ce should be OBJ_id_ce. 6273 Also some ip-pda OIDs in crypto/objects/objects.txt were 6274 incorrect (cf. RFC 3039). 6275 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 6276 6277 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 6278 returns early because it has nothing to do. 6279 [Andy Schneider <andy.schneider@bjss.co.uk>] 6280 6281 *) [In 0.9.6c-engine release:] 6282 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 6283 [Andy Schneider <andy.schneider@bjss.co.uk>] 6284 6285 *) [In 0.9.6c-engine release:] 6286 Add support for Cryptographic Appliance's keyserver technology. 6287 (Use engine 'keyclient') 6288 [Cryptographic Appliances and Geoff Thorpe] 6289 6290 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 6291 is called via tools/c89.sh because arguments have to be 6292 rearranged (all '-L' options must appear before the first object 6293 modules). 6294 [Richard Shapiro <rshapiro@abinitio.com>] 6295 6296 *) [In 0.9.6c-engine release:] 6297 Add support for Broadcom crypto accelerator cards, backported 6298 from 0.9.7. 6299 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 6300 6301 *) [In 0.9.6c-engine release:] 6302 Add support for SureWare crypto accelerator cards from 6303 Baltimore Technologies. (Use engine 'sureware') 6304 [Baltimore Technologies and Mark Cox] 6305 6306 *) [In 0.9.6c-engine release:] 6307 Add support for crypto accelerator cards from Accelerated 6308 Encryption Processing, www.aep.ie. (Use engine 'aep') 6309 [AEP Inc. and Mark Cox] 6310 6311 *) Add a configuration entry for gcc on UnixWare. 6312 [Gary Benson <gbenson@redhat.com>] 6313 6314 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 6315 messages are stored in a single piece (fixed-length part and 6316 variable-length part combined) and fix various bugs found on the way. 6317 [Bodo Moeller] 6318 6319 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 6320 instead. BIO_gethostbyname() does not know what timeouts are 6321 appropriate, so entries would stay in cache even when they have 6322 become invalid. 6323 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 6324 6325 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 6326 faced with a pathologically small ClientHello fragment that does 6327 not contain client_version: Instead of aborting with an error, 6328 simply choose the highest available protocol version (i.e., 6329 TLS 1.0 unless it is disabled). In practice, ClientHello 6330 messages are never sent like this, but this change gives us 6331 strictly correct behaviour at least for TLS. 6332 [Bodo Moeller] 6333 6334 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 6335 never resets s->method to s->ctx->method when called from within 6336 one of the SSL handshake functions. 6337 [Bodo Moeller; problem pointed out by Niko Baric] 6338 6339 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 6340 (sent using the client's version number) if client_version is 6341 smaller than the protocol version in use. Also change 6342 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 6343 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 6344 the client will at least see that alert. 6345 [Bodo Moeller] 6346 6347 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 6348 correctly. 6349 [Bodo Moeller] 6350 6351 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 6352 client receives HelloRequest while in a handshake. 6353 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 6354 6355 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 6356 should end in 'break', not 'goto end' which circuments various 6357 cleanups done in state SSL_ST_OK. But session related stuff 6358 must be disabled for SSL_ST_OK in the case that we just sent a 6359 HelloRequest. 6360 6361 Also avoid some overhead by not calling ssl_init_wbio_buffer() 6362 before just sending a HelloRequest. 6363 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 6364 6365 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 6366 reveal whether illegal block cipher padding was found or a MAC 6367 verification error occured. (Neither SSLerr() codes nor alerts 6368 are directly visible to potential attackers, but the information 6369 may leak via logfiles.) 6370 6371 Similar changes are not required for the SSL 2.0 implementation 6372 because the number of padding bytes is sent in clear for SSL 2.0, 6373 and the extra bytes are just ignored. However ssl/s2_pkt.c 6374 failed to verify that the purported number of padding bytes is in 6375 the legal range. 6376 [Bodo Moeller] 6377 6378 *) Add OpenUNIX-8 support including shared libraries 6379 (Boyd Lynn Gerber <gerberb@zenez.com>). 6380 [Lutz Jaenicke] 6381 6382 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 6383 'wristwatch attack' using huge encoding parameters (cf. 6384 James H. Manger's CRYPTO 2001 paper). Note that the 6385 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 6386 encoding parameters and hence was not vulnerable. 6387 [Bodo Moeller] 6388 6389 *) BN_sqr() bug fix. 6390 [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 6391 6392 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 6393 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 6394 followed by modular reduction. 6395 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 6396 6397 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 6398 equivalent based on BN_pseudo_rand() instead of BN_rand(). 6399 [Bodo Moeller] 6400 6401 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 6402 This function was broken, as the check for a new client hello message 6403 to handle SGC did not allow these large messages. 6404 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 6405 [Lutz Jaenicke] 6406 6407 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 6408 [Lutz Jaenicke] 6409 6410 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 6411 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 6412 [Lutz Jaenicke] 6413 6414 *) Rework the configuration and shared library support for Tru64 Unix. 6415 The configuration part makes use of modern compiler features and 6416 still retains old compiler behavior for those that run older versions 6417 of the OS. The shared library support part includes a variant that 6418 uses the RPATH feature, and is available through the special 6419 configuration target "alpha-cc-rpath", which will never be selected 6420 automatically. 6421 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 6422 6423 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 6424 with the same message size as in ssl3_get_certificate_request(). 6425 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 6426 messages might inadvertently be reject as too long. 6427 [Petr Lampa <lampa@fee.vutbr.cz>] 6428 6429 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 6430 [Andy Polyakov] 6431 6432 *) Modified SSL library such that the verify_callback that has been set 6433 specificly for an SSL object with SSL_set_verify() is actually being 6434 used. Before the change, a verify_callback set with this function was 6435 ignored and the verify_callback() set in the SSL_CTX at the time of 6436 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 6437 to allow the necessary settings. 6438 [Lutz Jaenicke] 6439 6440 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 6441 explicitly to NULL, as at least on Solaris 8 this seems not always to be 6442 done automatically (in contradiction to the requirements of the C 6443 standard). This made problems when used from OpenSSH. 6444 [Lutz Jaenicke] 6445 6446 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 6447 dh->length and always used 6448 6449 BN_rand_range(priv_key, dh->p). 6450 6451 BN_rand_range() is not necessary for Diffie-Hellman, and this 6452 specific range makes Diffie-Hellman unnecessarily inefficient if 6453 dh->length (recommended exponent length) is much smaller than the 6454 length of dh->p. We could use BN_rand_range() if the order of 6455 the subgroup was stored in the DH structure, but we only have 6456 dh->length. 6457 6458 So switch back to 6459 6460 BN_rand(priv_key, l, ...) 6461 6462 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 6463 otherwise. 6464 [Bodo Moeller] 6465 6466 *) In 6467 6468 RSA_eay_public_encrypt 6469 RSA_eay_private_decrypt 6470 RSA_eay_private_encrypt (signing) 6471 RSA_eay_public_decrypt (signature verification) 6472 6473 (default implementations for RSA_public_encrypt, 6474 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 6475 always reject numbers >= n. 6476 [Bodo Moeller] 6477 6478 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 6479 to synchronize access to 'locking_thread'. This is necessary on 6480 systems where access to 'locking_thread' (an 'unsigned long' 6481 variable) is not atomic. 6482 [Bodo Moeller] 6483 6484 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 6485 *before* setting the 'crypto_lock_rand' flag. The previous code had 6486 a race condition if 0 is a valid thread ID. 6487 [Travis Vitek <vitek@roguewave.com>] 6488 6489 *) Add support for shared libraries under Irix. 6490 [Albert Chin-A-Young <china@thewrittenword.com>] 6491 6492 *) Add configuration option to build on Linux on both big-endian and 6493 little-endian MIPS. 6494 [Ralf Baechle <ralf@uni-koblenz.de>] 6495 6496 *) Add the possibility to create shared libraries on HP-UX. 6497 [Richard Levitte] 6498 6499 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 6500 6501 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 6502 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 6503 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 6504 PRNG state recovery was possible based on the output of 6505 one PRNG request appropriately sized to gain knowledge on 6506 'md' followed by enough consecutive 1-byte PRNG requests 6507 to traverse all of 'state'. 6508 6509 1. When updating 'md_local' (the current thread's copy of 'md') 6510 during PRNG output generation, hash all of the previous 6511 'md_local' value, not just the half used for PRNG output. 6512 6513 2. Make the number of bytes from 'state' included into the hash 6514 independent from the number of PRNG bytes requested. 6515 6516 The first measure alone would be sufficient to avoid 6517 Markku-Juhani's attack. (Actually it had never occurred 6518 to me that the half of 'md_local' used for chaining was the 6519 half from which PRNG output bytes were taken -- I had always 6520 assumed that the secret half would be used.) The second 6521 measure makes sure that additional data from 'state' is never 6522 mixed into 'md_local' in small portions; this heuristically 6523 further strengthens the PRNG. 6524 [Bodo Moeller] 6525 6526 *) Fix crypto/bn/asm/mips3.s. 6527 [Andy Polyakov] 6528 6529 *) When only the key is given to "enc", the IV is undefined. Print out 6530 an error message in this case. 6531 [Lutz Jaenicke] 6532 6533 *) Handle special case when X509_NAME is empty in X509 printing routines. 6534 [Steve Henson] 6535 6536 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 6537 positive and less than q. 6538 [Bodo Moeller] 6539 6540 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 6541 used: it isn't thread safe and the add_lock_callback should handle 6542 that itself. 6543 [Paul Rose <Paul.Rose@bridge.com>] 6544 6545 *) Verify that incoming data obeys the block size in 6546 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 6547 [Bodo Moeller] 6548 6549 *) Fix OAEP check. 6550 [Ulf M�ller, Bodo M�ller] 6551 6552 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 6553 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 6554 when fixing the server behaviour for backwards-compatible 'client 6555 hello' messages. (Note that the attack is impractical against 6556 SSL 3.0 and TLS 1.0 anyway because length and version checking 6557 means that the probability of guessing a valid ciphertext is 6558 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 6559 paper.) 6560 6561 Before 0.9.5, the countermeasure (hide the error by generating a 6562 random 'decryption result') did not work properly because 6563 ERR_clear_error() was missing, meaning that SSL_get_error() would 6564 detect the supposedly ignored error. 6565 6566 Both problems are now fixed. 6567 [Bodo Moeller] 6568 6569 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 6570 (previously it was 1024). 6571 [Bodo Moeller] 6572 6573 *) Fix for compatibility mode trust settings: ignore trust settings 6574 unless some valid trust or reject settings are present. 6575 [Steve Henson] 6576 6577 *) Fix for blowfish EVP: its a variable length cipher. 6578 [Steve Henson] 6579 6580 *) Fix various bugs related to DSA S/MIME verification. Handle missing 6581 parameters in DSA public key structures and return an error in the 6582 DSA routines if parameters are absent. 6583 [Steve Henson] 6584 6585 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 6586 in the current directory if neither $RANDFILE nor $HOME was set. 6587 RAND_file_name() in 0.9.6a returned NULL in this case. This has 6588 caused some confusion to Windows users who haven't defined $HOME. 6589 Thus RAND_file_name() is changed again: e_os.h can define a 6590 DEFAULT_HOME, which will be used if $HOME is not set. 6591 For Windows, we use "C:"; on other platforms, we still require 6592 environment variables. 6593 6594 *) Move 'if (!initialized) RAND_poll()' into regions protected by 6595 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 6596 having multiple threads call RAND_poll() concurrently. 6597 [Bodo Moeller] 6598 6599 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 6600 combination of a flag and a thread ID variable. 6601 Otherwise while one thread is in ssleay_rand_bytes (which sets the 6602 flag), *other* threads can enter ssleay_add_bytes without obeying 6603 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 6604 that they do not hold after the first thread unsets add_do_not_lock). 6605 [Bodo Moeller] 6606 6607 *) Change bctest again: '-x' expressions are not available in all 6608 versions of 'test'. 6609 [Bodo Moeller] 6610 6611 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 6612 6613 *) Fix a couple of memory leaks in PKCS7_dataDecode() 6614 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 6615 6616 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 6617 the default extension for executables, if any. Also, make the perl 6618 scripts that use symlink() to test if it really exists and use "cp" 6619 if it doesn't. All this made OpenSSL compilable and installable in 6620 CygWin. 6621 [Richard Levitte] 6622 6623 *) Fix for asn1_GetSequence() for indefinite length constructed data. 6624 If SEQUENCE is length is indefinite just set c->slen to the total 6625 amount of data available. 6626 [Steve Henson, reported by shige@FreeBSD.org] 6627 [This change does not apply to 0.9.7.] 6628 6629 *) Change bctest to avoid here-documents inside command substitution 6630 (workaround for FreeBSD /bin/sh bug). 6631 For compatibility with Ultrix, avoid shell functions (introduced 6632 in the bctest version that searches along $PATH). 6633 [Bodo Moeller] 6634 6635 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 6636 with des_encrypt() defined on some operating systems, like Solaris 6637 and UnixWare. 6638 [Richard Levitte] 6639 6640 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 6641 On the Importance of Eliminating Errors in Cryptographic 6642 Computations, J. Cryptology 14 (2001) 2, 101-119, 6643 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 6644 [Ulf Moeller] 6645 6646 *) MIPS assembler BIGNUM division bug fix. 6647 [Andy Polyakov] 6648 6649 *) Disabled incorrect Alpha assembler code. 6650 [Richard Levitte] 6651 6652 *) Fix PKCS#7 decode routines so they correctly update the length 6653 after reading an EOC for the EXPLICIT tag. 6654 [Steve Henson] 6655 [This change does not apply to 0.9.7.] 6656 6657 *) Fix bug in PKCS#12 key generation routines. This was triggered 6658 if a 3DES key was generated with a 0 initial byte. Include 6659 PKCS12_BROKEN_KEYGEN compilation option to retain the old 6660 (but broken) behaviour. 6661 [Steve Henson] 6662 6663 *) Enhance bctest to search for a working bc along $PATH and print 6664 it when found. 6665 [Tim Rice <tim@multitalents.net> via Richard Levitte] 6666 6667 *) Fix memory leaks in err.c: free err_data string if necessary; 6668 don't write to the wrong index in ERR_set_error_data. 6669 [Bodo Moeller] 6670 6671 *) Implement ssl23_peek (analogous to ssl23_read), which previously 6672 did not exist. 6673 [Bodo Moeller] 6674 6675 *) Replace rdtsc with _emit statements for VC++ version 5. 6676 [Jeremy Cooper <jeremy@baymoo.org>] 6677 6678 *) Make it possible to reuse SSLv2 sessions. 6679 [Richard Levitte] 6680 6681 *) In copy_email() check for >= 0 as a return value for 6682 X509_NAME_get_index_by_NID() since 0 is a valid index. 6683 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 6684 6685 *) Avoid coredump with unsupported or invalid public keys by checking if 6686 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 6687 PKCS7_verify() fails with non detached data. 6688 [Steve Henson] 6689 6690 *) Don't use getenv in library functions when run as setuid/setgid. 6691 New function OPENSSL_issetugid(). 6692 [Ulf Moeller] 6693 6694 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 6695 due to incorrect handling of multi-threading: 6696 6697 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 6698 6699 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 6700 6701 3. Count how many times MemCheck_off() has been called so that 6702 nested use can be treated correctly. This also avoids 6703 inband-signalling in the previous code (which relied on the 6704 assumption that thread ID 0 is impossible). 6705 [Bodo Moeller] 6706 6707 *) Add "-rand" option also to s_client and s_server. 6708 [Lutz Jaenicke] 6709 6710 *) Fix CPU detection on Irix 6.x. 6711 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 6712 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 6713 6714 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 6715 was empty. 6716 [Steve Henson] 6717 [This change does not apply to 0.9.7.] 6718 6719 *) Use the cached encoding of an X509_NAME structure rather than 6720 copying it. This is apparently the reason for the libsafe "errors" 6721 but the code is actually correct. 6722 [Steve Henson] 6723 6724 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 6725 Bleichenbacher's DSA attack. 6726 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 6727 to be set and top=0 forces the highest bit to be set; top=-1 is new 6728 and leaves the highest bit random. 6729 [Ulf Moeller, Bodo Moeller] 6730 6731 *) In the NCONF_...-based implementations for CONF_... queries 6732 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 6733 a temporary CONF structure with the data component set to NULL 6734 (which gives segmentation faults in lh_retrieve). 6735 Instead, use NULL for the CONF pointer in CONF_get_string and 6736 CONF_get_number (which may use environment variables) and directly 6737 return NULL from CONF_get_section. 6738 [Bodo Moeller] 6739 6740 *) Fix potential buffer overrun for EBCDIC. 6741 [Ulf Moeller] 6742 6743 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 6744 keyUsage if basicConstraints absent for a CA. 6745 [Steve Henson] 6746 6747 *) Make SMIME_write_PKCS7() write mail header values with a format that 6748 is more generally accepted (no spaces before the semicolon), since 6749 some programs can't parse those values properly otherwise. Also make 6750 sure BIO's that break lines after each write do not create invalid 6751 headers. 6752 [Richard Levitte] 6753 6754 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 6755 macros previously used would not encode an empty SEQUENCE OF 6756 and break the signature. 6757 [Steve Henson] 6758 [This change does not apply to 0.9.7.] 6759 6760 *) Zero the premaster secret after deriving the master secret in 6761 DH ciphersuites. 6762 [Steve Henson] 6763 6764 *) Add some EVP_add_digest_alias registrations (as found in 6765 OpenSSL_add_all_digests()) to SSL_library_init() 6766 aka OpenSSL_add_ssl_algorithms(). This provides improved 6767 compatibility with peers using X.509 certificates 6768 with unconventional AlgorithmIdentifier OIDs. 6769 [Bodo Moeller] 6770 6771 *) Fix for Irix with NO_ASM. 6772 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 6773 6774 *) ./config script fixes. 6775 [Ulf Moeller, Richard Levitte] 6776 6777 *) Fix 'openssl passwd -1'. 6778 [Bodo Moeller] 6779 6780 *) Change PKCS12_key_gen_asc() so it can cope with non null 6781 terminated strings whose length is passed in the passlen 6782 parameter, for example from PEM callbacks. This was done 6783 by adding an extra length parameter to asc2uni(). 6784 [Steve Henson, reported by <oddissey@samsung.co.kr>] 6785 6786 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 6787 call failed, free the DSA structure. 6788 [Bodo Moeller] 6789 6790 *) Fix to uni2asc() to cope with zero length Unicode strings. 6791 These are present in some PKCS#12 files. 6792 [Steve Henson] 6793 6794 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 6795 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 6796 when writing a 32767 byte record. 6797 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 6798 6799 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 6800 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 6801 6802 (RSA objects have a reference count access to which is protected 6803 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 6804 so they are meant to be shared between threads.) 6805 [Bodo Moeller, Geoff Thorpe; original patch submitted by 6806 "Reddie, Steven" <Steven.Reddie@ca.com>] 6807 6808 *) Fix a deadlock in CRYPTO_mem_leaks(). 6809 [Bodo Moeller] 6810 6811 *) Use better test patterns in bntest. 6812 [Ulf M�ller] 6813 6814 *) rand_win.c fix for Borland C. 6815 [Ulf M�ller] 6816 6817 *) BN_rshift bugfix for n == 0. 6818 [Bodo Moeller] 6819 6820 *) Add a 'bctest' script that checks for some known 'bc' bugs 6821 so that 'make test' does not abort just because 'bc' is broken. 6822 [Bodo Moeller] 6823 6824 *) Store verify_result within SSL_SESSION also for client side to 6825 avoid potential security hole. (Re-used sessions on the client side 6826 always resulted in verify_result==X509_V_OK, not using the original 6827 result of the server certificate verification.) 6828 [Lutz Jaenicke] 6829 6830 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 6831 SSL3_RT_APPLICATION_DATA, return 0. 6832 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 6833 [Bodo Moeller] 6834 6835 *) Fix SSL_peek: 6836 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 6837 releases, have been re-implemented by renaming the previous 6838 implementations of ssl2_read and ssl3_read to ssl2_read_internal 6839 and ssl3_read_internal, respectively, and adding 'peek' parameters 6840 to them. The new ssl[23]_{read,peek} functions are calls to 6841 ssl[23]_read_internal with the 'peek' flag set appropriately. 6842 A 'peek' parameter has also been added to ssl3_read_bytes, which 6843 does the actual work for ssl3_read_internal. 6844 [Bodo Moeller] 6845 6846 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 6847 the method-specific "init()" handler. Also clean up ex_data after 6848 calling the method-specific "finish()" handler. Previously, this was 6849 happening the other way round. 6850 [Geoff Thorpe] 6851 6852 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 6853 The previous value, 12, was not always sufficient for BN_mod_exp(). 6854 [Bodo Moeller] 6855 6856 *) Make sure that shared libraries get the internal name engine with 6857 the full version number and not just 0. This should mark the 6858 shared libraries as not backward compatible. Of course, this should 6859 be changed again when we can guarantee backward binary compatibility. 6860 [Richard Levitte] 6861 6862 *) Fix typo in get_cert_by_subject() in by_dir.c 6863 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 6864 6865 *) Rework the system to generate shared libraries: 6866 6867 - Make note of the expected extension for the shared libraries and 6868 if there is a need for symbolic links from for example libcrypto.so.0 6869 to libcrypto.so.0.9.7. There is extended info in Configure for 6870 that. 6871 6872 - Make as few rebuilds of the shared libraries as possible. 6873 6874 - Still avoid linking the OpenSSL programs with the shared libraries. 6875 6876 - When installing, install the shared libraries separately from the 6877 static ones. 6878 [Richard Levitte] 6879 6880 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 6881 6882 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 6883 and not in SSL_clear because the latter is also used by the 6884 accept/connect functions; previously, the settings made by 6885 SSL_set_read_ahead would be lost during the handshake. 6886 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 6887 6888 *) Correct util/mkdef.pl to be selective about disabled algorithms. 6889 Previously, it would create entries for disableed algorithms no 6890 matter what. 6891 [Richard Levitte] 6892 6893 *) Added several new manual pages for SSL_* function. 6894 [Lutz Jaenicke] 6895 6896 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 6897 6898 *) In ssl23_get_client_hello, generate an error message when faced 6899 with an initial SSL 3.0/TLS record that is too small to contain the 6900 first two bytes of the ClientHello message, i.e. client_version. 6901 (Note that this is a pathologic case that probably has never happened 6902 in real life.) The previous approach was to use the version number 6903 from the record header as a substitute; but our protocol choice 6904 should not depend on that one because it is not authenticated 6905 by the Finished messages. 6906 [Bodo Moeller] 6907 6908 *) More robust randomness gathering functions for Windows. 6909 [Jeffrey Altman <jaltman@columbia.edu>] 6910 6911 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 6912 not set then we don't setup the error code for issuer check errors 6913 to avoid possibly overwriting other errors which the callback does 6914 handle. If an application does set the flag then we assume it knows 6915 what it is doing and can handle the new informational codes 6916 appropriately. 6917 [Steve Henson] 6918 6919 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 6920 a general "ANY" type, as such it should be able to decode anything 6921 including tagged types. However it didn't check the class so it would 6922 wrongly interpret tagged types in the same way as their universal 6923 counterpart and unknown types were just rejected. Changed so that the 6924 tagged and unknown types are handled in the same way as a SEQUENCE: 6925 that is the encoding is stored intact. There is also a new type 6926 "V_ASN1_OTHER" which is used when the class is not universal, in this 6927 case we have no idea what the actual type is so we just lump them all 6928 together. 6929 [Steve Henson] 6930 6931 *) On VMS, stdout may very well lead to a file that is written to 6932 in a record-oriented fashion. That means that every write() will 6933 write a separate record, which will be read separately by the 6934 programs trying to read from it. This can be very confusing. 6935 6936 The solution is to put a BIO filter in the way that will buffer 6937 text until a linefeed is reached, and then write everything a 6938 line at a time, so every record written will be an actual line, 6939 not chunks of lines and not (usually doesn't happen, but I've 6940 seen it once) several lines in one record. BIO_f_linebuffer() is 6941 the answer. 6942 6943 Currently, it's a VMS-only method, because that's where it has 6944 been tested well enough. 6945 [Richard Levitte] 6946 6947 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 6948 it can return incorrect results. 6949 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 6950 but it was in 0.9.6-beta[12].) 6951 [Bodo Moeller] 6952 6953 *) Disable the check for content being present when verifying detached 6954 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 6955 include zero length content when signing messages. 6956 [Steve Henson] 6957 6958 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 6959 BIO_ctrl (for BIO pairs). 6960 [Bodo M�ller] 6961 6962 *) Add DSO method for VMS. 6963 [Richard Levitte] 6964 6965 *) Bug fix: Montgomery multiplication could produce results with the 6966 wrong sign. 6967 [Ulf M�ller] 6968 6969 *) Add RPM specification openssl.spec and modify it to build three 6970 packages. The default package contains applications, application 6971 documentation and run-time libraries. The devel package contains 6972 include files, static libraries and function documentation. The 6973 doc package contains the contents of the doc directory. The original 6974 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 6975 [Richard Levitte] 6976 6977 *) Add a large number of documentation files for many SSL routines. 6978 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 6979 6980 *) Add a configuration entry for Sony News 4. 6981 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 6982 6983 *) Don't set the two most significant bits to one when generating a 6984 random number < q in the DSA library. 6985 [Ulf M�ller] 6986 6987 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 6988 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 6989 the underlying transport is blocking) if a handshake took place. 6990 (The default behaviour is needed by applications such as s_client 6991 and s_server that use select() to determine when to use SSL_read; 6992 but for applications that know in advance when to expect data, it 6993 just makes things more complicated.) 6994 [Bodo Moeller] 6995 6996 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 6997 from EGD. 6998 [Ben Laurie] 6999 7000 *) Add a few more EBCDIC conditionals that make `req' and `x509' 7001 work better on such systems. 7002 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 7003 7004 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 7005 Update PKCS12_parse() so it copies the friendlyName and the 7006 keyid to the certificates aux info. 7007 [Steve Henson] 7008 7009 *) Fix bug in PKCS7_verify() which caused an infinite loop 7010 if there was more than one signature. 7011 [Sven Uszpelkat <su@celocom.de>] 7012 7013 *) Major change in util/mkdef.pl to include extra information 7014 about each symbol, as well as presentig variables as well 7015 as functions. This change means that there's n more need 7016 to rebuild the .num files when some algorithms are excluded. 7017 [Richard Levitte] 7018 7019 *) Allow the verify time to be set by an application, 7020 rather than always using the current time. 7021 [Steve Henson] 7022 7023 *) Phase 2 verify code reorganisation. The certificate 7024 verify code now looks up an issuer certificate by a 7025 number of criteria: subject name, authority key id 7026 and key usage. It also verifies self signed certificates 7027 by the same criteria. The main comparison function is 7028 X509_check_issued() which performs these checks. 7029 7030 Lot of changes were necessary in order to support this 7031 without completely rewriting the lookup code. 7032 7033 Authority and subject key identifier are now cached. 7034 7035 The LHASH 'certs' is X509_STORE has now been replaced 7036 by a STACK_OF(X509_OBJECT). This is mainly because an 7037 LHASH can't store or retrieve multiple objects with 7038 the same hash value. 7039 7040 As a result various functions (which were all internal 7041 use only) have changed to handle the new X509_STORE 7042 structure. This will break anything that messed round 7043 with X509_STORE internally. 7044 7045 The functions X509_STORE_add_cert() now checks for an 7046 exact match, rather than just subject name. 7047 7048 The X509_STORE API doesn't directly support the retrieval 7049 of multiple certificates matching a given criteria, however 7050 this can be worked round by performing a lookup first 7051 (which will fill the cache with candidate certificates) 7052 and then examining the cache for matches. This is probably 7053 the best we can do without throwing out X509_LOOKUP 7054 entirely (maybe later...). 7055 7056 The X509_VERIFY_CTX structure has been enhanced considerably. 7057 7058 All certificate lookup operations now go via a get_issuer() 7059 callback. Although this currently uses an X509_STORE it 7060 can be replaced by custom lookups. This is a simple way 7061 to bypass the X509_STORE hackery necessary to make this 7062 work and makes it possible to use more efficient techniques 7063 in future. A very simple version which uses a simple 7064 STACK for its trusted certificate store is also provided 7065 using X509_STORE_CTX_trusted_stack(). 7066 7067 The verify_cb() and verify() callbacks now have equivalents 7068 in the X509_STORE_CTX structure. 7069 7070 X509_STORE_CTX also has a 'flags' field which can be used 7071 to customise the verify behaviour. 7072 [Steve Henson] 7073 7074 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 7075 excludes S/MIME capabilities. 7076 [Steve Henson] 7077 7078 *) When a certificate request is read in keep a copy of the 7079 original encoding of the signed data and use it when outputing 7080 again. Signatures then use the original encoding rather than 7081 a decoded, encoded version which may cause problems if the 7082 request is improperly encoded. 7083 [Steve Henson] 7084 7085 *) For consistency with other BIO_puts implementations, call 7086 buffer_write(b, ...) directly in buffer_puts instead of calling 7087 BIO_write(b, ...). 7088 7089 In BIO_puts, increment b->num_write as in BIO_write. 7090 [Peter.Sylvester@EdelWeb.fr] 7091 7092 *) Fix BN_mul_word for the case where the word is 0. (We have to use 7093 BN_zero, we may not return a BIGNUM with an array consisting of 7094 words set to zero.) 7095 [Bodo Moeller] 7096 7097 *) Avoid calling abort() from within the library when problems are 7098 detected, except if preprocessor symbols have been defined 7099 (such as REF_CHECK, BN_DEBUG etc.). 7100 [Bodo Moeller] 7101 7102 *) New openssl application 'rsautl'. This utility can be 7103 used for low level RSA operations. DER public key 7104 BIO/fp routines also added. 7105 [Steve Henson] 7106 7107 *) New Configure entry and patches for compiling on QNX 4. 7108 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 7109 7110 *) A demo state-machine implementation was sponsored by 7111 Nuron (http://www.nuron.com/) and is now available in 7112 demos/state_machine. 7113 [Ben Laurie] 7114 7115 *) New options added to the 'dgst' utility for signature 7116 generation and verification. 7117 [Steve Henson] 7118 7119 *) Unrecognized PKCS#7 content types are now handled via a 7120 catch all ASN1_TYPE structure. This allows unsupported 7121 types to be stored as a "blob" and an application can 7122 encode and decode it manually. 7123 [Steve Henson] 7124 7125 *) Fix various signed/unsigned issues to make a_strex.c 7126 compile under VC++. 7127 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 7128 7129 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 7130 length if passed a buffer. ASN1_INTEGER_to_BN failed 7131 if passed a NULL BN and its argument was negative. 7132 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 7133 7134 *) Modification to PKCS#7 encoding routines to output definite 7135 length encoding. Since currently the whole structures are in 7136 memory there's not real point in using indefinite length 7137 constructed encoding. However if OpenSSL is compiled with 7138 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 7139 [Steve Henson] 7140 7141 *) Added BIO_vprintf() and BIO_vsnprintf(). 7142 [Richard Levitte] 7143 7144 *) Added more prefixes to parse for in the the strings written 7145 through a logging bio, to cover all the levels that are available 7146 through syslog. The prefixes are now: 7147 7148 PANIC, EMERG, EMR => LOG_EMERG 7149 ALERT, ALR => LOG_ALERT 7150 CRIT, CRI => LOG_CRIT 7151 ERROR, ERR => LOG_ERR 7152 WARNING, WARN, WAR => LOG_WARNING 7153 NOTICE, NOTE, NOT => LOG_NOTICE 7154 INFO, INF => LOG_INFO 7155 DEBUG, DBG => LOG_DEBUG 7156 7157 and as before, if none of those prefixes are present at the 7158 beginning of the string, LOG_ERR is chosen. 7159 7160 On Win32, the LOG_* levels are mapped according to this: 7161 7162 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 7163 LOG_WARNING => EVENTLOG_WARNING_TYPE 7164 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 7165 7166 [Richard Levitte] 7167 7168 *) Made it possible to reconfigure with just the configuration 7169 argument "reconf" or "reconfigure". The command line arguments 7170 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 7171 and are retrieved from there when reconfiguring. 7172 [Richard Levitte] 7173 7174 *) MD4 implemented. 7175 [Assar Westerlund <assar@sics.se>, Richard Levitte] 7176 7177 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 7178 [Richard Levitte] 7179 7180 *) The obj_dat.pl script was messing up the sorting of object 7181 names. The reason was that it compared the quoted version 7182 of strings as a result "OCSP" > "OCSP Signing" because 7183 " > SPACE. Changed script to store unquoted versions of 7184 names and add quotes on output. It was also omitting some 7185 names from the lookup table if they were given a default 7186 value (that is if SN is missing it is given the same 7187 value as LN and vice versa), these are now added on the 7188 grounds that if an object has a name we should be able to 7189 look it up. Finally added warning output when duplicate 7190 short or long names are found. 7191 [Steve Henson] 7192 7193 *) Changes needed for Tandem NSK. 7194 [Scott Uroff <scott@xypro.com>] 7195 7196 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 7197 RSA_padding_check_SSLv23(), special padding was never detected 7198 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 7199 version rollback attacks was not effective. 7200 7201 In s23_clnt.c, don't use special rollback-attack detection padding 7202 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 7203 client; similarly, in s23_srvr.c, don't do the rollback check if 7204 SSL 2.0 is the only protocol enabled in the server. 7205 [Bodo Moeller] 7206 7207 *) Make it possible to get hexdumps of unprintable data with 'openssl 7208 asn1parse'. By implication, the functions ASN1_parse_dump() and 7209 BIO_dump_indent() are added. 7210 [Richard Levitte] 7211 7212 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 7213 these print out strings and name structures based on various 7214 flags including RFC2253 support and proper handling of 7215 multibyte characters. Added options to the 'x509' utility 7216 to allow the various flags to be set. 7217 [Steve Henson] 7218 7219 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 7220 Also change the functions X509_cmp_current_time() and 7221 X509_gmtime_adj() work with an ASN1_TIME structure, 7222 this will enable certificates using GeneralizedTime in validity 7223 dates to be checked. 7224 [Steve Henson] 7225 7226 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 7227 negative public key encodings) on by default, 7228 NO_NEG_PUBKEY_BUG can be set to disable it. 7229 [Steve Henson] 7230 7231 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 7232 content octets. An i2c_ASN1_OBJECT is unnecessary because 7233 the encoding can be trivially obtained from the structure. 7234 [Steve Henson] 7235 7236 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 7237 not read locks (CRYPTO_r_[un]lock). 7238 [Bodo Moeller] 7239 7240 *) A first attempt at creating official support for shared 7241 libraries through configuration. I've kept it so the 7242 default is static libraries only, and the OpenSSL programs 7243 are always statically linked for now, but there are 7244 preparations for dynamic linking in place. 7245 This has been tested on Linux and Tru64. 7246 [Richard Levitte] 7247 7248 *) Randomness polling function for Win9x, as described in: 7249 Peter Gutmann, Software Generation of Practically Strong 7250 Random Numbers. 7251 [Ulf M�ller] 7252 7253 *) Fix so PRNG is seeded in req if using an already existing 7254 DSA key. 7255 [Steve Henson] 7256 7257 *) New options to smime application. -inform and -outform 7258 allow alternative formats for the S/MIME message including 7259 PEM and DER. The -content option allows the content to be 7260 specified separately. This should allow things like Netscape 7261 form signing output easier to verify. 7262 [Steve Henson] 7263 7264 *) Fix the ASN1 encoding of tags using the 'long form'. 7265 [Steve Henson] 7266 7267 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 7268 STRING types. These convert content octets to and from the 7269 underlying type. The actual tag and length octets are 7270 already assumed to have been read in and checked. These 7271 are needed because all other string types have virtually 7272 identical handling apart from the tag. By having versions 7273 of the ASN1 functions that just operate on content octets 7274 IMPLICIT tagging can be handled properly. It also allows 7275 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 7276 and ASN1_INTEGER are identical apart from the tag. 7277 [Steve Henson] 7278 7279 *) Change the handling of OID objects as follows: 7280 7281 - New object identifiers are inserted in objects.txt, following 7282 the syntax given in objects.README. 7283 - objects.pl is used to process obj_mac.num and create a new 7284 obj_mac.h. 7285 - obj_dat.pl is used to create a new obj_dat.h, using the data in 7286 obj_mac.h. 7287 7288 This is currently kind of a hack, and the perl code in objects.pl 7289 isn't very elegant, but it works as I intended. The simplest way 7290 to check that it worked correctly is to look in obj_dat.h and 7291 check the array nid_objs and make sure the objects haven't moved 7292 around (this is important!). Additions are OK, as well as 7293 consistent name changes. 7294 [Richard Levitte] 7295 7296 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 7297 [Bodo Moeller] 7298 7299 *) Addition of the command line parameter '-rand file' to 'openssl req'. 7300 The given file adds to whatever has already been seeded into the 7301 random pool through the RANDFILE configuration file option or 7302 environment variable, or the default random state file. 7303 [Richard Levitte] 7304 7305 *) mkstack.pl now sorts each macro group into lexical order. 7306 Previously the output order depended on the order the files 7307 appeared in the directory, resulting in needless rewriting 7308 of safestack.h . 7309 [Steve Henson] 7310 7311 *) Patches to make OpenSSL compile under Win32 again. Mostly 7312 work arounds for the VC++ problem that it treats func() as 7313 func(void). Also stripped out the parts of mkdef.pl that 7314 added extra typesafe functions: these no longer exist. 7315 [Steve Henson] 7316 7317 *) Reorganisation of the stack code. The macros are now all 7318 collected in safestack.h . Each macro is defined in terms of 7319 a "stack macro" of the form SKM_<name>(type, a, b). The 7320 DEBUG_SAFESTACK is now handled in terms of function casts, 7321 this has the advantage of retaining type safety without the 7322 use of additional functions. If DEBUG_SAFESTACK is not defined 7323 then the non typesafe macros are used instead. Also modified the 7324 mkstack.pl script to handle the new form. Needs testing to see 7325 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 7326 the default if no major problems. Similar behaviour for ASN1_SET_OF 7327 and PKCS12_STACK_OF. 7328 [Steve Henson] 7329 7330 *) When some versions of IIS use the 'NET' form of private key the 7331 key derivation algorithm is different. Normally MD5(password) is 7332 used as a 128 bit RC4 key. In the modified case 7333 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 7334 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 7335 as the old Netscape_RSA functions except they have an additional 7336 'sgckey' parameter which uses the modified algorithm. Also added 7337 an -sgckey command line option to the rsa utility. Thanks to 7338 Adrian Peck <bertie@ncipher.com> for posting details of the modified 7339 algorithm to openssl-dev. 7340 [Steve Henson] 7341 7342 *) The evp_local.h macros were using 'c.##kname' which resulted in 7343 invalid expansion on some systems (SCO 5.0.5 for example). 7344 Corrected to 'c.kname'. 7345 [Phillip Porch <root@theporch.com>] 7346 7347 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 7348 a STACK of email addresses from a certificate or request, these look 7349 in the subject name and the subject alternative name extensions and 7350 omit any duplicate addresses. 7351 [Steve Henson] 7352 7353 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 7354 This makes DSA verification about 2 % faster. 7355 [Bodo Moeller] 7356 7357 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 7358 (meaning that now 2^5 values will be precomputed, which is only 4 KB 7359 plus overhead for 1024 bit moduli). 7360 This makes exponentiations about 0.5 % faster for 1024 bit 7361 exponents (as measured by "openssl speed rsa2048"). 7362 [Bodo Moeller] 7363 7364 *) Rename memory handling macros to avoid conflicts with other 7365 software: 7366 Malloc => OPENSSL_malloc 7367 Malloc_locked => OPENSSL_malloc_locked 7368 Realloc => OPENSSL_realloc 7369 Free => OPENSSL_free 7370 [Richard Levitte] 7371 7372 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 7373 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 7374 [Bodo Moeller] 7375 7376 *) CygWin32 support. 7377 [John Jarvie <jjarvie@newsguy.com>] 7378 7379 *) The type-safe stack code has been rejigged. It is now only compiled 7380 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 7381 by default all type-specific stack functions are "#define"d back to 7382 standard stack functions. This results in more streamlined output 7383 but retains the type-safety checking possibilities of the original 7384 approach. 7385 [Geoff Thorpe] 7386 7387 *) The STACK code has been cleaned up, and certain type declarations 7388 that didn't make a lot of sense have been brought in line. This has 7389 also involved a cleanup of sorts in safestack.h to more correctly 7390 map type-safe stack functions onto their plain stack counterparts. 7391 This work has also resulted in a variety of "const"ifications of 7392 lots of the code, especially "_cmp" operations which should normally 7393 be prototyped with "const" parameters anyway. 7394 [Geoff Thorpe] 7395 7396 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 7397 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 7398 (The PRNG state consists of two parts, the large pool 'state' and 'md', 7399 where all of 'md' is used each time the PRNG is used, but 'state' 7400 is used only indexed by a cyclic counter. As entropy may not be 7401 well distributed from the beginning, 'md' is important as a 7402 chaining variable. However, the output function chains only half 7403 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 7404 all of 'md', and seeding with STATE_SIZE dummy bytes will result 7405 in all of 'state' being rewritten, with the new values depending 7406 on virtually all of 'md'. This overcomes the 80 bit limitation.) 7407 [Bodo Moeller] 7408 7409 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 7410 the handshake is continued after ssl_verify_cert_chain(); 7411 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 7412 can lead to 'unexplainable' connection aborts later. 7413 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 7414 7415 *) Major EVP API cipher revision. 7416 Add hooks for extra EVP features. This allows various cipher 7417 parameters to be set in the EVP interface. Support added for variable 7418 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 7419 setting of RC2 and RC5 parameters. 7420 7421 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 7422 ciphers. 7423 7424 Remove lots of duplicated code from the EVP library. For example *every* 7425 cipher init() function handles the 'iv' in the same way according to the 7426 cipher mode. They also all do nothing if the 'key' parameter is NULL and 7427 for CFB and OFB modes they zero ctx->num. 7428 7429 New functionality allows removal of S/MIME code RC2 hack. 7430 7431 Most of the routines have the same form and so can be declared in terms 7432 of macros. 7433 7434 By shifting this to the top level EVP_CipherInit() it can be removed from 7435 all individual ciphers. If the cipher wants to handle IVs or keys 7436 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 7437 flags. 7438 7439 Change lots of functions like EVP_EncryptUpdate() to now return a 7440 value: although software versions of the algorithms cannot fail 7441 any installed hardware versions can. 7442 [Steve Henson] 7443 7444 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 7445 this option is set, tolerate broken clients that send the negotiated 7446 protocol version number instead of the requested protocol version 7447 number. 7448 [Bodo Moeller] 7449 7450 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 7451 i.e. non-zero for export ciphersuites, zero otherwise. 7452 Previous versions had this flag inverted, inconsistent with 7453 rsa_tmp_cb (..._TMP_RSA_CB). 7454 [Bodo Moeller; problem reported by Amit Chopra] 7455 7456 *) Add missing DSA library text string. Work around for some IIS 7457 key files with invalid SEQUENCE encoding. 7458 [Steve Henson] 7459 7460 *) Add a document (doc/standards.txt) that list all kinds of standards 7461 and so on that are implemented in OpenSSL. 7462 [Richard Levitte] 7463 7464 *) Enhance c_rehash script. Old version would mishandle certificates 7465 with the same subject name hash and wouldn't handle CRLs at all. 7466 Added -fingerprint option to crl utility, to support new c_rehash 7467 features. 7468 [Steve Henson] 7469 7470 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 7471 [Ulf M�ller] 7472 7473 *) Fix for SSL server purpose checking. Server checking was 7474 rejecting certificates which had extended key usage present 7475 but no ssl client purpose. 7476 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 7477 7478 *) Make PKCS#12 code work with no password. The PKCS#12 spec 7479 is a little unclear about how a blank password is handled. 7480 Since the password in encoded as a BMPString with terminating 7481 double NULL a zero length password would end up as just the 7482 double NULL. However no password at all is different and is 7483 handled differently in the PKCS#12 key generation code. NS 7484 treats a blank password as zero length. MSIE treats it as no 7485 password on export: but it will try both on import. We now do 7486 the same: PKCS12_parse() tries zero length and no password if 7487 the password is set to "" or NULL (NULL is now a valid password: 7488 it wasn't before) as does the pkcs12 application. 7489 [Steve Henson] 7490 7491 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 7492 perror when PEM_read_bio_X509_REQ fails, the error message must 7493 be obtained from the error queue. 7494 [Bodo Moeller] 7495 7496 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 7497 it in ERR_remove_state if appropriate, and change ERR_get_state 7498 accordingly to avoid race conditions (this is necessary because 7499 thread_hash is no longer constant once set). 7500 [Bodo Moeller] 7501 7502 *) Bugfix for linux-elf makefile.one. 7503 [Ulf M�ller] 7504 7505 *) RSA_get_default_method() will now cause a default 7506 RSA_METHOD to be chosen if one doesn't exist already. 7507 Previously this was only set during a call to RSA_new() 7508 or RSA_new_method(NULL) meaning it was possible for 7509 RSA_get_default_method() to return NULL. 7510 [Geoff Thorpe] 7511 7512 *) Added native name translation to the existing DSO code 7513 that will convert (if the flag to do so is set) filenames 7514 that are sufficiently small and have no path information 7515 into a canonical native form. Eg. "blah" converted to 7516 "libblah.so" or "blah.dll" etc. 7517 [Geoff Thorpe] 7518 7519 *) New function ERR_error_string_n(e, buf, len) which is like 7520 ERR_error_string(e, buf), but writes at most 'len' bytes 7521 including the 0 terminator. For ERR_error_string_n, 'buf' 7522 may not be NULL. 7523 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 7524 7525 *) CONF library reworked to become more general. A new CONF 7526 configuration file reader "class" is implemented as well as a 7527 new functions (NCONF_*, for "New CONF") to handle it. The now 7528 old CONF_* functions are still there, but are reimplemented to 7529 work in terms of the new functions. Also, a set of functions 7530 to handle the internal storage of the configuration data is 7531 provided to make it easier to write new configuration file 7532 reader "classes" (I can definitely see something reading a 7533 configuration file in XML format, for example), called _CONF_*, 7534 or "the configuration storage API"... 7535 7536 The new configuration file reading functions are: 7537 7538 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 7539 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 7540 7541 NCONF_default, NCONF_WIN32 7542 7543 NCONF_dump_fp, NCONF_dump_bio 7544 7545 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 7546 NCONF_new creates a new CONF object. This works in the same way 7547 as other interfaces in OpenSSL, like the BIO interface. 7548 NCONF_dump_* dump the internal storage of the configuration file, 7549 which is useful for debugging. All other functions take the same 7550 arguments as the old CONF_* functions wth the exception of the 7551 first that must be a `CONF *' instead of a `LHASH *'. 7552 7553 To make it easer to use the new classes with the old CONF_* functions, 7554 the function CONF_set_default_method is provided. 7555 [Richard Levitte] 7556 7557 *) Add '-tls1' option to 'openssl ciphers', which was already 7558 mentioned in the documentation but had not been implemented. 7559 (This option is not yet really useful because even the additional 7560 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 7561 [Bodo Moeller] 7562 7563 *) Initial DSO code added into libcrypto for letting OpenSSL (and 7564 OpenSSL-based applications) load shared libraries and bind to 7565 them in a portable way. 7566 [Geoff Thorpe, with contributions from Richard Levitte] 7567 7568 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 7569 7570 *) Make sure _lrotl and _lrotr are only used with MSVC. 7571 7572 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 7573 (the default implementation of RAND_status). 7574 7575 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 7576 to '-clrext' (= clear extensions), as intended and documented. 7577 [Bodo Moeller; inconsistency pointed out by Michael Attili 7578 <attili@amaxo.com>] 7579 7580 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 7581 was larger than the MD block size. 7582 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 7583 7584 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 7585 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 7586 using the passed key: if the passed key was a private key the result 7587 of X509_print(), for example, would be to print out all the private key 7588 components. 7589 [Steve Henson] 7590 7591 *) des_quad_cksum() byte order bug fix. 7592 [Ulf M�ller, using the problem description in krb4-0.9.7, where 7593 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 7594 7595 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 7596 discouraged. 7597 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 7598 7599 *) For easily testing in shell scripts whether some command 7600 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 7601 returns with exit code 0 iff no command of the given name is available. 7602 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 7603 the output goes to stdout and nothing is printed to stderr. 7604 Additional arguments are always ignored. 7605 7606 Since for each cipher there is a command of the same name, 7607 the 'no-cipher' compilation switches can be tested this way. 7608 7609 ('openssl no-XXX' is not able to detect pseudo-commands such 7610 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 7611 [Bodo Moeller] 7612 7613 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 7614 [Bodo Moeller] 7615 7616 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 7617 is set; it will be thrown away anyway because each handshake creates 7618 its own key. 7619 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 7620 to parameters -- in previous versions (since OpenSSL 0.9.3) the 7621 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 7622 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 7623 [Bodo Moeller] 7624 7625 *) New s_client option -ign_eof: EOF at stdin is ignored, and 7626 'Q' and 'R' lose their special meanings (quit/renegotiate). 7627 This is part of what -quiet does; unlike -quiet, -ign_eof 7628 does not suppress any output. 7629 [Richard Levitte] 7630 7631 *) Add compatibility options to the purpose and trust code. The 7632 purpose X509_PURPOSE_ANY is "any purpose" which automatically 7633 accepts a certificate or CA, this was the previous behaviour, 7634 with all the associated security issues. 7635 7636 X509_TRUST_COMPAT is the old trust behaviour: only and 7637 automatically trust self signed roots in certificate store. A 7638 new trust setting X509_TRUST_DEFAULT is used to specify that 7639 a purpose has no associated trust setting and it should instead 7640 use the value in the default purpose. 7641 [Steve Henson] 7642 7643 *) Fix the PKCS#8 DSA private key code so it decodes keys again 7644 and fix a memory leak. 7645 [Steve Henson] 7646 7647 *) In util/mkerr.pl (which implements 'make errors'), preserve 7648 reason strings from the previous version of the .c file, as 7649 the default to have only downcase letters (and digits) in 7650 automatically generated reasons codes is not always appropriate. 7651 [Bodo Moeller] 7652 7653 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 7654 using strerror. Previously, ERR_reason_error_string() returned 7655 library names as reason strings for SYSerr; but SYSerr is a special 7656 case where small numbers are errno values, not library numbers. 7657 [Bodo Moeller] 7658 7659 *) Add '-dsaparam' option to 'openssl dhparam' application. This 7660 converts DSA parameters into DH parameters. (When creating parameters, 7661 DSA_generate_parameters is used.) 7662 [Bodo Moeller] 7663 7664 *) Include 'length' (recommended exponent length) in C code generated 7665 by 'openssl dhparam -C'. 7666 [Bodo Moeller] 7667 7668 *) The second argument to set_label in perlasm was already being used 7669 so couldn't be used as a "file scope" flag. Moved to third argument 7670 which was free. 7671 [Steve Henson] 7672 7673 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 7674 instead of RAND_bytes for encryption IVs and salts. 7675 [Bodo Moeller] 7676 7677 *) Include RAND_status() into RAND_METHOD instead of implementing 7678 it only for md_rand.c Otherwise replacing the PRNG by calling 7679 RAND_set_rand_method would be impossible. 7680 [Bodo Moeller] 7681 7682 *) Don't let DSA_generate_key() enter an infinite loop if the random 7683 number generation fails. 7684 [Bodo Moeller] 7685 7686 *) New 'rand' application for creating pseudo-random output. 7687 [Bodo Moeller] 7688 7689 *) Added configuration support for Linux/IA64 7690 [Rolf Haberrecker <rolf@suse.de>] 7691 7692 *) Assembler module support for Mingw32. 7693 [Ulf M�ller] 7694 7695 *) Shared library support for HPUX (in shlib/). 7696 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 7697 7698 *) Shared library support for Solaris gcc. 7699 [Lutz Behnke <behnke@trustcenter.de>] 7700 7701 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 7702 7703 *) PKCS7_encrypt() was adding text MIME headers twice because they 7704 were added manually and by SMIME_crlf_copy(). 7705 [Steve Henson] 7706 7707 *) In bntest.c don't call BN_rand with zero bits argument. 7708 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 7709 7710 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 7711 case was implemented. This caused BN_div_recp() to fail occasionally. 7712 [Ulf M�ller] 7713 7714 *) Add an optional second argument to the set_label() in the perl 7715 assembly language builder. If this argument exists and is set 7716 to 1 it signals that the assembler should use a symbol whose 7717 scope is the entire file, not just the current function. This 7718 is needed with MASM which uses the format label:: for this scope. 7719 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 7720 7721 *) Change the ASN1 types so they are typedefs by default. Before 7722 almost all types were #define'd to ASN1_STRING which was causing 7723 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 7724 for example. 7725 [Steve Henson] 7726 7727 *) Change names of new functions to the new get1/get0 naming 7728 convention: After 'get1', the caller owns a reference count 7729 and has to call ..._free; 'get0' returns a pointer to some 7730 data structure without incrementing reference counters. 7731 (Some of the existing 'get' functions increment a reference 7732 counter, some don't.) 7733 Similarly, 'set1' and 'add1' functions increase reference 7734 counters or duplicate objects. 7735 [Steve Henson] 7736 7737 *) Allow for the possibility of temp RSA key generation failure: 7738 the code used to assume it always worked and crashed on failure. 7739 [Steve Henson] 7740 7741 *) Fix potential buffer overrun problem in BIO_printf(). 7742 [Ulf M�ller, using public domain code by Patrick Powell; problem 7743 pointed out by David Sacerdote <das33@cornell.edu>] 7744 7745 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 7746 RAND_egd() and RAND_status(). In the command line application, 7747 the EGD socket can be specified like a seed file using RANDFILE 7748 or -rand. 7749 [Ulf M�ller] 7750 7751 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 7752 Some CAs (e.g. Verisign) distribute certificates in this form. 7753 [Steve Henson] 7754 7755 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 7756 list to exclude them. This means that no special compilation option 7757 is needed to use anonymous DH: it just needs to be included in the 7758 cipher list. 7759 [Steve Henson] 7760 7761 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 7762 EVP_MD_type. The old functionality is available in a new macro called 7763 EVP_MD_md(). Change code that uses it and update docs. 7764 [Steve Henson] 7765 7766 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 7767 where the 'void *' argument is replaced by a function pointer argument. 7768 Previously 'void *' was abused to point to functions, which works on 7769 many platforms, but is not correct. As these functions are usually 7770 called by macros defined in OpenSSL header files, most source code 7771 should work without changes. 7772 [Richard Levitte] 7773 7774 *) <openssl/opensslconf.h> (which is created by Configure) now contains 7775 sections with information on -D... compiler switches used for 7776 compiling the library so that applications can see them. To enable 7777 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 7778 must be defined. E.g., 7779 #define OPENSSL_ALGORITHM_DEFINES 7780 #include <openssl/opensslconf.h> 7781 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 7782 [Richard Levitte, Ulf and Bodo M�ller] 7783 7784 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 7785 record layer. 7786 [Bodo Moeller] 7787 7788 *) Change the 'other' type in certificate aux info to a STACK_OF 7789 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 7790 the required ASN1 format: arbitrary types determined by an OID. 7791 [Steve Henson] 7792 7793 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 7794 argument to 'req'. This is not because the function is newer or 7795 better than others it just uses the work 'NEW' in the certificate 7796 request header lines. Some software needs this. 7797 [Steve Henson] 7798 7799 *) Reorganise password command line arguments: now passwords can be 7800 obtained from various sources. Delete the PEM_cb function and make 7801 it the default behaviour: i.e. if the callback is NULL and the 7802 usrdata argument is not NULL interpret it as a null terminated pass 7803 phrase. If usrdata and the callback are NULL then the pass phrase 7804 is prompted for as usual. 7805 [Steve Henson] 7806 7807 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 7808 the support is automatically enabled. The resulting binaries will 7809 autodetect the card and use it if present. 7810 [Ben Laurie and Compaq Inc.] 7811 7812 *) Work around for Netscape hang bug. This sends certificate request 7813 and server done in one record. Since this is perfectly legal in the 7814 SSL/TLS protocol it isn't a "bug" option and is on by default. See 7815 the bugs/SSLv3 entry for more info. 7816 [Steve Henson] 7817 7818 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 7819 [Andy Polyakov] 7820 7821 *) Add -rand argument to smime and pkcs12 applications and read/write 7822 of seed file. 7823 [Steve Henson] 7824 7825 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 7826 [Bodo Moeller] 7827 7828 *) Add command line password options to the remaining applications. 7829 [Steve Henson] 7830 7831 *) Bug fix for BN_div_recp() for numerators with an even number of 7832 bits. 7833 [Ulf M�ller] 7834 7835 *) More tests in bntest.c, and changed test_bn output. 7836 [Ulf M�ller] 7837 7838 *) ./config recognizes MacOS X now. 7839 [Andy Polyakov] 7840 7841 *) Bug fix for BN_div() when the first words of num and divsor are 7842 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 7843 [Ulf M�ller] 7844 7845 *) Add support for various broken PKCS#8 formats, and command line 7846 options to produce them. 7847 [Steve Henson] 7848 7849 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 7850 get temporary BIGNUMs from a BN_CTX. 7851 [Ulf M�ller] 7852 7853 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 7854 for p == 0. 7855 [Ulf M�ller] 7856 7857 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 7858 include a #define from the old name to the new. The original intent 7859 was that statically linked binaries could for example just call 7860 SSLeay_add_all_ciphers() to just add ciphers to the table and not 7861 link with digests. This never worked becayse SSLeay_add_all_digests() 7862 and SSLeay_add_all_ciphers() were in the same source file so calling 7863 one would link with the other. They are now in separate source files. 7864 [Steve Henson] 7865 7866 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 7867 [Steve Henson] 7868 7869 *) Use a less unusual form of the Miller-Rabin primality test (it used 7870 a binary algorithm for exponentiation integrated into the Miller-Rabin 7871 loop, our standard modexp algorithms are faster). 7872 [Bodo Moeller] 7873 7874 *) Support for the EBCDIC character set completed. 7875 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 7876 7877 *) Source code cleanups: use const where appropriate, eliminate casts, 7878 use void * instead of char * in lhash. 7879 [Ulf M�ller] 7880 7881 *) Bugfix: ssl3_send_server_key_exchange was not restartable 7882 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 7883 this the server could overwrite ephemeral keys that the client 7884 has already seen). 7885 [Bodo Moeller] 7886 7887 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 7888 using 50 iterations of the Rabin-Miller test. 7889 7890 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 7891 iterations of the Rabin-Miller test as required by the appendix 7892 to FIPS PUB 186[-1]) instead of DSA_is_prime. 7893 As BN_is_prime_fasttest includes trial division, DSA parameter 7894 generation becomes much faster. 7895 7896 This implies a change for the callback functions in DSA_is_prime 7897 and DSA_generate_parameters: The callback function is called once 7898 for each positive witness in the Rabin-Miller test, not just 7899 occasionally in the inner loop; and the parameters to the 7900 callback function now provide an iteration count for the outer 7901 loop rather than for the current invocation of the inner loop. 7902 DSA_generate_parameters additionally can call the callback 7903 function with an 'iteration count' of -1, meaning that a 7904 candidate has passed the trial division test (when q is generated 7905 from an application-provided seed, trial division is skipped). 7906 [Bodo Moeller] 7907 7908 *) New function BN_is_prime_fasttest that optionally does trial 7909 division before starting the Rabin-Miller test and has 7910 an additional BN_CTX * argument (whereas BN_is_prime always 7911 has to allocate at least one BN_CTX). 7912 'callback(1, -1, cb_arg)' is called when a number has passed the 7913 trial division stage. 7914 [Bodo Moeller] 7915 7916 *) Fix for bug in CRL encoding. The validity dates weren't being handled 7917 as ASN1_TIME. 7918 [Steve Henson] 7919 7920 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 7921 [Steve Henson] 7922 7923 *) New function BN_pseudo_rand(). 7924 [Ulf M�ller] 7925 7926 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 7927 bignum version of BN_from_montgomery() with the working code from 7928 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 7929 the comments. 7930 [Ulf M�ller] 7931 7932 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 7933 made it impossible to use the same SSL_SESSION data structure in 7934 SSL2 clients in multiple threads. 7935 [Bodo Moeller] 7936 7937 *) The return value of RAND_load_file() no longer counts bytes obtained 7938 by stat(). RAND_load_file(..., -1) is new and uses the complete file 7939 to seed the PRNG (previously an explicit byte count was required). 7940 [Ulf M�ller, Bodo M�ller] 7941 7942 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 7943 used (char *) instead of (void *) and had casts all over the place. 7944 [Steve Henson] 7945 7946 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 7947 [Ulf M�ller] 7948 7949 *) Retain source code compatibility for BN_prime_checks macro: 7950 BN_is_prime(..., BN_prime_checks, ...) now uses 7951 BN_prime_checks_for_size to determine the appropriate number of 7952 Rabin-Miller iterations. 7953 [Ulf M�ller] 7954 7955 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 7956 DH_CHECK_P_NOT_SAFE_PRIME. 7957 (Check if this is true? OpenPGP calls them "strong".) 7958 [Ulf M�ller] 7959 7960 *) Merge the functionality of "dh" and "gendh" programs into a new program 7961 "dhparam". The old programs are retained for now but will handle DH keys 7962 (instead of parameters) in future. 7963 [Steve Henson] 7964 7965 *) Make the ciphers, s_server and s_client programs check the return values 7966 when a new cipher list is set. 7967 [Steve Henson] 7968 7969 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 7970 ciphers. Before when the 56bit ciphers were enabled the sorting was 7971 wrong. 7972 7973 The syntax for the cipher sorting has been extended to support sorting by 7974 cipher-strength (using the strength_bits hard coded in the tables). 7975 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 7976 7977 Fix a bug in the cipher-command parser: when supplying a cipher command 7978 string with an "undefined" symbol (neither command nor alphanumeric 7979 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 7980 an error is flagged. 7981 7982 Due to the strength-sorting extension, the code of the 7983 ssl_create_cipher_list() function was completely rearranged. I hope that 7984 the readability was also increased :-) 7985 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 7986 7987 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 7988 for the first serial number and places 2 in the serial number file. This 7989 avoids problems when the root CA is created with serial number zero and 7990 the first user certificate has the same issuer name and serial number 7991 as the root CA. 7992 [Steve Henson] 7993 7994 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 7995 the new code. Add documentation for this stuff. 7996 [Steve Henson] 7997 7998 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 7999 X509_*() to X509at_*() on the grounds that they don't handle X509 8000 structures and behave in an analagous way to the X509v3 functions: 8001 they shouldn't be called directly but wrapper functions should be used 8002 instead. 8003 8004 So we also now have some wrapper functions that call the X509at functions 8005 when passed certificate requests. (TO DO: similar things can be done with 8006 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 8007 things. Some of these need some d2i or i2d and print functionality 8008 because they handle more complex structures.) 8009 [Steve Henson] 8010 8011 *) Add missing #ifndefs that caused missing symbols when building libssl 8012 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 8013 NO_RSA in ssl/s2*.c. 8014 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 8015 8016 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 8017 has a return value which indicates the quality of the random data 8018 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 8019 error queue. New function RAND_pseudo_bytes() generates output that is 8020 guaranteed to be unique but not unpredictable. RAND_add is like 8021 RAND_seed, but takes an extra argument for an entropy estimate 8022 (RAND_seed always assumes full entropy). 8023 [Ulf M�ller] 8024 8025 *) Do more iterations of Rabin-Miller probable prime test (specifically, 8026 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 8027 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 8028 in crypto/bn/bn_prime.c for the complete table). This guarantees a 8029 false-positive rate of at most 2^-80 for random input. 8030 [Bodo Moeller] 8031 8032 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 8033 [Bodo Moeller] 8034 8035 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 8036 in the 0.9.5 release), this returns the chain 8037 from an X509_CTX structure with a dup of the stack and all 8038 the X509 reference counts upped: so the stack will exist 8039 after X509_CTX_cleanup() has been called. Modify pkcs12.c 8040 to use this. 8041 8042 Also make SSL_SESSION_print() print out the verify return 8043 code. 8044 [Steve Henson] 8045 8046 *) Add manpage for the pkcs12 command. Also change the default 8047 behaviour so MAC iteration counts are used unless the new 8048 -nomaciter option is used. This improves file security and 8049 only older versions of MSIE (4.0 for example) need it. 8050 [Steve Henson] 8051 8052 *) Honor the no-xxx Configure options when creating .DEF files. 8053 [Ulf M�ller] 8054 8055 *) Add PKCS#10 attributes to field table: challengePassword, 8056 unstructuredName and unstructuredAddress. These are taken from 8057 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 8058 international characters are used. 8059 8060 More changes to X509_ATTRIBUTE code: allow the setting of types 8061 based on strings. Remove the 'loc' parameter when adding 8062 attributes because these will be a SET OF encoding which is sorted 8063 in ASN1 order. 8064 [Steve Henson] 8065 8066 *) Initial changes to the 'req' utility to allow request generation 8067 automation. This will allow an application to just generate a template 8068 file containing all the field values and have req construct the 8069 request. 8070 8071 Initial support for X509_ATTRIBUTE handling. Stacks of these are 8072 used all over the place including certificate requests and PKCS#7 8073 structures. They are currently handled manually where necessary with 8074 some primitive wrappers for PKCS#7. The new functions behave in a 8075 manner analogous to the X509 extension functions: they allow 8076 attributes to be looked up by NID and added. 8077 8078 Later something similar to the X509V3 code would be desirable to 8079 automatically handle the encoding, decoding and printing of the 8080 more complex types. The string types like challengePassword can 8081 be handled by the string table functions. 8082 8083 Also modified the multi byte string table handling. Now there is 8084 a 'global mask' which masks out certain types. The table itself 8085 can use the flag STABLE_NO_MASK to ignore the mask setting: this 8086 is useful when for example there is only one permissible type 8087 (as in countryName) and using the mask might result in no valid 8088 types at all. 8089 [Steve Henson] 8090 8091 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 8092 SSL_get_peer_finished to allow applications to obtain the latest 8093 Finished messages sent to the peer or expected from the peer, 8094 respectively. (SSL_get_peer_finished is usually the Finished message 8095 actually received from the peer, otherwise the protocol will be aborted.) 8096 8097 As the Finished message are message digests of the complete handshake 8098 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 8099 be used for external authentication procedures when the authentication 8100 provided by SSL/TLS is not desired or is not enough. 8101 [Bodo Moeller] 8102 8103 *) Enhanced support for Alpha Linux is added. Now ./config checks if 8104 the host supports BWX extension and if Compaq C is present on the 8105 $PATH. Just exploiting of the BWX extension results in 20-30% 8106 performance kick for some algorithms, e.g. DES and RC4 to mention 8107 a couple. Compaq C in turn generates ~20% faster code for MD5 and 8108 SHA1. 8109 [Andy Polyakov] 8110 8111 *) Add support for MS "fast SGC". This is arguably a violation of the 8112 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 8113 weak crypto and after checking the certificate is SGC a second one 8114 with strong crypto. MS SGC stops the first handshake after receiving 8115 the server certificate message and sends a second client hello. Since 8116 a server will typically do all the time consuming operations before 8117 expecting any further messages from the client (server key exchange 8118 is the most expensive) there is little difference between the two. 8119 8120 To get OpenSSL to support MS SGC we have to permit a second client 8121 hello message after we have sent server done. In addition we have to 8122 reset the MAC if we do get this second client hello. 8123 [Steve Henson] 8124 8125 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 8126 if a DER encoded private key is RSA or DSA traditional format. Changed 8127 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 8128 format DER encoded private key. Newer code should use PKCS#8 format which 8129 has the key type encoded in the ASN1 structure. Added DER private key 8130 support to pkcs8 application. 8131 [Steve Henson] 8132 8133 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 8134 ciphersuites has been selected (as required by the SSL 3/TLS 1 8135 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 8136 is set, we interpret this as a request to violate the specification 8137 (the worst that can happen is a handshake failure, and 'correct' 8138 behaviour would result in a handshake failure anyway). 8139 [Bodo Moeller] 8140 8141 *) In SSL_CTX_add_session, take into account that there might be multiple 8142 SSL_SESSION structures with the same session ID (e.g. when two threads 8143 concurrently obtain them from an external cache). 8144 The internal cache can handle only one SSL_SESSION with a given ID, 8145 so if there's a conflict, we now throw out the old one to achieve 8146 consistency. 8147 [Bodo Moeller] 8148 8149 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 8150 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 8151 some routines that use cipher OIDs: some ciphers do not have OIDs 8152 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 8153 example. 8154 [Steve Henson] 8155 8156 *) Simplify the trust setting structure and code. Now we just have 8157 two sequences of OIDs for trusted and rejected settings. These will 8158 typically have values the same as the extended key usage extension 8159 and any application specific purposes. 8160 8161 The trust checking code now has a default behaviour: it will just 8162 check for an object with the same NID as the passed id. Functions can 8163 be provided to override either the default behaviour or the behaviour 8164 for a given id. SSL client, server and email already have functions 8165 in place for compatibility: they check the NID and also return "trusted" 8166 if the certificate is self signed. 8167 [Steve Henson] 8168 8169 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 8170 traditional format into an EVP_PKEY structure. 8171 [Steve Henson] 8172 8173 *) Add a password callback function PEM_cb() which either prompts for 8174 a password if usr_data is NULL or otherwise assumes it is a null 8175 terminated password. Allow passwords to be passed on command line 8176 environment or config files in a few more utilities. 8177 [Steve Henson] 8178 8179 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 8180 keys. Add some short names for PKCS#8 PBE algorithms and allow them 8181 to be specified on the command line for the pkcs8 and pkcs12 utilities. 8182 Update documentation. 8183 [Steve Henson] 8184 8185 *) Support for ASN1 "NULL" type. This could be handled before by using 8186 ASN1_TYPE but there wasn't any function that would try to read a NULL 8187 and produce an error if it couldn't. For compatibility we also have 8188 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 8189 don't allocate anything because they don't need to. 8190 [Steve Henson] 8191 8192 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 8193 for details. 8194 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 8195 8196 *) Rebuild of the memory allocation routines used by OpenSSL code and 8197 possibly others as well. The purpose is to make an interface that 8198 provide hooks so anyone can build a separate set of allocation and 8199 deallocation routines to be used by OpenSSL, for example memory 8200 pool implementations, or something else, which was previously hard 8201 since Malloc(), Realloc() and Free() were defined as macros having 8202 the values malloc, realloc and free, respectively (except for Win32 8203 compilations). The same is provided for memory debugging code. 8204 OpenSSL already comes with functionality to find memory leaks, but 8205 this gives people a chance to debug other memory problems. 8206 8207 With these changes, a new set of functions and macros have appeared: 8208 8209 CRYPTO_set_mem_debug_functions() [F] 8210 CRYPTO_get_mem_debug_functions() [F] 8211 CRYPTO_dbg_set_options() [F] 8212 CRYPTO_dbg_get_options() [F] 8213 CRYPTO_malloc_debug_init() [M] 8214 8215 The memory debug functions are NULL by default, unless the library 8216 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 8217 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 8218 gives the standard debugging functions that come with OpenSSL) or 8219 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 8220 provided by the library user) must be used. When the standard 8221 debugging functions are used, CRYPTO_dbg_set_options can be used to 8222 request additional information: 8223 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 8224 the CRYPTO_MDEBUG_xxx macro when compiling the library. 8225 8226 Also, things like CRYPTO_set_mem_functions will always give the 8227 expected result (the new set of functions is used for allocation 8228 and deallocation) at all times, regardless of platform and compiler 8229 options. 8230 8231 To finish it up, some functions that were never use in any other 8232 way than through macros have a new API and new semantic: 8233 8234 CRYPTO_dbg_malloc() 8235 CRYPTO_dbg_realloc() 8236 CRYPTO_dbg_free() 8237 8238 All macros of value have retained their old syntax. 8239 [Richard Levitte and Bodo Moeller] 8240 8241 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 8242 ordering of SMIMECapabilities wasn't in "strength order" and there 8243 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 8244 algorithm. 8245 [Steve Henson] 8246 8247 *) Some ASN1 types with illegal zero length encoding (INTEGER, 8248 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 8249 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 8250 8251 *) Merge in my S/MIME library for OpenSSL. This provides a simple 8252 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 8253 functionality to handle multipart/signed properly) and a utility 8254 called 'smime' to call all this stuff. This is based on code I 8255 originally wrote for Celo who have kindly allowed it to be 8256 included in OpenSSL. 8257 [Steve Henson] 8258 8259 *) Add variants des_set_key_checked and des_set_key_unchecked of 8260 des_set_key (aka des_key_sched). Global variable des_check_key 8261 decides which of these is called by des_set_key; this way 8262 des_check_key behaves as it always did, but applications and 8263 the library itself, which was buggy for des_check_key == 1, 8264 have a cleaner way to pick the version they need. 8265 [Bodo Moeller] 8266 8267 *) New function PKCS12_newpass() which changes the password of a 8268 PKCS12 structure. 8269 [Steve Henson] 8270 8271 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 8272 dynamic mix. In both cases the ids can be used as an index into the 8273 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 8274 functions so they accept a list of the field values and the 8275 application doesn't need to directly manipulate the X509_TRUST 8276 structure. 8277 [Steve Henson] 8278 8279 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 8280 need initialising. 8281 [Steve Henson] 8282 8283 *) Modify the way the V3 extension code looks up extensions. This now 8284 works in a similar way to the object code: we have some "standard" 8285 extensions in a static table which is searched with OBJ_bsearch() 8286 and the application can add dynamic ones if needed. The file 8287 crypto/x509v3/ext_dat.h now has the info: this file needs to be 8288 updated whenever a new extension is added to the core code and kept 8289 in ext_nid order. There is a simple program 'tabtest.c' which checks 8290 this. New extensions are not added too often so this file can readily 8291 be maintained manually. 8292 8293 There are two big advantages in doing things this way. The extensions 8294 can be looked up immediately and no longer need to be "added" using 8295 X509V3_add_standard_extensions(): this function now does nothing. 8296 [Side note: I get *lots* of email saying the extension code doesn't 8297 work because people forget to call this function] 8298 Also no dynamic allocation is done unless new extensions are added: 8299 so if we don't add custom extensions there is no need to call 8300 X509V3_EXT_cleanup(). 8301 [Steve Henson] 8302 8303 *) Modify enc utility's salting as follows: make salting the default. Add a 8304 magic header, so unsalted files fail gracefully instead of just decrypting 8305 to garbage. This is because not salting is a big security hole, so people 8306 should be discouraged from doing it. 8307 [Ben Laurie] 8308 8309 *) Fixes and enhancements to the 'x509' utility. It allowed a message 8310 digest to be passed on the command line but it only used this 8311 parameter when signing a certificate. Modified so all relevant 8312 operations are affected by the digest parameter including the 8313 -fingerprint and -x509toreq options. Also -x509toreq choked if a 8314 DSA key was used because it didn't fix the digest. 8315 [Steve Henson] 8316 8317 *) Initial certificate chain verify code. Currently tests the untrusted 8318 certificates for consistency with the verify purpose (which is set 8319 when the X509_STORE_CTX structure is set up) and checks the pathlength. 8320 8321 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 8322 this is because it will reject chains with invalid extensions whereas 8323 every previous version of OpenSSL and SSLeay made no checks at all. 8324 8325 Trust code: checks the root CA for the relevant trust settings. Trust 8326 settings have an initial value consistent with the verify purpose: e.g. 8327 if the verify purpose is for SSL client use it expects the CA to be 8328 trusted for SSL client use. However the default value can be changed to 8329 permit custom trust settings: one example of this would be to only trust 8330 certificates from a specific "secure" set of CAs. 8331 8332 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 8333 which should be used for version portability: especially since the 8334 verify structure is likely to change more often now. 8335 8336 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 8337 to set them. If not set then assume SSL clients will verify SSL servers 8338 and vice versa. 8339 8340 Two new options to the verify program: -untrusted allows a set of 8341 untrusted certificates to be passed in and -purpose which sets the 8342 intended purpose of the certificate. If a purpose is set then the 8343 new chain verify code is used to check extension consistency. 8344 [Steve Henson] 8345 8346 *) Support for the authority information access extension. 8347 [Steve Henson] 8348 8349 *) Modify RSA and DSA PEM read routines to transparently handle 8350 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 8351 public keys in a format compatible with certificate 8352 SubjectPublicKeyInfo structures. Unfortunately there were already 8353 functions called *_PublicKey_* which used various odd formats so 8354 these are retained for compatibility: however the DSA variants were 8355 never in a public release so they have been deleted. Changed dsa/rsa 8356 utilities to handle the new format: note no releases ever handled public 8357 keys so we should be OK. 8358 8359 The primary motivation for this change is to avoid the same fiasco 8360 that dogs private keys: there are several incompatible private key 8361 formats some of which are standard and some OpenSSL specific and 8362 require various evil hacks to allow partial transparent handling and 8363 even then it doesn't work with DER formats. Given the option anything 8364 other than PKCS#8 should be dumped: but the other formats have to 8365 stay in the name of compatibility. 8366 8367 With public keys and the benefit of hindsight one standard format 8368 is used which works with EVP_PKEY, RSA or DSA structures: though 8369 it clearly returns an error if you try to read the wrong kind of key. 8370 8371 Added a -pubkey option to the 'x509' utility to output the public key. 8372 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 8373 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 8374 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 8375 that do the same as the EVP_PKEY_assign_*() except they up the 8376 reference count of the added key (they don't "swallow" the 8377 supplied key). 8378 [Steve Henson] 8379 8380 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 8381 CRLs would fail if the file contained no certificates or no CRLs: 8382 added a new function to read in both types and return the number 8383 read: this means that if none are read it will be an error. The 8384 DER versions of the certificate and CRL reader would always fail 8385 because it isn't possible to mix certificates and CRLs in DER format 8386 without choking one or the other routine. Changed this to just read 8387 a certificate: this is the best we can do. Also modified the code 8388 in apps/verify.c to take notice of return codes: it was previously 8389 attempting to read in certificates from NULL pointers and ignoring 8390 any errors: this is one reason why the cert and CRL reader seemed 8391 to work. It doesn't check return codes from the default certificate 8392 routines: these may well fail if the certificates aren't installed. 8393 [Steve Henson] 8394 8395 *) Code to support otherName option in GeneralName. 8396 [Steve Henson] 8397 8398 *) First update to verify code. Change the verify utility 8399 so it warns if it is passed a self signed certificate: 8400 for consistency with the normal behaviour. X509_verify 8401 has been modified to it will now verify a self signed 8402 certificate if *exactly* the same certificate appears 8403 in the store: it was previously impossible to trust a 8404 single self signed certificate. This means that: 8405 openssl verify ss.pem 8406 now gives a warning about a self signed certificate but 8407 openssl verify -CAfile ss.pem ss.pem 8408 is OK. 8409 [Steve Henson] 8410 8411 *) For servers, store verify_result in SSL_SESSION data structure 8412 (and add it to external session representation). 8413 This is needed when client certificate verifications fails, 8414 but an application-provided verification callback (set by 8415 SSL_CTX_set_cert_verify_callback) allows accepting the session 8416 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 8417 but returns 1): When the session is reused, we have to set 8418 ssl->verify_result to the appropriate error code to avoid 8419 security holes. 8420 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 8421 8422 *) Fix a bug in the new PKCS#7 code: it didn't consider the 8423 case in PKCS7_dataInit() where the signed PKCS7 structure 8424 didn't contain any existing data because it was being created. 8425 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 8426 8427 *) Add a salt to the key derivation routines in enc.c. This 8428 forms the first 8 bytes of the encrypted file. Also add a 8429 -S option to allow a salt to be input on the command line. 8430 [Steve Henson] 8431 8432 *) New function X509_cmp(). Oddly enough there wasn't a function 8433 to compare two certificates. We do this by working out the SHA1 8434 hash and comparing that. X509_cmp() will be needed by the trust 8435 code. 8436 [Steve Henson] 8437 8438 *) SSL_get1_session() is like SSL_get_session(), but increments 8439 the reference count in the SSL_SESSION returned. 8440 [Geoff Thorpe <geoff@eu.c2.net>] 8441 8442 *) Fix for 'req': it was adding a null to request attributes. 8443 Also change the X509_LOOKUP and X509_INFO code to handle 8444 certificate auxiliary information. 8445 [Steve Henson] 8446 8447 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 8448 the 'enc' command. 8449 [Steve Henson] 8450 8451 *) Add the possibility to add extra information to the memory leak 8452 detecting output, to form tracebacks, showing from where each 8453 allocation was originated: CRYPTO_push_info("constant string") adds 8454 the string plus current file name and line number to a per-thread 8455 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 8456 is like calling CYRPTO_pop_info() until the stack is empty. 8457 Also updated memory leak detection code to be multi-thread-safe. 8458 [Richard Levitte] 8459 8460 *) Add options -text and -noout to pkcs7 utility and delete the 8461 encryption options which never did anything. Update docs. 8462 [Steve Henson] 8463 8464 *) Add options to some of the utilities to allow the pass phrase 8465 to be included on either the command line (not recommended on 8466 OSes like Unix) or read from the environment. Update the 8467 manpages and fix a few bugs. 8468 [Steve Henson] 8469 8470 *) Add a few manpages for some of the openssl commands. 8471 [Steve Henson] 8472 8473 *) Fix the -revoke option in ca. It was freeing up memory twice, 8474 leaking and not finding already revoked certificates. 8475 [Steve Henson] 8476 8477 *) Extensive changes to support certificate auxiliary information. 8478 This involves the use of X509_CERT_AUX structure and X509_AUX 8479 functions. An X509_AUX function such as PEM_read_X509_AUX() 8480 can still read in a certificate file in the usual way but it 8481 will also read in any additional "auxiliary information". By 8482 doing things this way a fair degree of compatibility can be 8483 retained: existing certificates can have this information added 8484 using the new 'x509' options. 8485 8486 Current auxiliary information includes an "alias" and some trust 8487 settings. The trust settings will ultimately be used in enhanced 8488 certificate chain verification routines: currently a certificate 8489 can only be trusted if it is self signed and then it is trusted 8490 for all purposes. 8491 [Steve Henson] 8492 8493 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 8494 The problem was that one of the replacement routines had not been working 8495 since SSLeay releases. For now the offending routine has been replaced 8496 with non-optimised assembler. Even so, this now gives around 95% 8497 performance improvement for 1024 bit RSA signs. 8498 [Mark Cox] 8499 8500 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 8501 handling. Most clients have the effective key size in bits equal to 8502 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 8503 A few however don't do this and instead use the size of the decrypted key 8504 to determine the RC2 key length and the AlgorithmIdentifier to determine 8505 the effective key length. In this case the effective key length can still 8506 be 40 bits but the key length can be 168 bits for example. This is fixed 8507 by manually forcing an RC2 key into the EVP_PKEY structure because the 8508 EVP code can't currently handle unusual RC2 key sizes: it always assumes 8509 the key length and effective key length are equal. 8510 [Steve Henson] 8511 8512 *) Add a bunch of functions that should simplify the creation of 8513 X509_NAME structures. Now you should be able to do: 8514 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 8515 and have it automatically work out the correct field type and fill in 8516 the structures. The more adventurous can try: 8517 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 8518 and it will (hopefully) work out the correct multibyte encoding. 8519 [Steve Henson] 8520 8521 *) Change the 'req' utility to use the new field handling and multibyte 8522 copy routines. Before the DN field creation was handled in an ad hoc 8523 way in req, ca, and x509 which was rather broken and didn't support 8524 BMPStrings or UTF8Strings. Since some software doesn't implement 8525 BMPStrings or UTF8Strings yet, they can be enabled using the config file 8526 using the dirstring_type option. See the new comment in the default 8527 openssl.cnf for more info. 8528 [Steve Henson] 8529 8530 *) Make crypto/rand/md_rand.c more robust: 8531 - Assure unique random numbers after fork(). 8532 - Make sure that concurrent threads access the global counter and 8533 md serializably so that we never lose entropy in them 8534 or use exactly the same state in multiple threads. 8535 Access to the large state is not always serializable because 8536 the additional locking could be a performance killer, and 8537 md should be large enough anyway. 8538 [Bodo Moeller] 8539 8540 *) New file apps/app_rand.c with commonly needed functionality 8541 for handling the random seed file. 8542 8543 Use the random seed file in some applications that previously did not: 8544 ca, 8545 dsaparam -genkey (which also ignored its '-rand' option), 8546 s_client, 8547 s_server, 8548 x509 (when signing). 8549 Except on systems with /dev/urandom, it is crucial to have a random 8550 seed file at least for key creation, DSA signing, and for DH exchanges; 8551 for RSA signatures we could do without one. 8552 8553 gendh and gendsa (unlike genrsa) used to read only the first byte 8554 of each file listed in the '-rand' option. The function as previously 8555 found in genrsa is now in app_rand.c and is used by all programs 8556 that support '-rand'. 8557 [Bodo Moeller] 8558 8559 *) In RAND_write_file, use mode 0600 for creating files; 8560 don't just chmod when it may be too late. 8561 [Bodo Moeller] 8562 8563 *) Report an error from X509_STORE_load_locations 8564 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 8565 [Bill Perry] 8566 8567 *) New function ASN1_mbstring_copy() this copies a string in either 8568 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 8569 into an ASN1_STRING type. A mask of permissible types is passed 8570 and it chooses the "minimal" type to use or an error if not type 8571 is suitable. 8572 [Steve Henson] 8573 8574 *) Add function equivalents to the various macros in asn1.h. The old 8575 macros are retained with an M_ prefix. Code inside the library can 8576 use the M_ macros. External code (including the openssl utility) 8577 should *NOT* in order to be "shared library friendly". 8578 [Steve Henson] 8579 8580 *) Add various functions that can check a certificate's extensions 8581 to see if it usable for various purposes such as SSL client, 8582 server or S/MIME and CAs of these types. This is currently 8583 VERY EXPERIMENTAL but will ultimately be used for certificate chain 8584 verification. Also added a -purpose flag to x509 utility to 8585 print out all the purposes. 8586 [Steve Henson] 8587 8588 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 8589 functions. 8590 [Steve Henson] 8591 8592 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 8593 for, obtain and decode and extension and obtain its critical flag. 8594 This allows all the necessary extension code to be handled in a 8595 single function call. 8596 [Steve Henson] 8597 8598 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 8599 platforms. See crypto/rc4/rc4_enc.c for further details. 8600 [Andy Polyakov] 8601 8602 *) New -noout option to asn1parse. This causes no output to be produced 8603 its main use is when combined with -strparse and -out to extract data 8604 from a file (which may not be in ASN.1 format). 8605 [Steve Henson] 8606 8607 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 8608 when producing the local key id. 8609 [Richard Levitte <levitte@stacken.kth.se>] 8610 8611 *) New option -dhparam in s_server. This allows a DH parameter file to be 8612 stated explicitly. If it is not stated then it tries the first server 8613 certificate file. The previous behaviour hard coded the filename 8614 "server.pem". 8615 [Steve Henson] 8616 8617 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 8618 a public key to be input or output. For example: 8619 openssl rsa -in key.pem -pubout -out pubkey.pem 8620 Also added necessary DSA public key functions to handle this. 8621 [Steve Henson] 8622 8623 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 8624 in the message. This was handled by allowing 8625 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 8626 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 8627 8628 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 8629 to the end of the strings whereas this didn't. This would cause problems 8630 if strings read with d2i_ASN1_bytes() were later modified. 8631 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 8632 8633 *) Fix for base64 decode bug. When a base64 bio reads only one line of 8634 data and it contains EOF it will end up returning an error. This is 8635 caused by input 46 bytes long. The cause is due to the way base64 8636 BIOs find the start of base64 encoded data. They do this by trying a 8637 trial decode on each line until they find one that works. When they 8638 do a flag is set and it starts again knowing it can pass all the 8639 data directly through the decoder. Unfortunately it doesn't reset 8640 the context it uses. This means that if EOF is reached an attempt 8641 is made to pass two EOFs through the context and this causes the 8642 resulting error. This can also cause other problems as well. As is 8643 usual with these problems it takes *ages* to find and the fix is 8644 trivial: move one line. 8645 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 8646 8647 *) Ugly workaround to get s_client and s_server working under Windows. The 8648 old code wouldn't work because it needed to select() on sockets and the 8649 tty (for keypresses and to see if data could be written). Win32 only 8650 supports select() on sockets so we select() with a 1s timeout on the 8651 sockets and then see if any characters are waiting to be read, if none 8652 are present then we retry, we also assume we can always write data to 8653 the tty. This isn't nice because the code then blocks until we've 8654 received a complete line of data and it is effectively polling the 8655 keyboard at 1s intervals: however it's quite a bit better than not 8656 working at all :-) A dedicated Windows application might handle this 8657 with an event loop for example. 8658 [Steve Henson] 8659 8660 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 8661 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 8662 will be called when RSA_sign() and RSA_verify() are used. This is useful 8663 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 8664 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 8665 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 8666 This necessitated the support of an extra signature type NID_md5_sha1 8667 for SSL signatures and modifications to the SSL library to use it instead 8668 of calling RSA_public_decrypt() and RSA_private_encrypt(). 8669 [Steve Henson] 8670 8671 *) Add new -verify -CAfile and -CApath options to the crl program, these 8672 will lookup a CRL issuers certificate and verify the signature in a 8673 similar way to the verify program. Tidy up the crl program so it 8674 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 8675 less strict. It will now permit CRL extensions even if it is not 8676 a V2 CRL: this will allow it to tolerate some broken CRLs. 8677 [Steve Henson] 8678 8679 *) Initialize all non-automatic variables each time one of the openssl 8680 sub-programs is started (this is necessary as they may be started 8681 multiple times from the "OpenSSL>" prompt). 8682 [Lennart Bang, Bodo Moeller] 8683 8684 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 8685 removing all other RSA functionality (this is what NO_RSA does). This 8686 is so (for example) those in the US can disable those operations covered 8687 by the RSA patent while allowing storage and parsing of RSA keys and RSA 8688 key generation. 8689 [Steve Henson] 8690 8691 *) Non-copying interface to BIO pairs. 8692 (still largely untested) 8693 [Bodo Moeller] 8694 8695 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 8696 ASCII string. This was handled independently in various places before. 8697 [Steve Henson] 8698 8699 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 8700 UTF8 strings a character at a time. 8701 [Steve Henson] 8702 8703 *) Use client_version from client hello to select the protocol 8704 (s23_srvr.c) and for RSA client key exchange verification 8705 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 8706 [Bodo Moeller] 8707 8708 *) Add various utility functions to handle SPKACs, these were previously 8709 handled by poking round in the structure internals. Added new function 8710 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 8711 print, verify and generate SPKACs. Based on an original idea from 8712 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 8713 [Steve Henson] 8714 8715 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 8716 [Andy Polyakov] 8717 8718 *) Allow the config file extension section to be overwritten on the 8719 command line. Based on an original idea from Massimiliano Pala 8720 <madwolf@comune.modena.it>. The new option is called -extensions 8721 and can be applied to ca, req and x509. Also -reqexts to override 8722 the request extensions in req and -crlexts to override the crl extensions 8723 in ca. 8724 [Steve Henson] 8725 8726 *) Add new feature to the SPKAC handling in ca. Now you can include 8727 the same field multiple times by preceding it by "XXXX." for example: 8728 1.OU="Unit name 1" 8729 2.OU="Unit name 2" 8730 this is the same syntax as used in the req config file. 8731 [Steve Henson] 8732 8733 *) Allow certificate extensions to be added to certificate requests. These 8734 are specified in a 'req_extensions' option of the req section of the 8735 config file. They can be printed out with the -text option to req but 8736 are otherwise ignored at present. 8737 [Steve Henson] 8738 8739 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 8740 data read consists of only the final block it would not decrypted because 8741 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 8742 A misplaced 'break' also meant the decrypted final block might not be 8743 copied until the next read. 8744 [Steve Henson] 8745 8746 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 8747 a few extra parameters to the DH structure: these will be useful if 8748 for example we want the value of 'q' or implement X9.42 DH. 8749 [Steve Henson] 8750 8751 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 8752 provides hooks that allow the default DSA functions or functions on a 8753 "per key" basis to be replaced. This allows hardware acceleration and 8754 hardware key storage to be handled without major modification to the 8755 library. Also added low level modexp hooks and CRYPTO_EX structure and 8756 associated functions. 8757 [Steve Henson] 8758 8759 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 8760 as "read only": it can't be written to and the buffer it points to will 8761 not be freed. Reading from a read only BIO is much more efficient than 8762 a normal memory BIO. This was added because there are several times when 8763 an area of memory needs to be read from a BIO. The previous method was 8764 to create a memory BIO and write the data to it, this results in two 8765 copies of the data and an O(n^2) reading algorithm. There is a new 8766 function BIO_new_mem_buf() which creates a read only memory BIO from 8767 an area of memory. Also modified the PKCS#7 routines to use read only 8768 memory BIOs. 8769 [Steve Henson] 8770 8771 *) Bugfix: ssl23_get_client_hello did not work properly when called in 8772 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 8773 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 8774 but a retry condition occured while trying to read the rest. 8775 [Bodo Moeller] 8776 8777 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 8778 NID_pkcs7_encrypted by default: this was wrong since this should almost 8779 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 8780 the encrypted data type: this is a more sensible place to put it and it 8781 allows the PKCS#12 code to be tidied up that duplicated this 8782 functionality. 8783 [Steve Henson] 8784 8785 *) Changed obj_dat.pl script so it takes its input and output files on 8786 the command line. This should avoid shell escape redirection problems 8787 under Win32. 8788 [Steve Henson] 8789 8790 *) Initial support for certificate extension requests, these are included 8791 in things like Xenroll certificate requests. Included functions to allow 8792 extensions to be obtained and added. 8793 [Steve Henson] 8794 8795 *) -crlf option to s_client and s_server for sending newlines as 8796 CRLF (as required by many protocols). 8797 [Bodo Moeller] 8798 8799 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 8800 8801 *) Install libRSAglue.a when OpenSSL is built with RSAref. 8802 [Ralf S. Engelschall] 8803 8804 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 8805 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 8806 8807 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 8808 program. 8809 [Steve Henson] 8810 8811 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 8812 DH parameters/keys (q is lost during that conversion, but the resulting 8813 DH parameters contain its length). 8814 8815 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 8816 much faster than DH_generate_parameters (which creates parameters 8817 where p = 2*q + 1), and also the smaller q makes DH computations 8818 much more efficient (160-bit exponentiation instead of 1024-bit 8819 exponentiation); so this provides a convenient way to support DHE 8820 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 8821 utter importance to use 8822 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 8823 or 8824 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 8825 when such DH parameters are used, because otherwise small subgroup 8826 attacks may become possible! 8827 [Bodo Moeller] 8828 8829 *) Avoid memory leak in i2d_DHparams. 8830 [Bodo Moeller] 8831 8832 *) Allow the -k option to be used more than once in the enc program: 8833 this allows the same encrypted message to be read by multiple recipients. 8834 [Steve Henson] 8835 8836 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 8837 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 8838 it will always use the numerical form of the OID, even if it has a short 8839 or long name. 8840 [Steve Henson] 8841 8842 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 8843 method only got called if p,q,dmp1,dmq1,iqmp components were present, 8844 otherwise bn_mod_exp was called. In the case of hardware keys for example 8845 no private key components need be present and it might store extra data 8846 in the RSA structure, which cannot be accessed from bn_mod_exp. 8847 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 8848 private key operations. 8849 [Steve Henson] 8850 8851 *) Added support for SPARC Linux. 8852 [Andy Polyakov] 8853 8854 *) pem_password_cb function type incompatibly changed from 8855 typedef int pem_password_cb(char *buf, int size, int rwflag); 8856 to 8857 ....(char *buf, int size, int rwflag, void *userdata); 8858 so that applications can pass data to their callbacks: 8859 The PEM[_ASN1]_{read,write}... functions and macros now take an 8860 additional void * argument, which is just handed through whenever 8861 the password callback is called. 8862 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 8863 8864 New function SSL_CTX_set_default_passwd_cb_userdata. 8865 8866 Compatibility note: As many C implementations push function arguments 8867 onto the stack in reverse order, the new library version is likely to 8868 interoperate with programs that have been compiled with the old 8869 pem_password_cb definition (PEM_whatever takes some data that 8870 happens to be on the stack as its last argument, and the callback 8871 just ignores this garbage); but there is no guarantee whatsoever that 8872 this will work. 8873 8874 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 8875 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 8876 problems not only on Windows, but also on some Unix platforms. 8877 To avoid problematic command lines, these definitions are now in an 8878 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 8879 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 8880 [Bodo Moeller] 8881 8882 *) MIPS III/IV assembler module is reimplemented. 8883 [Andy Polyakov] 8884 8885 *) More DES library cleanups: remove references to srand/rand and 8886 delete an unused file. 8887 [Ulf M�ller] 8888 8889 *) Add support for the the free Netwide assembler (NASM) under Win32, 8890 since not many people have MASM (ml) and it can be hard to obtain. 8891 This is currently experimental but it seems to work OK and pass all 8892 the tests. Check out INSTALL.W32 for info. 8893 [Steve Henson] 8894 8895 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 8896 without temporary keys kept an extra copy of the server key, 8897 and connections with temporary keys did not free everything in case 8898 of an error. 8899 [Bodo Moeller] 8900 8901 *) New function RSA_check_key and new openssl rsa option -check 8902 for verifying the consistency of RSA keys. 8903 [Ulf Moeller, Bodo Moeller] 8904 8905 *) Various changes to make Win32 compile work: 8906 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 8907 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 8908 comparison" warnings. 8909 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 8910 [Steve Henson] 8911 8912 *) Add a debugging option to PKCS#5 v2 key generation function: when 8913 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 8914 derived keys are printed to stderr. 8915 [Steve Henson] 8916 8917 *) Copy the flags in ASN1_STRING_dup(). 8918 [Roman E. Pavlov <pre@mo.msk.ru>] 8919 8920 *) The x509 application mishandled signing requests containing DSA 8921 keys when the signing key was also DSA and the parameters didn't match. 8922 8923 It was supposed to omit the parameters when they matched the signing key: 8924 the verifying software was then supposed to automatically use the CA's 8925 parameters if they were absent from the end user certificate. 8926 8927 Omitting parameters is no longer recommended. The test was also 8928 the wrong way round! This was probably due to unusual behaviour in 8929 EVP_cmp_parameters() which returns 1 if the parameters match. 8930 This meant that parameters were omitted when they *didn't* match and 8931 the certificate was useless. Certificates signed with 'ca' didn't have 8932 this bug. 8933 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 8934 8935 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 8936 The interface is as follows: 8937 Applications can use 8938 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 8939 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 8940 "off" is now the default. 8941 The library internally uses 8942 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 8943 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 8944 to disable memory-checking temporarily. 8945 8946 Some inconsistent states that previously were possible (and were 8947 even the default) are now avoided. 8948 8949 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 8950 with each memory chunk allocated; this is occasionally more helpful 8951 than just having a counter. 8952 8953 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 8954 8955 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 8956 extensions. 8957 [Bodo Moeller] 8958 8959 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 8960 which largely parallels "options", but is for changing API behaviour, 8961 whereas "options" are about protocol behaviour. 8962 Initial "mode" flags are: 8963 8964 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 8965 a single record has been written. 8966 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 8967 retries use the same buffer location. 8968 (But all of the contents must be 8969 copied!) 8970 [Bodo Moeller] 8971 8972 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 8973 worked. 8974 8975 *) Fix problems with no-hmac etc. 8976 [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 8977 8978 *) New functions RSA_get_default_method(), RSA_set_method() and 8979 RSA_get_method(). These allows replacement of RSA_METHODs without having 8980 to mess around with the internals of an RSA structure. 8981 [Steve Henson] 8982 8983 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 8984 Also really enable memory leak checks in openssl.c and in some 8985 test programs. 8986 [Chad C. Mulligan, Bodo Moeller] 8987 8988 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 8989 up the length of negative integers. This has now been simplified to just 8990 store the length when it is first determined and use it later, rather 8991 than trying to keep track of where data is copied and updating it to 8992 point to the end. 8993 [Steve Henson, reported by Brien Wheeler 8994 <bwheeler@authentica-security.com>] 8995 8996 *) Add a new function PKCS7_signatureVerify. This allows the verification 8997 of a PKCS#7 signature but with the signing certificate passed to the 8998 function itself. This contrasts with PKCS7_dataVerify which assumes the 8999 certificate is present in the PKCS#7 structure. This isn't always the 9000 case: certificates can be omitted from a PKCS#7 structure and be 9001 distributed by "out of band" means (such as a certificate database). 9002 [Steve Henson] 9003 9004 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 9005 function prototypes in pem.h, also change util/mkdef.pl to add the 9006 necessary function names. 9007 [Steve Henson] 9008 9009 *) mk1mf.pl (used by Windows builds) did not properly read the 9010 options set by Configure in the top level Makefile, and Configure 9011 was not even able to write more than one option correctly. 9012 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 9013 [Bodo Moeller] 9014 9015 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 9016 file to be loaded from a BIO or FILE pointer. The BIO version will 9017 for example allow memory BIOs to contain config info. 9018 [Steve Henson] 9019 9020 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 9021 Whoever hopes to achieve shared-library compatibility across versions 9022 must use this, not the compile-time macro. 9023 (Exercise 0.9.4: Which is the minimum library version required by 9024 such programs?) 9025 Note: All this applies only to multi-threaded programs, others don't 9026 need locks. 9027 [Bodo Moeller] 9028 9029 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 9030 through a BIO pair triggered the default case, i.e. 9031 SSLerr(...,SSL_R_UNKNOWN_STATE). 9032 [Bodo Moeller] 9033 9034 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 9035 can use the SSL library even if none of the specific BIOs is 9036 appropriate. 9037 [Bodo Moeller] 9038 9039 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 9040 for the encoded length. 9041 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 9042 9043 *) Add initial documentation of the X509V3 functions. 9044 [Steve Henson] 9045 9046 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 9047 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 9048 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 9049 secure PKCS#8 private key format with a high iteration count. 9050 [Steve Henson] 9051 9052 *) Fix determination of Perl interpreter: A perl or perl5 9053 _directory_ in $PATH was also accepted as the interpreter. 9054 [Ralf S. Engelschall] 9055 9056 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 9057 wrong with it but it was very old and did things like calling 9058 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 9059 unusual formatting. 9060 [Steve Henson] 9061 9062 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 9063 to use the new extension code. 9064 [Steve Henson] 9065 9066 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 9067 with macros. This should make it easier to change their form, add extra 9068 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 9069 constant. 9070 [Steve Henson] 9071 9072 *) Add to configuration table a new entry that can specify an alternative 9073 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 9074 according to Mark Crispin <MRC@Panda.COM>. 9075 [Bodo Moeller] 9076 9077#if 0 9078 *) DES CBC did not update the IV. Weird. 9079 [Ben Laurie] 9080#else 9081 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 9082 Changing the behaviour of the former might break existing programs -- 9083 where IV updating is needed, des_ncbc_encrypt can be used. 9084#endif 9085 9086 *) When bntest is run from "make test" it drives bc to check its 9087 calculations, as well as internally checking them. If an internal check 9088 fails, it needs to cause bc to give a non-zero result or make test carries 9089 on without noticing the failure. Fixed. 9090 [Ben Laurie] 9091 9092 *) DES library cleanups. 9093 [Ulf M�ller] 9094 9095 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 9096 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 9097 ciphers. NOTE: although the key derivation function has been verified 9098 against some published test vectors it has not been extensively tested 9099 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 9100 of v2.0. 9101 [Steve Henson] 9102 9103 *) Instead of "mkdir -p", which is not fully portable, use new 9104 Perl script "util/mkdir-p.pl". 9105 [Bodo Moeller] 9106 9107 *) Rewrite the way password based encryption (PBE) is handled. It used to 9108 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 9109 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 9110 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 9111 the 'parameter' field of the AlgorithmIdentifier is passed to the 9112 underlying key generation function so it must do its own ASN1 parsing. 9113 This has also changed the EVP_PBE_CipherInit() function which now has a 9114 'parameter' argument instead of literal salt and iteration count values 9115 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 9116 [Steve Henson] 9117 9118 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 9119 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 9120 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 9121 KEY" because this clashed with PKCS#8 unencrypted string. Since this 9122 value was just used as a "magic string" and not used directly its 9123 value doesn't matter. 9124 [Steve Henson] 9125 9126 *) Introduce some semblance of const correctness to BN. Shame C doesn't 9127 support mutable. 9128 [Ben Laurie] 9129 9130 *) "linux-sparc64" configuration (ultrapenguin). 9131 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 9132 "linux-sparc" configuration. 9133 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 9134 9135 *) config now generates no-xxx options for missing ciphers. 9136 [Ulf M�ller] 9137 9138 *) Support the EBCDIC character set (work in progress). 9139 File ebcdic.c not yet included because it has a different license. 9140 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 9141 9142 *) Support BS2000/OSD-POSIX. 9143 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 9144 9145 *) Make callbacks for key generation use void * instead of char *. 9146 [Ben Laurie] 9147 9148 *) Make S/MIME samples compile (not yet tested). 9149 [Ben Laurie] 9150 9151 *) Additional typesafe stacks. 9152 [Ben Laurie] 9153 9154 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 9155 [Bodo Moeller] 9156 9157 9158 Changes between 0.9.3 and 0.9.3a [29 May 1999] 9159 9160 *) New configuration variant "sco5-gcc". 9161 9162 *) Updated some demos. 9163 [Sean O Riordain, Wade Scholine] 9164 9165 *) Add missing BIO_free at exit of pkcs12 application. 9166 [Wu Zhigang] 9167 9168 *) Fix memory leak in conf.c. 9169 [Steve Henson] 9170 9171 *) Updates for Win32 to assembler version of MD5. 9172 [Steve Henson] 9173 9174 *) Set #! path to perl in apps/der_chop to where we found it 9175 instead of using a fixed path. 9176 [Bodo Moeller] 9177 9178 *) SHA library changes for irix64-mips4-cc. 9179 [Andy Polyakov] 9180 9181 *) Improvements for VMS support. 9182 [Richard Levitte] 9183 9184 9185 Changes between 0.9.2b and 0.9.3 [24 May 1999] 9186 9187 *) Bignum library bug fix. IRIX 6 passes "make test" now! 9188 This also avoids the problems with SC4.2 and unpatched SC5. 9189 [Andy Polyakov <appro@fy.chalmers.se>] 9190 9191 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 9192 These are required because of the typesafe stack would otherwise break 9193 existing code. If old code used a structure member which used to be STACK 9194 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 9195 sk_num or sk_value it would produce an error because the num, data members 9196 are not present in STACK_OF. Now it just produces a warning. sk_set 9197 replaces the old method of assigning a value to sk_value 9198 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 9199 that does this will no longer work (and should use sk_set instead) but 9200 this could be regarded as a "questionable" behaviour anyway. 9201 [Steve Henson] 9202 9203 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 9204 correctly handle encrypted S/MIME data. 9205 [Steve Henson] 9206 9207 *) Change type of various DES function arguments from des_cblock 9208 (which means, in function argument declarations, pointer to char) 9209 to des_cblock * (meaning pointer to array with 8 char elements), 9210 which allows the compiler to do more typechecking; it was like 9211 that back in SSLeay, but with lots of ugly casts. 9212 9213 Introduce new type const_des_cblock. 9214 [Bodo Moeller] 9215 9216 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 9217 problems: find RecipientInfo structure that matches recipient certificate 9218 and initialise the ASN1 structures properly based on passed cipher. 9219 [Steve Henson] 9220 9221 *) Belatedly make the BN tests actually check the results. 9222 [Ben Laurie] 9223 9224 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 9225 to and from BNs: it was completely broken. New compilation option 9226 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 9227 key elements as negative integers. 9228 [Steve Henson] 9229 9230 *) Reorganize and speed up MD5. 9231 [Andy Polyakov <appro@fy.chalmers.se>] 9232 9233 *) VMS support. 9234 [Richard Levitte <richard@levitte.org>] 9235 9236 *) New option -out to asn1parse to allow the parsed structure to be 9237 output to a file. This is most useful when combined with the -strparse 9238 option to examine the output of things like OCTET STRINGS. 9239 [Steve Henson] 9240 9241 *) Make SSL library a little more fool-proof by not requiring any longer 9242 that SSL_set_{accept,connect}_state be called before 9243 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 9244 in many applications because usually everything *appeared* to work as 9245 intended anyway -- now it really works as intended). 9246 [Bodo Moeller] 9247 9248 *) Move openssl.cnf out of lib/. 9249 [Ulf M�ller] 9250 9251 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 9252 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 9253 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 9254 [Ralf S. Engelschall] 9255 9256 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 9257 handle PKCS#7 enveloped data properly. 9258 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 9259 9260 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 9261 copying pointers. The cert_st handling is changed by this in 9262 various ways (and thus what used to be known as ctx->default_cert 9263 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 9264 any longer when s->cert does not give us what we need). 9265 ssl_cert_instantiate becomes obsolete by this change. 9266 As soon as we've got the new code right (possibly it already is?), 9267 we have solved a couple of bugs of the earlier code where s->cert 9268 was used as if it could not have been shared with other SSL structures. 9269 9270 Note that using the SSL API in certain dirty ways now will result 9271 in different behaviour than observed with earlier library versions: 9272 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 9273 does not influence s as it used to. 9274 9275 In order to clean up things more thoroughly, inside SSL_SESSION 9276 we don't use CERT any longer, but a new structure SESS_CERT 9277 that holds per-session data (if available); currently, this is 9278 the peer's certificate chain and, for clients, the server's certificate 9279 and temporary key. CERT holds only those values that can have 9280 meaningful defaults in an SSL_CTX. 9281 [Bodo Moeller] 9282 9283 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 9284 from the internal representation. Various PKCS#7 fixes: remove some 9285 evil casts and set the enc_dig_alg field properly based on the signing 9286 key type. 9287 [Steve Henson] 9288 9289 *) Allow PKCS#12 password to be set from the command line or the 9290 environment. Let 'ca' get its config file name from the environment 9291 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 9292 and 'x509'). 9293 [Steve Henson] 9294 9295 *) Allow certificate policies extension to use an IA5STRING for the 9296 organization field. This is contrary to the PKIX definition but 9297 VeriSign uses it and IE5 only recognises this form. Document 'x509' 9298 extension option. 9299 [Steve Henson] 9300 9301 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 9302 without disallowing inline assembler and the like for non-pedantic builds. 9303 [Ben Laurie] 9304 9305 *) Support Borland C++ builder. 9306 [Janez Jere <jj@void.si>, modified by Ulf M�ller] 9307 9308 *) Support Mingw32. 9309 [Ulf M�ller] 9310 9311 *) SHA-1 cleanups and performance enhancements. 9312 [Andy Polyakov <appro@fy.chalmers.se>] 9313 9314 *) Sparc v8plus assembler for the bignum library. 9315 [Andy Polyakov <appro@fy.chalmers.se>] 9316 9317 *) Accept any -xxx and +xxx compiler options in Configure. 9318 [Ulf M�ller] 9319 9320 *) Update HPUX configuration. 9321 [Anonymous] 9322 9323 *) Add missing sk_<type>_unshift() function to safestack.h 9324 [Ralf S. Engelschall] 9325 9326 *) New function SSL_CTX_use_certificate_chain_file that sets the 9327 "extra_cert"s in addition to the certificate. (This makes sense 9328 only for "PEM" format files, as chains as a whole are not 9329 DER-encoded.) 9330 [Bodo Moeller] 9331 9332 *) Support verify_depth from the SSL API. 9333 x509_vfy.c had what can be considered an off-by-one-error: 9334 Its depth (which was not part of the external interface) 9335 was actually counting the number of certificates in a chain; 9336 now it really counts the depth. 9337 [Bodo Moeller] 9338 9339 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 9340 instead of X509err, which often resulted in confusing error 9341 messages since the error codes are not globally unique 9342 (e.g. an alleged error in ssl3_accept when a certificate 9343 didn't match the private key). 9344 9345 *) New function SSL_CTX_set_session_id_context that allows to set a default 9346 value (so that you don't need SSL_set_session_id_context for each 9347 connection using the SSL_CTX). 9348 [Bodo Moeller] 9349 9350 *) OAEP decoding bug fix. 9351 [Ulf M�ller] 9352 9353 *) Support INSTALL_PREFIX for package builders, as proposed by 9354 David Harris. 9355 [Bodo Moeller] 9356 9357 *) New Configure options "threads" and "no-threads". For systems 9358 where the proper compiler options are known (currently Solaris 9359 and Linux), "threads" is the default. 9360 [Bodo Moeller] 9361 9362 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 9363 [Bodo Moeller] 9364 9365 *) Install various scripts to $(OPENSSLDIR)/misc, not to 9366 $(INSTALLTOP)/bin -- they shouldn't clutter directories 9367 such as /usr/local/bin. 9368 [Bodo Moeller] 9369 9370 *) "make linux-shared" to build shared libraries. 9371 [Niels Poppe <niels@netbox.org>] 9372 9373 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 9374 [Ulf M�ller] 9375 9376 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 9377 extension adding in x509 utility. 9378 [Steve Henson] 9379 9380 *) Remove NOPROTO sections and error code comments. 9381 [Ulf M�ller] 9382 9383 *) Partial rewrite of the DEF file generator to now parse the ANSI 9384 prototypes. 9385 [Steve Henson] 9386 9387 *) New Configure options --prefix=DIR and --openssldir=DIR. 9388 [Ulf M�ller] 9389 9390 *) Complete rewrite of the error code script(s). It is all now handled 9391 by one script at the top level which handles error code gathering, 9392 header rewriting and C source file generation. It should be much better 9393 than the old method: it now uses a modified version of Ulf's parser to 9394 read the ANSI prototypes in all header files (thus the old K&R definitions 9395 aren't needed for error creation any more) and do a better job of 9396 translating function codes into names. The old 'ASN1 error code imbedded 9397 in a comment' is no longer necessary and it doesn't use .err files which 9398 have now been deleted. Also the error code call doesn't have to appear all 9399 on one line (which resulted in some large lines...). 9400 [Steve Henson] 9401 9402 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 9403 [Bodo Moeller] 9404 9405 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 9406 0 (which usually indicates a closed connection), but continue reading. 9407 [Bodo Moeller] 9408 9409 *) Fix some race conditions. 9410 [Bodo Moeller] 9411 9412 *) Add support for CRL distribution points extension. Add Certificate 9413 Policies and CRL distribution points documentation. 9414 [Steve Henson] 9415 9416 *) Move the autogenerated header file parts to crypto/opensslconf.h. 9417 [Ulf M�ller] 9418 9419 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 9420 8 of keying material. Merlin has also confirmed interop with this fix 9421 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 9422 [Merlin Hughes <merlin@baltimore.ie>] 9423 9424 *) Fix lots of warnings. 9425 [Richard Levitte <levitte@stacken.kth.se>] 9426 9427 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 9428 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 9429 [Richard Levitte <levitte@stacken.kth.se>] 9430 9431 *) Fix problems with sizeof(long) == 8. 9432 [Andy Polyakov <appro@fy.chalmers.se>] 9433 9434 *) Change functions to ANSI C. 9435 [Ulf M�ller] 9436 9437 *) Fix typos in error codes. 9438 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 9439 9440 *) Remove defunct assembler files from Configure. 9441 [Ulf M�ller] 9442 9443 *) SPARC v8 assembler BIGNUM implementation. 9444 [Andy Polyakov <appro@fy.chalmers.se>] 9445 9446 *) Support for Certificate Policies extension: both print and set. 9447 Various additions to support the r2i method this uses. 9448 [Steve Henson] 9449 9450 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 9451 return a const string when you are expecting an allocated buffer. 9452 [Ben Laurie] 9453 9454 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 9455 types DirectoryString and DisplayText. 9456 [Steve Henson] 9457 9458 *) Add code to allow r2i extensions to access the configuration database, 9459 add an LHASH database driver and add several ctx helper functions. 9460 [Steve Henson] 9461 9462 *) Fix an evil bug in bn_expand2() which caused various BN functions to 9463 fail when they extended the size of a BIGNUM. 9464 [Steve Henson] 9465 9466 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 9467 support typesafe stack. 9468 [Steve Henson] 9469 9470 *) Fix typo in SSL_[gs]et_options(). 9471 [Nils Frostberg <nils@medcom.se>] 9472 9473 *) Delete various functions and files that belonged to the (now obsolete) 9474 old X509V3 handling code. 9475 [Steve Henson] 9476 9477 *) New Configure option "rsaref". 9478 [Ulf M�ller] 9479 9480 *) Don't auto-generate pem.h. 9481 [Bodo Moeller] 9482 9483 *) Introduce type-safe ASN.1 SETs. 9484 [Ben Laurie] 9485 9486 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 9487 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 9488 9489 *) Introduce type-safe STACKs. This will almost certainly break lots of code 9490 that links with OpenSSL (well at least cause lots of warnings), but fear 9491 not: the conversion is trivial, and it eliminates loads of evil casts. A 9492 few STACKed things have been converted already. Feel free to convert more. 9493 In the fullness of time, I'll do away with the STACK type altogether. 9494 [Ben Laurie] 9495 9496 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 9497 specified in <certfile> by updating the entry in the index.txt file. 9498 This way one no longer has to edit the index.txt file manually for 9499 revoking a certificate. The -revoke option does the gory details now. 9500 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 9501 9502 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 9503 `-text' option at all and this way the `-noout -text' combination was 9504 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 9505 [Ralf S. Engelschall] 9506 9507 *) Make sure a corresponding plain text error message exists for the 9508 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 9509 verify callback function determined that a certificate was revoked. 9510 [Ralf S. Engelschall] 9511 9512 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 9513 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 9514 all available cipers including rc5, which was forgotten until now. 9515 In order to let the testing shell script know which algorithms 9516 are available, a new (up to now undocumented) command 9517 "openssl list-cipher-commands" is used. 9518 [Bodo Moeller] 9519 9520 *) Bugfix: s_client occasionally would sleep in select() when 9521 it should have checked SSL_pending() first. 9522 [Bodo Moeller] 9523 9524 *) New functions DSA_do_sign and DSA_do_verify to provide access to 9525 the raw DSA values prior to ASN.1 encoding. 9526 [Ulf M�ller] 9527 9528 *) Tweaks to Configure 9529 [Niels Poppe <niels@netbox.org>] 9530 9531 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 9532 yet... 9533 [Steve Henson] 9534 9535 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 9536 [Ulf M�ller] 9537 9538 *) New config option to avoid instructions that are illegal on the 80386. 9539 The default code is faster, but requires at least a 486. 9540 [Ulf M�ller] 9541 9542 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 9543 SSL2_SERVER_VERSION (not used at all) macros, which are now the 9544 same as SSL2_VERSION anyway. 9545 [Bodo Moeller] 9546 9547 *) New "-showcerts" option for s_client. 9548 [Bodo Moeller] 9549 9550 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 9551 application. Various cleanups and fixes. 9552 [Steve Henson] 9553 9554 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 9555 modify error routines to work internally. Add error codes and PBE init 9556 to library startup routines. 9557 [Steve Henson] 9558 9559 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 9560 packing functions to asn1 and evp. Changed function names and error 9561 codes along the way. 9562 [Steve Henson] 9563 9564 *) PKCS12 integration: and so it begins... First of several patches to 9565 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 9566 objects to objects.h 9567 [Steve Henson] 9568 9569 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 9570 and display support for Thawte strong extranet extension. 9571 [Steve Henson] 9572 9573 *) Add LinuxPPC support. 9574 [Jeff Dubrule <igor@pobox.org>] 9575 9576 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 9577 bn_div_words in alpha.s. 9578 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 9579 9580 *) Make sure the RSA OAEP test is skipped under -DRSAref because 9581 OAEP isn't supported when OpenSSL is built with RSAref. 9582 [Ulf Moeller <ulf@fitug.de>] 9583 9584 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 9585 so they no longer are missing under -DNOPROTO. 9586 [Soren S. Jorvang <soren@t.dk>] 9587 9588 9589 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 9590 9591 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 9592 doesn't work when the session is reused. Coming soon! 9593 [Ben Laurie] 9594 9595 *) Fix a security hole, that allows sessions to be reused in the wrong 9596 context thus bypassing client cert protection! All software that uses 9597 client certs and session caches in multiple contexts NEEDS PATCHING to 9598 allow session reuse! A fuller solution is in the works. 9599 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 9600 9601 *) Some more source tree cleanups (removed obsolete files 9602 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 9603 permission on "config" script to be executable) and a fix for the INSTALL 9604 document. 9605 [Ulf Moeller <ulf@fitug.de>] 9606 9607 *) Remove some legacy and erroneous uses of malloc, free instead of 9608 Malloc, Free. 9609 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 9610 9611 *) Make rsa_oaep_test return non-zero on error. 9612 [Ulf Moeller <ulf@fitug.de>] 9613 9614 *) Add support for native Solaris shared libraries. Configure 9615 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 9616 if someone would make that last step automatic. 9617 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 9618 9619 *) ctx_size was not built with the right compiler during "make links". Fixed. 9620 [Ben Laurie] 9621 9622 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 9623 except NULL ciphers". This means the default cipher list will no longer 9624 enable NULL ciphers. They need to be specifically enabled e.g. with 9625 the string "DEFAULT:eNULL". 9626 [Steve Henson] 9627 9628 *) Fix to RSA private encryption routines: if p < q then it would 9629 occasionally produce an invalid result. This will only happen with 9630 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 9631 [Steve Henson] 9632 9633 *) Be less restrictive and allow also `perl util/perlpath.pl 9634 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 9635 because this way one can also use an interpreter named `perl5' (which is 9636 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 9637 installed as `perl'). 9638 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 9639 9640 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 9641 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 9642 9643 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 9644 advapi32.lib to Win32 build and change the pem test comparision 9645 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 9646 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 9647 and crypto/des/ede_cbcm_enc.c. 9648 [Steve Henson] 9649 9650 *) DES quad checksum was broken on big-endian architectures. Fixed. 9651 [Ben Laurie] 9652 9653 *) Comment out two functions in bio.h that aren't implemented. Fix up the 9654 Win32 test batch file so it (might) work again. The Win32 test batch file 9655 is horrible: I feel ill.... 9656 [Steve Henson] 9657 9658 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 9659 in e_os.h. Audit of header files to check ANSI and non ANSI 9660 sections: 10 functions were absent from non ANSI section and not exported 9661 from Windows DLLs. Fixed up libeay.num for new functions. 9662 [Steve Henson] 9663 9664 *) Make `openssl version' output lines consistent. 9665 [Ralf S. Engelschall] 9666 9667 *) Fix Win32 symbol export lists for BIO functions: Added 9668 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 9669 to ms/libeay{16,32}.def. 9670 [Ralf S. Engelschall] 9671 9672 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 9673 fine under Unix and passes some trivial tests I've now added. But the 9674 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 9675 added to make sure no one expects that this stuff really works in the 9676 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 9677 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 9678 openssl_bio.xs. 9679 [Ralf S. Engelschall] 9680 9681 *) Fix the generation of two part addresses in perl. 9682 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 9683 9684 *) Add config entry for Linux on MIPS. 9685 [John Tobey <jtobey@channel1.com>] 9686 9687 *) Make links whenever Configure is run, unless we are on Windoze. 9688 [Ben Laurie] 9689 9690 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 9691 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 9692 in CRLs. 9693 [Steve Henson] 9694 9695 *) Add a useful kludge to allow package maintainers to specify compiler and 9696 other platforms details on the command line without having to patch the 9697 Configure script everytime: One now can use ``perl Configure 9698 <id>:<details>'', i.e. platform ids are allowed to have details appended 9699 to them (seperated by colons). This is treated as there would be a static 9700 pre-configured entry in Configure's %table under key <id> with value 9701 <details> and ``perl Configure <id>'' is called. So, when you want to 9702 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 9703 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 9704 now, which overrides the FreeBSD-elf entry on-the-fly. 9705 [Ralf S. Engelschall] 9706 9707 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 9708 [Ben Laurie] 9709 9710 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 9711 on the `perl Configure ...' command line. This way one can compile 9712 OpenSSL libraries with Position Independent Code (PIC) which is needed 9713 for linking it into DSOs. 9714 [Ralf S. Engelschall] 9715 9716 *) Remarkably, export ciphers were totally broken and no-one had noticed! 9717 Fixed. 9718 [Ben Laurie] 9719 9720 *) Cleaned up the LICENSE document: The official contact for any license 9721 questions now is the OpenSSL core team under openssl-core@openssl.org. 9722 And add a paragraph about the dual-license situation to make sure people 9723 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 9724 to the OpenSSL toolkit. 9725 [Ralf S. Engelschall] 9726 9727 *) General source tree makefile cleanups: Made `making xxx in yyy...' 9728 display consistent in the source tree and replaced `/bin/rm' by `rm'. 9729 Additonally cleaned up the `make links' target: Remove unnecessary 9730 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 9731 to speed processing and no longer clutter the display with confusing 9732 stuff. Instead only the actually done links are displayed. 9733 [Ralf S. Engelschall] 9734 9735 *) Permit null encryption ciphersuites, used for authentication only. It used 9736 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 9737 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 9738 encryption. 9739 [Ben Laurie] 9740 9741 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 9742 signed attributes when verifying signatures (this would break them), 9743 the detached data encoding was wrong and public keys obtained using 9744 X509_get_pubkey() weren't freed. 9745 [Steve Henson] 9746 9747 *) Add text documentation for the BUFFER functions. Also added a work around 9748 to a Win95 console bug. This was triggered by the password read stuff: the 9749 last character typed gets carried over to the next fread(). If you were 9750 generating a new cert request using 'req' for example then the last 9751 character of the passphrase would be CR which would then enter the first 9752 field as blank. 9753 [Steve Henson] 9754 9755 *) Added the new `Includes OpenSSL Cryptography Software' button as 9756 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 9757 button and can be used by applications based on OpenSSL to show the 9758 relationship to the OpenSSL project. 9759 [Ralf S. Engelschall] 9760 9761 *) Remove confusing variables in function signatures in files 9762 ssl/ssl_lib.c and ssl/ssl.h. 9763 [Lennart Bong <lob@kulthea.stacken.kth.se>] 9764 9765 *) Don't install bss_file.c under PREFIX/include/ 9766 [Lennart Bong <lob@kulthea.stacken.kth.se>] 9767 9768 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 9769 functions that return function pointers and has support for NT specific 9770 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 9771 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 9772 unsigned to signed types: this was killing the Win32 compile. 9773 [Steve Henson] 9774 9775 *) Add new certificate file to stack functions, 9776 SSL_add_dir_cert_subjects_to_stack() and 9777 SSL_add_file_cert_subjects_to_stack(). These largely supplant 9778 SSL_load_client_CA_file(), and can be used to add multiple certs easily 9779 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 9780 This means that Apache-SSL and similar packages don't have to mess around 9781 to add as many CAs as they want to the preferred list. 9782 [Ben Laurie] 9783 9784 *) Experiment with doxygen documentation. Currently only partially applied to 9785 ssl/ssl_lib.c. 9786 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 9787 openssl.doxy as the configuration file. 9788 [Ben Laurie] 9789 9790 *) Get rid of remaining C++-style comments which strict C compilers hate. 9791 [Ralf S. Engelschall, pointed out by Carlos Amengual] 9792 9793 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 9794 compiled in by default: it has problems with large keys. 9795 [Steve Henson] 9796 9797 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 9798 DH private keys and/or callback functions which directly correspond to 9799 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 9800 is needed for applications which have to configure certificates on a 9801 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 9802 (e.g. s_server). 9803 For the RSA certificate situation is makes no difference, but 9804 for the DSA certificate situation this fixes the "no shared cipher" 9805 problem where the OpenSSL cipher selection procedure failed because the 9806 temporary keys were not overtaken from the context and the API provided 9807 no way to reconfigure them. 9808 The new functions now let applications reconfigure the stuff and they 9809 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 9810 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 9811 non-public-API function ssl_cert_instantiate() is used as a helper 9812 function and also to reduce code redundancy inside ssl_rsa.c. 9813 [Ralf S. Engelschall] 9814 9815 *) Move s_server -dcert and -dkey options out of the undocumented feature 9816 area because they are useful for the DSA situation and should be 9817 recognized by the users. 9818 [Ralf S. Engelschall] 9819 9820 *) Fix the cipher decision scheme for export ciphers: the export bits are 9821 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 9822 SSL_EXP_MASK. So, the original variable has to be used instead of the 9823 already masked variable. 9824 [Richard Levitte <levitte@stacken.kth.se>] 9825 9826 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 9827 [Richard Levitte <levitte@stacken.kth.se>] 9828 9829 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 9830 from `int' to `unsigned int' because it's a length and initialized by 9831 EVP_DigestFinal() which expects an `unsigned int *'. 9832 [Richard Levitte <levitte@stacken.kth.se>] 9833 9834 *) Don't hard-code path to Perl interpreter on shebang line of Configure 9835 script. Instead use the usual Shell->Perl transition trick. 9836 [Ralf S. Engelschall] 9837 9838 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 9839 (in addition to RSA certificates) to match the behaviour of `openssl dsa 9840 -noout -modulus' as it's already the case for `openssl rsa -noout 9841 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 9842 currently the public key is printed (a decision which was already done by 9843 `openssl dsa -modulus' in the past) which serves a similar purpose. 9844 Additionally the NO_RSA no longer completely removes the whole -modulus 9845 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 9846 now, too. 9847 [Ralf S. Engelschall] 9848 9849 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 9850 BIO. See the source (crypto/evp/bio_ok.c) for more info. 9851 [Arne Ansper <arne@ats.cyber.ee>] 9852 9853 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 9854 to be added. Now both 'req' and 'ca' can use new objects defined in the 9855 config file. 9856 [Steve Henson] 9857 9858 *) Add cool BIO that does syslog (or event log on NT). 9859 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 9860 9861 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 9862 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 9863 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 9864 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 9865 [Ben Laurie] 9866 9867 *) Add preliminary config info for new extension code. 9868 [Steve Henson] 9869 9870 *) Make RSA_NO_PADDING really use no padding. 9871 [Ulf Moeller <ulf@fitug.de>] 9872 9873 *) Generate errors when private/public key check is done. 9874 [Ben Laurie] 9875 9876 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 9877 for some CRL extensions and new objects added. 9878 [Steve Henson] 9879 9880 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 9881 key usage extension and fuller support for authority key id. 9882 [Steve Henson] 9883 9884 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 9885 padding method for RSA, which is recommended for new applications in PKCS 9886 #1 v2.0 (RFC 2437, October 1998). 9887 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 9888 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 9889 against Bleichbacher's attack on RSA. 9890 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 9891 Ben Laurie] 9892 9893 *) Updates to the new SSL compression code 9894 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 9895 9896 *) Fix so that the version number in the master secret, when passed 9897 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 9898 (because the server will not accept higher), that the version number 9899 is 0x03,0x01, not 0x03,0x00 9900 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 9901 9902 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 9903 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 9904 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 9905 [Steve Henson] 9906 9907 *) Support for RAW extensions where an arbitrary extension can be 9908 created by including its DER encoding. See apps/openssl.cnf for 9909 an example. 9910 [Steve Henson] 9911 9912 *) Make sure latest Perl versions don't interpret some generated C array 9913 code as Perl array code in the crypto/err/err_genc.pl script. 9914 [Lars Weber <3weber@informatik.uni-hamburg.de>] 9915 9916 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 9917 not many people have the assembler. Various Win32 compilation fixes and 9918 update to the INSTALL.W32 file with (hopefully) more accurate Win32 9919 build instructions. 9920 [Steve Henson] 9921 9922 *) Modify configure script 'Configure' to automatically create crypto/date.h 9923 file under Win32 and also build pem.h from pem.org. New script 9924 util/mkfiles.pl to create the MINFO file on environments that can't do a 9925 'make files': perl util/mkfiles.pl >MINFO should work. 9926 [Steve Henson] 9927 9928 *) Major rework of DES function declarations, in the pursuit of correctness 9929 and purity. As a result, many evil casts evaporated, and some weirdness, 9930 too. You may find this causes warnings in your code. Zapping your evil 9931 casts will probably fix them. Mostly. 9932 [Ben Laurie] 9933 9934 *) Fix for a typo in asn1.h. Bug fix to object creation script 9935 obj_dat.pl. It considered a zero in an object definition to mean 9936 "end of object": none of the objects in objects.h have any zeros 9937 so it wasn't spotted. 9938 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 9939 9940 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 9941 Masking (CBCM). In the absence of test vectors, the best I have been able 9942 to do is check that the decrypt undoes the encrypt, so far. Send me test 9943 vectors if you have them. 9944 [Ben Laurie] 9945 9946 *) Correct calculation of key length for export ciphers (too much space was 9947 allocated for null ciphers). This has not been tested! 9948 [Ben Laurie] 9949 9950 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 9951 message is now correct (it understands "crypto" and "ssl" on its 9952 command line). There is also now an "update" option. This will update 9953 the util/ssleay.num and util/libeay.num files with any new functions. 9954 If you do a: 9955 perl util/mkdef.pl crypto ssl update 9956 it will update them. 9957 [Steve Henson] 9958 9959 *) Overhauled the Perl interface (perl/*): 9960 - ported BN stuff to OpenSSL's different BN library 9961 - made the perl/ source tree CVS-aware 9962 - renamed the package from SSLeay to OpenSSL (the files still contain 9963 their history because I've copied them in the repository) 9964 - removed obsolete files (the test scripts will be replaced 9965 by better Test::Harness variants in the future) 9966 [Ralf S. Engelschall] 9967 9968 *) First cut for a very conservative source tree cleanup: 9969 1. merge various obsolete readme texts into doc/ssleay.txt 9970 where we collect the old documents and readme texts. 9971 2. remove the first part of files where I'm already sure that we no 9972 longer need them because of three reasons: either they are just temporary 9973 files which were left by Eric or they are preserved original files where 9974 I've verified that the diff is also available in the CVS via "cvs diff 9975 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 9976 the crypto/md/ stuff). 9977 [Ralf S. Engelschall] 9978 9979 *) More extension code. Incomplete support for subject and issuer alt 9980 name, issuer and authority key id. Change the i2v function parameters 9981 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 9982 what that's for :-) Fix to ASN1 macro which messed up 9983 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 9984 [Steve Henson] 9985 9986 *) Preliminary support for ENUMERATED type. This is largely copied from the 9987 INTEGER code. 9988 [Steve Henson] 9989 9990 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 9991 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 9992 9993 *) Make sure `make rehash' target really finds the `openssl' program. 9994 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 9995 9996 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 9997 like to hear about it if this slows down other processors. 9998 [Ben Laurie] 9999 10000 *) Add CygWin32 platform information to Configure script. 10001 [Alan Batie <batie@aahz.jf.intel.com>] 10002 10003 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 10004 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 10005 10006 *) New program nseq to manipulate netscape certificate sequences 10007 [Steve Henson] 10008 10009 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 10010 few typos. 10011 [Steve Henson] 10012 10013 *) Fixes to BN code. Previously the default was to define BN_RECURSION 10014 but the BN code had some problems that would cause failures when 10015 doing certificate verification and some other functions. 10016 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10017 10018 *) Add ASN1 and PEM code to support netscape certificate sequences. 10019 [Steve Henson] 10020 10021 *) Add ASN1 and PEM code to support netscape certificate sequences. 10022 [Steve Henson] 10023 10024 *) Add several PKIX and private extended key usage OIDs. 10025 [Steve Henson] 10026 10027 *) Modify the 'ca' program to handle the new extension code. Modify 10028 openssl.cnf for new extension format, add comments. 10029 [Steve Henson] 10030 10031 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 10032 and add a sample to openssl.cnf so req -x509 now adds appropriate 10033 CA extensions. 10034 [Steve Henson] 10035 10036 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 10037 error code, add initial support to X509_print() and x509 application. 10038 [Steve Henson] 10039 10040 *) Takes a deep breath and start addding X509 V3 extension support code. Add 10041 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 10042 stuff is currently isolated and isn't even compiled yet. 10043 [Steve Henson] 10044 10045 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 10046 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 10047 Removed the versions check from X509 routines when loading extensions: 10048 this allows certain broken certificates that don't set the version 10049 properly to be processed. 10050 [Steve Henson] 10051 10052 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 10053 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 10054 can still be regenerated with "make depend". 10055 [Ben Laurie] 10056 10057 *) Spelling mistake in C version of CAST-128. 10058 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 10059 10060 *) Changes to the error generation code. The perl script err-code.pl 10061 now reads in the old error codes and retains the old numbers, only 10062 adding new ones if necessary. It also only changes the .err files if new 10063 codes are added. The makefiles have been modified to only insert errors 10064 when needed (to avoid needlessly modifying header files). This is done 10065 by only inserting errors if the .err file is newer than the auto generated 10066 C file. To rebuild all the error codes from scratch (the old behaviour) 10067 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 10068 or delete all the .err files. 10069 [Steve Henson] 10070 10071 *) CAST-128 was incorrectly implemented for short keys. The C version has 10072 been fixed, but is untested. The assembler versions are also fixed, but 10073 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 10074 to regenerate it if needed. 10075 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 10076 Hagino <itojun@kame.net>] 10077 10078 *) File was opened incorrectly in randfile.c. 10079 [Ulf M�ller <ulf@fitug.de>] 10080 10081 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 10082 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 10083 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 10084 al: it's just almost always a UTCTime. Note this patch adds new error 10085 codes so do a "make errors" if there are problems. 10086 [Steve Henson] 10087 10088 *) Correct Linux 1 recognition in config. 10089 [Ulf M�ller <ulf@fitug.de>] 10090 10091 *) Remove pointless MD5 hash when using DSA keys in ca. 10092 [Anonymous <nobody@replay.com>] 10093 10094 *) Generate an error if given an empty string as a cert directory. Also 10095 generate an error if handed NULL (previously returned 0 to indicate an 10096 error, but didn't set one). 10097 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 10098 10099 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 10100 [Ben Laurie] 10101 10102 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 10103 parameters. This was causing a warning which killed off the Win32 compile. 10104 [Steve Henson] 10105 10106 *) Remove C++ style comments from crypto/bn/bn_local.h. 10107 [Neil Costigan <neil.costigan@celocom.com>] 10108 10109 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 10110 based on a text string, looking up short and long names and finally 10111 "dot" format. The "dot" format stuff didn't work. Added new function 10112 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 10113 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 10114 OID is not part of the table. 10115 [Steve Henson] 10116 10117 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 10118 X509_LOOKUP_by_alias(). 10119 [Ben Laurie] 10120 10121 *) Sort openssl functions by name. 10122 [Ben Laurie] 10123 10124 *) Get the gendsa program working (hopefully) and add it to app list. Remove 10125 encryption from sample DSA keys (in case anyone is interested the password 10126 was "1234"). 10127 [Steve Henson] 10128 10129 *) Make _all_ *_free functions accept a NULL pointer. 10130 [Frans Heymans <fheymans@isaserver.be>] 10131 10132 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 10133 NULL pointers. 10134 [Anonymous <nobody@replay.com>] 10135 10136 *) s_server should send the CAfile as acceptable CAs, not its own cert. 10137 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 10138 10139 *) Don't blow it for numeric -newkey arguments to apps/req. 10140 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 10141 10142 *) Temp key "for export" tests were wrong in s3_srvr.c. 10143 [Anonymous <nobody@replay.com>] 10144 10145 *) Add prototype for temp key callback functions 10146 SSL_CTX_set_tmp_{rsa,dh}_callback(). 10147 [Ben Laurie] 10148 10149 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 10150 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 10151 [Steve Henson] 10152 10153 *) X509_name_add_entry() freed the wrong thing after an error. 10154 [Arne Ansper <arne@ats.cyber.ee>] 10155 10156 *) rsa_eay.c would attempt to free a NULL context. 10157 [Arne Ansper <arne@ats.cyber.ee>] 10158 10159 *) BIO_s_socket() had a broken should_retry() on Windoze. 10160 [Arne Ansper <arne@ats.cyber.ee>] 10161 10162 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 10163 [Arne Ansper <arne@ats.cyber.ee>] 10164 10165 *) Make sure the already existing X509_STORE->depth variable is initialized 10166 in X509_STORE_new(), but document the fact that this variable is still 10167 unused in the certificate verification process. 10168 [Ralf S. Engelschall] 10169 10170 *) Fix the various library and apps files to free up pkeys obtained from 10171 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 10172 [Steve Henson] 10173 10174 *) Fix reference counting in X509_PUBKEY_get(). This makes 10175 demos/maurice/example2.c work, amongst others, probably. 10176 [Steve Henson and Ben Laurie] 10177 10178 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 10179 `openssl' and second, the shortcut symlinks for the `openssl <command>' 10180 are no longer created. This way we have a single and consistent command 10181 line interface `openssl <command>', similar to `cvs <command>'. 10182 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 10183 10184 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 10185 BIT STRING wrapper always have zero unused bits. 10186 [Steve Henson] 10187 10188 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 10189 [Steve Henson] 10190 10191 *) Make the top-level INSTALL documentation easier to understand. 10192 [Paul Sutton] 10193 10194 *) Makefiles updated to exit if an error occurs in a sub-directory 10195 make (including if user presses ^C) [Paul Sutton] 10196 10197 *) Make Montgomery context stuff explicit in RSA data structure. 10198 [Ben Laurie] 10199 10200 *) Fix build order of pem and err to allow for generated pem.h. 10201 [Ben Laurie] 10202 10203 *) Fix renumbering bug in X509_NAME_delete_entry(). 10204 [Ben Laurie] 10205 10206 *) Enhanced the err-ins.pl script so it makes the error library number 10207 global and can add a library name. This is needed for external ASN1 and 10208 other error libraries. 10209 [Steve Henson] 10210 10211 *) Fixed sk_insert which never worked properly. 10212 [Steve Henson] 10213 10214 *) Fix ASN1 macros so they can handle indefinite length construted 10215 EXPLICIT tags. Some non standard certificates use these: they can now 10216 be read in. 10217 [Steve Henson] 10218 10219 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 10220 into a single doc/ssleay.txt bundle. This way the information is still 10221 preserved but no longer messes up this directory. Now it's new room for 10222 the new set of documenation files. 10223 [Ralf S. Engelschall] 10224 10225 *) SETs were incorrectly DER encoded. This was a major pain, because they 10226 shared code with SEQUENCEs, which aren't coded the same. This means that 10227 almost everything to do with SETs or SEQUENCEs has either changed name or 10228 number of arguments. 10229 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 10230 10231 *) Fix test data to work with the above. 10232 [Ben Laurie] 10233 10234 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 10235 was already fixed by Eric for 0.9.1 it seems. 10236 [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 10237 10238 *) Autodetect FreeBSD3. 10239 [Ben Laurie] 10240 10241 *) Fix various bugs in Configure. This affects the following platforms: 10242 nextstep 10243 ncr-scde 10244 unixware-2.0 10245 unixware-2.0-pentium 10246 sco5-cc. 10247 [Ben Laurie] 10248 10249 *) Eliminate generated files from CVS. Reorder tests to regenerate files 10250 before they are needed. 10251 [Ben Laurie] 10252 10253 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 10254 [Ben Laurie] 10255 10256 10257 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 10258 10259 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 10260 changed SSLeay to OpenSSL in version strings. 10261 [Ralf S. Engelschall] 10262 10263 *) Some fixups to the top-level documents. 10264 [Paul Sutton] 10265 10266 *) Fixed the nasty bug where rsaref.h was not found under compile-time 10267 because the symlink to include/ was missing. 10268 [Ralf S. Engelschall] 10269 10270 *) Incorporated the popular no-RSA/DSA-only patches 10271 which allow to compile a RSA-free SSLeay. 10272 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 10273 10274 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 10275 when "ssleay" is still not found. 10276 [Ralf S. Engelschall] 10277 10278 *) Added more platforms to Configure: Cray T3E, HPUX 11, 10279 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 10280 10281 *) Updated the README file. 10282 [Ralf S. Engelschall] 10283 10284 *) Added various .cvsignore files in the CVS repository subdirs 10285 to make a "cvs update" really silent. 10286 [Ralf S. Engelschall] 10287 10288 *) Recompiled the error-definition header files and added 10289 missing symbols to the Win32 linker tables. 10290 [Ralf S. Engelschall] 10291 10292 *) Cleaned up the top-level documents; 10293 o new files: CHANGES and LICENSE 10294 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 10295 o merged COPYRIGHT into LICENSE 10296 o removed obsolete TODO file 10297 o renamed MICROSOFT to INSTALL.W32 10298 [Ralf S. Engelschall] 10299 10300 *) Removed dummy files from the 0.9.1b source tree: 10301 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 10302 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 10303 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 10304 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 10305 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 10306 [Ralf S. Engelschall] 10307 10308 *) Added various platform portability fixes. 10309 [Mark J. Cox] 10310 10311 *) The Genesis of the OpenSSL rpject: 10312 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 10313 Young and Tim J. Hudson created while they were working for C2Net until 10314 summer 1998. 10315 [The OpenSSL Project] 10316 10317 10318 Changes between 0.9.0b and 0.9.1b [not released] 10319 10320 *) Updated a few CA certificates under certs/ 10321 [Eric A. Young] 10322 10323 *) Changed some BIGNUM api stuff. 10324 [Eric A. Young] 10325 10326 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 10327 DGUX x86, Linux Alpha, etc. 10328 [Eric A. Young] 10329 10330 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 10331 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 10332 available). 10333 [Eric A. Young] 10334 10335 *) Add -strparse option to asn1pars program which parses nested 10336 binary structures 10337 [Dr Stephen Henson <shenson@bigfoot.com>] 10338 10339 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 10340 [Eric A. Young] 10341 10342 *) DSA fix for "ca" program. 10343 [Eric A. Young] 10344 10345 *) Added "-genkey" option to "dsaparam" program. 10346 [Eric A. Young] 10347 10348 *) Added RIPE MD160 (rmd160) message digest. 10349 [Eric A. Young] 10350 10351 *) Added -a (all) option to "ssleay version" command. 10352 [Eric A. Young] 10353 10354 *) Added PLATFORM define which is the id given to Configure. 10355 [Eric A. Young] 10356 10357 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 10358 [Eric A. Young] 10359 10360 *) Extended the ASN.1 parser routines. 10361 [Eric A. Young] 10362 10363 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 10364 [Eric A. Young] 10365 10366 *) Added a BN_CTX to the BN library. 10367 [Eric A. Young] 10368 10369 *) Fixed the weak key values in DES library 10370 [Eric A. Young] 10371 10372 *) Changed API in EVP library for cipher aliases. 10373 [Eric A. Young] 10374 10375 *) Added support for RC2/64bit cipher. 10376 [Eric A. Young] 10377 10378 *) Converted the lhash library to the crypto/mem.c functions. 10379 [Eric A. Young] 10380 10381 *) Added more recognized ASN.1 object ids. 10382 [Eric A. Young] 10383 10384 *) Added more RSA padding checks for SSL/TLS. 10385 [Eric A. Young] 10386 10387 *) Added BIO proxy/filter functionality. 10388 [Eric A. Young] 10389 10390 *) Added extra_certs to SSL_CTX which can be used 10391 send extra CA certificates to the client in the CA cert chain sending 10392 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 10393 [Eric A. Young] 10394 10395 *) Now Fortezza is denied in the authentication phase because 10396 this is key exchange mechanism is not supported by SSLeay at all. 10397 [Eric A. Young] 10398 10399 *) Additional PKCS1 checks. 10400 [Eric A. Young] 10401 10402 *) Support the string "TLSv1" for all TLS v1 ciphers. 10403 [Eric A. Young] 10404 10405 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 10406 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 10407 [Eric A. Young] 10408 10409 *) Fixed a few memory leaks. 10410 [Eric A. Young] 10411 10412 *) Fixed various code and comment typos. 10413 [Eric A. Young] 10414 10415 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 10416 bytes sent in the client random. 10417 [Edward Bishop <ebishop@spyglass.com>] 10418 10419